The Defence Of Scientific Realism

The realist turn begins in philosophy of science since the early 1960s. The philosophical positions implicated in the debate about scientific realism turn from the analysis of language onto the epistemic issues. One thesis of scientific realism can be called’epistemic optimism’, that is, scientific theories describe the real picture of the world, and can be seen as (approximately) true. As science develops, we would be more and more close to the final truth.The main argument for epistemic optimism is the No-Miracle Argument which is also the main positive argument for scientific realism. The standard formulation of No-Miracle Argument is Putnam’s slogan:’Scientific realism is the only philosophy that does not make the success of science a miracle’. What appear to be variants of No-Miracle Argument had been put forward long by J.J.C.Smart and others before Putnam’s slogan appears, yet Smart talks about’no cosmic coincidence’ instead of ’no miracle’. Lots of prominent philosophers, such as Richard Boyd, also use No-Miracle Argument as the main strategy to make defence for scientific realism after Putnam. No-Miracle Argument has been developed to’explanationist defence of realism’ by Boyd. Boyd suggests that a realist epistemology of science should be thoroughly naturalistic, that is a realist epistemology of science should employ no methods other than those used by scientists.However, there are several important critics on the No-Miracle Argument. As well known, one of them is the’Pessimistic Meta-induction’ proposed by Larry Laudan. Laudan suggests that it is unwarranted for scientific realists to think the success of scientific theories as the evidence for their truth. He makes a list, which gives theories that were once empirically successful yet were false, such as the catastrophist geology and the humoral theory of medicine, ect. Laudan suggests, we can induce from this fact that our current best theories may be proved to false in the future. So, it is unreasonable to believe in the truth of scientific theories only on their empirical success.’Pessimistic Meta-induction’ can also be called’Pessimistic Induction’. It is the most pertinacious enemy for No-Miracle Argument. Some philosophers once pointed out, the debate about scientific realism in last few decades can be seen as swinging between No-Miracle Argument and Pessimistic Meta-induction. Yet the insight in Pessimistic Meta-induction is that we should never accept scientific theories uncritically. It forces scientific realism to confine their opinions so as to be more moderate.Nevertheless, we can find Pessimistic Meta-induction to be problematic. In 2001, a paper titled as Why the Pessimistic Induction is Fallacy was published on Sythese. The work was done by Peter J.Lewis. Lewis says that, Pessimistic Induction makes a mistake which can be called’False Positives Fallacy’. There are lots of once successful but false theories in the history of science, and Pessimistic Induction infers from this fact to say that empirical success can be a reliable means to judge on truth. But it is unjustified, for we evaluate the reliability of our means on their false positives rate which is the proportion of false positives in all the positive cases. Lower the false positives rate is, more reliable our means is. So we can’t infer only from the large amount of the past successful but false theories to a large false positives rate, for the rate should be decided by another element which is the number of the past successful theories.Moreover, in his work named’baseball, pessimistic inductions and the turnover fallacy’ published on Analysis 62, March Lange suggests Pessimistic Induction to be a turnover fallacy. We can’t infer from the number of the past successful but false theories to the instability of all the scientific theories. For all these false theories can be successors of a few scientific theories, that is to say, most theories in science may be stable, only a few are not. A small quantity of unstable theories can lead to the emergence of large amount of successors, only if the successors are proved to wrong soon after and the new ones come out frequently. Most theories in science can be seen as stable, and not to be turned over.Besides the two fallacies above, we can do more to rebut the Pessimistic Meta-induction. Scientific realists can point out that the empirically successful theories should be confined not to all the once accepted ones. We should notice that many theories given in Laudan’s list can be seen as genuinely successful theories. Empirical success should include both the explanatory success and the predictive success, especially the novel prediction. Only the theories which made a novel prediction can be seen as genuinely successful. Hence, we can powerfully undermine Laudan’s induction. Scientific realists can also suggest that not all the constituents in the past successful theories are true. The empirically successful theories can have both true constituents and false constituents. They are successful precisely because of the true constituents, and abandoned precisely because they have false constituents. Though they are proved to be false, they can be seen as approximately true.Structural realism is such a strategy attempt to show scientific theories have true constituents and to block Pessimistic Meta-induction. Unlike scientific realists, structural realists suggest, only the part of scientific theory which describes the structure of the world can be seen as true, the other part about the nature of the world is false. Since John Worrall’s work Structural Realism:the Best of Both Worlds in 1989, structural realism gets warm discussions. All these discussions generate three versions of structural realism:epistemic structural realism, ontic structural realism and synthetic structural realism. All these three versions have difficulties in answering the following questions:how to distinguish the structure of world and the nature of world? How to part the structural knowledge in scientific theories from the ontic knowledge? What can guarantee the epistemic priority of the structural knowledge? And each of the three versions has special difficulties itself. For instance, Worrall, the main advocate of epistemic structural realism, once took mathematic equations in scientific theories as the representation of the structure of world, hence got serious critics. Cao Tianyu criticizes that pure mathematic equations without physical interpretation have no sense. Then, Worrall