1. Introduction
Amongst other things, a theory of perceptual experience is supposed to explain what it’s like to have perceptual experiences. A theory that can’t account for perceptual phenomenology ought to be rejected; and theories that provide better explanations of perceptual phenomenology than their competitors ought to be preferred to those competitors. Accordingly, much of the debate between naïve realism—the theory that to have a veridical perceptual experience is to stand in a non-representational awareness relation to ordinary objects and their properties—and the content view—the theory that to have a perceptual experience is to mentally represent ordinary objects and their properties—has focused on phenomenological considerations. More specifically, this debate has focused on a particular aspect of perceptual phenomenology. What it’s like to have a perceptual experience of some object is very different from what it’s like to imagine, or to merely consciously think about, that same object. For instance, when I see the book on the table in front of me, the book itself is just there before my mind in a way it never is when I visualize the book being on the table, or when I consciously think that the book is on the table. That is, perceptual experiences possess a phenomenal immediacy or presence that imagery experiences and conscious thoughts lack. The claim that naïve realism provides a more satisfying explanation of perceptual experience’s immediacy than does the content view has been perhaps the most influential motivation for that theory (see, for example, Crane Reference Crane, Gendler and Hawthorne2006: 139–141; Fish Reference Fish2009: 19–23; and Kennedy Reference Kennedy2009: 578–580).
Contrary to what is commonly assumed, the content theorist has a distinct advantage when it comes to explaining this aspect of perceptual phenomenology—an advantage due to the fact that hallucinatory experiences possess the same phenomenal immediacy that veridical experiences possess. The naïve realist says that there is a non-representational mental relation, acquaintance, such that when you stand in that relation to ordinary objects and their properties, those objects and properties seem to be immediately present to you. That is, the naïve realist explains why perceptual experiences possess their characteristic immediacy by positing a mental relation that, by its very nature, makes it the case that perceptual experiences possess this phenomenal feature. Unsurprisingly, the content theorist can employ the same strategy—call it the nature-of-the-relation strategy. The content theorist can say that there is a representational mental relation, perceptual representation, such that when you stand in that relation to a proposition (or property-complex, or what have you), the objects and properties you thereby represent seem to be immediately present to you—I’ll call this specific variety of the content view representationism (for content theorists who consider or endorse this strategy for explaining immediacy, see Chalmers Reference Chalmers and Leiter2004; Pautz Reference Pautz and Hawthorne2007: 519, and 2021: chap. 3; Schellenberg Reference Schellenberg2011; and Schroer Reference Schroer2012: 599). But, while the naïve realist can’t appeal to acquaintance in order to explain why hallucinations possess immediacy, the content theorist can appeal to perceptual representation in order to explain this fact. Consequently, the content theorist can construct a revised version of the argument from hallucination against naïve realism.Footnote 1 Very briefly: both hallucinatory and veridical perceptual experiences possess immediacy; the content view provides a unified explanation of this fact, while naïve realism can’t; we should prefer theories that provide unified explanations to theories that don’t; so we should prefer the content view to naïve realism.
Now, if immediacy were the only phenomenal feature that a theory of perceptual experience ought to explain, then we could leave matters there; but we can’t because perceptual experiences are phenomenally complex. In particular, while veridical and hallucinatory experiences share a phenomenal feature that neither shares with conscious thoughts or imagery experiences, perceptual experiences also share a phenomenal feature with imagery experiences that neither shares with conscious thoughts. That is, perceptual experiences and imagery experiences always instantiate sensory phenomenology; conscious thoughts are often accompanied by sensory phenomenology, but they need not be. For instance, when I visualize the book being on the table in front of me, just as when I see the book on the table, the book is presented to me in a concrete, sensory way; whereas when I merely consciously think that the book is on the table (in the absence of any accompanying imagery), my thought represents the book in an abstract, non-sensory way. Even though the book itself doesn’t seem to be just there before my mind when I visualize it located on the table, my imagery experience, like my perceptual experience and unlike my conscious thought, visually presents the book. One way to express this point would be to say that imagery experiences involve presentation without presence. Accordingly, I’ll call this phenomenal feature that perceptual and imagery experiences share and that conscious thoughts lack, presentational character, or presentationality (this notion will be discussed at greater length in §3.1 below).
The moral of the argument from hallucination is that when two sorts of experience share a given phenomenal feature, we should prefer theories that provide a unified explanation of this fact. Accordingly, we can construct an analogous argument from imagination targeting any theory that employs the nature-of-the-relation strategy to explain immediacy. Very briefly: both perceptual experiences and imagery experiences possess a presentational character that distinguishes them from conscious thoughts; neither the naïve realist nor the representationist can provide a unified explanation of this fact; so, if some version of the content view can provide a unified explanation of why perceptual and imagery experiences possess presentationality, then we ought to prefer that theory to both naïve realism and representationism.Footnote 2 Now, if the relevant version of the content theory can’t also provide a unified explanation of immediacy, we won’t have made any progress. So, the overarching argument is: if some version of the content view can provide unified explanations of why hallucinatory and veridical perceptual experiences possess immediacy, and why perceptual and imagery experiences possess presentationality, then that theory provides a better explanation of the phenomenal facts than does either representationism or naïve realism.
