(Forthcoming in Dialogue: Canadian Philosophical Review)

Strategic Reliabilism and the Replacement Thesis in Epistemology

ANDREI A. BUCKAREFF Marist College ABSTRACT: In their recent book, Epistemology and the Psychology of Human Judgment, Michael Bishop and J.D. Trout have challenged Standard Analytic Epistemology (SAE) in all its guises and have endorsed a version of the “replacement thesis”—proponents of which aim at replacing the standard questions of SAE with psychological questions. In this paper I argue that Bishop and Trout offer an incomplete epistemology that, as formulated, cannot address many of the core issues that motivate interest in epistemological questions to begin with and so is not a fit replacement. RÉSUMÉ: Dans leur ouvrage récent, Epistemology and the Psychology of Human Judgment, Michael Bishop et J.D. Trout contestant toutes les formes du Standard Analytic Epistemology (SAE) et, contrairement, soutiennent une version de la <>. Cette nouvelle version de la thèse de remplacement vise à replacer les questions du SAE avec les questions psychologiques. Dans mon texte, je dispute les conclusions de Bishop et Trout. Je démontre que leur épistémologie est incomplète et ne peut répondre aux questions épistémologiques fondamentales. Michael Bishop and J.D. Trout (2005a and 2005b) have recently challenged Standard Analytic Epistemology (henceforth, SAE) in all its guises.

In this paper, I show that what Bishop and

Trout offer as a replacement is inadequate. I will proceed as follows. First, I will present the core features of Bishop and Trout’s proposed alternative to SAE. Second, I will show that they offer an incomplete epistemology that fails to be a viable alternative to SAE.1 Their Proposal Bishop and Trout endorse a species of epistemic reliabilism that emphasizes making epistemic evaluations of reasoners and reasoning-strategies. On their view, “Epistemic excellence involves the efficient allocation of cognitive resources to robustly reliable reasoning strategies applied to significant problems” (2005a, p. 71).

Our normative concerns in

1

epistemology should shift from assessing belief tokens for epistemic justification and towards assessing the voluntary processes that generate beliefs for epistemic excellence (2005a, pp. 18285).2 They note that while the “fundamental aim of SAE is to deliver an account of epistemic justification or knowledge…. The fundamental aim of our approach to epistemology is to provide an account of reasoning excellence” (Bishop and Trout 2005a, p. 16).3 They confess that this is a deep difference between themselves and practitioners of SAE (including other reliabilists) (2005a, p. 16). In this respect, the variety of reliabilism they endorse, “Strategic Reliabilism,” differs radically from traditional versions of reliabilism that focus on providing a theory of epistemic justification and knowledge. 4 Bishop and Trout offer a theory of epistemic rationality, but it is best regarded as a theory of procedural or methodological rationality—i.e., the emphasis is upon diachronic evaluations of processes (such as reasoning) rather than evaluations of the end state (in this case, individual belief tokens). But they point out that since employing reasoning strategies typically results in the production of belief tokens, applying a theory of reasoning excellence would normally recommend a belief token, albeit “at one remove” (2005a, p. 17). Bishop and Trout endorse a version of the “replacement thesis”—proponents of which aim at replacing the standard questions of SAE with psychological questions (Quine 1994; Kornblith 1994, pp. 3-7). Their investigation into reasoning strategies that promote epistemic excellence is dominated by considering data from empirical research on reasoning. And a significant proper part of doing epistemology according to them is constituted by Ameliorative Psychology. “Ameliorative Psychology” refers to “the branches of psychology that provide normative recommendations” for how best to reach judgments (Bishop and Trout 2005a, p. 3). Ameliorative Psychology is “the science of applied epistemology” and “Much of the point of

2

[Ameliorative Psychology] is to provide advice that will help people reason better about the world” (Bishop and Trout 2005a, p. 17). Ameliorative Psychology recommends using Statistical Prediction Rules in theoretical reasoning. These are “simple, formal rules that have been shown to be at least as reliable, and typically more reliable, than the predictions of human experts on a wide variety of problems” (Bishop and Trout 2005a, p. 3). There are three main features of the epistemological framework that guides the prescriptions of Ameliorative Psychology and the use of Statistical Prediction Rules that Bishop and Trout discuss (2005a, p. 70).

