Fundamental and Derivative Epistemic Value
With respect to any given domain of assessment, Sosa notes in his recent Locke Lectures (2007, 70), that we can distinguish fundamental values from derivative ones. A fundamental value in a domain may be understood as something that is good relative to the domain but not by virtue of something else’s being good relative to that domain. A derivative value in a domain can then be characterized as something that is good relative to the domain by virtue of something else’s being good relative to it. In the domain of computer assessment, for example, a computer’s having a high processing power would likely count as a fundamental value, whereas the computer’s being produced by a reputable manufacturer would be a derivative value. In gastronomy, a dish’s deliciousness, nutritiousness, and presentation may be fundamental values, and its production by traditional means—mise en place, relevant knife skills, and marinading technique, for example—merely derivative ones. If in ethics we can take morally right action as a fundamental value, perhaps action that conforms to certain conventional behavioural norms will be a derivative value.
In the sense just specified, a fundamental value needn’t be good relative to a domain simply by virtue of its intrinsic properties. A nutritious dish may be a fundamental value in gastronomy and what makes it good relative to that domain may be such relational properties as its tending to produce certain bodily effects in those who consume it. All that is required is that a fundamental value is good relative to a given domain and that whatever makes it good relative to that domain not include the goodness of something else relative to the domain. Hence, at least on one construal of “intrinsic value”—a Moorean one (see Moore Reference Moore1922, 260; cf. Bradley)—the fundamental values in a domain needn’t be limited to its intrinsic values; non-intrinsic “final values” or “end values” of the sort discussed by Korsgaard, Kagan, and Rabinowicz and Rønnow-Rasmussen may also count. Moreover, as Sosa (Reference Sosa2007, 74) suggests, although instrumental values—things that are good in a domain by virtue of causing or promoting other good things in the domain—are the most salient examples of derivative values, it is not obvious that the derivative is restricted to the instrumental. We can admit that a dish’s receiving a positive assessment from an expert—the renowned food critic, say, or the master chef—is a derivative value in gastronomy without insisting that it always brings about other gastronomic values.
It’s also worth noting that a fundamental value in one domain may be a derivative value in another. That a computer’s having a high processing power is a fundamental value in the computer assessment domain is consistent with its being a derivative value in some further domain: the computer’s high processing power may be good relative to a practical, information acquisition domain of assessment by virtue of such things as an individual’s ability to acquire information quickly being good, but it need not be good relative to the computer assessment domain by virtue of something else’s being so good.
Consonantly with the general distinction between fundamental and derivative values, we can say that a fundamental epistemic value is something that is epistemically good—good relative to the domain of assessing our representational cognitive states, but not by virtue of anything else’s being epistemically good. A derivative epistemic value, then, will be something that is epistemically good by virtue of something else’s being epistemically good.
Against Epistemic Value Monism
According to a popular view, there is but one fundamental epistemic value, viz. truth, or true belief:
Epistemic Value Monism (Truth): the sole fundamental epistemic value is true belief.
This view finds considerable support across a wide epistemological spectrum. Witness, for example, BonJour:
What makes us cognitive beings at all is our capacity for belief, and the goal of our distinctly cognitive endeavors is truth: we want our beliefs to correctly and accurately depict the world. . . . The basic role of justification is that of a means to truth, a more directly attainable mediating link between our subjective starting point and our objective goal. . . . If epistemic justification were not conducive to truth in this way, if finding epistemically justified beliefs did not substantially increase the likelihood of finding true ones, then epistemic justification would be irrelevant to our main cognitive goal and of dubious worth [emphasis BonJour’s]. (7)
Or listen to Goldman (1986):
The central epistemological concepts of appraisal. . . . invoke true belief as their ultimate aim. So the evaluation of epistemic procedures, methods, processes, or arrangements must appeal to truth-conduciveness, an objective standard of assessment [emphasis Goldman’s]. (3)
[C]oherence is not the highest standard of justificational rightness but a derivative, subsidiary standard. The fundamental standard concerns the formation of true belief. Coherence enters the picture only because coherence considerations are generally helpful in promoting true belief. (100)
True belief is the value that J-rules should promote—really promote—if they are to qualify as [epistemically] right. (103)
Even Alston’s attempt to reorient our epistemological focus toward a plurality of epistemic desiderata beyond justification proceeds against the backdrop of his insistence that true belief remains “the basic goal of cognition” (29) and “the master epistemic desideratum” (42).
