Is the Mind a Magic Trick? Illusionism about Consciousness in the “Consciousness-Only” Theory of Vasubandhu and Sthiramati

  • Amit Chaturvedi orcid logo (University of Hong Kong)


Illusionists about consciousness boldly argue that phenomenal consciousness does not fundamentally exist—it only seems to exist. For them, the impression of having a private inner life of conscious qualia is nothing more than a cognitive error, a conjuring trick put on by a purely physical brain. Some phenomenal realists have accused illusionism of being a byproduct of modern Western scientism and overzealous naturalism. However, Jay Garfield has endorsed illusionism by explicitly drawing support from the classical Yogācāra Buddhist philosopher Vasubandhu. In this paper, I assess the degree to which Garfield’s illusionist interpretation accurately captures the views of Vasubandhu and his commentator Sthiramati in their extant Sanskrit works. As it turns out, Vasubandhu and Sthiramati converge with contemporary illusionists in taking an unconscious causal basis of cognitive/linguistic processes to be responsible for generating the illusion of representational states with apparently phenomenal contents. Within their constitutive understanding of the mind as the “imagination of what is non-existent” (abhūtaparikalpa), I raise possible candidates for what might seem to be real instances of phenomenality—mental appearances (pratibhāsa), affective sensory experience (vedanā), and “intrinsic luminosity” (prakṛtiprabhāsvara)—and consider possible responses on behalf of an illusionist interpreter. I conclude that Vasubandhu and Sthiramati really do appear to be strong illusionists about phenomenal consciousness, particularly if phenomenal states are assumed to be essentially representational.

How to Cite:

Chaturvedi, A., (2024) “Is the Mind a Magic Trick? Illusionism about Consciousness in the “Consciousness-Only” Theory of Vasubandhu and Sthiramati”, Ergo an Open Access Journal of Philosophy 10: 52. doi:



Published on
29 Feb 2024
Peer Reviewed

“Suppose, bhikkhus, that a magician or a magician’s apprentice would display a magical illusion at a crossroads. A man with good sight would inspect it, ponder it, and carefully investigate it, and it would appear to him to be void, hollow, insubstantial. For what substance could there be in a magical illusion? So too, bhikkhus, whatever kind of consciousness (viññāna) there is . . . a bhikkhu inspects it, ponders it, and carefully investigates it, and it would appear to him to be void, hollow, insubstantial. For what substance could there be in consciousness?”

Saṃyutta Nikāya 22.95 (Bodhi 2000: 952)

The illusionist theory of consciousness boldly claims that phenomenal consciousness only seems to exist, but does not in fact exist. There is really nothing “it is like” to have phenomenally conscious experiences, because there are no phenomenally conscious experiences. The outputs of our sensory systems are not actually accompanied by any such thing as qualia, that is, inner states which possess a qualitative phenomenal character in virtue of instantiating phenomenal properties. A subject may have the impression of being directly acquainted with a private inner life of phenomenal qualia, but this impression is nothing more than a conjuring trick played by the brain on itself. The primary task of illusionists, then, is not to answer the “Hard Problem” of how phenomenal consciousness arises from a purely physical system—instead, it is to solve the “Illusion Problem” of how a purely physical system could generate the appearance of being phenomenally conscious even though it is not (Frankish 2016a).

Many find illusionism highly implausible. To some realists about phenomenal consciousness, the root of illusionism’s error partly lies in an overzealous adherence to scientism or radical naturalism. Given the modern successes of natural science in explaining the world through third-person observation and describing phenomena in mathematical language, it’s tempting to think that the same types of observation and language should be capable of capturing every other remaining phenomenon in the world. Given the failure of science so far to bridge the “explanatory gap” between the third-person knowledge of objective physical facts and first-person knowledge of subjective phenomenal facts, it’s further tempting to reject the existence of the latter facts for being incompatible with a scientific worldview. But, this suggests that the illusionist choice to deny phenomenal consciousness may be forced by the limitations of a “scientistic” attitude that is itself conditioned by its intellectual and historical context: As phenomenal realists like Katalin Balog and Phillip Goff suggest, the scientism motivating illusionism may be variously traceable to “the visceral impact of technology on philosophical inclinations” (Goff 2016: 98), or to the fact that “modern life supports a tendency toward objectivity” (Balog 2016: 42) and away from irreducibly subjective realities like phenomenal consciousness.

So, the possibility that there were anti-realists about phenomenal consciousness well before the advent of modern science would stand to bolster the independent plausibility of illusionism, and dispel the impression that it is merely an artifact of its historical context. To Keith Frankish, the possible pre-modern existence of illusionism suggests that, “The fact that many Western philosophers find illusionism utterly implausible may say more about their cultural horizons than about the nature of consciousness itself” (2016b: 259). Just such a possibility is identified by Jay Garfield (2016), who endorses illusionism by drawing support from the early Yogācāra Buddhist philosopher Vasubandhu (4th–5th cent.) and his Treatise on the Three Natures (Trisvabhāvanirdeśa).1 According to Yogācāra Buddhists, unenlightened beings are afflicted by the basic illusion that a world of external objects is presented to a real inner subject. From this claim, Garfield draws an even more radical implication: One should abandon not only the idea that consciousness presents us with a world of objects outside our minds, but also the notion that there is an inner subjectivity to which these objects are presented. For Garfield, Vasubandhu shows that to admit the existence of internal conscious states with a subjective phenomenal character is to fall prey to still another pernicious illusion: the Myth of the Given. It is a myth to think is phenomenal experience or primitive subjectivity that intrinsically possesses “something it is like” to appear, independent of how we imagine or cognitively construct such experience.

Two issues with Garfield’s illusionist reading of Vasubandhu call for extra scrutiny, and hence motivate this essay. To start, the text Garfield cites—the Trisvabhāvanirdeśa—is likely not composed by the 4th–5th century Vasubandhu who is taken to be the author of seminal works and commentaries on early Yogācāra philosophy (Kapstein 2018). More importantly, the construal of Vasubandhu as an illusionist about consciousness would seem to contradict a standard reading of Vasubandhu as an idealist who claims that the whole world is “nothing but mind” (cittamātra), that the apparently mind-independent objects of experience are “nothing but mental representations” (vijñaptimātra), and who has been historically labelled as a proponent of vijñānavāda, or the theory that there only exists consciousness (“consciousness” here being used to provisionally translate “vijñāna”). Given that Garfield elsewhere interprets Vasubandhu’s Yogācāra texts as endorsing a kind of idealism (2002; 2015), it appears strange that he would then advocate that we follow Vasubandhu’s denial of phenomenal subjectivity precisely to avoid the slippery slope to idealism that looms if our contact with the world is mediated by an immediate experience of internal qualia (2016: 74).

This essay therefore investigates whether Garfield’s illusionist reading accurately captures Vasubandhu’s views as expressed in the Yogācāra works that are more reliably attributed to him, and for which we have extant Sanskrit texts and commentaries. In particular, I focus on Vasubandhu’s Thirty Verses (Triṃśikā), and his commentary on The Analysis of the Middle and Extremes (Madhyāntavibhāgabhāṣya). I further examine the extensive elaborations of these works by the commentator Sthiramati (6th cent.).2 Indeed, my analysis will center around Sthiramati’s comments, as they most explicitly address the issues of interest here—namely, whether phenomenal consciousness is real, and, if not, how the illusion of phenomenal consciousness might be generated.

In what follows, Section 1 rehearses some distinctions between types of illusionism and eliminativism about consciousness that will help us reconstruct Vasubandhu and Sthiramati’s views. Section 2 discusses why it would appear that Vasubandhu and Sthiramati (henceforth, V&S) regard consciousness (vijñāna) to be fundamentally real. Then, Sections 3 and 4 consider the case for interpreting V&S as illusionists. Section 3 elaborates their view that vijñāna is fundamentally a process of conceptual fabrication (vikalpa) called the “imagination of what is non-existent” (abhūtaparikalpa). This process is responsible for generating the mental appearances (pratibhāsa) of both objective representational contents and subjective representational vehicles like an internal self and representational mental states. V&S argue that the mental appearances of these entities are all illusory, because no awareness-state can actually represent something through appearing as it. They go on to claim that these dualistic mental appearances are generated by the activity of discursive attention (jalpamanaskāra), that is, an activity whereby the mind registers, classifies, and fixates upon some entity under the influence of inner speech. V&S’s story of how mental appearances are fabricated thus resonates in several ways with solutions to the Illusion Problem proposed by illusionists like Daniel Dennett (2017), Keith Frankish (2016a), and François Kammerer (2021). All these theorists can be read as converging on the idea that the illusion of mental appearances is cognitive or linguistic in nature, and constructed through a process of introspective misrepresentation.

Section 4 considers whether illusory mental appearances or anything else in V&S’s theory of mind can still have a real phenomenal character. I suggest that a pratibhāsa in their account wouldn’t instantiate the sorts of phenomenal properties—viz., intrinsicness, ineffability, and infallible, immediate introspectibility—that are typically attributed to conscious qualitative states. I further argue that a pratibhāsa wouldn’t have any fundamentally real phenomenal character of its own if this phenomenal character is exhausted by an appearance-based representational character that is ultimately imaginary. Next, I raise two other possible instances of real phenomenal consciousness within V&S’s theory of mind: hedonically valenced sensory experience, or “feeling” for short (vedanā), and the mind’s “intrinsic luminosity” (prakṛtiprabhāsvara). I try to make plausible an illusionist reading for each of these possibilities. Section 5 takes stock, concluding that that V&S really seem to be illusionists about phenomenal consciousness, particularly if phenomenality and intentionality are essentially linked. I end with some brief reflections about the historical and philosophical implications of V&S’s apparent illusionism.

1. Illusionist Preliminaries

There are two ways in which V&S could be illusionists about phenomenal consciousness, depending on whether they’d reject the reality of one or both possible conceptions of it. A maximalist conception holds that a mental state is phenomenally conscious only if it instantiates what Frankish (2012) calls “classic qualia,” which are the qualitative properties that have traditionally posed the most trouble for physicalism: intrinsicality, ineffability, and subjectivity. Starting with the latter properties, ineffability entails that no amount of linguistic description can convey to another subject what it is like to experience a phenomenal state. The feature of subjectivity can be broken into further features like privacy and direct knowability (Dennett 1988): subjective qualia are private insofar as they are only knowable through a subject’s having an immediate and perhaps infallible acquaintance with them.

A mental state’s phenomenal properties are thought to be intrinsic or non-relational in the sense that their qualitative character doesn’t constitutively depend on factors outside the mental state, such as an extramental environment, the propositional attitudes (e.g., belief, doubt) one may take toward that state, or the state’s causal origins. One qualification: If a mental state’s representational content does not determine what it is like to experience that state, then the state could have intrinsic phenomenal properties that are non-intentional. On the other hand, some phenomenal realists are phenomenal intentionalists or narrow representationalists (e.g., Graham & Horgan 2008; Chalmers 2010) who claim that a state’s phenomenal properties determine what that state represents, in which case the state could have intrinsic qualia that are also intrinsically intentional. If V&S are fully committed to illusionism, then they should deny the reality of both intrinsically intentional and non-intentional phenomenal states.

As for the minimalist conception of phenomenal character, it isn’t necessarily committed to ascribing any of the above properties to phenomenally conscious states. On this conception, a mental state is phenomenally conscious just if there is “something it is like” to undergo it—it isn’t essential that the state also be intrinsic, ineffable, infallibly introspectible, and so on. Thus, one can be a “weak illusionist” (Frankish 2016a: 15–16) who accepts the existence of phenomenally conscious states but denies that these states have “classic” qualitative properties, appearances perhaps to the contrary. Here, we are concerned with whether V&S would be “strong illusionists” who are anti-realists about the minimalist conception as well. Strong illusionists think it is an illusion that there’s really something it is like for mental states to phenomenally appear or be experienced in any sense.

