Evidence and Inquiry: Toward Reconstruction in Epistemology.
Susan Haack's Evidence and Inquiry: Towards Reconstruction in Epistemology is a wide-ranging, lively, and provocative contribution to recent epistemology. The book is primarily a defense of Haack's "foundherentist" theory of epistemic justification. Foundherentism is a mixture of foundationalism and coherentism. Haack also defends a modest form of naturalism, claiming only that empirical results can contribute to the resolution of some traditional philosophical questions. Along the way, she subjects a number of prominent contemporary epistemological views to detailed criticism.
In the first chapter Haack describes the controversy between foundationalists and coherentists and then provides a preliminary account of her foundherentist compromise. Her central criticism of foundationalism is that it incorrectly requires that basic beliefs both justify all other justified beliefs and be justified by experience without the aid of other beliefs. Her central criticism of coherentism is that it mistakenly limits the justifiers of beliefs to other beliefs and excludes experiences from consideration. She then turns, in chapter two, to the version of foundationalism defended by C. I. Lewis and in chapter three to the versions of coherentism defended by Laurence Bonjour and Donald Davidson. She contends that their arguments fail to support their own views and that, when developed properly, they really lead to foundherentism. In chapter four Haack presents a more detailed discussion of foundherentism, emphasizing a crossword puzzle analogy according to which there are two sorts of evidence for the identity of the each entry in the puzzle: its clue (comparable to experiential input) and its fit with surrounding entries (comparable to coherence). Thus, foundherentism makes room for both experience and coherence as determinants of epistemic justification for each belief, thereby retaining what's good and overcoming the defects of its traditional rivals.
In chapters five through nine Haack criticizes the views of a variety of recent epistemologists. Chapter five takes on Karl Popper's deductivism and anti-psychologism, and a recent defence of it by J. W. N. Watkins. In chapter six she distinguishes several versions of naturalism in epistemology, criticizes the revolutionary form of naturalism Quine supports, and defends a weaker version of naturalism. Chapter seven is devoted to a critical discussion of Alvin Goldman's various versions of reliabilism and of his attempts to link psychology and epistemology more closely than Haack thinks appropriate. In chapters eight and nine Haack responds to recent attacks on epistemology by Steven Stich, Patricia and Paul Churchland, and Richard Rorty. Finally, in chapter ten she returns to foundherentism, arguing that if any criteria of justification are truth-indicative, the criteria put forth by foundherentism are.
Haack takes up many of the views and issues prominent in contemporary epistemology. She has something interesting and sensible to say about all these topics. The book is full of arguments that merit careful consideration. Haack does not shy away from rhetorical combat with her opponents. For example, she refers to "the wool that Rorty is trying to pull over our eyes" (p. 189), she describes some of Stich's arguments as "remarkably feeble" (p. 192) and she suggests that some of the appeal of epistemological naturalism comes from the prestige and excitement Al and cognitive psychology enjoy and says she will "resist the temptation" to describe this as "opportunistic naturalism" (p. 157). This makes for lively reading and I found it refreshing to read her clear-headed efforts to refute the arguments of the anti-epistemology crowd.
Although I am in general agreement with Haack about many of the general themes of her book, I found some of her arguments a bit too quick and thus less than fully convincing.
I will try to illustrate in more detail her style of argument and her theory and to explain some of my reservations by considering two of her critical points and one aspect of her defense of foundherentism.
Haack's critical points may provide the basis for significant criticisms, but I think that she does not develop some of them in enough detail to make her case. In her discussion of Goldman's reliabilism Haack traces Goldman's views through their changes over the years. An early version held that a belief is justified if and only if it results from an unconditionally reliable belief-independent belief-forming process or from a conditionally reliable belief-dependent belief-forming process which is operating on justified beliefs. In his original paper Goldman recognized the possibility of reliably produced beliefs which are not justified because they are defeated. He added a condition implying that a belief that p meeting the original conditions is not justified if the believer has available a reliable process which would have resulted in not believing that p. In other words, a belief is not justified if there is a reliable defeater process for it. As an example of such a process, Goldman mentions calling to mind and assessing previously acquired evidence.
