We wrap up the current series of posts on justified beliefs by considering Ernest Sosa’s theory known as ‘virtue perspectivism’. Broadly speaking, this is a position that seeks to combine internalist and externalist positions into a single account of justified belief.
Sosa starts by noting the general position of process reliabilism and the three common criticisms that arise: (1) the generality problem (processes must not be defined either too broadly or narrowly), (2) the new evil demon problem (reliabilism cannot account for the victim of the evil demon, who follows a reliable process but still does not have knowledge), and (3) the meta-incoherence problem (the subject has followed a reliable belief but still does not consider the belief to be justified, such as the genuine clairvoyant who dismisses her own ability). In his response, Sosa does not address the generality problem, but he does try to explain how his theory deals with the other two problems.
Similar to Steup’s internalist theory, Sosa’s theory relies on the idea of an ‘intellectual virtue’. This is an ability to accept a belief when the evidence is good, but dismiss the belief if the evidence is not reliable. Similar to moral virtues, holding an intellectual virtue leads the subject to make more correct decisions. Examples of intellectual virtues are vision, hearing, introspection and deduction. The intellectual virtues are heavily dependent on their context, such as vision only being appropriate when the light is good. According to Sosa, a justified belief is one that was formed by an intellectual virtue in the right context.
By defining the virtue to be context-dependent, this opens the way for Sosa to address the evil demon problem. When we look at the ‘victim’ in the evil demon world, we see someone who seems to have all the usual intellectual virtues (eyesight, thought, etc), but is unaware of the fact that these virtues are being fed by a third party. However, the virtues are still appropriate for the ‘normal world’ scenario. The subject is merely unlucky by finding herself in an evil demon world. Therefore, Sosa suggests that if the subject’s intellectual virtues would be entirely appropriate in the normal world (and produce more true beliefs than false), then the virtues are reliable in the evil demon world as well, so the subject can be internally justified, even if they don’t have knowledge as many of the beliefs are not actually true. Goldman proposed something similar with his ‘normal worlds’ theory (we should assess the reliability of a process based on how it works in the normal world).
In terms of the meta-incoherence problem, Sosa notes that the main requirement is that the subject has acquired the knowledge on the basis of an intellectual virtue. It does not matter if the subject himself thinks that the knowledge was unreliable. Therefore, similar to process reliabilism, this part of Sosa’s theory is external. The subject does not need to have any knowledge that their belief was acquired by a reliable intellectual virtue. However, Sosa is about to argue that in terms of reflective knowledge, acquiring the knowledge via an intellectual virtue alone is not sufficient.
One of the dilemmas for any theory of knowledge is how to explain the idea of animal knowledge. Animals and small children can definitely have knowledge, even if they cannot justify their ideas. How does the internalist deal with this scenario? Sosa notes that animal knowledge is simply acquiring a belief via the intellectual virtues, so is purely a matter of external justification. On the other hand, reflective knowledge requires both the intellectual virtue (as per animal knowledge) along with a second requirement: the new belief fits coherently into the subject’s existing set of beliefs. Sosa describes these two requirements as first order externalist aptness and second order internalist justification. In other words, this is Sosa’s attempt to merge both external and internal views into a single theory.
Referring again to the meta-incoherence problem – and the sceptical clairvoyant who obtains knowledge – Sosa would argue that the clairvoyant has animal knowledge (the knowledge was obtained via an apt process), however the clairvoyant would lack reflective knowledge, as they don’t have the belief themselves that their knowledge fits into their existing set of beliefs. On the contrary, the clairvoyant is sceptical that the knowledge was acquired reliably.
One question that Sosa doesn’t address is whether his theory deals with the Gettier cases, such as the fake barn scenario. Under Sosa’s theory, someone could have acquired knowledge via an apt process, and the subject believes the knowledge fits in with their existing set of beliefs, however we still don’t accept that this is knowledge. In other words, the idea of reflective knowledge alone does not have a sufficient connection to truth. Having a coherent belief does not mean that the belief is actually true. One counter-response to this is that the process to acquire knowledge wasn’t apt to begin with, as it resulted in false knowledge. Therefore, we could reject the fake barns as being an example of knowledge.