Consciousness is perhaps the hottest of all hot topics in contemporary philosophy of mind. The range of accounts of consciousness on offer is breathtaking, as is the panoply of distinctions between different kinds of consciousness. To survey, let alone to adjudicate, the range of options would require a substantial volume of its own (and fortunately, there are many such substantial volumes, such as Blackmore 2004, Chalmers 2010, Block, Flanagan and Güzeldere 1997, Dennett 1991, Zelazo, Moscovitch and Thompson 2007). Since these volumes exist, I don’t need to do that hard work here. Instead, I am going to rehearse what I see as the most important specific debates and alternatives, the complex set of presuppositions they involve and questions they raise.
As it is, this chapter addresses a range of issues, each of which might merit its own chapter, or indeed its own book. We will consider questions about the nature of perceptual experience, about the nature of introspection, the status of qualia and even the possibility of zombies! I don’t pretend to offer the last word on any of these matters, or even the last word on the way that Buddhist philosophy can contribute to discussions about these matters. Instead, I want to draw attention to a thread that links them, and to the ways in which adding Buddhist fibers as we spin that thread can make it stronger. That thread is the understanding of what it is to say that we are conscious and what the nature of consciousness, in its various senses, might be.
There is a widely accepted distinction in the Western literature between two senses of the term “consciousness,” that is, between what is I will call “access consciousness” and what is often called “phenomenal consciousness.” By “access consciousness” I will mean introspectibility.1 A psychological state or process is conscious in this sense if we are able to report on it, to take awareness of it as input into our reasoning processes. My belief that I am typing right now is access conscious in this sense. I can introspect and determine that I do believe this.
“Phenomenal consciousness,” on the other hand, denotes something more like the felt character of experience, often denoted by the problematic phrase, due to Thomas Nagel, “what it is like” to be in a particular state. So, unlike my belief, which doesn’t feel like anything, my typing right now is phenomenally conscious. It feels like something to be typing. At least to a first approximation, an approximation from which we may backtrack later. Phenomenal consciousness might also be called qualitative consciousness, and that of which we are phenomenally conscious is sometimes referred to as “qualitative character,” or even hypostasized into qualia.
This distinction is already a bit too crude, however, even to capture the questions debated in this literature.2 To these we might add two more senses. We might want a term to denote the property of being responsive to a stimulus, even if that responsiveness is not introspectible and has no phenomenal character. This might be the right way to think about low-level proprioception or our responsiveness to happenings on the road when our primary attention is on the radio news broadcast, and such things. Call this “responsive consciousness.”
Finally, we might want to discuss the pre-reflective background horizon of awareness, or bare subjectivity, that many, especially in the Husserlian tradition, regard as the most basic precondition of any kind of conscious activity. This kind of consciousness might be thought of as the non-thematized awareness of ourselves as subjects of our conscious states, and of those states as our states. To be conscious in this sense is simply to be a subject and, at least pre-reflectively or transcendentally, to take oneself as a subject. In order not to beg any philosophical or historical questions regarding the precise character of whatever might be denoted by “consciousness” when used in this sense, let us call this “subjective consciousness.”3
So, when we ask what consciousness is, how and whether it can be explained, what its relationship is to the body, whether it is discursive or not, or even whether there is such a thing, we need to be clear in each case and each context, in which of these four (and more may be forthcoming) senses we are using the term. It is one thing merely to be subjectively conscious. We might experience this in the barest moment of awakening in the morning. It is another to be responsive: we orient toward a sound that we have not even registered in any introspectible awareness; the sound then acquires a qualitative existence in our mental life, as in the case of the sound of our alarm clock and we are phenomenally conscious of it. A moment later we find that we can introspect and report on our state of mind—all of this is now accessible.
It might be tempting to think of this unfolding of awareness in the first few seconds of our day as a simple increase in a single dimension—as we wake up, we simply become progressively more conscious. If, however, this set of distinctions among senses of the term “consciousness” (leaving aside the question of whether the term in any or all of these senses denotes anything at all—at this point we are simply mapping the way these terms are used in contemporary philosophical discourse) is right, then we aren’t talking about the same thing when we say: first I was simply conscious, then conscious of something, I know not what, then of a sound, and then that I was waking up. Instead, new kinds of relations between subject and object (or at least kinds of states of subjects), perhaps interrelated, emerge as our engagement with our environment, inner and outer, becomes progressively more complex.
It is also worth noting that responsive and access consciousness are each characterized as explicitly relational. That is, when we say that we are conscious of a particular stimulus in the sense that we are responsive to it, or that we are conscious of an inner state in the sense that we can introspect and report it, we are explicitly describing relations we as subjects bear to particular objects. There is not even a whiff of an idea that we are describing a monadic or intrinsic property of subjectivity itself. And indeed, the relation we describe here is not even prima facie different from that we would describe between a thermostat and the temperature of a room in the case of responsiveness, or that between a fuel gauge and a fuel tank in a car in the case of the second.
It is tempting to respond to this observation, as have many who are interested in the nature of consciousness, by saying that these kinds of consciousness are not what we are really talking about when we ask about the nature of consciousness—in the now familiar jargon due to Chalmers (2010), they do not pose the hard problem. Zombies might be conscious in these senses. The hard problem of explaining consciousness, both analytically and scientifically, this line of thinking continues, concerns phenomenal and subjective consciousness. These are the kinds of consciousness, Zombologists argue, that we have, but that Zombies—beings behaviorally and functionally identical to us, but without any genuine inner life or phenomenology—lack. There is something that it is like to be us, they urge, and that something is captured by our experience of ourselves as subjects and by the phenomenal character of our perceptual and affective experience.
This line of argument has a kind of plausibility to it, and we will return to it below, with attention both to the conceptions of consciousness at work and with respect to the possibility of zombies. But for now, it is worth registering one preliminary reason for skepticism about the wedge the argument requires between responsive and access consciousness, on the one hand, and phenomenal and subjective consciousness, on the other. That wedge, as we have seen, finds its narrow edge in the distinction between relational and non-relational senses of consciousness. If we are to take this seriously, we have to take seriously the idea of a fundamental property of our mind that is intrinsic to it, a kind of mere objectless awareness. This is not as obviously a given, a self-presenting explanandum, as it might appear to be.
In this context, it is worth considering the complex role of affect in consciousness, another matter about which we will see that Buddhism has much to say. In contemporary discussions of consciousness in the West, affect often enters into the picture as an object of awareness, typically of phenomenal consciousness.4 On this view, represented perhaps most prominently by Chalmers (2010), our affective states have a phenomenal feel, and we are conscious of these much as we are conscious of sensory qualities.5,6
There is an immediate problem with this view, suggested by Buddhist analyses, to which we will turn below. On the one hand, we should countenance the possibility of affect that is neither accessible to introspection nor phenomenally present—what we might call in a Freudian moment, “subconscious” affect. Emotions that circulate below the radar may nonetheless determine much of our conscious life. And this leads us to the second problem: It may be more important to think of affect as a cause, rather than an effect or an object of the character of our conscious engagement with the world. As noted in passing above, this is very much part of a Heideggerian (and Freudian) picture in the West, but is in tension with views according to which we have privileged access to our affective lives, and so according to which all emotion is to some degree accessible. It is also a central part of a Buddhist understanding of the relationship between the five skandhas and of the nature of the origin of dukkha.
Buddhist discussions of consciousness feel radically different from contemporary Western discussions. Nonetheless, we will see that they can be valuable sources of viable alternatives, both with respect to positions on the topic and, more fundamentally, with respect to how questions and debates are framed in the first place. But first, we should be clear about terminological issues, as it is easy for participants in these atemporal, cross-cultural discussions either to talk past each other or to find their positions seriously distorted by misleading translation. We will have to acknowledge that to a certain extent Buddhist and contemporary Western discussions of consciousness occur in different intellectual contexts and are pursued for different reasons, in different vocabularies. But if we are clear about this, we will still find important points of contact.
It is methodologically useful to recall that our own philosophical vocabulary, as well as that of any other tradition, often structures, rather than reflects, the ontology of a domain or inquiry. This is particularly true of the psychological domain, as Wittgenstein notes in Philosophical Investigations and as Sellars notes in “Empiricism and the Philosophy of Mind.” In this domain, we become so accustomed to the use of a language, and a way of mapping the inner domain, that we take it for granted that the language we use and in which we think carves the inner world at its joints, joints that are simply apparent to our introspective gaze. In doing so, we forget the origins of our terms, which often lie in theoretical frameworks we explicitly reject, and forget that other traditions carve the mind in radically different ways, which they equally take as obvious. Taking another tradition seriously hence requires taking its vocabulary seriously and adopting a conversation horizon that allows us to problematize our own—and their—way of taking the landscape.
Let us begin with some basic Sanskrit terminology and rough glosses to get a sense of the lexical terrain and move from there to a more conceptual overview. There is a plethora of Sanskrit terms (and Pāli, Chinese and Tibetan equivalents or rough equivalents—for the sake of convenience I will focus on the Sanskrit here) that denote phenomena in the conceptual neighborhood of the already semantically complex English “consciousness.” I choose a few that are most relevant to the present discussion.
The most central term is probably vijñāna and the various compounds in which it figures. The term has a complex semantic range across Indian philosophical literature. We can read it as indicating that which enables knowledge. In Buddhist literature its most important locus for present purposes is in its denotation of the fifth skandha, or aggregate underlying personhood. The functions of that skandha are manifold, including conceptual thought, judgment, attention, reflection, etc…
We can get a sense of how vijñāna functions when we see it in the context of the psychology of the skandhas, in which we begin with the action of the physical body (rūpa), as in the functioning of a sense organ, leading to sensation (vedanā), which in turn gives rise to perception (saṃjñā)7—note again the root jñā (cognate with the kn or gn in terms like knowledge or gnosis in European languages, denoting awareness)—in which that which is sensed is initially, but in a sense preconsciously, brought to cognitive determinacy, and as a possible object of knowledge through the mediation of our affective and habitual predispositions, which, again, function prior to consciousness.
A second set of terms relevant to this discussion includes saṃveda (a term that figures prominently in the compound svasaṃvedanā, denoting reflexive awareness, to be discussed below) and anubhava. Saṃveda connotes perceptual awareness, the kind of consciousness of something we have when we perceive it, a consciousness that is both phenomenal and recognitional. When I am conscious of something in this sense, I both sense it and categorize it. Anubhava has a similar semantic range, denoting consciousness of something immediately present.
Samprajaña, a term often translated as mindfulness, is used to denote attentiveness to one’s own current state, a higher-order, introspective awareness. By distinguishing samprajaña from saṃveda or anubhava, the Sanskrit Buddhist tradition presupposes that it makes sense to talk about qualitatively rich perceptual consciousness, consciousness capable of guiding behavior, without presupposing that one is explicitly aware that one is having such consciousness.
One final term in this large family may be relevant for present purposes. Cetanā is most often translated as intention and figures most prominently in discussions of action and responsibility.8 And this is fair enough. But it shares a good part of the broad philosophical semantic range of the technical English (Latin) term intentional, indicating as well directedness on an object. Cetanā can hence also refer to the state of being conscious of something in the barest sense of being directed toward it, of registering or being responsive to it, tracking the sense of responsive consciousness. (See Meyers 2010, Heim 2013.)
I belabor this range of terms because it is important to remember that simply translating a conversation or a literature from a canonical Buddhist language into English, and using the term consciousness to translate all of the terms it most naturally translates, can mislead the unwary into thinking that there is a single sense of “consciousness” at work in the Buddhist literature. Or, with equal infidelity to the tradition, one might simply presume that the same set of distinctions among senses of consciousness in English is replicated in the relevant Buddhist language or literature. We must be alert to the fact that the distinctions encoded at the outset in language serve in part to define the conceptual landscape, and that the Buddhist and Western conceptual landscapes may not be entirely commensurable.
When consciousness is identified as one of the skandhas, it is, as we have noted, vijñāna that is meant. But one of the questions of importance in Buddhist philosophy of mind, as we saw in the previous chapter, is what constitutes the basis of designation of the self. In a Buddhist framework, as we have seen, there is no self. But this is not, of course, a simple matter. After all, from a Buddhist perspective, there is a persistent illusion of self to be explained as well as a reasonable convention of individuation permitting pronominal and nominal reference, ascription of agency and other relevant personal conventions. We must therefore ask what the role of vijñāna is in the propagation of the illusion of self as well as what its role is in the establishment of these conventions.
It may seem more plausible to think that our conventions of identification that allow us to designate ourselves and others, to re-identify persons over time, to ground morality and so forth—all that constitute persons in the everyday world—take ordinary vijñāna, or the aggregate of conscious cognitive processes, as the basis of designation. That is, one might think that even though this skandha is not the illusory self against which so many Buddhist arguments and analyses are ranged, it is the conventional self left untouched by those analyses.
But that has to be wrong. After all, especially in the identification of others, but also in reference to ourselves, our bodies (rūpa), personalities (saṃskāra) and feelings (vedanā) play significant roles as well. The whole point of skandha theory in the first place is that the basis of designation of self even if it is intrapersonal, is not a single thing but a complex, interwoven set of continua. This insight raises the possibility that much of who we are is unconscious and so the possibility that our focus on the conscious in our self-conception may be unwarranted.
There is another reason to think that from any Buddhist point of view, to take the skandha of vijñāna to constitute a self, or even to be the referent of a name or personal pronoun, would be—although tempting—a serious error. The skandha cannot be a genuine self because of its constant change and because of its dependence on the other skandhas. Most importantly, when we posit a self in the relevant sense, we posit it not as vijñāna, but rather as that which has vijñāna, as a subject more primordial than consciousness itself, modally independent of any particular skandha. As we pointed out in the previous chapter, the same consideration that shows that we don’t identify ourselves with our bodies—the imaginability of a body transplant—applies to our minds.9 Just as I can wish that I had the body of Usain Bolt, I can wish to have the mind of Stephen Hawking.
