Philosophical Studies

, Volume 159, Issue 3, pp 447–455

Conscious states as objects of awareness: on Uriah Kriegel, Subjective consciousness: a self-representational theory


    • Corcoran Department of PhilosophyUniversity of Virginia

DOI: 10.1007/s11098-011-9763-9

Cite this article as:
Gertler, B. Philos Stud (2012) 159: 447. doi:10.1007/s11098-011-9763-9

1 Introduction

It is a real pleasure to discuss this book, which makes a significant contribution to consciousness studies. Self-Representationalism is a bold and promising new theory. While the theory builds on some key insights of familiar theories of consciousness, especially higher-order theories, it goes far beyond existing theories in its development of these insights. A particular strength of the book is its sustained focus on phenomenology. Kriegel never loses sight of the real explananda of his theory, and he tirelessly tests the metaphysical and ontological details of the account against the phenomenological data. The combination of technical sophistication and phenomenological sensitivity makes for an exceptionally compelling, nuanced position.

My critical remarks will focus on a thesis that underpins Self-Representationalism: namely, that phenomenally conscious states are states the subject is aware of. This thesis is crucial to Kriegel’s defense of his theory, since the idea that we are aware of our conscious states is what motivates the claim that conscious states are represented. That claim forms a central premise in the argument to show that conscious states are self-represented.

Here is Kriegel’s formulation of this thesis, which he calls the awareness thesis (or “AT”).

Necessarily, for any mental state M of a subject S, M is phenomenally conscious (at a time t) only if S is aware of M (at t). (p. 300)1

Kriegel supports this thesis throughout the book, using a variety of arguments. Some of these arguments are conceptual; some are phenomenological; and some are abductive.

I will first elucidate the AT, by considering Kriegel’s conceptual case for it. I will argue that the conceptual case for the AT does not rule out an alternative, less radical picture of the relation between consciousness and awareness. After sketching this alternative, I will examine a few of Kriegel’s empirical arguments for the AT, showing how my alternative accommodates or explains the relevant data at least as well as the AT. My goal is not to present a competing theory of consciousness, but rather to suggest that Self-Representationalism’s central benefits can be secured without embracing the AT.

2 The conceptual argument for the AT

Crucially, the AT does not function in Kriegel’s argument as a stipulation, intended to fix what is meant by “phenomenally conscious”. Rather, he defines phenomenal consciousness as that phenomenon that generates the explanatory gap. The truth of the AT thus depends on whether the states that generate the explanatory gap—that is, those responsible for the “hard problem” of consciousness—are states of which the subject is aware.

Kriegel makes the following conceptual case for the AT.2

Conscious experiences are not states which we may host, as it were, unawares. Freudian suppressed states, sub-personal states, and a variety of other unconscious states may occur within us completely unbeknownst to us, but the intuition is that conscious experiences are different. … Conscious states are not states that just happen to take place in us, whether or not we are aware of their taking place; they are also for us, precisely in the sense that there is something it is like for us to have those states. Mental states that merely occur in us, but of which we are completely unaware, are not conscious experiences. (p. 16)

This conceptual case for the AT consists of two main steps. The first is explicit: conscious states are “for us”, in that “there is something it is like for us to have those states”. This first step seems unproblematic, since it is relatively uncontroversial that what poses the hard problem is that there is “something it is like” to be in a conscious state.

My doubts concern the second step, which is implicit. According to this step, if there is something it is like for one to have a state, then one is aware of that state. To appreciate what is involved in being aware of something, look at an object in your environment and consider your awareness of that object. Insofar as you are aware of that object—a water glass on a table before you, say—it is an object of your awareness. Similarly, on Kriegel’s view every conscious state is an object of awareness. So when you eat a forkful of blueberry pie, your conscious gustatory experience is a state of which you are aware—an object of your awareness—just as the water glass is an object of your (perceptual) awareness.