I maintain that a particular variety of the content view—the sensory vehicle theory—can provide satisfying, unified explanations of both immediacy and presentationality. In §2, I’ll make the case that the argument from hallucination establishes that we should prefer representationism to naïve realism (since the former provides a better explanation of immediacy than the latter). In §3, I’ll make the case that the argument from imagination establishes that a variety of the content view that provides a unified account of presentationality should be preferred to both naïve realism and representationism. Finally, in §4, I will outline the sensory vehicle theory and describe how it provides unified explanations of the relevant phenomenal features. I won’t attempt to show that no other variety of the content view can provide unified explanations of both immediacy and presentationality. The point to be established is just that at least some variety of the content view can provide unified explanations of these phenomenal features; and as such, we have compelling reasons to endorse such a theory and to reject both naïve realism and representationism.
2. Representationism Explains Immediacy Better than Naïve Realism Does
Some philosophers suggest that only naïve realism can provide a satisfying account of what it’s like to have a perceptual experience—that naïve realism is the only theory that takes the deliverances of introspection at face value. One way to defend this thesis is to claim that when you reflect on, for instance, your visual experience of the book on the table, the book doesn’t seem to be something distinct from, and merely causally related to, your experience—rather, the book seems to be a constituent of your experience (see, for example, Nudds Reference Nudds2009: 334–335; Langsam Reference Langsam2017: 110–111; and Allen Reference Allen2020: 644–645; for discussion, see Pautz Reference Pautz2021: §5.3). And, the naïve realist claims that to have a veridical perceptual experience is to stand in a non-representational awareness relation—the acquaintance relation—to ordinary objects and their properties; so, on this view, when you see the book on the table, that very book is a constituent of your experience. Conversely, according to the content view, to have a perceptual experience is to mentally represent ordinary objects and their properties; and you could have a perceptual experience that represents that the book is on the table even if there were no such book.
However, the claim that introspection reveals that ordinary objects are constituents of experiences is implausible: a perceptual experience is about some object, and this fact is evident from introspection; if the object an experience concerns is a component of the experience, and this fact is evident from introspection, then introspection reveals that the experience is about some part of itself—but it’s not the case that introspection reveals that a perceptual experience is about some part of itself. More generally, we should deny that simple introspection directly supports naïve realism in this fashion. When you reflect on your experience of the book, you seem to be directly aware of the book; and, as we’ve seen, the book itself seems to be present to you in a way it never does when you imagine or consciously think about the book. But the content view entails that you are directly aware of the book, just as naïve realism does; and at least certain versions of the content view entail that the book itself should seem present to you in a way it never does when you imagine or consciously think about the book. So, introspective observations of the sort at issue simply don’t entail that perceptual experiences constitutively involve acquaintance rather than representation.
Rather than claiming that naïve realism is the only theory consistent with the deliverances of introspection, then, a more plausible suggestion is that naïve realism simply provides a more satisfying explanation of perceptual experience’s immediacy than the content view does. The naïve realist understands acquaintance to be an awareness relation more basic than mental representation (see, for example, Fish Reference Fish2009: 14–15; and Raleigh Reference Raleigh, Knowles and Raleigh2019: 1–2). Mental representation allows for error: mental representations can ascribe properties to objects that they don’t possess; and represented objects need not even exist. Conversely, for you to stand in the acquaintance relation to a particular object is for that object itself simply to be given or revealed to your mind, just as it is. So, because the naïve realist maintains that to have a veridical perceptual experience is to be acquainted with some ordinary physical object and its properties, she has a ready explanation of why such experiences possess immediacy: standing in the acquaintance relation to some object makes it the case that that object itself seems to you to be just there before your mind. And because the content theorist maintains that to have a perceptual experience is to mentally represent some ordinary object and its properties, one might suggest that she won’t be able to explain why such experiences possess immediacy—after all, imagery experiences and conscious thoughts are representational mental episodes, and they don’t possess immediacy.
However, the content theorist is at least as well positioned as the naïve realist to explain perceptual experience’s immediacy because he can employ the naïve realist’s own strategy. The crucial point here is that while the naïve realist appeals to acquaintance to explain why perceptual experiences possess immediacy, it’s not possible to explain what acquaintance is supposed to be without appealing to this very phenomenal feature. The naïve realist tells us that perceptual experiences seem to make ordinary objects immediately present to our minds because, when we’re acquainted with ordinary objects, they really are immediately present to our minds. But we don’t have any independent grip on the kind “immediate presence” acquaintance involves—the sense in which acquaintance makes objects immediately present to us is just the sense in which conscious perception, unlike imagery or conscious thought, makes objects immediately present to us. That is, in order to explain the immediacy of perceptual experience, the naïve realist employs the nature-of-the-relation strategy: she explains why a certain sort of experience possesses a certain phenomenal feature by positing a mental relation which she characterizes as the thing that makes it the case that experiences possess the relevant phenomenal feature.
