5. Percepts and concepts.
According to pre-Mahayana Buddhist (and other Indian) philosophers, the world we experience and think about is composed of “dharmas”. This term has various meanings, but the one focused on here seems to be equivalent to what we would call a phenomenon, or perhaps more broadly an appearance. A phenomenon is an object of experience; an appearance is an object of cognition of any kind, whether perceptual (phenomena), intuitive (objects of ‘self-knowledge’) or conceptual (objects of rational knowledge, ‘universals’). Dharmas are “momentary, particular and multiple”; they are “not supported by substance or self” yet have their “own or independent nature”; they are “distinct and separate, yet appear and disappear in accordance with the principle of causality”. Nagarjuna denies the “reality” and intelligibility of dharmas, using the following main arguments.
(a) He argues, “a momentary entity or impermanent dharma” can be “divided into non-enduring or non-abiding” segments, each of which “has, analytically, no duration whatever. It disappears as soon as it appears. Therefore, it cannot be said to have true existence.” Yet, it is “supposed to have some duration.” Whence, “to say that an entity is impermanent is tantamount to saying that what abides is non-abiding,” which is “a contradiction in terms”.
(b) Against the contention that “impermanence” does not signify “non-duration”, but refers to “the reality of the phenomenal” that each thing “arises, endures for a moment and then ceases to be”, Nagarjuna replies: “how [does each of these three] characteristics characterize a dharma?” Is it “simultaneously or successively”? It cannot be simultaneously, because “origination, duration and cessation are opposed by nature: at the time of cessation there should not be duration, and at the time of duration there should not be cessation.” It cannot be successively, because if the characteristics occur at different times, there would be three different phenomena” and “how can different phenomena be true of the ‘same thing’?”
(c) Furthermore, he argues: these three characteristics – origination, duration and cessation – must be either “created” or “non-created”. If they are the created, then each of them should in turn “have the three characteristics”, each of which in turn, “like other created things,” should have them, and so on ad infinitum. If, on the other hand, “each characteristic is non-created, how can it characterize a created thing?” In either case, then, we have a “conceptual problem”.
(d) Moreover, he argues: “what is the relation between an object and characteristics?”. Are they “identical” or “different”? “If identical, there would be no distinction between them, and it would be absurd to say that the one is object and the other, characteristics.” Nor could one say that they are identical in the sense that the object is “the substance of” the characteristics, and the characteristics are “the manifestation of” the object. For to do so would, according to Nagarjuna, imply their relation to be “reflexive”, and therefore that “a thing would be subject and object at the same time”, which is “clearly impossible, because subject and object are different”. If, on the other hand, an object is “different from” its characteristics, “there would be no internal connection between them.” Therefore, “characteristics characterize objects” cannot be said.
(e) He also argues, “whatever can be conceived to exist has a cause. All things are produced by a combination of various causes and conditions. When the conditions change, things will also change and even disappear. To exist means to be caused, conditioned, generated or dependent on something. But by definition a dharma is an entity which has its own or independent nature.” Whence, he concludes, “to say that a dharma exists would be the same as saying that an independent thing is dependent”, i.e. the claim “dharmas exist” is “a contradiction in terms” and “absurd”.
Nagarjuna concluded from these arguments that the concept of dharma upheld by his predecessors, Buddhist or otherwise, was confused and untenable. Reality could not, therefore, be understood through such conceptual tools. But let us now look at his arguments more closely and critically. As we shall see, they are far from conclusive, and generally fallacious.
Argument (a) is simply a claim that when a duration of time (moment) is infinitely divided, its constituent points of time (instants) have zero duration and, therefore, cannot be said to exist. This argument is already known to Western philosophy through the paradoxes of Zeno of Elea (born c. 490 BCE), and has been amply contested since then by many philosophers, mathematicians and physicists, on various grounds. My own (additional, yet essential) objection to it would be that Nagarjuna here fails to analyze how and in what order the concepts he uses arise.
