Let us therefore consider how we might argue in favor of a soul, consisting of a Subject and his consciousness and an Agent and his will. If I do not mention feelings much here, it is only because I consider them derivatives of the other two powers of the soul; but the soul as author of evaluations (value-judgments, choices, affections) is intended here too.
As already stated, I agree that the soul has in itself no perceptible (i.e. visual, auditory, olfactory, gustatory or tactile) qualities, comparable to those in or around the ‘body’ (matter) or in mental projections (imaginations, dreams). This can be taken to simply mean that it is not made of material or mental substance, granting that “matter” (in a large sense, here, including physical and imaginary concrete phenomena) is whatever has these qualities; for this reason, let us say that soul is made of some distinctive substance, call it spirit. All we have done here is hypothesized, by analogy to the phenomenal realm, an entity (soul) of different stuff (spirit); this is logically legitimate, provided we go on and justify it further.
This concept of a soul is constructed to explain certain phenomena, on the basis of a mass of observations and theory-building. The soul is posited as the Subject of consciousness (or cognition) of, first, concrete phenomena (percepts) and, second, abstract appearances (concepts); and at a later stage as the Agent of will, the presumed cause (in a special sense) of certain perceptible actions of bodily organs (eye movements, speech, motions of arms and legs, and so on) as well as of intellectual organs (imagination, attention, thought processes, and so on). But if soul is reduced to such a conceptual construct, we only succeed at best in giving a general description of its powers and activities.
Such a theoretical approach leaves us without justification for our day-to-day propositions concerning our own particular thoughts and deeds at any given time. For conception cannot proceed from a single event; it is the outcome of comparisons and contrasts between two or more events. Whereas, statements about an individual person’s present situation are not made in comparison and contrast to other persons or situations. A general proposition can serve as major premise of a syllogism, but to obtain a particular conclusion, we need a particular minor premise. Indeed, to obtain the general proposition in the first place, we need to admit some particular cases of the same kind, which we can then generalize and apply to other particular cases (that is what syllogistic inference is all about).
That is, when we say, for instances, “I believe so and so” or “I choose so and so” or “I wish so and so”, we are evidently not referring to phenomena perceptible at the moment (belief, choice, wishing have no immediate concrete manifestations, though they may eventually have perceptible effects), and we are evidently not conceptually inferring such propositions from any perceptual phenomena (i.e. what these propositions refer to are not abstract appearances). Yet these propositions are significant to each of us, and can fairly be declared true or false by us. Their truth or falsehood is, to repeat, not exclusively based on experience and on rational considerations, as Buddhists suggest, but is immediately, directly known by introspection.
This is what I would call ‘self-knowledge’; and since this type of cognition is neither perception nor conception, it deserves a special name – say, ‘intuition’. My use of this term should not be taken to imply acceptance of knowledge of other people’s souls, thoughts, wills or emotions (which is another issue, open to debate, solipsism not being excluded) – it is here restricted to self-intuition. I do not use the term ‘introspection’, because this may be used with reference to perceptible phenomena, such as one’s mental imaginations or bodily feelings.
Thus, in this view, the soul is cognized by three types of cognition: directly by intuition, and indirectly by conceptualization based on the soul’s perceptual effects and its intuited states and activities. Of course, ‘cognition’ is one and the same in all three cases; only the object of cognition differs in each case. If we limit our consideration only to perceptual effects and concepts derived from them, we can only construct a theoretical ‘soul’ and refer to ‘powers’ of soul. To obtain and claim knowledge of an individual soul and of its actual perceptions, conceptions, beliefs, intentions, acts of will, value-judgments, affections, etc., we have to admit a direct cognition other than perception, namely ‘intuition’.
Thus, we could refer to soul with several terms: the ‘I’ of my own intuitions, the ‘self’ when assuming that others have an ‘I’ like mine (on the basis of similar perceptible effects), and the ‘soul’ when referring to the conceptual construct based on my ‘I’, your ‘I’ and their perceptually evident (presumed) effects. Granting all this, it is no wonder that if we seek definition or proof of the ‘I’ in phenomenal effects, we will not find it!
Let us now return to these intuited propositions, for a moment. Consider this well. If I say to you “I believe (or disbelieve or am unsure about) so and so” – did I infer this from anything and can you deny me? Sure, I have to mean what I say to you, be sincere. Sometimes, too, I may lie to myself, and claim to believe something (e.g. some complimentary claim about myself, or some religious or political claim), when in fact I do not really believe it. The human psyche has its complexities, and we can hide and not admit things even to oneself. In such cases, the truth of the statement can be verified with reference to a larger context, checking if my feelings and actions are consistent with my claimed belief. But this does not mean that all such personal claims are known by reference to perceptible side-effects, as Buddhists claim. It only means that, just as in the perceptual and conceptual fields, appearances have an initial credibility but have to be faced off with other appearances, so in the field of intuition, an inductive process of verification goes on, through which some intuitions are found to be doubtful (due to their conflicts with other intuitions, and/or perceptible phenomena and conceptual considerations).
Furthermore, it should be stressed that not all statements of the form “I-verb-object” (object being optional) are based on intuition alone. Some have perceptual and/or conceptual basis only, or also. For example, “I am thinking that we should go there” involves perceptual factors, perhaps a mental image of our bodies (mine and yours) walking along in some direction, as well as conceptual factors, perhaps a reasoning process as to why we should go there. But some such statements are purely intuitive, e.g. “I believe so and so” is final and independent, whatever the reasoning that led up to the belief. Furthermore, such statements need not be verbalized. The words “I”, “believe” etc. involved in the statement are of course products of conceptualization; but the intent of the sentence as a whole is a particular intuition, which the words verbalize.