Of course, there are important differences between your awareness of the water glass and the type of awareness that, on Kriegel’s view, is implicated in phenomenal consciousness. When you visually attend to the glass, it is an object of your focal awareness, whereas Kriegel claims that phenomenal consciousness requires only peripheral awareness. And in perceiving the glass your state of awareness, which is mental, is distinct from its object, the glass itself. By contrast, on Kriegel’s view phenomenal consciousness involves a state of awareness that is one and the same as its object. This is why he says that a phenomenally conscious experience, such as the experience of eating blueberry pie, is self-representing: an experience constitutes awareness of itself (that very experience).

But while the “awareness” relation at issue in the AT differs from ordinary perceptual awareness, there are also crucial points of similarity. Perhaps most striking is that both types of awareness are epistemically salient. According to Self-Representationalism, a state’s being phenomenally conscious is (at least partly) a matter of the subject’s bearing a particular epistemic relation to that state, one broadly similar to the epistemic relation I bear to the water glass when I am even peripherally aware of it. Kriegel explicitly characterizes the “awareness of” relation in epistemic terms.

Since this awareness is awareness-of, it involves an of-ness relation to the experience. …[I]t involves essentially the subject bearing an epistemic relation to her experience. (p. 104)

So Self-Representationalism portrays phenomenally conscious states as playing two roles. First, they are states (or vehicles) of awareness: that is, they constitute awareness of something. Second, they are objects of awareness: that is, they are states of which the subject is aware.

3 An alternative to the AT

Kriegel provides a range of arguments to show that conscious states must be both states of awareness and objects of awareness. I will present some reasons to doubt that phenomenally conscious states must play the latter role. I will suggest that a proper construal of the first role, being a state of awareness, will accommodate the conceptual and empirical data marshaled in favor of Self-Representationalism. If I am correct, a suitable understanding of what is involved in being a state of awareness will obviate the perceived need for supposing that conscious states are also objects of awareness.

Imagine eating a slice of blueberry pie. If you were to attend to the taste of blueberry, your experience would be an object of your awareness: you would be (introspectively) aware of it.3 But suppose that you are not, in fact, paying any attention to the taste of the pie. Instead, you are absent-mindedly eating while your attention is entirely occupied by a byzantine tax form you are desperately trying to complete.

Even in this case, there is something it is like (for you) to taste blueberry; the taste is part of your overall phenomenology. Self-Representationalism accommodates this fact by saying that you are aware of your blueberry-pie-eating experience; the fact that you are not attending to the experience means that your awareness is peripheral rather than focal. But an alternative is to say that your experience contributes to your phenomenology simply in virtue of being a state of awareness. The experience need not also be an object of awareness.

For the purpose of explicating this alternative, let’s assume that mental states are property instantiations. (Kriegel mentions this as one of two plausible ontologies of mental states—the other is that mental states are bare particulars.) When you absent-mindedly munch a forkful of blueberry pie, you have a “qualitative blueberryness” experience. This experience is, we are supposing, an instantiation of qualitative blueberryness: it is a concrete tokening of a qualitative property. On the alternative I have in mind, to have a phenomenally conscious experience of qualitative blueberryness is to be in a particular state of awareness, namely, a state that constitutes awareness of qualitative blueberryness. What is crucial here is that the object of this awareness is not the state itself—for you are not aware of your instantiation of that property. Rather, in instantiating qualitative blueberryness you are aware of the property qualitativeblueberryness.

Now a natural view is that, in your phenomenally conscious qualitative blueberryness experience, the object of your awareness is the pie, which has the property represented by the qualitative character of your experience. (This claim might be supplemented with Kriegel’s construal of qualitative character as representing response-dependent properties.) I prefer to think of the object of your awareness as the property qualitativeblueberryness itself. But the crucial point is that your experience, which is an instantiation of the relevant qualitative property, does not represent itself. It instead represents something else: perhaps the pie (an external object) or qualitative blueberryness (a property rather than an instantiation thereof).