What is under discussion here (viz. the dharmas), are primarily phenomena, empirical givens. In fact, at any one moment of experience, what we perceive is one holistic phenomenon; the ‘cutting up’ of that total phenomenon into smaller, individual phenomena (different shapes, colors, sounds, etc.) is not in itself perception, but one of the first rational acts. We experience things in flux – coming, staying a while, going. To understand such motion, we construct a concept of time, which we gradually refine (with measurements, mathematics, Relativity theory). Motion is an experience, but time is a concept. The concept of time arises in response to the experience of motion, so it has to be tailored to fit and cannot be used to deny such experience. If a conflict occurs between the two, it is the concept and not the experience that has to be put in doubt and adjusted.
Now, what is the ‘infinite division’ of a phenomenon that Nagarjuna appeals to? It is not a physical act of slicing a phenomenon with a knife, or anything of the sort. For we have no experience of infinite division in the physical realm; we may subdivide a material body or draw lines on a piece of paper or a computer screen only so far, not ad infinitum. Infinite division is an imaginary act. If the phenomenon is of the ‘material’ kind, the division may occur on a ‘mental’ image of it; if the phenomenon is already of the mental kind, the division can occur directly on it. But even in our heads, we do not in fact divide infinitely. We may slice the image, then mentally ‘zoom in’ and slice that slice, then zoom in and slice again a few more times, then we stop.
Now, the zooming in is merely production of a new image – so we are not even, in fact, repeatedly subdividing the same image; we merely say ‘suppose this image is a detail of the preceding’. The new image has the same size as the preceding, but its scale is declared different. Furthermore, the subdivision process takes time, and we do not anyway have an infinity of time – so we have to stop it after a few sample shots, and then say ‘suppose I repeat this to infinity’. Thus, infinite division is not even a real act in the mental field, but a mere verbal statement – i.e. at best, a concept referring to the intention to ‘cut’ and memory of recurrent events, projected to a hazy ‘infinity’.
Furthermore, when we imagine division of a (two-dimensional) phenomenon, we imagine (one dimensional) line drawn somewhere in the middle of it. But how is the geometrical entity known as a line (length devoid of width) first conceived? It is derived from experience of the visible boundaries of phenomena (with length and width) in relation to their surrounds; there has to be some difference between the two sides of a boundary for it to be visible. A line in the middle of an extended phenomenon is thus partly a concept, and not a pure percept. We never entirely see a line, we always have to some extent think it. We have to effectively accompany it with the thought ‘this line has no width’. Thus, the visualization of division does not in itself prove infinite divisibility.
Nagarjuna, for all his supposed meditative introspection, has clearly not paid attention to how his concept of ‘infinite division’ arose in detail. His argument or ‘thought experiment’ is without substance, because he has in fact certainly not engaged in ‘infinite division’. He has not shown experientially that dharmas of zero extension in time are the building blocks of dharmas with duration in time. He has therefore not demonstrated that a contradiction exists in the concept of momentary dharma.
Let us now move on to argument (b). It is true of all phenomena that they are momentary. It does not follow that all existents are momentary, but that need not concern us here. Nagarjuna’s predecessors or opponents are quite correct in their analysis of the momentary as something that appears, endures awhile then disappears. Nagarjuna is correct in saying that these three characteristics are opposed, i.e. cannot occur simultaneously. But his definition of simultaneity as “at the time of” is vague and misleading. His definition of succession as occurrence “at different times” is also incorrect. Both premises of his dilemma are therefore confused, as we shall now see.
For the arising and the ceasing are conceived as at the temporal boundaries of the duration, and so not as in it nor quite as outside it. Arising occurs at the instant (the unextended point of time) the duration starts, and ceasing occurs at the instant the duration ends. The concept of arising refers to just that instant of flip-over from absence to presence, and the concept of ceasing to just that instant of passing from presence to absence. The coming, staying and going are successive, in the sense that the arising and the ceasing are not simultaneous with each other. But each of the latter is instantaneous and contiguous (and in that sense only, simultaneous) with the duration (at either end of it). They cannot therefore be said to be ‘at different times’ from it. The arising cannot be said to be ‘before’ the duration and the ceasing cannot be said to be ‘after’ the duration; they are not time-consuming processes (though such processes may precede and cause them). The two limits of duration (be it brief or long) cannot actually be dissociated from it. The phenomenon remains one, even as we conceptually distinguish three ‘characteristics’ of it.
