Imagining What You Can Do: The Brain, Free Will, and Art
by Patrick Colm Hogan
July 18, 2005
This essay treats imagination and freedom, arguing that they are inseparable from one another and from the creation and experience of art. It discusses the conditions in which we experience freedom, the relation of these conditions to the varieties of imagination, and the relation of both to our experience of nothingness. It illustrates this analysis by reference to Zhang Yimou’s Hero. Specifically, elaborative imagination, which develops long-term trajectories, involves prefrontal cortex. Generally, we experience unimpeded initiatives of prefrontal cortex as free. But these initiatives are circuits of neuronal activation, thus determined. A sort of free will enters at the limit of causal analysis, for the observer cannot be included in his or her account of those neuronal circuits. However, the fact of death suggests that the observer is always subservient to those circuits, even if, for a time, his or her imagination is precisely what allowed their recognition and articulation.
Imagining What You Can Do: The Brain, Free Will, and Art1
Patrick Colm Hogan
About half-way through Zhang Yimou’s Hero (the top-grossing Chinese film in Chinese history), we are told a story about two rebels, Broken Sword and Flying Snow, who were also lovers. An anonymous warrior—"Nameless," as he is called—fights with Snow and kills her. This is part of a shared political stratagem and Snow has agreed to die. Nonetheless, Broken Sword feels that, because of Snow’s death, he too must fight Nameless. There is a lake, with a tiny island, where Snow lies on a bier. The warriors fly about the lake. They touch the water lightly with a foot or the tip of a sword and bounce back into the air. Circling around Snow’s lifeless body, they fight like dancers in a ferocious pas de deux. At one point, the spray of water from their blades touches Snow’s cheek with a single drop. There is no animating force in her body to make this a sign of sorrow. But Sword cannot bear to see the drop roll down her cheek like a tear. He alights and reaches out toward her face. Seeing his opportunity, Nameless swoops down on Sword. But, coming near, he realizes that his opponent is clearing the tear from Snow’s face. He turns, faltering, and almost descends into the water.
The scene is strange. Men fly like birds. They seem capable of doing anything they imagine. Indeed, it is just imagination, if somehow imagination that has real consequences, for the b attle, we are told, takes place in their minds. Yet, at the same time, Sword and Nameless cannot do everything, nor do they wish to. It may seem that this sequence is very Chinese, very much a matter of the conventions of the swordplay film. But I believe it is paradigmatic of literature, and of the relation of literature to two peculiar features of consciousness, features that are intimately connected, indeed inseparable—free will and fancy.
In "Literature, God, and the Unbearable Solitude of Consciousness," I argued that a crucial feature of consciousness—or, rather, self-consciousness--is that it is unbreachably alone. To be self-conscious is to experience utter isolation. I argued further that a central function of both religion and literature is to make us feel that the wall of consciousness has been breached, that we have indeed shared some experience, that we are not inescapably alone. In the following pages, I wish to consider two other aspects of consciousness, aspects that are bound up with one another and with solitude—imagination and freedom. Moreover, I shall argue that these aspects of consciousness are inseparable from our creation and experience of art.
On the Experience of Nothingness: Why We Are Alone
As Sartre might tell us, a table does not feel alone. But we do. When someone dies, the chair he sat in every day does not notice his absence. But, looking at the chair, all I can see is something that is not there. The absence is more real to me than anything present. Why is this? Sartre distinguishes between the chair as merely in-itself and me as for-self. But this does not explain just what is going on, at least not to the satisfaction of someone whose model for explanation is cognitive neuroscience. In terms of standard cognitive architecture, the experience of nothingness seems to be the result of several factors. First, our minds spontaneously produce expectations of immanent experiences. Pascal Boyer explains that "we cannot operate normally without a keen sense of the possible dangers that might result from intervening circumstances as we are performing an action." For example, "Opening the car door before the car has come to a complete stop" may be considered by one cognitive "system and rejected by running a quick scenario of what might happen and giving it a negative emotional valence before it is considered" self-consciously (239). These ongoing, spontaneous expectations often involve an emotionally consequential sort of sensory projection. Robert Jourdain points out that "the brain perceives by anticipation . . . . imagery arises from the unfolding of . . . anticipatory schemes in the absence of actual perceived objects" (163). Indeed "the brain makes the body move not merely by shouting commands down the corridors of the nervous system, but also by anticipating the sensations that will result from those commands. We project a flux of such anticipations before us in whatever we do, testing them against incoming sensation" (302). Such anticipations may be produced by the priming of memories relevant to places, circumstances, and so forth. In the case just mentioned, I look at the deceased person’s chair; this triggers memories; the memories generate expectations. The contrast between the expectation and the actual experience has effects on subcortical areas of the brain, producing emotions—in this case, sorrow. This process of sensory projection is a form of what we commonly call "imagination." Thus we may say that I imagine the presence of the person who is not there and that this imagination produces my feeling of grief.
There are two ways in which processes of imagination are important for art. The first, and more obvious way is that art itself is the product of imagination. This introduces us to a different sort of imagination. Imagining a film is a more complex process than imagining that a deceased friend will be in his usual chair when I enter the room. Specifically, the former is a type of elaborative and self-conscious imagination. The latter, in contrast, is spontaneous and unself-conscious. The second way in which imagination bears on the arts is that representational arts, and particularly arts bound up with narrative, address the topic of imagination with great insistence. Imagination—both elaborative and spontaneous--is one of the most important elements thematized in literature, whether we are speaking of Romeo’s imagination of union with Juliet, Othello’s imagination of Desdemona’s infidelity, Macbeth’s imagination of kingship and betrayal, or Prospero’s imagination of revenge.
