When I have a visual, auditory, or other perceptual experience, or when I experience certain bodily sensations, such as pain or an itch, there is clearly something it is like for me. That there is anything it is like for me I will call the “subjectivity” of the state. It is a state a subject experiences, and experience involves some kind of awareness, or consciousness. Specifically what it is like for me to see or hear something, or experience a pain or an itch, I’ll call the phenomenal character of the state. Phenomenal consciousness, then, involves two crucial components, both of which are reflected in the standard, Nagelian phrase “what it is like for the subject”: subjectivity, indicated by the prepositional phrase “for the subject,” and phenomenal character, indicated by the nominal phrase “what it is like.”Footnote ¹

Any theory of consciousness must say something about both of these elements, but some theories concentrate more on one than on the other. So, for instance, higher-order theories, such as Rosenthal’s “higher-order thought” (Rosenthal Reference Rosenthal, Block, Flanagan and Güzeldere1997) or Lycan’s “higher-order perception” (Lycan Reference Lycan1996), primarily address subjectivity. On their view, conscious mental states are those we are conscious of, and this in turn means that we mentally represent these states with higher-order states. Nonetheless, both Rosenthal and Lycan have something to say about what qualitative character is, though it isn’t determined by their theory of subjectivity.

On the other hand, first-order representationalist theories are primarily theories of phenomenal character. According to the standard, externalist version of representationalism, championed by Byrne (Reference Byrne2001), Dretske (Reference Dretske1995), Harman (Reference Harman and Tomberlin1990), and Tye (Reference Tye1995), the phenomenal character of an experience is identical to (or constituted by) the representational content of the relevant mental state. For example, the color phenomenal character of my visual experience of a ripe tomato is given by the feature of the tomato – a surface spectral reflectance, to a first approximation – that is represented by my perceptual system. But, of course, not every representation of that surface spectral reflectance is a conscious experience of it, so other conditions must be added to ensure that the representational state in question constitutes a genuine perceptual experience.

For some time now a debate has raged concerning the nature of conscious thought, or, more generally, conscious cognition. When I consciously deliberate, judge, or just entertain thoughts, is there something it is like to experience these mental states? Some philosophers claim that conscious thoughts have a phenomenal character just like conscious sensory experiences, while others deny this. I have already weighed in once on this controversy (Levine Reference Levine, Bayne and Montague2011), but I want to plunge in again and try to bring a different perspective to bear on the debate.

Let’s call the doctrine that there is a special, proprietary phenomenology for cognitive states “CP.” CP has been formulated in various ways by its adherents, but I will use this version by David Pitt. Here is how he puts it:

As far as I can tell there are two principal arguments for CP. The first is Pitt’s argument to the effect that without appeal to CP there is no good account of how we possess the kind of first-person knowledge of our thoughts – in particular, their contents – that we clearly do. I addressed that argument in Levine (Reference Levine, Bayne and Montague2011), objecting that it only worked if we built the doctrine of CP itself into the conception of the kind of first-person knowledge that we have of our thoughts, so the argument is question-begging.

The second form of argument is often called, after Susanna Siegel (Reference Siegel, Gendler and Hawthorne2006), the “phenomenal contrast” argument. The idea is to isolate two mental states that apparently differ only in their cognitive contents – leaving all sensory content the same – and noting that there is a difference in phenomenal character. Given the presumed identity of sensory content between the two states, the phenomenal difference can only be accounted for by appealing to the difference in cognitive contents, and so therefore CP must hold.

A typical exchange in the context of this argument goes like this:

At this point the debate largely reduces to a clash of intuitions; some find this elusive cognitive phenomenal character, and others only find what’s restricted to what Tye and Wright (Reference Tye, Wright, Bayne and Montague2011), following Lormand (Reference Lormand1996), call the “Quartet” of phenomenological states: perceptual experiences, bodily sensations, visual, auditory, and other sensory imagery, and linguistic imagery (“inner speech”).Footnote ² In what follows I will refer to the members of the Quartet as “sensory features.”

One extremely puzzling feature of the debate over CP is that there is a clash of intuitions at all. Consider the history of the debate about conscious experience as it developed over the last half of the twentieth century. From Smart’s (Reference Smart1959) concern about a “yellowy-orange after-image” through my own (Levine Reference Levine2001) discussion of the reddish character of visual experiences of my red diskette case (remember those?), the qualia debate focused on sensory-perceptual features like color, sound, and bodily sensation. Not coincidentally, these features are just the ones that exercised early modern philosophers when speaking of “secondary qualities.” Their task was to banish such qualities from the material world, relegating them to the interior of conscious experience. Once corralled within the mind, it was then up to materialist philosophers of mind to find a way to reduce them to the remaining nonqualitative features that had been left to the material world.