We thus see that Nagarjuna’s argument is based on a stupid or deliberate fuzziness of definition. The confusions involved in his dilemma are entirely of his own fabrication; he sows them to have pretexts for criticism. He uses ‘at the same time’ to mean ‘in overlapping durations’ and ‘at different times’ to mean ‘in separate durations’, whereas what is under discussion is instants which are the edges of a duration. No wonder then that he concludes that there is either contradiction or separation.
Now study argument (c). Two arguments are intermingled in it – one relates to the hierarchy of concepts and percepts, the other relates to causation.
Nagarjuna claims that the three stages (arising, staying, ceasing) of each phenomenon may be viewed as in turn a phenomenon. What he relies on here is a reification of the first and last stages; he tacitly implies that because they have separate names they too have durations. The distinctions between the three are thus erased. If we consider the conceptual development involved, we see that, in a first phase, ‘phenomenon’ refers to a unit of perception (a piece of the perceptual field isolated by mental projection, to be exact); in a second phase, we distinguish within this event or thing an instantaneous beginning, a momentary middle and an instantaneous end, and accordingly form concepts of arising, enduring and ceasing. The latter are abstract aspects of the concrete phenomenon, and therefore in a sense ‘present in’ it and ‘part of’ it.
But contrary to what Nagarjuna suggests, it does not follow that arising and ceasing in turn have a beginning, a middle and an end – since they are instantaneous. It does not even follow that the middle part of the initial phenomenon has another beginning, middle and end – since we have already abstracted the two ends of the phenomenon away from its middle. We thus have no basis for an infinite regression of concepts; we remain only justified in having one concrete phenomenon and only three abstract aspects of it. The “characteristics” are phenomenal in the sense of being distinguishable in a phenomenon; but not being themselves ‘divisible’ in the same way as it, they cannot rightly be called phenomenal in the same sense as it. One cannot say that arising both arises and ceases at once, or say the same about ceasing; because neither of them has duration; that which arises has to be absent for a while then present for a while, and similarly in the opposite direction with ceasing.
Furthermore, whatever produces the primary phenomenon also produces the three aspects of it we have distinguished in it; they do not require additional causes that will separately produce them. Even if we regard, as did Nagarjuna’s philosophical forerunners, everything in the phenomenal as having been “created” (in the sense at least of being produced by preceding causes and conditions), perhaps in an infinite chain, it does not mean that such causality forks out repeatedly and endlessly.
The “thing” caused, with all its characteristics, is one. Ordinarily, the cause causes arrival, a minimum stay, and if the event is momentary thereafter a departure. We may in some cases identify something as causing the arrival of that thing; a second as causing its staying on; and a third as causing its departure. But even then the cause of the arrival is also partially a cause of the staying on and of the departure, since without arriving a momentary event would not be able to stay or depart. Also, the cause of the staying on is a partial cause – in a negative sense of a hindrance – of the eventual departure. In such cases, however, ‘the cause’ of the phenomenon as a whole would simply be composed of a series of three subsidiary ‘causes’ – one determining the arising and a minimum momentary stay, the next prolonging the duration after arrival and preventing ceasing, and the last interrupting duration and determining ceasing. This is merely an analysis of causation and not a multiplication of causes ad infinitum.
Thus, we have replied to Nagarjuna that the thing characterized is not apart from its three characteristics, and they do not in turn each have three characteristics. Also, the respective causes of the three characteristics together sum up to the cause of what they characterize, and its cause is not apart from their causes. Nagarjuna gives the impression of making logical analyses, but in fact he glosses over details and nuances at his personal convenience. His arguments give an appearance of structure and order, but beneath them lies a great carelessness in observation.