Hero is a film that recurs continually to the topic of imagination. I find it deeply moving that Broken Sword stops his battle to wipe the drop of water from the face of Flying Snow. Without reflecting, we know immediately what is going on here. He cannot help but imagine spontaneously that she is weeping. In wiping the tear from her cheek, he is doing two things. First, he is trying to comfort her. Of course, he knows that she is not weeping, that she is not even alive, and thus cannot be comforted. Yet he spontaneously imagines that she is hurt. But that is not all he is doing. He is also comforting himself. He is acting in such a way as to erase from the world just the thing that triggers his imagination of Snow’s pain, and thus his own anxiety. We also know what happens with Nameless. We know it just as readily, just as immediately. He was never in love with Snow. But when he sees Broken Sword sliding his hand across her cheek, he understands what Sword imagines; he spontaneously imagines Sword’s feelings. That makes him sorrowful in just the way the tear made Sword sorrowful. Imagination fills both with compassion. Now, neither can fight. Nameless falters in retreat, his attention divided between concern for Broken Sword and the task of retreating to the shore.
How, then, do we know all this? We too experience the same sort of spontaneous imagination. There are complex, neurobiological reasons why we automatically imagine other people’s internal states. For our present purposes, these reasons are unimportant. The important thing is that we do imagine these states, and we do so spontaneously. In short, our imagination is the same as that of Broken Sword or Nameless. But, of course, in another way it is not at all the same. We exist. Broken Sword and Nameless are mere fictions. They do not have inner states any more than a chair. How strange, then, that we imagine inner states for them. This suggests a catastrophic cognitive failure on our part. We imagine this inner life with complete emotional certainty. But we are entirely mistaken in doing so.
This may seem to be a problem unique to literature and film. It is not. It is a problem of everyday life. Every day, we imagine what other people feel and think. Every day, we try to communicate to them what we feel and think. And every day we find ourselves failing more or less. My imagination of a fictional character is only a particularly extreme form of this quotidian failure. Indeed, it is extreme in precisely the same way that Broken Sword’s failure is extreme. We imagine his feelings, though he has none, just as he imagines the feelings of Flying Snow, though she has none.
More exactly, there are two sorts of loneliness. One is the ordinary loneliness of separation. The other is the existential loneliness that results from our recognition that we can never truly share experiences, feelings, ideas with anyone else (see my "Literature" 129). Both are the direct result of imagination. Ordinary loneliness results from our spontaneous imagination of someone’s presence—the empty chair, the unanswered telephone. The other, greater loneliness results from our imagination of other people’s imagination—our spontaneous sense that we do experience what they experience, and our realization that we do not and never can.
If loneliness is the basic emotional condition of consciousness, its most definitive or characteristic operation is, perhaps, imagination—crucially including the imagination of other people’s own emotions and imaginations.
Fancy Unbound: Imagining Free Will
But there is more to the scene in Hero than imaginative compassion. Indeed, the compassion interrupts something else, something that at least appears to be the main point of the scene—the battle. The compassion apparently operates as a sort of constraint, disrupting the smooth flow of action. Is imagination, then, a sort of impediment? On the one hand, the scene suggests this. But, at the same time, it indicates the opposite. Again, the battle takes place in the minds of the warriors, in their imagination. Technically, they do not move, but the visual and motor areas of their brains are activated almost as if they did; their visuo-spatial working memory is filled with images of the unfolding conflict. Here, we come to one more feature of imagination, a feature that leads us to another crucial property of consciousness. Perhaps the most salient feature of the battle is that it is entirely unrealistic. The swordsmen bounce from the surface of the lake as if it were a trampoline. They fly through the air as if they had wings. This, I think, suggests what we mean by "free will."
Philosophers have spent a great deal of time debating what free will is or might be—just what relation it might have to the ordinary causal sequences of the material world.2 I will consider this issue below. However, I do not believe it is the best place to begin. Rather, I believe it is best to begin with just what we ordinarily mean by free will. When chemists determined that water is H2O, they began with a good sense of what we ordinarily mean by "water" and refined that into a scientific concept. One problem with discussing "free will" is that, in my view, we do not have a good sense of the ordinary language use of the idea. Thus we do not have a clear concept to refine. Moreover, in isolating this ordinary language use we need to follow the principles of ordinary language. One crucial principle is that we do not, in general, understand terms by reference to necessary and sufficient conditions. Rather, our ordinary meanings tend to be a matter of prototypes—sometimes empirical prototypes (i.e., roughly, averages of our actual experience), sometimes ideal prototypes or what I refer to as "paradigms." Our default way of thinking about birds is by reference to prototypical birds, such as robins, not by reference to an abstract definition of birdness. Our default way of thinking about diet foods is not by reference to our empirical experience of diet foods, but by reference to paradigmatic cases, such as lettuce (see Kahneman and Miller 143).3 The same point holds for "free will." This is important because understanding paradigmatic cases of free will should help us to understand how we come to have the idea of free will. This, in turn, should give us an idea of what free will might be, should we decide that it exists.
My contention here is that our paradigm for freedom is imagination. The presumption of this paradigm is that, while our actions may be constrained in many ways, our imagination is unfettered. Of course, we simultaneously believe that our imagination is tightly constrained. We cannot get certain thoughts out of our heads; we cannot help but see a friend in his habitual chair even months after he has died. Thus in order to isolate the paradigm of freedom, we need to distinguish different sorts of imagination. I have already drawn the most important distinction here—that between spontaneous and elaborative imagination.
