2. The epistemic domain

The locus of the debate, to which this paper is a contribution, is the domain of epistemic evaluation and normativity. This is the domain where we find evaluations of individual beliefs, belief-sets, methods, agents, traits, and so on as epistemically good or bad, epistemically right or wrong, epistemically justified or unjustified, and where we find considerations of what one should believe or ought to do, epistemically speaking. A common assumption is that normative domains are structured in the sense that most of the normative properties can be explained in terms of some smaller set of basic normative properties. Think of the moral domain. According to some theorists, the most fundamental thing in that domain is some rule or norm, and other normative properties such as virtuousness and goodness are explained in terms of this, e.g. it is good to φ because φ-ing is in accordance with a rule that no one could reasonably reject. According to other theorists, the most fundamental thing in that domain is goodness, and other normative properties such as moral rightness and moral virtuousness are explained in terms of goodness, e.g. it is morally right to φ because φ-ing stands in some appropriate relation to goodness. The latter is a value-based conception of the moral domain. In the epistemic domain, the corresponding view takes epistemic value to be explanatorily prior to other epistemic normative properties, so that e.g. S's belief that p is justified because it (has some property that) stands in some appropriate relation to epistemic goodness. To conceive of the epistemic domain as value-based is common in the contemporary debate on epistemic normativity.Footnote ¹

The most central debate in contemporary epistemic axiology concerns the question of what has fundamental epistemic value. The idea here is that while lots of things are epistemically good, only a select few have fundamental epistemic value. The task of whatever is postulated at the level of fundamental epistemic value is to explain this rich array of epistemic goodness, in the sense that all epistemic goodness on the non-fundamental level must be derived from the goodness at the fundamental level. Monists think there is only one good on the fundamental level, whereas pluralists concede that there are at least two distinct fundamental epistemic goods. The arguably most common view in epistemic axiology is veritistic monism, or veritism for short, according to which only truth (or more generally, accuracy) has fundamental epistemic value.Footnote ² Veritists are committed to explaining the epistemic goodness of rationality, knowledge, understanding, curiosity, coherence, and whatever else we rightly consider epistemically valuable in terms of the value of truth. Differently put, according to veritists, all epistemically good things derive their epistemic value from the value of truth, through some relation or other. Given a value-based conception of the epistemic domain – which veritists endorse – truth is the most fundamental explainer in the domain of epistemic normativity.

Whatever one takes to be basic in the epistemic domain, there are at least two different ways to conceive of it. I will focus here on epistemic value, and on veritism in particular, but both positions could be occupied by those who hold different views. Consider the veritist claim:

Such a claim admits of an attributive and a predicative reading. As Ridge (Reference Ridge2013) points out, we can use Geach's (Reference Geach1956) distinction between attributive and predicative uses of ‘good’ in order to distinguish these two views on epistemic value. The general idea here is that modifying adjectives can be used in different ways, and this is relevant for what a locution involving said modifying adjectives entail. For instance, compare ‘x is a grey elephant’ and ‘x is a small elephant’. The former locution entails that x is both an elephant and grey. This is an example of a predicative use of the modifier ‘grey’. The latter locution, in contrast, does not entail that x is small. That's an example of an attributive use of the modifier ‘small’. Ridge extends Geach's distinction to locutions of the form ‘is F-ly G’, which count as predicative just in case they entail ‘is G’, and attributive if not. For instance, ‘is brightly coloured’ entails ‘is coloured’, but ‘is superficially reasonable’ does not entail ‘is reasonable’.

Now, we can read ‘epistemically’ in a claim like V so that it entails that truth has genuine value, or so that it entails no such thing. Theorists who are inclined towards the first, predicative reading, may be so because they think of epistemic value as a sui generis subcategory of value simpliciter – much like how many think of moral and/or eudaimonic value. But it can also be because they think that epistemic value is necessarily connected to some kind of genuine value. For instance, some think that epistemic goods are constitutively valuable in virtue of being a necessary part of the good life, which has eudaimonic value (see e.g. Lynch Reference Lynch2004 for an argument to this effect). Of course, on this kind of view, theorising about epistemic axiology, including defending a position like veritism, requires engaging with the ways in which epistemic value interacts with and relates to non-epistemic value.

This is in contrast to how theorists who prefer the second, attributive reading of ‘epistemically’ conceive of questions about epistemic value. Such theorists endorse what I call insulationism about epistemic value. We may say that they think of epistemic value as merely formal value. That term is not widely used, but the corresponding notion of formal normativity (in contrast to robust or genuine or categorical normativity) is familiar from metaethics and metanormativity debates more generally (e.g. Baker Reference Baker2017; Finlay Reference Finlay, Plunkett, Shapiro and Toh2019; Maguire and Woods Reference Maguire and Woods2020; Wodak Reference Wodak2019). It is, of course, open to these veritists to argue that truth also has genuine value. But notably, this will not be in virtue of it having epistemic value.