Despite such prominent support, Epistemic Value Monism (Truth) turns out to be difficult to accept upon reflection, for it seems to sit ill at ease with some very deep intuitions about the epistemic value of knowledge itself—a point that has become widely recognized in the growing literature on epistemic value.Footnote 1 Knowledge is not only an epistemic value, but a greater epistemic value than mere true belief: it’s epistemically better to know that p than merely to believe truly that p. And it’s very difficult to square this with Epistemic Value Monism (Truth). Thus, consider the following argument:
The Epistemic Value of Knowledge Argument
P1Knowledge is an epistemic value.
P2If something is the sole fundamental value in a domain of assessment δ, then any other value in δ is a derivative value in δ.
C1Therefore, if true belief is the sole fundamental epistemic value, then knowledge is a derivative epistemic value.
P3A derivative value in a domain of assessment δ cannot be a greater value in δ than the sole fundamental value in δ.
C2Therefore, if true belief is the sole fundamental epistemic value, then knowledge cannot be a greater epistemic value than true belief.
P4But knowledge is a greater epistemic value than true belief.
C3Therefore, true belief is not the sole fundamental epistemic value.
P1 is well-nigh an analytic truth. At any rate, if knowledge fails to count as an epistemic value, it’s hard to see how any other cognitive state—including those that have traditionally been thought to be among its constituents, such as true belief—could. And P2 follows from the point that the fundamental-derivative value distinction in a given domain is exhaustive: every value in a domain is either a fundamental or a derivative value in that domain. If, then, there is only one fundamental value in a domain, every other value in it is derivative.
What about constitutive values in a given domain of assessment? Might they serve as counter-examples to the claim that every value in a domain is either fundamental or derivative? The answer is no, for every constitutive value is derivative. To say that a value in a domain is constitutive is to say that it is good relative to the domain and that its goodness relative to the domain is constituted by the goodness of other things relative to that domain. But if something is good relative to a domain and its goodness is constituted by that of other things relative to the domain, then it is good relative to the domain by virtue of other things being good relative to the domain. If we take a dog’s being a well-groomed purebred to be a constitutive value in the dog confirmation show domain of assessment, then we admit that it is good relative to the show domain and that its goodness is constituted by the goodness, relative to that domain, of the dog’s being well-groomed and of its being a purebred. A dog’s being a well-groomed purebred is thus a derivative value in the show domain, because it is good relative to that domain by virtue of other things—a dog’s being well-groomed and its being a purebred—being good relative to the domain.
The plausibility of P3 is most commonly illustrated in the literature by the coffee example originally raised by Zagzebski. In one specifiable domain of coffee assessment, the sole fundamental value consists of a cup of coffee’s being delicious; a cup’s having been produced by a reliable espresso machine—assuming it is a coffee-assessment value—is thus merely a derivative coffee-assessment value. And neither that derivative value, nor any constitutive value that includes it can, it seems, be a greater coffee-assessment value than that of a delicious cup of coffee. Simply by virtue of its having been produced by a reliable espresso machine, a cup of coffee cannot be a better cup of coffee than one that is delicious. One of two equally delicious cups of coffee is not a better cup of coffee because it, unlike the other, came from a high-end Saeco.
P4, moreover, is nearly if not equally as plausible as P1. Indeed, some (DePaul, 174; cf. Jones, 423-4) have claimed that it is so plausible as to be trivial. Something is epistemically lacking, as Gettier may be read as having pointed out to us, if your justified true belief fails to amount to knowledge: your mere justified true belief that p is not as good, from the epistemic point of view, as knowledge that p. So too, then, for mere true belief: it fails to ascend as high on the scale of epistemic value as knowledge.
