AN A PRIORIST RESPONSE
Now, you may object, give the a priorist a fair chance. Surely he will have a devastating response. Let me choose my favorite a priorist, Rudolf Carnap. Earlier I claimed that there is a crucial relationship between evidence and a good reason for belief.
For Carnap there is a relationship between evidence and what one is justified in believing. This relationship, which is implied by his famous “requirement of total evidence,” is this:If e is evidence that h, or if e confirms h to degree r, and if e represents one's total (relevant) empirical knowledge, then one is (at least to some extent) justified in believing h on the basis of e, or one is justified in believing h to the degree r on the basis of e.
Now, Carnap’s claim is this: If e is the entire set of empirical propositions you know to be true, then whether you are justified in believing h to the degree r on the basis of e is an a priori question. For Carnap whether you are justified in believing h depends solely on what you know to be true. And if you put together everything you know to be true, then you can determine from this by a priori calculation whether, or to what extent, you are justified in believing some further empirical proposition.
Now putting together everything you know to be true, or even everything relevant, is a pretty tall order. Yet for a Carnapian this tall order must be filled if we are to use a concept of evidence and determine a priori whether, or to what extent, we are justified in believing h on the basis of e. But even if we could fill this tall order, this won’t suffice because Carnap is mistaken in an important respect. What you are justified in believing depends not only on what you know but also on what someone in your position is capable of knowing or ought to know, even if you don’t know it. Suppose that a detective seeing a smoking gun lying next to the victim believes that the victim was shot.
It turns out, however, that there are no bullet wounds in the victim’s body, something that the detective could have, and should have, discovered by looking more carefully. An amateur or scared bystander seeing the smoking gun may be justified in believing the victim was shot. But I don’t think the detective is. It is his job, and part of his training, to examine the scene more closely, even if, for some reason, he fails to realize that someone in his position should examine the situation with more care.An a priorist might agree with this but argue as follows. Using as an example our hasty detective, let us formulate a set of propositions S about the detective’s “situation,” namely, his training, his abilities, his responsibilities, and whatever other factors that affect whether he is justified in believing something on the basis of what he observes. The a priorist can modify an earlier Carnapian claim in this manner. Suppose that e represents a person’s total (relevant) empirical knowledge, and suppose that the person is in a situation S (his abilities, training, responsibilities, etc.). Then whether, or to what extent, e confirms h for such a person depends just on e, h, and S. Moreover, knowing just e, h, and S one can calculate a priori whether, or to what extent, e confirms h. Instead of writing c(h/e) = r, as Carnap does, we can write c (h/e) = r, which means that for a person in a situation S the degree of confirmation of h on e is r. And statements of this form are a priori true (or false).
I don't want to pursue the question of whether such evidential statements are, or can always be made to be, a priori but rather whether such evidential statements are of interest to science. I don't see how they could be. They are too particular for that. They are swimming in specificity. In the detective case the evidential statement would include not only facts the detective observed—the fallen victim, the smoking gun—but also facts about the detective's particular situation: his responsibilities, training, intelligence, and so forth.
Or, reverting to the case of Hertz's experiments with cathode rays in 1883, if we claim, as does the distinguished historian of physics Jed Buchwald (1994), that Hertz was justified in believing that cathode rays are electrically neutral, and we seek to make the justification a priori, then we will need to introduce more facts than simply ones about Hertz's experimental set-up and his results. We will need to include facts about Hertz's particular situation, including his inability, and that of others in 1883, to evacuate more gas from the cathode tube. What's wrong with such a concept of evidence? Why should it be of so little interest to scientists?To begin with, using such a concept, you can't ask a question such as “Were Hertz's experimental results evidence that cathode rays are electrically neutral?” You need to ask whether they were evidence for Hertz, or for Thomson, or for someone in either of their situations, or for someone else. Keep in mind this is not what someone takes to be, or believes is, evidence. It is supposed to be a set of facts that makes a person in a certain situation justified in believing the hypothesis, whether or not such a person realizes this or believes the hypothesis. Now even if scientists were to have some interest in answering this question, I think they are much more interested in answering the unrelativized question “Are (or were) Hertz's results evidence that cathode rays are neutral?” Not evidence for Hertz, or for Thomson, or for anyone else. They don't particularly want to know whether Hertz, or Thomson, or someone else was justified in those beliefs on the basis of Hertz's results. They want to know something more general that transcends a particular person or type of person. (More about this in a moment.)
Second, to answer the Carnapian evidential question and produce an evidential claim that is true, and a priori, and that will justify one's beliefs, one needs to include a lot of facts about the particular situation to which it is relativized.
To know whether Hertz's experimental results were evidence for Hertz or someone in his situation that cathode rays are neutral one needs to know a lot of facts about Hertz and his situation. This is because, on a Carnapian view, one needs to know whether Hertz, or someone in his situation, was justified in believing the hypothesis on the basis of his results. And one needs to know enough about the situation to render such a justifiction a priori. Such facts about an individual's situation are often difficult to ascertain, and scientists are not usually in a position to know them, and hence to produce true a priori evidence claims of the sort in question.By contrast, what I think scientists want from evidence is a good reason for believing something, not a Carnapian justification. This notion of “good reason” is not relativized to the circumstances of a particular or type of person. A change in barometric pressure is a good reason to believe in a change in the weather. Your owning 950 of the 1,000 lottery tickets sold is a good reason to believe one of your tickets will win. This is objective—it does not depend on what you or anyone else believes. And it is not, or does not need to be, relativized to anyone. One does not need to know facts about Hertz's training, his intelligence, or his ability, to determine whether the results he obtained in his experiments provide a good reason for believing that cathode rays are electrically neutral. This is something J. J. Thomson determined fourteen years after Hertz's experiments without investigating Hertz's particular epistemic situation in 1883. Moreover, that Hertz's results do not provide a good reason for believing that cathode rays are neutral is a determination that Thomson made empirically, not by a priori calculation.
Finally, suppose one tries to pack enough information into a good reason to believe h to make the fact that it is such a reason a priori. For example, one might try to include enough facts to transform it into a valid deductive argument, or into a valid a priori Carnapian evidential claim. But why do so? Is this the only, or the best, way to discover whether something is a good reason to believe? I don't know whether taking these pills for a headache is a good reason to believe the headache will disappear. Surely a good, indeed the most obvious, way to find out is to conduct a simple appropriate experiment: have one group of headache sufferers take the pills and another group take a placebo.
5.