The view that epistemic value is merely formal value has become increasingly popular in recent years.Footnote ³ The perhaps most influential proponent of this approach is Ernest Sosa, who memorably wrote that

He suggests that each of these are examples of normative domains, and each of them have some kind of internal structure along the lines suggested in the first paragraph of this section: they involve norms and values, some fundamental and the rest derivative. In value-based domains, there will be some fundamental, domain-specific good(s), in relation to which things are evaluated from the perspective of the domain. To illustrate, consider espionage. The fundamental good around which evaluations in that domain is centred is plausibly clandestine collection of confidential information. It is as derivative relative to this that the goodness or rightness, from the point of view of espionage, of anything else in the domain – spies, operations, organisations, agent handling practices, norms for how to behave when one is under cover – is to be understood. The espionage domain is a plausible example of an insulated domain: its existence and the evaluative and normative practices within it leave open whether clandestine collection of confidential information has any value simpliciter, and hence also whether anything that derives its espionage value through a relation to such information collection has ‘any real value’ (Sosa Reference Sosa2007: 73). Differently put, that x is good from the espionage point of view does not entail that x is good, and espionage value is only formal value.

A popular way (not the only one, although it is arguably Sosa's way) to flesh out the insulationist idea is to think of epistemic value as having a teleological structure: a spying operation is good insofar as it fulfils the function of a spying operation, i.e. collects intelligence without detection. We may then say that it has ‘spy value’, but from that we can't conclude anything about what other value it might have, including nothing about its real value (if there is such a thing). Similarly, a belief (or reasoning procedure) is good to the extent that it fulfils whatever function that a belief (or a reasoning procedure) is supposed to have, i.e. being true (or e.g. being truth conducive) – if we are veritists. Spying operations, beliefs, and reasoning procedures, like archers and assassins, are good to the extent that they perform well relative to achieving their formal ends. But it remains an open question whether those formal ends have value simpliciter.

Several prominent veritists are insulationists in the above sense (Hazlett Reference Hazlett2013; Pritchard Reference Pritchard, Pritchard, Millar and Haddock2010, Reference Pritchard, Matheson and Vitz2014; Sylvan Reference Sylvan2018, Reference Sylvan2020), and hence read V as not informing us about whether or not truth has any real value. Pritchard, for instance, writes that on his view, V is ‘entirely consistent with arguing that the general value of truth is merely instrumental (or indeed, that it has no value at all)’ (Reference Pritchard, Matheson and Vitz2014: 113).Footnote ⁴

3. A virtue-first approach?

A value-based, veritist approach to the epistemic domain is commonplace. Many would arguably agree that if there is a default, or at least a majority view, such veritism is it – indeed, it has often been assumed to be correct without further argument, following from the definition of ‘the epistemic’ (cf. David Reference David and Steup2001; Engel Reference Engel, Coliva and Pedersen2017; Pritchard Reference Pritchard, Matheson and Vitz2014: 112). That said, the tenability of veritism has been one of the main topics of discussion in recent debate over epistemic value. Although it has many prominent defenders, it has also come under fire from epistemic value pluralists who argue that truth cannot be the only fundamental epistemic good (DePaul Reference DePaul and Steup2001; Gardiner Reference Gardiner2012; Kelp Reference Kelp2021; Madison Reference Madison2014; Pedersen Reference Pedersen, Coliva and Pedersen2017, Reference Pedersen, Graham and Pedersen2020; Ranalli Reference Ranalli2021) and from other epistemic value monists who argue that the sole fundamental epistemic good is not truth but something else, such as knowledge (Aschliman Reference Aschliman2020; Millar Reference Millar2011) or evidential support (Bjelde Reference Bjelde2021).

Veritists attempt to defend themselves from these attacks in various ways. One particular kind of defence proceeds from the complaint that veritism has been construed in the wrong way by its critics (and sometimes by its proponents), and that the most challenging objections can be met once we construe the position differently. The virtue-first approach, recently proposed by Pritchard (Reference Pritchard2021a, Reference Pritchard2021b, Reference Pritchard2021c, Reference Pritchard2021d) is an example of that strategy.Footnote ⁵ The suggestion at issue here is that we should approach questions in epistemic axiology on a virtue-first basis. In particular, Pritchard claims that such an approach is the key to rendering veritism truly attractive.