As I said, for a long time that’s where the debate stayed. Everyone knew what anti-materialist philosophers were talking about when they referred to qualia – even eliminativists knew what was intended and could identify them in their own experience. Their claim that qualia don’t really exist was understood to be very counterintuitive – even to themselves – and was based on the idea that we somehow suffered from a kind of cognitive illusion in believing in them. But they understood the illusion and had no doubt of its presence in experience.

Contrast that situation with the debate over CP. The debate begins with a basic clash of intuitions: CP advocates claim to experience it clear as day, while CP opponents claim not to find such features in their experience. Of course, the debate develops further, with various thought experiments trotted out by both sides, as described above, along with replies that try to show that the examples don’t have the force they were thought to. But what’s still puzzling to me is that first, initial clash of intuitions. Why is this an issue with CP but not with the traditional problem of qualia?

One common response on the part of CP advocates is to say that the phenomenology of thought is quite subtle, so those who fail to find it in their own experience are just not looking hard enough.Footnote ³ I think there is probably something to this, but it just raises the question of why cognitive states should contribute in this subtle manner to phenomenal character in contrast with sensory states. It would be nice to have a deeper explanation, one that engages with a theory of what phenomenal character, and conscious experience more generally, really is.

That is what I want to do in this chapter. From my perspective, too much of the debate over CP has been conducted with the idea that one puts forward one’s position and arguments in a way that stays as neutral as possible among the various theories of consciousness that have been proposed over the decades – in particular, neutral between materialist and anti-materialist, or reductionist and nonreductionist, theories. I can see the advantage of not engaging these large issues when focusing on this one doctrine, but to my mind this attitude has only muddied the waters and led philosophers to talk past each other.

Therefore, I want to begin my discussion by looking at a couple of representative materialist/reductionist theories of conscious experience – representationalism and higher-order theory, mentioned above – and one nonreductionist theory that I am very sympathetic to. I will argue that the way the CP debate plays out on the reductionist theories is very different from the way it plays out on the nonreductionist one. In particular, I will argue that, when one looks closely at what conscious experience really amounts to on the most popular reductionist theories, it turns out there is not much to really argue about. On the other hand, on the nonreductionist theory, the question of CP strikes at the core of what conscious experience is. Furthermore, on this theory, it also makes sense why there should be such discontinuity between the intuitive consensus over sensory qualia and the clash of intuitions regarding cognitive phenomenal character.

By a “reductive theory” of consciousness I don’t just mean one that includes a commitment to physicalism. As I see it, there are theories that might justifiably consider themselves physicalist without being reductive in my sense. So, for instance, a commitment to supervenience of the mental on the physical might be sufficient to count as physicalist, but a commitment to supervenience alone doesn’t constitute a reductive theory of the mental.

Rather, what I mean by a reductive theory is captured by Jerry Fodor’s (Reference Fodor1987) famous quip about intentionality: If intentionality is real, he argued, then it must be “really something else.” Any theory that provides an account of conditions that are constitutive of conscious experience in terms that do not advert to consciousness itself is reductive in my sense. Hence the traditional identity theory, traditional functionalism, representationalism, and higher-order theories are all reductive, as they specify conditions constitutive of conscious experience without mentioning conscious experience.

As I said above, I will focus my discussion of reductive theories on representationalism and higher-order theory. Before looking at them in some detail, let me outline what I take to be the unifying naturalistic and reductionist framework within which both theories are located: the computational-representational theory of mind (CR). According to CR, the mind is a functional system with inputs, outputs, and internal states that mediate between them. The internal states are constituted by physical states that embody relations to token representations that bear semantic and structural properties. To occupy an intentional mental state, on this view, is to occupy a state that consists of a computational relation to a physically embodied representation, and mental processes are operations defined over these representations.

Let’s first consider Tye’s (Reference Tye1995) version of representationalism, which he calls by the acronym PANIC: Poised Abstract Nonconceptual Intentional Content. As I said earlier, representationalism is primarily a theory of phenomenal character. It is intended to answer questions such as: What distinguishes a greenish phenomenal character from a reddish one? By virtue of what facts is it the case that the scent of a rose has the character it does? The answer to all these questions, as briefly mentioned earlier, is that the phenomenal character of the state is determined by (or supervenient on) the representational content of the state.Footnote ⁴

Taking my visual experience of a ripe tomato as our example, a theory of phenomenal character needs to tell us more than what the constitutive conditions are for its being reddish rather than some other color. We also need to know what makes it the case that there is something it is like at all to occupy that perceptual state, and we also need to distinguish the perceptual experience of the reddish color of the tomato from a thought about its color. After all, when I think <That tomato sure is red>Footnote ⁵ the content of my thought could be the same as the content of the correlated perceptual state, yet Tye doesn’t want to accord the thought any phenomenal character, much less the reddish content. Even an advocate of CP wouldn’t want to identify the phenomenal character of the thought with that of the relevant perceptual state.