Yogācāra Buddhism, as noted in the previous chapter, does suggest a way in which a particular level of consciousness could be identified with the self, namely the alaya-vijñāna, or foundation consciousness. On certain Yogācāra views, this is the basis of identity over time and the unified ground of all other conscious states. It is the pre-reflective ground of experience.10 For this reason, in part, Mādhyamika philosophers such as Candrakīrti in India and Tsongkhapa in Tibet criticize this view as a smuggling of the self back into the basement after hard analysis has kicked it out of the front door. Yogācāra philosophers typically reply that the foundation consciousness is not a self in the pernicious sense—it is not an unchanging, substantial ground of experience; it is not that which we imagine ourselves to be in the act of self-grasping or in the narratives we tell about ourselves and the world; it is not even phenomenologically accessible. The self that is the object of negation of Buddhist analysis is a self like that. If this is right, though, then even if the construct of the foundation consciousness made sense, it would not constitute an identification of consciousness of any kind with self. Buddhist reflection on consciousness also suggests that we are best not focusing on consciousness—in any sense—as a singular thing. Even the skandha of vijñāna decomposes on most Buddhist analyses into a number of different faculties. For each of the six perceptual faculties (the five external senses and the introspective faculty) there is a corresponding faculty of sensory consciousness.
On a standard Buddhist model, perception first involves physical contact of the organ with the sensory object, then the registration in the sense faculty of the perceptual object, and only finally the consciousness of the object. But that consciousness is particular to the relevant sensory faculty. Visual consciousness is a different faculty from auditory consciousness. The introspective consciousness, on this picture, unifies the deliverances of the sensory consciousnesses as it takes them as its proper object. (See the discussion of Śāntarakṣita on the unity of consciousness in the previous chapter.) But importantly, consciousness is not unitary; it is present in each sense faculty prior to the action of the sixth sense.
If we accept the alaya-vijñāna in addition to these so-called “evolving consciousnesses” as a more fundamental level of consciousness, things get even more complex. This level of consciousness, which we might characterize as pure subjectivity, is regarded by the Yogācāra as the unifying principle that gives rise to the other forms of consciousness and in which the consequences of actions and events are stored, with the potential to give rise to future episodes of consciousness. But it is not introspectible and is not the seat of the subjectivity of any particular conscious episode, but rather stands as a transcendental condition on any awareness being conscious at all. It is more like the horizon provided by subjective consciousness, not something of which we are aware, but that in virtue of which we are aware.
On any Buddhist account, then, consciousness is neither singular, nor a monadic property, nor something that is simply present or not. It is a many-layered set of phenomena that are principally relational. Leaving the foundation consciousness aside for the moment, to be conscious is to be conscious of something. Consciousness is always a relation between a sensory faculty and its object. To understand consciousness is to understand that complex relation, the processes and conditions that make it possible, and those it in turn enables.
Even the foundation consciousness is, according to those who posit it, in ordinary circumstances relational in character.11 It is described as a kind of consciousness because it is the metaphysical ground of the relation to the evolving consciousnesses as objects of introspection. It can be thought of as the possibility of taking those forms of consciousness as one’s own.
David Chalmers lists a number of conscious phenomena that pose what he calls “easy problems” (2010, 3–13):
• The ability to discriminate, categorize and react to environmental stimuli
• The integration of information by a cognitive system
• The reportability of mental states
• The focus of attention
• The deliberate control of behavior
• The difference between wakefulness and sleep. (4)
These correspond, at least roughly, and probably exactly, to the second of the senses of “consciousness” scouted at the beginning of this chapter. Explaining them, Chalmers argues, is easy, because they are functional capacities that we can understand as being subserved by physical mechanisms. They contrast with the sense of “consciousness” in which the hard problem is posed:
The really hard problem of consciousness is the problem of experience. When we think and perceive, there is a whirl of information processing, but there is also a subjective aspect. As Nagel (1974) has put it there is something that it is like to be a conscious organism. This subjective aspect is experience. When we see, for instance, we experience visual sensations: the felt quality of redness, the experience of dark and light, the quality of depth in the visual field. Other experiences go along with perception in different modalities: the sound of a clarinet, the smell of mothballs. …What unites all of these states is that there is something that it is like to be in them. All of them are states of experience. (Chalmers 2010, 5)
The “hard problem”—explaining how experience is possible in a physical organism—Chalmers argues, is the problem of explaining consciousness in this sense. Seeing the problem as “hard” requires us to acknowledge two things: first, that the sense of experience—of there being something that it is like to be in a state—makes clear sense. Second, that it is possible to do all of the “easy to explain” things, and yet for it to be possible to differ from those with experience just in that one has no experience, the possibility of zombies.
We will come back to zombies below. But let us first consider the problem of understanding phenomenal experience, that the having of which is supposed to be like something. Of course there are many different theories on offer regarding the nature of such experience. But once again, Chalmers’s account can be taken as paradigmatic for present purposes. Phenomenal experience, according to those who deploy this notion (whether perceptual or not)12 involves the formation of phenomenal beliefs, which in turn involve the deployment of phenomenal concepts, concepts not of properties, but of the supposed inner correlates of those properties. Chalmers puts the point this way:
Phenomenal beliefs always involve phenomenal concepts, concepts of the phenomenal character of experience. When one believes that one is having a red experience, one deploys a phenomenal concept of a red experience. The most important phenomenal concepts are those we acquire directly from having experiences with that sort of phenomenal character. For example, when one first learns what it is like to experience an orgasm, one acquires a phenomenal concept of the experience of orgasm. (251)
…
I look at a red apple and visually experience its color. This experience instantiates a phenomenal quality R, which we might call phenomenal redness. It is natural to say that I am having a red experience even though experiences are not red in the same sense in which apples are red. Phenomenal redness (a property of experiences or subjects of experience) is a different property from external redness (a property of external objects), but both are respectable properties in their own right. (254)
Perhaps the most important sentence to note here is the last. If we are to take phenomenal consciousness as a special kind of consciousness, not simply the awareness of properties in the external world, but awareness of a peculiar family of inner properties, there must be a difference between these phenomenal properties and the properties of external objects. This distinction, and the consequent positing of consciousness as a special singular phenomenon characterized by properties of the second type, demanding a special theory, is driven by the idea that there is nothing “that it is like” merely to see a patch of blue; a non-conscious zombie might see patches of blue. But there is something “that it is like” to experience seeing a patch of blue. This involves special inner states that are blue in a different sense, phenomenally blue.
Now, the idea that inner states have such properties as blueness, redness, or other such properites, albeit in some special sense, has legs in Western philosophy, and has properly been subjected to critique before. (See Sellars 1992a, b for an elegant discussion.) Indeed, many contemporary theorists of phenomenal properties argue that such properties as phenomenal blueness are only contingently related to actual blueness, and that terms such as phenomenal blueness are as semantically distant from such terms as blueness as the term blunderbuss is from buss. On this account, phenomenal properties are simply those of our inner states that are causally connected to typical causes of the properties whose names sound like them. I don’t want to take us too far into this terrain, but suffice it to say that even if these accounts avoid the Sellarsian critique, they fall from that frying pan into the Wittgensteinian fire of critiques both of private language and of the very idea of special introspectible properties of inner processes. Now, these critiques may not be decisive, but they are weighty, and indeed, as Thurman (1980) has argued, both Candrakīrti and Tsongkhapa anticipate those critiques in precisely this context.
This thought—and the debate it occasions—has ancestry in the Indian and Tibetan Buddhist traditions. Dignāga and his followers in the Yogācāra tradition argued—against the earlier Vaibhāśika school’s direct realist theory of perception—that perception must be mediated by a representation (ākāra/rnam pa [nampa]) and that this representation is the intentional object (artha/don [dön]) of a perceptual state. The perception of a patch of blue, on such a model, is the immediate apprehension of a representation that is blue. Dignāga puts the point this way: “What we call an intentional object is what a cognition ascertains to be the entity itself. This is because its representation arises. …An object produced by self-presenting awareness is understood as a percept” (Autocommentary to Examination of the Percept commentary on verse 1). Perception on Dignāga’s model is always the immediate awareness of a representation; the qualities perceived are qualities not of external objects—indeed in this text he argues that they cannot be—but rather of internal representations.13 His 18th-century Tibetan commentator Gungthang (Gun thang dkon mchog bstan pa’i sgron me) comments on this as follows:
Suppose someone asks, “…[H]ow could [a representation] appear?” The representation does not appear as it exists. For example, when an image of the moon appears in a mirror, spatiality also appears. The moon appears to be different from its action of reflecting. Although it is apprehended with an appearance of externality, the object is that which exists internally. (4)
Gungthang’s principal claim here is that according to Dignāga, even though we might think of distal objects as the objects of perception, the objects of immediate awareness are always representations. To the extent that distal objects figure at all in our consciousness, it is in virtue of the indirect relation they bear to us, mediated by these representations. On the other hand, Gungthang also emphasizes that the mode of appearance of these representations is itself deceptive: just as a reflection of the moon appears to be located in space, and not to be a mere process of reflection, our own representations appear to us to exist as internal objects that we observe, when in reality they are merely cognitive processes. Ngawang Dendar (Ngag dbang bstan dar), writing later in that century, amplifies the argument that perception engages immediately with representations and only immediately with their objects:
Regarding the way in which the Sautrāntikas and more advanced schools maintain that cognition is representational, Introduction to the Middle Way states, “Whatever the mind represents is its intentional object.” [VI:13] When a representation of a blue intentional object arises for a visual cognition apprehending blue, because it thereby apprehends blue, it is said that there is an apprehension and an apprehender of blue. Other than by way of subjective cognition becoming like its object, there is absolutely no way for cognition to apprehend its object.
For example, when a crystal is placed on a blue cloth, all facets of the crystal turn blue. In the same way, when a visual cognition apprehending blue sees blue, the color blue is transferred to the visual perspective and so the subjective visual cognition becomes just like the blue object. The visual cognition apprehending blue is referred to as “a visual cognition that arises as a representation of blue.” This is the criterion for the direct apprehension of blue. (227–228)
Dendar emphasizes here that the consequence of this view is that the qualities we perceive—those of which we are immediately aware—are qualities of appearances, not of external things. It is qualitative blue, the blueness of an experience that I perceive, not any property of any external blue object. We will see soon enough that this view comes in both for extension and for critique in the Buddhist tradition. Some of the arguments and doctrines we will encounter will anticipate some we find in contemporary Western debates about phenomenal properties and phenomenal concepts. Some are unique to the Buddhist tradition. But all are bound up in a Buddhist context with a more general worry about the nature of consciousness itself, and it will be useful to put aside worries about the nature of phenomenal properties themselves for now and to explore the questions in the philosophy of mind they engender. Many are also bound up with issues concerning the epistemology of the inner.
Indian and Tibetan Buddhist philosophers share with their contemporary Western colleagues a concern for the structure of self-knowledge, and roughly the same range of positions that emerge in the West are found in Indian Buddhist debates. Moreover, just as we see in contemporary Western debates, the positions regarding the structure of self-knowledge do double duty as models of consciousness itself. We hence see reflexive models of self-consciousness and self-knowledge, higher-order thought models, higher-order perception models and self-luminosity models.
Reflexive models take consciousness always to involve two aspects: a directedness toward the manifest object of consciousness and a self-directed aspect that makes possible both knowledge of the conscious state itself and its status as a conscious, as opposed to an unconscious state. Higher order thought and higher order perceptual models take states to be conscious in virtue of other states that take them—either cognitively or perceptually—as their objects, and which enable our epistemic access to them. Self-luminosity models take conscious states to be primitively conscious and self-manifesting without being intentionally directed upon themselves, and without being the intentional objects of other states, revealing their subjectivity as a primitive property.
The idea that perception is, au fond, an engagement with phenomenal properties, Dignāga and his followers realized, issues in an intriguing dilemma. Given that it is the perceptual engagement with phenomenal properties that is meant to be constitutive of consciousness, the apprehension of these phenomenal properties must either itself have phenomenal properties apprehended by a higher-order state in order for it to be conscious, or it must be reflexively conscious—conscious of itself at the same time that it is conscious of its perceptual object (svasaṃvedena/svasaṃvittiḥ/rang rig). Otherwise, phenomenal properties themselves would not be apprehended, and would constitute at best a mysterious fifth wheel in the explanation of perception. The latter position (anticipating that of Husserl, a position inspiring the work of Coseru 2012 and Zahavi 2004, 2005) is adopted by all Indian Yogācāra thinkers and Śāntarakṣita, and in Tibet by eminent Sakya scholars such as Gorampa and Sakya Chokden. The former position, prescient of that of Sellars and his followers, is adopted by most Indian Mādhyamikas and in Tibet by Geluk figures such as Tsongkhapa.
There are three principal arguments for the reflexivity of awareness in the Buddhist literature, two that seem unique to that tradition, and one that anticipates Husserl’s argument for reflexivity pretty directly. Dignāga advances what has become known as the “memory argument” for reflexivity. He argues that when I remember yesterday’s sunshine I remember myself seeing yesterday’s sunshine. The memory of the sunshine and the memory of my seeing it are one and the same. Moreover, he argues, one cannot remember something of which one was not conscious. Therefore, since the memory of the sunshine and my seeing the sunshine are identical, and since my memory of the sunshine presupposes that I was aware of the sunshine, my memory of my apprehension of the sunshine presupposes my consciousness of that apprehension. And, since the memory of the two phenomena is identical, the two phenomena must be identical. Hence my apprehension of the sunshine was at the same time my awareness of the apprehension of the sunshine, and perceptual consciousness is therefore reflexive.
