Responses to Andrea Blomkvist, Fabrizio Calzavarini and Wade Munroe
Bence Nanay
I have always thought of philosophy as a conversation. I know, not a very original take, Socrates and all, but it’s true. A conversation not just with other philosophers but also a conversation between disciplines. That is part of why I wrote this book that aims to bring together philosophy, psychology and neuroscience.
As a result, I have also always thought of my contributions to philosophy not as the last and definitive word, but as the first word that, if I’m lucky, sparks a conversation. In a sense, what I want is not necessarily being right. What I want is that others think and worry about those questions I think and worry about.
So it is especially gratifying to receive these three commentaries, all of which challenge what I wrote in my book Mental Imagery: Philosophy, Psychology, Neuroscience and develop it in three very different and very interesting directions in a way that I hope will start something new.
Aphantasia
The first of these directions is the research on aphantasia. Some people, when they close their eyes and visualize an apple, see nothing ‘in the mind’s eye’. How can we explain this? Andrea Blomkvist and I have very different explanations, me emphasizing the importance of unconscious mental imagery and her emphasizing the importance of episodic memory. We probably agree, however, that aphantasia is a heterogenous phenomenon. Just to give one example, some aphantasics don’t dream in images at all. Others have vivid visual dreams. There are clearly very different underlying mental processes here, but they get lumped together as neither of these groups can conjure up vivid visual imagery at will.
In the last couple of years, we are finally getting some information about the neural underpinnings of aphantasia, and this has extremely important philosophical implications. The study Blomkvist focuses on is this one. I should say that is an elegant gesture from Blomkvist to focus on this study as not all recent neuroscientific studies of aphantasia are as supportive of my account as this one. This specific study is as close to my own account that emphasizes unconscious mental imagery as it gets (I once discussed with one of the authors of this study how one could talk about unconscious mental imagery without saying ‘unconscious mental imagery’, which is a term that rubs many referees the wrong way and we discussed ‘imageless imagery’ as an alternative, which this article went with). And the results here do indeed seem to suggest that at least some aphantasics do seem to have unconscious mental imagery: the contents of their attempted imagery can be decoded from their primary visual cortex.
But Blomkvist draws attention to an issue that is, to my mind, independent of the question about unconscious mental imagery: the failure to cross-decode perception and imagery in aphantasics. This is a crucial data point with regards to a hotly debated topic in the neuroscience of mental imagery, which concerns the degree of similarity between perception and mental imagery. While both perception and mental imagery seems to be coded in the primary visual cortex, in the case of aphantasics, they seem to be coded differently. The authors of the study frame this as imagery and perception having different formats, but it is important that what they mean is not format in the philosophical sense of the term (as in imagistic vs. propositional format), but basically what philosophers would call the ‘vehicle’.
Sense modalities
About one fifth of my book is devoted to what I call ‘multimodal mental imagery’: mental imagery in one sense modality (say, vision) induced by sensory input in another sense modality (say, audition). For this to make sense, I need to have a workable way of distinguishing different sense modalities – one grand topic in the philosophy of perception. And while I devote a chapter to this in the book, Fabrizio Calzavarini is absolutely right to put pressure on me to clarify this issue.
Calzavarini and I agree on most things. We agree (albeit for slightly different reasons) that it would be a mistake to identify sense modalities by brain regions. We also agree about the importance of recent findings about how some brain regions that we used to think of as unimodal processing centers (including the primary cortices) take their input from a variety of sense organs. We do frame these findings differently though. I think of these as examples of multimodal mental imagery – say, visual imagery triggered by auditory input. I don’t think this is a major departure from Calzavarini’s preferred ‘supramodal’ brain terminology (although my label may sound more pedestrian than his…).
Finally, to go really big picture, we also agree that the difference between sense modalities may not be as important as the illustrious history of this question would warrant. As I argued in this other book I published shortly after the one on mental imagery, sense modalities don’t work independently from one another: they influence each other at the earliest possible stages of perceptual processing. If we take the multimodality of perception seriously, the distinction between different sense modalities will be of secondary importance.
Language
One of the more surprising claims I argue for in the book is that mental imagery is essentially involved in language processing. This is a surprising claim as imagistic cognition (involving imagery) and linguistic cognition (involving language) have long been taken to be different and sometimes even opposing forms of our mental life. Wade Munroe challenges this claim about the essential involvement of mental imagery in language processing. I am not convinced. His main argument comes from studies about the congenitally blind, which show no significant difference between the language processing of sighted and congenitally blind people. But given that many congenitally blind people do have visual imagery (and the vast majority of them have non-visual imagery), I don’t think this establishes the lack of involvement of mental imagery in language processing.
It is important that my claim is not that language processing is (exclusively) mental imagery. What I argue for in the book is that mental imagery is essentially involved in language processing. Not that that’s the only thing that is involved in language processing. Munroe sometimes seems to read me as holding the stronger claim, and I may be partially guilty of this misunderstanding given my repeated use (going back to my 2013 book) of the metaphor that language is the icing of the cake of the human mind.
And this, to go in full circle, is a beautiful demonstration of the power of mental imagery on the way we think about the world (and the human mind). Munroe and I may have different mental imagery when visualizing a cake. When I visualize a cake, the icing is not this detachable clump that sits on top of the cake. It seeps into the dough, making it tastier and more interesting. And the same goes for language and the rest of the human mind. Language is, I maintain, the icing on the cake, and we can’t understand the cake if we start out from the icing. But to understand the whole thing, we also need to take the icing seriously.