To distinguish perceptual experiences from thoughts, Tye appeals to the nonconceptual condition, and to distinguish unconscious perceptual states (such as those induced in masking experiments) from conscious ones, he appeals to the poised condition. Let me elaborate on each of these conditions.

There is a large and confusing literature on the distinction between so-called conceptual and nonconceptual contents, or representations. The terminology I prefer puts the distinction in terms of “iconic” versus “discursive” representations, where the former are “picture-like” and the latter are “language-like.” Some philosophers, like Block (in an unpublished manuscript) pin the division between Cognition and Perception on just this distinction. Perception operates on iconic representations while Cognition operates on discursive (propositionally structured) representations. It’s notoriously difficult to provide a rigorous account of the distinction, but here are two ways philosophers have tried to explicate iconicity:

1. 1. The Picture Principle – with iconic representations, every (primitive) part of the representation represents a part of what’s represented. Discursive representations with propositional constituency relations do not satisfy this condition. (Fodor Reference Fodor, McLaughlin and Cohen2007)

2. 2. Continuity Mapping – with iconic representations, some of the spatial/temporal features of the representation represent the corresponding spatial/temporal features of what’s represented. Nothing like this occurs with discursive representations. (Block unpublished)

For each of these conditions there are complications. But clearly something of the sort is applicable; we do seem to get the difference between picturing, depicting a scene, and describing it in discursive terms. (In what follows I will use the verb “depict” for the way iconic representations represent and “describe” for discursive representations.) I will proceed on this intuitive understanding in what follows.

With regard to distinguishing, say, a masked perceptual representation of a red object, of which the subject claims to be totally unaware (but which exists, as demonstrated by subtle effects on future behavior), from one that is clearly experienced, Tye appeals to the poised condition. The idea is that a perceptual representation is conscious just in case it is poised to enter central or executive processes such as deliberation or report. When I visually experience the ripe tomato on the kitchen counter I can use this information to plan my dinner. But if the perceptual state is masked, say, then it isn’t available for my dinner planning or for report, even if it may have effects on my subsequent behavior. To be conscious, on this view, is to be “access conscious,” in Block’s (Reference Block1995) terms.

So now, with Tye’s first-order representationalism before us, what should we say about CP? Usually the question is asked this way: Do conscious thoughts have phenomenal character? Well, if we define phenomenal character, as Tye does, as the representational content of iconic representations, then the answer is obviously “no.” Thoughts involve discursive representations, so therefore their contents cannot give rise to phenomenal character. But suppose we ask the question this way: Is there something it is like to entertain thoughts (understood in the proprietary sense of Pitt Reference Pitt2004)? If instead of putting substantive conditions on the notion of phenomenal character, as Tye does, we thought of phenomenal character as whatever it is that answers to the “what” in “what it is like for the subject,” then the question about subjectivity becomes much more central than the question about phenomenal character.

One suspicion I have about the source of the intuitions conflict regarding CP is that, if you start with the question about phenomenal character, and you have a very “qualitative” or “qualia-like” idea of what that is, so you’re “looking for” something akin to sensory qualia when you look inside for cognitive phenomenology, then you're not likely to find anything answering to that conception when entertaining cognitive states. But if the question is asked initially about subjectivity – whether there is something it is like for the subject to entertain thoughts and the like – then you will be more open to the kind of phenomenal character that might be relevant to determining what it’s like to experience thoughts.

Let’s return then to Tye’s theory and ask: In terms of this theory, what would it take for there to be something it is like to entertain a thought? Well, according to PANIC theory, a mental state is conscious just in case the representation to which that mental state is related is “poised” for use by executive functions such as deliberation and report. Are thoughts ever poised in this way? Of course they are, all the time! Given the fact that thoughts clearly do often meet the constitutive condition for being conscious, for there being something it is like to have them, it’s quite unclear why Tye or any other representationalist would reject CP.

As far as I know, reductionists within the CR framework have only two options for distinguishing conscious from unconscious states, and they both involve some version of “access”: Either use some dispositional idea like being “poised” or use a more categorical idea like actually being employed by executive functions in various ways. I think of higher-order theory as a version of the latter. On this view, a state is conscious if that very state is the intentional target of another state, one that says, in effect, that the subject is currently occupying the target state. Again, what should we say about CP if we ask whether, according to higher-order theory, there is something it is like to entertain a thought? Well, it’s not quite as self-evident as on PANIC theory, but still it seems like there should be no in-principle bar to cognitive states being conscious in just the way perceptual states are. All we need is that higher-order states can target cognitive states as well as perceptual states. It seems this is a completely empirical question and there is no reason for a higher-order theorist to take an anti-CP stand.

One might object as follows. Yes, if we start by asking the subjectivity question – what is it for a state to be like something to the subject? – as opposed to the phenomenal character question – what determines what it’s like for the subject? – it appears that there is no good philosophical reason to deny subjectivity to cognitive states, at least on these reductive theories. But you do have to answer the phenomenal character question eventually. After all, what would go in for the “what it’s like” in the case of a thought?