Candrakīrti represents Dignāga’s argument as follows:
Even those who do not [mount the regress argument] will inevitably have to accept reflexive awareness. This is because otherwise, although when a memory arises at a later time of the form “…was seen,” an object would thereby be remembered, when a thought of the form “I saw…” occurs, the object that was seen would not thereby be remembered. Why is this, you might ask? Since the memory is to be experienced subjectively, and since consciousness is not perceived, it can’t be remembered. It follows that if there were no reflexive awareness, nothing could experience itself. And it makes no sense to say that it is experienced by another consciousness. Why is this? Because if it had to be experienced by another consciousness, an infinite regress would ensue. (Candrakīrti 2012, 166–167)
There are two arguments here. The first is the memory argument. My memory of yesterday’s sunshine could come to me in the form, “I saw the sunshine yesterday” or in the form, “Yesterday was a sunny day.” These seem to be the same memory, but if awareness were not reflexive—that is, if my memory of the sunshine were not identical to my memory of seeing the sunshine—the first would simply be a memory of a subjective state, and not of its object, and the second a memory of the object, and not of the experience, and they would be distinct and unrelated. They are not; so awareness must be reflexive.
At the end, the argument is reinforced by the second argument, a regress argument also proffered by Dignāga and his commentator Dharmakīrti. They argue that every conscious state must be known to its subject. Otherwise, it would not be conscious. If it is known, it must be known either reflexively, or by some higher-order state. If a cognitive state could be known only by a higher-order state, then either that higher-order state would be conscious or unconscious. If the former, we are off on a regress, requiring an infinitude of higher-order states in order for any state to be conscious; if the latter, there is no explanation of how an unconscious state could make another state conscious. Therefore, every conscious state must be reflexively aware. Dignāga offers a related argument for the necessity of reflexive awareness, one that emphasizes the need to be aware that one is aware in order to be aware, which, coupled with the regress argument, entails that any genuine awareness must involve reflexive awareness.
Although reflexive action is not conventionally presented as the object of the sense faculties, and such things as objects and attachments are presented without reflexive thought, meditators and lamas without exception whenever they perceive or think of an object maintain that there is reflexive awareness. …
Superficial perception is obscured for it is involved with conceptual thought. It is the effect of epistemic engagement, and although it is said to be authoritative, it is not actually discriminative. However, reflexive awareness is its effect. As a consequence of that, its object is ascertained, for reflexive awareness is authoritative with respect to the appearance of its object, and thus one says that the object is epistemically engaged.
Therefore, the appearance is the apprehended object. Epistemic authority is consequent on that. Therefore, apprehending consciousness arises from these three distinct things: the consciousness of the object; the consciousness of that consciousness; retaining these two modes of consciousness in memory. Through these two considerations we prove that awareness is reflexive. If a second consciousness was needed to perceive the first, there would be an infinite regress, and so an infinite number of distinct objects, and this is not what is observed. (Encyclopedia of Epistemology [Pramāṇasamucccaya] sDe dge 15b–16a)
This view is, of course, well known in the contemporary literature on consciousness as well, and is championed most notably by Kriegel (2009). Kriegel puts it this way: “[M]y conscious experience of a blue sky is the conscious experience it is in virtue of its bluishness, but it is a conscious experience at all in virtue of its for-me-ness.” (1) Kriegel clarifies this notion of “for-me-ness,” which he calls subjective character as follows:
[T]o say that my experience has subjective character is to point to a certain awareness I have of my experience. Conscious experiences are not states that we may host, as it were, unawares. …A mental state of which one is completely unaware is not a conscious experience. In this sense, my conscious experience is not only in me, it is also for me. (8)
Continuing in his recapitulation of Dignāga, Kriegel sums up his own argument for reflexive awareness as follows:
1. C is conscious in virtue of S’s being suitably aware of C.
2. For S to be suitably aware of C is for C to be suitably represented by S; therefore,
3. C is conscious in virtue of being suitably represented by S; therefore,
4. C is conscious in virtue of being suitably represented;
5. it is not the case that C is conscious of being unconsciously represented; and
6. it is not the case that C is conscious in virtue of being consciously represented by a distinct state; therefore,
7. it is not the case that C is conscious in virtue of being represented by a distinct state; therefore,
8. C is conscious of being represented by itself; that is,
9. C is conscious in virtue of suitably representing itself. (21)
The first premise is a recapitulation of Dignāga’s opening move. He then recapitulates Dignāga’s and Dharmakīrti’s regress argument.14
One comment on these arguments is in order before turning to some Buddhist critiques of them. These are not arguments for a broad pre-thetic awareness of one’s own subjectivity of the kind advanced by Husserl, and much later defended by Zahavi (2005) and Coseru (2012), as well as the late 19th-century to early 20th-century Tibetan philosopher, Mipham, who seems to endorse that broader reading of reflexivity. We will return to this below. Instead, these philosophers defend the claim that every conscious mental episode, in virtue of being a conscious episode, has a dual intentional content: on the one hand it is directed upon its nominal intentional object; on the other hand it is directed back upon itself, a view more like those of Gennaro (2011) and Kriegel (2009), views advanced on somewhat different grounds. This view instead is aimed at explaining how our responsiveness to the phenomena we apprehend can be conscious, and the explanation is that it is the reflexivity that makes this possible.
There are a lot of problems with these argument (as noted by contemporary proponents of Higher Order Thought [HOT] and Higher Order Perception [HOP] theories of consciousness), and its faults were pointed out by a number of Buddhist scholars. For one thing, there is a problem with the first premise. Given the enormous variety of prima facie conscious states, it is simply not true that all are known by their subjects. Much of our consciousness is opaque to us much of the time. And there are so many forms of consciousness to consider. In the Buddhist tradition, this is part of what motivates meditative practice: one thereby becomes increasingly aware of aspects of one’s consciousness of which one is generally unaware. We do not know our own minds flawlessly, but only through fallible perceptual and inferential means (Tsongkhapa 1991a). In modern cognitive science this is simply the familiar phenomenon of the opacity of consciousness to introspection and the prevalence of illusions of consciousness (Noë 2002, Schwitzgebel 2011 and Gennaro 2011). We will return below to the importance of these phenomena for understanding a Buddhist contribution to contemporary debates.
Second, the regress-generating premise fails to generate the requisite regress. Even if a conscious state must be known by another state, it does not follow that that higher-order state must itself be conscious. Just as my perception of a blue sky does not need to be perceived in order to be a perception, my consciousness of that perception does not have to be consciously observed in order to be conscious, even if it must be the possible object of some higher order state (See Rosenthal 2005b, Gennaro 2011). At best the proponent of the argument might get the premise that it must be possible to become consciously aware of any conscious state through a higher-order state. That generates a regress, but a benign regress. It just means that I can iterate higher order states of consciousness, not that I must, just as the fact that every natural number has a successor means that I always can count higher, not that I must.
Candrakīrti (2012) and Śāntideva (1997) each criticize the memory argument. Candrakīrti points out that the account of memory at work in the memory argument is problematic.
[168] Now, suppose that this is supposed to concern that which exists substantially. Then, since there is neither arising either from self or other, memory is not possible at all. So how could a non-existent memory demonstrate the existence of a non-existent reflexive awareness? On the other hand, suppose it is supposed to exist in terms of mundane conventions. Even in that case it is impossible to use memory as a reason to prove the existence of reflexive awareness.
Why is this, you might ask? Suppose that, just as in the case of fire, you had to establish the existence of reflexive awareness by appeal to another cognition. If this were the case, just as the existence of fire is proven by smoke that is observed at a later time, it would have to be proven to exist by a later memory. But since the cogency of reflexive awareness hasn’t been proven, how could a memory that itself is not reflexively aware prove the existence of reflexive awareness? In the same way, just by seeing water you can’t prove the existence of a water-producing crystal; [169] nor can you prove the existence of a fire-producing crystal just by seeing fire.
After all, water comes about without crystals, as a result of such things as rain, and fire comes about without crystals from such things as matches being struck. In the same way, memory can occur without reflexive awareness. Therefore, memory can never serve as a premise to prove the reflexivity of awareness. Since it has not already been proven to exist, in order to prove the existence of reflexive awareness, you would need a memory that was already proven to be reflexive, and this has not been proven to exist. (2012, 168–169)
This is, of course, not an argument against the reflexivity of awareness, but a rebuttal of the memory argument, an argument that ends up looming large—perhaps too large—in the Indian context, simply because it was one of the first arguments advanced for the reflexivity position. Subsequent theorists all feel the obligation to take it on. First, Candrakīrti notes that the domain of this debate is conventional reality, not ultimate reality; this debate is not about fundamental ontology but about psychology and the philosophy of mind. Candrakīrti then notes that the argument can’t proceed just by noting that reflexivity would explain the effect in question, for all kinds of things could explain the phenomenon of memory. Moreover, if awareness is necessarily reflexive, then the memory to which one appeals should independently be taken to be reflexive. If that were the case, nothing else would need to be said; if it is not, the conclusion is false.
All that is necessary for memory, Candrakīrti argues, is that the cognition that counts as a memory is caused in the right way by a previous cognitive episode. So I might experience today’s sunshine, fail to be reflexively aware that I am experiencing it, and tomorrow recall the sunshine in an episode caused (in the right way) by today’s perception, but in neither case thematize my own subjectivity in the matter.
Śāntideva goes further, arguing that the illusion of reflexivity arises only in virtue of the inferential processes used in memory reconstruction, using as an analogy the fanciful example of a bear bitten by a rat while hibernating, who wakes up to discover a painful, infected wound. The bear, Śāntideva says, knows on the basis of the painful wound that he suffered a rat bite, even though at the time of the bite he was not aware of the bite. The point of the analogy is this: (1) One can come to be aware of a previous event through causal sequelae of that event even though one was not aware of that event at the time of its occurrence; (2) Those sequelae can induce a cognitive state intentionally directed at that previous occurrence even if one was not aware of that occurrence at the time. And we might note that contemporary cognitive science, with its reconstructive, rather than storage-and-retrieval models of memory, is on the side of Candrakīrti and Śāntideva in this debate.
It hence follows not only that one could come to remember an event of which one was not explicitly aware at the time of its occurrence, but also, and most importantly, that one could have a memory with the intentional structure “I experienced…” even if one was not aware of experiencing that event when it occurred, despite being aware of the event. Suppose, to take a more contemporary event, one was engrossed in participation in a sporting event, diving to catch a goal in a game of Ultimate. At the time of the event, one might be aware of the flight of the disk, but not aware of one’s awareness of it; that is, the awareness of the disk is not accompanied by self-awareness. Nonetheless, after the game, one reports, “I remember diving for the disk, hoping I would catch it. …” This later memory is caused by the previous event, and although the memory presents itself as reflexive, the event itself need not have been one involving any reflexivity. (See Garfield 2006b for an extended discussion of these arguments.)
All of this is grounded in Candrakīrti’s more general concern with two issues—the importance of interdependence and the absence of any intrinsic nature in anything on the one hand, and the role of linguistic convention, and hence intersubjectivity, in constituting meaning and knowledge on the other (Thurman 1980). For Candrakīrti, reflexivity as a mechanism for self-knowledge would render cognitive states intrinsically known, independent of any other cognitive apparatus, and so would violate interdependence, and would amount to assigning cognitive states a kind of essence. And for the mechanism of their knowledge to be essentially inner would violate the principle that the epistemic instruments—perception, inference, testimony and analogy—are constituted as justificatory intersubjectively, by convention. The so-called knowledge that would be constituted by reflexivity, on this account, would not be knowledge in any sense at all, given its essential privacy and its isolation from the rule-governed, public epistemic practices that confer warrant on our beliefs. This view, once again, as do so many Madhyamaka analyses, nicely anticipates both Sellars’s attack on the givenness of mental content and Wittgenstein’s emphasis on the social dimensions of knowledge in On Certainty.
There is a final issue raised by Candrakīrti, and this may be the most profound, anticipating the more contemporary “rock” problem raised by Stubenberg (1998) for Western Higher Order theories, but mobilizing it against First Order theories that appeal to reflexivity. Stubenberg points out that if having a distinct conscious state directed on an object were sufficient to render it conscious, then my thinking about a rock should render the rock conscious; if it cannot, there is no reason to think that a higher-order state directed on a state not already conscious on its own could make that state conscious (185–195).
So, if there is a regress problem, reflexivity can’t be the answer to it. For the regress problem suggests that a state that is merely perceptually directed on an object but not the object of another state would not itself be conscious. Grant that claim for the sake of argument. Then consider a reflexively aware perceptual state. If that state were not reflexive, it would not be conscious. But to add reflexivity is then only to add the already unconscious awareness of an unconscious state. It is hard to see how that could issue in consciousness if an unconscious higher order state could not. Candrakīrti replies to the regress in a Sanskrit grammatical register with a clear point:
If one says that a conscious state knows itself then the nature of the process of knowledge would be as follows: Since there must be an agent of that action of knowledge, the account of that action would make no sense, since it would follow absurdly that the agent, action and object would be identical. Such a unity is never observed. For instance, the woodcutter, the wood and the cutting are not identical. For this reason we can see that there is no reflexive awareness, and that we do not become aware of a cognitive state through that state itself. As the Entry into Lanka sūtra says,
A sword
Cannot cut its own blade,
And a finger cannot cut itself.