First, Ameliorative Psychology “identifies successful

reasoning strategies in terms of their reliability” (Bishop and Trout 2005a, p. 55). A reasoning strategy should be robustly reliable. A robustly reliable reasoning strategy is reliable on a wide range of problems (Bishop and Trout 2005a, p. 55).5 Second, simpler reasoning strategies that minimize costs and maximize epistemic benefits are desirable. Finally, the Statistical Prediction Rules should tackle significant problems. An Incomplete Epistemology Bishop and Trout offer an incomplete epistemology. An especially glaring lacuna in their theory is evinced by their failure to elaborate on the connection between epistemic excellence and the rationality of belief tokens. A theory of epistemic excellence should require more than a theory of reasoning excellence. This is so especially if one intends (as they do) to replace SAE.

While

they address various objections to their project that may come from the direction of SAE, they owe us a fuller account of the relationship between reliable reasoning strategies and rational beliefs. They may object that this is not part of their project and that such complaints represent little more than an unreflective commitment to the concerns and projects of SAE. As they note, “Strategic Reliabilism is not a theory of justification” (2005a, p. 116). In fact, Bishop and Trout

3

repeatedly make it clear that they “do not believe that constructing theories that render judgments of justification on belief tokens [to be] a fruitful endeavor in epistemology” (2005a, p. 116). But we need a theory that can distinguish between belief tokens whose epistemic status is positive (whether we use the language of justification or not) and those belief tokens with a negative epistemic status that are not the product of explicit reasoning strategies. If Strategic Reliabilism fails to deliver this, then it does not afford us the means to determine the epistemic status of many belief tokens. And providing the means for making such distinctions should be a proper part of the task of theorizing about epistemic excellence.

If these are tasks worthy of

undertaking by philosophers, then something must be said about beliefs not formed as the consequence of reasoning processes—e.g., those acquired directly on the basis of testimony, perceptual experience, introspection, etc. Consider the following examples that Strategic Reliabilism as articulated by Bishop and Trout lacks the resources to address. Jane forms the belief that she is hungry on the basis of her being aware of her experiencing hunger pangs. Juan sees Jane ordering a bagel and forms the belief that Jane is hungry on the basis of her overt intentional behavior. Consider another example. Shusaku witnesses a car crash first-hand and sees Maria’s husband, Ivan, die. Maria learns of the crash from Shusaku. But Shusaku decides to lie and so he tells Maria that Ivan is okay. Jane and Shusaku are clearly in a better epistemic position than either Juan or Maria. Consider two different ways of accounting for the differences between the epistemic agents. A process reliabilist may argue that Shusaku and Jane’s beliefs are formed by belief-forming mechanisms that are more reliable than those responsible for the beliefs of either Juan or Maria.6 An evidentialist may claim that the quantity and quality of evidence upon which Jane and Shusaku base their beliefs is greater than that of either Juan or Maria.7 These are both rough

4

characterizations of how two leading theories of epistemic justification would account for the differences between the beliefs of the epistemic agents.

They are both attractive ways of

accounting for the differences. But Strategic Reliabilism tells us nothing about how to account for the differences in the epistemic status of the beliefs of the subjects in either scenario. Suppose we follow a good bit of Bishop and Trout’s advice and aim at developing theories of epistemic excellence, it does not follow that the traditional epistemological tasks of theorizing about knowledge, offering epistemic evaluations of belief tokens not formed on the basis of conscious reasoning, addressing the problem of skepticism, and other projects are not worth undertaking.

If such problems are such that they should be of no concern to

epistemologists, Bishop and Trout owe us substantive arguments in defense of such a strong claim. Of course, if they agree that theorizing about these matters lies within the province of epistemology, then the character of their epistemological project will quickly look more like a version of SAE than Bishop and Trout may want to admit. Any objections from the direction of Strategic Reliabilism notwithstanding, if we dig around a bit, it seems like Bishop and Trout assume more about matters addressed by proponents of SAE than they admit. As mentioned earlier, Bishop and Trout offer disparaging remarks about theorizing about justification. However, they offer advice by example about how to evaluate the epistemic status of belief tokens not formed on the basis of conscious reasoning. For instance, I suspect Bishop and Trout would agree that the following is true. If agents S1 and S2 have the same evidence about the reliability of Statistical Prediction Rules versus the unaided intuitive judgments of experts, and S1 forms the belief that Statistical Prediction Rules are more reliable and S2 forms the belief that the unaided judgments of experts are more reliable, S1’s belief is epistemically better than S2’s. Whether one wants to use terms like “justified” or not to