There is an easy response, but this leads immediately to another objection. The easy response to the phenomenal character question is this: What it’s like to occupy a cognitive state is a matter of its content, just like with sensory/perceptual states. So, for instance, when I (consciously) entertain the thought that it sure is a beautiful day today, the content of that thought – that it’s a beautiful day today – is (partly constitutive of) what it’s like to think the thought, in the same way as the blue of the sky is (partly constitutive of) what it’s like to see the sky.Footnote ⁶

When put this way, however, one might find it extremely counterintuitive. After all, the blue is right there in my experience, “presented” to me, as many are inclined to say. But in what way is the propositional content that it is a beautiful day today presented to me, right there in my experience where I can “see” it or experience it? This of course is just another expression of the intuitive conflict at the base of this debate that is itself somewhat baffling.

I do think there is something to this intuitive response, but it’s not something I believe that reductionist theorists such as representationalists and higher-order theorists have a right to. Before saying why, however, let me bolster the objection by appealing to the iconic-discursive distinction discussed above. Let’s assume, with Tye and many other philosophers of mind/psychology (such as Block), that a distinguishing feature of sensation/perception is its iconic representational format. Perception, we might say, depicts, while Cognition describes. We can then add that, when it comes to experience, it is only what’s depicted that can serve as what it’s like for the subject, not what’s described. Furthermore, only sensory qualities, of the sort traditionally included under the rubric of “qualia,” can be depicted, which is why they are the only occupants of phenomenally conscious experience. After all, how do you capture that it’s a beautiful day today in a picture? Sure, you can picture a bright sun, blue sky, green leaves, and so on, but without “commentary” it isn’t clear that this “says” that it’s a beautiful day, much less that it’s today. On the other hand, the spatial layout and colors can be depicted. Therefore, according to this argument, when considering values for the “what” in “what it’s like,” we are limited to Tye, Wright, and Lormand’s “quartet” of sensory features.

As I just said, I do think there is something to the idea that the depiction-description distinction is relevant to the intuitive difference between sensory and nonsensory features as candidates for phenomenal character. I will explore this more in Section 1.3. However, unless one goes down the route of just stipulating that by “phenomenal” one means “represented by depiction,” I don’t see that there is any basis for drawing a philosophically principled distinction between what’s depicted and what’s described on the CR-type reductionist theories just surveyed. After all, on these theories, all it means for the content of a mental state to contribute to what it’s like is for the relevant representation to be functionally accessible in the right way. True, some contents are described while others are depicted. But, given that both kinds of representation meet the constitutive conditions for there being something it’s like to entertain them, why should this difference matter for the aptness of what they represent for being values of what it’s like?

When I consider the intuitive difference between consciousness of sensory features and consciousness of cognitively represented features (which of course includes sensory features but only when they’re thought about, not sensed/perceived), that difference seems to me to be a matter of what is directly presented to the subject, not about what is represented by the subject. While some might want to capture this “presentation-to/represented-by” distinction in terms of the iconic-discursive distinction, I don’t think that can be done. The problem is that representational format alone can’t do the job. What is really being distinguished here is a mode of access, and on the CR-inspired functional accounts of subjectivity, there is no principled difference between the relevant modes of access to these different kinds of representation. Both kinds can be poised or represented by higher-order representations. Rather, I will argue, only if one introduces a sui generis non-CR-type relation, one that relates subjects directly to contents as opposed to mediately through relations to representations, can the significance of the depiction-description distinction play a role in both explaining the intuitive conflict over CP and rendering it a doctrine worthy of philosophical dispute. Presenting this case is my task for Section 1.3.

Suppose, then, that we relax what I’ve elsewhere (Levine Reference Levine, Alter and Walter2007) called the “materialist constraint” – that is, the doctrine that any property/relation must be either a basic nonmental property/relation or realized in basic nonmental properties/relations. Instead, let’s posit a basic, mental, intentional relation of conscious awareness – call it “Acquaintance” – that holds between conscious subjects of experience and the objects/properties/relations the subject is consciously aware of, or acquainted with. Consciousness just is the phenomenon of subjects being Acquainted with whatever fills their streams of consciousness. This relation is not realized in other relations but is (most likely) a causal product of neurally realized psychological processes studied in cognitive science.

I use the term “Acquaintance” to signal a connection with the traditional notion of acquaintance, as it was handed down from Russell (Reference Russell1912) and others, but I am not interested in the epistemological and semantic questions to which Russell applied the notion. In earlier work (Levine Reference Levine2010), I called it the “AA” relation, for Acquaintance-Appearance. The idea is that, from the direction of the subject, we can speak of what the subject is acquainted with (or consciously aware of, these are the same for me), while from the direction of the object we can speak of what appears to the subject. Appearance, then, is just the flip side of Acquaintance. Appearances are what populate the stream of consciousness.

