1 Introduction

In the intellectual milieu of ancient India where the Hindu views dominated the philosophical landscape, the Buddha put forward a revisionary metaphysics that lacks a “self” to provide an intellectually and morally preferred picture of the world. This view is deeply counterintuitive and the Buddhists are acutely aware of this fact. Accordingly, the Abhidharma-Buddhist writings are replete with attempts to explain the phenomenology of experience in a no-self world. To evaluate the merits of the Buddhist-Abhidharma worldview, the first task is to understand the target of the no-self doctrine. This task is urgent in the contemporary context as the interest in the notion of self is restricted not only to philosophers and phenomenologists but also exists across a variety of other disciplines including psychology and developmental psychology, neuroscience, psychiatry, anthropology, cultural studies and so on. Consequently, numerous conceptions of self are found in the contemporary literature, each with its corresponding no-self view. I am not going to attempt yet another taxonomy of the different notions of self.Footnote 1 No such taxonomy is exhaustive or immune to revision and, in addition, questions about relations between these various notions, some of which are complimentary and others conflicting, cannot be decisively answered. Nor will I make the assumption, as some do, that the target of the Buddhist no-self account is the Hindu view of self as a substantial, independent entity that exists apart from mental and physical states.Footnote 2 Rather than stipulating the notion of self that is the target of the Abhidharma no-self account, I look at the arguments against self in Vasubandhu’s Abhidharmakośa-bhāsya—the most authoritative Abhidharma text – and then work backwards to explicate the kind of self that the Abhidharma philosophers reject. It is in the Chapter 9 of the Abhidharmakośa-bhāsya that Vasubandhu enunciates the kinds of self or person that must be rejected and accounts for the way in which the work of the self is instead carried out within the Buddhist worldview.Footnote 3

My focus is on two related tasks. First, as said above, I offer a reconstruction of the Abhidharma arguments for the no-self doctrine and work backwards to interpret the self that is the target of this doctrine. I argue that Buddhists are not simply denying a diachronically unified, extended or narrative self but also the notion of a minimal self associated with a sense of ownership and a sense of agency. The second task is to reconstruct and defend the Buddhist-Abhidharma explanation of lack of agency using contemporary resources. I argue that since there is nothing that it’s like to be an agent, there is no onus on the Buddhist-Abhidharma philosophers to account for a sense of agency.

Briefly, the plan of the paper is as follows. In section 1 below, I discuss Vasubandhu’s argument in Abhidharmakośa-bhāsya for the claim that there is no self. Vasubandhu takes a lot of care to give an account of how we are supposed to explain the phenomenological facts of memory, moral responsibility, agency, etc., in the absence of selves or persons. The ancient Hindu philosophers, particularly the Naiyāyiakas exhort their Buddhist no-self theorists to address questions like: How, although we are not selves, we can apprehend an object, or remember it?; How, without a self, can there be an agent of action or a subject that experiences their results?; How, without a self to possess it, there can be a mind that conceives an “I”?; and, how, without a self, there is an underlying support for desire, cognition, feelings of pleasure and pain, etc.? These issues are covered in the original Nyāyasutras and various commentaries by subsequent Nyāya philosophers.Footnote 4 Issues about memory, ownership, etc. have received a considerable amount of attention in the contemporary literature on the debate between the Hindus and Buddhists.Footnote 5 I cannot possibly cover all of these issues in the scope of this paper, so I will limit myself in Section 2, to the Buddhist explanation of phenomenological facts associated with agency (if there is any such thing). I begin with Thompson’s (2014) reconstruction of the no-self view that endorses a self as the subject of experiences and agent of actions. I argue that such a reconstruction is not true to the spirit of the Buddhist views in general and mature Buddhist-Abhidharma view in particular. The denial of agency certainly leaves the Buddhist-Abhidharma philosopher with the burden of explaining the phenomenological sense of agency. But is there really such a sense of agency? Section 3 argues that there’s nothing like being an agent in the sense that there’s no experiential phenomenology associated with agency; our sense of agency is a conceptual construct. The upshot is that the Buddhist-Abhidharma philosopher need not burden herself anymore with explaining the sense of agency as it will be explained away by the end of this paper.

2 Self and no-self in Vasubandhu’s philosophy

Chapter 9 of the Abhidharmakośa-bhāsya begins with the customary homage to the Buddha and then Vasubandhu proceeds directly to state his main argument for no-self. Footnote 6 The question is: how do we know that the term ‘self’ refers to a series of aggregates of mental and physical states (skandhas) and not to something else? Vasubandhu responds by saying that we know this because no proof establishes the existence of a self apart from the aggregates. There is no proof for the existence of the self by direct perception, nor by inference. He elaborates further that we can know objects of the five senses and the objects of mental consciousness by perception. And we can know about the existence of the five external sense organs on the basis of inference from the fact that even in the presence of all other causes of perception—e.g., external objects, light, attention, etc.—the blind and the deaf cannot perceive. Thus we infer the existence of the sense organs as a cause whose presence, together with other factors, brings about a perception. However, we cannot perceive a self, nor are there any considerations that would lead us to infer or postulate a self; so we can conclude that there is no self (Pruden 1988, 1313–4).Footnote 7