The content theorist can employ this strategy by introducing a unique representational relation—the perceptual representation relation—that individuals stand in to propositions. There are many different representational relations: the belief relation, the desire relation, the imagining relation, and so on. And the content theorist can say that (at least part of) what distinguishes the perceptually-representing-relation from the rest is that experiences constituted by perceptually representing propositions possess immediacy. That is, the perceptually-representing-relation is characterized as the kind of relation to a proposition such that, when you stand in this relation to a proposition, the objects and properties that you thereby represent seem to be immediately present to you. Accordingly, the content theorist who employs this strategy—the representationist—can readily explain why perceptual experiences possess immediacy: perceptual experiences are constituted by perceptually representing propositions, and perceptually representing propositions makes it the case that one’s experience possesses immediacy.Footnote 3
By design, the representationist’s explanation of perceptual experience’s immediacy is just as satisfying as the naïve realist’s explanation. The naïve realist can’t complain that the representationist’s strategy is unmotivated or obscure: the motivation for positing the perceptually-representing-relation is precisely the same as the motivation for positing the acquaintance relation; and our understanding of perceptual representation is at least as clear as our understanding of acquaintance. Granted, the representationist’s explanation of immediacy is uninformative: the perceptually-experiencing-relation can’t be characterized independently of the phenomenal feature it is supposed to explain. But, of course, the naïve realist’s explanation is uninformative for precisely the same reason.
However, while these explanations are equally satisfying, they are not equally good: relative to naïve realism, representationism explains more relevant facts—specifically, facts concerning the phenomenology of hallucinatory perceptual experiences. As such, the representationist can construct a compelling argument from hallucination against naïve realism. First, individuals sometimes have hallucinatory perceptual experiences that possess a phenomenal immediacy that distinguishes them from imagery experiences and conscious thoughts. Naïve realists sometimes deny this claim (Martin 2004; Fish Reference Fish2009); but it is supported by considerable introspective, behavioral, and neurological evidence (for a review of that evidence, see Millar Reference Millar2014). Given the weight of the evidence, then, a naïve realist who denies that some hallucinations possess immediacy is at a significant disadvantage relative to representationism. Second, naïve realism can’t provide a unified explanation of the immediacy of veridical and hallucinatory experiences. The naïve realist maintains that veridical experiences possess immediacy because such experiences constitutively involve acquaintance with ordinary physical objects; but when you hallucinate, you are not acquainted with any relevant ordinary physical object. So, any explanation of hallucinatory experiences’ immediacy that the naïve realist employs must be wholly novel—for instance, she might claim that hallucinations constitutively involve mental representation, or acquaintance with mind-dependent objects, or mental imagery.Footnote 4 Accordingly, third, if the content theorist can provide a unified explanation of the immediacy of veridical and hallucinatory experiences, then the content view ought to be preferred to naïve realism. And, fourth, the representationist can do just that. According to the representationist, to hallucinate is to stand in the perceptually-representing-relation to a proposition in the absence of any appropriate causal connection to an ordinary object. As such, hallucinatory experiences possess immediacy for the very same reason that veridical experiences do: they constitutively involve perceptually representing propositions.
3. Presentationality Presents a Challenge to Both Naïve Realism and Representationism
Representationism, then, has a significant advantage over naïve realism with respect to explaining perceptual experience’s immediacy. But, if we want to determine which theory of perceptual experience explains perceptual phenomenology best, we shouldn’t focus on immediacy exclusively. Immediacy is a perspicuous phenomenal feature that distinguishes perceptual experiences from conscious thoughts and imagery experiences; presentationality is a perspicuous phenomenal feature that distinguishes perceptual and imagery experiences from conscious thoughts. And while the fact that both veridical and hallucinatory experiences possess immediacy creates difficulties for naïve realism, the fact that both perceptual and imagery experiences possess presentationality creates difficulties for naïve realism and representationism alike.
To demonstrate that presentationality creates difficulties for both naïve realism and representationism, two points need to be established. First, that perceptual and imagery experiences do in fact share the phenomenal feature at issue. And second, that neither naïve realism nor representationism can provide a unified explanation of the presentationality of perceptual and imagery experiences. I will address each point in turn.
(I should note that while I’m using “imagery experiences” to cover both sensory imagination and episodic memory, I will typically focus on the former (on the relevant notion of episodic memory, see Byrne Reference Byrne2010: §2). Sensory imagination occurs in every perceptual modality and can be either voluntary or involuntary. So, the sorts of experiences that are relevant for present purposes are: the experience you have when you close your eyes and visualize the setting sun; the experience you have when an image of a face spontaneously appears before your “mind’s eye” as you’re reading a novel; or, the experience you have when you imagine how the white wall you’re currently looking at would appear if it were red.)
