Is All Phenomenology Presentational?

3.1. 

I’m now going to describe a standard view of cognitive phenomenology, and argue that such phenomenology (if it exists) cannot be presentational (in my sense from the previous section). If cognitive phenomenology exists and is non-presentational, then being presentational cannot be a mark of the phenomenal.

The argument can be laid out as follows.

• P1 (Assume for the sake of argument that) CP experiences exist and are transparent to conscious thoughts contents.

• P2 If CP experiences exist and are transparent to conscious thought contents then, if the CP is presentational, these thought contents must be presented.

• P3 The contents of conscious thoughts are not presented.

• C1 CP is not presentational.

• C2 It’s not the case that all phenomenology is presentational.

3.2. 

The sort of CP theory I have in mind accepts that cognitive phenomenal character is transparent in the following way (see Forrest 2017: 407):

• CP Transparency: there exists a kind of experience that necessarily accompanies (or is identical to) a phenomenally conscious thought, such that the experience’s cognitive-phenomenal features and the thought’s contents are introspectively indistinguishable: when one introspects on what the experience is like (setting aside mental imagery) one inevitably just attends to the thought’s contents; and all the elements that make up the thought content of which one is conscious constitute aspects of what it is like to undergo the CP experience.

CP Transparency—and the argument of this section—is not intended to apply to every purported kind of cognitive phenomenology. In particular, it does not apply to the alleged phenomenology of attitude one holds towards the thought entertained—at least not without further argument. But CP Transparency is meant to apply to the kind of phenomenology described in the following passage:

There is something common phenomenologically, something that remains the same in consciousness when one passes from, say, believing that rabbits have tails to wondering whether rabbits have tails, or vice versa. It is the distinctive phenomenal character of holding before one’s mind the content rabbits have tails, apart from the particular attitude type, be it, say, wondering, hoping, or believing. (Horgan & Tienson 2002: 522)

Although they stop just short of saying it explicitly, it is natural to read Horgan and Tienson as implying CP Transparency with respect to the phenomenology they are highlighting, which we can dub “the phenomenology of entertaining thought content,” or just “the phenomenology of entertaining.” The phenomenology of entertaining is endorsed in one form or another by many proponents of popular “Phenomenal Intentionality” theories, according to which having an experience with a given phenomenal character is identical to, grounds, or at least metaphysically necessitates, being in a conscious state with a given intentional content.³ Phenomenal Intentionality theorists should find the claim that the phenomenology of entertaining is transparent to be congenial to their view. For according to the transparency claim, the thought contents entertained are constitutive of what it’s like to think that thought, so enjoying an experience with that CP character amounts to entertaining a thought with that content. Furthermore, Phenomenal Intentionality with respect to CP is a natural way to make sense of the popular idea that cognitive phenomenology is “proprietary” to thought. This idea, though rarely spelled out, is plausibly understood as the notion that this phenomenology is (non-accidentally) unique to and inseparable from thought: while there could be thoughts, perhaps even conscious thoughts, without these phenomenal properties, the phenomenal properties could never be instantiated in the mind without an accompanying thought. The phenomenal character “belongs to” the thought. But why would that be, unless the instantiation of cognitive phenomenal properties somehow had the power to bring about the entertaining of a thought? Thus, CP Transparency gains credibility from its connection to these other popular views about cognitive phenomenology.

Moreover, CP Transparency is independently compelling. Suppose we accept that there is something distinctive it is like for us to think a thought, beyond the phenomenology of sensory imagery (and any phenomenology of attitude, if such there be). What is it like? The only way I know how to answer that question is to appeal to those contents of my thoughts of which I am conscious when I introspect. It is extremely tempting to conclude that the experience itself is utterly transparent.

One of the classic arguments for CP, the so-called phenomenal contrast argument, acknowledges this point. The experiential contrast between not comprehending and comprehending written or spoken speech, which CP is introduced to explain, is often described as the difference between merely hearing (properly parsed) sentences, on the one hand, and consciously grasping the meanings of those sentences, on the other.⁴ It is the meaning, the content itself, that is added, and that supposedly makes the phenomenal difference, when we do not merely perceive a string of words but also apprehend a thought prompted by the words. So it is clear from their own arguments that many CP theorists’ views are broadly in line with CP Transparency.

3.3. 

Now that I have explained why many CP proponents should accept Premise 1, I need to defend the rest of the argument. Premise 2 states that if there is CP that is both transparent and presentational, then a thought’s contents should be what is presented. The reasoning here is that according to CP Transparency, certain cognitive phenomenal features that make up an experience of thinking are indistinguishable from the elements of the thought’s content. So if the experience presents the former to the subject, it thereby presents the latter as well.

