Luke Roelofs is a postdoc at New York University. When not working on philosophy of mind, they try to introduce their cat Sebastian to the so-called ‘external world’. Sebastian remains unimpressed.
A post by Luke Roelofs
Imagining is often emotionally charged. And it often serves to represent emotions. In this post I’m interested in cases where it’s both: where an emotional state felt by the imaginer (what I’ll call the ‘imaginative emotion’) represents an emotion which the imaginer or someone else either will feel, did feel, is feeling, or would have felt (what I’ll call the ‘imagined emotion’).
For example, in trying to empathise with a friend’s difficult situation, I might imagine myself in their situation so as to simulate their feelings. In trying to predict someone’s decision, I might do the same with an eye to what actions the simulated emotion might lead to; and in trying to make a decision I might imagine myself experiencing the consequences of one choice or the other, and see how it makes me feel. These efforts at emotional imagining are fallible, especially in difficult circumstances, but they are nevertheless common, and often better than nothing (cf. some past posts here).
There’s disagreement about whether imaginative emotions, in some or all cases, are ‘real’ emotions, or mere imaginative copies of emotions (e.g. Walton 1990, p.201ff, Currie and Ravenscroft 2002, pp.187-204, Gendler and Kovakovich 2006). I won’t address that question, because even those who doubt their genuineness agree that they feel like real emotions, and that’s enough to raise the question that interests me:
Interesting Question: What is the relationship between the intensity of the imaginative emotion, as felt by the imaginer, and the intensity that the imagined emotion is thereby represented as having?
This question employs the notion of ‘affective intensity’, which is not at all easy to define. We readily distinguish ‘stronger’ and ‘weaker’ joy, fear, anger, and so on, but it’s not clear what common dimension they all vary on (cf. disputes about how to define ‘vividness’ in mental imagery, e.g. Kind 2017, Tooming and Miyazono 2020). For the purposes of this post, I’ll take intensity to be a cluster concept, comprising at least three aspects:
More intense emotions are more salient: it’s harder to ignore them, and easier to attend to them or to what they identify as important
More intense emotions score higher on any applicable qualitative scale: the pleasure is ‘more pleasant’, the tension is ‘more tense’, the joy feels ‘more joyous’, etc.
More intense emotions have larger effects: if an emotion makes you blush and sweat and want to leave the room, a more intense version will make you blush more, sweat more, and want to leave the room more.
This cluster notion of intensity may not allow precise quantification, but hopefully it’s workable enough to give meaning to my question: how does the intensity of the imaginative emotion relate to that of the imagined emotions?
Here’s one fairly simple answer:
Identity: The imagined emotion is represented as having the same intensity as the imaginative emotion.
This answer has some initial appeal. It seems to mirror the behaviour of other features of imaginative and imagined mental states: it seems like the right answer for emotion-type and object. When I empathically imagine someone’s fear of a large dog, it seems like I do so by recreating a state whose emotion-type is fear, and whose object is that large dog, and by doing so I represent their emotion as having those same features.
Moreover, Identity seems to capture some intuitive expectations we have. If two people are saddened by my sadness, but one is greatly saddened and other only slightly saddened, then other things being equal I might expect the former to estimate my sadness as greater than the latter does. But I think Identity is too simple - hence ‘other things being equal’. Let me sketch a slight refinement to it, modelled on an analogy with another property: duration.
We imagine things that take time, and our imagining takes time. Sometimes the two seem to correspond. Indeed, one piece of evidence that subjects mentally recreate a motor process to solve a problem (as in Shepherd and Metzler’s famous experiments, 1971) is that the time they take to give an answer is directly proportional to the time that the motor process would take.
For this reason, it seems plausible that imaginings, by default, represent imagined processes as having the same duration as the imagining itself. But this can’t be true across the board, because it would radically limit our ability to imagine events taking long periods of time. It seems perfectly sensible for us to say to one another ‘try to imagine how horrible it would be to be stranded in the rain for a whole hour!’, and to respond by forming an imagining that, although it recreates many sensory details (the cold, the uncomfortable cling of wet clothing, the obscured vision, etc.) takes much less than an hour to complete.