Use of the term “appearances” might bring Immanuel Kant to mind, and the association is not unintended. Kant was one of the first theorists to posit a fundamental, architectural distinction between Perception and Cognition – he used the terms “Sensibility” and “Understanding,” a distinction that is still a matter of controversy today. On the Kantian picture, Sensibility is the faculty of Intuition, which seems to be a relation of the kind I have in mind by Acquaintance. What we Intuit are “appearances,” objects that exist in space and time, along with the properties that determine their spatial and temporal features. All intuitions are structured by space and time, what Kant called the “forms” of Intuition.

Understanding, for Kant, or Cognition, in modern-day terms, is that faculty that employs conceptualization in order to think and deliberate. The relevant states are propositionally structured and they possess propositional truth conditions. Intuition yields appearances, while Understanding yields conceptualized representations, or thoughts and judgments. Another important difference between Sensibility and Understanding, for Kant, is that the former is largely passive, or “receptive,” while the latter is active, “spontaneous.” When we perceive something, intuit it, we don’t call it forth but rather receive the information of its presence to us. However, our thought is not restricted to what acts on us and involves processes that we initiate independently of what is around us.

That was obviously a very sketchy account of the Kantian model, and I’m not too concerned if it accurately reflects Kant’s actual view. It is certainly Kantian in flavor, and the sketchy version will do for my purposes here. What I want to do now is see if attention to this model can make sense of the intuitive difference we feel between the role of the traditional qualia, the sensory features, and the role of nonqualitative properties in constituting phenomenal character. In other words, how would the question of cognitive phenomenology play out on this model?

If we think of subjectivity, there being something it’s like, as constituted by a subject standing in the Acquaintance relation to objects and their features, then the question becomes what kinds of features the Acquaintance relation can take as its object – with what sorts of objects and properties can we be Acquainted? Colors, sounds, tastes, pains, shapes, motion – these are the sensory features that everyone agrees are in the range of this relation. The question then becomes whether or not nonsensory properties are also included. How do we resolve this, and also how do we explain why it is so much more obvious intuitively that sensory features are included and a matter of controversy and conflicting intuitions when it comes to the nonsensory?

Earlier we discussed the distinction between what we called “depiction” and “description,” where the former is how iconic representations relate to their contents and the latter is how discursive representations relate to their contents. I argued that, though there is a clear intuitive difference between the two kinds of representations (though hard to make precise), the CR account of consciousness doesn’t have the resources to exploit this distinction to explain the intuitive difference we feel between sensation and cognition regarding conscious experience. But let’s dig a bit deeper into that distinction and see if the nonnaturalistic theory of Acquaintance can make use of it.

As I conceive of the Acquaintance theory, conscious experience is itself one big depiction – a movie, if you like, in the Cartesian Theater. Though I haven’t said much about the objects and properties that populate the stream of consciousness, as I conceive it the world presented to us in the stream of consciousness is populated with what I call “virtual” objects and properties. You can think of them as like Chalmers’s (Reference Chalmers2010) “Edenic” properties. They are what naive perception takes to be the colors, sounds, feelings, and so on, just as they appear, not as they are “in themselves,” as Kant would say. There are two reasons I have for treating appearances as “virtual” – that is, not metaphysically really “out there” in the world but rather a creation of the conscious mind.

The first reason is the problem of hallucination. If Acquaintance is a relation, then there have to be relata. If conscious experience is constituted by a subject’s standing in the Acquaintance relation to appearances, and if hallucination is a form of conscious experience, then hallucination must involve a subject’s standing in the Acquaintance relation to something. But if when I’m hallucinating pink elephants on parade there are no elephants, and nothing pink in my vicinity, then the pink elephants, the “appearances,” must be virtual objects and features. (I know “naive realists” such as Martin Reference Martin2004 don’t buy the argument from hallucination, but I can’t address their objections in this chapter.)

The second reason has to do with the long-standing controversy, going back to the Early Modern philosophers, about the status of the so-called “secondary qualities.” In Levine (Reference Levine2008) I argued that none of the views that place color and the other secondary qualities in the objective world are sustainable. Rather, while the objective world, as it is “in itself,” is composed of the stuff and properties that are represented in the natural sciences – so the analogue to color is something like spectral reflectance – the qualities themselves are creatures of the mind. I can’t defend that position here, but it is one of the bases on which my claim that appearances are virtual objects is grounded.

Suppose, then, that conscious experience is a virtual depiction of the world, populated by virtual objects and properties. As in a picture, in order to depict a scene, or event, you need to start with the basic building blocks – colors, sounds, shapes, and motions – that fill in space and time. You can’t depict, say, Lee’s surrender to Grant at Appomattox Court House without painting some colors and shapes. You can describe it in mere words, but to depict it you start with the colors and shapes. These basic building blocks of depiction, the necessary means of implementation for all depiction, are the very sensory features everyone is willing to recognize as forms of phenomenal character.