This is not to say that instantiating a qualitative property simply consists in an awareness of that property. There are various ways to think about qualitative properties, and hence to be aware of them, without instantiating them: e.g., I may think of qualitative blueberryness just as the property that I would instantiate were I to eat blueberry pie. So perhaps instantiating a qualitative property involves a particularly direct type of awareness, one that could enable demonstrative reference to the property.4 My point is only this. We can grant that phenomenal consciousness is conceptually linked with awareness without insisting that phenomenally conscious states must play two roles, as both states of awareness and objects of awareness. The first role seems to do justice to this conceptual link on its own.

I propose, then, that we replace AT with a necessary condition along the following lines.

Necessarily, for any mental state M of a subject S, M is phenomenally conscious (at a time t) only if there is a qualitative property Q such that M constitutes S’s awareness of Q (at t).

Depending on how we construe qualitative properties, S’s awareness of Q may consist in her awareness of an external object (such as the pie), her awareness of Q’s instantiation in an external object, or her awareness of a property (such as qualitative blueberryness). There are other possibilities as well.

This proposal will not accommodate all of the intuitions Kriegel cites in support of Self-Representationalism. Perhaps most significant is that, in allowing that a subject may be in a phenomenally conscious state of which she is unaware, the proposal conflicts with one of the principal motivations for Self-Representationalism, namely:

Mental states that merely occur in us, but of which we are completely unaware, are not conscious experiences. (p. 16)

But the proposal preserves intuitions closely related to this. It implies that phenomenal consciousness is conceptually linked with awareness. And as I will argue in a moment, it is compatible with the idea that we are generally able to become aware of our phenomenally conscious experiences—and, hence, that a state’s being phenomenally conscious implies that it bears an epistemic, accessibility relation to the subject.

Kriegel’s original awareness thesis, AT, is much stronger than my proposed alternative, AT*. But the appeal of AT may largely derive from the intuition that consciousness is conceptually tied to epistemic awareness. And AT* accommodates that intuition.

Let us now turn to the empirical arguments for AT.

4 Empirical arguments for the AT

To soften resistance to the idea that we are aware of all of our phenomenally conscious states, Kriegel offers a variety of empirical arguments to show that peripheral awareness of our own states—what he calls “peripheral inner awareness”—is ubiquitous in our experience. Some of these arguments are phenomenological; some are abductive. If successful, they show that AT not only accommodates the phenomenological evidence but is the best explanation of a wide range of empirical data.

I will consider a few of Kriegel’s arguments and argue that my proposed alternative, (AT*), fares as well as AT as regards them.

4.1 The Argument from Introspection

Kriegel recognizes that we engage in introspection only rarely. To explain how introspective awareness differs from the relatively ubiquitous awareness required for phenomenal consciousness, he adopts the “attention-shifting” model of introspection. On this model, introspection involves focusing one’s attention onto a particular state, and thereby transforming one’s awareness of that state from merely peripheral awareness (which is ubiquitous) to the much rarer focal awareness.

Whereas Kriegel’s view construes introspection as a shift in the focus of awareness, on the alternative I’ve outlined introspection involves a shift in the object of awareness. This alternative would say that, as I absent-mindedly munch blueberry pie, my instantiation of qualitative blueberryness is not an object of my awareness at all. I become aware of this mental state only when I introspect my blueberry-pie-eating experience. When this occurs, the object of my awareness shifts from the qualitative property (or, perhaps, the pie) to my mental state, which is the instantiation of that property.

Kriegel argues that the phenomenology of introspection favors his “attention-shifting” model over alternatives that, like my proposal, take introspection to introduce a new object of awareness.

[The attention-shifting model] has the added advantage of illuminating the fact that introspecting does not feel, phenomenologically, like performing a “dramatic” mental act, an act that creates an altogether new representation (as, say, visualizing a cat does). Instead, it feels more like shifting around one’s attention, and attending more carefully to contents that were already previously there. (p. 183)

To test this phenomenological claim, I invite you to reflect on how your feet feel in your shoes just now. In doing this, you are shifting your attention to an aspect of your current phenomenology. And there is a definite sense of familiarity—it seems to me, at least, that this feet-in-shoes feeling has been present for a while, though until this moment I hadn’t attended to it. So we can agree with Kriegel that introspection involves a shift of attention, and that the introspected experience typically appears familiar.