2.1 Vasubandhu’s epistemological argument for no-self

Kapstein (2000) and, more recently, Kellner and Taber (2014) interpret Vasubandhu’s argument as an epistemological argument. The latter suggest that it is an example of an argument from ignorance, a general argumentative strategy used by Vasubandhu in the Viṃśika to refute the existence of external objects and in the Abhidharmakośa-bhāsya to refute the existence of selves or persons. The argument from ignorance may seem like a bad strategy. It is often seen as a logical fallacy of the general form: since P is not known or proved to be true, P is false. But from the fact that the general form of the argument is fallacious or invalid, it does not necessarily follow its every instantiation is necessarily so. Arguments instantiating fallacious forms can be sound because of other features, for example the semantic meanings of the terms, contextual features, etc. Kellner and Taber (2014) emphasise that some arguments from ignorance are successful when they function as arguments to the best explanation especially in contexts where there are agreed-upon standards of verification. For example, the medical community agrees that the most accurate and sensitive test for typhoid is testing the bone marrow for salmonella-typhi bacteria. If it turns out that it cannot be proven that one has typhoid (because of the absence of this bacterium in one’s bone marrow), then it is false that one has typhoid. No matter how suggestive the symptoms are, if the specific bacteria do not show up in the bone marrow within a specific time period, then one does not have typhoid. So, then, the question is: is Vasubandhu’s argument from ignorance successful in refuting the existence of self? However, it is fair to say that Vasubandhu (and his Hindu opponents) are in broad agreement that there are at least three basic sources of knowledge (pramāṇas): perception, inference and scripture. And, in Abhidharmakośa-bhāsya, Vasubandhu covers all these three sources to rule out epistemic evidence for a self: Chapter 9 is neatly divisible into these three corresponding parts summarised in the sub-sections below. The first and second parts question the fellow-Buddhist Vātsīputrīyas (Personalists) view that the self can be known by perception and also by an appeal to the scripture. The third section examines the inferential proof for the self offered by the Nyāya-Vaiśeṣikas.

2.2 There is no perceptual basis for belief in pudgala (person)

The Personalists claim that pudgala (usually translated as person), which is akin to self, exists. The primary motivation for postulating pudgala is to account for continuity across lifetimes, since the aggregates in the two distinct lifetimes are completely different. So, Vasubandhu’s first question is: is the person ultimately real or only conceptually real? This question refers back to Buddhist doctrine of two-truths introduced earlier in the Chapter 6 of Abhidharmakośa-bhāsya. Vasubandhu introduces the distinction between relative truth (samvṛti-sat) and ultimate truth (paramārtha-sat) by way of examples. Things like pot and water are only relatively existent because they cannot be cognised when divided into parts. This division might be literal—smashing the pot into shards – or mental—analysing water by its component properties (colour, taste, etc.) such that the idea or concept of water falls away. In contrast things that are ultimately true continue to exist as such even when broken into parts or mentally deconstructed. Physical and psychological atoms—the dharmas – are the only things true in this latter sense. Persons or selves, on the other hand, are only relatively true in that they are conceptual constructions for they are deconstructed by the mind into aggregates and ultimately into momentary atoms. Ganeri (2007, 171) notes that these definitions suggest that what is really at issue for Vasubandhu is what we might call ‘stability under analysis’. So, in effect the Abhidharma view amounts to saying that only things that are stable under analysis are ultimately real, everything else is relatively or conceptually true. If the pudgala is ultimately real, it must be essentially separate and thus distinct from the mental and physical atoms. Furthermore, if it is a distinct thing, as the Personalists insist, they need to show whether it is unconditioned or is conditioned by its causes. If it is the former, then the Personalist can be charged with defending a non-Buddhist view. If on the other hand, the Personalist claims that the pudgala is conditioned by the aggregates, then their view is no different from that of the Abhidharma. The Personalists’ answer is that pudgala is neither ultimately nor conceptually real, but arises dependently on the aggregates. To unpack what dependence means in this context, she offers an example: it is like the dependence of fire on fuel. Vasubandhu says that if the pudgala arises in dependence on the aggregates, the term pudgala applies to the aggregates, not to a person for a person is not something that is perceived.

The Personalist rejects this argument and claims that pudgala is actually perceived; it is the object of all the six consciousnesses in accordance with the Abhidharma view of perception. According to this view each of the five sense organs have their own objects and domains, they do not experience the domain and the object of any other. The mind, the sixth faculty, experiences the domains and objects of the five faculties, it acts as a support for the sense faculties. And, thus even the mind is invariable with respects to its objects. Vasubandhu argues that only specific sensory qualities (e.g., colour, shape, etc. for the eye) are proper objects of perception and thus pudgala cannot be an object of any of the six consciousnesses; therefore, pudgala is not perceived. In other words, since the pudgala is not an object of consciousness, it cannot be ‘found’ which in Buddhist orthodoxy amounts to the claim that the pudgala does not exist. Thus the Personalist fails to explain what it means to say that persons are perceived as distinct from the sensory qualities of the aggregates.

What can we claim about the notion of self or person from this debate with the Personalists? They want to cash out the idea that persons are relatively permanent entities that arise depending on impermanent aggregates of mental and physical states. The person is ‘neither the same nor different from’ the aggregates of mental and physical states. The first disjunct ‘not the same as’ is meant to rule out a reductionist view of persons. The second disjunct ‘not different from’ is meant to show that persons are not causally independent of aggregates. Buddhists Personalists, Ganeri (2012) rightly argues, are emergentists about persons in as much as they reject that persons are reducible to aggregates. Vasubandhu tries to show that this view is inexplicable. The claim that pudgala does not exist as it cannot be perceived is part of the larger argument from ignorance against the Personalists in this part of the Abhidharmakośa-bhāsya. However, Vasubandhu raises another important concern about the status of persons as emergent properties. His claim that Personalists fail to explain what it means to say that persons are ‘causally founded’ on aggregates is to anticipate Kim’s argument against emergent properties. Jaegwon Kim (1992, 2005) argues that if emergent properties are causally inefficacious then they are epiphenomenal and thus cannot be said to exist. So, in arguing against the Personalists view, Vasubandhu dismisses the view that pudgala or persons are relatively permanent entities that emerge from the aggregates but are not reducible to or deducible from those aggregates.