Now study argument (d). Are an object and its characteristics “identical” or “different”? An individual object could be regarded as the sum total of all characteristics, permanent and transient, observable in it. More precisely, if (or so long as) one or several, or one or several combination(s), of these characteristics is observed in the object and never in any other, we may consider every such single or collective characteristic as a sign of the object, i.e. as signifying its individuality or essence. The single or collective characteristic(s) exclusive to an object could thus be regarded as “identical” with it for all intents and purposes, without however wholly equating it/them to the object. For the object as a whole should be taken to include its non-distinctive attributes or actions, as well as its distinctive essences.
So the answer to Nagarjuna’s question is as follows. His terminology is as usual lacking in nuances; for this reason, the choices he gives us seem restrictive and force us into dead ends. We have to first distinguish essential (distinctive) characteristics (or sets of them) from common (non-exclusive) ones. The individual object is the totality of its facets and history, including both these types of characteristics. The essential characteristics could be considered as the “substance” of the object; the non-essential ones, as its “manifestation”. This would avoid any implication of “reflexive” relation. Thus, we can regard some characteristics as “identical” with the object (without however meaning equal to it); and others as “different” from it (which does not imply them disconnected from it). And we can well say that “characteristics characterize objects”, while remaining aware that the subject and verb of this proposition are of variable meaning.
Of course, none of this tells us what the “relation” between an object and its characteristics precisely is, i.e. in what sense the later ‘belong’ to the former. We have above just accepted that there are relations, which we can in practice identify by observation and distinguish between statistically. To better understand the relational aspect, we need to develop a theory of ‘universals’ – what are these things and how do we know them? What we perceive are concrete objects; the ‘universals’ are abstractions from these phenomena.
Abstraction is performed by comparisons and contrasts between present phenomena and/or presumed memories of past phenomena. Abstracts are apparent as the various measures or degrees in the wave motions that constitute phenomena. Phenomena of light, sound, etc. have various intensities, frequencies, etc. These quantitative or mathematical variations are inherent in the phenomena of perception; some are measured roughly and ‘instinctively’, others, through conscious experiment and careful calculation. In either case, rational work is required to distinguish them out from their perceptual context, and from each other; and to name, interrelate and classify them; and to keep our theses concerning them logically consistent. For this reason, we regard them as objects of another level of cognition, the conceptual, and say that abstracts are known by conception.
In the Buddhist tradition preceding Nagarjuna, “dharmas” are already said to be “empty”. This can be rationally understood to mean, not that objects are devoid of essential characteristics (in the sense above defined), but that there is nothing non-phenomenal (or noumenal) to consider behind the phenomenal. I would agree with this proposition, and submit that when other Buddhist philosophers combat the idea of “essences”, they are not denying that abstract characteristics are distinguishable within phenomena and that some of those are distinctive, but are denying a particular philosophical development, namely the notion that “an object” is more than (or even other than) its evident phenomenal aspects and the inductively justifiable abstractions therefrom (which, to repeat, are merely measurements). The doctrine of “emptiness” initially opposed such fanciful reification as sidetracking our attention, and recommended we remain focused on what is in fact apparent to us. Knowledge is knowledge of actual phenomena, not of some imagined ‘reality’ behind them.
A lot of the confusion in this issue is due to failure to make two distinctions. If we perceptually knew all the phenomena ever existing in the universe, we obviously could not logically claim that there might be any further phenomenon hidden behind them. But because we conceptually know (having memory of our changing scope of knowledge, and in any case the uncertainty at all times that we have perceived everything) that we have access to only some of the phenomena in the universe, we can legitimately suppose that there might be yet unknown phenomena to consider, and that these might in yet unknown ways affect known phenomena. Furthermore, even if the totality of existents appeared to us, i.e. even if we experienced everything that ever is, was or will be, on a concrete level, we could still additionally abstract their similarities and differences, and their statistical regularities and irregularities, and point to such abstract aspects as underlying substrata or causes.