Reflexive awareness is exactly the same. (2012, 171–172)
Now, this argument is cast in a classical grammatical mode, arguing that agent, action and object are necessarily distinct on the grounds of case, and in that form, the argument is hardly compelling. As many proponents of reflexive awareness in the Indian and Tibetan traditions point out, for instance, lamps do reveal themselves, and to the extent that we think of awareness as a kind of revelation of its objects—a very frequent image in Indian philosophy of mind and encapsulated in the term svaprakaśa (self-illumination)—that is not a bad counterexample. But Candrakīrti also makes a deeper point. If it is in virtue of reflexivity that a state is conscious, the state that is the object of the reflexive awareness must already be conscious, since it is identical with the awareness that makes it conscious. But if that is so, reflexive awareness is awareness of an already conscious state, and so cannot be that which makes it conscious. The very idea of reflexivity thus presupposes the very property it is meant to explain. An unconscious state directed on itself would hardly make itself conscious.
Candrakīrti’s critique may have been what led the 9th-century Buddhist philosopher Śāntarakṣita to maintain that reflexivity is simply definitional of subjective consciousness as such, that it is what differentiates sentience from insentience. Here is what he has in mind. Both sentient and insentient matter is subject to effects, and indeed, we might add, many of those are information-rich effects. My footprints in the sand are indicative of the shape of my foot just as is my visual perception of my foot. So no merely relational account of the interaction between my visual faculty, my eye and the foot will distinguish between the interaction between my foot and the sand in a way that will explain why the one is conscious and the other is not. There must therefore be some intrinsic character of consciousness itself, namely its reflexivity, the fact that it presents both the object and our subjectivity of that object, that makes sentience sentience. Here is how Śāntarakṣita puts the point in his Ornament of the Middle Way:
Consciousness arises as diametrically opposed
In nature to insentient matter.
Its nature as non-insentient
Just is the reflexivity of awareness. [16]
Since it makes no sense for that which is unitary and partless
To have a threefold nature,
The reflexivity of awareness
Does not have an agent-action-structure. [17] (verses in Blumenthal 2004)
Śāntarakṣita eschews any talk of self-representation or of the explanatory force of reflexivity in favor of a brute definitional move coupled with phenomenological reflection. In (16) he simply claims that reflexivity—awareness of oneself as a subject—is constitutive of consciousness, and in (17), conceding Candrakīrti’s critique, he asserts that reflexivity is non-representational and pre-thetic. This view anticipates those of Husserl and his followers, such as Zahavi (2005), Thompson (2007) and Coseru (2012), suggesting a primordial pre-thetic character to subjectivity. As we will see, it also comes in for trenchant critique by Tsongkhapa, following Candrakīrti.
Just as in the case of the theory of consciousness, debates about reflexivity play a large role in Buddhist debates about self-knowledge. Dignāga introduces this issue in his account of pramāṇa, or epistemic warrant, and the nature of epistemic instruments in his Encyclopedia of Epistemology. There he argues that reflexive awareness is the mechanism by means of which we know our own inner states. Since this is a kind of perception, on Dignāga’s view it gives us immediate, non-conceptualized knowledge of our inner episodes, and this, as we saw above, is why he thinks that perception, despite its ordinary fallibility, is, in the end, authoritative. While this might appear to be a classical Buddhist version of a contemporary higher-order perception model of self-knowledge (HOP) (Armstrong 1968, Lycan 1997), it is not. For the point of a reflexive model of awareness in Dignāga’s framework—and in subsequent Buddhist articulations of this doctrine—is that each mental state is itself reflexive, requiring no higher-order state to make itself known.
On the other hand, there are homologies between this reflexive model of self-knowledge and contemporary HOP theories. In both accounts, self-knowledge is taken to be direct, unmediated by theory, and hence available even to those who lack the capacity for sophisticated theory, such as small children and animals. Both views recognize the spontaneous, non-reflective nature of self-report, and the fact that we seem not to theorize, when we know our own minds. A great virtue—or a great vice—of this account, depending on one’s perspective, is the fact that it represents self-knowledge as very different in kind from knowledge of other minds. Whereas we know other minds only by inference, we know ourselves, on the other hand, directly. This kind of view explains both the apparent asymmetry with respect to mode of access and epistemic security between self- and other-knowledge as well as the apparent symmetry between this kind of knowledge and ordinary perceptual knowledge.
But there are also important differences between Dignāga’s theory and HOP theories. For Dignāga, reflexivity is not limited to the case of explicit introspection: it is also implicated in the process of external perception becoming conscious. Reflexive awareness generating self-knowledge is indeed the means by which we know ourselves, but this same reflexive awareness is implicated in every conscious state, and explains not only self-knowledge but consciousness itself. Once again, this may constitute a theoretical virtue or a vice depending on one’s perspective. On the one hand, to the extent that one thinks that there are no unknown—or at least no unknowable—conscious states, this model explains that fact; on the other hand, to the extent that one thinks that there are, this model might make the wrong prediction regarding our epistemic access to our own minds. Moreover, because reflexive awareness is not contaminated by conceptual superimposition, Dignāga is committed to the view that our awareness of our own mental states, at least until the moment that we conceptualize it in verbal judgment, is veridical. (See Yao 2005 for more detail.) He thus provides the platform for a strong doctrine of the transparency of mind to itself, although not one committed to a kind of higher-order model of cognitive access.
Now Dignāga’s account is a starting point, and perhaps most interesting for its historical role as the beginning of a tradition of Buddhist reflection on the epistemology of the inner, a topic which, while central to the Buddhist soteriological project, receives surprisingly little attention prior to Dignāga’s reflections. Things get more interesting as we move forward in the history of that reflection.
Dignāga’s commentator and successor in the development of Buddhist epistemology, Dharmakīrti, develops Dignāga’s account of reflexivity and self-knowledge in terms of self-luminosity. He argues that all consciousness is necessarily dual in aspect, incorporating both an awareness of the object of consciousness, whether external or internal, and an awareness of the act of consciousness as such. Mipham endorses this view in his commentary on Śāntarakṣita (Mipham 2004, 273).
On this model, we know our own mental states simply in virtue of being in them, for it is of the very nature of a cognitive state to present itself as the state that it is in its manifestation. While this view is very close to Dignāga’s, and is presented as a gloss on that earlier view, it is slightly different. On Dignāga’s account, every mental state takes itself as object, providing a representation of itself, as well as of its object, to consciousness; on Dharmakīrti’s view, each state is instead internally complex, with its subjective aspect self-presenting as subjectivity, not as the object of any state, including itself. While its primary intentional content is its apparent object, it takes its own subjectivity as a secondary intentional object. This is a pre-thetic awareness of subjectivity itself, by means of which we know ourselves and our states immediately.
A similar self-luminosity position is also well represented in the West. As noted above, it has its modern origins in Husserl’s account of consciousness (Zahavi 2005, Kriegel 2009). According to this position, our cognitive states are known immediately because it is of their very nature to present themselves to consciousness as the states they are. My belief that 7 + 5 = 12 is, on this view, at the same time a cognitive state directed on an arithmetic fact and a state which presents myself as subject. My knowledge that I have that belief is part and parcel of the belief itself. Kriegel, Thompson and Zahavi, each in his own way, defend just such a view.
To take a belief to be self-luminous is to take it that to be in a cognitive or affective state (or at least to be in a conscious cognitive or affective state, with the appropriate sense of conscious to be filled in, or perhaps an occurrent state) is to know that state immediately, not only without the need for any higher-order thought or perception of that state, but also without any perception of that state even by itself. This is possible, on this view, because conscious states are self-illuminating, requiring no second state in order to be known. Just as we do not need a second light bulb to see a glowing light bulb, on this view, we do not need a second state to allow us to become aware of a first state. We may need a light bulb to illuminate that which is not self-luminous, of course, like the furniture in the room, and so we may need conscious states to mediate our awareness of the world around us, but to require a higher-order state, on this view, is to invite a vicious regress. On this account, just as to be self-luminous in this sense is constitutive of and criterial for consciousness, since cognitive states are conscious states, self-luminosity is criterial for cognitive states as well (Zahavi 2005, Kriegel 2009).15
Debates about self-knowledge and reflexivity extend into Tibet, with its rich tradition of analytic epistemology. The 13th-century Tibetan scholar Sakya Paṇḍita argued that self-knowledge is achieved through reflexive awareness, relying heavily on Dharmakīrti’s analysis of cognitive states as self-luminous. He argues, however, that since in any perceptual or cognitive apprehension, subject and object must be distinct, in the case of reflexive awareness, since there is no such distinctness, there is no object of that awareness. Such knowledge, he argues, is in fact objectless. Now, of course this raises the question of just in what sense reflexive awareness counts as knowledge at all. Dreyfus (1997, 402) compares this account to that of European phenomenologists like Sartre, who argue that all consciousness presents subjectivity pre-thematically as a precondition of any awareness. On the other hand, while this might vouchsafe the very general self-knowledge that one is aware, it would hardly take one to knowledge of the contents of one’s own mind, and may in the end be too thin to count as any kind of knowledge worth having, in addition to being so odd, that it is hard to see it as knowledge in the sense that anything else we count as knowledge is knowledge.
This worry about both the sui generis nature and the thinness of reflexivity was very much in the mind of Sakya Paṇḍita’s 15th-century Sakya successor Gorampa. Gorampa argues from the fact that reflexive awareness is the mechanism of self-knowledge that it must therefore be intentional, and therefore must have an object of knowledge (Dreyfus 1997, 403). Now, Gorampa’s own view is that all knowledge is mediated by representation, and this raises further difficulties for the account that reflexivity is the basis of self-knowledge. For if we grant the necessity of the mediation by representation of all knowledge, then self-knowledge must be mediated by representation as well, and if this is so, a regress looms. If we need a representation in order to know our own inner state, then presumably to know this representation a further representation is needed, and so on.
Gorampa is aware of this potential problem, and cuts the regress off by following Dharmakīrti in another respect, that is, in distinguishing two aspects of any cognitive state, a subjective and objective aspect.
The representation is the object of apprehension. To say that the instrument of knowledge, its effect and the apprehension are three different things therefore makes no sense. With regard to the object, consciousness and the consciousness of that consciousness appear to consciousness in a twofold manner, and this how it is in memory at a later time as well. Reflexive awareness is this twofold appearance. If there were no reflexive awareness there would be no perception. And if a distinct consciousness was required to perceive it there would be an infinite regress. And memory would always be of another object. (Elucidation of Epistemology [Tsad ma rnam ‘grel] 300). (See also an extended treatment in Illumination of the Ascertained Object [Ngas don rab gsal] 577–582)
A reflexive awareness, he argues, takes, in its subjective aspect, the objective aspect of the state as its object. This may cut of the regress, but once again there is a cost. Subjectivity now takes objectivity as an object, but to the degree that this is all that is involved in reflexivity, it is a thin sense of self-knowledge, sounding more like a restatement of the fact that subjective states have objects.
Geluk scholars, following Tsongkhapa, take Nāgārjuna and Candrakīrti more as touchstones in this debate than they do Dharmakīrti. Tsongkhapa (1988) argues, following Candrakīrti, that reflexivity makes no sense in the first place.
Suppose someone asked, “If according to your view there is no assertion of reflexive awareness, how does memory occur?” According to mundane convention, the mind does not experience itself. But the previous state of consciousness perceives a previous object, and this is the cause of the effect, which is the later memory. (289)
Here Tsongkhapa follows Candrakīrti in asserting that the simplest explanation of memory is simply a causal process, requiring no special higher-order intentionality. He now turns to a possible objection. Does the denial of reflexivity undermine the possibility of ordinary introspective knowledge? He argues that it does not, but that our ordinary understanding of introspective knowledge is in terms of higher-order perception and thought (which he does not clearly distinguish here, but which can easily be in a Buddhist framework in terms of their respective objects—particulars vs. universals).
Suppose one thought as follows: Since it would be to deny that one experiences such things as pleasure and pain through the introspective consciousness, how could there be no reflexive awareness? We commit no such error, because the denial of reflexive awareness is consistent with the distinction between subject and object with respect to all cognitive states that are directed inward. …According to mundane nominal conventions as well, the experience of pleasure and pain occurs in this way. Since the perceiver and the perceived appear distinctly, there is no need to posit reflexive awareness as it is accepted in our opponents’ position. (Ibid., 297)
Tsongkhapa now comes to the heart of the issue. He takes as his target the position recently defended by Kriegel, a self-representational model of reflexive awareness: The only cogent model of reflexivity, Tsongkhapa notes, is representational. After all, reflexive awareness has to have an intentional structure, and its intentional content must be known via a representation. Hence every cognitive state must represent itself on this view. But once we grant that representational structure, Tsongkhapa says, we can now ask about the epistemic status of that reflexive state. If it contains itself as object, it must be completely authoritative. How could it be mistaken about its own nature? But if it is authoritative, we have the consequence that we would have omniscient access to our own minds.
Now Tsongkhapa, like many contemporary philosophers, takes this to constitute a reductio on reflexivity. Of course any proponent of strong privileged access to our own doxastic and perceptual states—and there are such people—would take this as a welcome consequence. Blumson (personal communication) argues for such a view, advancing two distinct arguments, which I consider here because I think that they express exactly the intuition against which Tsongkhapa is concerned to argue. First, Blumson urges that the Cogito provides a clear example: If I am thinking, I know immediately and irrefutably that I am thinking. Hence, I have infallible access in this case to my own mind. Blumson may be right about the premise, but it hardly establishes the conclusion he wants. For I know that I am thinking because of the logic of thinking, not because of the content of any thought. It does not follow, for instance, that if I am thinking that roses are red that I know that I am thinking that. And that is what the strong reflexivist position requires.