The core idea is that we should ‘take the notion of an intellectual virtue as primary, and then understand claims about epistemic value in terms of this notion’ (Pritchard Reference Pritchard2021b: 5522) (see also Reference Pritchard2021d: 1354). It is not entirely clear what questions the virtue-first approach is supposed to help us answer, but one key question that we obviously want an answer to – and to which veritism is one proposed answer – is: what has fundamental epistemic value? I am going to assume for the sake of argument that the virtue-first approach is supposed to help us answer this question.

It is clear that Pritchard thinks the virtue-first approach is also supposed to elucidate the veritist position, i.e. tell us what it means to say that truth has fundamental epistemic value: ‘rather than trying to determine what is involved in treating the truth as a fundamental good in the abstract (…) we should instead let it be determined by considering what an intellectually virtuous inquirer values’ (Reference Pritchard2022: 12). Indeed, this is crucial to his argument for the claim that veritism is in fact not vulnerable to certain classical objections.Footnote ⁶ There is the possibility that such elucidation of the veritist position is all the virtue-first approach is supposed to do for us, i.e. that it is not supposed to also offer a direct argument for veritism. This would make the approach less interesting and powerful, as it would require us to assume from the outset that veritism is true – an increasingly controversial assumption. I will return to this fallback position towards the end of the paper, but my focus is on the idea that a virtue-first approach to epistemic value is supposed to help resolve the question of what has fundamental epistemic value in favour of veritism.

4. Intellectual virtue: insulated or not?

A question which presents itself in relation to value from virtue is whether we are to read ‘intellectual virtue’ as an insulationist normative notion or not. That is, are we supposed to take claims like ‘S behaves intellectually virtuously’ and ‘T is an intellectual virtue’ to licence conclusions like ‘S behaves genuinely virtuously’ and ‘T is a real, genuine virtue’, or do such inferences require independent argument? I will explore both options here and argue that neither will make value from virtue into an argument for veritism in the insulationist sense.

Let's first consider value from virtue under a predicative reading of ‘intellectual virtue’, where intellectual virtues really are genuine virtues. This is how neo-Aristotelian virtue epistemologists tend to understand intellectual virtue, as relevant to being a good person (Baehr Reference Baehr2011) or leading a good life (Zagzebski Reference Zagzebski1996).Footnote ¹¹ One may of course question why this particular evaluative epistemic concept should receive a different treatment from e.g. epistemic goodness or epistemic rightness, if one is otherwise an insulationist, but I will set that issue aside here because there is a more pressing problem with methodological value from virtue on this reading. The problem, basically, is that P1 now equivocates between formal and real normative notions. Properly spelled out, P1 on this reading states:

This reading of P1 is not supported by the neo-Aristotelian conception of virtue, which connects virtue with genuine, not merely formal, value. Moreover, by the insulationists’ own lights, P1 is false on this reading: one cannot draw conclusions about what has epistemic value in the insulated sense from truths about what has genuine value.

Let me briefly explicate this. First, it is true that intellectual virtues in the genuine sense are intentionally directed at what is good in itself, on the neo-Aristotelian picture. Moreover, virtue theorists in the neo-Aristotelian tradition often distinguish between epistemic and moral virtues on the basis of their intentional objects (Baehr Reference Baehr2011; Wilson Reference Wilson2017; Zagzebski Reference Zagzebski1996). That is, intellectual virtues are traits that are intentionally directed in the appropriate way towards specifically epistemic goods for their own sake. This seems initially promising for the current approach: if we can identify the intellectually virtuous, we can get at the epistemic good(s) by examining what they are finally motivated by. But of course, ‘epistemic goods’ in this sense are a subset of the genuine, real goods. These may or may not coincide with what is epistemically good in a formal, insulated sense, but that they do is not part of the neo-Aristotelian conception of intellectual virtue.

Second, insulationism about epistemic value explicitly denies an automatic connection between epistemic value and genuine value. And this insulationism works in two directions. As we have already seen, it does not follow that x has genuine value from the fact that x has epistemic value. But neither does it follow that x has epistemic value from the fact that is an epistemic kind of thing that has some genuine value (Pritchard Reference Pritchard2017: 1482). Let's consider a toy example to illustrate this. A research method is arguably an epistemic kind of thing, insofar as that kind of classification makes sense. If this particular research method – which we may stipulate is very unlikely to generate any true beliefs and fails to respect various norms of rationality – nevertheless makes someone very happy, and let's assume that pleasure is genuinely valuable in itself. This entails that the research method has real (if derivative) value. But it does not follow that the method has epistemic value, in the insulationist sense. And the same reasoning applies to the intentional objects of intellectual virtues. Genuine virtue, viz. intellectual virtue on the predicative reading, is motivationally directed at things that are good in themselves and that are furthermore ‘intellectual’ or ‘epistemic’ in nature rather than, say, moral. But even if the intentional objects of intellectual virtues are both valuable and in the category of things and properties we call ‘epistemic’, it does not follow that they are epistemic goods in the insulationist sense. In short, insulationism blocks precisely the kind of inference from intellectual virtue in the non-insulated sense to conclusions about epistemic value in the insulated sense, that P1 on this reading suggests.