Premise 3 states that our conscious thought contents are not presented to us. I think a little reflection on what we find when we introspect our conscious thoughts should make this truth evident. For instance, the other day for some reason the following thought popped into my head: Paul Newman and Robert Redford played ping-pong with each other during breaks on the set of Butch Cassidy and the Sundance Kid. Because the thought was conscious, I was, in one everyday sense, immediately aware of its content. But whatever such propositional contents are made of—regardless of their ontological status—it is obvious that my conscious thought in this instance does not essentially seem to include standing in a relation of immediate, object-involving awareness to a cluster of concrete particulars: Newman, Redford, instances of the game of ping-pong, and so on.⁵

It might be instructive to compare conscious thought to sensory imagery in this respect. The paradigmatic case of presentational phenomenology is found in normal perceptual experience: Robert Redford—and/or a number of properties characterizing Redford’s present physical appearance—is visually presented to me from across the red carpet at a movie premiere. When I experience a visual image of the late great Paul Newman, by contrast, Paul Newman is not presented to me. Nor are any (or most) of his physical appearance properties. We can grant that my mental image should count as an image of or about Paul Newman, presumably in virtue of some causal connection that traces back to the real person. But the image all by itself is not sufficient to make it seem to me as if I am at that moment standing in a relation to any concrete, flesh-and-blood human being. The proof of this fact is that my experience of the image alone does not supply me with the least bit of evidence that there is or ever was any such person.

However, I do seem to stand in an awareness relation to some concrete particulars when I have a visual image of Paul Newman, namely those sensory-imagistic properties that characterize what it’s like for me to undergo the experience. Apparently, these are properties of my experience, instantiated within my current stream of consciousness. This experience is a dateable episode in the ongoing mental life of a particular concrete subject, and the image—or at least those of its shadowy features of which I’m fleetingly aware—seems to me to be located at a particular time (and arguably place). Thus, both sensory-perceptual experiences and sensory images present their phenomenal features to the subject. Though in the case of imagery, the properties that are presented are not the same ones that we typically use to individuate an image, since what imagistic phenomenology presents is something other than what it’s about.

Now consider again the cognitive phenomenology of entertaining. Thoughts made conscious by CP would have to be the opposite of images in both of the above respects. On the one hand, CP Transparency tells us that (unlike imagery) if these CP experiences present anything, they present a thought’s contents, since that’s what I’m immediately conscious of when I introspect. On the other hand, since CP clearly doesn’t present these contents to me—a thought about Paul Newman does not present that individual to me any more than a mental image does—then (unlike imagery) CP experiences just aren’t presentational at all. Our thoughts simply are not like that for us.

3.4. 

Let us review. In this section I have presented an argument for thinking that cognitive phenomenology is not presentational. I have not argued that CP exists. Instead, I have argued that many proponents of CP do or should accept the version of CP that my argument requires, since they do or should accept that the phenomenology of entertaining thought content exists and is transparent to thought content. My argument shows that, to the degree that we have reason to accept this sort of transparent, proprietary phenomenology of thought, we thereby have reason to deny that all phenomenal character is presentational in nature. Since acceptance of cognitive phenomenology is currently a popular position in philosophy of mind, this argument has significant implications for any efforts to unearth the marks of the phenomenal.

In what follows, I’m going to turn this argument on its head. As the old saying goes, every ponens can become a tollens. Since the foregoing argument begins with the assumption that cognitive phenomenology exists and reasons to the conclusion that not all phenomenology is presentational, if all phenomenology is presentational, then, by the same reasoning, CP must not exist—at least not in this popular transparent form. In the next section, I set out my argument that all phenomenology is presentational.

4.1. Why suppose that all phenomenal character is presentational?

Here is the sketch of an argument. A necessary condition on a property being a phenomenal feature of experience is that it could be successfully grasped via a phenomenal concept. Plausibly, not every phenomenal property at the moment it is experienced is conceptualized using a phenomenal concept. But every phenomenal property, and no non-phenomenal property, could have a phenomenal concept correctly applied to it. Further, a necessary condition on a property being graspable by a phenomenal concept is that it is available to be picked out by an introspective demonstrative. In other words, the conscious subject must be able to mentally “point” to that feature of her experience. Finally, a necessary condition on a property being available for introspective demonstration is that that property is presented to the subject by her experience. In a slogan: no pointing without presentation. So, all phenomenal properties must be presented to the conscious subject in experience.

This argument requires we accept three transitions: (1) from being phenomenal to being a candidate referent of a phenomenal concept; (2) from being a candidate referent of a phenomenal concept to being available for introspective demonstration; (3) from being available for introspective demonstration to being presentational. Let us take these each in turn.

4.2. From Being Phenomenal to Being a Candidate Referent of a Phenomenal Concept

Consider this passage from Brie Gertler:

in some introspective judgments about experience, (phenomenal) reality intersects with the epistemic, that is, with the subject’s grasp of that reality. This thesis—or something close to it—is implied by the claim that we sometimes grasp our experiences directly, by using an experience’s defining phenomenal quality to form an epistemically substantive conception of the experience itself. (2012: 94)

Our focus is her claim that a mark of the phenomenal features of our conscious stream is the ability we have to form certain “direct” and “epistemically substantive” conceptions of them. Such conceptions allow us to make introspective judgments of the following forms:

• I am now undergoing this kind of experience

• Such-and-such phenomenal property is instantiated (in me, now)

• That (feature of experience) is $X$

(Here ‘this’, ‘such-and-such’, and ‘$X$’ are stand-ins for some phenomenal feature that the subject is presently introspecting.)