Spontaneous imagination occurs without choice. It is closely related to emotional response and is part of our ongoing (or "on-line") engagement with events in the world. It is short-term in its projections; it dissipates quickly; and it is usually bound up with our moment-to-moment activity. Elaborative imagination, in contrast, is what we (apparently) choose to do. It does not occur in critical situations of ongoing action, but in contemplative removal from action (i.e., it operates "off-line"). It is related to emotion, but that relation is attenuated, and more open to control in the sense that we can (again, apparently) change what we are imagining, or alter its course and development. Finally, it involves long-term projections and may be extended almost indefinitely.
For the moment, I do not wish to take a stand on whether our intuitions about this difference are reasonable. I merely wish to note that elaborative imagination as we understand it is the paradigm for the exercise of free will. While our spontaneous imagination may be constrained in much the same way as our bodily actions are constrained, our elaborative imagination does not seem to be constrained.
Of course, there is a way in which our elaborative imagination may be constrained, and constrained in precisely the way our physical action is constrained. Specifically, we often imagine possible courses of action in order to anticipate outcomes realistically. The entire point of such imaginations is to obey the restrictions imposed by the material world; the entire point is not to be entirely free. Yet the fact is that we can, and often do, violate these constraints in imagination. It is precisely that possibility and the practice of violation that make imagination a paradigm of freedom. Imagination is, indeed, the way we recognize constraint, just as it is the way we recognize isolation. We recognize constraint by imagining what is impossible.
We may refer to these two sorts of imagination as pragmatic (the sort that conforms to material constraints) and unbounded (the sort that does not conform to material constraints). This distinction is only partially aligned with that between spontaneous and elaborative imagination. Spontaneous imagination is always pragmatic and unbounded imagination is always elaborative.4 But elaborative imagination may be either pragmatic or unbounded.
Unbounded elaborative imagination is, I believe, our paradigm for freedom. When we judge our practical freedom in the real world, we judge it relative to this paradigm. Put differently, freedom of action aspires to the condition of unbounded imagination. The point has obvious bearing on literature. Literature routinely treats pragmatic imagination, both spontaneous and elaborative. Indeed, to a great extent, one could see realistic literature as an operation of pragmatic imagination.5 But literature crucially includes unbounded imagination. Indeed, one purpose of literature seems to be that of allowing us to explore freedoms beyond those that we experience in actual life or in pragmatic imagination.
Hero provides us with an apt instance. In this film, we find imagination at two levels. Zhang and his co-writers have followed pragmatic imagination in setting out much of the progress of this story about an attempted assassination. However, they have also relied on unbounded imagination in developing the fighters’ skills. Perhaps more importantly, the film thematizes the operation of both pragmatic and unbounded imagination. The plans followed out by Nameless, Broken Sword, and others are all instances of pragmatic imagination. The story relies on Nameless having imagined the ruler’s response when he tells him a particular story. Specifically, Nameless imagines that the ruler will allow him to come within ten paces of the throne, thereby providing the opportunity for assassination. Similarly, Broken Sword relies on his imagination of Nameless’s response when he tries to dissuade Nameless from assassinating the ruler. Most poignantly, at the end of the film, Flying Snow finds that Broken Sword has succeeded in preventing the assassination. (The earlier story of her death turns out to have been false.) Snow had longed for the death of the ruler as revenge for her own father’s death. In grief and anger, she challenges Sword to a fight. At one point, she lunges at him, as if to kill him. First, Sword holds up his blade to block her thrust. But then, as she passes through the air toward him, he tosses the weapon to the ground. When she has pierced him and he is dead, she can only stare in bewilderment, tears flowing down her cheeks, and mutter—why didn’t you block my sword? why didn’t you block it? Though she called for a fight, she had imagined the outcome even from the start. It was a standoff—or perhaps her own defeat. It is clear that her pragmatic imagination never allowed his death as a possibility. When she plunged toward him, as if to deliver a death blow, her spontaneous imagination produced an image of him, expertly deflecting her thrust.
Unbounded imagination too is examined continually in the course of the film. Sword and Nameless escape the ordinary laws of physics and biology when they fight in the story’s reality, but more importantly they escape those laws when they fight only in their minds. Zhang emphasizes that their fight took place deep in the minds of the warriors. This foregrounds the topic of unbounded imagination. It points to the openness of that imagination. Indeed, the imaginative freedom of the warriors in the film exceeds that of even the director and screenwriters. Here is the remarkable thing about the battle in their minds: It is just one battle, not two. The imaginations of the warriors should be disjointed. Nameless should imagine things one way while Sword imagines them another. But that is not what happens. Their imaginations transcend that boundary between one consciousness and another—which, of course, is even less possible than flying like birds or walking on water.
What Is Will?: The Freedom of Gods and Machines
I have been maintaining that unbounded imagination is a paradigm of freedom. We certainly experience such imagination as free. We certainly judge it to be free. But is our feeling anything other than delusion? Is our judgment in this case reasonable? Indeed, what would it mean to say that imagination is free? One obvious meaning would be that imagination is uncaused or, equivalently, that it is the first cause of all that follows. Imagination is a bit like God, then—it creates out of nothing. But this seems clearly wrong. It would appear that all our thoughts and actions are objectifiable and identifiable as brain states. These brain states follow from other brain states and lead to subsequent brain states. There is an ongoing causal continuum here, just as there is in any physical system. There is no point at which causality pauses and a new sort of cause enters from nowhere. In terms of standard brain architecture, it is simply a matter of circuits of neurons. Certain sequences of neuronal firings give rise to other sequences of neuronal firings, and that is all there is to it. In terms of a more intuitively comprehensible cognitive architecture, we could say that some current sequence of experiences triggers procedural and representational memories, along with various semantic structures. These structures enter working memory in combinations that yield what we think of as unbounded imagination. But there is nothing "free" about it. It is just as constrained as anything else.
