Before turning to Popper's “Darwinian model” of science, I want to begin our next class by briefly talking about Goodman’s “New Riddle of Induction”. Let me contextualize this important contribution a bit here so I can lecture to you less on Tuesday. So . . . we have this problem justifying inductive inference. It seems that we cannot show that it is a reliable form of inference (any more than I can show you that I am telling the truth by merely proclaiming “I’m telling the truth!”). That might not give us reason for doubting its reliability — or suspecting that we’d be better off with other inductive methods —, but it’s disquieting nonetheless. At this point, Goodman comes onto the scene, points out that attempts to solve the justificatory problem have a certain air of pathetic desperation (my phrase) about them (p. 61), and proposes to dissolve the problem rather than solve it. He writes:
Come to think of it, what precisely would constitute the justification we seek? If the problem is to explain how we know that certain predictions will turn out to be correct, the sufficient answer is that we don't know any such thing. If the problem is to find some way of distinguishing antecedently between true and false predictions, we are asking for prevision rather than for philosophical explanation. (Goodman 1983, 62)
Here’s where he makes the comparison between justifying induction and justifying deduction (see Foster p. 15 on this general strategy). “Principles of deductive inference are justified,” Goodman says, “by their conformity with accepted deductive practice. Their validity depends upon accordance with the particular deductive inferences we actually make and sanction” (63). In other words: rules of deductive logic are sanctioned only by the fact that they give us the results that we expect from them! Isn’t this circular!? Yes, “but this circle is a virtuous one” (64). So goes the thought. These are deep waters.
But suppose that this is on the right track. . . . The question then becomes What are the rules of inductive inference? Well, this is the problem of describing (rather than justifying) inductive inference. And as we’ve seen, it’s pretty hairy. The Hypothetico-Deductive model faces the “Tacking Problem”, the Instantial Model faces the “Ravens Paradox” . . . and yet these both seem like decent descriptions of how inductive inference operates in science. Scientists very commonly take instances of a generalization as supporting (confirming to at least some degree) the truth of that generalization. Observing that the consequences of our hypotheses are in fact borne out does seem to lend support to those hypotheses. Goodman’s “New Riddle” (see in particular §4 of his chapter) is another, arguably deeper puzzle about the instantial model. But it turns out to be one that has suggested new directions for addressing the justificatory problem in a more robust way — a subject we will return to later in the course once we have a bit more conceptual apparatus built up. . . .
But for the bulk of next week we will consider the perspective of Karl Popper — arguably one of the two most influential philosophers of science, the other being Thomas Kuhn — on two issues. The first issue (for Tuesday) pertains to the philosophical trauma we’ve so far endured on the justificatory problem of induction. Suppose neither the purported solutions nor Goodman’s dissolution satisfy us. What would happen if we just accepted the conclusion? This is Popper’s move. He writes: “My own view is that the various difficulties of inductive logic here sketched are insurmountable” (Popper 1959, 6).
Now shouldn’t this just scuttle science once and for all? Isn’t science up to its neck in induction? If what I claimed before about the triviality of deductive logic is right, wouldn’t this make science a trivial enterprise? Popper’s clever idea is to articulate a deductive model of theory testing. We can never confirm scientific theories (even in the weak sense we’ve been considering). What scientists do is attempt to falsify theories. And as we’ve seen, this is apparently a deductive business. If my hypothesis H implies that I should observe O, then if O is not observed, I know as a matter of deductive logic that H is false. Suppose now that I have a range of hypotheses: H1, H2, H3, . . . . If I falsify all but H1, what am I going to do? Probably pursue H1 a bit more — not in an effort to confirm it, but in more and more stringent attempts to falsify it. If I fail, time after time, Popper says that we should think of this theory as “corroborated” (rather than confirmed). So the model of science resembles natural selection: theories are proposed like mutations; then testing weeds out the less fit theories. We cannot say what remains is true, but can we not place more faith in it?
There are a number of worries about this approach, however. One we’ve already gestured to: the Duhem-Quine problem. Another is articulated in the optional paper by Wesley Salmon: what is it to “corroborate” a theory? Does a well-corroborated theory license any predictions? If the answer is ‘no’, then don’t we still have a problem?
The second issue (for Thursday) is whether Popper’s focus on falsification will allow us to answer a tricky question that we’ve been conveniently avoiding so far: What is science? Is it possible to decisively separate science from non-science or pseudo-science? We’ll continue discussing this issue in the context of important social issues in week 7 when we discuss the evolutionism/creationism/intelligent-design controversy as a case study. In that case, what we see is philosophy of science being played out in the courtroom!
Tuesday (9/27): The Evolutionary Model of Science
• French, pp. 49–59
• Popper, selections from The Logic of Scientific Discovery [PDF] ← Note that this was optional reading from (9/15); it is required for this class, however.
• Salmon, “Rational Prediction” [PDF] *
Questions: (respond to one)
Questions: (respond to one)
- Attempt to explain in your own words how Popper proposes to do science without induction.
- Why does the Duhem-Quine thesis (discussed in French, pp. 47–48) pose a problem for his “Darwinian model” of science.
- Consider Lakatos’s objection (quoted in French, pp. 58–59): “There is no falsification before the emergence of a better theory.” How might Popper respond to this objection?
- Salmon asks of Popper’s account why it should give us a guide to “rational prediction”. What is Salmon’s criticism here? *
Thursday (9/29): Falsification as Demarcation
• Popper, "Conjectures and Refutations"
• Kuhn, "Logic of Discovery of Psychology of Research" [PDF] <— This single file contains both articles.
Questions: (respond to one)
- At first glance, it might seem like a good thing for a theory to have a great deal of “explanatory power” or receive a lot of experimental/observational verification. Why does Popper think that it’s not (necessarily)?
- Suppose that you are a proponent of one of Popper’s pseudoscientific fields. Explain two distinct ways in which you might respond to Popper’s verdict about the value of your chosen field.
- Describe Kuhn’s criticism of Popper’s theory of demarcation.