Also to note well is that a proposition like “I believe so and so” cannot be based on a coded message from the brain, to the effect that “so and so should be declared as ‘your belief’ at this time”, for the simple reason that we have no awareness of any perceptible message of this sort. Therefore, such a statement is not a translation in words of a special kind of percept (just as conceptual statements are not). Perhaps the statement “I believe so and so” itself is the perceptible message from the brain? If so, we would be justified in denying any intuition of soul and its states and activities. But it is evident from introspection that we know what we want to say before we put it in words. The words merely verbalize an object already cognized; and this cognition must be ‘intuition’, since it is neither perception (having no perceptible qualities) nor conception (since it is particular).
It seems justified, in conclusion, to hypothesize, in addition to perception and conception, a third source of knowledge, called intuition, a direct cognition whose objects are the self (I) and its actual cognitions (I know what I am seeing, hearing, imagining, thinking, etc., right now), volitions (I know what I choose, decide, want, intend, will, etc., at this moment) and affections (I know what I like or dislike or am indifferent to, what I hope or fear, etc., at this time). I know these most intimate of things – who can tell me otherwise, how would they know better than me what the imperceptible contents of my consciousness are? Soul and its presumed powers – cognition, volition, affection – cannot be conceived by comparison, since I do not see any souls other than my own; it can only be conceived by inference from perceptible and intuitive phenomena that we hypothesize to be its effects. The objects of intuition may be “empty” of perceptible qualities; but they may still have an existence of sorts, just as abstracts are not themselves perceptible but may credibly be affirmed.
Suppose, for example, I meditate, watching my breath; my random thoughts cause my attention to stray for awhile; I drag my attention back to the object of my meditation, my breath. Here, the direction and intensity of my attention require an act of will. The straying away of attention from the breath is not my will; my will is what makes it return to the breath. Phenomenally, the attention on the breath and the loss of this attention (or rather the breath phenomenon and the lack of it) are on an equal plane. What allows me to regard the one as mine and the other as not mine, is the awareness that I had to make an effort in the one case and that no effort was involved in the other case. This ‘effort’ is the intuited volition and that it is ‘mine’ signals intuition of soul. I may focus on the effort alone, or by an additional act of will focus on the fact that it is mine. There is no ‘reflexive act’ involved in this self-consciousness, because it is one part of me watched by the rest of me.
Of course, this is all very mysterious. When we say “I think this” or “I will that”, we have no idea where this or that event came from or how it popped up. Certainly the deep source and manufacture of a thought or will of the soul is unknown to us, so we cannot claim to wholly own it. We do not have a plan of action before the thought or will, through which we consciously construct the latter. Each thought or will, finally, just is. There are no steps or stages, we just do it. But it is still not just happenstance; there is an author, ourselves. We are able to distinguish, in most cases, between thoughts or wills that just ‘happen to us’, and others that ‘we author’; we may even identify them as voluntary or involuntary to various degrees.
All this to say that Nagarjuna’s critique of soul and its powers, and of the knowability of these things, is far from conclusive. Buddhists are justified in doubting and inquiring into the issues, but from a purely philosophical point of view the Madhyamika conclusion of “emptiness” may be considered too radical and extreme. It may be obviously valid from the perspective of someone who has reached some higher form of consciousness (which, I know, I have not), but their rational arguments are not decisive. Most important, as we have seen, Nagarjuna bases his denial on one particular theory of soul (the atman theory), and has not considered all conceivable theories. To rebut (or more precisely, to put in doubt) his arguments, it is therefore sufficient to propose one alternative theory (as above done) that he has ignored; the alternative does not need to be proved – if it is just conceivable (coherent, consistent), that is enough.
Nagarjuna does not, in my view, satisfactorily answer questions like ‘who is it that perceives, thinks, desires or acts?’, ‘who is it that meditates in pursuit of liberation or eventually reaches it?’, when he explains away the soul as a mere cluster of percepts or concepts, as something (illegitimately) inferred from perceptible phenomena by a presumed cause-effect relation.
In passing, it is worth noting that, although the doctrine of no-self is fundamental to Buddhism, not all Buddhists have interpreted it as a total rejection of soul (in some sense of the term). One Theravada school, known as the ‘Personalists’, dating back to about 300 BCE, whose adepts in the 7th century CE included almost one third of all Buddhist monks in India, “motivated by commonsense, maintained that in addition to impersonal events, there is still a ‘person’ to be reckoned with.” According to the Abhidharmakosha, a Mahayana work by Vasubhandu (4th century CE), the Personalists interpreted the no-self doctrine of the Buddha as signifying simply that “something which is not the true Self is mistaken for the true Self”.
It is thus possible to understand the doctrine of not-self as a rejection, not of ‘soul’ (‘real or deep self’), but rather of ‘ego’ (‘conventional or superficial self’). The ego is a confused construct of ‘selfhood’ by the soul, due to the latter’s self-identification with delusive opinions (acquired by itself and through social influences), and consequently with certain attitudes and actions it engages in, in the way of a self-protective reaction. By predefining itself and its world, the soul imprisons itself in patterns of response appropriate to that definition. It is up to the soul to rid itself of the ego-centered viewpoint, by realizing the stupidity and avoidability of it.
 We can leave as an open issue, parenthetically, the possibility that matter and spirit are respectively coarse and fine manifestations of one and the same substance.
 As we meditate, countless thoughts pop up, tempting us to follow them. Eventually, one manages to hook us, grabbing our interest and hurtling us through a series of associations. Thus totally absorbed, we forget our object of meditation for a while, until we realize we have been distracted.
 The thoughts I strayed into may have involved voluntary processes, but my straying into them was involuntary.
 According to Edward Conze, in Buddhist Scriptures (Penguin: England, 1959). See pp. 190 and 192-7.