One more distinction helps clarify whether V&S are anti-realists about phenomenal consciousness. This sort of anti-realism has been known as “eliminativism,” but Liz Irvine and Mark Sprevak (2020) point out that there are two distinct senses in which one might seek to “eliminate” phenomenal consciousness. “Entity eliminativism” about x claims that x does not exist, whereas “discourse eliminativism” about x claims that we should remove x-related talk or concepts from serious scientific discourse and practice. While these two forms of eliminativism can be adopted together, they need not be. For example, a theistic scientist may think that God exists, but wish to eliminate discourse about God within the study of evolutionary biology. Conversely, the distinction between inorganic and organic compounds might ultimately be unreal, but chemists may wish to retain the distinction for pragmatic purposes. As for V&S, they might replace “inclusion in serious science” with “pragmatic or soteriological efficacy” as their criterion for retaining or rejecting discourse about some entity. For instance, Vasubandhu in his Abhidharma work acknowledges that talk about persons is usefully retained in the context of ordinary activity and discourse, even though persons should be eliminated from an ontology of ultimate reality (e.g., see Sangpo 2012: 2551). Likewise, V&S’s assertions about the existence or non-existence of consciousness may be context-dependent: As an example, one should adopt the belief that consciousness alone exists in order to counteract one’s attachment to external objects—however, at a later stage on the Yogācāra path to liberation, one should also abandon this belief due to its being a potential source of attachment itself.3 For V&S to count as strong illusionists, they should be entity eliminativists, and not merely discourse eliminativists, about phenomenal consciousness.

Finally, the mark of an illusionist is that they offer a solution to the Illusion Problem, that is, they give some explanation of how the illusion of seeming to have phenomenal consciousness arises. Illusionists have typically identified the source of error to be some mechanism in the brain for introspectively misjudging its own physical/functional states as having a phenomenal character. A visual perception of a red object does not actually have a qualitative property of phenomenal redness—rather, the perceptual state only has the “quasi-phenomenal” property of being disposed to trigger a false introspective judgment that it has the property of phenomenal redness (Frankish 2016a: 15).

Dennett similarly claims that qualia only exist as fictional intentional objects of our judgments about them; failing to exist independently of such judgments, it can’t be that I first experience a subjective red quale and then introspectively judge myself to have that experience of red. As he writes,

It is your ability to describe ‘the red stripe,’ your judgment, your willingness to make the assertions you just made, and your emotional reactions (if any) to ‘the red stripe’ that is the source of your conviction that there is a subjective red stripe. (2017: 359)

As to why we are mistakenly convinced that we have consciousness in the first place, Dennett pinpoints cultural evolution as installing in our brains “a bounty of words and many other thinking tools” that in turn bestow us with “a user illusion,” or a first-person perspective on our neural events that we interpret as presenting both external properties and our internal responses to them (2017: 370). According to him, the human brain’s representation of itself as having a first-personal perspective on the world evolved out of the need to communicate with others for the sake of social coordination. One becomes introspectively aware of one’s own mind through learning to communicate with others, and this introspective awareness leads us to attribute our brain states with simple phenomenal properties out of our need to standardize communication, or “compare notes,” across minds. In sum, Dennett claims, “It is like something to be you because you have been enabled to tell us—or refrain from telling us—what it’s like to be you” (2017: 344–45).

Frankish and Kammerer also explain how the brain introspectively judges its states to have phenomenal properties by invoking certain evolved conceptual abilities for mental state attribution. Developing a suggestion by Frankish (2016a: 36–37), Kammerer (2021) takes introspective judgments of phenomenality to involve the application of a hybrid phenomenal concept made up of two conceptual components. The judgment that misrepresents a visual perception of a red object as being a phenomenal experience of red contains two concepts. One is a recognitional concept <red> that picks out the sensible property of redness belonging to some external surface. The other is a theoretical concept <phenomenal experience> whose content is determined by sub-personal principles encoded in an innate theory-of-mind module used for attributing mental states to oneself and others. Briefly, these implicit theoretical principles hold that a certain state of affairs can appear to a subject in virtue of the subject’s being internally and receptively/passively affected by that state of affairs—this internal receptive affection being what it is to have an “experience.” Further principles dictate that a state of affairs appears through an experience insofar as the experience maximally resembles it; and, that experiences are individuated by what they make appear. Together, these theoretical principles enable a subject to judge that it or some other subject has an experience of an object: When I have a receptive affection resembling a red rose that makes the red rose appear to me, I can form an introspective judgment using phenomenal concepts to represent myself as having an experience of a red rose. Of course, the illusionist claims that nothing genuinely satisfies the principles underlying our theoretical concept <phenomenal experience>—that is, “nothing in reality is such that it is a receptive affection which makes things appear to a subject in virtue of its maximal resemblance to the appearing thing” (Kammerer 2021: 856).

Abstracting away from the details of these particular theories of introspective illusion, there are some basic commitments that we might expect V&S to share if they are indeed illusionists about phenomenal consciousness. V&S should be strong illusionists and entity eliminativists—not just discourse eliminativists—about even a minimalist conception of phenomenal consciousness, and they should offer an explanation of how mental states falsely appear to have “something it is like” to undergo them. Though, before considering whether V&S are actually illusionists, we should first address why it would seem that they are not.

2. Why It Would Seem that V&S Are Phenomenal Realists

Just as illusionism about phenomenal consciousness is deeply counter-intuitive, so too is an interpretation of Vasubandhu’s Yogācāra works as being congruent with the contemporary illusionist program of denying the fundamental reality of phenomenal consciousness and explaining the illusion of phenomenality in purely physical terms. A quick overview of these works will suggest why. The main argument put forward by Vasubandhu’s Twenty Verses is this: Everything in the world is “nothing but mind” (cittamātra)—“mind” here being a synonym of “consciousness” (vijñāna)—because the seemingly mind-independent objects that appear to awareness are in fact non-existent, as illustrated by obvious cases of illusion like the floating hairs that appear in the vision of someone suffering from myodesopsia. Hence, a central task of the Twenty Verses is to show that objects do not exist outside of the mind’s representation (vijñapti) of them. Vasubandhu attempts to accomplish this task by arguing that the spatiotemporal, intersubjective, and causal regularities underlying phenomenal appearances are all parsimoniously explained without positing real mind-independent objects. He also argues that the notion of a physical object composed of atoms is incoherent. In the Thirty Verses, Vasubandhu claims that the self and all other entities are nothing but “metonymic” constructions (upacāra), which are fictional notions projected onto an otherwise impersonal causal series of dependently co-arisen mental states called the “transformation of consciousness” (vijñānapariṇāma).

Sthiramati’s commentary highlights how Vasubandhu intends to give a privileged ontological status to consciousness. Sthiramati suggests that the Thirty Verses sets out to reject two extreme views that both mistakenly put consciousness on an ontological par with the objects of consciousness (vijñeya): one view takes the objects of consciousness to exist fundamentally (dravyataḥ) like consciousness, whereas the other takes both consciousness and the objects of consciousness to only exist conventionally.4 The former external realist view is false because none of the things that we can be conscious of actually exist outside of consciousness. All these purportedly mind-independent things instead have only a nominal existence and are essentially imaginary—they don’t exist apart from how the mind conceptually fabricates them. Nevertheless, this denial of mind-independence to the objects of consciousness doesn’t commit V&S to a pan-fictionalist view that everything which exists is merely a conceptual fabrication and only conventionally real. The dependently co-arisen process of consciousness must exist fundamentally because it’s the causal basis of all conceptual fabrications.5

So, it seems curious that Garfield would cite Vasubandhu as an advocate of illusionism if one of Garfield’s main motivations for rejecting phenomenal realism is a desire to avoid a commitment to idealism. He claims that by believing there are phenomenal appearances with which a subject is immediately acquainted, we are led down a slippery slope to the belief that qualia mediate our perceptual contact with the external world. Once there, we naturally slide into the idealist belief that we only have evidence for the existence of qualia and not of the external world itself. To avoid this slide into idealism, it’s better to acknowledge that our perceptual experience is transparently intentional, presenting only the properties of objects rather than the properties of subjective qualia (Garfield 2016: 74–75). Of course, Garfield recognizes that Vasubandhu himself rejects the transparency of experience, along with any naïve realist claim that we directly perceive external objects as they exist in themselves—for Vasubandhu, it’s an error to take our “sensory systems as transparent windows onto a world existing with sensible properties independent of our mode of apprehension” (Garfield 2016: 80). But then, Vasubandhu’s thesis that we lack epistemic access to an apprehension-independent world appears to contradict Garfield’s own appeal to experiential transparency in rejecting phenomenal realism.

Another seeming incongruity within Garfield’s illusionist interpretation of Vasubandhu concerns his allegation that phenomenal realism is committed to the Myth of the Given, defined as

the myth that there is some level of our experience that is immediate, immune from error, given to us, as opposed to constructed, and that that level of experience constitutes the foundation or transcendental condition of the possibility of knowledge of anything else. (Garfield 2016: 81)

The problem is that V&S intend their arguments to culminate in a type of transcendent or supramundane knowledge which should qualify as mythical under Garfield’s definition. This knowledge consists in an immediate, non-conceptual vision of the ultimate reality or “absolute nature” (pariniṣpannasvabhāva) of all dependently co-arisen entities. Specifically, one sees that all entities are essentially devoid of a fictitious subject-object duality. That is, one directly realizes that nothing actually has the nature of being an external object or internal subject which could be cognitively grasped by, and thus become the object of attachment for, an impersonal series of dependently arisen states of consciousness. This direct perception in which all things are given just as they are (tathatāmātradarśana), free from conceptual fabrication (nirvikalpa), in turn serves as the basis for an enlightened being’s subsequent knowledge of and engagement with the world.6

All that said, there is still another apparent contradiction to confront. On the one hand, V&S are evidently committed to the fundamental reality of consciousness; on the other, they also claim that consciousness is itself a “metonymic” construction (upacāra), and that the “transformation of consciousness” onto which all such constructions are projected is actually just a dependently arisen process for generating mental fictions. Indeed, one basic insight granted by the transcendent vision of ultimate reality mentioned above is that consciousness, like most everything else, is an illusion akin to a magic trick or mirage, in that it ultimately does not exist in the way that it appears.7 Here, two questions arise: (1) Given that V&S identify consciousness as a process of conceptual fabrication (vikalpa) and the “imagination of what is non-existent” (abhūtaparikalpa), what does it mean to say that consciousness is constitutively a process that generates illusions about itself; and (2) do the illusory appearances that consciousness generates nonetheless have a real phenomenal character? Which leads to a third question: if the answer to (2) is negative, then in what sense is vijñāna in V&S’s “vijñānavāda” really a state of consciousness at all? We will consider these questions in the next section.