2.3 There is no scriptural basis for a belief in the self

Vasubandhu argues that there is no statement of the Buddha affirming that self or person exist. In fact there is evidence to the contrary. This evaluation is important in the Indian context as all Buddhist thinkers (irrespective of their School or tradition) saw themselves as offering the right interpretation and defence of Buddha’s words. Since the term pudgala is often used in the Buddhist scriptures, Vasubandhu needs to address the import of these passages to refute the Personalist. Since scripture is regarded as one of main pramāṇas, this part of Abhidharmakośa-bhāsya is important because it can be conceived as part of the overall argument from ignorance. But because it does not carry much philosophical weight, I will not look at the detail of the interpretations of various texts offered by Vasubandhu. Just one example to give the reader a taste of the style of interpretation can suffice. The Personalists ask, how else are we to interpret Buddha’s use of the phrase ‘the bearer of the burden’. Vasubandhu explains away this talk of ‘bearer’ by pointing to other sections in the Nikayas (the original dialogues of the Buddha) where the Buddha explicitly says that such expressions are purely conventional devices. Hence the talk of ‘burden’ and ‘bearer of the burden’ can be explicated in terms of the aggregates and the causal affectation of the later by the earlier aggregates. No other inexplicable and relatively permanent entity, such as pudgala, need be introduced.

2.4 There is no inferential basis for belief in the self

In this part of Abhidharmakośa-bhāsya, Vasubandhu directly attacks the inferential proof for the self by the Hindu philosophers in Nyāya-sūtra 1.1.10. The Nyāya argument is that desire, volition, etc., would not be possible without a single agent that cognises and recognises the object. This single agent is the self. The weight of the argument rests on memory to defend the reality of a diachronically extended single agent. In his response, Vasubandhu outlines an alternative explanation of memory in terms of causal connections between momentary mental events in a series. Next, the Hindu asks: how can we make sense of agents of physical actions and that of knowledge without there being a self? The questions go on: cognition, happiness and pain are qualities had by a substratum, what is the substratum of these qualities?; who is the referent of the notion of “I”?; who is the one who is happy or unhappy?; and, finally who is the agent of karma and the enjoyer of the results of karma?

Vasubandhu’s strategy is to respond to each of these questions by giving an alternative explanation of the phenomena at issue by appeal to nothing but (the only ultimately real) momentary events and the relations of cause and effect combined with conventional practices. So, for example, the need to postulate an agent for bodily actions like eating, bathing, walking, etc., is explained in the following manner. We do not need a self as the agent of an action of the body, since we cannot infer it as a cause. A self contributes nothing to the arising of an action, for the desire to eat, say a mango, arises from a memory of enjoying a mango in the past. From this desire arises a consideration as to how to satisfy this desire, and from that consideration arises an intention to move the body for the sake of satisfying the desire. This movement, of the hand to acquire and cut a mango, finally leads to the action of eating a mango. There is no need to invoke the self as an agent at any point in this explanation. For the Abhidharma Buddhist the self is an ontological dangler without a causal role or an explanation. Vasubandhu says that by the very fact that we cannot apprehend the capacity of the self, any more than the capacity of the various chants uttered by a quack doctor when it is established that the effect has been brought about by the use of certain herbs, we must conclude that the hypothesis of the self is untenable. The point of these explanations is not just that there is a better alternative explanation of the phenomena, but that these alternative explanations show that there is no need to postulate or infer a self to explain these phenomena. Thus, Vasubandhu concludes, there is no inferential basis for a belief in a self.

What can we claim about the notion of the self from this debate with the Nyāya-Vaiśeṣikas? The self, according to Vasubandhu, is not a subject of experiences, nor an owner of memory or other cognitive states, nor even an agent of actions because the subjects, owners, and agents are not separate from the cognitive states themselves. And there is no permanent or a persisting self required to explain phenomena like memory and karmic causality. Furthermore, Vasubandhu also denies the need to postulate a self as the substratum of qualities. All these phenomena, which are regarded by the Nyāya-Vaiśeṣikas as inferential marks of the self, can be explained without postulating a self. His strategy in defeating each of the inferential marks is similar to his argument against the Personalist. Vasubandhu argues if as the Nyāya-Vaiśeṣikas claim that the self they are talking about is separate from and causally independent of the series of mental and physical states, then such a self is causally inefficacious and thus cannot explain anything. There is no need to posit such a self.

I will not evaluate the success of Vasubandhu’s argument at this stage. I am well aware of the many concerns that may arise in the mind of the reader about the outright dismissal of emergentism as a serious candidate for explaining the self, the suspected incoherence in explicit denial of the need for an owner of experiences or agent, the adequacy of the explanation of memory, etc. But I will briefly note a point about the structure of the argument. Kellner and Taber point out that Dharmakīrti, a leading philosopher of the mature Buddhist-Abhidharma tradition, introduces restrictions on the argument from ignorance to the effect that non-apprehension of a thing proves the non-existence of a thing only when the thing in question would have to be apprehended were it to exist (2014, 734). The point is that arguments from ignorance cannot work with the single premise that P is not known or proved to be true, to the conclusion that P is false. In an extended study of arguments from ignorance, Walton (1995) adds a second premise: if P were true, it would have been known that P. Dharmakīrti’s restriction adds exactly this second premise. Arguments from ignorance have a presumptive status, their conclusion can be presumed to be true given that it is reasonable to assume that there is no counter evidence. For now, we will presume that there is good reason to think that a relatively permanent self, which is distinct from and independent of the aggregates of mental and physical states, does not exist.