Let us stipulate that for a concept to count as a phenomenal concept in our sense, it must meet the following three conditions:

1. It must be used to refer to a (previous or current; token or type) phenomenal property of the subject’s experience, when the subject is conceiving of her experience in a manner that is proprietary to the first-person perspective (Levine 2006: 149)

2. It must be conceptually direct

3. It must be epistemically substantive

For the subject’s conception of a phenomenal property to be proprietary to the subjective perspective requires that the subject typically acquires this way of conceiving of the property when and only when she enjoys privileged first-person access to that property as an occurrent feature of her own experience. This means, inter alia, that phenomenal concepts are not theoretical concepts of experience. For theoretical concepts have descriptive contents that can be grasped and applied in exactly the same way by different subjects, regardless of whether these individuals presently instantiate the properties being thought about, or indeed have ever experienced them firsthand.

Our cognitive grasp of the phenomenal features in these judgments is meant to be referentially or conceptually direct, in that the way we conceive of these phenomenal properties does not rely on any additional properties of those properties—as it would if we were picking out the referent descriptively via a contingent mode of presentation. Rather, when we introspect on our conscious experience we just have a lucid grip on the essential and distinctive character of the phenomenal property we are presently experiencing, and our concept of the property picks out all and only instances that share that character. The way Brian Loar puts the point in his influential discussion of phenomenal concepts is that the phenomenal property serves as its own mode of presentation: both referent and reference fixer (Loar 2002; Chalmers 2010: 182).

For our grasp of a phenomenal property to be epistemically substantive, it must genuinely constrain the space of epistemic possibilities. For example, learning what cheddar cheese tastes like—such that one is in a position to judge, I’m having this kind of taste experience, where ‘this’ refers to the particular subjective quality of cheddar’s taste—allows one to eliminate the possibility that one is having an experience of some other phenomenal type. If in addition one knows that one is a normal taster and is eating cheddar, one can rule out alternatives in which cheddar tastes some other way, for example, the way Emmental tastes.

The point of requiring that phenomenal concepts be substantive is to guard against views on which our grasp of a phenomenal feature is direct—securing reference to the property by somehow latching onto the property itself, stripped of any descriptive adornment—at the cost of remaining empty from the subjective perspective. A useful analogy is to an act of “blind” demonstration, which succeeds in referring to something in the subject’s environment without providing her with any informative internal conception of the referent (cf. Wittgenstein 1958: §279). Clearly, one’s grasp of one’s phenomenal reality does not seem to be empty in this way, but rather to capture rich and varied epistemic possibilities.

Now, one reason to accept step (1) of our argument is that such a view seems to be the current philosophical orthodoxy. Many philosophers who have written about phenomenal consciousness believe that our ability to think about phenomenal properties using phenomenal concepts is key to explaining what is important and distinctive about phenomenology.⁶ An implicit assumption of this work is that this explanation is sufficiently general to account for all phenomenal properties in this way (hence the name “phenomenal concepts”). One common motivation for the view—from physicalists and dualists alike—is the belief that these concepts are crucial to accounting for the “explanatory gap” that allegedly separates phenomenal properties from physical theories (Levine 1983). Thus, if one accepts that general theories of phenomenal consciousness that appeal to phenomenal concepts are on the right track—especially if one believes that all phenomenal character is implicated in the presence of the explanatory gap—then one has a powerful reason to accept that all phenomenal properties can be picked out by phenomenal concepts.

However, it is debatable whether all phenomenal character gives rise to an explanatory gap. For even within perceptual experience, experiences of primary qualities such as spatial properties arguably do not create the same kind or degree of mystery as do experiences of secondary qualities such as tastes, smells, and colors.⁷ Thus, if bridging the explanatory gap is the only reason to posit phenomenal concepts for every kind of phenomenal property, we might doubt whether this step is warranted.

Fortunately, there is a further argument for this conclusion. The argument appeals to a characterization of phenomenal character about which virtually everyone agrees: it is treated as a matter of definition that phenomenal character is “what it’s like” for the subject to consciously experience the world. Of course, this phrase is a term of art the usefulness of which extends only so far as we can expound on what it means in ways that at least most of us understand and accept. But for our purposes, we only need to endorse the idea that what it is like to be in a certain mental state is always what it is like for the conscious subject, and thus that what-it’s-like-ness, whatever else we say about it, is a characteristic of experience that (partly or wholly) captures the distinctive subjective access we have to our own conscious minds.

To see the significance of this idea for our argument, we can begin by considering the contrast between the way we recognize many of the properties we perceive in our environment, on the one hand, and the phenomenal properties of perceptual experience, on the other. As Levine puts it,

I judge that there’s a dog in front of me because it appears to me that there’s a dog in front of me… . if someone were to challenge my claim about the dog … my justification would certainly involve mentioning how things visually appear to me. Now, when it comes to phenomenal judgments—say, I’m having a reddish visual experience, I’m having a headache—there don’t seem to be any epistemic liaisons of this sort to serve as evidence. (2006: 151–52)

A dog and the visual appearance of a dog are clearly distinct; a headache and the phenomenal “appearance” of a headache are not. So in the headache case the only epistemic basis for my judgment is the headache, the experience itself.⁸

And the ability to identify these phenomenal properties directly in this way goes hand-in-hand with the ability to refer to them directly in thought as well. Call the former kind of directness epistemic directness or E-directness, to contrast it with the latter, conceptual- or C-directness, discussed above. The same introspective access we have to perceptual experience that allows us to identify these properties as we experience them (E-directness), should also allow us to conceive of them simply in terms of how we experience them as being (C-directness). For it is highly implausible to think that my only conception of a phenomenal property may be indirect, reliant on a cluster of other properties to pick it out, even if my identification of the property is neither indirect nor blind, but directly due to a substantive awareness of the property’s character. On the contrary, it is natural to suppose that my possessing a substantive, C-direct grasp of a phenomenal property in thought is required in order for me to form a substantive recognitional judgment about instances of that property without first needing to identify a bunch of associated properties.