Consider, again, the film—specifically, Zhang’s unbounded imagination of the battle scenes. He is at the point of composing the scene surrounding Snow’s corpse. He is making a sword film in the Chinese tradition. These facts prime various episodic memories, lexical items, procedural schemas, and so forth. (These primed memories, etc., are themselves ultimately circuits of neuronal activation.) When he thinks of the fight, a number of primed elements receive further activation. For example, the use of wires and the defiance of gravity are commonplace in the genre. These, then, have a very high degree of activation. Indeed, a wide range of Zhang’s episodic memories (e.g., memories of particular scenes in other sword films), lexical items, and so forth, have some degree of activation. The elements with the highest degrees of activation are the ones that enter into his imagination and thereby enter into the film. This is not because Zhang chooses them, and still less because he creates them from nothing. Rather, it is a simple result of the degree of activation. That is why he makes use of wires in the film, why he has his characters defy gravity. Of course, Zhang still has to work out the details of the movement. One obvious lexical item that is activated by the idea of flying is, of course, birds. When this is paired with a location that involves water, it is likely that ideas about particular water birds will be activated. The swooping movements, the skimming of the water, and other features of the battle are undoubtedly derived from this association (thus memories of birds, schemas of bird flight, and so forth)—an association produced by ordinary paths of neuronal activation.
In this respect, then, we cannot really say that imagination, even unbounded imagination, is free—at least not in the sense of being an uncaused cause. But perhaps there is another sense in which we could see imagination as free. Indeed, there is a second sense of "free" that is almost as obvious as "uncaused cause." Moreover, this second sense seems more reasonable, at least prima facie. In this second sense, an action is free if the causation is primarily internal to the agent rather than external. Thus Zhang’s imagination may be considered free insofar as it is the result of his own brain processes rather than the result of external circumstances. But, in the end, this is no less problematic than the uncaused cause definition. If the uncaused cause definition requires that every person be God, the "internal cause" definition allows every system (e.g., every mechanism) to be free. If a machine is merely following its internal principles, then, by this definition, it is proceeding freely. Alternatively, the internal cause definition may be understood in such a way that it eliminates all agents. Our current actions are always a partial response to external conditions. Moreover, our "internal" orientation is not merely internal. It is inseparable from a history of experiences. Indeed, our "internal" memories are, for the most part, memories of external events that affected us.
In short, neither definition seems adequate. However, both play some role in our quotidian deliberations about free will. Our ordinary talk about freedom commonly stresses the uncaused nature of our choice. But it also stresses the source of decisions in our own preferences or character, and it does this no less often. The two together suggest a sort of solipsistic understanding of free will, an understanding in which I act in a certain way because of who I am, but who I am is uncaused.
Why We Act: The Brain and Its Motives
Before continuing with our tacit ideas about free will, it may be useful to pause for a moment and consider more concretely the structure of our "decisions" as they operate in the causal sequences of brain activity. The question is why do we act in the particular ways we do act. At any given moment, we do some things and not others. Why?
The standard answer throughout human history has been given by folk psychology. We have certain desires or aims; we have certain beliefs; and we have reasoning capacities. The desires or aims motivate us. The beliefs, suitably integrated with the reasoning capacities, guide us in implementing our aims. Though neuroscience complicates this picture considerably, the resulting account remains fundamentally similar. Specifically, it appears that our motivations are the result of subcortical, emotional processes (where "emotion" is understood in a very broad sense to include not only anger and fear, but hunger and lust as well). As Panksepp notes, "neurological evidence . . . indicates that mammals possess highly specific emotional and motivational systems in subcortical regions" (46). Zajonc explains that "cognitions of themselves are incapable of triggering an instrumental process, unless they first generate an emotion that mobilizes a motivational state capable of recruiting action" (47). Thus cortical processes on their own do not supply any motivating impulse. I take this to be one lesson of Antonio Damasio’s research. Damasio discovered that patients with emotion-inhibiting brain damage are deficient in their ability to make and act on decisions. One patient, for example, repeatedly behaved in self-destructive ways, despite a fine intellect that one might have expected would lead him to make better decisions (see chapter three of Damasio). Damasio sees this as suggesting that emotion is important to reasoning. I see it, perhaps differently, as suggesting that reasoning is largely irrelevant to motivation, and that motivation is largely a matter of emotion.
On the other hand, subcortical motivations may be modified by higher cortical processes. As Joseph LeDoux’s work indicates, our experiences feed directly into our emotion circuits. But the link between perception and emotion is very rough and inaccurate. As LeDoux explains, referring to fear, there is a "low road" of perception. This low road quickly transmits a rough "sketch" of perceptual information to the amygdala. If it happens that this sketch contains apparently fear-relevant information, it allows us to act quickly, thereby avoiding the danger. For example, if I am walking in a forest and glimpse a twisty, glistening, mottled branch on the path, I am likely to leap back. This is because a limited version of this perceptual information is passed swiftly to the amygdala. The amygdala responds with a fear signal, roughly because it has received information that is consistent with the presence of a snake. But, of course, I do recognize that this is a branch. The more detailed information passes via the "high road" (as LeDoux puts it) to the cortex. Particular cortical systems then identify the object as a branch, sending this information to the amygdala, which then stops emitting the fear signal. Though most cases of cortical/subcortical interaction are more complicated,6 this provides a good illustration of the relation between the two. The cortex does not inspire or motivate action. Rather, it modifies the action-motivating processes of the subcortical emotion systems.