Thus, two distinctions are called for. The first is a distinction between a theoretical perceptual omniscience, from which viewpoint by definition no hidden phenomena are conceivable, and a practical relativity of knowledge to limited perceptual context, which viewpoint allows for supposition of unknown but subterraneously operative phenomena. In the former case, ‘existent’ and ‘apparent’ are co-extensive, but in the latter case ‘existent’ is a genus of ‘apparent’. Secondly, neither of these absolute and relative positions excludes a category of being and knowing other than the perceptual, viz. the conceptual, from being appealed to. In both cases, abstracts can still be posited as ‘underlying’ concretes. Here, the concept of ‘apparent’ is enlarged to include not only concretes (phenomena) but also abstracts (universals).
On this basis, we can ask what Buddhism means when it says that “dharmas” are “empty”. Does it mean that phenomena have no other phenomena behind them? This may be affirmed by a proven omniscient Subject, but the rest of us have to always concede that there are probably phenomena hidden to us (as we often discover later), which may impinge on those known to us. Does it, alternatively, mean that concrete appearances (phenomena) have no abstract appearances behind them? This cannot logically be claimed without self-contradiction, since such a claim is itself manifestly abstract; the fact of the claim must itself be taken into consideration. One may legitimately argue, discursively, about the objectivity or subjectivity of the abstract, but not about its ultimate validity in some way. Also, whether the abstract is present in the object or in the subject, it still abides – at least in the sense that there is no time duration when it is absent from existence.
Nagarjuna’s doctrine of “emptiness” includes not only the previous denial of a noumenal world, but equally denial of the phenomenal world. It is an attempted one-upmanship on his predecessors. They were anti-rationalist, in the sense of rejecting a certain excess of rationalism, a sickness or error of rational projection that ignores, obscures or eclipses experience. He typically takes a more radical and extreme posture and rejects all rationalism indiscriminately. But this is really a rejection of experience, a claim that ultimate reality is beyond it – i.e. it is in effect another form of noumenalism, a return to the sickness his predecessors combated. He pretends that his conclusion can be reached by logical means; but his means are evidently not logical.
Finally, consider argument (e). Nagarjuna takes as one of his premises that all conceivable existents have causes of some sort. But that is debatable. We might accept a statement that all phenomena (i.e. perceived existents, concretes) have causes – though even that is debatable. For such a general statement can only at best be known inductively, by hypothetical generalization from cases where causality has specifically been established; strictly speaking, it is also conceivable that some phenomena (or perhaps some unperceived concrete existents) are eternal or spontaneous or free (i.e. uncaused in some sense). But what of conceived existents (abstracts) – do they also, as he claims, all have causes? That is even more debatable. When we speak of a kind of thing causing another kind of thing, we more precisely mean that instances of the former cause instances of the latter. As for large abstractions, like God or the universe as a whole, or even just existence, we can conceive them as existing without cause.
As a second premise Nagarjuna takes the idea of his philosophical predecessors or opponents that a “dharma” has “its own or independent nature” as meaning that it is independent of causes. But this is not their intended meaning, which is only that dharmas are “distinct and separate”, i.e. each have a specific nature of their own. This is evident in their explicit position that, as we have seen, dharmas “appear and disappear in accordance with the principle of causality”. So Nagarjuna is playing on the equivocation of the term “independent”. He does so to load the dice in favor of his desired conclusion, making it seem as if they made self-contradictory claims about dharmas.
Nagarjuna thus has not disproved the statement that dharmas exist. And in fact such a statement has no need of rational proof, if it is understood to mean that phenomena exist, for that is empirically evident. We know for sure of the existence of “existence” only through the experience of phenomena. The concept of existence is based on that of phenomena, enlarging the latter to include hypothetical unperceived concretes, and at a later stage hypothetical abstracts and hypothetical objects of intuition (self-knowledge).