5

describe the beliefs of S1 and S2 is irrelevant. What is important is that Bishop and Trout do seem to assume that the evidence supports the truth of the proposition expressed by “Statistical Prediction Rules are more reliable than the unaided judgments of experts,” and that in normal circumstances it is epistemically better for people to believe this (see Bishop and Trout 2005a, pp. 169-70). Minimally, this suggests that they think it is generally better for people to believe what the preponderance of their evidence supports. So Bishop and Trout do seem to be in the business of evaluating belief-tokens absent any consideration of procedural rationality, even if they are not public about it and merely share an unarticulated assumption that it is epistemically better to believe some things rather than others. So they tacitly assume something like the following epistemic principle for determining when one proposition is epistemically better to believe than another (henceforth, EB): (EB) Ceteris paribus, for any agent S and any propositions p and ~p, it is epistemically better for S to believe that p if the preponderance of S’s evidence favors the truth of p rather than ~p.8 Of course, more needs to be said about what counts as evidence, as well as what if any other considerations should factor into making epistemic evaluations. But such work requires that we clarify concepts and do what looks a lot like SAE. For now, notice that an epistemologist who endorses EB, whether or not she endorses SAE explicitly, is offering advice to people about how to evaluate beliefs. And the advice that is offered looks like the sort of advice that we get from deep within our armchairs and is typical of SAE, not Strategic Reliabilism. Granted, while I cannot imagine them denying EB, they might do so on the grounds that it is just the sort of epistemic principle they want to dispense with. But doing so will be costly. I suspect they would not want to accept the costs, especially if it means denying that persons who agree with

6

them about the reliability of Statistical Prediction Rules have beliefs that are epistemically better than those of their opponents who believe differently on the basis of the same body of evidence. Bishop and Trout are not without a reply. They could argue that any concerns about evidence are parasitic upon considerations about robustly reliable belief-forming mechanisms, such as reasoning in accordance with Statistical Prediction Rules.9 As it stands, such a reply does not strike me as all that promising. Defenders of SAE will insist that Bishop and Trout tell us why reliability is the epistemic good-making property that confers positive epistemic status upon an agent’s beliefs. Again, a satisfactory story will require first doing something that looks more like SAE and less like the sort of alternative that some defenders of versions of naturalized epistemology such as Bishop and Trout purport will replace traditional epistemology. So whether we like it or not, it looks like we cannot escape the range of topics that have been the province of SAE and which Bishop and Trout do not believe we need to bother ourselves with. Moreover, it seems they must assume exactly what they say we do not really need to bother ourselves with, viz., a theory of epistemically justified belief tokens. So Strategic Reliabilism seems incomplete; and, insofar as Strategic Reliabilism works, it must assume some of the features, including the methods, of SAE necessary to make it complete. Bishop and Trout may claim that SAE is incomplete insofar as it fails to incorporate Ameliorative Psychology. But SAE can easily incorporate Ameliorative Psychology without turning into something else (see Feldman 1999). However, as I noted above, if Strategic Reliabilism is to be made complete, it will resemble SAE—a consequence I suspect Bishop and Trout want to resist. Bishop and Trout are not without a possible reply to the foregoing. They may argue that Strategic Reliabilism is not incomplete. This is so for two reasons. 10 First, some questions that defenders of SAE have found most vexing are not really as important as analytic epistemologists