One might wonder if the simplicity of this particular case explains the limited roles of these two parts of the brain. There is some validity to such a concern. Cortical processes are not confined to inhibiting subcortical emotion circuits. They may have excitatory effects also. Again, emotion is a matter of certain subcortical systems being aroused. These systems do not become aroused on their own. There appear to be three sources for such arousal. The most obvious is direct, perceptual experience. The second is imagination, part of the neocortical working memory system. When we imagine certain events, our brains are activated in much the same was as when we experience those events (see Kosslyn 295, 301, 325; Rubin 41-6 and 57-9). As a result, when imagining events, we experience attenuated versions of the relevant emotions. The third source of emotional arousal is emotional memory. This involves "implicit" or unself-conscious recollections of earlier emotional experiences—which were themselves presumably triggered by perception or imagination (see LeDoux Emotional 182; for a fuller discussion of these issues, see chapter seven of my Cognitive). Obviously, imagination is particularly crucial for our present concerns.
Before returning to our main topic, however, we need to develop the account of human action a bit further. Specifically, we need to distinguish between the ongoing, moment-to-moment execution of actions and the larger trajectories that give those moment-to-moment actions a function. Take a simple example. Suppose I want to eat a snack. I engage in a series of actions—getting up from the chair, walking to the kitchen, opening the refrigerator door, and so forth. At any moment, I may encounter some unexpected event. For example, a dog may be lying down in the kitchen doorway. In response, I step over the dog. In each case, my specific actions are oriented and given sense by the larger trajectory.
Now, let me complicate this picture slightly. Suppose that the dog is lying in the doorway to the kitchen and suppose further that I am afraid of dogs. In order to avoid irrelevant concerns, suppose also that I know the dog is harmless. In this case, I encounter not a minor obstacle, but a serious, emotion-triggering obstacle to accomplishing my goal of eating a snack.
The case of the feared dog in the doorway allows us to consider the relation between brain systems and the components of action. It may seem at first that moment-to-moment actions are governed by subcortical impulses while encompassing trajectories are governed by higher cortical routines, perhaps from the prefrontal cortex. However, this is not precisely the case. After all, my hunger is just as emotional as my phobia of dogs. In other words, at every level, my motivation remains emotional, thus subcortical. However, it is the case that the role of elaborative imagination increases as the orientational framework goes beyond sub-routines to larger trajectories. Moreover, the autonomy of imagination increases with this transition also, for it is less affected by moment-to-moment changes in emotional state. This returns us to the difference between spontaneous and elaborative imagination.
Clearly, spontaneous imagination operates in relation to moment-to-moment action. In keeping with this, it appears that spontaneous imagination may not be part of working memory at all. For example, Jourdain states that kinesthetic anticipations "occur in somatosensory and parietal cortex" (227). In some cases, anticipations result from brainstem processes, such as those involving automatic calculations to maintain "a constant retinal position for images during movement" (Matthews 223).7 Given this, it seems inappropriate to refer to these anticipations as forms of imagination with its location in the working memory system--roughly, the lateral prefrontal cortex and the anterior cingulate cortex (see LeDoux Emotional 277).8 I will therefore reserve the term "imagination" for an aspect of working memory, referring to these anticipations as spontaneous, short-term, perceptual and motor projections or, more simply, spontaneous projections.
In contrast with spontaneous projections, elaborative imagination functions primarily in relation to larger trajectories. Indeed, it seems likely that this is a crucial part of its evolutionary function. This function comes into play most consequentially when subcortical emotion systems are likely to be overwhelmed by short-term emotion triggers that work against the organism’s long-term interests. Most importantly, then, elaborative imagination serves to represent emotionally-rich situations that contradict immediate, moment-to-moment emotion triggers from other sources. In other words, it seems that our brains are designed in such a way as to routinely place elaborative imagination (and its associated emotions) in opposition to experience, emotional memory, and short-term perceptual projection. Put simply, I see the dog in the doorway and my immediate impulse is to flee. The impulse is triggered by the sight of the dog, by the spontaneous projection of the dog leaping up with bared fangs, by the childhood memory of a friend being bitten in the crotch by a dog, and so on. It is the function of my elaborative imagination to contrast with all this the image of a tasty pie that is waiting for me on the other side of Cerberus.
The Limits of Autonomy: Variations on the Theme of Pie
Does this mean that the elaborative imagination operating through the lateral prefrontal cortex and the anterior cingulate cortex is the free part of my mind? Is the struggle between my desire for pie and my fear of dog-bites a conflict between choice and determinism? There are different ways of answering these questions. From the perspective of neuroscience, the answer to both is a straightforward "no." The elaborative imagination is no less a result of circuits of neuronal activation than is the stimulation of my amygdala. The difference between working memory and emotion/motivation systems is simply a matter of the types of information that produce the activations, the constraints on the activations, and so forth. It is not a matter of caused versus uncaused causation. Both sequences, "high" and "low," result from causal antecedents.
In case this seems counterintuitive, consider the following (necessarily much simplified) scenario. Various aspects of bodily monitoring signal to my brain that I am hungry. This signal serves to trigger recent food memories. Several memories occur to me, all indicating what is in the refrigerator—broccoli, raw chicken, canned cranberry sauce from Thanksgiving five years earlier, and pie. Here, it may seem that I freely choose among the objects presented in memory. Again, from a neurocognitive perspective, this is not what happens. What appears to be me choosing freely is in fact a process of multiple circuits producing a highest level of activation for a given item. Suppose, then, that pie receives the highest degree of activation. This, in turn, activates spatial memories (regarding the location of the refrigerator), procedural schemas (how to get up from the sofa, how to walk to the kitchen), and so on. Now, I reach the doorway and find the dog. The image of the dog immediately activates that part of my amygdala—or, more generally, the subcortical emotion system--that generates fear. However, at the same time, the perception also activates memories of non-threatening interactions with this dog. These memories are integrated into working memory. Moreover, the hunger signals presumably continue, as does the imagination of the pie (in this case, by way of reactivation circuits operating through working memory). Cortical systems, bound up with working memory, send inhibitory signals to the fear system, including the amygdala. If adequately activated, these various systems send signals to areas of the motor cortex. Whatever produces the strongest level of activation produces the action. Roughly speaking, if the cortical areas bound up with working memory and the hunger system manage to inhibit the fear activation and excite the motor areas, then I step over the dog. If the fear system, prominently including the relevant amygdala circuits, produces adequate activations to inhibit the procedures being enacted by the motor areas, or to substitute other procedures (e.g., withdrawal), then I do not step over the dog. From the neurocognitive perspective, there is no issue of free choice here. There are only different circuits of ordinary, material causality.