3 Abhidharma deconstruction of the self: contemporary views

So far we have discussed the kind of self that is the target of Vasubandhu’s refutation. On this distinctive Abhidharma view there are no selves or persons, there are only aggregates or the sequential psycho-physical processes that supervene on collections of ultimately real momentary atoms (dharmas) (Chadha 2015). This deeply counterintuitive view drives contemporary Buddhist philosophers to qualify the rejection of self as the denial of a substantial self that is independent of the mental and physical aggregates that constitute us. This tendency is further exacerbated by the later Abhidharma-Yogācāra epistemologists’ introduction of the notion of self-awareness or reflexive awareness (svasaṃvedana).Footnote 8 In the absence of a self this raises the question: what are we aware ‘of’ in self-awareness? The Buddhist-Abhidharma answer is to say that self-awareness is not to be understood as awareness of a subject having or possessing different experiences, rather it is simply a conscious state being aware of itself or being given to itself in a first-person way. Dharmakīrti can be interpreted as saying that there are “numerically distinct minimal selves: dependently conditioned, temporary subjects that arise, exist, and pass away within the span of an occurrent episode of consciousness” (Kruger 2011, 51). Add to this the Buddhist-Abhidharma claim that occurrent episodes of consciousness are momentary events, and it follows that the so-called minimal selves are far too minimal to be ‘good enough deservers of the name’ (Lewis 1995, 140). In Lewisian-speak, I concede that the search for prefect deservers of our folk-psychological, semi-theoretical notion of ‘self’ is futile, because there aren’t any perfect occupants of the role and hence no perfect deservers of the name. But the very minimal Buddhist-Abhidharma selves are so imperfect that they are anyway not good enough deservers of the name. This notion of self-awareness in Buddhist-Abhidharma philosophy is in some ways similar to the phenomenologists’ notion of pre-reflective awareness. However, it is important to distinguish the Buddhist-Abhidharma notion of self-awareness from the various notions of minimal self in the literature. Zahavi equates a minimal self with the ‘very subjectivity of experience’ (2005; 2012), but this is a notion thicker than the one that a Buddhist-Abhidharma philosopher can willingly endorse. For Zahavi, conscious experiences and consciousness itself have temporal structure and an extension in time, unlike the Buddhist-Abhidharma universe in which conscious experiences, like everything else, are only momentary events. The same applies to Gallagher’s (2000) notion of minimal self and Damasio’s (2012) notion of core self as they both include the sense of ownership and agency which is not part of the Buddhist-Abhidharma notion of self-awareness. Similarly, the notion of minimal phenomenal self proposed by Blanke and Metzinger (2009) is too rich to be endorsed by the Buddhist-Abhidharma philosophers as it involves self-location and self-identification.

Contemporary Buddhist philosophers, for example, Duerlinger (2003), Ganeri (2007) and Thompson (2014) are inclined to argue that to say that the self is not ultimately real is not to say that the self is an illusion. Ganeri says that persons are conventionally real or ‘real with reference to conception’ and, therefore, are not illusion. Person-involving conceptual schemes are subject-specific or interest-specific; they are positional observations, but not for that reason subjective illusions (2007, 173). They are ways of thinking about the real; not false but certainly imperfect. I agree with this view, insofar as the self is a construction, just like pots, we should not think of persons as real but as artificial kinds. The person-involving conceptual schemes are artificially constructed; persons are real in the sense in which the Menzies building is real. The construction of the Menzies Building is, however, imperfect; we could do better.Footnote 9 So too, it is with persons and selves. According to the Buddhists, the conceptual schemes containing persons and selves are morally and intellectually inadequate. These schemes are ways of thinking about the world as organised into persons, divided into me, you, ours and others. There are no such strict divisions and boundaries at the level of reality. Furthermore, there is no reason to endorse such divisions; it only leads to suffering. The Abhidharma view, and the Buddhist view more generally, is not that the self is an illusion but that it is a delusion that needs to be deconstructed as we are better off without it. If we are able to get rid of this delusion we will reduce suffering, which is the overarching aim of Buddhism.

Thompson, however, argues that a minimal notion of self, which involves thinking or experiencing of the stream of consciousness as “mine”, is not a delusion (2014, 359). According to him, the Abhidharma no-self doctrine is the denial of a substantial independently existing thing, but not of the minimal sense of self that is the subject of experiences and agent of actions (2014, 358–360). Thompson agrees with the Abhidharma-Yogācāra view that our sense of self is mentally constructed but does not believe that it follows from that that the self is nothing but an illusion, since all illusions are constructions but the converse is false. Thompson claims that the Abhidharma-Yogācāra offer an account of how the self is constructed. The Yogācāra introduces two new notions of consciousness, namely afflictive mind (kliśṭa-manas) and storehouse consciousness (ālaya-vijñāna), over and above the five sensory consciousnesses and mental perception. The basic or storehouse consciousness (ālaya-vijñāna) is a neutral, baseline consciousness that serves as a repository of all basic habits, tendencies and karmic latencies accumulated by an individual. This basic consciousness is misapprehended by the afflictive mind (kliśṭa-manas) as self (Dreyfus and Thompson 2007, 97). The afflictive mind is responsible for generating a sense of self, which is articulated in “I-Me-Mine”. This sense of mineness is not based on introspective attention, which requires picking out a given thought or experience and identifying it as one’s own. Rather, my experiences are given to me as mine in a more basic, preattentive and nonidentifying way. According to the Abhidharma-Yogācāra this sense of mineness, in turn leads to the generation of I-thoughts which are necessarily mistaken. Thompson explains that the mistake consists in assuming that the “I” essentially refers to a substantial self that exists independently of the psycho-physical stream of consciousness. However, a minimal notion of self that can be said to arise dependently from the stream of consciousness cannot be faulted in the same manner. Rather, such a minimal notion of self as a subject of experience and an agent of action provides a legitimate and valuable notion of self. Legitimate because it allows us to experience ourselves as neither the same as nor different from the stream of consciousness and valuable because it allows us to individuate my experiences and actions as belonging to me as subject and agent without thinking of myself as a substantial entity (Thompson 2014, 361). Furthermore, in his defence of the minimal notion of self, Thompson appeals to the Mādhyamika view, rather than restricting himself to the Abhidharma position. Another reason offered by him in support of the minimal self is that the function of the term “I” is not to refer. Rather, following the Mādhyamika philosopher Candrakīrti, he says that the term “I” serves an appropriative function. The appropriation is to be thought of as an activity of laying claim to rather than asserting ownership of experiences and thoughts within one’s conscious stream:

One individuates oneself as a subject of experience and agent of action by laying claim to thoughts emotions and feelings—as well as commitments and social practices – and thereby enacts a self that is no different from the self-appropriating activity itself. Again, the self isn’t an object or a thing; it’s a process—the process of “I-ing” or ongoing self-appropriating activity (Thompson 2014, 363)

Thompson’s rich notion of the self as the subject of experiences and the agent of actions may well offer intuitively acceptable notion of self as process, but it is certainly not an Abhidharma notion of self. The reason why such a notion of self is unacceptable to the Abhidharma-Buddhist philosophers is because the notion of a minimal self as a subject and an agent presupposes a diachronically extended and thus relatively permanent self. Thompson (2014) does not talk explicitly about the temporal extension of the minimal self, but he does not subscribe to the doctrine of momentariness as he does not think that the Abhidharma views of momentariness can account for the temporality of consciousness.Footnote 10 It is important to note that for the Abhidharma-Buddhist philosophers the process of “I-making” is thought to be afflicted and erroneous precisely because it takes the storehouse consciousness, itself a momentary series of conscious moments, and transforms it into a relatively permanent self. A good deserver of the name ‘self’ brings in the idea of continuity and at least some temporal width. But it is precisely for this reason that any notion of self must be rejected by the Abhidharma. Thompson may simply say that he is not defending an Abhidharma view, his aim is to use Buddhist ideas and materials in the service of constructing his own view. Strictly speaking, Thompson’s view is not an Abhidharma view. However insofar as Thompson’s view is a reconstruction which in his words “combines elements from Buddhist philosophy (specifically from the Mādhyamika school), biology, cognitive science and the neuroscience of meditation” (Thompson 2014, 24), he is not justified in ignoring the central Mādhyamika thesis that the sense of self is a delusion. In his seminal text Ratnavali, Nagarjuna, the founder of Mādhyamika endorses the view that our sense of self is a delusion:

The psycho-physical complex originated from the sense of self, but this sense of self is in reality false (anatta). How can the sprout be true when the seed is false. (Ganeri 2004, 68).

Similarly, Candrakīrti develops this line of thought:

…for those who are far removed from viewing the nature of self and own as they really are, who are caught in the cycle of birth and death, in the grip of the misbelief of primal ignorance, for such, a false thing—the self as hypostatized on the basis of the skandhas—manifests itself as real. But for those close by who see truth of these matters, no such false thing manifests itself. (Ganeri 2004, 68)

The Abhidharma-Buddhist philosophers’ rejection of self as the subject of experiences and the agent of actions is underwritten by their denial of any entities that exist over time. Also, as explained in section 1 above, Vasubandhu denies the need to postulate an agent of action; causal connections among series of mental states are adequate to explain mental and physical actions. The Buddhist concern is that the sense of self as the agent of actions is responsible for the delusion of a diachronically unified self that not only coordinates the mind-body complex but also is able to mobilize emotional resources for actions necessary to maintain the integrity of the organism (Dreyfus, forthcoming). Our actions are aimed at self-preservation of this mind-body complex and are directed by the special, though unwarranted, concern one has for one’s future self. The strength of my emotions contributes to the overriding and asymmetrical concern I have for myself, and endows me with a sense of ‘specialness’ that makes me, in the words of William James, “the home of interest” (1983[1890], 285). But such a sense of self, the Buddhist argues is an illusion because it depends on the diachronically unified and bounded nature of the self separate from the rest of the world. This deluded sense of agency, insofar as it is based on an egocentric view of the world with a special referent for “I” over and above the psychological and physical processes, is not a faithful representation of how things really are. The denial of self as an agent is a special case of the general Abhidharma position that nothing stands over and above the physical and psychological states. However, the Buddhist-Abhidharma are committed to giving an account of phenomenology of experiences. So, our question is: is there anything like a sense of agency which is grounded in the subconscious activity of the afflictive mind?Footnote 11 I will argue that there is not.