The Epistemic Value of Knowledge Argument, accordingly, provides a compelling reason to reject Epistemic Value Monism (Truth). But if we reject this veritistic form of monism about epistemic value, what can we put in its place? We might be tempted to hang on to the monism bit and change the veritistic bit. Consider, then, the following version of epistemic value monism:
Epistemic Value Monism (Justification): the sole fundamental epistemic value is justified belief,
where “justified belief” has the broad sense of the kind of well-formed belief that is entailed by knowledge but does not entail true belief, and where we thus leave it open whether justified belief is to be given an externalist or internalist treatment.
Unfortunately, this form of monism seems all too easily ruled out by a slight revision to the Epistemic Value of Knowledge Argument: just substitute “justified belief” for “true belief” throughout, and we get the conclusion that justified belief is not the sole fundamental epistemic value. The key difference between the original version of the argument and the revised one pertains to the fourth premise of each. And if it’s plausible that knowledge is a greater epistemic value than mere true belief, it’s surely also plausible that knowledge is a greater epistemic value than mere justified belief.
There is, however, yet another variety of epistemic value monism worth considering, one that takes neither truth nor justification as uniquely pre-eminent in the domain of epistemic assessment, but rather knowledge itself:
Epistemic Value Monism (Knowledge): the sole fundamental epistemic value is knowledge.
Note that this form of monism is not susceptible to refutation by a variation on the Epistemic Value of Knowledge Argument. Such a variation would have to hold either that true belief is a greater epistemic value than knowledge or that justified belief is a greater epistemic value than knowledge, and that would run directly contrary to the intuitions we have about the greater epistemic value of knowledge.
Even so, according to Epistemic Value Monism (Knowledge), the epistemic goodness of any state other than knowledge supervenes on the epistemic goodness of knowledge, such that if there were no hope at all of obtaining knowledge in a given situation, there would be no hope at all of obtaining any epistemic goodness in that situation. Take away the possibility of the state on whose epistemic goodness that of all other states depends in a given situation and you take away the possibility of the epistemic goodness of all other states in the situation. Indeed, a more general principle seems operative here: If the domain-relative goodness of any state in a situation depends upon a particular state’s being so good in that situation, then the impossibility of the particular state’s obtaining in the situation means that nothing can be so good in the situation.Footnote 2 If in a given situation all coffee-assessment goodness really does depend upon the coffee-assessment goodness of a delicious cup of coffee, then if a delicious cup of coffee cannot be had in the situation there’s no point, from the coffee-assessment point of view, to turning on the espresso machine in that situation, for nothing it can there produce will be of any coffee-assessment goodness.
But if the preclusion of knowledge in a given situation precludes any epistemic goodness in that situation, it precludes any epistemic goodness of justified belief in that situation as well: whether or not one can obtain justified belief in the situation, it can be of no epistemic goodness in that situation. This result, however, translated into the concrete, yields some highly counterintuitive consequences. Imagine that on some deeply difficult matter of contemporary physics, we just cannot know the truth. Does it follow that, from the point of view of assessing the representational cognitive attitudes of physicists on the matter, we cannot hold some of their beliefs—justified ones—to be epistemically good and others—unjustified ones—to be epistemically bad? It seems not. Intuitively, we can still rightly assess some of their beliefs about the matter as being of some epistemic goodness, despite the preclusion of the possibility of knowledge.