With this idea of sensory-qualitative features as basic building-blocks of depictions, we can both explain why there is universal acknowledgment of the phenomenal presence of traditional qualia while the inclusion of nonsensory features provokes intuitive clashes and at the same time outline various theses regarding CP that are ordered in terms of their “strength” – the degree to which they add to the population of features found in the stream of consciousness.

To begin with, the Acquaintance theory shares a certain model with naive realists-relationalists, as already presented. Naive realists think of conscious perception as constituted by a relation of the subject, the mind, to objects and properties in the world. In this they reject the representational theory that breaks the conscious perception relation into two relations: a representation relation between perceptual representation and objects and their properties, on the one hand, and a computational-functional relation between the subject and the perceptual representation, on the other. The Acquaintance theory shares this model of the relation between subject and object in conscious perception but takes the objects to be virtual constructs of the mind itself. On both theories, though, conscious perception of a red surface, say, involves a direct, immediate relation between the mind and redness, not an attitude toward a mental representation of redness.

In characterizing Acquaintance as a matter of direct presentation of a depicted scene, as in a movie, I’m following Kant’s idea that Intuition (which I’m interpreting as Acquaintance) relates the mind to individuals: objects, property instances, events, and the like. Concepts, according to Kant, give us access to the general, but appearances are always of individuals, the kinds of entities that are apt for picking out by demonstratives. It is the depiction of individuals – as in the surrender scene at Appomattox Court House – that requires the traditional qualities as building blocks. But can we be acquainted with universals, propositions, and the like as well? Can we be acquainted with the property of redness, not just an instance of it? What about the property of justice? What about the proposition or fact that 2 + 2 = 4?

What I think of as the maximal version of CP – what I called in Levine (Reference Levine, Bayne and Montague2011) “pure CP” – says that we are acquainted with not only concrete objects and their qualities but also abstract objects such as universals and propositions; we are acquainted, on this view, with the actual contents of thoughts and concepts, not merely their representations. Put another way, on this view, when we entertain a thought, or make a judgment, we are acquainted not only with our state of mind but also with what that state of mind is about – again, not merely with another form of representation of what it is about but the very objects and properties it is about.

To give you the flavor of what I mean, let me give you an example due to Elijah Chudnoff, who I think of as an advocate of this maximal CP, or pure CP. Here is how he puts it:

So, you are trying to figure out why this mathematical statement (call it A) is true and at some point you come to “see” why it’s true. This event of coming to see why it’s true is, according to him, an experience with phenomenal character, and this phenomenal character is constitutively determined by the cognitive state of seeing how A is true. Now I don’t know if Chudnoff would buy any part of the Acquaintance theory I’ve outlined, but we could put his point in terms of that theory. On his view, when I come to see how A is the case, I am acquainted with a complex numerical state of affairs, or something like that.

Notice that the state of affairs, or proposition, that if a < 1, then 2 – 2a > 0, is not something that can be depicted. There is no picture that expresses this proposition. If indeed phenomenal consciousness is Acquaintance, and Acquaintance essentially involves depicting a scene of individuals and their properties, then it is easy to see why many fail to share the intuition that there is any distinctive, or individuative, phenomenal character arising from our cognizing this algebraic fact. I must admit to being one of those who fail to share this intuition. I honestly can’t make intuitive sense of consciously grasping this fact in anything like the sense in which I know what it is to consciously grasp the perceived scene around me. If it isn’t composed of these depiction building blocks, I don’t know what it is to experience it.

Of course, Chudnoff and others (see Siewert Reference Siewert1998) claim to experience these contents. In fact, Siewert specifically refers to them as “non-iconic” experiences. As usual, appeals to a phenomenal contrast argument are employed. On this issue I side with CP-skeptics who appeal to iconic material that explains the relevant contrast. So, just to focus on Chudnoff’s mathematical case, he appeals to one’s intuition about what happens when the transition from not seeing how A is true to seeing how it’s true occurs. Having myself not seen it immediately when first reading it and then coming to see how it’s true after a moment of reflection, I find that what is “before my mind” the entire time are images of mathematical representations. I start imagining certain fractions as values for “a,” and then perform the calculations, and then infer that whatever fraction (<1) I put in will yield the desired result. I just don’t know what it would be to be Acquainted with the fact itself, as opposed to representations of it. The representations, of course, are depictable.

I don’t expect pure CP advocates to find this conclusive, by any means. I don’t want to linger on this now, but just note that pure CP is what I think of as the strongest version of CP, and, within the framework of Acquaintance theory, it seems to me unsustainable. But that doesn’t mean CP itself must be rejected. There is another way that cognition might help determine the phenomenal character of an experience, without involving one’s Acquaintance with abstracta like propositions, universals, and states of affairs.