3.1. Presentationality
First, introspection provides compelling evidence that perceptual and imagery experiences present their subject matter in a concrete, sensory manner, rather than an abstract, non-sensory manner. For instance, just as you can’t see an object that is both triangular and circular, neither can you visualize such an object; conversely, you have no difficulty consciously entertaining the proposition that some object is both triangular and circular (in other words, Pautz’s (Reference Pautz2020) “laws of appearance” apply to both perceptual and imagery experiences). Or, consider the fact that when you visualize a triangle, just as when you see a triangle, your experience always includes details such as how the triangle is oriented in space, or whether it is equilateral; conversely, your thoughts about triangles often leave such features unspecified. Another way to illustrate the present point: whenever you have either visualized or seen a triangle, you can sensibly compare the triangle your experience presented to a picture of a triangle—you would know how to go about answering the question, “does the triangle in this picture resemble the triangle your experience presented?” Conversely, with respect to your thoughts about triangles, such comparisons are not always sensible.
Introspection also reveals that we can compare our imagery experiences to our perceptual experiences with respect to their vividness. Because conscious thoughts present their subject matter in an abstract, non-sensory manner, there is no sense in which our conscious thoughts are more or less vivid than our perceptual experiences. But we readily judge that our imagery experiences are typically less vivid than our perceptual experiences; and we can readily determine whether a given imagery experience’s vividness is more or less similar to that of a typical perceptual experience. Vividness is not a matter of how clear or detailed an experience is: a visual experience of an object in low light, or at the periphery of the visual field, can still be more vivid than an imagery experience that presents that same object more clearly or in greater detail.Footnote 5 Rather, vividness judgments concern intensity; and while we can’t provide an informative definition, we can make the relevant notion of intensity tolerably clear with the right examples. For instance, suppose that you first view a projection of a red circle on a screen, and then later view a much fainter projection of the circle on the same screen; next, suppose that you visualize the red circle so successfully that you would describe your imagery experience as “almost like seeing the circle”; and finally, suppose that you visualize the red circle so unsuccessfully that you would describe your experience as “not at all like seeing the circle.” There is an evident sense in which each experience in this sequence is less vivid or intense than the last. And we can only make these intensity comparisons because perceptual and imagery experiences share a specific phenomenal feature that neither shares with conscious thoughts. Moreover, Pearson, Rademaker, and Tong (Reference Pearson, Rademaker and Tong2011) showed that an individual’s judgments concerning the vividness of her imagery experiences correlates with that experience’s influence on a subsequent perceptual experience. And brain imaging studies have shown that when individuals judge imagery experiences to be more vivid, the neurological activity in their brains is more similar to that of perceptual experiences (Dijkstra, Bosch, and van Gerven Reference Dijkstra, Bosch and van Gerven2017; Fulford et al. Reference Fulford, Milton, Salas, Smith, Simler, Winlove and Zeman2018). The best explanation of these results is that imagery experiences can be more or less similar to perceptual experiences with respect to their intensity, and that individuals can typically determine an imagery experience’s intensity via introspection.
In addition, there is a good deal of behavioral and neurological evidence for the claim that perceptual and imagery experiences share a certain phenomenal feature. For instance, researchers have found that imagery experiences are subject to at least some prototypically perceptual illusions. The horizontal-vertical illusion occurs when individuals view a figure consisting of two lines, one horizontal and the other vertical; although the lines are the same length, the vertical line appears to be longer than the horizontal line. When individuals are asked to visualize this figure and then determine the length of the visualized lines, their imagery experiences mislead them: they judge the vertical line to be longer than the horizontal line—in fact, the magnitude of the illusion is the same whether the figure is visualized or perceived (Blanuša and Zdravković Reference Blanuša and Zdravković2015). Presumably this error is only possible because imagery experiences present the lines in the same concrete, visual way that perceptual experiences do.
The claim that imagery and perceptual experiences are phenomenally similar in the way at issue—that both present their subject matter in a concrete, sensory manner that contrasts with the abstract, non-sensory manner characteristic of conscious thoughts—is also suggested by the fact that imagery experiences have been shown to influence our perceptual experiences in ways that conscious thoughts do not. For instance, Pearson, Clifford, and Tong (Reference Pearson, Clifford and Tong2008) showed that visualization has the same influence on binocular rivalry experiments as perceiving “weak visual images.” Binocular rivalry involves one image being presented to one eye while a quite different sort of image is presented to the other; in such cases, one or the other of these images dominates the subject’s resulting perceptual experience. When the binocular rivalry stimulus is presented to a particular subject repeatedly under controlled conditions, each image will dominate the subject’s experience for roughly half of the trials. However, Pearson and colleagues showed that asking a subject to visualize one of the two images beforehand makes it significantly more likely that that image will dominate the subject’s subsequent experience—just as perceiving one of the two images beforehand does. Relatedly, visualization has been shown to have perception-like impacts on perceptual learning. For instance, Tartaglia et al. (Reference Tartaglia, Bamert, Mast and Herzog2009) showed that when learning to judge the distance between two lines, visualizing a line to occupy a certain location improves performance to the same extent that perceiving a line in that location does.