Or consider the “New Evil Demon Problem.” Although originally characterized as a problem specifically for reliabilist theories (Lehrer and Cohen; cf. Foley,150 and Sosa Reference Sosa1991, 132), it is easily seen to be a problem for any account according to which justified belief is merely a derivative epistemic value. The problem is that it is very intuitive that victims of a Cartesian evil demon, or brains in vats if you prefer, can be evaluated as doing well epistemically, despite the fact that true belief and hence knowledge are precluded for them, and any view that takes justified belief to be a mere derivative epistemic value seems unable to account for this. “Compare,” Sosa suggests, “two evil demon victims”:
The first victim takes in fully and flawlessly the import of her sense experience and other states of consciousness, to an extent rarely matched by any human, and reasons therefrom with equal prowess to conclusions beyond the reach of most people, and retains the results in memory well beyond the normal. The other victim is on the contrary extensively handicapped in her cognitive faculties and performs with singular ineptness. (2003, 175)
It’s obvious, Sosa goes on to point out, that the first of these victims is faring epistemically better than the second. It’s equally obvious, we may add, that the justified beliefs of the first are of some epistemic goodness. The leading intuition of the New Evil Demon Problem, as well as that of the physics example, then, implies that even in situations where knowledge is precluded (i.e., cannot obtain), justified belief can be epistemically good. And this can now be used as the basis for an argument against Epistemic Value Monism (Knowledge):
The Epistemic Value of Justification Argument
P1If knowledge is the sole fundamental epistemic value, then if knowledge is precluded in a situation σ nothing can be epistemically good in σ.
P2If nothing can be epistemically good in a situation σ, then justified belief cannot be epistemically good in σ.
C1Therefore, if knowledge is the sole fundamental epistemic value, then if knowledge is precluded in a situation σ, justified belief cannot be epistemically good in σ.
P3But justified belief can be epistemically good in a situation σ even if knowledge is precluded in σ.
C2Therefore, knowledge is not the sole fundamental epistemic value.
An Epistemic Value Pluralism
In the light of the arguments considered above, monism about epistemic value seems to consist of dead-end trails. Passing over an obviously implausible nihilism about epistemic value, according to which there are no fundamental epistemic values (and hence either no epistemic values at all or an infinite number of infinitely regressing derivative epistemic values), we are thus led to a pluralist alternative:
Epistemic Value Pluralism: there are multiple fundamental epistemic values.
In the remainder of this paper, I want to outline a particular version of Epistemic Value Pluralism and consider its plausibility.
I do not think a pluralist can afford to be stingy about the states she admits as fundamental epistemic values. One might try to take only true belief and justified belief as fundamental epistemic values, for example, and to treat knowledge as a derivative epistemic value because it is a constitutive epistemic value.Footnote 3 On this view, the epistemic goodness of knowledge would be constituted by that of true belief and that of justified belief; hence knowledge would be epistemically good partly by virtue of true belief’s being epistemically good and partly by virtue of justified belief’s being epistemically good. The trouble, however, with this stingy form of pluralism lies in the difficulty of squaring it with a central insight brought to bear in support of P4 of the Epistemic Value of Knowledge Argument, viz. that knowledge is a greater epistemic value not only than mere true belief, and not only than mere justified belief, but also than the two together: again, in the light of Gettier, something is epistemically lacking if your justified true belief fails to amount to knowledge. But then how can we account for the greater epistemic value of knowledge over true belief and justified belief together—over mere justified true belief—if the epistemic goodness of knowledge is derivative because constituted just by theirs together? It seems that we cannot, for a slightly different but equally plausible version of P3 of the Epistemic Value of Knowledge Argument blocks any such account: it’s not just that a derivative value in a domain of assessment δ cannot be a greater value in δ than the sole fundamental value in δ; it’s also that a derivative value in a domain of assessment δ cannot be a greater value in δ than both, or all, of the sole fundamental values in δ together. Consider once again Zagzebski’s coffee example, only now modify it such that the sole fundamental coffee-assessment values are two: a delicious cup of coffee and an aromatic cup of coffee. Now consider the following derivative value in the relevant domain of coffee-assessment: the constitutive coffee-assessment value of a delicious, aromatic cup of coffee. It clearly cannot be a greater coffee-assessment value than the value that obtains when the fundamental values of a delicious cup of coffee and an aromatic cup of coffee are put together—no more so than the derivative value of a cup’s being produced by a reliable espresso machine.
