Elisabeth Schellekens is Chair Professor of Aesthetics at the Department of Philosophy at Uppsala University and Vice-President of the Nordic Society for Aesthetics. Her research interests include aesthetic normativity and obligation, the relations between aesthetic, moral and cognitive value in art, non-perceptual art and theories of perception, the philosophy of cultural heritage, and more.
This week at The Junkyard, we're hosting a symposium on Greg Currie’s recent book: Imagining and Knowing: The Shape of Fiction. See here for an introduction from Greg. Commentaries and replies appear Tuesday through Thursday.
* * *
Commentary from Elisabeth Schellekens Dammann.
In many respects, some of the main philosophical achievements of the intricate and important work done in Imagining and Knowing lie in what it does not set out to do. It does not for example seek to define the notion of fiction, nor to provide conclusive evidence for the alleged relation between fiction and the development of empathic skills. It does not present a decisive case against the claim which constitutes its central target, namely that fiction can yield knowledge or understanding (hereafter Cognitivism), nor does it settle for it. What it does do however, is to carve out a distinct set of pressing concerns by developing this one striking point: if we grant that the imagination is central to our experience of fiction – as we ought to – then the numerous (hypothetical) connections which supporters of Cognitivism have relied on between works of fiction on the one hand and learning, truth and knowledge on the other can only be described as shaky. At best.[1] Where, we are repeatedly asked, is the empirical evidence for positing these connections?
Imagining and Knowing is a cautionary, at times even admonitory tale which judiciously lists the many challenges any viable version of Cognitivism ought to be able to address and overcome. In the author’s own words,
‘[i]t cannot count as the generation of genuine insight merely that people have the feeling that insight has been generated… We must offer some standards, no doubt fallible, by which to tell when people have arrived at whatever form of enlightenment is at issue… confirmation that the learner has gotten it right.’[2]
In this stimulating and well-argued book, defenders of Cognitivism are defied to fill the gaping philosophical hole Greg draws the outline of with some precision.
My reflections will in this context have to be limited to brief observations about how we ought to conceive of the epistemic gain one stands to make from engaging imaginatively with fiction, and arise primarily in connection with arguments presented in Chapters 5-9. The first has to do with whether such gain always fits into the established distinctions between kinds of knowledge. The second spells out how the epistemic standards of an alternative conception may not always allow for useful comparison with the standards of science. I do not believe that Greg is entirely unsympathetic to all the issues highlighted below. At the same time, the points I raise could be used by Cognitivists to develop a model set to refute some of the book’s fundamental assumptions, and thereby side-step the central challenge.
Anna Karenina, Crime and Punishment, and Othello are but some of the great classic works of fiction carefully examined by Greg as standard examples of works we tend to think of as yielding knowledge or insight. Why, then, ‘[i]f someone says they learned so much from Anna Karenina it [nonetheless] seems poor form to ask: “What exactly did you learn?” ’.[3] The underlying assumption, he suggests, is that literature educates us in ways that are ‘too subtle, too pervasive, to be discovered by the crass methods of the sciences’[4]. We may well, on Greg’s view, learn something about the nature of longing and desire from Tolstoy and the many facets of jealousy from Shakespeare, and such insights can come either in the form of ‘knowing-that’, ‘knowing-how’ or knowing by acquaintance. But crucially, we ought still to insist that any knowledge so acquired ‘be judged by the very standards that are used in assessing the claim of science’[5].
In countless circumstances, the evasive response (“too subtle, too pervasive”, etc.) does indeed seem entirely inadequate. It won’t do to say that I have learnt something from reading a short story and then be unwilling (or worse, incapable) of saying anything about it. Suppose, however, that there is (also) another way of conceiving of our epistemic gain from fiction, one which is not modeled on the idea that knowledge is something that can be ‘extracted’ from our experience of particular works of fiction and recast in the terms of conventional models of knowing. On this alternative conception, the gain we stand to make is instead understood as primarily borne out of a reader’s long-term engagement with fiction. Let’s call this the ‘subtractive conception’ and ‘cumulative conception’ respectively. What would be the main features of the latter?
i) Epistemic gain from fiction is not the work of one single novel, film, or poem.