Finally, the fact that the neural activity of a visualizing brain is very similar to that of a visually perceiving brain provides further evidence of the phenomenal similarity between imagery and perceptual experiences. Because visualization relies on memory rather than perceptual inputs, we wouldn’t expect visualization to involve all and only the brain regions that conscious visual perception involves. But if the brain regions that are central to visual perceptual experience are also central to visual imagery experiences, then this fact would support the thesis that visual imagery experiences present their subject matter in the concrete, sensory way that visual perceptual experiences do. And, in fact, there is considerable neuroscientific evidence, primarily from imaging studies, that visualization is subserved by neural activity in the brain’s visual areas.Footnote 6 In fact, as described above, the imagining evidence suggests that more vivid visual imagery experiences tend to involve more activity in the brain’s visual areas.
3.2. Naïve Realist and Representationist Explanations
Naïve realists are not able to provide a unified explanation of the presentationality of imagery and perceptual experiences because they maintain that perceptual phenomenology is grounded in acquaintance with ordinary objects and their properties. Just as naïve realists say that your perceptual experiences possess immediacy in virtue of the fact that they constitutively involve acquaintance with ordinary objects, they say that your perceptual experience possesses presentationality in virtue of the fact that they constitutively involve acquaintance with ordinary objects. (Presumably, a given naïve realist will explain the presentationality of hallucinations using whatever strategy he uses to explain the immediacy of hallucinations.) But, of course, imagery experiences don’t constitutively involve acquaintance with ordinary objects.
One might think that the naïve realist can provide a unified explanation of presentationality by maintaining that imagery experiences constitutively involve acquaintance with mental images and their properties. If ordinary objects and mental images instantiate the same properties, then the naïve realist might claim that it’s acquaintance with these properties that makes it the case that experiences possess presentationality—in which case, imagery experiences possess presentationality for the very same reason that perceptual experiences do. However, such an account is unacceptable for at least two reasons. First, our imagery experiences concern ordinary objects, not mental images—when you visualize a book on the table in front of you, you’re thinking about a book, not a mental image of a book (see Kind Reference Kind2001: 106–107). Second, the naïve realist characterizes acquaintance as the kind of non-representational awareness that confers immediacy; but imagery experiences don’t possess immediacy. For instance, when you visualize a book sitting next to a book that you see, it’s not the case that there are two entities—one an ordinary physical object, the other a mental image—that seem to be just there before your mind. (In order to address this last point, the naïve realist might claim that there are different ways of being acquainted with objects, and that imagery experiences constitutively involve a kind of acquaintance that doesn’t confer immediacy. However, the resulting theory would not provide a unified explanation of presentationality: according to this theory, one acquaintance relation provides perceptual experiences with presentationality, and a distinct acquaintance relation provides imagery experiences with presentationality.)
Alternatively, one might think that the naïve realist can provide a unified explanation of presentationality by adopting Martin’s (Reference Martin2002) view that to imagine some object is to mentally represent a perceptual experience of that object (see also Soteriou Reference Soteriou2013: chap. 7). The suggestion would be that an imagery experience’s presentational character is derived from that of the perceptual experience it represents—visualizing possesses presentationality because it constitutively involves representing a visual experience that possesses presentationality. However, there are many representational mental states that represent experiences without thereby acquiring any of their phenomenal features. For instance, you might believe that you saw a certain famous painting ten years ago; but what it’s like for you to believe this proposition is in no way similar to what it was like for you to see the painting. So, as it stands, Martin’s proposal doesn’t explain why imagery experiences acquire some of the phenomenal features of perceptual experiences while other mental representations don’t—as Martin (Reference Martin2002, 406–407) explicitly acknowledges. Moreover, no matter what the naïve realist adds to her theory in order to explain how imagery experiences derive their presentationality from that of the perceptual experiences they represent, the resulting theory will not provide a unified explanation: according to any such theory, the feature that makes it the case that perceptual experiences possess presentationality is distinct from the feature that makes it the case that imagery experiences possess presentationality.