7

have thought they are. If Strategic Reliabilism cannot provide answers to some of the traditional problems in epistemology it is only because the problems are not really worth thinking about. They are pseudo-problems at worst. So Strategic Reliabilism has the tools built into it that are necessary to address the most important problems epistemologists have been concerned about. It is no less complete for failing to have the resources necessary to address pseudo-problems. Of course, such a response is ad hoc and seems suspect given the wide range of important everyday scenarios that Strategic Reliabilism appears to lack the resources to address (e.g., the foregoing examples of forming beliefs on the basis of introspection, perceptual experience, and testimony, not to mention some more rarefied scenarios involving brains in vats and Cartesian demons). The second reason they offer for the completeness of Strategic Reliabilism is more interesting and presents a more formidable challenge to the defender of SAE. Recall that Bishop and Trout assert that Strategic Reliabilism offers a means of determining the epistemic status of belief tokens, albeit “at one remove” (2005a, pp. 17, 116). In a section of their book where they compare Strategic Reliabilism to SAE, Bishop and Trout offer an “inessential codicil” to Strategic Reliabilism: “The beliefs that result from the reasoning strategies recommended by Strategic Reliabilism are justified” (2005a, p. 116). While they claim that “It is no objection to our view to show that there is a scenario in which it is the considered judgment of analytic epistemologists that the belief recommended by Strategic Reliabilism is not justified” the inessential codicil allows them to ask the following question: “Does Strategic Reliabilism recommend only beliefs that are justified (in whatever sense a proponent of SAE means by ‘justified’)?” (Bishop and Trout 2005a, p. 116). If God gave the proponent of SAE a theory of justification (Theory J) that satisfied all of the analytic epistemologist’s desiderata would Theory J and Strategic Reliabilism (with the inessential codicil added) always agree? (Bishop and Trout

8

2005a, p. 116). These are two ways of asking the same question. However the question is formulated, Bishop and Trout contend that no matter how the question is answered, Strategic Reliabilism is superior to any SAE theory offered to date. Their reasoning can be represented as follows.11 1. Either Strategic Reliabilism recommends only justified beliefs or it does not. 2. If it recommends only justified beliefs, then there is no need for any other theory of justification. 3. If it does not recommend only justified beliefs, then the concept of epistemic justification is unimportant.12 4. So either there is no need for any other theory of justification or the concept of epistemic justification is unimportant. In the remainder of this section I will focus on showing why premises (2) and (3) of this argument are false. Consider premise (2). Suppose that whatever beliefs are justified under Theory J are recommended by Strategic Reliabilism. So we do not need SAE (Bishop and Trout 2005a, p. 117). If two SAE theories are in conflict over a particular case and Strategic Reliabilism yields the same results as Theory J, then we can break any deadlock by simply determining what Strategic Reliabilism recommends. SAE is superfluous. This assumes, of course, that Strategic Reliabilism will yield the same results as Theory J. But in the actual world we do not know whether Strategic Reliabilism will yield the same results as Theory J. I am not confident that it will or that it can. For one, even with the inessential codicil there are a host of types of cases that Strategic Reliabilism as stated is ill prepared to address. Going back to my earlier examples of forming beliefs immediately on the basis of sense perception or testimony, it is not obvious what

9

Strategic Reliabilism will tell us about the status of the beliefs the agents form since it is not clear that any reasoning is involved in the formation of their beliefs. Their experiences cause them to acquire certain beliefs. So even if Theory J and Strategic Reliabilism will always agree with one another, it is not obvious that Strategic Reliabilism will provide an answer to our questions about a host of epistemological problems that cannot be written off as esoteric pseudo-problems. For instance, unlike Theory J, Strategic Reliabilism (even with the codicil) will prove ill-equipped to resolve a dispute between a process reliabilist and an evidentialist over why the person who is the subject of a visual experience may have a belief that is epistemically better than her friend who forms a belief with the same content on the basis of testimony. At this point it seems we can agree with Bishop and Trout about Theory J and Strategic Reliabilism and still insist that Strategic Reliabilism is incomplete. It is not obvious that if Strategic Reliabilism recommends only justified beliefs then there is no need for any other theory of justification. This is the case if for no other reason than Strategic Reliabilism does not provide us with the tools necessary to evaluate the beliefs of agents that are not formed on the basis of any reasoning strategies. So premise (2) is false. What about premise (3)? Suppose that Theory J and Strategic Reliabilism diverge. In this case S’s belief that p is the result of a reasoning strategy recommended by Strategic Reliabilism, but believing that p is not justified according to Theory J (one would justifiably believe that ~p). Which should one believe, p or ~p? Bishop and Trout answer as follows. The belief recommended by Strategic Reliabilism is the result of excellent reasoning on the part of S. Excellent reasoning maximizes S’s chances of coming to true beliefs about significant problems. This, in turn, will tend to lead in the long run to better outcomes

10

for S than if she had adopted reasoning strategies of lesser quality.