The second argument is meant to respond to this point, and appeals to what one might call the BB principle, that is, that if one believes that P one believes that one believes that P. Leave the regress issue aside—there are ways to finesse that by talking about potential but not actual regresses. The obvious response from Tsongkhapa’s perspective (and from that of much contemporary research in cognitive development—particularly that devoted to the acquisition of Theory of Mind) is simply that the principle is false. To believe something is a conceptually enriched act; to believe that one has a belief requires that one have the concept of belief, and many small children and infralingual animals don’t have it. And then there is the Freudian point, at the heart of Tsongkhapa’s concern: It also requires that one have a pretty good understanding of one’s own cognitive life, something that even many adults fail to have. On this perspective, self-knowledge is far from immediate, far from given. And while this might appear counterintuitive from certain other perspectives, it turns out to be surprisingly well-confirmed empirically. Here is how Tsongkhapa makes this point:
If any consciousness to which the object of that consciousness appears were also its own object, that consciousness would appear as a representation. If that consciousness were non-deceptive with respect to that, that mundane, non-deceptive consciousness just by being known as authoritative, would have to be authoritative. In that case, if the apparent object of knowledge were to be established by that consciousness, the subject would already have to have been. (Ibid., 298–299)
Knowledge, Tsongkhapa argues, is always a relationship between a subject and an object. If we know our own mental states, it is because of epistemic states that take our cognitive states as object, and moreover, states that are distinct from those that they know. Self-knowledge, on his view, if it is to mean anything, is an understanding of our own cognitive lives and perceptual experience achieved by the same mechanisms by means of which we know external events.
Moreover, the Geluk school is realist, not representationalist, about perception, taking perception to be a relation between a perceived object, a sensory system and a cognizer. Tsongkhapa’s demand that apperception and self-knowledge are constituted in the first instance by higher-order perception does not, therefore, engender a vicious regress, and at worst, only a regress of potential higher-order states, corresponding to potential higher-order states of knowledge. This position does raise another issue, though. On this analysis, the introspective power is mediated by the sixth sense faculty (the introspective faculty), epistemically aligned to the external sense faculties, as a perceptual state. But perception, at least for ordinary beings, is also epistemically complex and problematic on a Buddhist account. For while it may in principle be veridical, in practice it delivers us data that are infected by the imperfections of our sensory organs and our instinctive superimposition of conceptual categories—many or all of which may be intrinsically falsifying—on the deliverances of sensation. So, while higher-order perception may be a central part of the story, it may well be that any analysis of that perception must appeal to higher-order thought as well, and on this account, unlike those of proponents of a special self-revealing nature of first-order cognitive states, self-knowledge will always be fallible. That is, on this account the mind, like external objects, always is, at least to some degree, hidden to us.
To take self-knowledge to be mediated by HOT, as one must on Tsongkhapa’s view, is to see it as reflective, as theoretical. On this model we know ourselves in part by theorizing about ourselves, by inference to the best explanation. Our inner states are not perceptible, but our behavior is interpretable, and when we self-ascribe cognitive states, we engage in an act of interpretation, just as we do when we ascribe cognitive states to others in the absence of any ability to perceive them directly. This homogeneity of self-ascription and other-ascription is, depending on one’s perspective, a great virtue or a great vice of this account. On the one hand, it eliminates the mystery of a sui generis access to our own mental lives and explains our fallibility with respect to the mental. On the other hand, it suggests that creatures incapable of the requisite degree of reflective capacity, such as young children and animals, do not know their own minds at all. This is an important insight, for, as I have been emphasizing, it is easy to take for granted the view that our access to our own minds is immediate. That is probably not just a part of our Cartesian legacy; it seems to have an intuitive hold on many people, and in many traditions.
The Madhaymaka view defended by Tsongkhapa is hence not just common sense. That is not to say—as I have also been at pains to emphasize—that it has no counterparts in the contemporary Western literature. Rorty (1979) as well as Churchland (1979) make similar claims, though on different, more metaphysical grounds. It is interesting to contrast Tsongkhapa’s view with Lewis’s otherwise similar position (1995), as it is representative of this trend. Lewis argues that qualia are inconsistent with materialism and therefore should be rejected en route to a more general rejection of immediate self-knowledge. Fair enough, but if one is not antecendently materialist, as many contemporary qualiaphiles are not, this argument may seem patently question-begging, or at least ad hoc. For this reason, the Indo-Tibetan Buddhist tradition might be a welcome voice in this debate: it provides more principled reason for treating self-knowledge in the same way as knowledge of the external world, grounded in an analysis of perception, self-knowledge and subjectivity.
It might appear that there is something terribly flat-footed about any insistence such as that of Śāntarakṣita’s that nothing relational, nothing not already subjective, could ever differentiate conscious from unconscious phenomena. But it is a position remarkably prescient of new mysterians (those who regard the psychophysical relation as unknowable, either in principle or in practice) such as Stoljar (2006), Chalmers (2010) and McGinn (2004) and even of those who insist on some intrinsic feature of mental states that make them conscious, such as Searle (1997), Levine (2001) and Gennaro (2011). On the other hand, Śāntarakṣita raises an important issue. There is a burden of proof on the shoulders of the higher-order theorist. She must explain how some relations among subjects, their faculties and their objects constitute consciousness and why others do not. If she cannot do this, reflexivity looks like the only option, an argument taken up in a different register by Kriegel (2009).
Before turning to Madhyamaka responses to this challenge, let us spend a bit of time with contemporary higher-order theories. Peter Carruthers (1996) has offered the most extensively worked-out higher-order theory of consciousness. (There are many more, of course, most recently that of Gennaro [2011]. But to survey and address the contemporary literature on consciousness is well beyond the scope of this discussion. My aim is simply to show how Buddhist theory can inform that discussion.) Carruthers writes:
Consider routine activities, such as driving, walking, or washing up, which we can conduct with our conscious attention elsewhere. When driving home over a route I know well, for example, I will often pay no conscious heed to what I am doing on the road…I saw the vehicle double-parked at the side of the road, since I deftly turned the wheel to avoid it. Yet I was not conscious of seeing it, either at the time or later in memory. My perception of that vehicle was not a conscious one. (135)
…
Recall just how rich and detailed an experience can be. There can be an immense amount of which we can be consciously aware at any one time. For example, imagine looking down at a city from a window high up in a tower block…In this case I am consciously perceiving the complex distribution of trees, roads, and buildings; the colours on the ground and in the sky above; the moving cars and pedestrians; and so on. And I am conscious of all of this simultaneously. (166–167)
…[C]onsciousness is constituted by an accessibility-relation to occurrent thinkings, where those thinkings are conscious in turn (that is, where they are regularly made available to further occurrent thinkings). Conscious experiences, in particular, are those which are available to acts of thinking which are reflexively available to further thinkings…
What makes my perception of a glass on the desk to be conscious, on this account, is the fact that perceptual information about the glass is held in a short-term memory store whose function is to make that information available to conscious thinkings—where those thinkings are conscious, too, in virtue of each one of them fed back, reflexively, to the same (or similar) short-term memory store to be available to be thought about in turn. (194)
There is a lot to be said about this discussion, but I want to focus on three points here: first, the way that consciousness, for all of Carruthers’ philosophical care, is taken as a unitary, all-or-nothing phenomenon; second, the easy claims regarding what we are conscious of or not in perception (surveying the cityscape from the skyscraper vs. driving); and third, the implication of thought, and hence of language, in the story.
The term consciousness is used in this passage to denote our sensitivity to information in order to make the claim that we are conscious of the visual array when looking from the window of the skyscraper, to denote introspective availability (but not mere sensitivity) when denying conscious awareness of what we are doing when washing dishes or what we are seeing when driving, and conceptual availability when offering a more formal definition at the end of the passage. This easy slide is commonplace in this literature, and it is pernicious. It is pernicious not because there are many senses of consciousness at play; it is only appropriate that there are, as there are many cognitive phenomena denoted by this vague single term; rather, it is pernicious because the very singularity of the term and our familiarity with its use leads us both to take consciousness to be a unitary phenomenon, and one whose presence or absence is simply apparent to us.
From a Buddhist perspective, neither could be further from the truth. As we have seen, Buddhist psychology recognizes multiple kinds and levels of consciousness, including sensory and conceptual forms of consciousness; consciousness that is introspectible and consciousness that is too deep for introspection; consciousness that takes external phenomena as objects and consciousness that takes inner phenomena as objects; consciousness that is merely receptive and consciousness that is constructive and even projective. In general, the complex set of phenomena is opaque to casual introspection, and are knowable only theoretically or perhaps by highly trained meditators. Let us leave the meditators to the side, for a moment. To take seriously the idea that consciousness itself is both complex and opaque is to see that what Carruthers here takes as unproblematic data may well be nothing more than fantasy, akin to taking the flatness of the earth, so manifest to us, as a datum in astronomy.
The second point is closely connected. Carruthers blithely asserts that he is conscious simultaneously of all of the trees, cars and colors below his window, but also that when driving he is not conscious of the obstacles on the road he avoids when his mind is occupied in conversation. How on earth can he know either of these things? We already know that when surveying our visual field, or even attending voluntarily and carefully to part of it, that there is much that we do not see. We are inattentionally blind, change blind, blind to foveal gaps and to the monochromicity of our peripheral vision. Introspection is no guide to what is available to vision (Most 2010; Most et al. 2005; Mack 2003; Rensink 2001, 2002, 2004; Rensink, O’Regan and Clark 1997; Noë 2007; Noë, Pessoa and Thompson 2000; Simons 2000a, 2000b; Simons and Ambinder 2005; Fendrich, Demirel and Danziger 1999; Guirao and Artal 1999).
There is a neat dilemma here for one who wants simply to take the data of naive introspection as a starting point for inquiry in this domain: either take the term conscious to apply only to what is available to introspection or take it to be a complex, theoretically informed term that may well comprise a natural kind whose members are heterogeneous and whose natures are hidden from introspection. If we take the first option, all we are doing is investigating introspective awareness, something we know to be variable among subjects, prone to illusion and dependent on attentional variables. Characterizing that might be interesting as a matter of autobiography, but not as philosophy or psychology. If we take the second, casual introspective reports are irrelevant. And this is why most Buddhist theorists eschew naive introspection as a source of data for reflection on consciousness, taking it to be the report of illusion, not a veridical source of data.16 This is an issue for almost every current philosophical account of consciousness.
Finally, it is noteworthy that Carruthers’s account (like those of Rosenthal and Gennaro), as a higher-order thought account, implicates conceptual thought in its account of consciousness—at least in its account of what we might more precisely call introspectible conscious experience—and hence language, as the vehicle for the expression of concepts. (I leave aside the vexed question of whether there are inexpressible phenomenal concepts for now. We will return to phenomenal concepts and qualia below.)
Here there is real harmony with at least several important strands of Buddhist thought about apperception. Philosophers as different in perspective as Vasubandhu and Candrakīrti (the former a Yogācāra and the latter a Mādhyamika critic of the Yogācāra school) in India; Tsongkhapa and Gorampa in Tibet (the former a Geluk insistent on the importance of conventional truth and the latter a Sakya scholar insistent on its entirely deceptive character) and Dōgen in Japan (who rejects virtually all of the analytic philosophy developed in the Indian and Tibetan scholastic tradition) agree that introspective awareness is mediated by thought and therefore by language. Therefore, moreover, it is mediated by the conventions that determine linguistic meaning and the structures of our concepts, and hence is always a matter of convention, not a deliverance of the nature of mind to a lucid instrument observing it. But from a Buddhist perspective, this insight of the HOT school undermines, rather than facilitates, its more general program of developing a theory of consciousness, per se. At best, one could hope for a theory of our folk conception of the meaning of the word consciousness. Not much to write home about.
A very different strand of thought running through contemporary theories of consciousness is the idea that qualia or qualitative properties, which are posited or said to be discovered introspectively as the inner objects of experience, constitute or explain the fact that “there is something that it is like” to have conscious experience, and given the emphasis that Buddhist ethics places on the interests of sentient beings (see chapter 9 below), the question of the nature of sentience, and hence of experience, is central to the Buddhist philosophical project. The idea that something like qualia are essential to sentience is common to a number of theoretical perspectives that differ among themselves in countless other ways, including Higher-Order Thought, Higher-Order Perception, Intrinsic First-Order, Phenomenological and Reflexive models. We have already encountered Chalmers’s posit of phenomenal properties and concepts to go with them. Gennaro makes a similar claim:
There is significant disagreement over the nature, or even the existence, of qualia, but they are most often understood as the felt properties or qualities of conscious states. There is something it is like to have qualia or to be in a qualitative state. Most generally, perhaps, qualia are “introspectively accessible phenomenal aspects of our mental lives.” But even this can be misleading if it is taken to imply that only introspected, or introspectible, states have qualia. Surely first-order, or world-directed, conscious states also have qualia. …What it feels like, experientially, to see a red rose is different from what it feels like to see a yellow rose. (2011, 7)
There is a massive philosophical literature on qualia and related constructs. I will ignore most of it in the interest of focusing once again on what Buddhism brings to this debate. But first, a few preliminary remarks on qualia to open the field for another perspective. It is often taken for granted in recent literature on consciousness that “surely…conscious states…have qualia.” But given how little clarity there is on just in what qualia are supposed to consist, we should be suspicious of this consensus. As we have seen, qualitative redness is not redness. Roses may be red, and violets may be blue; but red and blue qualia are found only in me and in you. These are inner properties, or properties of inner states. As Nida-Rümelin puts it,
To have a particular phenomenal property is to have an experience with a specific subjective feel. If you have a phenomenal concept of a phenomenal property, then you know what it is to have an experience with that subject feel. (2007, 307)
The echoes of Dignāga are striking. Levine (2011) agrees that phenomenal properties are special properties of inner states, essentially grasped by having them (just as the facets of the crystal turn blue when placed on the blue paper) and that perception is characterized by the immediate awareness of these properties. So, we should ask ourselves what reason there is to posit this second layer of properties in order to understand consciousness or experience.