What’s going on here? I would say: we are exploiting our ability to intentionally manipulate what aspects of our imagining represent which aspects of the imagined thing. This is an ability we already need for other reasons: since any visual image could represent multiple objects equally well, we must be able to just ‘make’ an image be an image of a certain object. And we exercise this ability in more complex and flexible ways when engaged in ‘imaginative scaffolding’ (Kind 2020), imagining some experience we’ve never had by compiling experiences we have had: “it would feel like a mixture of this and that, but with a structure like this, and…”, though we needn’t assume there’s an explicit thought of this form.
Call this ability, however we end up analysing it, ‘representational control’. My suggestion is that we can use representational control to stipulate a break from the default equivalence of imaginative duration with represented duration. We can imagine ‘being in the rain for an hour’ by recreating an experience as of being in the rain, and taking the imagined experience to be, so to speak, ‘like this, but for a whole hour’.
But doing this requires compensating adjustments to any conclusions we draw from the imagining. If we’re trying to predict someone’s behaviour, it’s not enough to just imagine trying to avoid, or feeling irritable after, a momentary experience of cold, wetness, etc. - an hour in the rain is much more unpleasant, so they’ll be correspondingly more irritable, and more motivated to avoid it. Making these adjustments draws on background knowledge about the significance and effects of the difference between a moment and an hour.
It’s like we take a shortcut, imagining only a brief ‘sample’ of the experience and not the full hour of it, but we pay for saving time by expending more cognitive labour. We have to regulate the inferences we make, to make them approximate the inferences we would have made if we had slogged through imagining the full hour in real time.
I think the same goes for intensity. By default, the imagined emotion is represented as having the same intensity as the imaginative emotion, but this default can be broken using representational control. We can, for instance, consciously recognise that someone else is profoundly unhappy, even though the empathic sadness we feel for them is only quite mild. Indeed, one paradigmatic image of an appropriately ‘empathetic’ person is someone who conveys precisely this, whose own empathic sadness is manifest but who explicitly says ‘I can’t imagine how much this hurts’.
As with duration, these departures require the imaginer to make compensating adjustments, and these adjustments require drawing on background knowledge about the significance of changing intensity. If I feel mildly saddened by someone else’s severe grief, my predictions about their behaviour, and my own offers of aid, should reflect the severity of their emotion, not the mildness of mine. To ensure this I rely on my background knowledge of how intensity works, and how greater sadness differs from lesser.
If that’s right, we should be careful when talking about empathising with someone ‘more’ or ‘less’. We should distinguish the intensity of empathy (the extent to which the empathiser is themselves emotionally affected) from its strength (its epistemic and motivational effects - likelihood to offer help, keenness of understanding, solidity of resolution to change behaviour in the future). A very faint empathic emotion might be taken as representing a very powerful emotion, and if the imaginer makes the necessary compensating adjustments to the epistemic and motivational effects of their imaginative emotion, their empathy might be high in ‘strength’ despite being low in intensity.
Indeed, it might be useful sometimes to deliberately ‘tone down’ the intensity of our empathy. Overly intense emotion might be distressing, tiring, and distracting for the empathiser, and might actually get in the way of understanding or helping the other person. Someone trying to empathise as accurately and efficaciously as possible might therefore often not display much empathic affect.
We might draw an analogy to a volume knob on a song you’re trying to learn: too low and you can’t make out the words, too high and you can’t focus on them because it’s deafening. There’s a skill in being able to adjust the knob to the optimum volume, and someone listening at a higher volume isn’t necessarily listening better.
References:
Currie, Greg, and Ravenscroft, Iain. 2002. Recreative Minds: Imagination in Philosophy and Psychology. Oxford: Clarendon Press.
Gendler, Tamar, and Kovakovich, Karson. 2006. “Genuine Rational Fictional Emotions.” In Matthew Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art. Blackwell 241-253.
Kind, Amy. 2017. “Imaginative Vividness.” Journal of the American Philosophical Association 3 (1):32-50.
Kind, Amy. 2020. “What imagination teaches.” In Becoming Someone New: Essays on Transformative Experience, Choice, and Change, edited by J. Schwenkler and E. Lambert. Oxford University Press 133-146.
Shepard, R. N. & Metzler, J. 1971. “Mental rotation of three-dimensional objects.” Science 171, 701-703.
Tooming, Uku, and Miyazono, Kengo. 2020. “Vividness as a natural kind.” Synthese:1-21
Walton, Kendall. 1990. Mimesis as Make-Believe. Cambridge, MA: Harvard University Press.