A GLOBALIST RESPONSE

and philosophy, we can and should generalize. For example, we might say that although certain features of nature are complex, for the most part and to a considerable degree, na­ture is simple or is representable as such in a fairly accurate way.

Since Newton and Einstein make “global” rather than “local” claims about simplicity, how do they defend their claims? Newton offers no defense at all in the rules.^[71] Einstein writes that “our experience hitherto justifies us in trusting that nature is the realization of the simplest that is mathematically conceivable.” This suggests a defense of the claim that nature is simple (at least with respect to mathematical simplicity) on inductive grounds: theories that are mathematically simple have generally been successful. Since an even more general inductive argument will appear in sections 5 to 7 when I dis­cuss the epistemological role of simplicity, I will postpone a discussion of it until then. Whatever one's views on its validity, however, neither Newton nor Einstein is in a position to give such an argument. To inductively justify the claim that nature is simple, in the sense of having laws that are uniformly appli­cable, rather than applicable to some parts of nature but not others, Newton would need to use some version of his induc­tive Rule 3. But he justifies Rule 3 by appeal to the claim that “nature is always simple and ever consonant with itself.” For him, an inductive argument for the claim that nature is simple would be a circular argument.

What about Einstein? If he is indeed defending the claim that nature is simple, he cannot do so inductively by appeal to the empirical success of simple theories and the lack of such success in complex theories. His claim is that responsible for everything. These are very bold inductive generalizations, indeed! But they still don't get him to the simplicity of the world, since God could also be responsible for complex phenomena that may (or may not) be subject to complex causes and laws. Newton needs an additional premise that he does not “deduce from the phenomena” viz. since all the works created by God are perfect, they are simple. This is basically the the­ological or metaphysical assumption quoted at the beginning of this note. experience by itself (including presumably observing the success of simple theories) cannot justify a grand conclusion about the world, such as nature is simple. There are other conflicting hypotheses about the world compatible with the success of simple theories. To defend the claim that nature is simple on the empirical grounds that simple theories have been successful would be insufficient for Einstein, since em­pirical grounds alone are insufficient for such a defense. But what other way is there for Einstein?

It might be replied that the claim that nature is simple can be defended in an empirical way that is non-inductive, and non-causal, viz. by appeal to what is called “inference to the best explanation” (IBE). The basic idea, introduced by Harman,^[72] and developed much further by Lipton,^[73] is that the best explanation of the fact that the simple theories we have studied have been empirically more successful than the complex ones (assuming they have) is that nature is simple.

There is a deeper issue here. What are the criteria for the “best explanation”? Both Harman and Lipton appeal to simplicity, among other criteria.^[75] So, focusing just on the criterion of simplicity, which is central to both Harman and Lipton, the question to be raised is this: How is one to infer the (probable) truth of a hypothesis (as IBE theorists do) from the fact that it is the simplest? If an appeal is made to the fact that nature itself is simple, then the response should be: How do you know that? If the answer is that we know this because, using IBE, that is the simplest explanation of the past success of simple theories, then the argument is circular. We are using IBE to defend the claim that nature is simple, where our use of IBE presupposes that nature is simple.

Suppose, then, that the use of IBE is not based on the on­tological assumption that nature is simple. How, then, are we supposed to justify an inference from simplicity of a hypoth­esis or theory to its (probable) truth? Harman and Lipton offer no argument, but just assume it. Can an epistemological argument be constructed? For the answer we must turn to the aforementioned epistemological claim about simplicity.

5.