Jacopo Frascaroli is a doctoral candidate in the Department of Philosophy at the University of York, where he contributes to a Leverhulme-funded interdisciplinary project entitled “Learning from Fiction”. His work brings together aesthetics, philosophy of language, and cognitive science.
A post by Jacopo Frascaroli
My discussion here has been conveniently anticipated by last week’s post, as well as by a couple of older posts by Max Jones and Sam Wilkinson. It is about some recent developments in cognitive science that could be of great consequence for our understanding of imagination, in many of its varieties. The developments in question fall under the predictive processing (PP) framework (and related formulations: “active inference”, “free energy principle”). As a grand unifying theory of cognitive function, PP is one of the most hotly debated topics in cognitive science and philosophy of mind, one that arouses in equal measure enthusiasms and scepticisms. As an account of imagination in particular, PP is starting to gain traction (Clark 2015, Kirchhoff 2018) and to encounter objections (Jones & Wilkinson 2020). Here however I won’t try to assess the PP story about imagination in its details. Instead, I will try to trace this story back to its Kantian roots. As we shall see, PP seems to give new strength to a distinctively Kantian view of imagination as a ubiquitous mental capacity, a capacity far more pervasive than what is normally thought. This view of imagination could well be of value even if many of the details of the PP story turn out to be wrong or imprecise. Here I want to suggest, in an intuitive and informal way, what this view could entail and what scope and prospects it could have.
A useful starting point is what Kant says in the first Critique about the role of imagination in “synthesizing the manifold of intuition”. Kant’s story is complex, but let me condense it to its bare minimum for the sake of the matter at hand (for a more extended summary, see Matherne 2016). The problem that Kant addresses is one that still haunts perception science. In every moment of our waking life, we are presented with a constant flux of stimulations, which reach our sensory organs unorganized, unrelated and atomised. Yet the world that we experience is one of relatively stable objects unfolding in an orderly fashion. Since, pace the empiricist, this order is not to be found in the flux of stimulations itself, the mind must be actively structuring the manifold of intuition by means of some kind of conceptual apparatus. To recognise this fleeting, barking, ever-changing patch of shapes, colours and sounds as a “dog”, says Kant, we need to be able to synthesise the manifold of intuition in accordance with the appropriate concept. The problem is, however, that concepts are general, abstract entities, whereas sensory stimulations are unique and ephemeral. A faculty is needed, then, to bridge the gap between concepts and intuitions, adapting a general concept to a particular series of perceptual stimulations or, which is the same, subsuming various particular stimulations under the relevant general concept. This faculty, according to Kant, is imagination.
It is important to stress how radical Kant’s proposal is. Most contemporary philosophers of mind conceive imagination either by contrast with perception, or as complementing and enriching perception in important, but limited ways (see e.g. Nanay 2010, Macpherson 2012, Gregory 2018). For Kant, instead, perception as such is imaginative. The very process of organising the scattered stimulations we receive from the senses to obtain the relatively ordered and stable world of objects we inhabit is an imaginative process. Exactly how this wonderous process takes place, Kant tries to explain in the famous chapter on schematism (A137–47/B176–87). Kant’s discussion there is as suggestive as it is problematic. Ultimately, he is led to admit that the procedure by means of which the imagination brings the relevant concepts and intuitions together is “a hidden art in the depths of the human soul, whose true operations we can divine from nature and lay unveiled before our eyes only with difficulty” (A141/B180–1).
Here is where the PP story may come into the picture. For PP offers a way to operationalise Kant’s constructivism and explain, within a Bayesian framework, just how we are able to mediate between concepts and intuitions (see Hohwy 2013, Clark 2015 and Seth 2021 for the whole picture). “Concepts”, in the PP story, are to be understood in terms of the flow of predictions of a hierarchical generative model (instantiated by our brain structures and dynamics) that are tested against the incoming sensory data (the “manifold of intuition” in Kantian terms). These predictions play out across many spatial and temporal scales. A high-level prediction to see a dog gives rise to “lower-level predictions about limbs, eyes, ears and fur, which then cascade further down in terms of predictions about colours, textures and edges, and finally into anticipated variations of brightness across the visual field” (Seth 2021). At each level of the hierarchy, predictions are compared with – and constantly revised in light of – the incoming sensory data. In this way, by constantly updating our predictions in light of what we get from the senses, we can make contact with a structured world full of objects, people and places. According to PP, in sum, objects are always inferred, hypothesised as the most likely causes of the sensory data we are receiving. As a precursor of the PP approach (Kantian himself in many ways) would put it, “objects are always imagined as being present in the field of vision as would have to be there in order to produce the same impression on the nervous mechanism” (Helmholtz 2013/1867).
But whether the details of the PP story will turn out to be right or not, a crucial thing should be noted: while one can maintain that perception is always an imaginative process (in the sense that Kant envisaged and PP seems to operationalise), one need not say that perception is always imaginative to the same degree. Sometimes “synthesising the manifold” is more immediate, sometimes it is more difficult, sometimes it is just impossible. Consider for example these three different visual renditions of Kant’s dog-recognition scenario:
When we look at the picture on the left, perceptual organisation happens immediately and unproblematically: it is evident what kind of concept we should bring to bear on this manifold (in PP terms, the stimulus is easily and quickly explained away by an appropriate cascade of top-down predictions). In the picture on the right, on the other hand, we are not able to find any clear perceptual explanation: intuitions without concepts (that is, in PP terms, without an appropriate flow of top-down predictions) remain blind. In the middle image, on the other hand, our initial puzzlement at what seems just a random configuration of black and white patches might suddenly give way, with a characteristic feeling of insight that psychologists call an “aha!” experience, to the realisation that there is a dalmatian there too, sniffing the ground. It is in cases like these, where perceptual recognition is hindered, but not made impossible, that the inferential character of perception is made more manifest. And it is in cases like these that we are more inclined to think of perception as an imaginative activity.