The representationist’s explanatory options are similarly restricted. Because the representationist employs the nature-of-the-relation strategy to explain perceptual experience’s immediacy, it is natural to explain its presentationality the same way. That is, it is natural for the representationist to claim that (at least part of) what distinguishes the perceptually-representing-relation from other representational relations is that experiences constituted by perceptually representing propositions possess both presentationality and immediacy. But, then, imagery experiences don’t constitutively involve the perceptually-representing-relation: the perceptually-representing-relation is characterized as the kind of relation to a proposition such that, when you stand in this relation to a proposition, the objects and properties that you thereby represent seem to be immediately present to you; and the objects and properties your imagery experiences represent don’t seem to be immediately present to you (imagery experiences, unlike perceptual experiences, lack immediacy). So, the representationist will need to posit a distinct representational relation—we could call it the imaginative representation relation—which is such that experiences constituted by standing in this relation to propositions possess presentationality but not immediacy. The result of this natural development of the view is that the representationist doesn’t provide a unified explanation of presentationality: perceptual representation makes it the case that perceptual experiences possess presentationality, and imaginative representation makes it the case that imagery experiences possess presentationality.Footnote 7
Accordingly, in order to provide a unified explanation of presentationality—an explanation where the feature that makes it the case that perceptual experiences possess presentationality is also the feature that makes it the case that imagery experiences possess presentationality—the representationist must refuse to employ the nature-of-the-relation strategy to explain this phenomenal feature. That is, the representationist must refuse to include presentationality in his characterization of the perceptually-representing-relation, and explain why both imagery and perceptual experiences possess presentationality some other way. However, the most natural options for providing such an explanation are not promising. For instance, one might claim that imagery and perceptual experiences share a certain sort of representational content, and it’s this shared content that makes it the case that both kinds of experience possess presentationality (see Nanay Reference Nanay2015). But this proposal can’t succeed unless we have good reasons to insist that no conscious thought ever has the same content as any perceptual or imagery experience; and we don’t have such reasons. Alternatively, the representationist can posit non-representational phenomenal properties, or qualia, and claim that instantiating qualia is what makes it the case that perceptual and imagery experiences, unlike conscious thoughts, possess presentationality. But this proposal is unacceptable because the phenomenal properties it posits play no role in explaining immediacy. If perceptual experiences instantiate qualia, then surely it’s at least partly because my present visual experience instantiates a reddish quale (rather than a greenish quale, or none at all) that it seems to me that something red is immediately present to me. And yet, according to the present view, what makes it the case that something red seems to be immediately present to me is simply the fact that I stand in the perceptually-representing-relation to a proposition concerning a nearby red object; the non-representational red quale merely accompanies the experience’s representational component, and even if this red quale were absent, it would still seem that something red is immediately present to me.
We should conclude, then, that no theory that employs the nature-of-the-relation strategy to explain immediacy will be able to provide a unified explanation of presentationality; and as such, if some variety of the content view could provide a unified explanation of presentationality, then that fact would enable us to construct a compelling argument from imagination against both naïve realism and representationism. This argument would proceed as follows: First, perceptual experiences and imagery experiences possess a presentational character that distinguishes them from conscious thoughts. While this claim is nearly universally accepted amongst philosophers and psychologists (see Thomas Reference Thomas and Zalta1997/2014), some naïve realists and representationists might deny it—just as some naïve realists deny that hallucinatory experiences possess immediacy. However, given the weight of the evidence surveyed above (§3.1), a naïve realist or representationist who denies this assumption is at a significant disadvantage relative to any theory that can explain why our experiences possess presentationality. Second, because naïve realism and representationism employ the nature-of-the-relation strategy to explain immediacy, they can’t provide a unified explanation of presentationality. And so, third, if some version of the content view can provide a unified explanation of why perceptual and imagery experiences possess presentationality, then we have a compelling reason to prefer that theory to both naïve realism and representationism. Moreover, if that theory can also provide a unified explanation of immediacy, then that theory provides a better explanation of the phenomenal facts than does either representationism or naïve realism.
(It’s worth emphasizing that naïve realism and representationism are not on equal footing at this point in the argument. While representationism can’t provide a unified explanation of presentationality, it does provide a unified explanation of immediacy. Conversely naïve realism’s account of perceptual phenomenology is doubly disunified: the theory can’t provide unified explanations of either presentationality or immediacy. So, if there were no other alternatives, representationism would provide the best explanation of the relevant phenomenological facts. The point to be established in the next section is that some version of the content view can do better still.)
4. An Alternative Version of the Content View: Sensory Vehicle Theory
Consider a scenario in which viewing conditions are rather poor (light levels are low, you’re squinting, etc.) and all you can make out is a small black dot directly in front of you. Now, suppose you close your eyes and visualize the black dot just as it appeared a moment ago. And suppose that, after the visual image fades, you consciously think that there is a small black dot directly in front of you. Assuming that these three distinct experiences are representational episodes, they don’t appear to differ with respect to what they represent—we can stipulate that each accurately represents all and only the same details concerning the black dot. Rather, your perceptual and imagery experiences appear to differ from your conscious thought only with respect to how they represent the black dot. That is, although perceptual and imagery experiences are mental representations of different kinds, there is nonetheless a basic commonality with regards to how they represent their subject matter. Plausibly, then, any satisfying, unified explanation of this similarity will posit a similarity with respect to how perceptual and imagery experiences represent what they do. And a particularly natural proposal is that perceptual and imagery experiences represent what they do via similar representational vehicles.