3. Why It Really Seems that V&S Are Illusionists

3.1. Types of Unconscious and Conscious Vijñāna

We should first unpack V&S’s many-layered notion of vijñāna in order to determine whether anything about their notion entails a commitment to phenomenal realism. To start, the so-called “transformation of consciousness,” which constitutes our mental lives and the worlds we experience, is understood as a process of causal interactions between three kinds of awareness. (Though I’ve been translating vijñāna as “consciousness” thus far, I’ll now shift to a more neutral translation as “awareness,” because we will soon see why there may be states of vijñāna that are unconscious.) At the root of this process is the “storehouse awareness” (ālayavijñāna), which is said to store dispositional memory traces (vāsanā) left by past awareness-states and karmic acts. In turn, these memory traces serve as “seeds” (bīja) whose maturation yield the consequences of past karmic acts in the form of future awareness-states of a similar type. One kind of awareness-state that emerges from the activation of memory traces in the storehouse awareness is “occurrent awareness” (pravṛttivijñāna), consisting in five forms of external sensory awareness and a sixth form of cognitive awareness. The other kind of awareness is often known as the “afflicted mind” (kliṣṭamanas), but we might also call it “afflicted thought,” since this type of cognitive awareness essentially involves thinking (manana). Under the influence of egoistic craving and cognitive delusions, the afflicted mind entertains de se thoughts about the storehouse awareness that represent it as being an individual self and owner of mental events, actions, and external objects.8

Would V&S consider any or all of these types of awareness to be phenomenally conscious? First, there is reason to suspect that they take at least ālayavijñāna to be a form of unconscious awareness present in every moment of one’s unenlightened existence (cf. Waldron 2003). Among the main functions of the storehouse awareness is to represent internal “aggregates of appropriation” (upādāna) as well as a surrounding external environment (bhājanaloka). The appropriation-aggregates are all the elementary mental states to which one erroneously clings as constituting one’s apparent psycho-physical individuality. According to V&S, these states are essentially dispositional memory traces left by past conceptual representations of imaginary entities (like the self and mind-independent objects), as well as by one’s attachment to such entities. When these memory traces are activated, they in turn generate further representations of conceptual fictions and corresponding attachments. For our purposes, the important point is that Vasubandhu takes the storehouse awareness’s representational states to be “unknown” or “unperceived” (asaṃviditaka).9 Sthiramati explains that the appropriation-aggregates—that is, the memory traces left by the representation of conceptual fictions—remain unknown insofar as it is impossible to have an experience that demonstratively identifies them; put another way, one can never mentally point to a memory trace. The basic reason Sthiramati gives is that the appropriation-aggregates are “extremely subtle,” making them difficult to be ascertained even by wise scholars who, presumably accepting the existence of such aggregates, would know what to look for.10

Another relevant feature of storehouse awareness is that it is only ever associated with a neutral hedonic valence, unlike the occurrent forms of sensory and cognitive awareness that can feel pleasurable and painful as well as neutral. Like all states of awareness, feelings (vedanā) are also said to have two essential features: 1) an “objective support” (ālambana), or some object which appears in awareness by virtue of causing that awareness; and 2) a representational aspect (ākāra), or a mode whereby that object is apprehended as being a certain way. Sthiramati claims that, whereas painful and pleasurable feelings have a determinate, discernible objective support and representational aspect, the neutral feelings associated with storehouse awareness only have an objective support and representational aspect that is indeterminate or indiscernible (aparicchinna).11 (This indeterminate and indiscernible representational character also belongs to the storehouse awareness’s representation of an external physical environment.12) Furthermore, neutral feelings leave no subjective trace in the sense that they don’t give rise to any desires or attitudes of craving and aversion, whereas Vasubandhu’s “desire view” of pain and pleasure defines these feelings in terms of the desires they respectively generate.13

A last piece of evidence for the unconscious character of storehouse awareness is that it is present in otherwise “mindless” (acittaka) states which lack most or all other forms of mental activity. These states include two rarified forms of meditative attainment—the attainment of non-ideation (asaṃjñikasamāpatti) and the attainment of cessation (nirodhasamāpatti)—as well as two more mundane states of extreme sleepiness and syncope. Each of these states can be called “mindless” insofar as they lack any occurrent cognitive awareness or thought (manovijñāna), but they also involve the absence of occurrent sensory awareness.14 Indeed, since death was partly defined by earlier Yogācāras as the departure of awareness (vijñāna) from the body, one of the main reasons they postulated the existence of storehouse awareness was to explain why these states of almost total mental inactivity do not result in death (Schmithausen 1987: 19–20).

We can thus surmise from the storehouse awareness’s introspective invisibility, its lack of subjectively discernible representational content and hedonic valence, as well as its similarity to states of mental inactivity, that it is a kind of subconscious awareness which lacks a phenomenal character—in short, it should not “feel like” anything to undergo a state of storehouse awareness on its own. What, then, about the occurrent states of sensory and cognitive awareness that populate our ordinary lived experience of the world? If V&S would take ālayavijñāna to be a form of unconscious awareness, then it would be reasonable to identify their notion of conscious awareness with the six types of pravṛttijñāna. These occurrent states are characterized as representations of intentional objects (viṣaya) due to their having the appearance of those objects.15 Given that phenomenality is etymologically/conceptually rooted in the notion of appearance (phainomenon in Greek), the mental states in which objects appear would be natural candidates for having a phenomenally conscious character. Though, one immediate caveat is that if we are correct in taking the storehouse awareness to be phenomenally unconscious, then an awareness’s merely having the appearance of some represented object cannot be sufficient for that appearance to be phenomenally conscious, since the storehouse awareness is also said to represent its objects in virtue of appearing as them.16 More significantly, once we examine in the next section V&S’s account of how mental appearances (pratibhāsa) arise within even occurrent sensory and cognitive awareness-episodes, we find them converging on the illusionist conclusion that the way these appearances seem to exist is not how they actually are.

3.2. The Fabrication of Mental Appearances

When V&S defend the thesis that everything in the world is nothing but mind or awareness, what they are defending is the fundamental reality of vijñāna understood as the “imagination of what is non-existent.” This thesis entails that everything which appears in awareness to exist—from subjective selves to objective entities—is just a conceptual fabrication of the mind.17 These conceptual fabrications are divided into two types: one involves the conception of some apprehended or grasped entity (grāhyagrāha), and the other involves the conception of an apprehending or grasping entity (grāhakagrāha). Sthiramati specifies that these two conceptions take the form of determinate judgments (adhyavasāya/niścaya). To conceive of an apprehended object is to judge that something exists independently of the mind when it’s in fact fictionally superimposed onto one’s own stream of mental states. To conceive of an apprehending entity is just to be certain that an apprehended object is indeed apprehended, known, or made into an object of awareness, by an awareness-state.18 In the Analysis and its commentaries, the two types of conceptual fabrication are further subdivided into four types of mental appearance (pratibhāsa).19 Apprehended entities appear either as external objects or embodied sentient beings, while apprehending entities appear as either an internal self or a representational state (vijñapti). Every state of awareness arises with one of these four kinds of appearance.

Sthiramati adds important clarification about how mental appearance and mental representation are essentially linked. He explains that an awareness-state possesses an appearance (pratibhāsa) or “reflected image” (pratibimba) of some object insofar as it takes on the “aspect,” “form,” or “structure” (ākāra) of that object.20 The object has its own form or structure, and so an awareness-state would appear as or represent that object through taking on a resembling form or structure. If the object that the awareness-state appears to resemble is also one of the causal conditions that generated the awareness-state, then it is distinguished as the awareness’s “objective support condition” (ālambanapratyaya).21 Apparent resemblance is a necessary requirement for being an objective support; something may causally co-vary with an awareness-state, but it won’t qualify as an objective support if the awareness-state has a different appearance.22 Sthiramati further glosses the representational aspect (ākāra) belonging to awareness as a “mode of apprehension” (grahaṇaprakāra) through which an awareness’s objective support is represented to exist in a certain way.23

Overall, Sthiramati’s account of mental representation seems to align with that of phenomenal intentionalism in taking an awareness-state’s representational content to be determined by the state’s appearance-content. While a mental state has the appearance of a certain object through its having a representational aspect, that aspect represents or apprehends some aspect of an object in virtue of resembling the object’s appearance. That being so, we might take Sthiramati’s view about the connection between mental appearance and representation to be expressed by David Chalmers’s statement that “consciousness and intentionality are intertwined, all the way down to the ground” (2010: 371). A phenomenal realist reader could then say that, for V&S, what it is like to undergo a conscious awareness-state is grounded in what the awareness-state appears like.

As it turns out, though, V&S will argue that the appearance-based representation or apprehension of awareness-independent objects is ultimately impossible. Instead, they think that no mental appearance has representational content or serves as a representational vehicle apart from being erroneously imagined as such. This suggests the following illusionist conclusion, to which we’ll return in Section 4.1: If awareness-states would have a phenomenal character in virtue of a pratibhāsa they contain, and this phenomenal character would be bound up with the representational character of a pratibhāsa, then since the representational character of a pratibhāsa is an illusory fabrication, the phenomenal character of awareness-states is also an illusory fabrication.

Briefly, one reason why V&S deny that awareness-states have a genuine representational character is that these states never really have an objective support. They give a range of arguments to show that all awareness-states are like a dream or sensory hallucination, in that the objects which appear in such states need not, or indeed cannot, have causally generated those states.24 Even if some object were actually the cause of an awareness, the Yogācāra assumption of radical momentariness (von Rospatt 1995: 67–93) means that the object will have ceased to exist by the time the awareness of it arises—an awareness-state will hence come into being with an appearance of something that doesn’t exist.25 If a representational state (vijñapti) is defined as the apprehension of an awareness-independent object (artha), then representational states are impossible in the absence of such objects.26 Strictly speaking, an awareness-state wouldn’t be a representation or appearance of anything.

Still, a phenomenal realist reader could allow even the non-veridical awareness of a non-existent objective support to have phenomenal intentional content, since such awareness nonetheless possesses a representational aspect with a corresponding mental appearance. In response, V&S will point to a basic incoherence within the very notion of a mental appearance that veridically or non-veridically represents some object in virtue of resembling the object’s form. Notably, Vasubandhu’s reason why the appearances of objective entities cannot be genuine appearances of awareness-independent objects is that these appearances are in fact “formless” (anākāra).27 According to Sthiramati, the problem is that an awareness-state’s representational aspect or form, being the mode through which something is apprehended or represented, is essentially a subjective apprehending entity (grāhaka); yet, external objects and embodied sentient beings are supposed to appear as being objective apprehended entities (grāhya). That is to say, if an awareness-state is going to represent an objective entity, it must transparently appear in the form of an object that is external to or independent of the awareness-state. However, the very nature of a mental appearance is that of a subjective, awareness-dependent mode of representation. Thus, if an awareness-state is to represent an external object through an ākāra, then it must appear to be something that it cannot be. In this way, the appearance in awareness of apprehended objects is an illusion: Because these appearances are subjective modes of apprehension, they cannot exist in the way that they supposedly appear to exist (i.e., as appearing to be apprehended, awareness-independent entities).28

If awareness-states can’t really appear as or represent objective entities, what about their appearing to be subjective representational states themselves? Vasubandhu claims that these appearances also cannot represent what they purport to represent, not because they are said to be “formless” like the appearances of apprehended objects, but because they are appearances which are false or misleading (vitathapratibhāsa).29 The six types of occurrent sensory and cognitive awareness are what take on the appearance of being representational states (vijñapti) that apprehend intentional objects. (States of afflicted thought are what take on the false appearance of an internal self.) Given the Yogācāra denial of mind-independent objects, an awareness-state’s appearance as being a representational state or vehicle will be misleading in the strict sense if there actually are no objective entities for it to apprehend or represent.

But, Sthiramati suggests another more subtle explanation for why an awareness-state’s appearance of being a representational state is misleading—namely, the awareness-state itself is a misleading objective support for that appearance. He illustrates with an example: Suppose that you are walking through a dense jungle, and you hear some rustling in the bushes, but out of fear you immediately think that the rustling sound is actually the growl of a tiger. The rustling sound was real, but the sound misled you into imagining it as having been something it wasn’t. Similarly, there really are occurrent sensory and cognitive mental states belonging to the causal series of vijñāna-transformations, but they mislead you into thinking that they actually represented or possessed the appearance of some objective entity. (As we’ve seen above, it’s incoherent for any awareness-state to ever appear as an awareness-independent object.) Occurrent awareness-states are instead only imagined as appearing to be representational states by other (themselves imaginary) representational/conceptual states (vijñaptyantara; vikalpāntara).30 Hence, even states of vijñapti—to which all mind-independent entities are reduced within the overarching Yogācāra view that everything is “nothing but mental representations” (vijñaptimātra)—are illusory. Representational states don’t exist in the way that they appear, that is, as appearance-based representations of something existing apart from themselves.

