Dr. Seth Goldwasser is a lecturer at the University of Miami. Seth’s research focuses primarily on skillful mental action with an emphasis on skillful remembering and imagining. He has also written on the ascription of normal-proper functions in cancer biology and on the epistemic status of traumatic memories.
This week at The Junkyard, we’re hosting a symposium on Rob Hopkins’ recent book: The Profile of Imagining (Oxford University Press, 2024). Yesterday we began with an introduction from Rob. Commentaries and replies follow today through Thursday.
* * *
Commentary from Seth Goldwasser
Robert Hopkins’s The Profile of Imagining articulates with great clarity and defends with sprezzatura a new vision of sensory imagining. I love this book. And I find much to like about Hopkins’s profiling account of sensory imagining, as I defend the claim that imagination in general is an agential power, namely, a skill (Goldwasser 2024). I hope to be the Tommy Oliver to Hopkins’s Jason Lee Scott.
My commentary focuses on Chapter 1, where Hopkins defends the claim that imagination is at root an agential power and, so, stems from spontaneity. Spontaneity is the self as agent. Hopkins says that it’s the source of control and that all action involves control by the agent. And he understands control in terms of Frankfurtian (1978) guidance: for an unfolding happening to be controlled is for something to guide that happening towards some end. For Hopkins, agential guidance is the agent’s guiding her bodily or mental movement towards some end, whether that end is set by her through an intention or by something “lower level” (2024: 33). Finally, Hopkins argues that imagining and, a fortiori, sensory imagining are fully spontaneous, despite it being the case that not every imagining is an action, because it always makes sense to try to control one’s imagining (2024: 36-38).
I worry about Hopkins’s characterization of spontaneity. Because it forms the metaphysical core of the profiling account, finding a problem here might spell trouble for the account as a whole. That said, I intend the following as a friendly invitation to say more.
The way I see it, Hopkins’s sketch of agential guidance is too threadbare.
Frankfurt (1978) sets out his guidance view of action in contrast to a form of causalism which maintains that the distinguishing mark of action is the presence of a prior psychological cause of the right kind. Frankfurt points out that this wrongly ignores cases where the exercise of agency isn’t prior but ongoing as an action unfolds. What ties such cases to action in general is that agency is possessed by virtue of mechanisms of the agent that function to keep actions on course towards achieving their ends. It’s the counterfactual dependence of the unfolding of the relevant happenings on the activities of those mechanisms that make those happenings actions.
But Frankfurt doesn’t fully spell out which mechanisms are guidance mechanisms and which distinguish agential guidance. And causalists have since adopted Frankfurt’s guidance view. Many causalists agree that, whatever the mechanisms are, they’re ones for guidance because their activity in shaping a heretofore unguided happening into an action consists in instantiating the appropriate causal loops. These mechanisms receive input from the happening and produce modifications which, in turn, result in corresponding changes in the happening which, in turn, are reflected in further input. The mechanisms relevant for action, then, are those the agent can make use of in adjusting her movements.
Hopkins’s sketch of agential guidance doesn’t go into depth except to say that it’s the agent who guides her movement towards some end and, when so guiding, controls her actions.
This leaves him with two options:
1) Adopt full-blooded Frankfurtian guidance.
2) Avoid adopting full-blooded Frankfurtian guidance.
Each option has unsavory upshots.
Starting with (1), there are problems with guidance views of action. I’ll focus on one. The kind of guidance at issue is agential. But what is agential guidance? I don’t think there are any good answers that rule out cases of, say, homeostatic function mediated by the central nervous system.
Consider the thirst response. It’s generated as a function of two mechanisms in the subfornical organ that respond to levels of angiotensin II and sodium in blood plasma (Matsuda et al. 2020). As levels of angiotensin II and sodium change, so too does the activity of these mechanisms. When levels of angiotensin II are high, “water neurons” in the subfornical organ are activated, generating the thirst response. When sodium levels are low, these neurons are inhibited by GABAergic interneurons. The ebb and flow of angiotensin II and sodium levels means that the firing rates of and the strength of inhibition on water neurons is constantly changing and that there’s some threshold of activity above which the thirst response is supposed to be generated. In other words, we have the kind of causal looping among neural mechanisms involved in the thirst response that characterizes guidance.