But these observations do not show that the shift of attention simply relocates your experience from the periphery to the focus of attention. For they are compatible with my alternative, which says that the shift of attention makes what was previously merely a state of awareness—the phenomenal “feet in shoes” experience—into an object of awareness. My alternative explains why the new representation will seem familiar. When you were merely having the feet-in-shoes experience, without attending to it, you were aware of the qualitative property. (The object of your awareness may have been the qualitative feet-in-shoes property itself; or it may have been the relation between your feet and your shoes; etc.) This is why attending to your experience “does not feel, phenomenologically, like performing a ‘dramatic’ mental act”: you were aware of its characteristic qualitative property all along, though until you turned your attention to your experience you were not aware of your experience.5

My proposal accommodates another argument of Kriegel’s, the “Memorability Argument”, in a similar way. The Memorability Argument is straightforward. Kriegel notes that he can remember how his morning orange juice tasted, and argues that this means he was aware of his experience of the orange juice when he had it. On my alternative, his ability to remember the experience is explained by the fact that, when drinking the orange juice, he was aware of the relevant qualitative property. He can therefore “call up” this property and replay the experience, so to speak, in episodic memory. Attending to the current episodic reliving of the original experience, he can make the orange-juice-drinking experience (the instantiation of qualitative orange-juiciness) an object of awareness even if the original experience was only a state of awareness.

A different issue arises in Kriegel’s “Sophisticated Argument from Blindsight”. Kriegel argues that the AT provides “the simplest explanation” of why a normally sighted individual has a conscious state when confronted with a particular visual stimulus, while a blindsighter presented with the same stimulus has an unconscious state. The explanation is simply that the normally sighted person is aware of his visual state, while the blindsighter is not. Kriegel adds an extra step to this argument (this is what renders it “sophisticated”), drawing on research suggesting that blindsighters are susceptible to the same priming effects as the normally sighted. Since it seems likely that qualitative properties are responsible for priming effects, this result implies that blindsighters instantiate the relevant qualitative properties. Kriegel concludes that we should allow that their visual states have these qualitative properties, while attributing their lack of phenomenal consciousness to the fact that they are unaware of these qualitative properties. According to the AT, this means that their visual states are not phenomenally conscious.

Interestingly, my alternative allows for an interpretation of the blindsight case that is very similar to this Self-Representationalist interpretation. I agree that what is lacking is awareness of the qualitative properties of experience; but on my view, this does not entail that the experience was not phenomenally conscious. It simply means that the experience is one the blindsighter cannot introspect. I am operating with Kriegel’s definition of phenomenal consciousness as the phenomenon that makes consciousness seem mysterious. So the issue here is whether the blindsighter’s experience exhibits the phenomenon that makes consciousness seem mysterious. It seems to me that a particular state, such as the blindsighter’s visual state, could exhibit that phenomenon even if the subject himself cannot introspect that state—and, hence, cannot appreciate the mystery through reflection on it.

This brings me to the source of my concerns about Kriegel’s Self-Representationalism. At bottom, my concerns are epistemic. It seems to me that one may have a phenomenally conscious state, in Kriegel’s sense, without standing in the sort of robust, actualized epistemic relation to that state required for awareness of it. This is not to say that phenomenal consciousness is epistemically inert. On my view, the epistemic significance of phenomenal consciousness is that phenomenally conscious states are a type of state that can be known with an especially high degree of certainty (though not every subject, in every circumstance, can achieve this degree of certainty as regards each of her phenomenally conscious states). They have this feature because, as it is sometimes said, there is no gap between appearance and reality as regards phenomenally conscious states. More precisely: phenomenal states are as they epistemically appear to doxastically cautious, careful subjects.6 This means that subjects can sometimes achieve especially certain knowledge of their phenomenal states. But it does not mean that one is aware of all of one’s phenomenal states.