It therefore seems for V&S that every mental appearance—which is to say, every appearance of apprehended objective and apprehending subjective entities, or of represented contents and representational vehicles—is an illusory conceptual fabrication generated by vijñāna qua the imagination of what is non-existent. Sthiramati makes explicit that these appearances are conceptual fabrications (vikalpa) in virtue of their not having an objective basis or support in awareness-independent reality. All appearances of purportedly awareness-independent entities (artha), in any of the four forms described above, are in fact erroneously superimposed onto an awareness-state by another awareness-state. Within what Yogācāras call the “three realms” of unenlightened existence, all states of awareness (citta/vijñāna) arise with some representational aspect or appearance of an objective support—to that extent, though, all these states are illusory.31 The basic insight of V&S’s “vijñaptimātra” view is that all entities—including representational states themselves (vijñapti)—are empty of the imagined nature that is superimposed on them by other representational states.32

3.3. Discursive Attention as the Source of Mental Appearances

How then do mental states come to be imagined as having a representational aspect through which they fictitiously appear as objective and subjective entities? Put another way, how are mental states misrepresented as having an appearance-based representational character? The answer V&S draw from Yogācāra texts is that the appearance of subject-object duality is generated through the activity of “discursive attention” (jalpamanaskāra).33

Sthiramati explains that the conceptually constructed appearance of apprehended objects and apprehending subjects arises from discourse (jalpa), which he glosses as the act of cognitive/linguistic “noting” (abhilapana). The notion of abhilapana has a number of importantly interrelated meanings. It can mean “talking” or “speaking” about something, with this meaning becoming more relevant in later Yogācāra definitions of conceptualization as that kind of awareness which is associated with a linguistic expression (abhilāpa).34 Within an Abhidharma framework, V&S use abhilapana to define “mindfulness” (smṛti), which is one of ten mental activities that accompany every state of awareness. In that context, abhilapana is understood as the repeated recalling of a previously experienced object, which concentrates the mind on the object and prevents distraction by other objects.35 Another suggestive gloss on the term is provided by Yaśomitra in his commentary on Vasubandhu’s Abhidharmakośabhāṣya and its discussion of insight’s (prajñā) role in the application of mindfulness. In whatever way an object is observed through some form of testimony-based, analytic, or meditative insight—say, as being momentary or conducive to suffering—it is in that very same way that the object is “noted,” or as per Yaśomitra’s gloss, “taken up” (udgṛhyate) through mindfulness.36 This idea of “taking up” is also used in V&S’s characterization of cognitive labeling (saṃjñā) as the “taking up” of an intentional object’s distinguishing mark. “Taking up” is more clearly characterized as the mind’s grasping or seizing upon—which is to say, abstracting and singling out—some specific feature of an object for the sake of discerning that object’s identity.37 Returning to jalpamanaskāra, the activity of discursive attention is said to be conditioned by cognitive labels that either arise from or give rise to linguistic expression.38

With this range of meanings for abhilapana in view, we can understand the act of noting as a cognitive/linguistic act of repeated, fixated ideation whereby mental events are labelled and classified in a basic sense as being either self or other, subjective or objective. The character of repetitive fixation is also evident in Sthiramati’s general definition of attention (manaskāra) as bringing about the mind’s retention of an objective support, where retention over time consists in the mind’s turning toward that objective support over and over again.39 The repetitive nature of discursive attention is also due to its arising from the memory traces left by past discursive mental activity. An act of discursive attention which generated the conceptually constructed appearances of subjects and objects in turn “nourishes its own seed,” that is, it reinforces the memory traces of its activity and conditions the arising of further discursive attention and conceptual fabrication. Stemming from the “seeds” or dispositional traces deposited in the storehouse awareness, the continual activity of discursive ideation and attentional conditioning is thus rooted in the deepest, subliminal level of the mind. As such, discursive attention is equated with the mind’s “dependent nature” (paratantrasvabhāva), or its fundamental nature as a dependently arisen causal process which fabricates the illusion of objective and subjective entities.40

All that said, what exactly gives discursive attention the power to generate the illusory appearance of objects and subjects? Why should subliminal inner speech and the attentional activity it conditions be responsible for our basic sense of being subjects that experience an external world? We can fill in some details of the story by considering Sthiramati’s brief discussions of vitarka and vicāra, or initial and sustained thought. According to Vasubandhu in the Abhidharmakośabhāṣya, these two specific types of mental discourse are forms of conceptualization or ideation that are intrinsic even to sensory awareness-states.41 Sthiramati explains that, from the first moment of sensory awareness, mental discourse in the form of vitarka is applied such that the mind attentionally selects a cognized object and seeks to discern its identity, asking “What is this?” This is a moment at which the mind starts to describe or “narrate” the object, giving it a discursive life of its own. Then, more subtle but sustained discourse in the form of vicāra retains this experienced object and recognizes it as being a certain way. Both vitarka and vicāra are said to provide the basis for “comfortable or uncomfortable abiding” (sparśāsparśavihāra), that is, the pleasure or pain that follow from recognizing the attended object as being desirable or undesirable. Being themselves a kind of mental volition, these moments of vitarka and vicāra set the mind into motion on the path of some karmically valenced pattern of activity.42

We therefore have a rough outline of how, through the activation of these forms of discursive mental activity at the onset of any given experience, awareness takes on the bifurcated appearance of objective representational contents and subjective representational vehicles. Habituated by the memory traces left by past mental discourse, the attentional activity of discursive noting fixates on some virtual aspect of experience and labels it as having an awareness-independent existence. Taking on a nominal existence as an extramental object, the virtual aspect would in some sense become both alienated from awareness, and nonetheless subject-implicating—for example, an awareness which appears as an apprehended object would further appear to afford an apprehending subject with possibilities for intentional action and desire-satisfaction. Once fabricated, the dualistic appearances of objective and subjective entities in awareness would thereby serve as the foundation for the attachment to these notional, nominal entities.43 The repeated and fixated character of discursive attentional activity toward such entities ultimately perpetuates the cyclical process of fabrication and attachment for as long as this activity does not cease.

3.4. A Yogācāra Answer to the Illusion Problem

V&S’s account of discursive attention offers an explanation of how the illusion of mental appearances arise, or how it is that states of awareness take on the appearance of represented objects or representational vehicles despite these states not actually existing in the way that they appear. To determine whether this account of discursive attention could also be addressing the contemporary Illusion Problem of how the appearance of phenomenal consciousness arises, we can note its many parallels with recent illusionist answers to the Problem. Of course, we have to look past the obvious differences in philosophical contexts and motivations: To put it simply, most contemporary illusionists are physicalists who seek to eliminate qualia from scientific discourse about reality, whereas V&S deny the existence of mind-independent physical objects on the way to cultivating a non-discursive, transcendent vision of reality. Regardless, there is much common ground between their respective explanations of how illusion of mental appearances originates.

For example, one basic point of agreement is that mental appearances conceal the complex causal processes underlying their construction. Briefly compare V&S’s account with another recent illusionist-adjacent account, namely Michael Graziano’s (2016; 2019) attention schema theory of consciousness. For V&S, the awareness of apparently apprehended objects and apprehending subjective entities arises from the activity of conceptual fabrication qua discursive attention. These apparent entities are imaginary fictions that are falsely superimposed onto the fundamentally real causal process which generated them, viz., the “transformation of vijñāna.” As for Graziano, he posits that conscious awareness is grounded in an attention schema, which is an internal model of the brain’s distribution of covert attention to external objects as well as stored self-specifying information (e.g., the body schema, autobiographical memories). This model leaves out information about the complex causal mechanisms involved in the selective processing of these objective and subjective representations. In so doing, the attention schema gives the brain a simplified description of its own attentional resources that allows for their efficient monitoring and control. More importantly, the attention schema gives the brain an impression not simply that there is some object in the environment, nor just that there is a psychophysical self, but that there also is an awareness of an external object on the part of a self. That is because, in Graziano’s view, awareness largely tracks attention—a subject typically has awareness of some object when, having directed attention to that object, the subject’s corresponding cognitive and behavioral responses are enabled. But, since the attention schema leaves out information about the underlying neurocomputational machinery, all the brain can report through accessing the attention schema is that it is a self who has an awareness or “mental grasp” (2019: 102) of an object (i.e., what V&S would call a “grahaṇa”). Like the brain’s object- and self-models, the awareness constituted by an attention schema is nothing more than a construct fabricated through abstracting away from the “base reality” (2020: 230) of the complex physical processes of attentional activity—this is why conscious awareness-states do not seem to have any physical properties of their own.

While Graziano shies away from the “illusionist” label because the attention schema is a useful caricature of real attentional mechanisms, it is still true that the attention schema distorts physical reality as it actually is, forming the subjective impression of there being a “magical mental possession” with a non-physical essence (2020: 228). Put in Yogācāra terms, this “subjective essence” engendered by the attention schema is merely imaginary (parikalpitasvabhāva); to the extent that the causal/computational processes involved in attention (part of the mind/brain’s paratantrasvabhāva) do not actually have such an essence, the appearance of this essence in the mind is an illusion. For Graziano as well as other contemporary illusionists (Humphrey 2011; Frankish 2016a; Dennett 2016), the mental appearance of subjective experiences that take possession of objects is akin to a magic trick: Brains attribute themselves with these mysterious non-physical states on the basis of a sleight-of-hand played by inherently inaccurate (and thus deceptive) internal representations that conceal their own causal origins (Graziano et al. 2020: 155, 158). So also for V&S: The appearance of subjective states of mental possession (along with the objective entities they take hold of) are like magical illusions in that these apparent states do not actually exist, whereas the real causal processes that generate these appearances do not exist in the way that they appear.44

V&S would also agree with illusionists like Dennett and Kammerer that the fiction of mental appearances is conceptual/linguistic in nature, and deeply rooted in our evolved cognitive tendency to introspectively misjudge our own mental states. Again, V&S think that awareness-states are endowed with the conceptually fabricated appearances of objective and subjective entities due to the activity of discursive attention conditioned by inner speech. Being a specific state of storehouse awareness, this activity functions at a subpersonal and subliminal level of mind, where it is conditioned by dispositional memory traces left by relevantly similar discursive/conceptual activity—specifically, acts of cognitive/linguistic labeling.45 Since these dispositions for constructing and clinging to mental appearances lack a distinct beginning within the perpetual cycle of rebirths, the activity of discursive attention can be plausibly interpreted as being innate in some sense, while also being developed and reinforced over a great number of lifetimes. If this story about discursive attention being conditioned by cognitive/linguistic activity in past lives were naturalized, then it would probably sound like Dennett’s suggestion (2017: 370) that sociocultural evolution is responsible for endowing our minds with an innate linguistic/cognitive architecture for constructing a first-person perspective, from which we further misrepresent the brain’s non-phenomenal states as appearing to present external objects and internal subjective responses to them. V&S additionally share with Dennett an emphasis on the role of inner speech in enabling conscious (i.e., quasi-phenomenal) representational states. Dennett thinks that there is no “consciousness of a stimulus in the absence of the subject’s belief in that consciousness” (1991: 132). So, by enabling a subject to form metacognitive thoughts about its representational states, inner speech enables a subject to have the sense that there is “something it is like something to be.” Absent “the ability to talk to yourself silently . . . the contents of events in one’s nervous system” would only have an “unconscious” effect on one’s cognition and behavior (Dennett 2005).

To highlight one further point of agreement with contemporary illusionists, V&S might broadly concur with Kammerer’s account of the theoretical principles that underpin the innate activity of conceptually fabricating mental appearances. According to Kammerer, an innate, sub-personal theory-of-mind module determines that some state of affairs x phenomenally appears to a subject just when the subject undergoes an internal state that has been receptively affected by x in such a way that the state maximally resembles x. These characteristics of the phenomenal concept <experience of x> align with the criteria assumed by V&S for an awareness-state’s taking something as its objective support (ālambana): an awareness is aware of an objective support x only if it has been both causally generated by x and takes on an appearance that resembles the form or structure of x.46 For the reasons discussed above in Section 3.2, V&S think that no state of vijñāna genuinely represents any objective or subjective entity in virtue of the mental appearance or representational aspect it contains. The ultimate emptiness of vijñāna, that is, its lacking any subject-object duality, is thus akin to the emptiness in Kammerer’s account of the concept <phenomenal experience>: there is no internal state of a subject whose appearance actually resembles the object that purportedly generated it. To that extent, V&S would agree with Kammerer that “nothing in reality is such that it is a receptive affection which makes things appear to a subject in virtue of its maximal resemblance to the appearing thing” (2021: 856).

4. Could Mental Appearances, or Anything Else, Still Be Really Phenomenal?

So far, we’ve seen how V&S’s treatment of vijñāna as the “imagining of what is non-existent” (abhūtaparikalpa) parallels the contemporary illusionist treatment of phenomenal consciousness as a “fiction of the impossible” (Humphrey 2011: 204). We’ve found within these treatments a shared commitment to two basic claims. First, mental appearances are illusions, insofar as the causal basis for these illusions does not exist in the way that it introspectively appears. Second, the illusion of mental appearances is primarily cognitive or linguistic in character.