Haack objects that this added condition "would require an explication of the notion of evidence and of criteria for the appraisal of its worth. But if we had that, surely, it could stand by itself as a theory of justification ... " (p. 144). 1 think that this objection fails. It's true that a good explication of the notion of evidence and a set of criteria for appraising the worth of evidence would constitute a theory of justification and render reliabilism pointless (or worse). But, whatever the merits of Goldman's proposal, I don't think it requires a theory of evidence of the sort Haack mentions. His proposal is that a belief produced by a reliable process is not justified if it is defeated. It is defeated if there is available to the believer a reliable process that would lead from other beliefs or experiences he has (his evidence) to the denial or withholding of the original belief. The theory is thus spelled out in terms of beliefs, experiences, processes, and the reliability of processes. While I don't endorse any such view, I don't see that it requires the full theory of evidence Haack mentions and thus I think that it escapes this particular criticism.
Part of Haack's criticism of foundationalism depends on exactly what foundationalism amounts to. Haack says that "a theory qualifies as foundationalist which subscribes to the theses:
(FD 1) Some justified beliefs are basic: a basic belief is justified independently of the support of any other belief,
(FD2) All other justified beliefs are derived; a derived belief is justified via the support, direct or indirect, of a basic belief or beliefs." (p. 14) Haack acknowledges that there are many versions of foundationalism, including weak foundationalism which holds that basic beliefs are justified merely prima facie but defeasibly, to some degree but not completely, by something other than a belief" (p. 3 1). She contends that foundationalism falls to the "up and back all the way down argument". This argument holds that even the foundationalist's allegedly basic beliefs can get some justification from other beliefs. For example, a person, A, might believe, on the basis of perceptual experience, that there is a dog present. Haack writes:
would not A be more justified, or more securely justified, in believing
that there is a dog before him if he also justifiedly believes that
his eyes are working normally, that he is not under the influence of
post-hypnotic suggestion, that there are no very lifelike toy dogs
around, etc., etc.) Surely he would. But the weak foundationalist
cannot allow this, for his story is that basic beliefs get their justification
exclusively from something other than the support of further beliefs;
to allow that they get some justification from experience and
some from the support of other beliefs would violate the one-directional
character of justification, on which, quafoundationalist, he insists.
And if this possibility were allowed, weak experientialist
foundationalism would be transmuted into a form of foundherentism.(pp.31-2)
I find this argument less than compelling. First, it is not clear that (FDI) and (FD2) limit foundationalism in the way Haack thinks they do. She says that weak foundationalists hold that basic beliefs are prima facie but not completely justified by things other than beliefs, namely experiences. So, (FD 1) should not be read to imply that basic beliefs are completely justified independently of experience. Thus, experiences may provide only some degree of justification, but not enough for complete (knowledge level) justification. This apparently leaves open the possibility that basic beliefs get additional support from other beliefs. The "other justified beliefs" mentioned in (FD2) would then be beliefs that don't get any justification in the basic way. Thus, it seems that a weak foundationalist might endorse three theses: (i) there are basic beliefs which get justified, but not fully justified, directly by experience; (ii) these basic beliefs can be made fully justified by support from other beliefs; and (iii) there are non-basic beliefs that get all their justification from support from other beliefs. Given the way Haack interprets (FDI), a philosopher holding this view seems to accept both (FDI) and (FD2), and thus counts as a foundationalist. This sort of foundationalist can accept Haack's claim that the person who believes that there is a dog before him might be better justified in that belief if he had various collateral supporting beliefs.
Haack apparently thinks that a foundationalist cannot accept claim (ii) above, since, she says, foundationalists hold that "basic beliefs get their justification exclusively from something other than the support of further beliefs." Perhaps she takes (FD2) to imply that any belief that gets any support from another belief is derived and therefore not basic. It is surely true that many foundationalists have accepted this constraint. But not all epistemologists who call themselves foundationalists accept it. (See, for example, Mark Pastin's "Modest Foundationalism and Self-Warrant," in Pappas, George S. and Swain, Marshall, eds., Essays on Knowledge and Justification, Ithaca: Cornell University Press, 1978 and Paul Moser's Knowledge and Evidence, Cambridge: Cambridge University Press, 1989, pp. 77-111.) These foundationalists take the distinguishing feature of foundationalism to be acknowledgment that things other than beliefs can be sources of justification, something typically denied by coherentists. Haack, in contrast, takes a defining feature of foundationalism to be the view that basic beliefs get all their justification from experience. This makes the weak foundationalist view she describes very peculiar, since it holds that (at least some) basic beliefs get only partial justification from experience, that they can't get justification from anywhere else, and that they are fully justified.