There are two principal kinds of reasons one might have: observational or theoretical. The only kind of observational evidence could be introspective, and, as we have already noted, introspection is a highly fallible instrument for limning the cognitive world. But suppose that, for the sake of argument, we grant its veridicality. Let us now ask what occurs when we see a red rose. In particular, what introspective evidence could there be for the claim that beyond the redness we perceive in the rose, there is a second phenomenal redness that we perceive immediately, or that permeates our perceptual experience? To have such evidence, both properties would have to be evident, and we would have to be able to distinguish them. But by the qualiaphile’s own lights, only the phenomenal property is directly perceived. So, perhaps the evidence is theoretical. But then we would imagine that, somewhere in the best psychological theory of perception, qualia, qualitative properties, or qualitative concepts appear. The fact that they do not should at least give philosophers pause. I conclude that the surely that precedes the claim that our conscious states have qualia is the protestation that evidence and argument are not needed, albeit for a claim nobody really understands. We will see why this is below.
It might be interesting to take a whirlwind tour through the doxography of Buddhist positions on perceptual experience to get a very different take on these matters. Buddhist doxographers distinguish a number of schools of thought organized in a hierarchy of degrees of supposed sophistication. On the one hand, this doxographic enterprise helps to systematize the history of Indian Buddhist philosophy. On the other hand, it provides a dialectical map for reflection and education, motivating each successive view on the ranked list as a reply to difficulties raised by the immediately preceding view.17
The most straightforward account of perceptual awareness in the Buddhist world is presented by the Vaibhāṣika school. A pre-Mahāyāna school, the Vaibhāṣikas presented a non-conceptual direct realist theory of perception. On this view, perceptual awareness arises from the direct contact of sensory apparatus with perceptible objects or qualities in the environment, unmediated by any representations. We might see this as a kind of proto-Gibsonian model of perceptual consciousness: consciousness is not a kind of mental accompaniment to sensory contact, or even a downstream consequence of it; it is the very fact of sensory contact with the world. According to the Vaibhāṣika, then, consciousness is not a property, a thing, a process, but rather a simple relation between a sensory/cognitive system and the world. In this system, note, nothing answers to qualia; there are no special qualitative properties. Instead, in perceptual experience we are immediately aware of ordinary external properties. This direct realism, however, has difficulty accounting for perceptual illusion, as perception itself is taken to be veridical.
The Sautrāntika system arises as a response to the difficulty the Vaibhāṣika system has accounting for perceptual error or illusion. Sautrāntika is a representational theory of mind, interposing a representation (ākāra/rnam pa [nampa])18 as the direct object of awareness between the external object and subjectivity. While for a Vaibhāṣika, when I perceive a patch of blue, I perceive the patch directly, and the consciousness of that patch of blue just is the relation between my perceptual apparatus and the patch of blue, for a Sautrāntika, that consciousness is immediately of a representation of that patch of blue. The representation is itself blue, but in a different sense from that in which the patch is blue. It has blue content, or resembles the blue patch in some way.
It is here, if anywhere in the Buddhist tradition, that we come close to a doctrine of phenomenal properties. For the Sautrāntikas, like the Vaibhāṣikas to whom they respond, inner cognitive phenomena like representations cannot be literally blue. On the other hand, if they are the immediate objects of perception, since in perception we are conscious of blueness, they must be blue somehow. Here we might say that the account of perceptual consciousness on offer trades directly on the presence of blue qualia as mediators of our perception. We have encountered this view in our discussion of Examination of the Percept and its commentaries.
So, it might seem that the Buddhist world is friendly to an account of perceptual consciousness in terms of qualia. But not so fast. We are only on the second rung of the doxographic ladder. Yogācāra is presented as the next step in sophistication, and as a response to an intractable problem posed by the Sautrāntika representationalism, one that directly concerns the qualitative representations it posits. The problem, as Buddhist theorists see it, is twofold: First, if the blue representation is blue in a different sense from that in which the blue patch is blue, or if it somehow resembles the blue patch, the only way that it could mediate perceptual knowledge of the blue patch would be if we were to know how it resembles the external patch, or how one sense of blue connects to the other. But the very point of the representationalism is that we are cut off from direct contact with the represented, and so have no idea what it might be like, and so no idea of what is represented, or how.19
This is, of course, an anticipation of Berkeley’s argument against representationalism in the second Dialogue. And like Berkeley, the Yogācāra wield the argument in favor of idealism, arguing that since direct realism is incoherent, as is representationalism, the direct and only object of conscious experience is an inner state. A second argument deployed against the Sautrāntika position is a simple regress argument: if the only way that we can be perceptually aware of an external patch of blue is by directly perceiving something qualitatively blue, then it would seem that the only way that we could ever perceive a patch of qualitative blueness would be by perceiving something that was qualitatively qualitative blue. And so on. That is, if the mediation of the perception of a property by the perception of a second property is needed in the first place, the perception of that second property should need similar mediation. If it does not, there is no reason to think that the first does, either.
The view that there are phenomenal properties corresponding to the properties of experienced objects, and that our perception of the properties of objects is mediated by those phenomenal properties is well-known in Buddhist literature as well. The eighth century Indian philosopher Dharmottara, in his Epitome of Philosophy (Nyāyabindu), a sub-commentary on Dharmakīrti’s Commentary on the Encyclopedia of Philosophy, considers and replies to a query regarding the causal basis of perceptual intentionality:
Question: How can a single cognition have the relation of intended and intentional?
Response: Since that cognition, experiencing the likeness of blue, is established as apprehending blue by a thought that is a judgment (niácaya-pratyaya), therefore the experienced likeness is the cause of intending. And that cognition, being established as an experience of blue by a thought that is a judgment, is what is intended. Therefore, a cognition’s resemblance, which is realized by way of exclusion of what is unlike (apoha), is a cause of intending; and the fact of being in the form of a thought of blue, which is realized by exclusion of thoughts of nonblue, is what is to be intended. (trans. Arnold 2009, 194–195)
The interlocutor here wonders how it could be that an internal state could have both subjective and objective character, that is, to be, in this case, at once the apprehended blue of an intentional object and the apprehender of that blue. The reply is that the apprehended blue is grasped by means of a subjective aspect of the intentional state which resembles that blue, which is subjectively blue. Now, I think it is plain that this account falls prey to the same difficulties that afflict contemporary perceptual theories according to which perception is mediated by inner phenomenal properties that correspond to external properties. Nonetheless, it is better off than most of these theories in one odd respect. Because this is an idealist theory of perception, according to which there is no genuine external object to be perceived, the two occurrences of blue are both mental. It is much easier to see how one mental occurrence of blueness could be involved in the perception of another in virtue of similarity than to see how a mental occurrence of blueness could be similar enough to physical blueness to constitute the mechanism by means of which an intentional state is directed on that property, as Berkeley pointed out.
In a Yogācāra analysis of perceptual consciousness, the external object and the distinction between subject and object drop out entirely. As a response to a mediated view of consciousness, Yogācāra reinstates an immediate theory, but one in which the only terms of experience are inner. On this view, perceptual consciousness has no external object; to be conscious is just for a conscious perceptual state to arise in cognition, and the analysis of the contents and character of consciousness involves only attention to inner states. This view has certain obvious advantages: it avoids the need to posit a layer of mysterious properties that are not obviously explanatory, and it avoids what looks like a nasty explanatory regress, with the claim that conscious states simply arise as a consequence of propensities in the foundation consciousness, or, as we might put it, our fundamental psychological architecture.
On the other hand, Yogācāra is not obviously successful as an explanation of perception or consciousness, inasmuch as what looks like awareness of external objects, and what looks like a distinctive mode of engagement with the world, simply are explained away as the sui generis arising of conscious states from a foundation consciousness. What looks like explanation seems more like the positing of brute fact; while it looks like perceptual consciousness should arise from a variety of causal factors, it ends up being a kind of explanatory surd.
Madhyamaka enters as the most sophisticated view, then, as a response to this explanatory predicament. The Madhyamaka view of perceptual consciousness is simple and sophisticated. Mādhyamikas analyze consciousness, as they analyze all phenomena, as a set of relations, not as an independent phenomenon or characteristic. In this case, perceptual consciousness is simply the fact of the relations between a perceived object, a sense organ, a sensory system and the conceptual and motor systems to which that sensory system is connected. Just as the illusion of a self is resolved in favor of a network of interconnected psychophysical processes, the illusion that there is a special property or center of consciousness is resolved in favor of a network of processes.
That they constitute consciousness is simply the fact that they are perceptual, conceptual, conative and affective. The properties they deliver are simply the properties of the external objects we perceive, as they are delivered by our perceptual processes.20 There is nothing more than that. This may feel disappointing, as deflationary accounts often do, but, from a Madhyamaka perspective, all that we lose in such an account is the illusion that there is more in conscious experience than the psychology and physiology of experience. In particular, reference to internal representations, qualia, phenomenal properties and other such ghostly mediators of our experience drop away. Ontology becomes cleaner, perhaps more naturalistic, and certainly more public, less private.
I review this Buddhist doxography in order to point out that what is taken as obvious in so much of the Western discourse on perceptual consciousness—that it centrally involves something like qualia, qualitative properties or qualitative concepts—is taken in this tradition as but one unstable moment early in a dialectic, and a moment that emerges not from naive introspection, but from reflection on illusion, as did the introduction of sense data in the early 20th century. This dialectic, moreover, is not aimed at providing an analysis of a special, singular property that makes our experience come alive, or confers subjectivity to it, but at dispelling the illusion that there is any such singular property. There are many kinds and degrees of consciousness on this view, and they reflect the many kinds of ways that cognitive and perceptual processes engage their objects. It is the engagement that constitutes, rather than gives rise to, or is accompanied by, consciousness on this view.
I do not claim that this Madhyamaka view—or any Buddhist view—is obviously correct, only that it is a voice with which to be reckoned. And reckoning with this voice forces the theorist who takes something like the qualitative character of experience to be real, and to be essential to consciousness, to defend and not to presuppose that view. And the route to a defense is not at all obvious. It also forces us to ask once again just whether when we propose a theory of consciousness, we even know what we are talking about, or whether the object of our theory exists. We may be doing something akin to the biology of unicorns.
There is a major strand of theory in contemporary consciousness studies that we have ignored so far, the neo-Husserlian phenomenological strand left hanging in the previous chapter. On this view, consciousness is essentially bound up with self-consciousness, and to understand what it is to be conscious is in the first instance to understand what it is to be self-conscious. Dan Zahavi (2005) is a sophisticated proponent of this view. He writes:
Although phenomenologists might disagree on important questions concerning method and focus, and even about the status and existence of the self, they are in nearly unanimous agreement when it comes to the relation between consciousness and self-consciousness. Literally all the major figures in phenomenology defend the view that the experiential dimension is characterized by a tacit self-consciousness. (11)
This is intriguing. Of course this is not an argument for the claim that all consciousness involves self-consciousness, but it does indicate a willingness to take this as read as a starting point for investigation. Of course to say that all consciousness involves self-consciousness is not yet to specify how it involves self-consciousness, or what the nature of that self-consciousness is. It may be, for instance, that self-consciousness is merely potential, and not actual; that it is the mere consciousness of subjectivity, as opposed to the consciousness of a self; it may be that it is the consciousness of a specific subject of experience. Let us see how Zahavi specifies the account:
Self-consciousness is not merely something that comes about at the moment one scrutinizes one’s experiences attentively…Rather, self-consciousness comes in many forms and degrees. It makes perfect sense to speak of self-consciousness whenever I am not simply conscious of an external object—a chair, a chestnut tree, or a rising sun—but acquainted with my experience of the object as well, for in such a case my consciousness reveals itself to me. Thus, the basic distinction to be made is the distinction between the case where an object is given (object-consciousness) and the case wherein consciousness itself is given (self-consciousness). In its most primitive and fundamental form, self-consciousness is taken to be a question of having first-personal access to one’s own consciousness; it is a question of the first-person givenness or manifestation of experiential life. (15)
Self-consciousness, then, is being “acquainted with my experience of the object as well,” or “first-person givenness or manifestation of experiential life.” It may seem obvious that when I see a blue sky I am not only acquainted with the blue sky but with my experience thereof, and that I know my own life from a first-person perspective. The latter fact, however, is a tautology. How else could I know my own life? The sleight of hand that gives content to this formulation is only apparent when we ask whether when an object is “given”—that is, when I become perceptually aware of an object—my experience is also given—that is, made aware to me—as a second object of awareness. This is not at all obvious, and no argument is offered for it. For my part, when I perceive a sky, my experience is exhausted by seeing the sky. If the experience of the object is distinct from the experiencing of the object, I sense one object too many, and we are pushed back to the Sautrāntika position, and the regress problems it raises.