4 Is there a sense of agency?

Intuitively, the sense of agency is the sense of me as the agent of my actions, the being that is the owner and controller of my mind and body, insofar as they are involved in action planning and execution. For normal well-functioning individuals, actions are accompanied by a sense of authorship and a sense of being in control of one’s actions. Historically, the intimate connection between sense of self and agency was first formulated in the works of Maine de Biran who has influences Ricoeur’s writings in the twentieth century.Footnote 12 Contemporary discussions of the sense of agency have been influenced by this work in the early nineteenth and twentieth century, particularly Ricoeur’s notion of selves as agents and the narrative self. Recent discussions reveal that the sense of agency is complex and ambiguous and the literature contains a variety of perspective. And, while philosophers do agree that it is hard to pin down an exact definition or meaning for the sense of agency, they insist that it “should not be taken as mere façon de parler” (Bayne 2011). Bayne and Pacherie (2014) include experiences of deliberation, experiences of intentionality, experiences of decision-making, experiences of freedom, experiences of mental causation, the awareness of movement, the awareness of intentions to act, the sense of control, the sense of effort and so on as aspects of agency. Gallagher (2012) talks about multiple contributories to the sense of agency, some of which are reflectively conscious, some pre-reflectively conscious, and some non-conscious. Earlier he had suggested that sense of agency is constituted of a sense of action control and monitoring (Gallagher 2000). de Vignemont and Fourneret (2004) take it to be the sense of initiation and the sense of one’s own movements, while Pacherie (2007) simply talks about it as the experience of being in control. Bayne, however, identifies agentive experience as having at “its core the experience of a particular movement or mental event as realizing one’s own agency” (2011, 357; emphasis added). In other words, simply the sense of authorship.

My interest is not in this whole complex but, following Bayne and Pacherie (2014), only in the core of the sense of agency as sense of authorship. Thus, I ignore reflective judgements and beliefs about agency. This is because the Buddhist-Abhidharma philosopher is interested in explaining the phenomenology of our experience of agency as the sense of authorship (if there is any such thing) and not reflective beliefs and judgments of agency. As agentive experience, the sense of authorship stands for a ‘special feeling’, or ‘a positive phenomenal content associated with acting’. Bayne (2011) is keen to insist on the analogy with perception. Just as we have sensory systems that function to inform us about how things are within and immediately outside our bodies, so too we have a sensory system whose function it is to inform us about features of our own agency. Bayne does not offer a direct argument for his perceptual model of agentive experience, but he does say that the existence of agentive experience can be highlighted by drawing attention to the pathologies of agency, e.g., anarchic hand and schizophrenic delusions of thought-insertion and alien control. They are pathologies of experience in virtue of the fact that there is good reason to believe that these delusions are at least partly grounded in abnormal experiences of agency (Pacherie 2006). That these syndromes involve experiences of alienated agency does not entail (though it provides some support for) the thesis that unimpaired agency is accompanied by experiences of intact agency, Bayne argues. I will argue this claim does not provide much support at all for the view that there is something like an agentive experience.

I am not denying that there are anomalous experiences of agency in pathological cases; rather, I want to deny that in the normal case, in which there is no such loss of authorship, agentive experience is present. From the fact that the anarchic hand patient appeals to the experience of lack of authorship to explain her denial of the authorship, it does not follow that there is the presence of agentive experience. The point is that we should not immediately infer the presence of an agentive experience from the behavioural capacity to distinguish the presence of authorship from the lack thereof. The behavioural capacity, and the resultant ability to judge reliably whether or not we are the authors of our actions is proof that the experience of being the author is remarkably different from the experience of not being the author. There is a positive phenomenology associated with the case of loss of authorship, which is absent in the case of authorship. There is no agentive experience since there is no phenomenal content exclusively associated with authorship. Rather, the sense of agency or authorship results from the absence of a feeling associated with loss of authorship. The claim is not that there is a feeling associated with being the author of one’s actions, but it is not normally accessed because it is recessive, like the sense of one’s body. It is also important to note that the absence of feeling of agency is not meant to suggest that there are no sub-personal processes associated with agency. Surely there are sub-personal and sub-doxastic processes underpinning agency, the claim here is merely that these processes are inaccessible to our awareness. The fact that experiences of alienated agency in cases of syndromes like anarchic hand are likely to be cognitively impenetrable, Bayne argues, adds weight to his argument that there is agentive experience because the agent’s experience of lack of agency does not go away in the face of rational explanation to believe otherwise. But again the resilience of experiences of lack of agency is no proof for the existence of agentive experience. Cognitive impenetrability of experiences of alienated agency give us good reason to believe that there are robust experiences of the lack of agency, but does not tell us anything directly about the existence or otherwise of agentive experiences.

My case against the presence of agentive experiences is draws strong support from Paglieri’s (2013) thesis that there is no specific feeling associated with free action or decision. Paglieri demonstrates that from the fact that we can easily and reliably judge whether or not our actions are free, it does not follow that there is ‘freedom attribute’ in our experience of acting freely. The judgments are based on the absence of coercion, not the extra ingredient of ‘feeling free’ in our experience. Similarly, I want to argue that from the fact that I can judge easily and reliably whether or not I am the author of a given action, it does not follow that there is an extra ingredient of ‘agentive experience’; it may just be an absence of a feeling of loss of authorship. What I am aiming at, following Paglieri (2013), is to defend a default theory of agency. This default option is to consider ourselves as authors of our actions with no need for any other proof or evidence from our own experience. In contrast, only a phenomenologically salient experience of loss of authorship can lift the default and force me to judge that I am not the author of my actions. Paglieri (2013, 155) suggests the following as desiderata of a default theory of sense of agency:

  1. 1)

    Evidence that the phenomenology specifically associated with authorship is thin and recessive and tends to be reported in ways that sound suspiciously close to ex-post reconstructions, if not outright fabrications;

  2. 2)

    Evidence that the lack of authorship, in contrast, associates with a clear and rich phenomenology;

  3. 3)

    Arguments to the effect that presence of the sense of authorship is typical and nonproblematic, whereas the absence of the sense of authorship is the exception that needs to be promptly detected, for the individuals to perform adequately;

  4. 4)

    Evidence that the known pathologies and distortions are adequately accounted for by the default theory.