Like Kriegel, I believe that phenomenal consciousness is epistemically significant. But my construal of this significance, unlike Kriegel’s, allows that someone could have a particular phenomenally conscious state that doesn’t epistemically appear to her at all. Since this possibility seems to me to be perfectly coherent, I regard compatibility with it as an advantage of my view.

Kriegel makes a provocative suggestion that bears on these epistemic issues. He suggests that a state’s subjective character, which partially consists in the subject’s awareness of the state, is the categorical basis for the dispositional property of access consciousness.

[S]ubjective character seems to play the right explanatory role vis-à-vis access consciousness. The reason why a mental state is poised for the subject’s free use in personal-level reasoning and action control, it is reasonable to suppose, is that the subject is already aware of it.

This approach to access consciousness is, I think, very promising. It distinguishes phenomenal from access consciousness while doing justice to the intimate relations between them. Most importantly, it enables consciousness to play a robust explanatory role: I can access my experience of blueberryness, and use it to guide reasoning and action (deciding to have another slice of pie, say), because it is phenomenally conscious. This is a major benefit of Self-Representationalism. It avoids a serious pitfall of theories that reduce consciousness to functional or dispositional properties: such theories deny, implausibly, that consciousness explains the functional or dispositional properties at issue.

But while the suggestion that phenomenal consciousness is the categorical basis of access consciousness is promising, I think that its promise would be better realized by adopting the sort of alternative I have proposed. As a candidate for the categorical basis of access consciousness, awareness of a qualitative property is more fitting than awareness of my experience—that is, my instantiation of that property. Taking the categorical basis of access consciousness to be a state of awareness that is not an object of awareness provides a more natural explanation of our use of mental states in reasoning and action. For instance, my disposition to infer that I may safely press the accelerator is explained by my awareness of the green light, not by my awareness of my experience of the green light. What secures this key benefit of the Self-Representationalist view is not the AT but, rather, the fact that conscious states are states of awareness. So my weaker alternative, AT*, suffices for this purpose.

5 Conclusion

I have raised some doubts about the plausibility of, and need for, the claim that phenomenally conscious states must be objects of awareness. Of course, space constraints have forced me to ignore some of Kriegel’s arguments for this claim. One in particular deserves mention. Kriegel argues (in the Appendix) that the AT is needed to explain cognitive phenomenology, since such phenomenology is not a matter of, say, imagery. He would presumably argue, on similar grounds, that a difference in the qualitative properties of which one is aware will not secure the relevant difference in cognitive phenomenology. The resolution of this question will depend on how one construes qualitative properties, and I have not addressed that issue here.

I’m grateful to Kriegel for inspiring me to think about these challenging questions, by addressing them with such clarity and rigor. I now have more doubts about these issues than I did when I first opened his book—and that is a clear mark of his achievement.


All page references are to Kriegel (2009).


This passage is directly concerned with subjective character. But since subjective character is what makes a state phenomenally conscious, on his view, these remarks support the AT.


Some philosophers moved by the “transparency of experience” doubt whether we are ever aware of experiences as such; but it seems to me clear that we do achieve such awareness. This claim is modest, relative to the current dialectic, since Kriegel thinks that we are at least peripherally aware of all of our conscious experiences.


This is connected to Kriegel’s claim that subjective consciousness requires essential self-reference.


Kriegel uses this point about familiarity in his “Argument from Surprise”, to which my response directly applies.


This is not to say that every aspect of their phenomenal reality is reflected in their epistemic appearance. As the speckled hen case has taught us, the epistemic appearance of a phenomenal state does not always exhaust its phenomenal reality. My point here is only that, for careful, doxastically cautious subjects, judgments about phenomenal states based on how those states epistemically appear can achieve an especially high degree of certainty. (I develop this view in Gertler forthcoming.)


Copyright information

© Springer Science+Business Media B.V. 2011