Even foundationalists who say that a basic belief must get all its justification from experiences may be able to escape Haack's objection. They might conclude from her argument that when a belief such as the belief that there is a dog present gets some support from other beliefs it is not a basic belief. When it, or any other belief, does get all its support from experience, then it is a basic belief. Whether it is basic depends upon how it is supported. Moreover, such foundationalists might contend that an external world belief, such as the one Haack uses in her example, is typically not basic. Whether Haack could run her argument when applied to beliefs about current experiences that foundationalists often regard as basic is far from clear, since it is less obvious that they do get collateral support from other beliefs. Of course, against versions of foundationalism that appeal these sorts of beliefs as basic, Haack is likely to object that they fail to provide adequate support for the rest of our beliefs. She takes up some of the issues related to this in her discussion in chapter two of C. I. Lewis's version of foundationalism.
Finally, Haack says in presenting her example that the belief that there is a dog present would be better justified if the person also justifiably believes that his eyes are working normally, etc. Here the argument turns on difficult matters about exactly what it is for a belief to be justified "independently of the support of any other belief". A person having an experience of a dog might have other simultaneous experiences that justify the belief that his eyes are working normally. And these additional experiences, rather than anything about other beliefs he has, may be what bolsters the justification of his belief that there is a dog present. To show that the belief that there is a dog present is supported by other beliefs Haack would have to show that the added justification comes from believing additional propositions, rather than from the experiences that justify such beliefs. If the added justification does come from these experiences rather than beliefs, then the example poses no threat to the claim that the belief that there is a dog present is justified entirely independently of the support of any other belief. These points lead me to think that foundationalism has not quite been put to rest yet.
As for Haack's defense of foundherentism, there is much that I find welcome and appealing, including its implication that epistemic justification is ultimately a matter of the quality of one's evidence. I can only give the outlines of Haack's theory here. She says that how good one's evidence for p is depends on how "favorable" it is with respect to p, how "secure" it is with respect to p, and how "comprehensive" it is with respect to p (p. 82). A body of evidence, E, is conclusively favorable toward a proposition,p, provided the conjunction of E and p is consistent while the conjunction of E and not-p is not consistent (pp. 82-3). Where one's evidence is less than conclusive, Haack's fundamental claim is that
E is supportive to some degree with respect to p just in case the addition
of p to it improves its explanatory integration; E is the more
supportive with respect to p the more the addition of p to it improves
its explanatory integration more than the more [sic] its competitors
do. (p. 83)
The security of a body of evidence relative to p is determined by how well supported that evidence is independent of any support it might get from p (p. 86). Haack gives the comprehensiveness condition less attention. She says that it is somewhat analogous to a total evidence requirement and that it is most clearly violated in cases in which one fails to take into account some relevant evidence, as for example by failing to take a closer look at something (p. 87). In a sense, this seems to make justification for a proposition depend not just upon the character of one's evidence, but also upon some features of one's investigation into the proposition.
One puzzling case for this view concerns the logical consequences of one's evidence. Suppose that one has a consistent body of very secure evidence that entails another proposition by means a complex and unnoticed derivation. This evidence is conclusively supportive, secure, and (I think) comprehensive evidence for p. If belief in that evidence causes one to believe p, then Haack's theory seems to imply that the belief in p is well justified. This strikes me as a mistake, given that the actual logical connection between the evidence and the conclusion is not appreciated.
Of greater interest, I think, is the notion of inconclusive evidential support. Here, the key notion is explanatory integration. I would like to have seen more discussion of this idea. Haack says that it has some of the appeal of inference to the best explanation and of explanatory coherence (p. 84). She acknowledges toward the end of the book that there is some vagueness" in her account (p. 217), but it seems to me that the problems here are potentially quite severe. One wants to know, for example, how "objective" a notion it is. That is, can the addition of a proposition to a body of evidence improve its explanatory integration whether or not the believer has any appreciation of this fact? If so, then her theory may have the implausible result that some complex conjunction of propositions, not well understood by a person, is justified for the person because it in fact enormously increases the explanatory integration of her beliefs. If, on the other hand, it is a more subjective notion, her theory may render justified bizarre theories that provide their defenders with feelings of explanatory completeness. I'm sure Haack favors a more objectivist reading and I do not claim that Haack's theory cannot be made to work. What's unfortunate is that the notion of explanatory integration does not receive more discussion in her book so that we could better assess how it deals with this and other questions.
The reservations I've expressed are about details of Haack's arguments. In spite of them, I think that Evidence and Inquiry is a valuable contribution to contemporary epistemology. It deserves the attention of epistemologists.