If we knew the probabilities of all things, good thinking would be apt calculation, and the proper subject of theories of good reasoning would be confined to the artful and efficient assembly of judgments in accordance with the calculus of probabilities. In fact, we know the probabilities of very few things, and to understand inductive competence we must construct theories that will explain how preferences among hypotheses may be established in the absence of such knowledge. More than that, we must also understand how the procedures for establishing preferences in the absence of knowledge of probabilities combine with fragmentary knowledge...

What is the main epistemic problem concerning science? I take it that it is the explication of how we compare and evaluate theories, as a basis either for theory acceptance or for practical action. This comparison is clearly a comparison in the light of the available evidence—whateverthatmeans.¹ My aim in this paper is very modest: to explore the logic of evidential support in the theory of Clark Glymour.² The question of how, and to what extent, overall theory comparison is to be related to the data from relevant tests, I broach in another paper.³

First I shall...

Clark Glymour has persuasively outlined the advantages of an account of the confirmation of scientific hypotheses in terms of what he calls a bootstrap strategy. The stragey relates sentences in an observational vocabulary to hypotheses in a theoretical vocabulary as follows: Auxiliary hypotheses are found that, when conjoined with the observation-sentences in question, entail in a certain way theoretical sentences that, in turn, confirm the hypotheses in some standard way. When the strategy succeeds, the hypotheses are confirmed by the observation-sentences relative to the auxiliary hypotheses (or any theory that contains them).¹

Despite its many virtues, though, the account of...

The singular feature of Clark Glymour's "bootstrap" account of evidence, that which primarily distinguishes it from both Bayesianism and the hypothetico-deductive (henceforth, H-D) method, is that a whole theory is not necessarily confirmed just because one of its parts is: neither evidence p, nor even the discovery of p, is sufficient to confirm the conjunction pAq. This is the intended result of his requirement that confirmation of a hypothesis involve derivations of instances of every quantity in it. For example, a set of observations might support

${\rm{y}}\;{\rm{ = }}\;{{\rm{x}}^2}\; + \;{\rm{z}}$

but not the conjunction

${\rm{y}}\; = \;{{\rm{x}}^2}\; + \;{\rm{z}}$

and

${\rm{x}}\; = \;{\rm{s}} \cdot {\rm{t}}$

and, therefore, not

${\rm{y}} = {({\rm{s}} \cdot {\rm{t)}}^2} + {\rm{z}}$

because values...

In the third chapter of his bookTheory and Evidence, Clark Glymour explains why he is not a Bayesian. I shall attempt to show, on the contrary, that he is a Bayesian, more so than many who march under that banner.

The central problem his book addresses is to explain how findings in one (observational) vocabulary can evidence propositions stated in a different (theoretical) vocabulary. The solution offered is that a hypothesis is confirmed with respect to a theory by deducing instances of that hypothesis from the evidence and assumptions of the theory, where these assumptions may include the very...

The Bayesian framework is intended, at least in part, as a formalization and systematization of the sorts of reasoning that we all carry on at an intuitive level. One of the most attractive features of the Bayesian approach is the apparent ease and elegance with which it can deal with typical strategies for the confirmation of hypotheses in science. Using the apparatus of the mathematical theory of probability, the Bayesian can show how the acquisition of evidence can result in increased confidence in hypotheses, in accord with our best intuitions. Despite the obvious attractiveness of the Bayesian account of confirmation,...

Well, I’m one, for example. But not according to Clark Glymour (1980, pp. 68-69) and some other definers of Bayesianism and personalism, such as Ian Hacking (1967, p. 314) and Isaac Levi (1980, p. xiv). Thus it behooves me to give an explicit account of the species of Bayesianism I espouse (sections 1 and 2) before adding my bit (section 3, with lots of help from my friends) to Daniel Garber’s treatment in this volume of the problem of new explanation of common knowledge: the so-called problem of old evidence.

With Clark Glymour, I take there to be identifiable canons...

Consider a model of learning in which we update our probability assignments by conditionalization; i.e., upon learning S, the probability of not-S is set at zero and the probabilities of statements entailing S are increased by a factor of one over the initial probability of S. In such a model, there is a certain peculiar sense in which we lose information every time we learn something. That is, we lose information concerning the initial relative probabilities of statements not entailing S.