So far this argument deals only with perceptual phenomenology. Why should we suppose that our introspective access to all phenomenal properties works the way that it does in perceptual experience, unless we are assuming that perceptual experience exhausts all phenomenal character? This is where the observation that phenomenally conscious states are those that there is something it is like to be in—and that there is a constitutive connection between our commonplace introspective access to our experiences and what it is like for us to experience them—becomes important.

The connection is this: (a) we can, at least in principle, enjoy immediate introspective knowledge of any given phenomenally conscious state simply based on our conscious awareness of what the state is like for us to be in. For it is overwhelmingly plausible that when undergoing an experience, if we are aware of what it is variously like for us to undergo it, then we are also able in principle to think about the various ways it is like of which we are aware (and when such thoughts are properly based on our conscious awareness, they are justified and can constitute knowledge). Crucially, our awareness of the way(s) it is like to have an experience is not reducible to an introspective judgment to the effect that this is what the experience is like—it is awareness of rather than awareness that—since otherwise the former would be identical to, and could not provide the basis for, the latter. But (b) what-it’s-like-ness is just a synonym for phenomenal character, which is the aspect of a phenomenal state or instantiated phenomenal property that individuates it as the state or property it is. So the very same aspect of a phenomenal property that marks it out as being the particular phenomenal type that it is, is also that aspect most plainly and immediately accessible to introspection. When I reflect upon what my occurrent phenomenal state is like, I am reflecting upon a particular phenomenal property type of which I am conscious. Thus, when my judgments about the presence and nature of phenomenal properties in experience are based on such introspective reflection, they are justified by my awareness of these very properties. And because every phenomenal property has phenomenal character (i.e., what-it’s-like-ness) that is in principle introspectively accessible in this way, every phenomenal property can be introspectively identified E-directly, which means it can be conceived of C-directly.

I just argued that every phenomenal property can be conceptualized in a substantive yet (epistemically and conceptually) direct manner. But there is a third condition a concept must meet to qualify as a phenomenal concept, namely it must capture the proprietary way that we conceive of a phenomenal property from the first-person perspective. That means (at a minimum) it must be distinct from theoretical concepts and other types of concept that are not proprietary to thinking about our own experiences. But it is plausible that theoretical concepts can be both C-direct and substantive, capturing the essence of a property (Chalmers 2010: 183; Loar 2002: 297–98, 305). So in order to establish that every phenomenal property can be grasped by a phenomenal concept, we need to rule out the possibility that the scenarios that I have just been describing can be explained just as well by appeal to non-phenomenal, theoretical concepts.

However, it is clear that the concepts we are considering are not theoretical in nature. First of all, for many of the phenomenal properties we form introspective judgments about, in particular for those that correspond to secondary qualities such as colors, smells, and bodily sensations, either (a) we lack theoretical concepts for these properties altogether, or else (b) the theoretical concepts that supposedly pick out these properties directly (e.g., the concepts of certain neural states) could never be confused with the concepts we utilize when conceiving of our experience phenomenally. Furthermore, even limiting our focus to primary qualities such as shapes and sizes, we often lack theoretical concepts for the precise phenomenal properties we can introspectively single out and think about (e.g., this particular trapezoidal shape phenomenology), since arguably the phenomenal character of experience is far too rich and fine-grained to be exhaustively captured by our theoretical concepts. And even in the case of more general phenomenal properties, we clearly can think about such properties after picking them out in introspection, without having any kind of theoretical understanding of what makes them the properties they are. For example, without knowing what a trapezoid is (or what a mental state is) I can form introspective judgments about the phenomenal property of visual trapezoid experience.

Finally, one might object to this last point by contending that although phenomenal concepts for perceptual experiences of primary qualities can be deployed in the absence of theoretical knowledge, the same could not be said of phenomenal concepts for the corresponding cognitive experiences. So in requiring this condition be met, I am again in danger of begging the question against the defenders of cognitive phenomenology. For instance, suppose for the sake of argument there is a distinctive cognitive-phenomenal property—and phenomenal concept of that property—for the experience of thinking about trapezoids. One cannot think about trapezoids without having some grasp of what a trapezoid is, and arguably the corresponding phenomenal character for this thought must bear a very close relationship (e.g., constitution, grounding) to the content trapezoid. If so, then entertaining thoughts about a cognitive phenomenal property, using a phenomenal concept, plausibly would require awareness of the thought content consciously entertained when undergoing that cognitive-phenomenal experience, which would involve a grasp of the non-phenomenal property about which one was originally thinking. For example, if I cannot think about trapezoids (or at least, not think of them as trapezoids) without knowing something about what a trapezoid is, then arguably I also cannot think about what it’s like to consciously think about trapezoids without knowing something about what a trapezoid is. So the phenomenal concept that would allow me to reflect on what it’s like to have this cognitive experience would need to include some kind of descriptive content, capturing my grasp of both the property of being a trapezoid and, by extension, the property of what it’s like to think about trapezoids.