To begin with, if we take the standard phenomenal contrast arguments seriously, it appears that kind-properties, causal relations, and maybe agency are features that make a difference to what it’s like. Contrast seeing a red, round shape with seeing a ripe tomato; seeing one event follow another with seeing the first as causing the second; and seeing and feeling a body part move with feeling it as one’s voluntary motion. I think a strong case can be made that adding kind, causal, and agency information can make a difference to the phenomenal character of an experience. How would this be accommodated by our Acquaintance model?

Properties such as being a tomato, or relations such as causation and agency, do not seem to be qualitative properties in our proprietary sense; that is, they aren’t basic, or primary, ways of filling space-time but rather features of certain ways of filling space-time. When you have space-time filled with a redness, roundness, and other qualitative properties in a certain way, you have what appears to be a ripe tomato. But the conditions for being a ripe tomato do not dictate any specific ways of filling space-time, nor are they restricted to these ways, as certain causal and historical conditions must be met as well. That is, being a depiction of a ripe tomato does not supervene on the depiction of the basic sensory features. So how do we account for their effects on what it’s like?

It’s interesting to think back to Empiricist debates about “abstract ideas” in this regard. I take it that the Empiricists shared something like this Acquaintance view of consciousness but then added the claim that this is really all there is to mentality. So, Hume doubted our ability to form an idea of causation based on impressions, since he couldn’t find an impression for the necessitation relation between events. We just see one billiard ball make contact with another and then the other moving away; we don’t really see, he claimed, the causal relation by virtue of which the contact necessitated the movement. Berkeley famously objected to the idea that we can form an idea of a triangle simpliciter, since any image of a triangle had to have specific dimensions and angles. If Acquaintance is all there is to mentality, and Acquaintance can only take sensory qualia as objects, then it doesn’t seem as if there’s any way for more abstract representations to affect what it’s like for us; indeed, there doesn’t seem to be any way for more abstract representations to enter the mind in the first place.

Now let’s relax the Empiricist restriction of all mentality to Acquaintance, allowing in a separate faculty of cognition – Kant’s Understanding – that can form conceptual, discursive representations of any sort of property or relation. The question to ask now is, what relation can Cognition, with its discursive representations, bear to Acquaintance? We already rejected the pure CP account on which conscious Cognition Acquaints us with universals, propositions, and states of affairs. But I think there is another way that cognitive activity can help determine phenomenal character, one that, to my mind, deserves to be called a version of CP.

When we perceive a certain collection of basic sensible properties, sometimes we apply certain concepts to them and sometimes not. So, if I see a red, round object I might see it as just that, or, on the other hand, as a ripe tomato. The redness and roundness are objects of Acquaintance, but what about the tomato-hood of the object? Is there an Acquaintance relation between me and that property? If we apply the phenomenal contrast method – imagining seeing the very same scenes in terms of basic sensibles but applying the category “tomato” to one and not the other – we may detect a difference in what it is like. That is, when we apply the categorical concept “is a tomato” to the scene, the experience feels different from the way it would have felt had the concept been withheld. In other words, I see it (or, better, am Acquainted with it) “as a tomato.”

A similar story might be told about applying the concept of causation. Consider the Michotte (1946/Reference Michotte1963) experiments in which subjects are shown video displays of two discs, one of which moves toward the other and then the second moves away. Depending on the details of the scene – whether the two discs touch, the relative velocities of the discs, the trajectories, and so on – subjects see the first as causing the second to move in some cases and not in others. It very much seems as if, in the cases where the first disc is seen as causing, or pushing, the second disc, that this isn’t merely a judgment but actually how we see it – part of what it is like.

There are two features that are essential to the version of CP I’m presenting here – call it “impure CP.” First, it is restricted to the application of concepts to objects of perception. Only when we have something perceived and apply a concept to it can that concept make a difference to experience. (Here I’m including hallucination as a kind of perception – something is given in experience as out there, even if it’s not really out there in the world.) Concepts that are deployed in purely cognitive states – entertaining thoughts, making judgments, and so on – cannot affect the stream of conscious experience (except, of course, through sound and visual imagery like “inner speech”).

The second feature is that the difference that the application of the relevant concept makes to the phenomenal character of the state is not a matter of the property or relation that constitutes the external content of the concept – that is, the object or property represented – serving as an object of Acquaintance or as being presented to the subject. Rather, the difference in what it’s like between the cases where the concept is applied and where it’s not is a purely internal matter. One of them feels different from the other, and what is responsible for this difference is not the presentation of the content of the concept but rather the detection of the application of the mental representation itself. In a sense, rather than being Acquainted with the causal connection itself that holds between the two events perceived, we’re Acquainted, in a sense, with the predication of “causal connection” to the percept. On this view, only the basic sensible properties and relations can be objects of Acquaintance, not properties and relations such as categories, causation, and agency, not to mention even more abstract properties.