The sensory vehicle theory is a variety of the content view that provides an account of just this sort. As the name suggests, the theory maintains, first, that perceptual experiences constitutively involve representational vehicles, and second, that these representational vehicles are constituted by an experience’s sensory phenomenology. (For defenders of this view, see Hall Reference Hall1961; Clark Reference Clark1973; and Millar Reference Millar2017, and Reference Millar2022: §6. For similar views, see Wishon Reference Wishon, Miguens and Preyer2012, and Papineau Reference Papineau2014 and Reference Papineau2021.) Sensory phenomenology is a particular class of phenomenal properties; specifically, it includes the kinds of phenomenal properties that perceptual and imagery experiences possess but which conscious thoughts lack. Sensory phenomenal properties can be divided into two classes: those aspects of sensory phenomenology associated with the experience of physical properties—which I will call sensory qualities—and those aspects of sensory phenomenology associated with the experience of the objects that instantiate those properties—which I will call sensations. So, for example, consider what your experience is like when you view a red object under white light and a white object under red light. These two experiences share a certain specific aspect of their phenomenology that neither shares with your experience of a white object under white light. This shared aspect of these two experiences’ phenomenology is a sensory quality we can call phenomenal redness. We could use more detailed examples to isolate the different shades that make up the class of red qualities; and the class of red qualities is included in the more inclusive class of colour qualities. A similar procedure can be used to identify the phenomenal properties characteristic of visual experiences of shape, size, and spatial location. Further, when you have a visual experience as of a particular object instantiating a variety of visible properties, the corresponding sensory qualities are united in a particular sensation. The sensation corresponding to the perceptual experience of an object is not simply a collection of sensory qualities; added to the relevant sensory qualities is a distinct phenomenological unity that makes a sensation a single phenomenal entity. (Using terminology introduced by Bayne and Chalmers (Reference Bayne, Chalmers and Cleeremans2003), we would say that a sensation is instantiated if and only if multiple sensory qualities are objectually phenomenally unified.)
These phenomenal properties are representational vehicles in the sense that when you have a perceptual experience of a particular object and its properties, your experience represents what it does in virtue of the fact that it instantiates sensations and sensory qualities related to that object and those properties in the right way. In other words, according to the present theory, sensory phenomenology functions in sensory experience in much the same way that linguistic symbols function in thought. For example, when you see a particular square object in front of you and you token the sentence “that is square” in inner speech, by tokening these linguistic symbols you thereby represent a proposition that concerns the relevant physical object and the property of being square. Similarly, when your visual experience instantiates a particular sensation partly composed of phenomenal squareness, that experience thereby represents a proposition that concerns a particular physical object and the property of being square.
Sensory vehicle theory doesn’t entail any particular theory of the metaphysical nature of sensory qualities and sensations. For all the theory says, these phenomenal properties might be non-physical properties of non-physical mental states, or they might be wholly physical properties of wholly physical states—certain neurological properties of brain states, for example. However, the theory is committed to a quite specific account of what sensory qualities and sensations do. Specifically, sensory vehicle theory maintains that sensory phenomenology is what you represent with rather than what you represent. For instance, the theory is incompatible with the traditional sense datum theory, since that theory maintains that to have a perceptual experience is to be aware of sense data—and you are aware of ordinary objects and properties in virtue of being aware of sense data. Conversely, according to sensory vehicle theory, you are not aware of a perceptual experience’s sensory phenomenology in the same sense that you are aware of the objects and properties your experience represents. By way of illustration, consider the analogy with inner speech again: when you think about objects and properties using inner speech, tokening names and adjectives constitutes a way of thinking about objects and properties that does not involve an explicit awareness of the words themselves—you think with those words, not about them.Footnote 8
I should perhaps also add that sensory vehicle theory doesn’t entail any particular theory about how sensory phenomenal properties come to represent what they do. For instance, for all the theory says, sensory phenomenology might be inherently representational, and the specific representational contents of particular sensory qualities might be innate. Or, these phenomenal properties might not be inherently representational. For example, the sensory vehicle theorist could endorse Papineau’s (Reference Papineau2021, 6) view that a brain in a vat created by a cosmic accident might undergo an experience instantiating all the same sensory phenomenal properties that my present visual experience possesses, without that experience representing anything at all. On the other hand, even assuming that the phenomenal properties at issue are not inherently representational, it doesn’t follow that they must be causally related in the right way to environmental properties in order to represent those properties. It might be, for instance, that at least some sensory phenomenal properties represent what they do via a satisfactional rather than relational mechanism (Millar Reference Millar2022). In any case, the success or failure of the argument from imagination doesn’t depend on these details of the theory.