But it isn’t the agent’s guidance. Agential guidance is supposed to be distinguished by some mechanism that “flexibly integrates and coordinates the workings of other subsystems in a characteristically agential way” (Hendrickx 2023: 3134; my emphasis). But what does this mean? It could mean that such mechanisms are “top-down” in their activity or that their activity somehow realizes individual-level intentions. But neither of these moves works. First, like the subfornical organ, activity in the relevant mechanisms—say, of the “executive system”—is driven by “bottom-up” inputs from lower-level systems and the environment that inform their subsequent activity. So, there’s no obvious line to draw between activity that’s top-down versus bottom-up. The second move assumes that talk of the relevant neuronal or subsystem-level representations or talk of such mechanisms or subsystems “realizing” individual-level mental representations is uncontroversial. But the jury is still out here (e.g., Favela and Machery 2023; cf. Shea 2018).
As it stands, a full-blooded Frankfurtian guidance view doesn’t illuminate spontaneity.
What about (2)? The problem here is that, without further elaboration, agential guidance as Hopkins sketches it isn’t necessary for intentional action and, so, isn’t necessary for action or agency. Consider a pro-golfer making a hole-in-one. Once the ball ceases to make contact with the club, the golfer cannot guide it towards the hole. Yet, if she makes the hole-in-one, she does so intentionally. Her action isn’t merely one of trying to make a hole-in-one, swinging the club, or hitting the ball. The ball’s soaring through the air, landing on the green, and rolling into the hole are all part of her action too! Despite having and exercising her ability to make a hole-in-one, she does these latter things without the ability to guide them as they unfold. So, agency and its exercise in action, intentional or otherwise, outstrips Hopkins’s agential guidance.
What does all this have to do with sensory imagining and perception?
If Hopkins goes with (1), it becomes difficult, if not impossible, to distinguish imagination from perception metaphysically in terms of spontaneity and receptivity, respectively. (2) leaves open the possibility that perception involves spontaneity in some way other than the possibility of control through agential guidance, e.g., in attending or observing, weakening the metaphysical distinction between it and sensory imagining. Moreover, (2) leaves open the possibility that unbidden imagery and “attending to” or “observing” what we sensorily imagine are, after all, instances of receptivity rather than defective or qualified forms of spontaneity (cf. Hopkins 2024: 133ff.).
I think we can characterize spontaneity and receptivity in a way that avoids these problems. Hopkins is right that sensory imagination is an agential ability. But I think we can take a page from Kant here: spontaneity and receptivity are dimensions of all of our knowledge-forming faculties. This would make perception and sensory imagining closer in kind. They might be a capacity and ability, respectively, that have their differential shares of spontaneity and receptivity, respectively.
What characterizations of spontaneity and receptivity would make sense of imagination and perception having differential shares of both? I’ll end by offering a start. I assume that actions are an agent’s interacting with the environment or herself where such interaction is constitutive of the agent’s (self-)creation and where agents are always already acting. Spontaneity is my contribution in an interaction constitutive of such creation, whereas receptivity is the impact on me of things other than myself as self-conscious subject in such an interaction. On the whole, then, sensory imagining is spontaneous despite its having a share of receptivity because my contribution in the creation of the imagined object is metaphysically primary. And, on the whole, perception is receptive despite having a share of spontaneity because what impacts me in the formation of my perceptual judgments is metaphysically primary.
References
Favela, L.H. and E. Machery (2023). Investigating the concept of representation in the neural and psychological sciences. Front. Psychol. 14:1165622. doi: 10.3389/fpsyg.2023.1165622
Frankfurt, H. G. (1978). The Problem of Action. American Philosophical Quarterly 15, no. 2: 157–62. http://www.jstor.org/stable/20009708.
Goldwasser, S. (2024). Imagining as a Skillful Mental Action. Synthese 204 (38):1-33.
Hendrickx, M. (2023). Agentially Controlled Action: Causal, not Counterfactual. Philosophical Studies 180 (10-11): 3121-3139.
Hopkins, R. (2024). The Profile of Imagining. Oxford: Oxford University Press.
Matsuda, T., T.Y. Hiyama, K. Kobayashi, K. Kobayashi, and M. Noda (2020). Distinct CCK-positive SFO neurons are involved in persistent or transient suppression of water intake. Nat Commun 11, 5692. https://doi.org/10.1038/s41467-020-19191-0
Reply to Goldwasser
Seth Goldwasser, in his thoughtful and generous-spirited comments, focusses on my discussion of the relation between imagining and agency. He's right to think this a theme on which much more might, and perhaps should, be said; and some of the questions he raises are perfectly fair.
There are, I think, two main challenges in Goldwasser's comments. The first is to say more about what I call 'the source of control'.