On the other hand, if we adopt a phenomenological or experiential perspective, the differences just outlined do suggest something about the way we experience and understand freedom. Specifically, in this case, I would say that I know the dog is harmless, but my phobia prevents me from getting the pie—or, more generally, from doing what I want to do. In other words, I experience the pursuit of pie as freely chosen and the withdrawal from dogs as a constraint. Of course, this has nothing to do with dogs and pie per se. I may be on a diet. I want pie, but I imagine my brother making fat jokes at the upcoming family reunion. Still, I am overwhelmed with my love for pie. I rush toward the kitchen, step over the dog, and plunge my fork into the flaky crust. Afterward, I am ashamed. I say that I have given in. Something has compelled me. I did not choose freely. I was "weak willed."
We sometimes say that emotion compels us and that unemotional, rational decision is what makes for a "free" choice. But clearly that is not the case. This is not how we experience freedom. The conflicts in both of the preceding examples were conflicts between different emotions. In the first case, the compelling motivational emotion was fear (of the dog) and the free motivational emotion was hunger (for the pie); in the second case, the compelling motivational emotion was hunger (for the pie) and the free motivational emotion was fear (of ridicule). Thus our commonsense interpretation of conflict in free choice as a conflict between reason and emotion appears to be mistaken. Similarly, it would also be mistaken to understand this conflict in terms of a simple division between subcortical and cortical impulses—because we do not have cortical impulses. The impulses are all subcortical.
However, the neurocognitive analysis of general patterns, combined with the phenomenological exploration of particular instances, suggests a way of understanding conflicts of will. It appears that we consider the "free" choice to be whatever option is most bound up with certain frontal lobe processes. More exactly, in cases of conflict, we tend to identify ourselves with those processes most fully involved with the lateral prefrontal cortex and the anterior cingulate cortex. In keeping with this, we tend to identify ourselves with long-term goals, rather than short-term goals (or short-term projections). Again, long-term goals are emotionally motivated. However, the emotional motivation is not the result of immediate experience. After all, the point of something being a long-term goal is that it is to be achieved (or sustained) in the future, and not simply the immediate future of spontaneous projection. In consequence, long-term goals are inseparable from elaborative imagination. Given all this, it is unsurprising that our paradigm for free choice is, precisely, elaborative imagination. Imagination is, after all, the neocortical activity that enables our long-term goals (even though, once again, the goals have motivational force only due to subcortical, emotional excitation).
More technically, certain frontal lobe systems bring together and organize information from various other brain systems. As Damasio explains, "there is a particular region in the human brain where the systems concerned with emotion/feeling, attention, and working memory interact so intimately that they constitute the source for the energy of both external action (movement) and internal action (thought animation, reasoning). This fountainhead region is the anterior cingulate cortex" (71). LeDoux explains that "Because prefrontal regions receive convergent inputs from sensory, memory, emotional, and motivational circuits, they are believed to be capable of performing the complex kind of information integration (binding) that must occur in the brain during a conscious experience" (Synaptic 194). The processes that occur in this area—perhaps an integration of the lateral prefrontal cortex and the anterior cingulate cortex—define what we consider to be free.
If my evolutionary analysis is correct, we would expect a degree of conflict between elaborative imagination and moment-to-moment emotional excitation deriving from other sources, such as perception. Indeed, we would expect this conflict to be of considerable importance for action, and not a sort of glitch in the neurocognitive system. In keeping with this, these conflicts have significant phenomenological consequences. Specifically, we experience these conflicts as conflicts between freedom and constraint. Of course, this makes no sense if we conceive of freedom as uncaused causality. (Again, no cortical activity is uncaused.) However, it makes some sense if we conceive of freedom as "internal" causality. I in effect consider certain frontal lobe processes to be "me." Or, rather, whatever conflicts with those processes is "not me." (My sense of self expands or contracts in relation to such conflicts.) Thus actions produced or inhibited through lateral prefrontal cortical/anterior cingulate cortical processes (or, more exactly, through the integrated coordination of relevant brain systems with such processes) are "internally" caused—hence "free," by the definition just given. In contrast, actions produced or inhibited in separation from or opposition to such processes are "constrained." Note that this predicts that we may extend our sense of constraint to other cortical areas as well, such as semantic memory. Consider a further alteration in my story of pie. Suppose I am at a restaurant in France. I have studied French, but I find myself unable to remember how to ask for pie. In this and related cases, I experience my own (cortical) memory systems as inhibitions on my free self—which is, again, elaborative imagination involving the lateral prefrontal and anterior cingulate regions.
But is Free Choice Truly Free?: On Causality and Death
We see, then, that there is a specifiable and explanatorily consequential pattern to the way we understand and experience freedom. But does this in fact tell us anything about whether or not we are genuinely free in any usual sense of the term. Again, it seems that we are free, by this account, only in the sense that a robotic vacuum cleaner is "free" when following its internal "impulses" (i.e., the motions produced by its circuitry) and is not "constrained" by being wedged between a chaise lounge and a wall. Is that all there is to our freedom? Is it just a much more complicated version of the robotic vacuum cleaner?