The sensory vehicle theory, then, has a ready explanation of presentationality: instantiating sensations and sensory qualities is what makes it the case that an experience possesses presentationality. And since both perceptual and imagery experiences instantiate sensations and sensory qualities, the theory provides a unified explanation: the feature that makes it the case that perceptual experiences possess presentationality is the very same feature that makes it the case that imagery experiences possess presentationality (for a brief discussion of a view of this sort, see Kind Reference Kind2001: 108–109). Moreover, this account is particularly satisfying because it captures the natural suggestion that what distinguishes perceptual and imagery experiences from conscious thoughts is how they represent what they do, rather than what they represent. And since sensory qualities are phenomenal properties that can vary in intensity, the theory provides a simple account of variations in vividness: due to the different processes via which they are generated, a perceptual experience’s sensory qualities are typically more vivid than an imagery experience’s sensory qualities.
Now, the sensory vehicle theory has a significant advantage over both naïve realism and representationism only so long as it can also provide a unified explanation of immediacy. In order to explain immediacy, then, the sensory vehicle theorist can point to how sensory phenomenal properties are generated and how they are used. In visual perception, sensations and sensory qualities are generated automatically by the visual system in response to current retinal stimulation and via a process over which the subject has no voluntary control; when visual sensations and sensory qualities are generated in this manner, they are being used as temporary labels for physical objects and properties that are currently causally impacting the subject’s retinas. Conversely, imagery experiences involve sensations and sensory qualities being generated internally (typically via a voluntary process) and not as a response to currently ongoing sensory input. According to the sensory vehicle theorist, when you have a perceptual experience, your experience constitutively involves the awareness (or the sense) that your experience is an externally-generated response to the current state of objects in your environment. And although the component of the experience representing this fact is distinct from the way the experience represents your environment to be (the way your experience represents your environment to be is determined by the sensations and sensory qualities it instantiates), it is an intrinsic representational component of the experience, and it grounds the intrinsic phenomenal component of immediacy. (It’s worth emphasizing that this strategy for explaining the immediacy of perceptual experiences entails that such experiences are not wholly composed of sensory phenomenal properties, as, for example, Papineau (Reference Papineau2021) claims.) In other words, the present proposal is that perceptual experiences possess immediacy in virtue of the fact that they are partly composed of a metacognitive representation with a distinctive content.Footnote 9 Conversely, imagery experiences lack immediacy because they do not constitutively involve a metacognitive representation with the same content.
An analogy with episodic memory might help to clarify the present proposal. Compare remembering what the Hale-Bopp comet looked like in the night sky to visualizing what Hale-Bopp looked like in the night sky. These are both imagery experiences, and in both cases you’re mentally representing the same past event. Yet, there is a clear phenomenal difference between them—what it’s like to remember is distinct from what it’s like to simply imagine. And since there isn’t any obvious difference between the imagery involved, or the content of these experiences, the most natural explanation of this phenomenal difference is that when you remember you’re aware that you’re remembering, and when you imagine you’re aware that you’re imagining. In other words, the most natural account of episodic memory is that such experiences are phenomenally complex: there’s something it’s like to undergo the mental imagery—that component is shared by imagining—and, in addition, your episodic memory just seems like a memory. One way to capture these facts is to say that episodic memories instantiate sensory phenomenology in virtue of which they represent some past event, and that they also constitutively involve a metacognitive representation of the fact that the experience is a memory (see, for example, Dokic Reference Dokic2014, and Michaelian Reference Michaelian2016: Chap. 11).
Finally, the sensory vehicle theorist can easily extend the present account of the immediacy of perceptual experiences to hallucinations. Hallucinations, just like veridical perceptual experiences, instantiate sensory qualities and sensations that function as temporary labels for the physical objects that are causally impacting one’s visual system; the only difference is that when one hallucinates, there are no such physical objects. And just like veridical perceptual experiences, hallucinations constitutively involve the sense that your experience is an externally-generated response to the current state of objects in your environment; the only difference is that when you hallucinate, the relevant representation is generated in error.Footnote 10 And since the relevant representation is generated by the same sub-personal, belief-independent cognitive processes that generate sensory phenomenology, a hallucinatory experience can possess immediacy even though you know that you’re hallucinating. So, once again, the sensory vehicle theory provides a unified explanation: the feature that makes it the case that veridical experiences possess immediacy is the very same feature that makes it the case that hallucinatory experiences possess immediacy.
5. Conclusion
The moral of the argument from hallucination is that we should prefer theories that provide a unified explanation of the immediacy of veridical and hallucinatory experiences. The moral of the argument from imagination is that we should prefer theories that provide a unified account of the presentational character of perceptual and imagery experiences. The naïve realist can’t provide a unified explanation of either of these crucial phenomenal features, and the representationist can’t provide a unified explanation of presentationality; conversely, the sensory vehicle theory provides a unified explanation of both presentationality and immediacy. Because phenomenological criteria are not the only criteria by which we ought to evaluate theories of perceptual experience, we can’t say conclusively that we should reject naïve realism and representationism. But the fact that the sensory vehicle theory explains perceptual phenomenology better than these theories do provides a compelling reason to reject naïve realism and representationism and to endorse the sensory vehicle theory.
Open access