There is agreement in the literature on sense of agency that the phenomenology associated with agency is thin and recessive. In noting the complexity of the experience of agency, Gallagher (2012, 19) says:

This complexity may be surprising in light of what is usually considered to be the “thin” phenomenology associated with agency, which means that the sense of agency is short-lived and phenomenologically recessive (i.e., it remains in the pre-reflective background of experience and so not very noticeable in ordinary experience), and therefore difficult to specify.

In a similar vein, Bayne (2008, 184) argues:

The advocate of agentive experience need not hold that agentive experiences are phenomenologically vivid or easy to discern; indeed, it is common for agentive experience to be described as recessive—as typically confined to the margins of consciousness.

In the course of arguing that sense of agency is generated in sensory-motor processes, Tsakiris et al. (2007, 660) also claim that:

The sense of ownership and sense of agency are part of a pre-reflective experience of embodied experience. They are generated in low-level, albeit complex, sensory–motor processes. They tend to remain phenomenologically recessive or attenuated. That is, they involve a thin or minimal although not necessarily simple phenomenology.

Thus, there seems to be widespread agreement, among those who advocate that there are agentive experiences, that the phenomenology is thin and recessive. Furthermore, the reports of such agentive experiences sound very close to ex-post rational reconstructions. Gallagher (2012, 29) is concerned whether the various aspects and elements of agency involved in a unified qualitative experience of agency can be articulated in action. He uses the example of the actions of a cliff climber José who takes on a challenging climb in the Himalayas. After months of meticulous planning José is finally at it and is totally immersed in the activity. Gallagher would agree that it is not clear that José’s experience of being the author is so articulated when he is in action. I think Gallagher is right to worry about this, but what is more important for our purposes is not just whether José’s experience of authorship is articulated, but also whether he has any sense of being the author when he is in action. Csikszentmihalyi (1978) has shown that when people are immersed in an activity, e.g., rock climbing, they retrospectively report that they were aware of the immediate situation but say that they cannot report the contents of their conscious awareness at the time. They also report having no sense of how much time passed during the activity in which they were immersed. Ex post facto reports of rock climbing and bushwalking in challenging terrains with their florid phrases sound a lot closer to fabrication than actual description of experience in action. For example:

Our rope slides through the anchor and flirts with gravity before eloping into the wind. Again we’re caught off guard by crashing and tearing noises overhead. Through lips chapped and bleeding, we debate the cause of these eerie sounds, rule out rockfall, and credit the anomaly to swirling gusts ripping through the canyon’s narrowing walls. (Blake Herrington: Climbing &Writing; http://blakeclimbs.blogspot.com/2010/10/red-rock-retrospective.html)

and,

Our reverie upon the great trail of dust and rock was soon shaken to its very core when our bricky crew came across what proved to be a great torrent of water. Surely this cataract was the resulting outburst of effusive melting of great snowfalls high above us in the cradle of the unreachable Sierra peaks. Old Sol, his heat and light an all too powerful blast, turning ice to water, and the subsequent cataclysm now found its way to lower elevations, threatening to end our journey before it began. One false step and our weak manflesh would be hurled downstream, only to find purchase upon hard boulders and snag-filled pools; our lifeless bodies broken and desecrated. (Trip Report from J M Jelak; http://www.summitpost.org/there-were-giants-a-mount-langley-trip-report/952861)

The foregoing presents some evidence to satisfy the first desideratum to defend the default theory of the sense of authorship. Evidence for the second desideratum is not difficult to come by as is shown by the vignettes from researchers or verbatim quotes by patients suffering from thought-insertion and anarchic hand syndromes:

I look at the window and I think that the garden looks nice and the grass look cool, but the thoughts of Eamonn Andrews come into my mind. There are no other thoughts there, only his … He treats my mind like a screen and flashes thoughts onto it like you flash a picture.’(Mellor 1970, 17)

Thoughts come into my head like “Kill God.” It’s just like my mind working, but it isn’t. They come from this chap, Chris. They’re his thoughts (Frith 1992, 16)

I got up during the middle of class one day, and without telling anyone, I started to walk home—which was about five miles away — and I felt that the houses were starting to communicate with me and that they were sending me messages. I didn't hear any voices, and I thought they were putting thoughts inside my head, things like, ‘Walk, repent, you are special, you are especially bad.' Accompanying this were feelings of intense loathing and fear”. (Eli Sacks, http://io9.com/5983970/im-elyn-saks-and-this-is-what-its-like-to-live-with-schizophrenia)

One of our patients (GP) once, at dinner, much to her dismay saw he r left hand taking some fish bone s from leftovers and putting them into he r mouth (Della Sala et al. 1994).

Another patient of ours (GC) often complained that he r hand did what it wanted to do, and tried to control its wayward behaviour by hitting it violently or talking to it in anger and frustration (Della Sala et al. 1991).

For the third desideratum, we need arguments to the effect that presence of the sense of authorship is typical and nonproblematic, whereas absence of the sense of authorship is the exception that needs to be promptly detected, for the individuals to perform adequately. My first argument draws attention to the notion of ‘naked intention’, or more precisely why we should reject it. The idea of naked intention suggests that there can be awareness of an action without an awareness of who the agent is (Jeannerod and Pacherie 2004)—or “agent-neutral” action experience (Pacherie 2007, 16). Jeannerod and Pacherie argue for it on the basis of neurological evidence. The same areas of the brain are activated when I engage in intentional action or when I see another engage in the same or similar intentional action. The reasoning involves some mirror neurons and shared representations. On this basis, they claim that the activation is neutral in regard to who is doing the action. So they postulate the ‘who’ system in the brain. But this argument is based on an invalid inference from sub-personal brain processes to phenomenological conclusions. Even if we grant that ‘who’ system identifies the agent of the intention, there is no reason to think that there is an isomorphism between sub-personal mechanisms at the level of brain hardware and phenomenal level of experience. The ‘who’ question is possibly relevant at the level of brain hardware but it never comes up at the phenomenal level because neural systems have facilitated the answer. Even if I am wrong about ‘who’ the agent of an action is, something that happens in anarchic hand syndrome and other cases of delusions of control, I am perceiving or experiencing the action as already specified with respect to agency. There is no experience of actions without an agent. This should give us some reason to think that the default option is to consider our actions as being authored by us as agents. That is to say that there is a presumption of authorship built into one’s actions, unless there is reason to believe otherwise.