But this requirement does not establish that such phenomenal concepts would or could be theoretical (or descriptive of any kind), because in this case, the requirement for the subject to have some theoretical knowledge of the property initially derives not from the ability to deploy the phenomenal concept in introspection, but from the ability to think the original thought that is then introspected (about trapezoids, not trapezoid thought experiences). Thus, while it may be true that one could not successfully entertain the thought, “I am now having this [i.e., a trapezoid-entertaining] kind of cognitive experience,” without a grasp of what it is to be a trapezoid or a trapezoid experience, the phenomenal concept used to pick out the phenomenal property in these cases would not be theoretical, because one would not apply it on the basis of a set of theoretical criteria that the phenomenal property must meet in order to fall under the concept. Rather, the subject could apply the concept simply by attending to the phenomenal property present in consciousness and either recognizing it as (a token of) a familiar type, or else forming a novel concept of it, on the basis of its introspected character. That the conscious appreciation of that character would involve a cognitive grasp of the content trapezoid is an artifact of the type of experience being introspected, and thus incidental to how the subject introspectively identifies the phenomenal property and applies the phenomenal concept. Contrast this with the way that we successfully apply a genuinely theoretical concept, for example, trapezoid, to an object in our environment: the competent use of the concept requires at least some familiarity with, and application of, a theoretical account of the nature of the property. So phenomenal concepts are not theoretical, but a genuinely distinct kind of concept, proprietary to the first-person way we learn and think about our own experiences.

4.3. From Being a Candidate for a Phenomenal Concept to Being Available for Introspective Demonstration

This leads us to the second step in our argument that all phenomenology is presentational: the link between a property being conceptualized with a phenomenal concept and being available to be picked out with an introspective demonstrative. Many of the major theories of phenomenal concepts developed in the literature—including indexical, demonstrative, recognitional, constitutive, and quotational theories—describe an act of mentally ostending a phenomenal property as a precursor to forming and applying phenomenal concepts.⁹ But why suppose these theories are right? I will present an argument from inference to the best explanation: first, by showing how an account of phenomenal concepts that appeals to inner demonstration offers a compelling explanation of how these concepts could have the distinctive features described in the last section; second, by considering a number of alternatives to this picture and arguing that they are unsatisfactory.

The key to explaining the conceptually direct yet substantive way that phenomenal concepts allow us to grasp phenomenal properties is the idea that a phenomenal concept is (partly) constituted by its referent. Token phenomenal features directly supply the concept’s content, by literally becoming a part of the conceptual representation. We can call this characteristic metaphysical (or M-) directness to distinguish it from both conceptual and epistemic directness.

Ordinarily, we should not expect that simply putting a property literally inside a representation of that property would give us any intimate access to the property, or do anything special for how it is represented. But phenomenal properties are no ordinary properties: built into any instantiation of a phenomenal property (at least that is minimally attended) is a primitive experiential awareness of what it’s like to have it—which is just to say an awareness of the property itself. Given this fact, if a token phenomenal property partly constitutes a conceptual representation of itself, then when we employ that representation we have an experience with that phenomenal feature, and it makes sense that our cognitive grasp of the property consists (at least in part) in that same basic experiential awareness of itself that instantiating the property always affords a subject.¹⁰ Thus, this account of the concept’s constitution—its metaphysically direct relationship to its referent—explains how the concept could pick out its referent directly in the conceptual sense, that is, without any other properties serving as descriptive intermediaries.

According to David Chalmers’s influential account, the direct constitution of a phenomenal concept by a phenomenal property is achieved “when a subject attends to the quality of an experience, and forms a concept wholly based on the attention to the quality, ‘taking up’ the quality into the concept” (2003: 235). Being “taken up” into a concept requires that the attended feature be used to characterize the experience’s epistemic appearance (i.e., what the subject is inclined to believe, and express in her introspective judgments, about what the experience is like), which in turn requires that one focuses on appreciating the lucid, substantive grasp one has of that phenomenal feature simply by experiencing and attending to it, thinking about it purely based on that grasp.

Moreover, it is hard to imagine that this focused cognitive appreciation could be achieved without a conscious act of inner demonstration. We need to be able to consciously isolate an occurrent feature from the rest of experience, singling it out as a feature of our experience, in order for it to uniquely constitute a way experience epistemically seems to us.¹¹ The act of attention, when suitably combined with the subject’s intentions, achieves this feat by functioning as an internal “pointing” at the feature to be conceptualized: “how does my experience now seem? Like this.”

Still, I have not shown that the link between conceptual and metaphysical directness provided by introspective demonstration is metaphysically necessary. In order to establish that this account is the best available explanation of the evidence, we need to consider two kinds of possible alternatives: first, the possibility that some other concepts could provide conceptual directness without being metaphysically direct; second, the possibility that phenomenal concepts could be M-direct in some way other than through introspective demonstration. Let us take each in turn.