So the belief

recommended by Strategic Reliabilism has a lot going for it. (2005a, p. 117) The only advantage of holding the belief recommended by Theory J (and, thus, SAE) is that it is deemed justified by analytic epistemologists. Even if a wider range of people who understand the concept of epistemic justification would deem believing that ~p as justified, it is not the belief that should be recommended from an epistemic perspective, according to Bishop and Trout. S’s belief that p “is the belief that comes with all the practically important empirical advantages of epistemic goodness and of epistemic success” (Bishop and Trout 2005a, p. 117). Bishop and Trout seem to be committed to the claim that broadly practical concerns (for instance, realizing one’s goal of forming one’s beliefs in accordance with reasoning strategies recommended by Strategic Reliabilism) can outweigh any deontic epistemic norms defenders of SAE might associate with epistemic justification (for instance, you ought to believe whatever is recommended by Theory J). They assume that the proponent of SAE will take epistemic norms associated with having justified belief tokens to outweigh any competing norms. But it is not obvious that the defender of SAE must prefer having one belief justified over realizing the goal of having many more beliefs justified as a consequence of believing what is recommended by excellent reasoning. Perhaps this is because epistemic norms are subservient to broadly practical norms in cases of conflict.13 But if we suppose that epistemic norms are not obviously trumped by practical ends or vice versa, it may be the case that there is a parity between epistemic norms and practical norms. Adjudicating between them would be difficult because neither type of norm obviously is privileged over the other. So while the prudent and practically rational thing for agents to do is to believe contrary to what Theory J recommends and doing so is epistemically

11

unjustified, it is not obviously better in some generic sense of ‘better’ to believe one way or the other.14 Finally, it is worth noting that there are competing senses of “epistemic rationality” that need to be clearly distinguished. If Theory J is a theory of epistemic justification, then it seems that it would provide a theory of end-state rationality, which is synchronic (and so applies to belief tokens). Strategic Reliabilism, on the other hand, is a theory of procedural rationality, which is diachronic. Suppose that Theory J is a version of evidentialism. Theory J would recommend believing p because the preponderance of evidence S has supports p (suppose that all of the expert testimony S considers favors p). However, Strategic Reliabilism may recommend believing ~p because following the best reasoning strategy would result in forming the belief that ~p. But believing ~p is recommended only because it is the output of the best reasoning strategy. Once again, we are faced with a situation where different types of rationality (synchronic and diachronic) and so different norms are in competition with each other. In this case it is assumed that these are both ways of being epistemically rational and the norms are all epistemic.15 It is not obvious that the defender of SAE must insist that synchronic rationality will always trump diachronic rationality. It seems that Bishop and Trout are sparring with straw opponents. I do not claim to speak on behalf of all analytic epistemologists. But it is just not true that the proponent of SAE must agree that a tension between Theory J and Strategic Reliabilism either favors Theory J or Strategic Reliabilism. And even if Strategic Reliabilism wins sometimes, this may be because of an agent’s ends, be they construed as broadly practical or epistemic. In any case, the concept of justification seems no less important even if it does not enjoy an exalted position. So premise (3) is false.

12

The upshot of the foregoing is that, even with the inessential codicil added, Strategic Reliabilism is incomplete. Strategic Reliabilism is not an adequate replacement for SAE. We need more if we are going to satisfy the desiderata of epistemological theorizing, construed broadly. Conclusion: (Mostly) Safe in our Armchairs Bishop and Trout have not shown that SAE is in its last throes and that its methods and projects are dispensable. Epistemology cannot just be Ameliorative Psychology. If it is, it will be incomplete.16

Notes 1

Recent reviews of Epistemology and the Psychology of Human Judgment have mentioned the incompleteness of

Bishop and Trout’s proposal (see Buckareff 2006 and Weinberg Forthcoming). But none explore this problem to the extent that I am here. Additionally, Alan Goldman (2005) raises worries similar to my own about Bishop and Trout’s critical comments about SAE and its shortcomings. But there are significant differences between the concerns Goldman raises and what I am taking up here. For one, I am going on the offensive, so to speak, emphasizing the incompleteness of Bishop and Trout’s proposed alternative to SAE. Moreover, Goldman does not directly address Bishop and Trout’s arguments for why their proposal does not result in an incomplete epistemology and should replace SAE. 2.

A referee for this journal asked whether, qua reliabilist theory of epistemic excellence, it is essential to Strategic

Reliabilism that it focus on voluntary processes such as conscious reasoning.