I seek to understand the relation between imagining and agency. I do so by considering imagining both under its aspect as a mental occurrence, and under its aspect as a power those occurrences actualize. Goldwasser concentrates on what I say under the first heading, and here, therefore, so will I; though I'd note I think we need an integrated view of both aspects if we are to grasp the truth about the way imagining and agency relate. My proposal is that, considered as an occurrence, imagining is not always an action (because not always actually controlled by the subject); but it always originates in that part of the self that is the source of action, the part that in action exercises control. As I put it, imagining is spontaneous.
But what, Goldwasser wants to know, is that part of the self? Or, to put the point in the form in which he develops it, given that many elements in our complex makeup exert control over various goings on in us (his example is the working of the subfornical organ in regulating thirst), which of those centres of control is the one in which imagining must originate, if it is to relate to agency in some noteworthy way?
This is a fair question. I am less certain, however, that it is one I need to answer in order to fulfil my ambitions in the book. I appealed to control in order to avoid a feature of much philosophy of action, which construes action in somewhat demanding terms: as, for instance, requiring relatively sophisticated mental states such as intention or considerations recognized by the subject as reasons. In my view, not all that counts as action meets that high bar: I cite Helen Steward's nice examples of such actions as absent-mindedly biting one's nails or twirling a pen. But, more saliently, such rich conceptions of action seem particularly ill-suited to imagining. The somewhat wayward nature of much of our visualizing, etc. (think of those versions of 'mind-wandering' that involve sensory imagining) renders them ill-suited to count as intentional or driven by reasons. And one can hold that much without yet thinking that this shows that imagining is simply something that happens to us, something as unconnected to agency as suddenly feeling thirsty, or sneezing. That limns the space I sought to occupy. The appeal to control constituted my attempt to establish rights of residency there.
That claim to residency will be vindicated provided there is some source of control we acknowledge as the source of that control which governs action. That there are other sources, such that control by them does not constitute exercises of agency, is by-the-by. We can recognize the control exercised by the subfornical organ without worrying that such control generates action. Of course, it would be nice to be told quite what does mark out the special source, the one control by which constitutes action. But that is the job of a philosophy of action. I sought not to provide such a philosophy but to lean on the key notions it must explicate.
That said, if I were pressed to say more to identify the key source of control, what I would offer might well be something with which Goldwasser would find himself in sympathy. The agential source of control is that source which can exhibit sensitivity to reasons. It can be the source of action even when not so sensitive, as when I absent-mindedly bite my nails. It can be the source of non-actions when it is not even exercising control, as in those imaginings that constitute mind-wandering. But what marks it out as the key point of origin is that on other occasions it exerts control over processes guided by the pursuit of ends that the subject has adopted (and may modify) in light of her sense of what she has reason to do. That is not to answer Goldwasser's challenge in the terms he tentatively suggests: appeal to reason is not appeal to self-creation. Nonetheless, it moves in a direction he may find congenial.
The other challenge Goldwasser lays down for me is to make out the way in which imagining is spontaneous but perception is not. Here he is, I think, motivated by his thought that many actions involve elements that are beyond our control (the hole-in-one example). Now, I'm not sure exactly how we should describe that case. (Holing out is something the golfer does, in some sense; and he certainly performs actions by way of doing it; but is one of his actions to hit a hole-in-one?) But let's just concede that, while an agent controls some elements constitutive of an action of hers, she need not control them all; and that they need not all even be such that she could control them. Turning to imagining, Goldwasser's worry is (I think) as follows. Even some imaginings will involve elements that cannot be controlled. (Imagining will sometimes be like hitting a hole-in-one.) But if they cannot be controlled, these elements not merely are not controlled (as I'm happy to concede is true, not only of many such elements, but even of some imaginings as a whole); but cannot be thought to arise in that part of the self that exercises control. That is, (1) they are not spontaneous. Therefore (2) they are receptive. So, (3) at least some imaginings combine spontaneity and receptivity. But, since that last is true of perception, (4) we are owed an account of how imagining and perceiving differ in these respects.
My response to this worry is to get off the escalator at stage (2). Spontaneity and receptivity are not contradictories. We know what spontaneity is (at least according to me). For a state to be receptive is for it to be determined by the things it represents (Profile p.20). So, failure to be spontaneous does not entail being receptive. Perception is always receptive, imagining never is. (See, for a defence of this last in the hardest case, Ch.5.2.) Whether or not perception is ever spontaneous (in my sense), the contrast between it and imagining can stand. There is no need to start apportioning out degrees of each in the way Goldwasser sketches.