My second argument for the presence of the sense of authorship as default option is based on an examination of Gallagher and Zahavi’s claim that there is a first-order, pre-reflective, non-conceptual, primitive experience of agency (2012, 189). Insofar as sense of authorship of one’s action is the core of the sense of agency, this may lead some to believe that there is some positive phenomenology or feeling associated with being the author of one’s actions. But this would be an error. I draw attention to Hume’s famous remark concerning the denial of self: “For my part, when I enter most intimately into what I call myself, I always stumble on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. I never can catch myself at any time without a perception, and never can observe anything but the perception” (Treatise, 1.4.6.3). Hume is complaining that he fails to find a phenomenological marker of the self in experience. Zahavi and Kreigel (2016) argue that this way of looking at Hume’s complaint suggests that there is a separate self-quale that one can consult in one’s phenomenology in isolation from the content of consciousness. This is a mistake; the pre-reflective self-awareness does not deliver a datum or a quality like the smell of fermented garlic. They claim,

… it is not supposed to be any specific feeling or determinate quale at all. Nor is it supposed to be a synchronic or diachronic sum of such contents of consciousness (or any other relation that might obtain among such contents). Our view is not that in addition to the objects in one’s experiential field—the books, computer screen, half empty cup of coffee, and so on—there is also a self-object. Rather the point is that each of these objects, when experienced, is given to one in a distinctly first-personal way. On our view, one does not grasp for-me-ness by introspecting a self-standing quale, in the same way one grasps the taste of lemon or smell of mint. … In other words, the ‘me’ of for-me-ness is not a separate and distinct item but rather a ‘formal’ feature of experiential life as such. (Zahavi and Kreigel 2016).

This implies that pre-reflective awareness of the self does not deliver self-standing quale. In a similar vein, we might say that the pre-reflective awareness of authorship is not a specific feeling or a determinate quale that we can consult in our phenomenology—an author that we can detect in performing in action. That there is an author is just a formal feature of the structure of agency. To say that the pre-reflective sense of authorship is a formal feature of the structure of agency is to deny that there is any positive phenomenal content interlaced into the experience of authorship. But it is also to say at the same time that the presence of the sense of authorship is the typical case of action.

Furthermore, there is reason to believe that the absence of the sense of authorship is the exception that needs to be promptly detected for the individuals to perform adequately. Parnas and Handest (2003) argue that the phenomenological manifestations of anomalous self-experience, where one’s experience of being the author of one’s physical and mental actions is distorted, are symptomatic of prodromal phases of schizophrenia and psychosis. They argue that familiarity with subtle, nonpsychotic anomalies of subjective experience of one’s status as the author of one’s actions, among other symptoms, is not just theoretically significant but crucial for early differential diagnosis (2003, 121). It is useful to quote some of the clinical descriptions of the alterations of self-experience (specifically cases that report lack of authorship) to make the case that such absence if promptly detected might allow for timely diagnosis and early interventions before the onset of psychosis. Some of these disturbances manifest themselves in motor performance. Random verbal or motor acts may occur as if they are interfering with one’s actions and speech without being clearly labelled as uttered or performed by oneself or some other external agency as exemplified in Cases 5 and 6 reported in Parnas and Handest (2003, 127):

  • Case 5: A former paramedic reported that many years prior to the onset of his illness he occasionally experienced—while driving in an ambulance and much to the driver’s surprise—uttering words entirely unconnected with his train of thoughts. He immediately continued to speak in a relevant way or make some cliché remark to cover for this embarrassing episode.

  • Case 6: A female library assistant reported that prior to the onset of her illness she was alarmed by a frequently recurring experience that replacing books from a trailer onto the shelves suddenly required attention: she had to think how she was to lift her arm, grasp a book with her hand, turn herself to the shelf etc.

These cases clearly demonstrate that such random bodily and linguistic actions, which are associated with the positive experience of the loss of authorship, need to be promptly detected to avoid full-blown symptoms of schizophrenia and psychosis.

The fourth, the last desideratum concerns the adequacy of the default theory to account for known pathologies and distortions. The most striking illustrations are delusions of alien control in schizophrenia where a subject is aware of the content of the action she is executing but denies being the agent of this action. According to the default theory there is a presence of an experience—albeit not a veridical experience – of loss of authorship. The patient has a positive experience of some other external agent that causes their hands to move (anarchic hand syndrome) or inserts thoughts into their head (thought insertion). The positive experience, although mistaken, accounts for the vividness and fine-grainedness of the delusional phenomenology. The patient has a positive experience of not being the author of her actions possibly because of neurological or other mental disorders.

This completes my case for a default theory of the core aspect of sense of agency. To conclude, there is no positive phenomenology associated with the experience of agency. There is nothing like to be an agent; no experience of agency. The Buddhist-Abhidharma philosopher can hold on to the no-self doctrine without the extra burden of having to explain the phenomenology of agency.