As reliabilists, they can avail

themselves of the resources offered by prior work on the various belief-forming modalities that are assessed for reliability. While this seems correct, and Bishop and Trout even write approvingly of the earlier work of Goldman (2005a, p. 183), epistemology as Ameliorative Psychology is emphasized throughout their book. Moreover, it will be evident to readers of Epistemology and the Psychology of Human Judgment that what Bishop and Trout offer, while having a rightful place among reliabilist theories in epistemology, is quite different from what has been offered to date by other reliabilists who offer reliabilist theories of epistemic justification and knowledge.

13

3.

It is worth noting that Bishop and Trout appear to be using ‘reasoning’ to refer to both voluntary processes and the

involuntary mechanisms involved in reasoning. But their emphasis is upon voluntary processes. Insofar as Strategic Reliabilism is Ameliorative Psychology, it aims at giving advice to persons about behavior they can exercise some voluntary control over. Evidence of their being primarily concerned with voluntary processes is found in their repeated remarks such as the following: “Much of the point of Ameliorative Psychology is to provide advice that will help people reason better about the world” (Bishop and Trout 2005a, p. 17). But they do not ignore involuntary belief-forming mechanisms that play a role in reasoning. Towards the end of the book they note, “Our efforts have typically been directed at voluntary reasoning strategies—strategies reasoners can choose to use or not to use. That’s not to say that involuntary reasoning processes should be completely ignored. epistemology must pay closer attention to such processes.

In fact, in our view,

For example, a practical epistemology will offer

voluntary reasoning strategies that correct involuntary reasoning processes (e.g., don’t trust your visual color experiences in artificial light)” (Bishop and Trout 2005a, p. 182). Thanks to a referee for this journal for pressing me on this point. 4

A referee for this journal has suggested that what Bishop and Trout are up to is quite similar to Alvin Goldman’s

project in Epistemology and Cognition (1988). I agree there are some similarities. Bishop and Trout approve of Goldman’s call for a more scientific epistemology. But the differences between Goldman’s work and Bishop and Trout’s are important. Bishop and Trout aim at replacing SAE, rejecting its methods and its goals. Goldman engages in the type of philosophical work that is characteristic of SAE in the first half of Epistemology and Cognition, defending a thesis about how to evaluate the beliefs of agents in reliabilist terms. Goldman is concerned about traditional problems focused on by practitioners of SAE—e.g., justification, skepticism, and providing a theory of knowledge. Bishop and Trout are not concerned with the core problems traditionally associated with SAE with which Goldman is concerned. Finally, what should be apparent to those familiar with Goldman’s work and Bishop and Trout’s work is that while the latter endorse a very strong version of the replacement thesis, Goldman endorses a more modest proposal of redirecting epistemology (and not the outright replacement of SAE). Goldman pushes us in the direction of doing scientifically informed work in epistemology, which is welcomed by many who do not consider themselves reliabilists and who are otherwise critical of Goldman’s project (e.g., Feldman 1999 and Pollock and Cruz 1999).

14

5

Since I am concerned with defending SAE broadly, I will avoid the problems raised against reliabilism in Conee

and Feldman 2004. For my purposes, it is best to remain neutral with respect to the debates among proponents of SAE. Incidentally, Bishop and Trout do take preemptive measures and respond to a possible version of the generality problem in the Appendix of Bishop and Trout 2005a. 6

See Goldman 1979.

7

See Feldman and Conee 2004.

8

In an earlier draft of this paper I offered EB without a ceteris paribus clause built into it. In his comments on an

earlier draft of this paper Michael Bishop noted that, as stated, absent a ceteris paribus clause, EB is a principle he and J.D. Trout would reject.

This is so because Strategic Reliabilism involves a cost-benefit approach to

epistemology. So there are instances where reasoning strategies and belief-policies that are less-than-ideal are recommended. This means that there are cases where it is epistemically better to accept ~p when the ideal strategy would have led one to accept p. If EB enjoys some plausibility as an epistemic principle Bishop and Trout would accept, then it seems we have provided an epistemic principle that at least enjoys prima facie plausibility and can be refined from the comfort of our armchairs. 9

Thanks to a referee for this journal for raising this objection.

10

Thanks to Michael Bishop for bringing this to my attention.