Against the first basic claim, however, a phenomenal realist may reply that, regardless of their causal origins or representational inaccuracy, the existence of mental appearances entails the existence of real phenomenally conscious states. As David Rosenthal (2011: 431) puts it, “A state’s being conscious is a matter of mental appearance—of how one’s mental life appears to one.” Likewise, the phenomenal realist reader of V&S may argue that, even if mental appearances (pratibhāsa) are illusory because there are no awareness-independent objective or subjective entities that genuinely appear to an awareness-state, the bare existence of these illusory mental appearances (in at least occurrent, if not storehouse awareness-states) entails that V&S would be realists about phenomenal consciousness. In support, the phenomenal realist reader can cite Sthiramati’s statement that although vijñāna doesn’t really exist as the apprehender of any vijñāna-independent entities—such entities being imaginary and ultimately non-existent—it still really exists in virtue of its appearing as those entities. Given that vijñāna is what is fundamentally real, and is in its essence the generator of illusory mental appearances, the total non-existence of these illusions would absurdly entail the total non-existence of vijñāna and hence everything at all.47 So, if phenomenal realists would take mental appearances to be constitutively phenomenal—which is to say, anything that appears to the mind must phenomenally appear—then the fact that the mental appearances contained in vijñāna are massively illusory with regards to their representations of external and internal worlds doesn’t mean that their phenomenal character is itself an illusion.

4.1. The Phenomenal Properties of a Pratibhāsa: Classic, Diet, or Neither?

Here, the illusionist interpreter can ask: If V&S would claim that illusory mental appearances are still phenomenally conscious, then what sorts of qualitative properties would they take a pratibhāsa to have? Consider first whether a pratibhāsa would satisfy the maximalist conception of phenomenal character, and instantiate the properties of “classic qualia”: intrinsicness, ineffability, direct knowability, and privacy.

A mental state’s phenomenal properties are intrinsic if their phenomenal character doesn’t constitutively depend on some relation to factors outside that state. Of special relevance to whether a pratibhāsa has an intrinsic phenomenal character is whether phenomenal properties are also intrinsically representational. Because a mental state’s representational character is often taken to be determined by its causal relations to an external environment, and its functional profile determined by causal relations to a subject’s other mental and behavioral states, intrinsic phenomenal qualities have commonly been construed as being intrinsically non-representational/non-intentional as well as non-functional (Tye 2021). As Tim Crane explains, a non-intentional mental state “has no intentional structure: it is not directed on anything, it has no intentional object, no aspectual shape, and no distinction can be made between anything like mode and anything like content” (2001: 77). Now, while a pratibhāsa is never veridical (because there are no external objects to represent, and because it only represents internal mental states falsely), its presence in a state of vijñāna would clearly qualify that state as intentional on Crane’s definition. An awareness-state has a pratibhāsa in virtue of arising with an aspectual shape—that is, an ākāra—which in turn gives the awareness the appearance of being intentionally directed toward some objective support. Accordingly, all awareness-states possessing a pratibhāsa are dualistically structured such that there appears to be a representational content that is distinctly apprehended under some subjective mode of apprehension. So if intrinsic qualia must be non-representational, then an awareness-state can’t have an intrinsic phenomenal character due to possessing a pratibhāsa.

Yet, if we grant with phenomenal intentionalists that intrinsic qualia are also intrinsically representational, then a pratibhāsa’s intrinsic intentional content could have an intrinsic phenomenal character regardless of its non-veridicality. In other words, since a pratibhāsa is intrinsically an appearance of something (viz., an objective or subjective entity), the pratibhāsa will intrinsically have something it is like to appear, even though what appears doesn’t actually exist. Indeed, this intentionally structured phenomenal appearance should be all the more intrinsic to awareness if it cannot be grounded upon a causal relation with a mind-independent objective support.

In response, the illusionist interpreter can point out that although it may be essential to awareness-states that they have appearance-based content, this essential nature of awareness is nonetheless imaginary (parikalpitasvabhāva). V&S understand the representational nature of each mental appearance—that is, what it is that the appearance is of, and thus what the appearance “is like”—to be imagined or fabricated by other representational/conceptual states (vijñaptyantara; vikalpāntara).48 Stated this way, their view seems to align with how illusionists like Dennett take the purportedly intrinsic phenomenal qualities of experience to be constituted by a subject’s attitudes toward and reactions to the experience (1988: 533). In claiming that all mental representations are fabricated by other representations, V&S would hence arrive at Dennett’s position (1996: 50–54) that all intentionality is derived, meaning that no states have their representational contents intrinsically. That being so, an awareness-state’s apparent phenomenal character shouldn’t be intrinsic to it either, because the way in which an awareness-state appears through its pratibhāsa or representational aspect is fabricated by other mental states. Again, states of vijñāna themselves do not actually exist in the dualistic, intentionally structured way that they are projected by other states to appear.49

Next, consider the other characteristics of classic qualia. A pratibhāsa would likely not be ineffable, given that all objective and subjective mental appearances are conceptually constructed through discursive mental activity that is conditioned by cognitive/linguistic classifications. For Sthiramati, everything that an awareness-state appears as is an imaginary fabrication, and at best exists at the level of conventional discourse.50 All imaginative fabrication arises by means of linguistic expressions, and everything which is conventionally real has a purely nominal existence—they are “nothing but names” (nāmamātra).51 So, since the things that mental appearances represent—which is to say, the things which would determine how these appearances appear—are all fictions of language, any phenomenal character belonging to mental appearances would seem to be linguistic in essence. As a result, a pratibhāsa would seem not to have a phenomenal character that is ineffable.

The two remaining properties of classic qualia—direct knowability and privacy—are also not apt to characterize pratibhāsa on V&S’s account. The unenlightened beings who undergo dualistically structured mental appearances wouldn’t seem to be infallibly acquainted with them, insofar as they erroneously take these appearances to actually be extramental objects and real internal selves rather than conceptually constructed fictions. They would also lack a direct acquaintance with their own mental appearances, since there is no evidence in their extant Sanskrit works that V&S accept the later Yogācāra view that all awareness is reflexively aware and non-erroneously, non-conceptually self-intimating (Yao 2005: 127; Sakuma 2020: 46–47). Without being reflexively aware of an awareness-state and its pratibhāsa immediately in the moment that it occurs, one could only be aware of that first-order state through another subsequent awareness-state. But, that subsequent state won’t have direct access to the first-order state—being momentary, the prior state will be non-existent by the time a subsequent awareness-state could retroactively represent it. Lastly, a pratibhāsa would perhaps not be private in the way that classic qualia are supposed to be. For one, Yogācāra ontology might be incompatible with the idea that mental appearances are only accessed subjectively and are unknowable through objective, third-personal methods, since it denies the fundamental existence of an objective mind-independent reality and internal subjects who first-personally own those appearances. Additionally, whereas the privacy of qualia is supposed to make interpersonal comparisons of phenomenal states impossible, Vasubandhu allows that there can be intersubjective agreement between the appearances generated by different mental streams, owing to the fruition of their similar karma.52

Even if a pratibhāsa doesn’t instantiate the properties of classic qualia, might there still be “something it is like” to undergo it? That is, could illusory mental appearances still count as diet qualia, in which case V&S would only be weak illusionists? Although the things that appearances appear as are conceptually fabricated and ultimately non-existent, states of vijñāna could nonetheless have a real phenomenal character just in virtue of their really having those appearances. The illusionist interpreter has several replies. To start, if an awareness-state’s phenomenal qualities are purely intentional and are identical to some property of the state’s representational content, then these qualities would be illusory and ultimately non-existent, since awareness-states don’t actually have representational contents in virtue of their mental appearances. If a state’s phenomenal qualities are impurely intentional, meaning that they’re identical with some property of the intentional attitude that the state takes toward its content (e.g., a belief about x will phenomenally differ from a doubt about the same x), then these qualities will also be illusory and ultimately non-existent, because V&S think that awareness-states don’t actually have subjective modes through which they apprehend representational contents. The only possibility seeming to remain is that mental appearances have some kind of non-representational phenomenal qualities, but it’s unclear what such qualities would be and whether V&S would ascribe them to a pratibhāsa. The way in which vijñāna appears through a pratibhāsa is evidently exhausted by its appearing as a represented content or representing vehicle, and both these types of appearance do not present vijñāna as it actually exists.

4.2. Is There Something Else Other than a Pratibhāsa that Is Really Phenomenal?

If a pratibhāsa cannot have a fundamentally real phenomenal character, because what it is an appearance of—and hence what it is like—is fundamentally non-existent, then does a state of vijñāna have some other property in virtue of which there really is something it is like to undergo it? There are two prima facie candidates in V&S’s theory of mind: sensation or feeling (vedanā), and vijñāna’s “intrinsic luminosity” or “natural radiance” (prakṛtiprabhāsvara).

4.2.1. Vedanā

V&S lend support to identifying vedanā as having a phenomenal character by glossing the notion with two terms that are straightforwardly translated as “experience”: “anubhava” and “upabhoga.” Sthiramati elaborates that the feeling or hedonic affect accompanying every awareness-state is said to be what experiences or directly manifests an object’s nature through a representational aspect of pleasure, pain, or a neutral state that is neither.53 He also states that feeling is experience insofar as it is what is experienced in one of those three ways.54 In addition to being connected with direct manifestation and experience, feeling is provocatively described as the “quintessential flavor of conditioned existence” (bhavarasasārarūpā). It is for the sake of experiencing this flavor that ignorant people become attached to objects (thereby binding themselves to the cycle of rebirths).55 The phenomenal realist reader would point out that our craving for the experience of pleasure, and thus our clinging to the objects that provide this pleasure, wouldn’t make any sense if pleasure were only ever unconscious and there were nothing it’s like to feel it. Lastly, while the presence of vedanā might not be sufficient for making an awareness-state phenomenally conscious insofar as it’s also associated with unconscious storehouse awareness, Sthiramati clarifies that only states of occurrent awareness are properly experiential insofar as they’re associated with feelings that are overtly introspectible.56

The illusionist interpreter can offer several responses. First, states of vedanā are also said to have objective supports and subjective representational modes; but, because these things don’t actually exist, they cannot account for whatever phenomenal character that vedanā is really supposed to have. Next, illusionists are perfectly comfortable admitting the existence of feelings and experiences understood as “quasi-phenomenal” states, or purely physical/functional states that are introspectively misrepresented as phenomenal (Frankish 2016a: 15). Similarly, the illusionist interpreter could characterize states of vedanā in purely functional terms of inputs and outputs: Just as Frankish defines “experience” as a mental state that is the direct output of sensory systems (2016a: 13), V&S define vedanā as originating from the six sense-faculties.57 In turn, the outputs of vedanā are the desires for more pleasure and less pain, as well as the motivation to attain the objects that would gratify those desires.

Finally, Sthiramati’s characterization of vedāna as the “quintessential flavor of conditioned existence” roughly coheres with some contemporary illusionist accounts regarding the evolutionary function of (quasi-phenomenal) consciousness. For Nicholas Humphrey (2011), the illusion of phenomenal consciousness originates from “sensations” understood as internalized evaluative responses to sensory stimulation. The basic story is this: Environmental stimuli provoke in simple organisms direct bodily responses (“expressive . . . ‘wriggles of acceptance or rejection’”58) that are biologically adaptive. With the evolutionary internalization of these responses, the signals from sensory and motor areas entered endogenous feedback loops that, when internally monitored, came to be represented as a subjective phenomenal feeling about objective facts in the environment. An organism would now represent not only how the world is, but how it evaluates or feels about the world. Moreover, these internalized sensations come to ground an organism’s sense of being a conscious substantial self, or something for whom there is an inner space of value-laden feelings that give the organism an interest in continuing to persist. The upshot of this story is that the illusion of phenomenal feelings has an evolutionary function in motivating an organism to promote its own survival: “If natural selection can arrange that you enjoy the feeling of existing, then existence can and does become a goal . . .” (2011: 86). Humphrey continues:

Any creature who has it as a goal to indulge its senses . . . will be likely to engage in a range of activities that promote its bodily and mental well-being. . . . Since you can reach these moments of intense existence only by doing all the other things required to stay alive, then . . . being alive as such will become a goal. . . . You will want a life because you want to feel. (2011: 89)

Sthiramati has made a similar claim to Humphrey’s regarding the function of vedanā: it is in order to keep experiencing feeling—“the quintessential flavor of existence”—that we cling to objects of awareness and further perpetuate our continued existence within the cycle of rebirths. Where they depart on the function of feeling is that, for Sthiramati, the indulgent enjoyment of feeling (“enjoyment” also being a translation of “upabhoga”) ends up being soteriologically maladaptive.

