The Metaphysics of Sensory Experience

Abstract

What is the nature of conscious sensory experience? In The Metaphysics of Sensory Experience David Papineau sets out to answer this question. He argues for the qualitative view: conscious sensory experiences are “intrinsic qualitative properties of people that are only contingently representational” (6).This book is instructive, engaging, original, full of argument, straight-talking, and it defends an interesting view! I enthusiastically recommend it to philosophers of mind and perception.Papineau’s central argument is a last view standing argument: wading through the detritus left after his assault on competitor views—naive realism and representationalism—one finds the qualitative view, standing strong. It is “the only option that makes good metaphysical sense” (8). Given this, I focus my critical attention in what follows on Papineau’s main negative arguments. First, a brief overview.The book has an introduction and four chapters. In chapter 1, Papineau argues against naive realism and begins a detailed critical investigation of representationalism. Papineau’s discussion of contingent and essential representationalism is particularly important, for it helps him to clarify his own disagreement with representationalism (30–32). Papineau thinks that conscious sensory properties can represent, but they need not. He accepts contingent representationalism but rejects essential representationalism—the mainstream position in the philosophy of perception. Consider words written on a page. In themselves these marks aren’t representationally significant. “Their representational power derives from the way they happen to be used by linguistic communities. But just those marks could easily have not been used in this way” (32). Papineau views conscious sensory properties as similar: such properties need not represent at all; they are representational only in the context of contingent external circumstances.Chapter 2 is where Papineau argues against essential representationalism. This chapter is a real highlight—a deep, original, and rewarding critical discussion. Papineau considers and rejects various ideas that might be used to motivate representationalism (e.g., the argument from the transparency of experience). He also helpfully weaves together themes that we don’t always find treated together in discussions of representationalism in the philosophy of perception. For instance, how general theories of representation interact with representationalism.In chapter 3, Papineau clarifies and develops his positive view. Like the previous chapters, this is jam-packed with interesting discussion. I’ll mention just two things. First, positioning. Papineau positions his view by aligning it with adverbialism (though not adverbialism about our ways of talking about sensory experience [83–84]). He also usefully compares his view to qualia views: these views typically hold that some sensory conscious properties conform to something like the qualitative model, but that these are additional to essentially representational properties. In contrast, Papineau denies that any conscious sensory properties are essentially representational (87).Second, Papineau attempts to explain why the structure and organization of experience can lead one to believe that experience is inherently directed on external items. For instance, elements of experience alter as I move around the external world—just what one would expect if those elements were inherently directed on stable items out there (compare afterimages, which, as it were, move with me). But none of this, Papineau suggests, means that these elements of experience really are inherently directed on external items. Here Papineau acknowledges that there is an “intuitive pull” to the idea that experiences somehow involve external items (91). Naive realists and representationalists explain this intuitive pull by claiming that experiences inherently relate one to or represent such external items. They will thus be motivated to question whether Papineau’s alternative explanation is adequate. Is this intuitive pull really intelligible if conscious sensory properties are purely qualitative, as Papineau suggests? I expect this aspect of the chapter to generate some critical discussion.Finally, in chapter 4 Papineau develops an account of introspection of sensory experiences, discusses how Jackson’s many properties problem for adverbialism (and related problems) don’t apply to his view, discusses the link between the qualitative view and the phenomenal concepts strategy against the knowledge argument, and develops new ideas about rich or high-level experience—among other things! This final chapter explores implications of the qualitative view and makes connections within and outside of the philosophy of perception.What then of Papineau’s negative arguments? (Note that Papineau develops various arguments against his key opponents. Here I have space to focus just on the main ones.) Let’s start with naive realism. Papineau’s main objection is the following.Papineau seems to be arguing that the naive realist commits to a radical disconnect between consciousness and introspection—pushing our very concept of consciousness to breaking point. This, he suggests, is a consequence of their commitment to conscious differences between veridical perceptions and matching hallucinations. The target here is naive realist disjunctivism, which holds that the conscious character of a veridical perception of a yellow ball (say) is constituted by acquaintance with aspects of the external world, whereas the conscious character of a subjectively matching hallucination is different—it has a very different nature.Papineau understands conscious character, standardly enough, in terms of what it is like for a subject to undergo an experience (16). We can understand, then, why it would be puzzling to claim that aspects of conscious character could be in principle introspectively undetectable. No doubt there is