and others appeal to Ramsey Sentence to identify the structural knowledge in scientific theories. I will argue that, Ramsey Sentence gives no help here. I will point out the difficulties they met, and conclude that all of them fail as the attempt to give a rational response to Pessimistic Meta-induction. Without enough time and place, I will focus on epistemic structural realism and ontic structural realism. Scientific realists usually appeal to theory of reference to make defence for approximate truth of theories. Once we have a good theory of reference, we can prove that substitute theory refer to same thing as the theory replaced. Then we can legitimately say that, there is the continuity of ontology in the shift of theories. The theory replaced can be thought as approximately true, because part of their description of the unobservable entity is correct. However, before turning to reference, we have to examine a strategy which uses the Quine-Davison holism to show the past theories can be approximately true. It is been proposed by Jack Ritchie in his work Structural Realism and Davison in 2008. Jack Ritchie suggests, we can interpret the theory replaced in the framework of the substitute theory, to show that the constituents responsible for empirical success of the old theory were retained in the new theory. Then from the view of the new theory, the old theory can be seen as approximately true. This strategy depends on Quine-Davison holism of meaning, especially the inscrutability of reference thesis. I will point out the five objections against the inscrutability of reference thesis, and conclude that the inscrutability of reference thesis is untenable, so the holism of meaning based on it. In turn, any strategy based on the holism of meaning to defence for approximate truth of scientific theories can’t work.Then we can turn to the theory of reference for approximate truth. The question here is how to give a satisfactory theory of reference to account for the stability of theoretic terms? That is, is there any theory of reference can be used to account for the community of ontology in science? If the answer is positive, what it is? Surely, whether the theoretic terms have reference or not don’t depend on the theory of reference, but a satisfactory theory of reference can help us to judge the existence of a theoretic term, to judge whether the theoretic terms in the substitute theory and the theory replaced can have same reference. I will point out the difficulties met by two traditional theories of reference, and advocate the two-dimensional multi-level theory which was proposed by David Braddon-Mitchell in 2005. The original theory given by Mitchell met some difficulties, and I will meliorate it in some aspects to avoid the difficulties. Then I make the conclusion that the theory meliorated can account for the stability of theoretic terms, and tell us when we can see the theory as approximate true. If we have the satisfactory theory of reference as tool, to show that there is a continuity of ontology in the change of scientific theories, is just a matter of practice.As far as it goes, we can fight back on Pessimistic Meta-induction. Nevertheless, there is another important critic on No-Miracle Argument which says that No-Miracle Argument is viciously circular and question-begging. This critic was mainly proposed by Arthur Fine. Fine suggests, if we accept No-Miracle Argument as valid, then its conclusion——scientific realism——would be true. What scientific realism says is that scientific theories are (approximately) true, if it is true, then we should admit that the methodology in science is reliable, that is, the abduction (inference to the best explanation) is reliable. But the No-Miracle Argument is an instance of abduction, and its validity presupposes the reliability of abduction. Now, we can see that No-Miracle Argument presupposes what its conclusion would imply. I will argue that, there are two kind of circular in argument:the circular of premise and the circular of rule. No-Miracle Argument is an instance of the circular of rule. However, the circular of rule is not the viciously circular, while the circular of premise is. I will make a further justification that the rule of inference in No-Miracle Argument, abduction, is reliable. The opponents of scientific realism would recede to say, scientific realism is not the best explanation for the empirical success of theories. If they are right, then one premise of No-Miracle Argument is wrong, so its conclusion——scientific realism is unwarranted. I will compare the explanation provided by scientific realism with the one provided by in instrumentalism at first, then with the Darwin explanation provided by van Fraassen, and conclude that scientific realism is the best explanation for the empirical success of scientific theory.Moreover, there is a third important critic on No-Miracle Argument:the Underdetermination of Theory by Evidence. It is used by many opponents of scientific realism to criticize on it, such as Pierre Duhem, Willard Van Orman Quine, Hilary Putnam, van Fraassen, and so on. Even more, van Fraassen established his constructive empiricism on the base of the Underdetermination of Theory by Evidence. What the Underdetermination of Theory by Evidence says is what we get from our experience can’t guarantee the truth of scientific theory, for there are two or more theories which are empirically equivalent. If two theories which imply incompatible content about ontology are empirically equivalent, then we can’t distinguish those using epistemic means, and we have no reason to accept one as true, not the other. We also have no reason to see one as better than the other. So, van Fraassen suggests, we can only accept the theory as empirically adequate, not as true. The Underdetermination of Theory by Evidence is based on two premises:the empirical equivalent thesis and the entailment thesis. I will rebut these two premises respectively. I will point out, even if there are empirically equivalent theories, it is local rather than global, so can’t be used to generally deny the rationality of belief in the truth of scientific theories. To be an evidence for some theory for certain observable phenomena is not equal to entail it from the theory. So, even if two theories are empirically equivalent, there is no sense to say they are confirmed equally by evidence.In sum, I will examine the three important critics on No-Miracle Argument in detail, and refute them respectively, to make defence for scientific realism.