Zahavi insists that his theory is no higher-order theory. Like Kriegel, he argues that self-consciousness is intrinsic to consciousness, and not the result of a moment of awareness itself becoming object of another state:
In contrast to higher-order theories, the phenomenologists explicitly deny that the self-consciousness that is present the moment I consciously experience something is to be understood in terms of some kind of reflection, or introspection, or higher-order monitoring. It does not involve an additional mental state, but is rather to be understood as an intrinsic feature of the primary experience. That is, in contrast to the higher-order account of consciousness, which claims that consciousness is an extrinsic property of those mental states that have it, a property bestowed upon them from without by some further states, the phenomenologists typically argue that the feature that makes a mental state conscious is located within the state itself; it is an intrinsic property of those states that have it. (20)
Thompson follows Zahavi closely:
Consider visual experience, When I see the bottle of wine in front of me on the table, I experience (I am visually aware of) the wine bottle. But I also experience my seeing. In experiencing my seeing in this way, I do not need to introspect or reflect; my awareness is instead an implicit and nonreflective one. I experience my seeing by living it nonreflectively. (2007, 285)
From a Madhyamaka perspective, this is seriously problematic stuff. When we say that a state is intrinsically conscious, we are stating that its being conscious is independent of its relation to any object, to any perceptual system, to any other psychological processes. Even philosophers like Dharmakīrti and Śāntarakṣita, with their respective accounts of the primitive manifestation of subjectivity in conscious states, took that subjectivity to be parasitic on their intentionality and directedness to an objective content. Zahavi’s and Thompson’s view, on the contrary, like Husserl’s, raises the possibility of a state being conscious, but not being conscious of anything, and of the possibility of seeing without seeing anything. The multiplicity of objects of experience is either gratuitous or incoherent, and in its attempt to rescue consciousness and intentionality from reduction, leave these phenomena as unexplained primitives. It is indeed difficult to see how this makes sense at all. At this point, the Buddhist insistence that consciousness is a relation, not a brute, intrinsic property is surely a voice necessary to bring clarity to the discussion. The confusions entailed by this version of phenomenology only multiply:
We are never conscious of an object simpliciter, but always of the object as appearing in a certain way; as judged, seen, described, feared, remembered, smelled, anticipated, tasted, and so on. We cannot be conscious of an object (a tasted lemon, a smelt rose, a seen table, a touched piece of silk) unless we are aware of the experience through which this object is made to appear (the tasting, smelling, seeing, touching). This is not to say that our access to, say, the lemon is indirect, or that it is mediated, contaminated, or blocked by our awareness of the experience; the given experience is not an object on a par with the lemon, but instead constitutes the access to the appearing lemon. The object is given through the experience; if there is no awareness of the experience, the object does not appear at all. (Zahavi 2007, 121)
At the end of this passage we have a conclusion that, at least from a Madhyamaka perspective, is a reductio on the entire project. “If there is no awareness of the experience, the object does not appear at all.” Once again, if the experience is the appearance of the object, this is a tautology: If there is no appearance of the object, there is no appearance of the object. The claim is only contentful, though, if the experience is more than the appearance of the object, if it is a special inner appearing. And there is simply no reason to think that when I am aware of a blue sky, I am also aware of a second thing, namely its appearance. This really does seem like a needless multiplication of entities, and even worse, entities with obscure properties, obscure identity conditions, and which turn no explanatory wheels. Once again, reflection on the Buddhist tradition of thinking of consciousness not as some intrinsic property, but as a relation between thinkers and their objects, suggests a route out of this morass.
All of this brings us back to the question of the meaning of the phrase “what it is like,” that looms so large in current discussions.Uriah Kriegel (2009), as we saw above, has helpfully distinguished two aspects of the intended meaning of this phrase, what he calls qualitative character and what he calls for-me-ness. He argues that any conscious experience must have both of these two characteristics, that that is what it is like to be conscious. We have already found reasons to doubt the need for consciousness to have some special qualitative character. The discussion of subjectivity above can help us in dismissing the need for for-me-ness.
Kriegel claims that consciousness consists in a kind of penumbral halo around every experience, whether perceptual or cognitive, that reveals it as mine. Now, he concedes that we have no introspective evidence for this. And we might add that there is no experimental psychological evidence, either. The argument would have to be philosophical. Rather than imagine such arguments, let us look once again at the conclusion. If the argument is meant to show that when I have an experience, it is mine, the claim is an empty tautology. But if the claim is that beyond the experience of the blue sky, there is an experience of me seeing the blue sky, and that that is not a higher-order or an introspective experience, we are back to a regress-generating, explanatorily impotent inner state for which there is no evidence, exactly the kind of position that Candrakīrti criticizes so trenchantly.
I conclude that one consequence of this Madhyamaka take on consciousness is that the phrase “what it is like” is simply devoid of any content, if it is meant to apply to anything other than the objects of experience, and does nothing to explicate consciousness. I can say what a blue sky is like: it is blue; what a red sunset is like: it is red. But to go further and to say that there is something more or different that it is like to see a blue sky or a red sunset is simply to obfuscate, and to posit an ineffable, mysterious nothing as a mediator of my awareness of the world. We will return to this issue and explore the sense of the phrase “what it is like” in greater depth in the next chapter.
When we turn to the question of for-me-ness, things are even worse. For the sense of being mine, and the sense of a self that is the subject of consciousness, as we have seen earlier, are, from any Buddhist perspective, necessarily illusory. For this character presupposes both the awareness of a subjective self and the appropriation of the experience as belonging to that self. On any Buddhist view, however, there is no such self, and there is no intrinsic possession relation between a self and its experiences or faculties. There are only psychological and physical states and processes, including sensory faculties that constitute subjectivity. Together they form the basis of imputation of a conventional person, but they do not constitute a self. So, to the extent that consciousness is implicated with self-consciousness, it is not revelatory of its character, an idea taken for granted in virtually all of this literature, but rather obscurational, or, in Sanskrit, samvṛti, concealing. So, even if, despite the absence both of introspective or scientific evidence, one felt compelled to posit this explanatorily impotent property of consciousness, if the general Buddhist framework for thinking about the self is at all correct, the consequence is only that consciousness is necessarily opaque, distorting its own character as it presents it and superimposing a deceptive subject-object duality. It might be better to rethink this posit.
We remain simply and necessarily strangers to ourselves, we do not understand ourselves, we must be confused about ourselves. For us this law holds for all eternity: “Each man is furthest from himself”—where we ourselves are concerned, we are not “knowledgeable people.”
—Nietzsche, Genealogy of Morals
No discussion of contemporary views of consciousness would be complete without a discussion of some of the central examples—real and imagined—that animate that discussion. For part of what makes it appear obvious that consciousness is some special thing or property with a nature either intrinsic or extrinsic is that we are invited to imagine the difference between conscious and unconscious states or beings, and to imagine in them in ways that they constitute what Dennett once called “intuition pumps” (1980). It is time to dispose of those once and for all, following the lead of the Buddhist-inspired analysis offered thus far.
Zombies perhaps play the leading role in this philosophical horror show. They are deployed effectively by Chalmers (2010) among others to argue that there is a special “hard problem” about consciousness, in virtue of the fact that there could be beings functionally exactly like us, but obviously unconscious—zombies. Now, nobody claims that these zombies are nomologically possible; the claim is rather that they are metaphysically possible. Let us leave aside for now (though we introduced the topic in chapter 1 and will return to this issue in chapter 7 below) what metaphysical possibility amounts to, and how we can determine what is metaphysically possible. Let us rather ask whether, even on a generous reading of the relevant modal metaphysics and epistemology, zombies so understood are metaphysically possible. We shall see that in taking them to be so a question is being begged, and that the begging of that question is obvious from a Buddhist perspective.
Zombies are nonsentient duplicates of us. There are two ways to characterize the sense in which they are duplicates, however. We can think of them as physical duplicates without qualitative states or as functional duplicates without qualitative states. There are good reasons to prefer the second option dialectically. If we are asked to imagine physical duplicates of us who lack any cognitive states at all, we might simply balk at the imaginative exercise. That is an exercise in special effects: we are asked to imagine something that appears to be a duplicate, but is not. I might as well be asked to imagine a triangle whose interior angles sum to 100º. The fact that I claim to be able to do so would be no guide to metaphysics, only to my own strange imaginative powers, real or illusory. It is at least prima facie more interesting to imagine zombies who are physical and functional duplicates of us, but who lack only qualitative states. If we can make sense of this, we might motivate a dualistic theory of consciousness, and this is the goal of the zombie exercise. It is for this reason that dialectically, even if zombie discussions begin with physical duplicates, they quickly move to a discussion of functional duplicates.
In this context, a functional duplicate must be understood as a being with the same beliefs, desires and other intentional states that we have, but who lacks any states with qualitative character. The idea here is that intentional states are relational, and can be characterized in terms of their typical causes and effects—their relations to stimuli, to behavior and to other intentional states; to have qualitative character, however, is taken to be monadic, to be intrinsic to the state itself. This is what gives zombies their plausibility, and the hope that their possibility can illuminate the nature of consciousness, and hence their starring role in contemporary Western philosophy of mind. But does it yield possibility? At least in a non-question-begging way?
Let us focus on the fact that a zombie is meant to be functionally—that is, cognitively—identical to a human being, but lacking in qualitative states. If it is not really cognitively like us, there is really nothing of interest here. For then, all we are really imagining is a special effect: something that appears to be human in interesting senses, but is not—a product of PIXAR, perhaps. Or we are asked to fall back on the implausible—and not obviously imaginable—target of a physical duplicate that behaves exactly like us, but with nothing inside. The important point, then, is that despite behavioral equivalence—and hence, at least prima facie, cognitive equivalence—there is nothing “that it is like” to be the zombie. It has no genuine experience, no consciousness. One can feel the pull of intuition here. (Though I note, non-probatively, but as an anecdotal aside—that nobody I have ever met from a Buddhist culture has found the case even vaguely intuitively compelling.)
Now, we can see the problem for this device: for the qualiaphile zombologist, (1) our beliefs about our perceptual states are caused by the qualitative character of our states, as are our introspective beliefs that we have experience; (2) zombies are cognitively/functionally identical to us; (3) zombies perceive the world (unconsciously) and believe that they have qualitative perceptual states; but (4) their beliefs are not caused by any qualitative states, because they lack them; their perceptual beliefs and their introspective beliefs that they have experience are false. It is obvious that these four claims are inconsistent with one another. Together they constitute a reductio on the hypothesis of the metaphysical possibility of zombies.
I do not want to go too far down the path of zombie debates, a path that leads through many thickets. I do note that Chalmers, among others, does reply to this argument. He argues (2010) that (1) is false, because the relation between qualitative states and qualitative beliefs is one of constitution, not causation. Hence, he argues, zombies don’t have beliefs about qualitative states. This reply is patently question-begging. First, it requires a unique set of intentional states that are in part constituted by their objects, but second, it denies the obvious possibility that false qualitative beliefs are possible, in which case there is no way to tell whether we are zombies. He denies (2) on the grounds that whereas our states are caused by (or constituted by) qualia, zombies’ states are not. But this would be to deny that zombies are enough like us to be duplicates, as opposed to mockups. It is thus hard to see how one can mount a non-question-begging argument for the possibility of zombies.
So, when we claim that zombies are devoid of qualitative states, while claiming that our own are cognitively relevant to our lives, we can no longer claim that zombies are cognitively like us. This conclusion only reinforces the Madhyamaka insight that consciousness is not a thing, or a unitary property, but a complex set of relations. Moreover, it shows that the very supposition that zombies are possible is equivalent to the supposition that there is a discrete something—qualitative character, or consciousness—that we can simply subtract from ourselves to get some other possible entity, the zombie. That supposition is not only false, but is the conclusion of the argument of which the possibility of zombies is meant to be a premise. (See Garfield 1996, as well as Braddon-Mitchell 2003, Hawthorne 2002 for more detailed discussion of this problem as well as Dennett 2003, Harman 1990 and Webster 2006 for other reasons to distrust zombie intuitions.)
It might appear that in a book about the importance of Buddhist philosophy to Western philosophy I am spending far too much time on this issue. There is a reason for this attention, though. Zombies don’t show up in Buddhist discussions of mind and consciousness, and there is a reason that they do not. To be sure, they are meant to provide arguments for the non-physical status of consciousness, and just as there are modern dualists about the conscious and the physical, there were, and are, Buddhist dualists about these matters. These Buddhist dualists should welcome this line of argument, were it sound. But Buddhist philosophers saw that one can’t simply treat consciousness as a simple thing or property that one either has or does not; it is a complex of states, processes and capacities; and Mādhyamika Buddhists like Candrakīrti and Tsongkhapa saw that we can’t just look inside and report on the nature of our psychological states; nor can we say a priori what kinds of things have or lack any of these kinds of consciousness. These lessons are both apposite here; either would have forestalled the descent into this peculiar rabbithole wonderland.
Illusions of consciousness confirm this sensible approach. The well-known phenomena of inattentional blindness (Most 2010; Most et al. 2005; Rensink 2000b; Rensink, O’Regan and Clark 1997; Noë 2007; Simons 2000a, 2000b), change blindness (Rensink 2001, 2002, 2004; Noë, Pessoa and Thompson 2000; Simons and Ambinder 2005) and blindness to the anisotropy of the visual field suggest that not only do we refer to a motley, complex set of processes and states as conscious, and so that there is no unitary phenomenon to be called consciousness, but also that we are not introspectively authoritative regarding the objects and properties to which we are responding. The well-known phenomenon of masked priming (see for instance Masson and Isaak 1999) demonstrates conclusively that our cognitive processes respond to stimuli of which we had to be conscious, but which we cannot report, and of whose effects on our thought we are completely ignorant. Our conscious life is, rather than being transparent, opaque; rather than revealing our subjectivity often occludes it and its nature.
All of this suggests that while it appears that we know ourselves and our inner life intimately from a first-person point of view, a point of view that can provide immediate data for a theory of that life, in fact, we are strangers to ourselves, and what we take to be immediate data may be nothing more than illusion. Not only is there nothing that it is like to be me, but even my judgments of what particular experiences are like are likely to be useless. If zombies are possible, we may well be zombies. If they are not, we learn nothing about consciousness from the thought experiment. Once again, the Madhyamaka insight that the mind, and even consciousness, are hidden, rather than manifest phenomena, known only by inference, and through imperfect processes, can be a useful corrective to the easy use of introspection, speculation and shallow phenomenological analysis.
Blindsight raises further problems. Blindsighters affirm sincerely that they cannot see, are introspectively unaware of any vision, and form no articulable beliefs about the visible objects in their immediate environment (Collins 2010, Danckert and Rossetti 2005, Ptito and Leh 2007). Nonetheless, they navigate successfully around obstacles using only visual cues and reliably reach in the right place and for the right objects in forced choice situations, while insisting that they are only guessing. If we take consciousness to be necessarily transparent, or introspectible, or available for verbal report, or revelatory of our subjectivity, or as “feeling like something,” or to satisfy any of the characterizations prevalent in the Western literature, we would have to conclude that blindsighters are not visually conscious.