4.2.2. Prakṛtiprabhāsvara

In finding a locus of real phenomenal character within V&S’s theory of mind, the last redoubt for the phenomenal realist reader is the notion of the mind’s “intrinsic luminosity” or “natural radiance.” We’ve seen how the representational character of vijñāna is an illusory conceptual fabrication that comprise vijñāna’s “imagined nature” (parikalpitasvabhāva): Being unreal fictions and ultimately non-existent, the mental appearances of representational contents and representational vehicles cannot grant the dependently arisen causal processes that constitute vijñāna with any non-illusory phenomenal character. The ultimately real “absolute nature” (pariniṣpannasvabhāva) of vijñāna is that it is empty of these representational fictions and thus devoid of apparent subject-object duality.

But, it’s the intrinsic luminosity of awareness that is said to be characterized by vijñāna’s absolute nature,59 meaning that if intrinsic luminosity has some phenomenal character, then it should be ultimately real. V&S’s descriptions of this luminosity seem to attribute it with the properties of classic qualia. Luminosity is intrinsic or essential to awareness given that it is equated with vijñāna’s “thusness” (tathatā), or the non-causally conditioned way that awareness always exists.60 Intrinsic luminosity qua the absolute nature of awareness, or indeed qua the realm of ultimate reality as such (dharmadhātu), is further described as: ineffable and free from conceptualization;61 non-illusory and free from error (so infallibly knowable);62 knowable through an immediate, non-conceptual vision;63 and, subjectively private in the sense that the non-conceptual knowledge of it has a form that is knowable just for oneself (svasaṃvedya).64 These descriptions of vijñāna’s intrinsic luminosity are reinforced by later Yogācāra thinkers who explicitly defend the existence of non-conceptual, non-dual, reflexive awareness (svasaṃvedana) as the ground of all phenomenal appearance (Kellner 2010; Kellner 2017: 112–13). Of note is the eleventh century Yogācāra thinker Ratnākaraśānti, who argues against the reality of mental appearances and representational contents while also equating the intrinsic luminosity of awareness with its fundamental reality as pure, contentless illumination or manifestation as such (prakāśamātra). He further suggests that intrinsic luminosity has a positive phenomenal character—what it’s like to undergo pure awareness is in part to experience great bliss (Tomlinson 2018: 367; 2019: 98–99).

Returning to V&S, we find Sthiramati stating in his commentary on the Thirty Verses that the non-conceptual vision of reality, which is a state of gnosis (jñāna) wherein one does not apprehend or cling to objects as existing outside of awareness, does not resemble a state of congenital blindness.65 A congenitally blind person would also fail to perceive extramental objects, but not for the same reason that an enlightened being fails to perceive them. To the phenomenal realist reader, non-conceptual gnosis is not a totally blind state in which the “lights are off” and nothing at all manifests in awareness; rather, there is still phenomenal manifestation, just not of anything as being an objective or subjective entity that exists apart from the awareness of it. Additionally, V&S also characterize the realm of ultimate reality—which they identify with the intrinsic luminosity of vijñāna—as being blissful.66

The illusionist interpreter has several exegetical responses available. To start, Sthiramati’s descriptions of non-conceptual gnosis do not prove that he concurs with later Yogācāras in admitting the existence of reflexive awareness (svasaṃvedana). To say that a non-conceptual vision of reality is to be “known for oneself” (“svasaṃvedya” or “pratyātmavedya67) is not equivalent to saying that every awareness-state is always non-conceptually and non-erroneously acquainted with itself. In fact, rather than believing that all awareness is infallibly and immediately self-aware, Vasubandhu suggests that unenlightened beings don’t know their own mental states any better than they know the mental states of others. As he puts it, sentient beings who haven’t abandoned the conceptual fiction of subject-object duality are misled by the appearance of that duality, and so remain ignorant about the true ineffable nature of both their own and others’ minds.68 As for Sthiramati, he says that what is to be “known for oneself” is the non-existence of subject-object duality within vijñāna qua the causal process of imagining what is non-existent.69 The illusionist interpreter would insist that this knowledge is of an absence rather than a positive phenomenal fact.

Still, enlightened beings do have a non-conceptual awareness of their own intrinsically luminous minds—why can’t this awareness have a positive phenomenal character? The illusionist interpreter could answer by pointing out that V&S characterize intrinsic luminosity, and its associated quality of blissfulness, in purely negative terms. Intrinsic luminosity qua realm of reality is blissful because it is eternal (i.e., not causally conditioned), whereas everything that is impermanent (i.e., causally conditioned) is said to have the character of suffering.70 In other words, this eternal bliss could just consist in the eternal absence of suffering—yet, there need not be anything it is like for suffering to be eternally absent. Likewise, intrinsic luminosity itself is essentially an absence that is hardly characterized by V&S in positive phenomenological terms. Instead of describing the mind’s luminosity as an effulgent state of pure phenomenal manifestation, V&S understand the notion—roughly in line with earlier Buddhist traditions (Anālayo 2017: 31–32)—as referring to the mind’s inherent lack of extraneous moral defilements and impurities, chief among them being ignorance, craving, and clinging.71 So too with the mind’s absolute nature, which V&S define as the absence of apprehended objects and apprehending subjects within the dependently arisen causal process that constitutes vijñāna.72 In all these cases, the illusionist interpreter’s point would be that absences have nothing it is like to exist.

The phenomenal realist reader could still object that the absence of positive phenomenological descriptions for the intrinsic luminosity of awareness does not constitute evidence for the absence of a positive phenomenology altogether. After all, how could we expect any positive descriptions of what intrinsic luminosity is like, given its ineffability?73 But, this objection cuts both ways—the total ineffability of intrinsic luminosity would entail that V&S’s comments about it lend as little support to the phenomenal realist reading as they do to the illusionist interpretation. And, the illusionist interpreter need not settle for this stalemate: As we will see in the next section, V&S’s negative descriptions of intrinsic luminosity suggest that it is a state empty of phenomenal appearances and images.

4.3. Are V&S Entity Eliminativists about Phenomenal Consciousness?

If the illusionist interpretation of V&S is correct in taking them to deny the existence of ultimately real phenomenal states, then the last questions to settle are these: Do they intend to eliminate phenomenal consciousness from their ontology full stop? Or, do they merely wish us to refrain from talking and thinking about phenomenal consciousness, perhaps because this talk and thought is philosophically and soteriologically counterproductive? Only if the answer to the former question is affirmative would they be entity eliminativists, and thus strong illusionists.

The phenomenal realist reader has reason to believe that V&S would settle for discourse eliminativism. While they both make statements to the effect that mental states (citta) and awareness-states (vijñāna) are conceptual fictions that are ultimately non-existent, these bold statements are always qualified so as to seemingly back away from entity eliminativism. True, Sthiramati claims that vijñāna, qua one of the five fundamental aggregates (skandha) within Abhidharma ontology, is only a “metonymic” construction that has a merely nominal existence and no objective support in reality.74 And yes, in Vasubandhu’s words, the middle way perspective with respect to the ontological status of citta is one wherein there is no citta nor vijñāna.75 The “realm of reality” is also said to be “non-mental” (acitta).76 However, there is a pragmatic motivation for such claims—namely, they serve as “antidotes” (pratipakṣa) to our engagement in, and attachment to, erroneous conceptual fabrication.77 Further, the description of ultimate reality as “non-mental” is glossed by Sthiramati as referring to the real absence of mental states that subjectively apprehend anything objective and mind-independent.78 This claim too has a pragmatic motivation within the context of Yogācāra meditation practice. Once one realizes that there are no extramental objects appearing in awareness, one does not rest content with the thought that whatever appears is merely mental, because this thought still contains a conceptual representation of the mind. It’s only when one also stops thinking about awareness that one properly dwells in the “reality of the mind,” a reality that is beyond all thought and rational analysis.79 Hence, if V&S are eliminativists about phenomenal consciousness, then they should merely be discourse eliminativists, just because they seek to eliminate all discourse about the inexpressible reality of the mind.

The illusionist interpreter has some final rebuttals at their disposal. First, they can grant that V&S’s claims about vijñāna’s non-existence enjoin us toward abandoning our attachment to the illusion that vijñāna exists in the way that it appears. Nevertheless, this pragmatic injunction follows from an entity eliminativism about ways of appearing. It is clear from Sthiramati’s glosses that the way in which vijñāna illusorily appears—namely, through a dualistic “mode of apprehension” or representational form/aspect/structure (ākāra)—is in itself non-existent (“svātmany avidyamāna”).80 Therefore, if an awareness-state has a phenomenal appearance (pratibhāsa) just in virtue of arising with some representational aspect or mode of apprehension,81 but modes of apprehension don’t actually exist because there are no genuinely apprehended or apprehending entities, then phenomenal appearances should not actually exist, either.

Of course, Sthiramati does say that every awareness-state must have a representational form,82 and that awareness does actually have the appearance of non-existent subjective and objective entities.83 Indeed, he claims that awareness must possess the representational form of some mind-independent object in order for that object to illusorily appear as being apprehended by awareness.84 Yet, we know that he also thinks mind-independent objective supports and the representational aspects through which they are apprehended are both non-existent. So, if phenomenal appearances are simply constituted by ultimately non-existent modes of apprehension that represent ultimately non-existent objects, then phenomenal appearances should also be ultimately non-existent, thereby taking their place on the entity eliminativist’s chopping block. Joining them there will also be representationally structured states of awareness—namely, states of vijñāna—since they can’t be described apart from their having the four types of illusory appearances mentioned above.85 What isn’t eliminated by V&S is the illusion that there are awareness-states that possess the appearance of subjective and objective entities—this illusion does actually exist, as does the causal process of vijñāna qua abhūtaparikalpa which brings it about.86 Any strong illusionist about phenomenal properties will say the same: For them, the existence of mental states with phenomenal properties is an illusion, and hence these properties should be eliminated from our ontology; but, both the illusion that there are states with such properties, as well as the causal processes which generate this illusion, really do exist.

Finally, would V&S extend their entity eliminativism about phenomenal appearances to the non-dual, non-conceptual vision of ultimate reality? In other words, would they assert that even this state of supramundane gnosis can’t be phenomenally conscious? Or, would they simply be discourse eliminativists about such a state because it is essentially ineffable? Consider the phenomenal realist interpretation by Hugh Urban and Paul Griffiths (1994) of V&S’s statements about the bodhisattva’s meditative “penetration” into emptiness. Urban and Griffiths suggest that even though all conceptually fabricated mental appearances are in some sense abandoned within the bodhisattva’s vision of emptiness, it needn’t be the case that this vision lacks any phenomenal properties at all. In their reading, “The possibility that appearances continue (though not, of course, erroneous judgments about them) for the bodhisattva who has penetrated to emptiness is thus left open: the occurrence of appearances does not entail the presence of error” (1994: 17).