What, anyway, do we mean by the “nature” of a thing? My understanding of the term refers to the ‘laws’ of behavior of the thing, signifying that things exhibit certain regularities of behavior (being or doing). For instances, something may have character X or do X always (while in existence), or only when Y occurs. Apparently, in our universe, things cannot be or do just anything we imagine for them. Maybe, if everything is just energy, they ultimately can; but the world as we observe it so far seems to contain things with limited behavior possibilities. We acknowledge this apparent fact by saying that existents have a ‘nature’. We do not thereby imply them independent of causes, as Nagarjuna suggests, but on the contrary say that if things have causes, they have a nature. Moreover, even something without causes may have a nature, if it has limited behavior patterns. Only something not subject to ‘law’ at all has no ‘nature’.
Phenomena may yet be ultimately not subject to ‘law’, i.e. devoid of ‘nature’. But to support that thesis, Nagarjuna ought rather to have emphasized, like his predecessors, the positivistic idea that phenomena exist in succession, each moment caused by a previous and causing the next, without an underlying continuity between them across time. This concept remains conceivable, if we gloss over our observations of regularity, arguing that regularity is only known by generalization. But generalization is justified as follows. We observe certain things that are X to always be Y; we infer that all X are Y, because we refuse to assume that there are Xs that are not Y until we have observed such negative cases. On the other hand, to refuse to generalize would be to admit such imagined changes in polarity without empirical basis.
Thus, generalization (duly controlled by particularization, when new observations belie it) is a more empirical rational act than non-generalization; it makes less assumptions. I have observed some Xs that are Y, and maintain that all are since some are; but I have not observed any Xs that are not Y, so how can I presume the latter possible without specific additional reasons? The notion that anything might become anything is thus a very hard thesis to prove – one would have to observe everything eventually turning into everything else, one could not appeal to any generalization whatsoever. One would also have to explain why different things were transformed in different sequences. One would therefore have to be omniscient to prove such a thesis. Or one would have to find some convincing indirect theoretical reason to believe it, such as experimental and mathematical evidence that all energies are convertible into all others (a unified field theory), which neither Nagarjuna nor anyone has succeeded in doing yet.
To summarize, all five arguments proposed by Nagarjuna in relation to the concept of dharmas are faulty (the three middle arguments being inconclusive dilemmas, the other two not self-contradictory), and indeed probably intentionally so. It is not the concepts he attacks that are absurd or contradictory, it is his own discourse that merits such condemnation. It may seem incredible that so many people for so many centuries have studied his work without crying ‘foul!’ – but, what can I say, that is the way of the human psyche. It can allow itself to be intimidated by someone’s prestige and submit unthinkingly to authority, or to gloss over incredulity in response to a promise of salvation dangled appetizingly before it.
 See Cheng, pp. 76-78, on this topic. He there refers to MT VII:1-2,23,25 and XV:1-2, as well as TGT IV:2, VI:1, VII, VIII:1.
 See Cheng, p. 128, for a list.
 It is not clear here whether specifically the three characteristics of arising, enduring and ceasing are meant, or more generally any characteristics. But it does not affect the argument.
 In Buddhist philosophy, causes are relatively internal or direct, conditions are relatively external or indirect. But the word ‘cause’ may also be taken more broadly, to include such causes and conditions indiscriminately. See Lotus in a Stream, by Hsing Yun (New York: Weatherhill, 2000), for more details (pp. 80-82).
 See Ralph E. Kenyon Jr, Atomism and Infinite Divisibility, a doctoral dissertation presented to the University of Massachusetts, Amherst, 1994. The full text is available on the Internet at http://www.xenodochy.org/rekphd/.
 See Appendix 1: fallacy G.
 See Appendix 1: fallacy E.
 See Appendix 1: fallacy F.
 See Appendix 1: fallacy G.
 See Appendix 1: fallacy E.
 See Appendix 1: fallacy E.
 See Appendix 1: fallacy G.
 See Appendix 1: fallacy D.
 See Appendix 1: fallacy F.
 See Appendix 1: fallacy G.
 See my Future Logic, ch. 50 and 54-55.
 See Appendix 1: fallacy B.