11

What follows is based on an argument offered by Bishop and Trout without any numbered premises (see 2005a,

pp. 117-18). 12

I admit that I am taking some liberties in my representation of the consequent of premise (3). They have “so much

the worse for justification” where I have “the concept of epistemic justification is unimportant.” I do not believe that the liberties I am taking result in an unfair presentation of the views of Bishop and Trout. If SAE recommends only justified beliefs where Strategic Reliabilism does not, then “the orthodox concept of justification … is a crude and insensitive instrument of evaluation…. Those devoted to offering a psychologically accurate picture of reasoning strategies … are likely to find such ‘justification talk’ quaint, were it not for the resources squandered by its primping” (Bishop and Trout 2005a, p. 118). 13

John Heil has argued in defense of this type of view. He writes that: “Because epistemic norms are instruments the

primary function of which is to satisfy extrinsic non-epistemic interests, rationality as a whole requires that we promote non-epistemic values at the expense of their epistemic counterparts when these are at odds” (1992, p. 48)

15

14

Richard Feldman (2004, pp. 191-94) has argued in defense of a position similar to this. See Scott Aiken 2006 for

a critique of Feldman from a perspective similar to that found in Heil 1992. 15

Of course, diachronic epistemic rationality may just be a species of practical rationality. If this is the case, then it

is obvious that at least the epistemic norms associated with diachronic rationality are subservient to practical ends. 16

Thanks to Michael Bishop, Nathan Nobis, and an anonymous referee for this journal for helpful comments on

earlier versions of this paper.

References Aiken, Scott 2006 “Modest Evidentialism.” International Philosophical Quarterly, 46: 327-43. Bishop, Michael and Trout, J.D. 2005a Epistemology and the Psychology of Human Judgment. New York: Oxford University Press. Bishop, Michael and Trout, J.D. 2005b “The pathologies of standard analytic epistemology.” Noûs, 39: 696-714. Buckareff, Andrei 2006

Review of Michael Bishop and J.D. Trout, Epistemology and the Psychology of Human Judgment. Philosophical Psychology, 19: 847-51.

Conee, Earl and Feldman, Richard 2004 “The Generality Problem for Reliabilism. Reprinted in their Evidentialism: Essays in Epistemology. New York: Oxford University Press, pp. 135-65. Feldman, Richard

16

2004 “The ethics of belief.” Reprinted in Earl Conee and Richard Feldman, Evidentialism: Essays in Epistemology. New York: Oxford University Press, pp. 166-95. Feldman, Richard 1999 “Methodological Naturalism in Epistemology.” In The Blackwell Guide to Epistemology. Edited by J. Greco and E. Sosa. Malden, MA: Blackwell, pp. 17086. Feldman, Richard and Conee, Earl 2004 “Evidentialism.” Reprinted in their Evidentialism: Essays in Epistemology. New York: Oxford University Press, pp. 83-107. Goldman, Alan 2005 Review of Michael Bishop and J.D. Trout, Epistemology and the Psychology of Human Judgment. Notre Dame Philosophical Reviews, URL = . Goldman, Alvin 1979 “What is justified belief?” In Justification and knowledge. Edited by G. Pappas. Dordrecht: D. Reidel, pp. 271-93. Goldman, Alvin 1988 Epistemology and cognition. Cambridge, MA: Harvard University Press. Heil, John 1992 “Believing reasonably.” Noûs, 26: 47-61. Kim, Jaegwon 1988 “What is ‘Naturalized Epistemology’?” Philosophical Perspectives, 2: 381-405.

17

Kornblith, Hilary 1994 “Introduction: What is Naturalistic Epistemology?” In Naturalizing Epistemology. Edited by H. Kornblith. Cambridge, MA: MIT Press, pp. 1-14. Pollock, John L. and Cruz, Joseph 1999 Contemporary Theories of Knowledge, 2nd ed. Lanham, MD: Rowman & Littlefield. Quine, Willard Van Orman 1994 “Epistemology Naturalized.” Reprinted in Naturalizing Epistemology. Edited by H. Kornblith. Cambridge, MA: MIT Press, pp. 15-31. Weinberg, Jonathan Forthcoming Review of Michael Bishop and J.D. Trout, Epistemology and the Psychology of Human Judgment. Philosophy of Science.

18

19