As we discussed above, there arguably are conceptually direct theoretical concepts, which secure directness in a very different manner than by being constituted by their referents. But the phenomenal concepts that are our focus are not theoretical concepts. Instead, one might try accounting for the conceptual directness of our phenomenal concepts wholly by appeal to what Chalmers (2003) calls “standing” phenomenal concepts: those concepts that conceive of phenomenal properties directly in terms of their essential natures even when the subject is not experiencing them. For example, I can conceive of, and know many truths about, phenomenal green, solely in terms of what it’s like to experience it, without presently enjoying a green experience, and thus without any phenomenal feature available to play the constitutive role in my concept required for M-directness. As Chalmers suggests, such concepts probably fix their content by some combination of (a) a subject’s dispositions to recognize, categorize, and discriminate between features of experience, (b) mental imagery that offers rough versions of more vivid and fine-grained non-imagistic phenomenal properties, and (c) dispositions to have such imagery (2003: 239).

However, to the degree that these concepts consist in dispositions, they are “blind” rather than substantive, unless those dispositions are supplemented by some appreciation of the phenomenal character in question that has cognitive significance for the subject. The concept could consist in dispositions to recognize the property when present in experience, provided that, in the event of recognition, the phenomenal character of the experience was not incidental, but rather constituted the property’s epistemic appearance to the subject, and thus her basis for categorizing it as a token of a recognized type: “it’s one of those” (cf. Gertler 2012: 117–21). So a property’s availability to be represented by a standing phenomenal concept would appear to be parasitic upon its availability to be represented by the corresponding metaphysically direct phenomenal concept.

If this line of reasoning is correct, then—absent a novel account of phenomenal concepts that has escaped me—the best explanation for the fact that all phenomenal properties can be grasped in a conceptually direct, substantive way is that they all can be taken up into metaphysically direct phenomenal concepts.

Next, the only alternative I see to the introspective demonstrative account of metaphysical directness is some kind of functionally characterized “quotational” mechanism for forming M-direct phenomenal concepts (of the sort found in Papineau 2006). In Chalmers’s terms (2006b), this account would be “bottom up” in the sense of being “specified in purely physical/functional terms, without building in any assumptions about consciousness”:

The basic idea will be that there are some neural states $N$ (those that correspond to phenomenal states, though we will not assume that) that can come to be embedded in more complex neural representations by a sort of ‘quotation’ process, which allows the original state to be incorporated as a constituent. Perhaps this will go along with some sort of demonstrative reference to the original neural state, so that the complex state has the form ‘That state: $N$.’ Of course, it is not obvious that one can explain any sort of demonstrative reference in physical/functional terms, but I will leave that point aside. (Chalmers 2006b: 190)

I don’t see how this strategy can work, at least not as an alternative to our account. For the “bottom up” quotational account faces a dilemma: either quotational PC’s are unconscious, in which case they cannot play the role in our conscious lives that we have claimed for M-direct PC’s; or they are conscious, in which case their “quotation” process is probably just a functional implementation of conscious introspective demonstration.

Suppose we take the first horn, viz. that quotational PC’s are unconscious. The difficulty for this view is that it betrays the whole point of positing M-direct PC’s in the first place, which is to explain a personal-level phenomenon. Recall that we motivated the claim that all phenomenal properties can be conceived of using metaphysically direct PC’s by discussing the unique epistemic characteristics of phenomenal concepts: conceptual and epistemic directness and epistemic substance. Yet this sort of conceptualization obviously occurs when we consciously introspect on and make judgments concerning aspects of our stream of consciousness as we experience it.

This brings us to the second horn of our dilemma: suppose that quotational PC’s are conscious. We still need to say how the unique characteristics of M-direct quotational phenomenal concepts are manifest at the level of the conscious person. For consider some of the important purposes to which we consciously put M-direct PC’s that distinguish them from standing PC’s.

For instance, sometimes we wish to form maximally precise beliefs about maximally determinate phenomenal properties, so we need PC’s for what we may dub the Precision Stance. But we only ever have access to such properties when they serve as features of our experience. By contrast, standing phenomenal concepts are manifestly inadequate to deliver the level of fine-grainedness that much of our experience possesses (Chalmers 2003). I may have a standing phenomenal concept of phenomenal blue or even of phenomenal azure, but I cannot hold onto a phenomenal concept of a maximally specific shade of phenomenal azure for more than a fleeting moment after the property leaves consciousness.

We also use M-direct PC’s for what we can call the Skeptical Stance: those situations in which one adopts a “a scrupulously cautious attitude” towards the deliverances of present experience, by suspending judgment about, or “bracketing”, background beliefs that might incline one to infer anything about one’s experience beyond simply what it’s like for one at the present moment “from the inside” (Gertler 2012: 111–12).¹² One does so not only when one is feeling genuinely skeptical, but whenever one relies on the phenomenal “appearances” of present experience to exclusively characterize how experience epistemically appears. For example, instead of conceiving of one’s speckled hen visual experience as an experience of a hen or of 103-speckledness, one conceptualizes it as that experience or that phenomenal feature. But then standing phenomenal concepts cannot adequately supply content for thoughts formed in this manner. For even in the unlikely event we have a standing PC for some very specific experiential feature we encountered in the past, there is always the threat of a mismatch between our present experience and the content of the standing PC that we apply to it, and thus of a gap opening up between what experience is really like as we consciously undergo it and how we conceive it to be. But this worry does not get off the ground for constitutive phenomenal concepts.