What about the other option? Reading intellectual virtue as insulated in the same sense as epistemically good, justified, and so on, would seem to be the most natural thing for an insulationist to do. But it is doubtful whether insulated virtues can generally realise the motivations-involving, neo-Aristotelian conception of virtue. And if they cannot, examining them cannot tell us about what has fundamental epistemic value.

Intellectual virtue in the insulated sense does not, without further argument, contribute to someone's being a good person or leading a good life. What might this then amount to? A plausible gloss might be in terms of being an excellent enquirer (or believer, perhaps). Chess is a favourite analogy that insulationists reach for when they want to illustrate their view, so let's start with the idea of an excellent chess player. Presumably, an excellent chess player is someone who is really good at chess, which involves regularly and reliably succeeding in reaching or approximating the constitutive aim of the activity, i.e. someone who tends to checkmate (or get close to doing so) her opponents when she plays. Indeed, checkmating one's opponent is said to be the fundamental good of the domain of chess evaluation. We might want to qualify this in certain ways, e.g. her winning cannot be the result of cheating or unbelievable luck – it must somehow be attributable to her skill as a player. Now, being an excellent chess player in this sense is arguably quite irrelevant to whether one is a good person in a moral or prudential sense – it neither adds nor detracts from it. Or at least we cannot say that it does without further argument, and our idea of what a good chess player is, is independent of any such arguments.

With this domain-specific, insulated sense of agential excellence in hand we can construct an analogous notion of ‘insulated’ intellectual virtue, in terms of the excellent enquirer: someone who, when she enquires, regularly succeeds in realising the constitutive aim of that activity, i.e. she ends up believing the truth. Moreover, this is not the result of a cosmic coincidence but attributable to her inquisitive skills and she enquires in line with generally approved epistemic norms. It is not wholly implausible that this could be irrelevant to whether or not she is a good person, morally or prudentially. She is just really good at enquiring. For the veritist, this means that she typically succeeds in getting (close) to the truth when she attempts to form beliefs.

I have nothing against this artificial conception of domain-specific virtue as such, but I doubt that it can support an argument like value from virtue. Consider the following two cases.

I propose that Alison is an excellent enquirer, and that Carrie is an excellent chess player. To deny this seems preposterous: they are both widely recognised as world-leading at what they are doing. Insofar as we can construct an insulationist notion of virtue built around domain-specific excellence, they are both virtuous in the insulated epistemic and chess sense, respectively.

If the virtue-first approach is on the right track, we should be able to study Carrie and Alison in order to find out what the fundamental goods of their respective domains are. In particular, we should look into what the final aims of their enquiries and chess activities are. But note that, while there is a sense in which Carrie is committed to checkmating her opponents and Alison is committed to epistemic goods, they are not committed to these things for their own sake. Carrie is committed to checkmating her opponents as a means to a non-chess-related end. Similarly, Alison does not care about the important truths of astrophysics for their own sake – she is committed to truth in her enquiries as a crucial means to academic glory and admiration. The upshot is that even if we do have an ability to identify domain-specifically virtuous individuals independently of knowing what the aims of the domains they are excellent with respect to are, we cannot hope to learn about the fundamental goods of these domains by looking at what these agents are motivationally oriented towards for its own sake.