This is to note what should already be apparent: imagination comes in degrees. It presents itself along a spectrum that runs from the most mundane and irreflexive cases, to the most inventive and deliberate, up to another extreme where imaginative activity becomes vague and indeterminate, and then simply impossible. If it is somewhere in the middle of this spectrum that we get what we tend to consider paradigmatic instances of imaginative activity, we should not fail to see that imagination might be active in all cases, albeit to different extents.
This is where this story gets interesting for people (like me) working in aesthetics. For there is a growing consensus that artworks might be designed to provide us with just the kind of open-ended, exploratory experience that we get in that intermediate area of the imaginative spectrum. (To give just an example, consider how impressionist paintings such as Monet’s Regatta at Argenteuil “leave space for the observer’s visual system to perform its interpretive work” (Seth 2021)). Here we might have a way to reformulate Kant’s intuition that, during aesthetic experiences, “the powers of cognition are in a free play” (KU 5:217), because the imagination is free to explore and play with different possible organisations of the manifold. Kant’s paradoxical characterization of the imagination’s state in these cases as one of “free lawfulness” (KU 5:240) seems to allude precisely to a condition in which perceptual organisation is neither too obvious and restrictive, nor too open and unstable to lead to any clear result. Pursuing this line of thought further, one can begin to see not only that aesthetics is (or ought to be) continuous with the study of perception, but also that perceptual experience has always in it, to a greater or lesser degree, something of the ingenuity of an aesthetic achievement.
In just the same vein, the Kantian/PP story I am sketching might be useful for understanding the role of imagination in scientific inquiry. In fact, according to PP the brain, as a relentless hypothesis-testing machine, approaches sensory data using principles analogous to those that govern the interrogation of scientific data. The perceptual inferences that we make in organising our sensorium are not different in kind from the more deliberate inferences that, on a more abstract level, allow the scientist to probe the causal structure of the world. The “aha!” moment that we get when we see the Dalmatian might well be the forerunner (in both phylogenetic and ontogenetic terms) of the sudden click of scientific discovery.
I grant that, as I have presented them here, the above claims have no higher status than that of vague suggestions. Nonetheless, I think that they are worth pursuing further. If developed carefully, they could yield a comprehensive picture that seamlessly unifies the role of imagination in perception, art, scientific inquiry, and probably other areas as well (action, practical problem-solving, empathy and theory of mind seem other obvious candidates). It might be objected that, conceived in such a broad way, “imagination” becomes such a general category that it ceases to be of any practical use. I prefer to think that a view of this kind, precisely because of its breadth, may help us appreciate the misguidedness of some of our theoretical distinctions and see relationships between areas of inquiry that we would not be able to grasp otherwise. Imagination, as we know, is a heterogeneous continent, very difficult to chart. To have a theory that is able to connect discussions on imagination in such a variety of areas, and that can boast at the same time a long philosophical past and a link to the latest acquisitions in cognitive science would be, I think, an interesting result.
References
Brown, D. H. (2018). Infusing perception with imagination. In F. Macpherson and F. Dorsch (Eds.), Perceptual Imagination and Perceptual Memory, 133–160. Oxford: Oxford University Press.
Clark, A. (2015). Surfing Uncertainty: Prediction, Action, and the Embodied Mind. Oxford: Oxford University Press.
Helmholtz, H. (2013/1867). Treatise of Physiological Optics, Vol. 3. Tr. A. Gullstrand, J. von Kries and C. Ladd-Franklin. New York: Dover Publications.
Hohwy, J. (2013). The Predictive Mind. Oxford: Oxford University Press.
Jones, M., & Wilkinson, S. (2020). From prediction to imagination. In Abraham (Ed.) The Cambridge Handbook of the Imagination, 94-110. Cambridge: Cambridge University Press.
Kant, I. (1998/1781-87). Critique of Pure Reason. Tr. P. Guyer and A. Wood. Cambridge: Cambridge University Press.
Kant, I. (2002/1790). Critique of the Power of Judgment. Tr. P. Guyer and E. Matthews. Cambridge: Cambridge University Press.
Kirchhoff, M. D. (2018). Predictive processing, perceiving and imagining: Is to perceive to imagine, or something close to it?. Philosophical Studies, 175(3), 751-767.
Macpherson, F. (2012). Cognitive penetration of colour experience: Rethinking the issue in light of an indirect mechanism. Philosophy and Phenomenological Research, 84(1), 24–62.
Matherne, S. (2016). Kant’s theory of the imagination. In A. Kind (Ed.), The Routledge Handbook of Philosophy of Imagination, 55-68. New York: Routledge.
Nanay, B. (2010). Perception and imagination: Amodal perception as mental imagery. Philosophical Studies, 150, 239–254.
Seth, A. (2021). Being You. A New Science of Consciousness. London: Faber & Faber.