Taking up either of these two “stances” towards experience is a conscious activity: we are aware that we intend the content of the phenomenal concept to be drawn from unalloyed experience, uncontaminated by background assumptions and conceptual interpretations (Skeptical Stance) and not approximated or generalized but rather captured in all its glorious detail (Precision Stance). So there should be a difference in consciousness between applying these constitutive PC’s and applying mere standing PC’s.¹³

There is an introspectively compelling explanation of the difference: only when applying a constitutive phenomenal concept do I consciously use my attention to the relevant feature of experience as a way of mentally “showcasing” the object of my attention to myself, in order that it might also become (my internal conception of) the object of my thought. After all, how could a subject consciously use a feature of occurrent experience to characterize the feature itself, without first isolating it from the rest of experience? And how else could the subject become conscious of (non-descriptively) isolating a feature in order to make it the object of thought, except by something that makes it seem salient to her? And what better way to efficiently indicate the salience of a feature in consciousness than to focus the subject’s attention on that feature? Moreover, we must add (something like) a conscious intention to refer, by attending, to that which is attended, for otherwise we could have no sense that what we were thinking about was being pulled directly from occurrent experience. The attending must mean “that phenomenal feature” to us. But this is just another way of describing introspective demonstration.

If what we just said is on the right track, then at best the quotational account of PC’s describes an implementation, at the sub-personal level, of the conscious attention to, and demonstration of, a phenomenal feature at the corresponding personal level. Indeed, it is very natural to understand the bottom up quotational account in this way. As the Chalmers passage above suggests, far from it being an alternative to demonstration, embedding the phenomenal state in a larger conceptual representation arguably would only count as a “quotation” process if it were accompanied by something like demonstrative reference to the embedded state.¹⁴ And given that introspective demonstratives must co-vary with some underlying functional process, it is plausible that this process would resemble a functionally defined quotation mechanism.

In sum, the quotational theory fails as an alternative account of metaphysically direct phenomenal concepts. If phenomenal concepts indeed are partly constituted by phenomenal features of experience, then using these concepts requires introspective demonstrative reference to those features.

4.4. From Being Available for Introspective Demonstration to Being Presentational

Finally, we can turn to the last step in our argument for the presentational nature of phenomenal character. I contend that successful demonstrative reference, of the kind used in M-direct phenomenal concepts, requires the demonstrated property be presented. The idea is not that the act of demonstration renders the property presentational, but rather that being presented to the conscious subject in this way (i.e., as a concrete object or property instance) is a prerequisite of being mentally demonstrated; it’s what makes the property available for the subject’s attention to select it.¹⁵ One can’t successfully point—mentally or otherwise—to what doesn’t seem to be there. And by “there” we must at least mean reality as spatiotemporally and causally continuous with one’s self, hence, presented.

One might wonder if the “objects” of introspective attention, and thus of the subject’s intentions to demonstratively refer, really must appear as concrete particulars. For example, suppose we accept my earlier claim that conscious thoughts and their contents are not presented to the subject in experience: why couldn’t it be the case that the conscious conceptual contents, for example, of some bit of mathematical reasoning, are available to be introspectively demonstrated?

I don’t think the possibility we are being asked to entertain here is conceivable. For one thing, I can’t perform the act of introspectively demonstrating my abstract thought contents so as to apply a phenomenal concept to them while my cognitive faculties are immediately engaged with entertaining the thought itself. And by the time my mind is free to entertain this second, metacognitive thought, the original one is long gone. By contrast, perceptual contents can remain suspended in consciousness, in relatively unchanged form, for a sustained period of time, allowing me to pick them out through an act of attention. Thus, the only way we might “capture” abstracta like our thought contents with our attention, so as to make them the targets of demonstration and form singular thoughts about them, is if, contrary to fact, we could both think a thought and focus attention on some aspect of that thought as a means of thinking about it (qua component of our cognitive experience), at the very same moment.

In reality, the best we might do is to consciously think through a quadratic equation, say, and then think, “all of that was realized by a concrete mental process in my mind.” But in this case ‘that’ really does not stand for the referent of a true demonstrative, but of a disguised description with indexical elements: something like, the mathematical contents I thought through just now. In applying this description to my thought, I certainly haven’t exercised a concept that has some part of my episode of mathematical reasoning as a constituent, since I am unable to attend to and “take up” into a concept any part of this episode or its contents once it is finished.¹⁶ And if memory allows me to recover something of the episode that could be demonstrated after the fact, it typically will be some bit of inner speech or imagery.