They take an admittedly difficult route to this claim by grappling directly with V&S’s descriptions of the bodhisattva’s vision of emptiness as lacking mental images (animitta), “mental image” being their translation of the Sanskrit term “nimitta.” They understand a nimitta in this context to be any image present to the mind that causes some affective or cognitive response on the part of the experiencing subject (1994: 18). As they note (1994: 8–9), there are numerous places where V&S describe emptiness as consisting in the cessation/absence of nimitta, and the bodhisattva’s non-conceptual vision of emptiness as a state in which nimitta are not active.87 However, if a mental image no longer causes erroneous conceptual judgments, then it no longer counts as a nimitta despite its still being phenomenally present in a bodhisattva’s experience. Urban and Griffiths therefore conclude,

There are no images in emptiness if by nimitta is meant an image with error-producing phenomenal properties; but there are images in emptiness if by that is meant a phenomenally rich flow of experience in which all objects are experienced directly and without distortion. (1994: 19)

The illusionist interpreter would reject their conclusion. Urban and Griffiths are evidently adopting an understanding of the bodhisattva’s supramundane vision of reality that aligns with later Yogācāra sākāravāda, or the theory that phenomenal images or representational aspects are real, ineliminable features of even an enlightened being’s non-conceptual awareness (Tomlinson 2019). But, Sthiramati, along with his predecessors Asaṅga (4th cent.) and Vasubandhu, were traditionally attributed with the nirākāravāda view that phenomenal images or representational aspects are ultimately unreal and absent from a buddha’s mind (Kajiyama 1965). While we cannot always trust traditional doxographies, I think the reading of V&S as nirākāravādins is apt, particularly given Sthiramati’s statements that the representational aspects (ākāra) of awareness, while being essential to the illusory appearance of subject-object duality, are non-existent in themselves.

That being so, we should take V&S to deny Urban and Griffiths’ claim (1994: 19) that mental images (nimitta), appearances (pratibhāsa), and representational states (vijñapti) would still appear in the non-conceptual experience of emptiness. Urban and Griffiths think that phenomenal appearances could be present in such experience provided they don’t generate dualistic conceptual errors. Yet, we’ve seen how V&S take the four types of dualistic mental appearance, including the appearance of subjective representational states, to be constituted by the cognitive/linguistic activity of discursive attention and inner speech. This activity is conditioned by the memory traces of past conceptual fabrications of subjective and objective entities. Upon reaching a state wherein such dualistic conceptual conditioning has ceased, the fabrication of dualistic appearances would presumably cease as well. Accordingly, it’s not that phenomenal appearances count as “images” within erroneous, conceptually conditioned experience but become “imageless” in non-erroneous, non-conceptual awareness. As Sthiramati says, the “middle way” perspective granted by non-conceptual gnosis does not make things become imageless (animitta)—all things just are imageless already.88 Thus, if nimitta are phenomenal mental images according to Urban and Griffiths, then to know through a supramundane gnosis the absolute empty nature of all things in terms of their imagelessness89 would be to know the ultimate reality of all things as lacking a phenomenal character. In that case, V&S’s entity eliminativism about phenomenal consciousness would not be a purely theoretical pursuit—the ultimate non-existence of phenomenal consciousness could be directly realized through ceasing the cognitive mechanisms responsible for producing the illusion of it.

5. Conclusion

Despite initial appearances to the contrary, V&S really seem to be strong illusionists about phenomenal consciousness. In the last analysis, though, is the illusionist interpretation actually correct? Here is what I think can be safely concluded: V&S are at least strong illusionists about phenomenal intentionality, if not phenomenality as such. They would certainly reject the idea that mental states can genuinely have representational content in virtue of possessing some phenomenal appearance. Though, given how their characterization of the intrinsic, ineffable reality of the mind aligns at some level with the properties of “classic qualia,” it is hard to conclusively rule out that there would be “nothing it is like” at all to have a non-conceptual, non-dual vision of reality. On the other hand, if V&S would take the concept of phenomenality to be necessarily tied with the concept of intentionality, so that a phenomenal awareness-state is one in which there always an appearance or manifestation “of”/“as” something,90 then they may indeed be entity eliminativists about phenomenality as such.

If the illusionist interpretation is correct, then it brings into focus some interesting historical and conceptual questions for further consideration. One question concerns the reasons why many Yogācāra thinkers after V&S would diverge so sharply from illusionism in taking the fundamental reality of the mind to be grounded in its reflexive, non-erroneous, non-conceptual, and non-dual phenomenal manifestation. A second question regards the reasons why contemporary illusionists would converge so strikingly with early Yogācāra illusionists in spite of their quite distinct ontological assumptions and philosophical motivations. This apparent convergence across such wide theoretical and temporal gaps suggests at least that phenomenal realists can’t undercut the plausibility of illusionism by claiming it to be a unique artifact of Western modernity.


  1. See also Huebner, Aviv, and Kachru (2022), which reconstructs from various Yogācāra sources a general account of how illusions about experience arise.
  2. Of course, there are at least two basic issues facing my proposal to read Sthiramati’s commentaries on the Triṃśikā and Madhyāntavibhāgabhāṣya as being representative of Vasubandhu’s views. One is whether the two commentaries are both works of Sthiramati; some brief reasons for skepticism are given by Kramer (2016: 55–58). In response, I would point to evidence for their shared authorship offered by Sakuma (2020: 51–53) and Kano (2017). The issue can’t be settled here, but I’m inclined to take textual divergences between the two commentaries not as definitive proof of a difference in authorship, but perhaps as being explainable by their having different root texts and/or having been written at different times. More significant is the argumentative overlap between the two commentaries, though I cannot elaborate for lack of space. These parallels don’t necessarily prove that two texts have the same author—they could be drawing arguments from a common canonical source. Nonetheless, concerning the arguments relevant to this paper, I don’t find major conceptual discrepancies between the two texts of the sort that lead Kapstein (2018) to doubt whether the same Vasubandhu authored the Triṃśikā and Trisvabhāvanirdeśa. Likewise regarding the second issue of whether Vasubandhu’s views can be accurately interpreted in light of Sthiramati’s commentaries: I don’t find any significant divergences in their views at least on the question of their possible illusionism about phenomenal consciousness (but cf. Kachru 2021: 56–62). Further textual study in another venue would be needed to fully justify these claims.
  3. Tr 134.12–13; TrB 134.10–136.04.
  4. TrB 38.18–20.
  5. TrB 42.15–17.
  6. TrB 128.21–22.
  7. TrB 128.17–18.
  8. Tr 62.07–08; TrB 62.14–15.
  9. Tr 52.06.
  10. TrB 52.05–06; PSkV 92.13–15.
  11. TrB 58.09–11.
  12. TrB 54.01–02.
  13. PSk 376.12–14; TrB 56.22–24.
  14. Tr 104.13–14; TrB 104.11–106.17; PSkV 94.11–13. Technically, Sthiramati only applies the label “mindless” to the states of sleepiness (middha) and syncope (mūrccha), but Vasubandhu in Tr verse 16 can be read as including the two types of meditative attainment, as well. The two attainments are also called “mindless” in other works of Vasubandhu—see Cox (1995: 115) and Brunnhölzl (2019: 293).
  15. TrB 50.07–08.
  16. MAVṬ 33.09–10.
  17. Tr 108.04–05; TrB 108.01–13.
  18. TrB 112.13–16.
  19. MAVṬ 14.01–03.
  20. MAVṬ 18.09–10, 25.25–26, 217.23–218.01.
  21. TrB 42.20–22; MAVṬ 25.02–03.
  22. MAVṬ 25.06–09; Stanley (1988: 30 fn. 149).
  23. MAVṬ 18.21–22, following AKB 401.18–21; see Kellner (2014).
  24. This is a central claim of the Twenty Verses—see also TrB 44.01–16, 108.14–110.03; and MAVṬ 19.03–14, 24.27–25.19.
  25. TrB 44.22–46.02; MAVṬ 24.11–23.
  26. MAVB 72.15–16; MAVṬ 248.01–02.
  27. MAVB 18.26–27.
  28. Sthiramati also offers another explanation for why the appearances of external objects and sentient beings are “formless” (anākāra), based on a simpler definition of a representational form or aspect (ākāra) as that which consists in the awareness of an objective support (ālambanasaṃvedana). Since there are no actually external entities which could serve as objective supports for awareness, there could be no ākāra qua awareness of them. So, the appearances of those entities in awareness must be “formless” (MAVṬ 18.23–24; Stanley 1988: 22 fn. 102).
  29. MAVB 18.27–19.01.
  30. Vi 193.06–11; MAVṬ 19.15–19, 20.23–24.01, 23.03. Of course, there is an evident tension in claiming that all representational states are only erroneously imagined as being such: The mental state that imagines another state to be representational must surely be a representational state itself, since to erroneously imagine something is still to represent it as being a certain way. A similar problem faces contemporary illusionists, who must explain how a mental state can erroneously appear as being phenomenal without that erroneous appearance itself being phenomenal. They resolve this tension by construing the type of representation responsible for this erroneous appearance in functional and non-phenomenal terms. Thus, mental state y can falsely represent another state x as being phenomenal without either y or x being phenomenal themselves. This sort of resolution to the tension facing V&S isn’t available to them, since they assume that a state’s representational character is appearance-based. As a result, if state y imagines state x to appear as a representational state, then y also must have the appearance of being a representational state, since imagining is a form of mental representation. So too, y’s appearance as a representational state must have been imagined by another apparently representational state z, and on into infinity. Nevertheless, V&S would not view this infinite regress of imaginary representational states as a vicious explanatory regress, insofar as the ignorance-driven cycle of existence (saṃsāra) is beginningless and endless (MAVṬ 133.15–16). Underlying this perpetual cycle of imaginary fabrication are the causal processes constituting the “transformation of vijñāna.” I thank an anonymous reviewer for pressing these points.
  31. TrB 108.07.
  32. Vi 193.04–09; MAVṬ 19.18–19.
  33. MAVB 66.06–10; MAVṬ 218.14–219.17.
  34. E.g., PVin 7.05–06: “abhilāpinī pratītiḥ kalpanā.”
  35. TrB 74.2–6; PSkV 37.14–38.2.
  36. AKV 530.23–24.
  37. AKB 10.15; TrB 56.26–28.
  38. MAVṬ 219.11–13.
  39. TrB 56.9–10.
  40. MAVṬ 225.19–226.1, 229.12–13.
  41. AKB 22.03–22.
  42. TrB 98.16–100.5.
  43. MAVṬ 17.10–11; Stanley (1988: 20 fn. 94).
  44. MAVṬ 113.09–16.
  45. MAVṬ 219.06–15.
  46. MAVṬ 25.02–03.
  47. MAVṬ 10.09–12, 20.02–04.
  48. Vi 193.06–11; MAVṬ 19.18–19, 20.23–21.01, 22.22–23.04.
  49. MAVṬ 113.12–14.
  50. MAVṬ 22.10–12.
  51. MAVṬ 221.01–02, 19–20.
  52. Vi 190.09–13; see Tzohar (2017).
  53. MAVṬ 34.14–15; PSkV 25.01–02.
  54. MAVṬ 34.15.
  55. MAVṬ 34.15–16; Stanley (1988: 42 fns. 225 & 226).
  56. MAVṬ 34.03–04.
  57. MAVB 45.15–16; MAVṬ 145.06–10 (Stanley 1988: 192 fn. 230 & 231).
  58. Humphrey (2016: 117).
  59. MAVṬ 231.24–232.01.
  60. MAVṬ 50.07–09, 122.01.
  61. MAVṬ 229.03–04, 229.15–17.
  62. MAVṬ 227.21–228.01, 231.23–232.01.
  63. MAVṬ 100.07–08, 100.15–16, 104.06–08, 104.11–12.
  64. MAVṬ 104.06–08.
  65. TrB 136.09–12.
  66. Tr 138.09; TrB 142.07–08.
  67. TrB 142.05.
  68. Vi 199.05–10.
  69. MAVṬ 27.21–23.
  70. TrB 142.07–08.
  71. MAVB 21.23, 27.06–07.
  72. MAVB 19.20.
  73. I thank an anonymous reviewer for raising this objection and possible response.
  74. TrB 40.07–08, 108.10–13.
  75. MAVB 70.13–15.
  76. Tr 138.06.
  77. MAVB 67.03–05; MAVṬ 236.11–12.
  78. TrB 138.12–13.
  79. TrB 134.02–136.12, 142.05.
  80. MAVṬ 15.01–03; for other places where Sthiramati says that vijñāna illusorily appears through a non-existent representational form (ākāra), see MAVṬ 13.05–06, 113.15–16, and 231.18.
  81. MAVṬ 18.09–10, 217.23–218.01
  82. TrB 50.20–21.
  83. MAVṬ 20.08.
  84. MAVṬ 26.01–02.
  85. MAVṬ 20.05.
  86. MAVṬ 16.05–08, 21.04.
  87. For example, see MAVB 23.21–22 and MAVṬ 105.25–106.01.
  88. MAVṬ 243.11–12.
  89. MAVṬ 50.15–21.
  90. MAVṬ 24.14–16.