Otherwise put, the stingy form of pluralism that takes true belief and justified belief to be the only two fundamental epistemic values, and knowledge to be a derivative epistemic value whose epistemic goodness is constituted by the conjunction of their epistemic goodnesses, seems subject to refutation by a revision to the Epistemic Value of Knowledge Argument significant enough to justify giving it its own name:
The New Epistemic Value of Knowledge Argument
P1Knowledge is an epistemic value.
P2If x1…xn are the sole fundamental values in a domain of assessment δ, then any other value in δ y is a derivative value in δ.
C1Therefore, if true belief and justified belief are the sole fundamental epistemic values, then knowledge is a derivative epistemic value.
P3A derivative value in a domain of assessment δ y cannot be a greater value in δ than the sole fundamental values in δ x1…xn together.
C2Therefore, if true belief and justified belief are the sole fundamental epistemic values, then knowledge cannot be a greater epistemic value than justified true belief.
P4But knowledge is a greater epistemic value than justified true belief.
C3Therefore, true belief and justified belief are not the sole fundamental epistemic values.
The version of pluralism I have in mind, by contrast to the stingy one just considered, is more liberal; it takes true belief, justified belief, and knowledge all to be fundamental epistemic values.Footnote 4 To that extent alone, my pluralist alternative avoids refutation by anything like the arguments we have considered against the varieties of epistemic value monism. Still, more must be added to yield a plausible explanation of the truth of two intuitions that play a central part in those arguments, viz. that knowledge is a greater epistemic value than mere true belief and that knowledge is a greater epistemic value than mere justified belief. The truth of the intuition that justified belief can be epistemically good and hence epistemically better than unjustified belief even in situations where knowledge is precluded is, I take it, already accounted for simply by admitting justified belief as a fundamental epistemic value. And for purposes of simplicity I will leave aside further discussion of the central intuition in the New Epistemic Value of Knowledge Argument, viz. the argument’s P4; I assume that an adequate explanation of it can be easily derived from an adequate explanation of the intuitions that knowledge is a greater epistemic value than true belief and that knowledge is a greater epistemic value than justified belief.
That is, we have the following claims:
1 Knowledge is a fundamental epistemic value.
2 True belief is a fundamental epistemic value.
3 Justified belief is a fundamental epistemic value.
4 Knowledge is a greater epistemic value than true belief.
5 Knowledge is a greater epistemic value than justified belief.
And we want some additional claims as part of our pluralist package to help us explain the connection between 1–3 and 4–5. If knowledge, true belief, and justified belief all count as fundamental epistemic values, how can we account for the greater epistemic value of knowledge over true and justified belief?
No account that appeals to true belief and justified belief as epistemic value constituents of knowledge will do. To illustrate, suppose we were to add to 1–5 the following three claims in our effort to explain the connection between 1–3 and 4–5:
6* If x and y are fundamental values in a domain of assessment δ, and the goodness of y relative to δ is a constituent of the goodness of x relative to δ, then x is a greater value than y in δ.
7* The epistemic goodness of true belief is a constituent of the epistemic goodness of knowledge.
8* The epistemic goodness of justified belief is a constituent of the epistemic goodness of knowledge.
The problem now is that 1–8* are incoherent, as can be seen by recalling the point that all constitutive values are derivative values. If either 7* or 8* is true, then knowledge is a derivative epistemic value because it is a constitutive epistemic value. But according to 1, knowledge is a fundamental epistemic value. Hence 1–8* jointly entail that knowledge is both a fundamental epistemic value and a derivative epistemic value—a contradiction.
We can still hold knowledge to be a constitutive cognitive state even if we refrain from affirming it as a constitutive epistemic value: we need merely to add that the epistemic goodness of knowledge does not supervene on the epistemic goodness of constituents of knowledge such as true belief and justified belief. There is conceptual room to insist, in other words, that true belief and justified belief are constituents of knowledge but that the epistemic goodness of true belief and the epistemic goodness of justified belief are not constituents of the epistemic goodness of knowledge. But this appeal to constitution, alone, is of no help in explaining the link between 1–3 and 4–5, because it tells us nothing at all about the greater epistemic value of knowledge over that of true belief and justified belief.