Nor do the abstract details of my mathematical reasoning offer themselves to me as potential objects of introspective demonstration even as I think them. The problem is not simply the result of our inability to think two thoughts simultaneously, but of the deeper fact that grasping abstract thought contents is just not the same sort of thing as holding something stable and available “before the mind” to be demonstrated. Of course, we regularly pause to deliberately attend to what we are thinking, in the sense of consciously rehearsing, with increased focus, a thought that we were about to think or had just thought unreflectively a moment earlier. Yet in spite of my being, in a natural sense, conscious of the quadratic equation as I think about it, the equation itself neither seems to me to be an object furnishing the concrete inner world of my mind, nor a datable event or process flowing through my stream of consciousness (still less a feature of my external environment). By contrast, when I set myself the task of imagining what it would be like for abstract objects—a quadratic equation, or the set of all round things—to be proper targets of my inner demonstration, I inevitably just imagine concrete entities—written numerals, mathematical symbols, or circles—located in external space or in my own mind. And this failure does not look like an accidental fact about our imaginative faculties, since if these abstract objects were available to be demonstrated, then by virtue of this availability they would be (apparently) connected to me in ways that would render them non-abstract. Thus, all my attempts to reach Plato’s Heaven and its inhabitants by introspective demonstration come crashing back down to earth.

At this juncture, someone might object as follows.¹⁷ Consider blindsight: in ordinary cases, the subject cannot perceptually demonstrate objects or properties in the blind part of her visual field, nor form demonstrative thoughts about these objects. But we can easily imagine an augmented blindsighter whose unconscious mental processes can track and single out unseen objects and properties in front of her so as to enable her to form (quasi-) demonstrative thoughts about them. After all, if there is nothing conceptually incoherent about Ned Block’s hypothetical “super-blindsighter,” who can confidently and spontaneously use unconscious perceptual information to guide her reasoning, speech, and behavior, then there is no obvious reason to think giving this subject the further ability to demonstratively refer to things in her blind field is a theoretical impossibility (Block 1995).

Now, if perceptual demonstratives are possible for the subject in a hypothetical case of augmented blindsight, what does this tell us about introspective demonstratives? The case would appear to provide a counterexample to the final premise of our argument. For the blindsighter has no conscious experience of the items that she demonstrates; and if demonstrative reference does not require that the referents be experienced at all, it a fortiori does not require that they be presented to the subject in experience!

One way to respond to this objection is to contend that while conscious experience may not be necessary for perceptual demonstratives, introspective demonstratives are different. After all, what is the introspective equivalent of the super-blindsight case supposed to look like? In order for there to be a successful case of introspective demonstrative reference, there must be features of experience to which the subject can refer, and on many accounts, any experience entails the subject’s minimal conscious awareness of the experience. So it is hard to see how the subject could be “blind” to her own experiences.

The problem with this response is that, with a little ingenuity, we can frame a hypothetical scenario in which the subject is unaware of her own experiences, in a sense that is arguably relevant to the present discussion. We then can add that the subject has the power to use information about features of her experience of which she is unaware, and to form demonstrative thoughts about them.

For example, we can adapt John Campbell’s Sea of Faces example (2002: 8–9; discussed at length in Smithies 2011). Campbell asks us to imagine being at a crowded dinner party and seeing a “sea of faces” without being able to focus on any particular one of them. The intuition one is invited to have is that the subject will not be able to entertain any demonstrative thoughts about the faces, because demonstrative thought requires not just perceptual experience but also conscious perceptual attention. But imagine a scenario in which the subject in this situation has a kind of blindsight ability just with respect to compensating for the lack of conscious attention. The subject (a) experiences the faces, yet (b) fails to consciously attend to any one of them, with a corresponding lack of the accompanying phenomenal changes brought about by shifts of visual attention—it’s still all just a sea of faces. Nevertheless, (c) she is able to unconsciously select faces from the sea, on the basis of which she can confidently and spontaneously form demonstrative thoughts about the object(s) selected (Smithies 2011: 9–15). Next, change the case to be about introspection rather than perception: even without the ability to introspectively attend to any of the details of her visual experience, the introspective “blindsighter” has unconscious processes that play the functional role of introspective attention and demonstration, allowing her to refer to phenomenal features (e.g., of visual face experiences). But there is a natural sense in which she is not aware, or not sufficiently aware, of these phenomenal features. At the very least, it is not obvious that if the subject demonstrates features of her experience without the aid of conscious attention, those features must be presented to her. So if this case describes a real possibility, it undermines the intuitive link between introspective demonstrative reference and the presentational nature of experience.

In light of this example, I think the best strategy for responding to the augmented blindsight challenge is to contend that such cases don’t touch the kind of demonstrative reference involved in M-direct phenomenal concepts. This response works regardless of whether or not we think the blindsighter really can form demonstrative thoughts about experiences in her conscious stream to which she cannot attend. For purported cases of introspective demonstrative reference without attention such as the modified Sea of Faces example at best only involve conceiving of the demonstrated objects in a way that is non-substantive, in contrast to the substantive grasp of phenomenal features afforded us by phenomenal concepts. Thus, the moral to draw seems to be this: features of experience need not be presented in order to be introspectively demonstrated, provided that the sort of introspective demonstration in question also does not require conscious attention to the features at all.¹⁸ We argued in the previous section that the use of direct phenomenal concepts requires a kind of introspective demonstration that essentially involves a conscious act of focusing attention onto a feature of experience, as opposed to a merely sub-personal “quoting” process. Thus, at this point in our argument, we can assume that the introspective demonstration used in the formation and application of phenomenal concepts requires attending to features of experience. As a result, we are left with no good objection to the conclusion that it also requires attending to something presented in experience.

This concludes my argument that all phenomenology is presentational.