The relationship between art and epistemology has been forever tenuous and fraught with much debate. It seems fairly obvious that we gain something meaningful from experiences and interactions with works of art. It does not seem so obvious whether or not the experiences we have with art can produce propositional knowledge that is constituted by true justified belief. In what follows I will give some historical background on the debate and flesh out some of the important issues surrounding the question "(What) can we learn from art?"
Continuing with the line of argument Aristotle began, all the way through the Renaissance and beyond, philosophers have defended the notion that we can learn from art, and that poetry and fiction engage the emotions in a helpful, rather than detrimental, way. The Romantics dealt with this question in a manner that the earlier rationalists and empiricists did not. The rationalists rejected the idea that the imagination could be considered a source of knowledge, with Descartes going so far as to dismiss what he called "the blundering constructions of the imagination." Returning to the ideals of Plato, the rationalists strictly employed a knowledge requirement involving justified true belief. Empiricist epistemology too is particularly unhelpful when it comes to explaining how we might gain justified knowledge from fictional or representational situations. For it seems impossible to learn actual things from fictional situations.
The Romantics provided the real beginnings of an argument against the passive accounts of knowledge for which the empiricists argued. Romantic epistemology emphasizes the role of the imagination in addition to (or over) reason. This allowed for the notion that there is not merely one right way to know, and that there is not only one right way to view, experience, understand, and construct the world.
The Romantics adopted three main tenets concerning the relationship between literature (and art more generally) and truth. The first denied that there is any one point of view from which Truth can be determined. The second began to question the Augustinian conviction that art and literature, like science, should concern only general features of nature. The third tenet, which the Romantics developed more fully, concerned the notion of transcendence, especially in association with growth. Natural science is able to describe the physical world, but only from a single point of view (Harrison 1998). Art and literature can describe the world in a myriad of ways, transcending experience of the physical world into the emotional and even the supernatural. Although art does not record truths about the world in the same way that science does, it can give insight into the different ways that we understand the world and with different degrees of accuracy. It is those degrees of accuracy that continue to be called into question.
David Novitz (1998) points out that there are three basic kinds of knowledge claims we can make about the arts, all of which are distinguished by their objects. The first concerns what we claim to know or believe about the art object itself and whatever imaginary or fictional worlds might be connected to that object. For example, I can claim to know things about the way the light reflects in Monet's Water Lilies. I can also claim to know things about Anna Karenina's relationships with her husband and with her lover, Vronsky. Beyond this, we may feel justified in our pity for Anna, because of the way Tolstoy's novel presents her story. Can my knowledge of Anna be meaningful, however, or be considered knowledge at all in the traditional sense (justified true belief) if Anna Karenina is a non-referring name? Further, how can one's interpretation of her situation be any more legitimate than anyone else's? Can single interpretations hold value over time and across cultures? Without the propositional content used to legitimize the standard analysis of knowledge, it seems that the knowledge claims we have about the content of an artwork will never have the same kind of validity. Whether or not that same kind of validity is required also needs to be called into question.
The second kind of knowledge claim we can make about art concerns what we know or believe to be an appropriate or warranted emotional response to the artwork. We often believe that works of art are only properly understood if we have a certain kind of emotional response to them. One problem here, of course, concerns how it is that we know what kind of response is appropriate to a particular work. On occasion we talk to friends about a response they had to a particular work of art that was manifestly different from the one we had. How is it possible to judge which response is more appropriate or justified? Even suggesting that one should respond as if a novel, for example, were to be taken as an account of true events, with responses following as if the events depicted therein were actually happening or had happened, does not solve the problem. For one thing, not all emotional responses to real events are taken as equally justified. For another, most novels are not meant to be taken as true (despite the "report model" of emotive response [see Matravers 1997]). The fact that we do respond emotively to art, and to fiction in particular, would seem to indicate that there is something in the artwork that is worth responding to, even if it is not the same thing possessed by the objects we respond to outside the art world.
The third kind of knowledge claim we can have about art concerns the sort of information art can provide about the world. That is, how is it that we can gain real knowledge from fictional or non-real events or activities? It is widely accepted that art does, in fact, convey important insight into the way we order and understand the world. It is also widely acknowledged that art gives a certain degree of meaning to our lives. Art, and literature in particular, can elicit new beliefs and even new knowledge about the world. But the concern is this: fiction is not produced in a way that is reflective of the world as it actually is. It might be quite dangerous, in fact, for one to obtain knowledge about human affairs only from fiction. For example, it could be downright unhealthy for me to get my sense of what it is like to be in love from romance novels alone.
We can easily be experientially misled by art. The so-called empathic beliefs, those we gain from experiencing art, should be based on and enhanced by our broader experience of the world and should not arise independently of our other beliefs. But here the problem of justification returns. That is, if the empathic beliefs we gain from our experience of art actually coincide with our experience of the real world, then they can pass as empathic knowledge (i.e., beliefs become true and justified when they are connected to other justified beliefs). The problem is that often the emotions and beliefs that we adopt empathically turn out to be temporary, since they are not grounded in concrete experience. Can the experience we have with a work of art be confirming in and of itself, or must there be another, external authority to make the experience, or at least the knowledge gained from the experience, legitimate? It seems that much of what we learn about the world does come from art, and thus the justificatory claims to knowledge must be reconsidered.
The propositional theory of knowledge holds that one must have justified true belief in the content of a proposition in order to have knowledge. This appears reasonable under normal circumstances, but seems not to work at all in the case of art. It seems odd, in fact, to hold that in order to show that one has learned from a work of fiction, one must show that the work has propositional content of a general or philosophical nature, or that it provides experience that cannot be gained in any other way. If we can learn from art, we must be able to do so in a manner that diverges from the traditional notion of justified true belief, but that still holds some sort of legitimate ground.
What kind of justification is needed to ground these potential knowledge claims that art provides? First of all, we must be at least somewhat aware of what the new knowledge consists of. Moreover, one's engagement with the artwork should provide at least some degree of justification (e.g., I feel pity for Anna Karenina because she is in an unfortunate set of circumstances that she feels she has no control over. I am justified in my emotional response to her if I can see that she is in a truly pitiable situation). It is important to distinguish learning from art from merely being affected or influenced by it, or even from being challenged by it. Accounts of knowledge provided by art should be able to identify clearly what it is about the artwork itself, qua artwork, which prompts knowledge. A cognitivist account in particular will require first that the content of the work be specifiable (what is it we learn?); second, that the demands for justification be respected; and third, that these accounts appeal directly to aesthetic experience (Freeland 1997).