If an appeal to constitution will not serve as the explanatory link between 1–3 and 4–5, then, what will? I think the answer turns out to be pretty straightforward: non-constitutive, non-mutual entailment. As Williamson reminds us in the context of his case against the tradition of seeing knowledge as a reductively analyzable state, entailment without being entailed does not entail being (in part) constituted by. Consider a simple example: “Although being coloured is a necessary but insufficient condition for being red,” Williamson points out, “we cannot state a necessary and sufficient condition for being red by conjoining being coloured with other properties specified without reference to red” (3). It is in other words possible—because actual—that something’s being red entails its being coloured, that its being coloured does not entail its being red, and yet that its being red is not (in part) constituted by its being coloured. Other examples abound: it’s possible that something’s being triangular entails its being a plane geometric figure, that its being a plane geometric figure does not entail its being triangular, and yet that its being triangular is not (in part) constituted by its being a plane geometric figure; and so on. Similarly, we can say that the epistemic goodness of knowledge entails that of true belief (likewise that of justified belief), the epistemic goodness of true belief (likewise that of justified belief) does not entail that of knowledge, and yet that the epistemic goodness of knowledge is not (in part) constituted by that of true belief (or by that of justified belief).
In which case, we can effect an explanatory link between 1–3 and 4–5 with the likes of 6–10 instead of the likes of 6*–8*:
6 If x and y are fundamental values in a domain of assessment δ, the goodness of x relative to δ entails that of y, and the goodness of y relative to δ does not entail that of x, then x is a greater value than y in δ.
7 The epistemic goodness of knowledge entails that of true belief.
8 The epistemic goodness of true belief does not entail that of knowledge.
9 The epistemic goodness of knowledge entails that of justified belief.
10 The epistemic goodness of justified belief does not entail that of knowledge.
With 1–10 together we have a view of epistemic value that not only avoids refutation by the sorts of arguments that can be marshalled against epistemic value monism, as we have seen, but also yields a satisfactory explanation of some central intuitions that motivate much of the relevant literature.
Do 1–10 imply that knowledge is an “organic unity” relative to the epistemic domain of assessment? Only, I think, if we add to them something like the view (mentioned above) that true belief and justified belief are constituents of knowledge but that the epistemic goodness of true belief and the epistemic goodness of justified belief are not constituents of the epistemic goodness of knowledge. And I do not here wish to commit myself to such an addition.
An organic unity, on one common construal (see Moore [1903] Reference Moore and Baldwin.1993, 82ff.; Chisholm, ch. 7; and Lemos, ch. 3), is a state whose goodness is greater than the goodness that obtains from the summation, or conjunction, of the goodnesses of its parts or constituents. The claim that something is an epistemic organic unity—an organic unity relative to the epistemic domain—then, can be glossed as the claim that its epistemic goodness is greater than the conjunction of the epistemic goodnesses of its constituents. Suppose we were to add to 1–10 the view that true belief and justified belief are the sole constituents of knowledge but that the epistemic goodness of true belief and the epistemic goodness of justified belief are not constituents of the epistemic goodness of knowledge. It would follow that knowledge is an epistemic organic unity. Still, the pluralism that consists of 1–10 alone needn’t commit us to the view that knowledge is an epistemic organic unity; such pluralism can be embraced even by those who, like Williamson, are inclined to see knowledge itself as a non-constitutive state and hence reject the suggestion that true belief and justified belief, whatever their epistemic value, are constituents of knowledge.
The pluralist position I have presented comes at a cost. It requires us to adopt a Williamson-inspired stance with respect to the epistemic value of knowledge: knowledge has a unique epistemic goodness that is to be explained in an expressly non-constitutive way. I say “Williamson-inspired,” not “Williamsonian,” because the attribution of this stance to Williamson is a matter of some interpretation, seeing that his own specific remarks about the value of knowledge (3.4) seem to pertain more strictly to the practical or psychological value of knowledge, not its epistemic value (cf. Sosa Reference Sosa2007, 71–2). It’s also worth nothing that there’s nothing about Williamson’s epistemology that commits him in any clear way to my pluralism. He is not, for example, obviously committed to true belief and justified belief as fundamental epistemic values along with knowledge.