Works of fiction are not (always) individual and direct sources of knowledge, capable of imparting or conveying independently packaged cognitive nuggets. Instead, we learn from the cumulative effect of works, sometimes by the same author, such as when we learn about loneliness from reading not only Naipaul’s Half a Life (2001) but also the sequel Magic Seeds (2004), but also by engaging with Coetzee and Conrad and seeing bridges, common influences, and thematic echoes between the three authors.[6] Such snowballing can also occur across artforms, such as when we understand something about the universal fate of refugees when we ‘join up’ Géricault’s Raft of the Medusa (1819) with Forensic Architecture’s Liquid Traces: The Left-to-Die Boat Case (2014).
ii) Epistemic gain from fiction cannot meaningfully be disentangled or isolated from other, previously acquired items of knowledge picked up in non-fictional (or non-artistic) contexts.
Beyond the obvious psychological point that we don’t come to fictions as blank mental slates, a work of fiction of epistemic interest should be seen as a solicitation, overture, or even provocation to imaginatively identify new cognitive possibilities building on one’s own previous experience. What I learn from reading Shikibu’s The Tales of Genji is largely dependent on, say, whether I have been to Japan myself, and can never be identical to anyone else’s private cognitive combinations. The insights that result can for that reason too be difficult to express to others (especially since one might not be clear about which aspects of one’s understanding stem from personal experience and which connect directly to the fiction).
iii) Epistemic gain depends greatly on the reader qua epistemic agent.
Imaginary experiences of fiction can enable us to take control of our epistemic trajectories. What we learn from engaging with fiction depends largely on what we ‘do’ with the epistemic potential of the work. It is up to us to connect the cognitive dots, and this is often a gradual process, prolonged over time. I might not know exactly what I am in the process of learning at time t1, and the character of my epistemic gain might evolve between t2 and t3. Much of the cognitive worth of fiction lies in our own propensity to put the literary work, say, to its best epistemic use. We tend to say that books or films ‘stay with us’, and this is an approximation of what that can mean.
If what we learn from fiction can be cumulative in the respects sketched above, another option arises about why asking of someone who just finished reading Anna Karenina, “What exactly did you learn?”, may seem incongruous or obtuse. On this conception, what we acquire from fiction is not so much an ability or a tool – improved empathic skills for example – as a part of the toolbox itself. By that I mean that we have improved the conditions which allow us to better structure, organize or connect the knowledge we in great part already possess.
Asking what we have gained from a single work of fiction might be rather like asking a friend ‘What did you learn in 1998?’ when really what we are after is what she learnt during her formative years. What we should be asking, then, is perhaps better captured thus: ‘What have you gained from a lifetime of reading Tolstoy or Austen?’. The answer might be something along the lines of ‘The scaffolding necessary for seeing the world differently’ - for making novel associations, for placing pieces into unfamiliar puzzles, for framing a manifold of personal perspectives. And is this not just an instance of knowing-how? Perhaps, albeit of a particularly complex kind. Perhaps setting oneself up to better know how to navigate one’s one life and better understand that of others is, in some respects, like knowing how to navigate London traffic. But is there a meaningful sense in which this is knowledge of which it can ever be said that we have ‘gotten it right’ in the same sense that we can expect when we have read a volume on the American Civil War or an article about tuberculosis in sub-Saharan Africa? No. The most appropriate standard here, surely lies in the very practice of continuously putting our insight to its best possible epistemic use.
Imagining and Knowing places considerable pressure on defenders of Cognitivism to go to much greater lengths to express, check, and test what we feel we learn from fiction. Philosophers on all sides of the debate will benefit from the book’s many positive insights. It is a much-needed and timely contribution which will no doubt shape the discussion about fiction, knowledge and imagination for a long time to come.