I cannot give a straightforward answer to this question. In fact, I believe the premises of the question are wrong.
Up to this point, I have been shifting back and forth between phenomenological and neurocognitive accounts as if they are directly parallel. In fact, I do not believe that they are parallel. I believe, rather, that they are radically different—or, more properly, radically incommensurable. Here I need to return again to my argument in "Literature, God, and the Unbearable Solitude of Consciousness." There I contend that no material causal account of the world can be complete for a causal account can never encompass itself. Put differently, there is always some observer who necessarily stands outside the system he or she is describing, thus outside the material world. Each of us is part of the material world when it is observed and explained by someone else. But each of us is outside that material world when acting as the observer. In one account, I am all neuronal sequences and material causality. But, from the subjective or observer position, I am free choice.
Note that the discontinuity of these accounts—the account of me as observed material object and the account of me as observer—does not determine just what is to count as the observing self. Indeed, it is clear that we can include parts of our "selves" in the causal/material account without undermining our status as observer. Most obviously, I may treat most of my body in terms of material causal sequences. For example, if I have vision problems, I can include my own eyes in the material/causal account, and not in the subjective/observer role.9 Indeed, I may count my impulses and spontaneous projections as part of the material world. What it seems I cannot include, however, is what thinks the world, what synthesizes it, what interprets it. This is not precisely imagination, but it is working memory and thus includes the capacities that allow elaborative imagination.
In this way, the self of our implicit phenomenological view is indeed the self of metaphysics and epistemology as well—the self that is necessary for there to be any causal account of the world at all. As such, we may reasonably say that this self defines a sort of free will. It defines free will in that our actions may be in keeping with its processes or not, as discussed above. Put differently, from the perspective of another observer, each of us is entirely caused, and talk of freedom makes no sense. If there is freedom, in this case, it is just the freedom of the robotic vacuum cleaner. However, from our own perspective, we act—or at least think and imagine—in accordance with the internal principles of a minimal self, the self of working memory, the self of ongoing subjective experience. That minimal self is uncaused in our own account of the world.10 It is uncaused not because it is self-caused, but because it is presupposed in any causal account I might give. Thus it is removed from any such causal account. In this way, my minimal self is a ground for free will—but only in my own account of the world. When I act according to my minimal self, then I am acting freely, not because that self is uncaused, but because it is itself the very possibility of a causal account. Note that this is different from the robotic vacuum cleaner, for the robotic vacuum cleaner can never define a causal account of the world. It is never the observer who marks the point of incompleteness in a material/causal analysis. Note also that, in this account, I cannot coherently identify my minimal self with my lateral prefrontal cortex and anterior cingulate cortex, or any other brain region or system. All brain processes are necessarily part of a causal sequence, and my minimal self is precisely the observer that is construing, and is thus outside, such a causal sequence (even if it may be fully identified with particular brain systems and their processes in a third person account).
So, some sort of free will does emerge here. On the other hand, it is a disappointing type of free will. Among other things, it is a form of free will that is enslaved to the very causal determinations it underwrites—for all perspectives stop with death. The third person view turns out to be correct. We end like things. That point too pervades literature and art as it pervades our emotional lives.
This returns us, one final time, to Zhang’s Hero. In a scene I have already discussed, Flying Snow is enraged by Broken Sword. She challenges him to a battle. But when she kills him, she cannot recover. She continues to address him as if he were a person with a perspective on the world. She asks him the same question over and over. In the end, she embraces him from behind, grasps her sword (which is still lodged in his chest) and plunges it to the hilt. It passes through his body and through hers. We see here that death is the ultimate limit on one’s freedom. But it is not the limit in precisely the way one might expect. Specifically, it is not one’s own death that is the unsurpassable limit on freedom. Once I die, I am gone; there no more self that could be free or could be constrained. The ultimate limit on a person’s freedom, then, is the death of someone else, the death of someone he or she loves. Again, we have freedom only insofar as we are utterly removed from the world. That isolation is the condition for our autonomy. But it is also what makes that autonomy futile. For even when two lovers are alive, they are, as The Tale of the Heike puts it, "remote from one another as the sky" (437). Nothing, not even imagination can truly join them. Death is the ultimate realization of this isolation and futility. That is one reason why consciousness leads us to imagine death so compulsively, and so elaborately, in life, in religion, and in art.
1 An earlier version of this essay was presented as a keynote address at the Consciousness, Theatre, Literature, and the Arts conference at the University of Wales, Aberystwyth, in May 2005. I am grateful to the organizer, Daniel Meyer-Dinkgräfe, and to the participants for stimulating questions and comments. I am also grateful to two anonymous referees for their generous and knowledgeable comments on this essay.
2 Moreover, some philosophers have done this in the context of neuroscience. I cannot overview this diverse and complex material in a short essay. Interested readers may consult Walter for a useful overview of some work on neuroscience and free will, as well as a treatment of these issues that is very different from my own. Some of my fundamental disagreements with the premises of much work in this area may be found in “Literature.”
3 In general, the more normatively weighted a term, the more reliant it is on paradigms rather than empirical prototypes.
4 There is one qualification to this. Spontaneous imagination operates along with elaborative imagination. When Nameless falters in his fight with Sword, I spontaneously imagine that he will plunge into the water, perhaps drowning. In this way, spontaneous imagination does sometimes operate in entirely unrealistic conditions. However, even in those cases, its function is to anticipate moment-to-moment events pragmatically, given the conditions of the encompassing, elaborative imagination—hence my sudden feeling of anxiety as Nameless begins to fall.