I do not see that the cost in question is particularly high, however, since the Williamson-inspired stance is compatible with a host of distinct substantive theories of knowledge—even those that purport to offer necessary and sufficient conditions for knowledge. Suppose, for example, one were to follow Sosa (Reference Sosa2007) and take knowledge to obtain just in case apt belief—belief that is true because adroit (i.e., true because the result of the exercise of an intellectual virtue or competence)—obtains, where adroit belief is understood as justified belief in a broad sense. This substantive virtue epistemology can be added as is to 1–10 without any fear of inconsistency. The same can be said of an earlier, Goldman-style process reliabilism (1979, 1986), which takes justified belief to be belief that is the product of a reliable belief-forming process and knowledge to obtain if and only if true belief that is the product of such a process obtains. Even an account according to which S knows that p if and only if S has a justified true belief that p whose justification does not essentially depend on a false lemma (e.g., Harman)—indeed, even one that construes justification along coherentist or evidentialist lines—can be coherently conjoined with 1–10. That the cost of adopting my pluralist view of epistemic value is so slight comports with Neta’s observation that reliabilism per se “predicts nothing about what is good, epistemically or otherwise” (2) and echoes Fricker’s complaint that the way the problem of the value of knowledge is typically construed in the literature erroneously “encourages players to look for the value of knowledge in something that distinguishes their theory of knowledge from their competitors’ theories, when in fact the basic value of knowledge may be better explained by reference to something less epistemologically specific”(123).
Generalizing: A Sketch of an Ethical Value Pluralism
The plausibility of my pluralism would be enhanced by its generalizability. Does it present a picture of epistemic value that can, along its core dimensions, be extended to other sorts of value, in other domains of assessment? I think the answer is yes, and I will close by tentatively suggesting one area of extension. Consider the ethical domain—the domain of assessing action from the moral point of view. We might with some justification take consequentialist and deontological ethical theories to be picking out different fundamental ethical values, respectively: broadly speaking, the consequentialists are concerned with beneficial action, and the deontological theorists with well-intentioned action. The various arguments on either side against the other might further suggest that neither sort of action can be fully identified with yet another, greater ethical value, viz. morally right action. We would in such a case be led to ethical analogues of 1–5 above: morally right action, beneficial action, and well-intentioned action are all fundamental ethical values, but morally right action is nonetheless a greater ethical value than both beneficial action and well-intentioned action. An account explaining the link between the claims about the fundamental ethical values and the claims about one of them being a greater ethical value than the others could then be given in terms of 6 and analogues of 7–10:
1’ Morally right action is a fundamental ethical value.
2’ Beneficial action is a fundamental ethical value.
3’ Well-intentioned action is a fundamental ethical value.
4’ Morally right action is a greater ethical value than beneficial action.
5’ Morally right action is a greater ethical value than well-intentioned action.
6 If x and y are fundamental values in a domain of assessment δ, the goodness of x relative to δ entails that of y, and the goodness of y relative to δ does not entail that of x, then x is a greater value than y in δ.
7’ The ethical goodness of morally right action entails that of beneficial action.
8’ The ethical goodness of beneficial action does not entail that of morally right action.
9’ The ethical goodness of morally right action entails that of well-intentioned action.
10’ The ethical goodness of well-intentioned action does not entail that of morally right action.
Such a form of pluralism about ethical value, vaguely stated though it may be, seems quite plausible on the face of things (though I concede that the final verdict on that score awaits consideration by fair-minded metaethicists). It too will carry the cost of a commitment to a non-constitutive view of a greater—perhaps the greatest—fundamental value in the relevant domain of assessment. But I doubt that the cost will be any heavier for the ethicists than the analogous cost is for the epistemologists.