I would like to thank several people for contributing to the development of this paper. David Reigle helped me obtain a copy of Sthiramati’s Madhyāntavibhāgaṭīkā. I am indebted to Fabien Muller and Nilanjan Das for discussing Vasubandhu and Sthiramati’s texts with me. Conversations with Bryce Huebner were especially useful in getting me to think more carefully about Vasubandhu and Sthiramati’s possible illusionism. I benefitted from having several occasions to present earlier versions of this paper at Sun Yat-Sen University (Zhuhai), Institut für Kultur- und Geistesgeschichte Asiens (Vienna), and Chonnam National University (Gwangju)—I am grateful to the organizers of these presentations, particularly Itay Shani, Patrick McAllister, and Hyoung Seok Ham. The paper also benefitted from discussions with Laura Guerrero, Cat Prueitt, and Mark Siderits. Lastly, I thank the two anonymous referees of this journal for providing constructive comments that helped me to greatly improve the clarity of my arguments. Research for this paper was undertaken as part of the grant project, “Classical Indian Theories of Perception and Consciousness: A Contemporary Reconstruction” (#27620419), funded by the Research Grants Council Early Career Scheme of the Hong Kong Special Administrative Region, China.


AKB Abhidharmakośabhāṣya (Pradhan 1967)
AKV Sphuṭārthā Abhidharmakośavyākhyā (Wogihara 1936)
MAVB Madhyāntavibhāgabhāṣya (Nagao 1964)
MAVṬ Madhyāntavibhāgaṭīkā (Yamaguchi 1934)
PSk Pañcaskandhaka (Shastri 1956)
PSkV Pañcaskandhakavibhāṣā (Kramer 2013)
PVin Pramāṇaviniścaya (Steinkellner 2007)
Tr Triṃśikā (Buescher 2007)
TrB Triṃśikābhāṣya (Buescher 2007)
Vi Viṃśikā (Silk 2018)

Anālayo, Bhikkhu (2017). The Luminous Mind in Theravāda and Dharmaguptaka Discourses. Journal of the Oxford Centre for Buddhist Studies, 13, 10–51.

Balog, Katalin (2016). Illusionism’s Discontent. Journal of Consciousness Studies, 23(11–12), 40–51.

Bodhi, Bhikkhu (2000). The Connected Discourses of the Buddha: A Translation of the Saṃyutta Nikāya. Wisdom Publications.

Brunnhölzl, Karl (2019). A Compendium of the Mahāyāna: Asaṅga’s Mahāyānasaṃgraha and its Indian and Tibetan Commentaries (Vol. 1). Snow Lion.

Buescher, Harmut (2007). Sthiramati’s Triṃśikāvijñaptibhāṣya. Austrian Academy of Sciences.

Chalmers, David J. (2010). The Character of Consciousness. Oxford University Press.

Cox, Collett (1995). Disputed Dharmas: Early Buddhist Theories on Existence: An Annotated Translation of the Section on Factors Dissociated from Thought from Saṅghabhadra’s Nyāyānusāra. International Institute for Buddhist Studies.

Crane, Tim (2001). Elements of Mind: An Introduction to the Philosophy of Mind. Oxford University Press.

Dennett, Daniel C. (1988). Quining Qualia. In Anthony J. Marcel and Edoardo Bisiach (Eds.), Consciousness in Contemporary Science (42–77). Oxford University Press.

Dennett, Daniel C. (1991). Consciousness Explained. Back Bay Books.

Dennett, Daniel C. (1996). Kinds of Minds: Toward an Understanding of Consciousness. Basic Books.

Dennett, Daniel C. (2005). 2005: What Do You Believe is True Even Though You Cannot Prove It? Edge.

Dennett, Daniel C. (2016). Illusionism as the Obvious Default Theory of Consciousness. Journal of Consciousness Studies, 23(11–12), 65–72.

Dennett, Daniel C. (2017). From Bacteria to Bach and Back: The Evolution of Minds. W. W. Norton & Company.

Frankish, Keith (2012). Quining Diet Qualia. Consciousness and Cognition, 21(2), 667–76.

Frankish, Keith (2016a). Illusionism as a Theory of Consciousness. Journal of Consciousness Studies, 23(11–12), 11–39.

Frankish, Keith (2016b). Not Disillusioned: Reply to Commentators. Journal of Consciousness Studies, 23(11–12), 256–89.

Garfield, Jay L. (2002). Empty Words: Buddhist Philosophy and Cross-Cultural Interpretation. Oxford University Press.

Garfield, Jay L. (2015). Engaging Buddhism: Why It Matters to Philosophy. Oxford University Press.

Garfield, Jay L. (2016). Illusionism and Givenness. Journal of Consciousness Studies, 23(11–12), 73–82.

Goff, Philip (2016). Is Realism About Consciousness Compatible with a Scientifically Respectable Worldview? Journal of Consciousness Studies, 23(11–12), 83–97.

Graham, George and Terence Horgan (2008). Qualia Realism, Its Phenomenal Contents and Discontents. In Edmond Wright (Ed.), The Case for Qualia (89–107). MIT Press.

Graziano, Michael S. A. (2016). Consciousness Engineered. Journal of Consciousness Studies, 23(11–12), 98–115.

Graziano, Michael S. A. (2019). Rethinking Consciousness: A Scientific Theory of Subjective Experience. W. W. Norton & Company.

Graziano, Michael S. A. (2020). Consciousness and the Attention Schema: Why It Has to be Right. Cognitive Neuropsychology, 37(3–4), 224–33.

Graziano, Michael S. A., Arvid Guterstam, Branden J. Bio, and Andrew I. Wilterson (2020). Toward a Standard Model of Consciousness: Reconciling the Attention Schema, Global Workspace, Higher-Order Thought, and Illusionist Theories. Cognitive Neuropsychology, 37(3–4), 155–72.

Huebner, Bryce, Eyal Aviv, and Sonam Kachru (2022). The Magic of Consciousness: Sculpting an Alternative Illusionism. In Itay Shani and Susanne K. Beiweis (Eds.), Cross-Cultural Approaches to Consciousness: Mind, Nature, and Ultimate Reality (223–44). Bloomsbury Academic.

Humphrey, Nicholas (2011). Soul Dust: The Magic of Consciousness. Princeton University Press.

Humphrey, Nicholas (2016). Redder Than Red: Illusionism or Phenomenal Surrealism? Journal of Consciousness Studies, 23(11–12), 116–23.

Irvine, Elizabeth and Mark Sprevak (2020). Eliminativism about Consciousness. In Uriah Kriegel (Ed.), Oxford Handbook of the Philosophy of Consciousness (348–70). Oxford University Press.

Kachru, Sonam (2021). Other Lives: Mind and World in Indian Buddhism. Columbia University Press.

Kajiyama, Yuichi (1965). Controversy between the Sākāra- and Nirākāra-vādins of the Yogācāra School – Some Materials. Journal of Indian and Buddhist Studies, 14(1), 26–37.

Kammerer, François (2021). The Illusion of Conscious Experience. Synthese, 198(1), 845–66.

Kano, Kano (2017). Some Remarks on the Sanskrit Titles of Sthiramati’s Works. In Seongcheol Kim and Jundo Nagashima (Eds.), Śrāvakabhūmi and Buddhist Manuscripts (191–208). Nombre.

Kapstein, Matthew T. (2018). Who Wrote the Trisvabhāvanirdeśa? Reflections on an Enigmatic Text and its Place in the History of Buddhist Philosophy. Journal of Indian Philosophy, 46(1), 1–30.

Kellner, Birgit (2010). Self-Awareness (Svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -Vṛtti: A Close Reading. Journal of Indian Philosophy, 38(3), 203–31.

Kellner, Birgit (2014). Changing Frames in Buddhist Thought: The Concept of Ākāra in Abhidharma and in Buddhist Epistemological Analysis. Journal of Indian Philosophy, 42(2–3), 275–95.

Kellner, Birgit (2017). Proofs of Idealism in Buddhist Epistemology: Dharmakīrti’s Refutation of External objects. In Joerg Tuske (Ed.), Indian Epistemology and Metaphysics (102–28). Bloomsbury Academic.

Kramer, Jowita (2013). Sthiramati’s Pañcaskandhakavibhāṣā, Part 1: Critical Edition. Austrian Academy of Sciences and China Tibetology Research Center.

Kramer, Jowita (2016). Some Remarks on Sthiramati and his Putative Authorship of the Madhyāntavibhāgaṭīkā, the *Sūtrālaṃkāravṛttibhāṣya and the Triṃśikāvijñaptibhāṣya. Buddhist Studies Review, 33(1–2), 47–63.

Nagao, Gadjin M. (1964). Madhyāntavibhāgabhāṣya. Suzuki Research Foundation.

Pradhan, Prahlad (1967). Abhidharmakosabhāṣya of Vasubandhu. Kashi Prasad Jayaswal Research Institute.

Rosenthal, David (2011). Exaggerated Reports: Reply to Block. Analysis, 71(3), 431–37.

Sakuma, Hidenori (2020). Was There Really Only One Commentator Named Sthiramati? Tetsugaku Shisō Ronshū, 45, 39–61.

Sangpo, Gelong Lodrö (2012). Abhidharmakośabhāṣỵa: The Treasury of Abhidharma and Its (Auto) Commentary (Vol. IV). Motilal Banarsidass.

Schmithausen, Lambert (1987). Ālayavijñāna: On the Origin and the Early Development of a Central Concept of Yogācāra philosophy. International Institute for Buddhist Studies.

Shastri, Shantibhiksu (1956). Pañcaskandhakaprakaraṇa of Vasubandhu. The Indian Historical Quarterly, 32(4), 368–85.

Silk, Jonathan A. (2018). Materials Toward the Study of Vasubandhu’s Viṁśikā (I): Sanskrit and Tibetan Critical Editions of the Verses and Autocommentary; An English Translation and Annotations. Harvard University Press.

Stanley, Richard (1988). A Study of the Madhyāntavibhāga-Bhāṣya-Ṭīkā [Doctoral dissertation]. Australian National University.

Steinkeller, Ernst (2007). Dharmakīrti’s Pramāṇaviniścaya: Parts 1 and 2. Austrian Academy of Sciences and China Tibetology Research Center.

Tomlinson, Davey K. (2018). The Tantric Context of Ratnākaraśānti’s Philosophy of Mind. Journal of Indian Philosophy, 46(2), 355–72.

Tomlinson, David K. (2019). Buddhahood and Philosophy of Mind: Ratnākaraśānti, Jñānaśrīmitra, and the Debate over Mental Content (Ākāra) [Doctoral Dissertation, University of Chicago]. Knowledge@Chicago.

Tye, Michael (2021). Qualia. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Fall 2021 ed.).

Tzohar, Roy (2017). Imagine Being a Preta: Early Indian Yogācāra Approaches to Intersubjectivity. Sophia, 56(2), 337–54.

Urban, Hugh B. and Paul J. Griffiths (1994). What Else Remains in Śūnyatā? An Investigation of Terms for Mental Imagery in the Madhyāntavibhāga-Corpus. Journal of the International Association of Buddhist Studies, 17(1), 1–25.

von Rospatt, Alexander (1995). The Buddhist Doctrine of Momentariness: A Survey of the Origins and Early Phase of this Doctrine up to Vasubandhu. Franz Steiner Verlag.

Waldron, William S. (2003). The Buddhist Unconscious: The Ālaya-Vijñāna in the Context of Indian Buddhist Thought. RoutledgeCurzon.

Wogihara, Unrai (1936). Sphuṭārthā Abhidharmakośavyākhyā by Yaśomitra (Part II). The Publishing Association of Abhidharmakośavyākhyā.

Yamaguchi, Susumu (1934). Madhyāntavibhāgaṭīkā. Librairie Hajinkaku.

Yao, Zhihua (2005). The Buddhist Theory of Self-Cognition. Routledge.