[1] ‘The idea that we learn from fictions of any kinds face difficulties not well recognized by those who want to argue that fictions are generally or often or sometimes “educative” ’ (2020: 3-4).
[2] 2020: 183.
[3] 2020: 6.
[4] 2020: 6-7.
[5] 2020: 8
[6] Sometimes, repeatedly revisiting the same work can have a similar effect.
Response from Greg Currie
Thanks to Elisabeth Schellekens for her comments on the final part of Imagining and Knowing, where I register doubts about learning from fiction. She asks what it is sensible to expect by way of such learning. She helpfully lays out three principles, which I will discuss in turn. Reflecting on them gives me an opportunity to say things I should have said in the book.
1. Epistemic gain from fiction is not the work of one single novel, film, or poem.
This is very plausible and the humanistic tradition has emphasised, I think, the role of exposure to literature as a cumulative phenomenon. But is that model (let’s call it accumulation) the right model? An alternative is that learning from fiction is more like the effects of eating: regular food maintains bodily energy at a roughly constant level and a day without food will significantly affect this; regular eating does not, on its own, make for long-term improvements in health (call that the regulation model). Some studies have claimed to show that reading a single short story will measurably raise your empathy levels, a result consistent with both the accumulation and the regulation models, but tending to favour the latter since measurable effects of a single dose suggest impossibly high accumulations over time. But such studies have been hard to replicate, and an empirical project I work on with psychologist Heather Ferguson and philosopher Stacie Friend has failed so far to support this idea (this work is currently under review). Still, if the effect of literature is cumulative and slow we ought to be able to find evidence that does not require looking at whole lifetimes. We can’t (I guess) measure the effects on health of one cigarette but we know something about the effects of smoking just one-to-four cigarettes a day; we can correlate years smoked, and years after smoking, with health risk. Studying the effects of literature may be hard but it is not impossible. Elisabeth, I think, comes close to denying that with her second principle:
2. Epistemic gain from fiction cannot meaningfully be disentangled or isolated from other, previously acquired items of knowledge picked up in non-fictional (or non-artistic) contexts.
If we can’t disentangle the effects of age and the effects of smoking on cancer rates we can’t say what, if any, contribution smoking makes to cancer. Fortunately, in this case and many others, we do have ways to unravel the effects of interacting causes. We know that susceptibility to Covid-19 rises dramatically beyond a certain age, but we can distinguish the effects of age itself from the effects of underlying health conditions closely associated with age. We may think that when it comes to literature the picture is just too variable from person to person, too exquisitely dependent on small differences in experience or temperament. I’m not sure; machine learning is pushing the boundaries of what can be predicted in chaotic systems. Still, it is likely that more time and effort will be invested in predicting earthquakes and solar storms than in tracking the evolution of a reader’s learning from literature. What is knowable in principle is one thing: what we will pay the cost of learning is another. In that sense Elisabeth has a point, and the prospects for disentanglement are not good.
Elisabeth’s third principle raises particularly interesting questions:
3. Epistemic gain depends greatly on the reader qua epistemic agent.
It is surely right that learning from fiction, when it occurs, is the product of many factors only one of which is what we might call the epistemic potential of the source. But that’s true, isn’t it, of all learning? Hartley Rogers’ Recursive Functions and Effective Computability sits on my bookshelf. I will never learn anything from it, lacking the right cognitive capacities. Elisabeth may say there is a difference: we can easily say what a more intelligent me would learn from that book, but not what I would learn from Anna Karenina. But then we must be careful. Attempting Recursive Functions taught me how little talent I have, but it would be a mistake to think this counts towards its epistemic value. We should distinguish between what you learn from a book and what you learn by reading it. The case for learning from literature will depend a good deal of whether the following claim can be made out: Learning of the latter “by reading” kind does, in some cases, count towards the epistemic value of the work in question. It does not count in the case of Recursive Functions; it might count in the case of Anna Karenina. Why?