a lot more to say, but let’s grant that there cannot be aspects of what it is like for a subject to undergo a sensory experience that are in principle introspectively undetectable.The question, then, is whether committing to conscious differences between veridical perceptions and matching hallucinations as above entails that there are aspects of conscious character which are in principle introspectively undetectable? I will now argue, applying some ideas from French and Gomes 2019, section 6, that this is doubtful.Call one’s veridical perception of the yellow ball V, and one’s matching hallucination H. Now let’s invoke a relational model of the conscious difference between these experiences: it consists in the relation of being nonidentical that holds between the conscious characters of V and H. Now, suppose that we grant that this difference is introspectively undetectable. This just means that one cannot know, purely on the basis of introspection, that V and H have nonidentical conscious characters.However, this does not entail a radical disconnect between consciousness and introspection: it does not entail that there is any aspect of what it is like for one to undergo V, or H, that is not in principle available to introspection. For the relation of nonidentity between the characters of V and H is not an aspect of what it is like for one to undergo either H or V. Even if the best-placed introspector is unable to detect that the characters of V and H are nonidentical, they can still introspectively reflect on the character of V and on the character of H.Suppose instead that we invoke a difference maker model of the conscious difference between V and H. Consider here Dretske’s (2004: 9–10) discussion of change blindness. Your clean-shaven friend grows a mustache. The mustache makes for a difference between your friend at one time and your friend later. Similarly, the conscious character of V constitutively involves acquaintance with a mind-independent yellow ball-shaped object. This makes for a difference between V and H. And it is an aspect of what it is like for one to undergo V.But is the conscious difference between V and H, understood in this way, introspectively undetectable? No! Just as you might see a difference maker in seeing your friend’s mustache even if you fail to notice that they have changed, one might introspect an aspect of conscious character that makes two experiences consciously different, even if one cannot tell by introspection that these experiences are different.So, the naive realist can agree with Papineau that we had better not radically disconnect sensory consciousness from introspection. But they can question whether the conscious differences that they admit between veridical and hallucinatory experiences really do amount to such a disconnect.Consider, then, representationalism. Papineau’s main argument here is that conscious sensory properties and representational properties are incommensurable. Now, Papineau understands representationalists to be committed to the idea that when we sensorily represent a yellow ball this involves a mental relation to abstract, uninstantiated yellowness—for how else could a representationalist plausibly understand sensory representation of a yellow ball in a case of hallucination where no yellowness is instantiated? He thus argues as follows. (1) Instantiations of conscious sensory properties constitute concrete facts with causes and effects.(2) Instantiations of representational properties constitute abstract facts that cannot feature as causes or effects.Therefore,(3) Conscious sensory properties are not representational properties.Papineau defines concrete facts as “constituted by some spatio-temporal particular (or particulars) instantiating some first-order property (or relation).... They are localized in space and time and have causes and effects” (66–67). With this, and assuming that epiphenomenalism is implausible, (1) looks secure. But what about (2)? Papineau says thatNow, it might be that my sensorily representing a yellow ball is a matter of me, a spatiotemporal particular, instantiating a first-order property or relation—that of being mentally related to (abstract) yellowness. But that is not enough to make this fact a concrete fact unless it is in the causal nexus and spatiotemporally local (“here-and-now”). Indeed, Papineau claims that this fact (call it FACT) is an abstract fact, not a concrete fact. The question is whether Papineau has done enough to establish this.My worry is that the main point that Papineau seems to offer for the claim that FACT is abstract is that it involves abstract yellowness. But why should the claim that FACT involves something abstract make FACT itself abstract? Does it make it nonlocal? This is not obvious, since despite involving abstract yellowness, FACT is still a matter of a spatiotemporal subject existing here-and-now instantiating a property here-and-now. Similarly, consider the position of FACT in the causal nexus. It seems perfectly intelligible that certain causes in the subject’s environment and brain conjoin to produce an experience that is constitutively a relation between them and abstract yellowness. Similarly, it seems perfectly intelligible that FACT might cause that subject to behave in certain ways (e.g., to say, “Ah, yellowness, how I love thee”). Or, at least, it is not clear why Papineau would find such causal claims problematic. They do not involve, after all, the claim that the mental relation in question is itself a causal relation—which would be problematic on the assumption that uninstantiated yellowness is abstract. Only that the experience, constituted by some sort of relation between a concrete subject and an abstract property, is in the causal nexus.These critical remarks about Papineau’s central arguments are clearly not decisive, but merely invitations for further discussion of the very interesting lines of thought that Papineau explores in this excellent book.Thanks to Anil Gomes for discussion.