5 This is the aspect of literature tacitly emphasized by evolutionary psychologists such as Tooby and Cosmides, and Pinker (538-43), who have argued that imagination allows us to test outcomes of various actions “off-line” and thus to increase our chances for reproductive success. On some of the problems with this approach, see my Cognitive 210-13.
6 Indeed, even this case is more complicated in that there are complex, interrelating paths of activation and inhibition. As one reader for this essay pointed out, “the orbitofrontal cortex acting through the uncinate fasciculus on the amygdala projecting to thalamus and back to orbitofrontal cortex inhibits amygdaloid reactions.” The preceding statements clearly leave out some aspects of this process. The crucial point, however, is that cortical systems are not themselves the source of emotive/motivational impulses, but particular cortical systems can affect emotive/motivational processes in an inhibitory fashion.
7 This probably involves the parietal cortex as well in the selection of visual objects (cf. Matthews 223-24 and 232-33).
8 In making these connections, my purpose is not to advocate strict localization (i.e., the isolation of a single, narrowly bounded, contiguous brain region) for working memory or any other psychological system. Clearly, psychological systems may involve the interaction of distinct brain regions. In part for this reason, correlating psychological systems with brain systems is not only an empirical issue, but a complex theoretical one as well (see my Cognitive 202-210 for discussion). However, it is possible to connect psychological systems and brain regions without claiming that the correlation is complete or that the brain regions in question are somehow walled off from one another. Indeed, for the present analysis, it is crucial that cortical working memory is not walled off from subcortical emotive/motivational systems. However, it is equally crucial that working memory, and other cortical systems, are not themselves sources of emotion/motivation. In other words, the systems are distinct, but they are not isolated from one another, nor need they be strictly localized. As we will see, the distinctness of the systems allows us to experience a particular emotion as a constraint on our freedom, an obviously common occurrence. At the same time, the motivational necessity of subcortical emotion systems means that there is no dichotomy between emotion and freedom. Thus free will cannot be identified with an absence of emotion. In fact, without emotion we could not experience anything as freely chosen or inhibited.
9 “Observer” here does not mean “someone who has sensory experience,” but someone who thinks of the world in material/causal terms.
10 Of course, we can adopt someone else’s perspective empathically. In doing so, we (imaginatively) locate him or her as the observer. Nonetheless, the point remains that it is the subjective/experiential, or first-person position that defines an observer. In connection with this, I should emphasize that this account of freedom does not require that the observer be self-conscious of himself or herself as an observer. Indeed, it would be fairly rare for this to be the case. The point is, rather, that there are two ways of understanding someone’s actions. One is a third-person perspective; the other is a first-person perspective. Neither is uniquely correct. But freedom appears only in the first-person perspective.
Boyer, Pascal. Religion Explained: The Evolutionary Origins of Religious Thought. New York: Basic Books, 2001.
Damasio, Antonio R. Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Avon, 1994.
Hogan, Patrick Colm. Cognitive Science, Literature, and the Arts: A Guide for Humanists. New York: Routledge, 2003.
---. "Literature, God, and the Unbearable Solitude of Consciousness." Journal of Consciousness Studies 11.5/11.6 (May/June 2004): 116-42.
Jourdain, Robert. Music, the Brain, and Ecstasy: How Music Captures Our Imagination. New York: Quill, 2002.
Kahneman, Daniel and Dale T. Miller. "Norm Theory: Comparing Reality to Its Alternatives. Psychological Review 93.2 (1986): 136-53.
Kosslyn, Stephen. Image and Brain: The Resolution of the Imagery Debate. Cambridge, MA: MIT P, 1994.
LeDoux, Joseph. The Emotional Brain: The Mysterious Underpinnings of Emotional Life. new York: Touchstone, 1996.
---. Synaptic Self: How Our Brains Become Who We Are. New York: Viking, 2002.
Matthews, Gary G. Neurobiology: Molecules, Cells, and Systems. 2nd ed. Malden, MA: Blackwell Science, 2001.
Panksepp, Jaak. Affective Neuroscience: The Foundations of Human and Animal Emotions. New York: Oxford UP, 1998.
Pinker, Steven. How the Mind Works. New York: W. W. Norton, 1997.
Rubin, David. Memory in Oral Traditions: The Cognitive Psychology of Epic, Ballads, and Counting-out Rhymes. New York: Oxford UP, 1995.
Tale of the Heike. Trans. Helen Craig McCullough. Stanford, CA: Stanford UP, 1988.
Tooby, John and Leda Cosmides. "Does Beauty Build Adapted Minds? Toward an Evolutionary Theory of Aesthetics, Fiction, and the Arts." Substance: A Review of Theory and Literary Criticism 30.1/30.2 (2001): 6-27.
Walter, Henrik. Neurophilosophy of Free Will: From Libertarian Illusions to a Concept of Natural Autonomy. Trans. Cynthia Klohr. Cambridge: MA: MIT P, 2001.
Zajonc, Robert B. "Feeling and Thinking: Closing the Debate Over the Independence of Affect." In Feeling and Thinking: The Role of Affect in Social Cognition. Ed. Joseph P. Forgas. Cambridge: Cambridge UP and Paris: Editions de la Maison des Sciences de l’Homme, 2000, 31-58.
Zhang Yimou, dir. Hero. Screenplay by Li Feng, Zhang Yimou, and Wang Bin. Story by Zhang Yimou, Li Feng, and Wang Bin. Music by Tan Dun. Cinematography by Christopher Doyle. Edited by Zhai Ru and Angie Lam. Produced by Bill Kong and Zhang Yimou. Elite Group Enterprises, 2002.
Received: January 1, 2005, Published: July 18, 2005. Copyright © 2005 Patrick Colm Hogan