The loss makes itself felt in various ways. Suppose that learning is meant to be corrigible. After conditionalizing on...

In Chapter VI ofTheory and Evidence(specifically pages 199-203) and in a subsequent paper, Clark Glymour develops an account of scientific explanation to go with his theory of relevant evidence.¹ Especially significant for me is his use of these ideas in support of his contention that we can have more reason to believe one theory to be true than another even in cases in which the two theories are empirically equivalent. For that contention poses a challenge to the empiricist account of science that I have proposed.² Although I do not conceive of the empiricist-realist debate concerning science as...

Newton consistently asserted that his method was not hypotheticodeductive.

I cannot think it effectuall for determining truth to examine the several! ways by wch Phaenomena may be explained, unless there can be a perfect enumeration of all those ways. You know the proper Method for inquiring after the properties of things is to deduce them from Experiments. And I told you that the Theory wch I propounded was evinced by me, not by inferring tis this because not otherwise, but by deriving it from Experiments concluding positively & directly.

. . . what I shall tell . . . ....

There is supposed to be a problem in the philosophy of science called “realism versus instrumentalism.” In the version with which I am concerned, this supposed problem is whether scientific theories in general are put forward as true, or whether they are put forward as untrue but nonetheless convenient devices for the prediction (and retrodiction) of observable phenomena.

I have argued elsewhere (1979) that this problem is misconceived. Whether a theory is put forward as true or merely as a device depends on various aspects of the theory’s structure and content, and on the nature of the evidence for it....

Philosophers of science concerned with theories and the nature of evidence tend currently to fall into several only partially overlapping groups. One group follows its logical empiricist ancestors at least to the extent of believing that there is a “logic” in the relation between theories and evidence. This logic is now most often embedded in the theory of a rational (scientific) agent. Bayesian agents are currently most popular, but there are notable dissenters from The Bayesian Way such as Henry Kyburg and Isaac Levi. Another group derives its inspiration from the historical criticisms of logical empiricism begun a generation ago...

1. By the deductive model, I mean the deductive model of scientific inquiry. There are plenty of deductive systems around, of course: arithmetic, number theory, set theory, probability theory, even axiomatizations of scientific theories, for example in physics, in biology, and even in psychology. What I have in mind in its simplest form is the good old hypothetico-deductive model: formulate your hypothesis clearly; draw testable implications from it; put those implications to the test; if they fail test, reject the hypothesis; and if they pass, the hypothesis is to that extent confirmed.

Everybody knows that things are a little more...

The repression-aetiology of neurotic disorders is the cornerstone of the psychoanalytic theory of unconscious motivations. Repressions are held to be not only the pathogens of the psychoneuroses but also the motives of dream construction, and the causes of various sorts of bungled actions (“parapraxes”) in which the subject is normally successful (e.g., slips of the tongue or pen). Thus even in the case of “normal” people, Freud saw manifest dream content and various sorts of “slips” as the telltale symptoms of (temporary) minieuroses, engendered by repressions.

Having modified Breuer’s cathartic method of investigationandtherapy, Freud arrived at the purported...

An alternative subtitle to this essay, which my non-Freudian Minnesota colleagues urged upon me, would have been, “Whose mind does the mindreader read?” To motivate discussion of a topic not deemed important by some today, consider the story of the last “Congress” between Freud and Fliess, the rupture of their relationship at Achensee in the summer of 1900—the last time the two men ever met, although an attenuated correspondence continued for a couple of years more. Setting aside the doubtless complex psychodynamics, and the prior indications (from both content and density of correspondence) that the relationship was deteriorating, I...

Most educated persons today, who have been repeatedly exposed since childhood to pictures of the famous “horse series,” are likely to think of it, or of the fossil dinosaurs they have seen in museums, rather than other lines of evidence for the theory of evolution. They differ in this respect from Charles Darwin, who did not consider the fossil data to be supportive of his theory. On the contrary, he viewed the paleontological findings available in his day as perhaps presenting “the gravest objection which can be urged against my theory” (Darwin 1859, p. 280), emphasizing the fossil record’s failure...