Virtue epistemology and the acquisition of knowledge

Abstract

The recent literature on the theory of knowledge has taken a distinctive turn by focusing on the role of the cognitive and intellectual virtues in the acquisition of knowledge. The main contours and motivations for such virtue-theoretic accounts of knowledge are here sketched and it is argued that virtue epistemology in its most plausible form can be regarded as a refined form of reliabilism, and thus a variety of epistemic externalism. Moreover, it is claimed that there is strong empirical support in favour of the virtue epistemic position so understood, and an empirical study regarding the cognitive processes employed by medical experts in their diagnosis and treatment of epilepsy is cited in this regard. In general, it is argued that one can best account for ‘expert’ knowledge in terms of a virtue-theoretic epistemology that retains key reliabilist features. It is thus shown that understanding knowledge along virtue-theoretic lines has important implications for our understanding of how knowledge is acquired, and thus for the philosophy of education.

Acknowledgements

I am grateful to Nick Boreham, Michael Luntley, and Finn Spicer for assistance in the writing of this paper. Thanks also to The Leverhulme Trust for the award for a Special Research Fellowship which has enabled me to conduct work in this area.

Notes

1. For more on the externalism/internalism distinction in epistemology in general, see Kornblith (2001) and Bonjour and Sosa (2003).

2. The early Goldman view receives its most refined expression in Goldman (1986).

3. The issue of defeaters is one reason why the famous 'clairvoyant' counterexamples to reliabilism proposed by Bonjour (e.g. 1985, chap. 4) are problematic, as one typically is in the possession—or at least ought to be, if one is sufficiently careful in how one forms one's beliefs—of a standing defeater regarding the reliability of this belief-forming process. In particular, modern psychology is often characterised, especially in the public realm, in terms of an opposition to the supposed reliability of belief-forming processes of this sort. Once one notes the presence of the defeater, however, then the example starts to lose its dialectical force since it is typically part of the process reliabilist view, at least when refined, that an agent who does not suitably revise her belief in response to defeaters can as a result lack knowledge of what she believes even if the underlying process which gave rise to the belief was as a matter of fact reliable. See, for example, Goldman (1986, 62–63) and Greco (forthcoming). The key reliabilist point is that mere reliability alone can sometimes suffice for knowledge, not that it always suffices.

4. It isn't always clear when this case is being discussed just what the content of the target belief is supposed to be. Is it, for example, the belief that the chick which the agent is holding is of a certain sex? This is possible, though I don't think the example should be saddled with quite so specific a belief content, and the usual way that the chicken-sexer's abilities are described tends to count against such a description in any case. Instead it seems that what is at issue is the epistemic status of the agent's belief that the chick which she is holding is of a different sex to some other specified chick. Having the ability to reliably tell male and female chicks apart is one thing; having the ability to reliably tell them apart and identify which set of chicks is male and which is female is more demanding.

5. For further discussion of the chicken-sexer example, see Foley (1987, 168–69), Lewis (1996), Zagzebski (1996, sections 2.1, 4.1), and Brandom (1998). See also the exchange between Sainsbury (1996) and Wright (1996). Note that it is entirely consistent with the externalist view that it grants that there is something epistemically desirable about being in a position to offer adequate reflectively accessible grounds in favour of one's beliefs. The point is simply that such a reflective capacity is not necessary for knowledge possession.

6. Some have argued that identifying which belief-forming mechanism is in play is itself a near-impossible task. This is the so-called 'generality' problem (see, for example, Conee and Feldman 1988). To my mind, however, this problem has been overstated. While it can obviously be sometimes very difficult to determine empirically which belief-forming mechanism an agent is employing, it is far from impossible (indeed, the cognitive science literature is full of studies which identify belief-forming mechanisms). For a subtle discussion of how reliabilists might meet the generality problem, see Goldman (1993).

7. Of course, it may be that the relevant process is one that the agent cannot master straight away, perhaps because it requires special training or abilities. In such cases, it may be that the optimum process in terms of reliability for this agent at this time is very different to the optimum process simpliciter. I return to this point below when I discuss 'expert' knowledge.

8. It is actually just Greco who describes his view as an agent reliabilist thesis, but the similarities between his view and Sosa's are strong enough to count them both as falling under this description for our purposes. See also the 'proper functionalist' theory presented by Plantinga (1993), which bears a number of key similarities to the basic agent reliabilist thesis. It should also be noted that Goldman now presents his version of reliabilism in a way that is broadly in line with agent reliabilism. See, for example, Goldman (1993).

9. For further discussion of these empirical studies, see Boreham (1994).

10. For more on this point, see Spicer (2004). For a useful survey of the main a priori and empirical considerations which count against internalist epistemologies, see Goldman (1999).

11. There is a third option in this respect, which is to allow that there might be two types of knowledge—one which is 'brute' and unreflective, and another which is reflective. This type of position is most usually associated with the work of Sosa (1995, 1997). For discussion, see Grimm (2001; cf. Sosa 2001) and Greco (2002, 298–301). For an overview of the two types of virtue epistemology—the early and broadly reliabilist and externalist view on the one hand, and the later and broadly responsibilist and internalist view on the other—see Axtell (1997) and Brady and Pritchard (2003).

12. See Sosa (1991, 271), for an explicit defence of the more inclusive conception of 'virtue'.

13. See also Boreham (1995) and Boreham, Foster, and Mawer (2000).

14. There is actually a further complication in this regard that we do not need to go into, which is that half of the final-year medical students had played a game called 'The Phenytoin Game' which is designed to aid correct diagnosis of the dosage of phenytoin needed. While playing this game didn't ensure that the students made the correct diagnosis, it did have an effect in that students who had played the game yet made sub-optimal diagnoses tended to opt for overly cautious, rather than hazardous, doses of the drug.

15. Note, however, that it is not immediately obvious that one should teach the more subtle rules followed by experts to medical students, since only the experts have the wealth of background knowledge available to them that can allow them to utilise these rules effectively. Indeed, the 'cautious' rules already followed by many final-year medical students in this regard may be better rules to follow to begin with. At the very least, though, it is useful to make this tacit knowledge explicit so that it can guide research and inform the education of medical students.

16. For more on the relationship between virtue epistemology and the issue of epistemic luck—i.e. the sense in which knowledge is non-lucky true belief—see the symposium between Axtell (2003), Greco (2003), and Pritchard (2003). For more on the topic of epistemic luck more generally, see Pritchard (2005). Recent studies, both in philosophy and cognitive psychology, have emphasised the importance of luck in understanding agent responsibility, whether cognitive or otherwise. For a critical survey of this literature, see Pritchard and Smith (2004).

17. Luntley discusses the status of expert knowledge in a number of works. See, for example, Luntley (2002). In general, his position in this regard relates to a novel theory of concept possession. I think that this project shares several key themes with the agent reliabilist conception of knowledge, though to argue for this in detail is a task for another occasion.

18. This is not to suggest, of course, that there are no problems of substance facing the agent reliabilist view, only that this thesis represents the most satisfactory theory of knowledge currently available. Pritchard (2003) discusses one key problem for virtue epistemology which concerns how strong the thesis should be formulated.