Let me elaborate some on what I mean by a “purely internal matter.” To begin with, I want to back up and address a likely reaction the reader will have to talk of “virtual objects and properties” – namely, that my Acquaintance model is just a version of “sense data” theory. After all, sense data were supposed to be purely mental objects and properties that are grasped by the mind, with no independent existence outside it. This is certainly how I am characterizing the virtual objects and properties that are the objects of Acquaintance. I have two responses: first, yes, guilty as charged. There is a clear sense in which what I’m proposing for the nature of conscious experience involves objects whose metaphysical status is akin to sense data.

On the other hand, as I understand what sense data were supposed to be, they would be restricted to the basic sensory building blocks, and that’s it. What one is presented with in conscious experience, on this view, is just the various sensory profiles involved in the scene around us. However, on impure CP, as I propose it, what conscious experience presents us with is an organized scene of objects, properties, and relations among them, and these are properties and relations that go beyond what supervenes on the basic sensory building blocks. In other words, what is presented to the subject in the stream of consciousness is an organized world, not merely a collage. The organization derives from the cognitive contribution, from the application of conceptual representations to the otherwise chaotic display of basic sensory features. As Kant famously declared, “Concepts without Intuitions are empty, and Intuitions without concepts are blind.” To see a ripe tomato, and not just co-located redness, roundness, and the like, one requires a conceptual, categorical representation under which to bring the relevant sensory profile. And, as per CP, this constitutes a genuine contribution to what it’s like for the subject.

What I am denying, however, is that even when what we experience is not merely a sensory profile but actually a tomato, not just a succession of events but one event causing the other, in these cases it is the property of being a tomato itself, or the causal relation itself, that we experience. These properties and relations do serve as the external contents of our conceptual representations, but I don’t see how they can contribute to the character of the subjective experience, as I argued above. However, there is an internal content – something like conceptual role – that I do think we can experience. It’s the entire dispositional set that comes with seeing something as a tomato, or one billiard ball hitting the other as a causal transaction – the host of expectations, potential inferences, and the like, that come along with certain conceptualizations by virtue of their rational connections to other conceptual representations. And it is that internal content that largely determines how we experience the sensory features that confront us in perceptual experience.

On this Acquaintance model, then, there are two components to our conscious experience. First, there are the sensory building blocks, sense-data-like entities that are the genuine objects of the Acquaintance relation, providing the raw material to construct the virtual world of the Cartesian theater. Second, there is the conceptual system that weaves these sensory features into an organized world, providing labels, as it were, that make rational sense out of the whirling, buzzing show of qualia. This system that binds the sensory building blocks into an organized world is determined by the pattern of application of conceptual and inferential relations to bundles of sensory features and the pattern of relations among the conceptual representations in our belief system. This is the world of conscious experience.

Since these two components play such different roles in making the world of conscious experience, it isn’t surprising that there should be much more consensus about one component than the other. We can’t help but notice the actual relata of our Acquaintance relation, the sensory features. But the organizing role of our concepts, when applied to the sensory building blocks, is more subtle, more difficult to discern. Given that the contents of the conceptual representations are not themselves direct objects of Acquaintance but rather mold its objects into a system for us, their effects on conscious experience are less obvious. Nevertheless, following the various phenomenal contrast arguments that have been put forward in the literature, I believe those cognitive effects on conscious experience are indeed real. Thus, I support my moderate, impure version of CP.

In this chapter I have contrasted naturalistic CR-style theories of conscious experience with the nonnaturalist Acquaintance theory that I laid out. I noted two advantages for the latter over the former. First, with regard to the troubling intuitive conflict that seems to surround the question of CP in a way that does not arise with the traditional sensory qualia, I showed how Acquaintance theory can make sense of that difference in a more satisfying manner than can the naturalist theories. On Acquaintance theory, sensory features play a constitutively different role in the construction of the (virtual) world of experience than do the features represented by concepts.

Second, even more important, when considering the naturalistic theories, it becomes unclear why there is even a philosophical controversy over CP at all. Given their accounts of subjectivity, there doesn’t seem to be any basis for distinguishing the sense in which there is something it is like to perceive from that in which there is something it is like to think. If subjectivity is understood in terms of functional availability, or activation of the relevant representation, there doesn’t seem to be any deep reason why this kind of availability should distinguish among representations in terms of their contents. After all, the contents only make a difference by way of the representations of them, and in both cognitive and perceptual representations we are talking about neural states that bear the contents, and it’s the functional operation on the neural states themselves that determines their availability. So, what would distinguish cognitive from perceptual representations that would explain why the latter can be available in the right way and not the former?

Are these two advantages enough to tip the scales in favor of Acquaintance theory? I doubt many would be convinced by these considerations alone. Still, it seems worthwhile to me to chart the way different controversies over the nature of conscious experience play out on different underlying conceptions of just what it is for there to be experience at all. And it doesn’t hurt if the theory I favor can make sense of the contours of a fundamental controversy, like the one over CP, in a way that seems much more illuminating than some of its most popular competitors.