Blindsighters’ actions, however, are modulated by sensorimotor processes that are shared by ordinary sighted people, and blindsight might well be exactly like the visual mechanisms that subserve the visual experience of infrahuman animals, who may well not introspect and who do not report their visual experience. And it would be odd to deny them any consciousness. From a Buddhist perspective, there is no immediate reason to believe that all visual consciousness is introspectible, or that we are capable of reporting and reflecting on all that we see. Blindsight on this model simply involves the presence of one kind of consciousness in the absence of other kinds. Visual consciousness is present—there is receptivity and responsiveness to visual information—but this is not transmitted to introspective consciousness. Once again, this perspective, like this phenomenon, suggests that the construct of consciousness as an entity or property is just too simple to do justice to that which it is called to explain.
Though clouds’ illusions I recall
I really don’t know clouds at all.
—Joni Mitchell.
Madhyamaka and Yogācāra accounts of the mind remind us of the complexity and limitations of introspection. Even the most optimistic and, we might say, naively realistic accounts of the introspective sense faculty we find in the Buddhist literature remind us that it is very much a sense faculty. As such, even if its objects are in some sense, or in principle, known by such a sense faculty, they are known fallibly. If they are intrinsically self-revealing, on the other hand, while they are not then known through the mediation of a human sensory system, to the degree that we take up that knowledge and make use of it, it is only subject to higher-order reflection. And such reflection is imperfect on a number of dimensions, and is always deeply inflected by our conceptual framework. As a consequence, we can’t think of introspection as a veridical means for simply delivering the inner world to subjectivity. Just how we can think of it is a complex matter.
Finally, and consequent upon this insight regarding the fallibility of the introspective sense, a review of Buddhist accounts of the project of self-knowledge reminds us that we may be massively confused about its structure, about its object and about the epistemic credentials of introspection or phenomenological reflection as tools for the use in the development of a theory of mind. (Schwitzgebel 2011 makes similar observations.) To inquire into the nature of self-knowledge in the first place is to grant that we don’t know how it works, and so to grant that these seemingly obvious instruments may be entirely deceptive. So not only are the objects of introspection delivered in an opaque way, but the instruments we use, and even the data delivered by those instruments are, in a deep sense, opaque to us. Self-knowledge becomes very elusive indeed.21
To use introspective data themselves, and to treat them as constituting an unproblematic given in that investigation, is then to treat the investigation as complete before it has begun. If we truly don’t know what the mind is, how it works, or how we might know about it, considerably more epistemic humility regarding what seems evident to us as we reflect on it may well be in order. If we learn nothing else from the Buddhist tradition of reflection on the mind than this, reflection on that tradition is well worth the trouble.
What does all of this suggest for current discussions of self-knowledge? While one cannot simply map Buddhist debates onto current debates, I do think that the dialectic we see developing from 5th-century India through 15th-century Tibet helps us to appreciate both the perils of immediate first-order accounts of self-knowledge, and the degree to which taking higher-order views seriously forces us closer to taking knowledge of our inner lives to be very much like knowledge of the external world. We are thus led to accept a serious falliblism about introspection, and to appreciate the complex and problematic role of inference and theory in making sense even of higher-order perceptual processes. All of this means that what we might take as unproblematic explananda for, or commonsense intuitions to be accommodated by, theories of self-knowledge, might instead be illusions or delusions to be explained away.
Long ago, Dennett (1991) made the same point, perhaps channeling Buddhist ideas unawares. He distinguishes between two approaches to phenomenology (44–45). The first takes as its task the mapping of our inner life, of inner episodes, processes and experiences as they are—of the structure of consciousness, we might say—taking as authoritative introspective reports of that inner world and our sense of what it must be like. The second takes as its task mapping the claims that people make about their inner life and the popular lore about what consciousness and its contents must be like—a kind of anthropology of phenomenological reflection.
While it might appear that the former is serious philosophy and the latter is merely a superficial survey of attitudes, Dennett points out that only the second has any respectable epistemic credentials. We can know what people say about their inner lives; but only if we take people to be oracular about those lives do we have any grounds for taking the inner appearances or experiences they report seriously as entities, and that requires an improbable model of introspection, one taken for granted by almost everyone in the contemporary consciousness literature in virtue of regarding our intuitions about our own inner experience as constituting the data to which a theory of consciousness must respond.
From a Buddhist perspective, we can recast this as the distinction between exploring an illusion versus taking that illusion for granted as reality. We know that our inner sense, like our outer sense, is fallible; we know that we tend to reify ourselves and to take our introspective awareness as well as our external perceptual faculties as more reliable than they are. We know that our understanding of our lives is deeply theoretically infected and that our theories are always tentative. To take what introspection delivers as reality is therefore, from a Buddhist perspective, simply to embrace delusion. On the other hand, to take what we believe about introspection seriously as a matter for reflection is to take delusion as an object for study, and that may be the first step not to an oracular understanding of an inner reality, but at least to an understanding of our own capacity to obscure whatever that reality might be.
The Chan/Zen tradition offers one more insight we might well consider in closing. In distinguishing between (1) acting thoughtlessly; (2) acting with thought and (3) acting without-thinking, Zen theorists challenge us to reconceive the omnipresence and necessity of self-consciousness in consciousness as well as the introspectibility and explicit availability to reflection of conscious experience. Consider the mastery of a skill, such as that of playing the piano or playing football. One first might bang away at the keys or kick the ball aimlessly about the pitch with no particular thought; we might even say that one is not conscious of what one is doing. Perhaps the mind is elsewhere; perhaps there just aren’t the conceptual resources to make sense of one’s activity. But with a bit of coaching and prompting, one begins to play with thought, attending to where one’s fingers are on the keys, to how one strikes the ball, to where one’s teammates are on the field. At this point, perhaps, we might think that one is acting with maximum consciousness.
But of course when real expertise clicks in, all of that thought that was once so necessary to guide action drops off. Action becomes automatic. One is no longer able to say where one’s fingers are, of what key is to be struck next, of the angle of one’s foot, or even who is where on the field. One just plays, spontaneously and expertly.22 Now, in one sense, that of explicit access to what is going on, that of phenomenal feel, or of self-consciousness, there is no consciousness. But this is a straightforward reductio on these criteria of consciousness. From the standpoint both of Chan theory and of common sense, it is in this kind of expert performance that consciousness is most manifest and most present, when our responsiveness and effectiveness are at their maximum, when we are most fully engaged, most fully present. Once we realize this, we might reflect on the fact that most of what we do—walking, interacting with one another, cooking, reading—we do with something like this expert consciousness. This is the stuff waking life is made of, and in this mode of engagement, the world is, as Heidegger would put it, ready-to-hand, not present-to-hand, in a mode of conscious engagement that is not introspected, in which the subject is not present even as periphery, and in which both higher-order and reflexive thought is absent. We will return to this topic in more detail in the next chapter.
This suggests that a Buddhist perspective can do a great deal to enrich contemporary discussions of consciousness. It brings a suite of insights that can refigure debates and ways of thinking about the topic. Instead of seeing consciousness as a singular phenomenon, a natural surd perhaps requiring non-natural explanations, the Buddhist asks us to think of consciousness as a family of relations that subjects might bear to their objects. Its different levels and manifestations may require very different kinds of explanations. But in each case they will take the form of natural explanations of the relations between psychological processes and their conditions or objects, not mysterious powers of a self or subjectivity.
1. It is important to note that I am using this term not in the original sense in which it was introduced by Block (1995), but in a sense that has become more common since then. Block used it to denote access to external information, as in the case of blindsighted individuals, who in his sense have access to information about objects in their environment despite lacking any of what he called (and what I will call) phenomenal consciousness. I instead use this term to denote access to one’s own inner states. Below I will call that to which Block refers as access consciousness as responsive consciousness. I think that this vocabulary is more intuitive, and accords more with ordinary usage these days.
2. And of course the contemporary literature is replete with alternative classifications of kinds of consciousness.
3. This is sometimes called creature consciousness. But this term may beg questions regarding whether there is a common phenomenon in this neighborhood across phylogenetic types.
4. The exception is that body of literature deriving from Heidegger’s attention to mood as structuring subjective consciousness.
5. But see Harman (1990, 1996), for instance, for a very different view of qualia, not as direct objects of consciousness, but as aspects of a transparent relation to distal objects.
6. Of course they are also the objects of access and responsive consciousness, but that is beside the present point.
7. This term as well deserves comment. saṃjñā, unlike perception, connotes a kind of discriminative awareness, and even a linguistic, or quasi-linguistic awareness, what we would call perception-as, as opposed to a mere undifferentiated awareness.
8. This term can also have the sense of volition, or intention to act. And so it shares the ambiguity of intention in English. But that sense is not relevant to the present discussion. For more on the range of the term, see Meyers (2010) and Heim (2013).
9. It is important to remember in this context, as we noted in that context, that the mere fact that we imagine such things is no guide to fundamental metaphysics. This doesn’t entail that we are something entirely distinct from our body, our mind or our aggregates, only that when we take ourselves to be selves in that sense, these reveries seem to make sense. The imagination exercise, then, tells us a great deal about what we take ourselves to be, but nothing at all about what we actually are. This does not mean, however, that considerations of these exercises are metaphysically idle. When we are interested in what a self would be, it is essential to fix on the object of negation, to make it clear, even if our purpose is to demonstrate its nonexistence.
10. Once again, it is worth recalling the similarity of this view to that of Husserl and his more recent followers, such as Zahavi (2004, 2005) and Thompson (2007).
11. We leave aside here the hard-to-characterize state of awakening, in which that consciousness undergoes a complete transformation and becomes an objectless awareness.
12. Whether there is non-perceptual phenomenal experience is another fascinating question we will leave to one side, as it has no bearing on the issues at hand here.
13. The material in this discussion related to Alaṃbanaparikṣā and its commentaries is from as yet unpublished joint work with John Powers, John Makeham, Sonam Thakchöe, Douglas Duckwkorth, M David Eckel and Dan Lusthaus, supported by the Australian Research Council.
14. Kriegel, at the time he wrote this, was unaware of the Buddhist antecedents to (and critiques of) his view. I do not want to suggest, however, that he remains unaware of them. In any case, it is instructive to read his argument in the context of those antecedents.
15. Expressivist views, such as those of Shoemaker 1994 and Bar-On 2004, share with the self-luminosity view the sense that self-knowledge is immediate, as opposed to being mediated by additional cognition, but these views distance the subject slightly from the state known, requiring an act of profession. One way of putting this point is that while on the self-luminosity view, it is impossible to be in a conscious mental state without knowing oneself to be in that state, on the professing view, it is possible to be in such a state, and only to come to know it when that state finds explicit inner or outer expression.
16. On the other hand, some Buddhist traditions, particularly meditative or tantric traditions, do rely heavily on the introspective reports of trained and experienced meditators, taking those reports to provide compelling evidence regarding the nature of cognition. Whether this is an epistemically warranted practice is a real question, both in Buddhist epistemology and in contemporary contemplative neuroscience. On the one hand, science requires observational data, and trained observers are better than untrained observers. On the other hand, trained observers can easily find only what they seek, and in the introspective domain calibrating the observer is difficult, if not impossible. I note this issue only to set it aside for another time.
17. Doxography of this kind, common in India, Tibet and China, might strike the Western philosopher or historian of philosophy as a bit exotic and odd. But of course we engage in doxography as well to organize our own canon, even if we don’t go in for the hierarchies that characterize Buddhist views. We have our Platonists and Aristotelians, rationalists, empiricists, Kantians, idealists, realists, and even HOT, HOP and FO theorists in the present debate. Doxographies are, of course, both useful heuristics for understanding the history of philosophy and polemical strategies for shaping it. Here we follow a fairly standard, and clearly polemical, Tibetan doxographic framework. (See Hopkins 1996 and Cabezón 1994, as well as Gregory 1991 for more on Buddhist doxography.)
18. This term has a very complex semantic range and functions in many ways as a technical term in Buddhist epistemology and philosophy of mind. In some contexts it is best translated as aspect; in others as image or form. (Think of the many ways in which idea is used in early modern philosophy as a parallel.) A complete exploration of Buddhist theories of ākāra would take us far afield and would require a substantial volume of its own. For a very fine exploration of the role of this term and the phenomena it denotes in Buddhist epistemology see Kellner and McClintock (2014).
19. Here especially (but throughout this discussion) I simplify considerably, eliding substantial disagreements among philosophers in these schools in favor of a broad brush. The reader interested in the details of these debates should refer to Coseru (2012), Dreyfus (1997) and Kellner and McClinock (2014).
20. It is important to note here that despite their important differences along other dimensions, the Yogācāra and Madhyamaka traditions are united in the view that there are not two, but instead one set of such properties. Despite the fact that Yogācāra offers an idealist account of perception and Madhyamaka a realist account, neither posits both a set of empirical and a set of phenomenal properties in order to explain perceptual experience.
21. We should note in this context, however, that there is a very different position also represented in classical Indian Buddhist philosophy, one which attracts a great deal of attention both in the Indian and Tibetan literature and in contemporary Buddhist Studies and cross-cultural phenomenology. (See especially Coseru 2012, Thompson 2011 and the essays in Siderits, Thompson and Zahavi 2013.) That is the position of Dignāga and his followers, according to which we have immediate verdical access to our own perceptual states, in virtue of their being self-presenting and reflexive. We will consider this position in detail in the next chapter. This is the position against which Candrakīrti and his followers react. See Coseru (2012) for a defense of this position.
22. Compare Merleau-Ponty (1962) on the embodiment of skill.