I expect some will want to push back here, saying that while Alison and Carrie are excellent at what they do, they are not really virtuous even in the domain-specific sense. To really be intellectually (or chess) virtuous, one really does need to love epistemic goods (or checkmating one's opponent) for their own sake, and one's excellency really must arise from that motivational orientation. But what motivates the insistence that virtue really requires the right motivation? I agree that we might well feel the pull to resist calling Alison and Carrie virtuous despite their excellence at what they do, but I also submit that the most plausible source of this pull is our idea of genuine virtue, i.e. the kind of traits relevant to being a good person or living a good life. We might want to say that a virtuous chess player should also have certain attitudes, e.g. respect others’ chess-related achievements, feel proper pride at their own chess-related accomplishments, and fostering a sense of fair play in younger players. And we might want to require corresponding things of a truly virtuous enquirer. But are we then not thinking about how a genuinely virtuous person would behave in chess-related or enquiry-related circumstances, rather than about a domain-specific sense of virtue? Recall that the domain-specific notion of virtue is, by insulationism, supposed to be completely orthogonal – without further argument, at least – to whether one is a good person or leads a good life. But if someone behaves in these ways with respect to chess or enquiry that does arguably contribute to their moral and/or eudaimonic status without further argument. It might not be enough on its own to make them good people or constitute eudaimonia, and it may be overridden by how they feel and behave more generally, but it is hardly irrelevant. Unlike the ‘pure’, viz. insulated, excellence which Carrie and Alison both possess.

A different way to make the same point is to consider versions of Award-Winning Astrophysics and Cynical Chess where Carrie and Alison are ultimately committed to non-epistemic things that have more plausibility as candidate simpliciter goods (which one may well doubt that fame, fortune, and academic glory are), such as pleasure or achievement. Let's assume that Alison is motivated by her wish to take pride in true achievement and reconstruct the case accordingly. We may now be more inclined to consider her virtuous, but this does not help the insulationist. It is still not the case that we can find out about the fundamental epistemic good(s) from enquiring into what she values for its own sake. While the prospective final value simpliciter of an agent's ultimate ends might be relevant to whether or not she is virtuous in a non-insulated sense, the very point of insulationism is to not commit to any claims about what is valuable simpliciter, so it should not matter to questions of insulated intellectual virtue. What we need in order to vindicate P2 is not that an epistemic good like truth is valuable for its own sake, but that an excellent enquirer qua excellent enquirer is motivationally oriented towards truth for its own sake. But the cases show that they need not be.

What cases like Award-Winning Astrophysics show is that, as some epistemologists of a pragmatist bent are fond of pointing out, the ends of well-conducted enquiry are often non-epistemic, and their epistemic end-results (truth, knowledge) are valued only instrumentally or constitutively to those ends (Crerar Reference Crerar2021; Hookway Reference Hookway, Zagzebski and DePaul2003; Olson Reference Olson2012). In a sense, this observation is friendly to insulationism: the very point of it is that we are supposed to be able to judge things as good or bad in the epistemic domain without taking a stance on what is good or bad outside of that domain. People can have all sorts of bad ends and personal traits, and still be good at enquiring (or chess). But the other side of the coin is that this also seems to make the ultimate motivations of enquirers irrelevant to what is good or bad in the epistemic domain. Hence the observation is unfriendly to the idea, underlying the virtue-first approach to epistemic value, that we can find out about what is of epistemic value by considering what the intellectually virtuous – in the insulated sense of being an excellent enquirer – value for its own sake.

The upshot is that there is no way for an insulationist about epistemic value to employ the methodological value from virtue argument. Either the notion of virtue in question is insulated, in which case virtue does not necessarily involve a commitment to any epistemic goods for their own sake. But then we cannot look to what the intellectually virtuous (qua excellent enquirers) value for its own sake, in order to find out what is epistemically valuable. Or the notion of virtue is not insulated, in which case it does necessarily involve a proper commitment to things that are in some sense epistemic and good for their own sake, but not to epistemic goods in the insulated sense.

I want to end by considering what I earlier referred to as a fallback position. As we have seen, some philosophers think that things which are good in themselves can demand different kinds of responses: it may be proper to seek to maximise and bring about some goods, but other goods rather demand e.g. preservation or respect. Studying how the virtuous treat epistemic goods will teach us how those goods ought to be treated as the goods they are. Pritchard clearly thinks that the virtue-first approach will do this for us (i.e. teach us about the right way to value truth) but the fallback position is that this is all it does. Since it turned out that a stronger virtue-first approach is incompatible with insulationism, what about this weaker view?

In fact, I think the reasoning above applies here too. First, if we study how intellectually virtuous people in the non-insulated sense treat e.g. truth, it is not obvious that this teaches us anything about what it means for truth to be epistemically valuable in the insulated sense.Footnote ¹³ Second, if we instead study excellent enquirers, viz. intellectually virtuous agents in the insulated sense, it is not clear that we will find that they treat truth in any particularly unified way. To be sure, they will all be disposed to enquire in ways that are conducive to getting to truth, but this says nothing about what they respect, desire, praise, take joy in, celebrate, foster in others, love, prize, and so on, nor why they are disposed to enquire in those ways. As noted above, many different motivational profiles seem compatible with being an excellent enquirer in the insulated sense.