Causality is a central concept in Buddhism. In Western philosophy, the term is applied generically to causation (a relation of “constant conjunction” between any two events – physical events, and likewise “psychological” events) and to volition (the relation between a conscious being and an action willed by it). For Indian and Buddhist philosophy, an additional connotation of causality is the moral concept that has become colloquial in the West under the name of karma (the belief that good deeds are ultimately rewarded and bad deeds punished, whether in a present lifetime or a later one – and indeed that we have to be reborn to bear the consequences of our actions, at least until we find “liberation” from this cycle). Buddhism additionally (and if I am not mistaken, originally and exclusively) has a concept of “co-dependency” (according to which, roughly put, nothing stands on its own, but everything exists only by virtue of its direct or indirect causal interrelationships with other things).
A definition of causality traditionally cited in Buddhism is: “When this is, that is; this arising, that arises; when this is not, that is not; this ceasing, that ceases.” It is an excellent definition of causation, or more precisely the strongest type of causation – namely complete and necessary causation. It is better than the definition “constant conjunction”, proposed by some Western philosophers, which only refers to complete causation. But the said traditional formula is not accurate. First because there are other, weaker types of causation, namely, complete but contingent, partial though necessary, and neither complete nor necessary – and derivatives of these. And second, because causation does not include volition. In truth, if we study the actual descriptions of “co-dependency” in Buddhist texts, it is easy to see that the causal relations referred to do not all fall under the stated definition of causality (as “when this is, that is”, etc.) but range far more widely over the many other senses of the term.
For earlier Buddhists, and Buddhists of other schools, causality is an objective fact, which gives rise to and implies “co-dependency” and thence “emptiness”. But for Nagarjuna and his school, all these concepts and tenets are ultimately mere “conventional” truths, without real validity. Thus, although they are Buddhist doctrines, and he admits their value as initial teaching tools, he regards it as necessary to ultimately disown them, so as to go beyond the discourse they involve, into non-discursive consciousness of actual emptiness. For him, it is useless and counterproductive to talk about emptiness, to analyze and reason it – it has to be lived. It should not surprise us, therefore, that he tries to disprove causality, to show all concepts of it to be confused and absurd.
(a) Let us first consider Nagarjuna’s argument concerning “production”, as presented by Cheng. It has the same dilemmatic form as some of his arguments about motion and rest. He divides the “process of production” into three parts. The part “already produced” is “finished” and the part “yet to be produced” is “not yet”; so neither of these can be “established”. The part “being produced” can be “established” only if the aforementioned two parts have been; so it too cannot be “established”. Hence, “the act of producing is impossible”; and therefore, “there cannot be a producer”. They are both “unreal” and “involve contradictions or absurdities”.
We can reply as follows. For a start, let us note that Nagarjuna (in Cheng’s account, at least) does not even define what he means by “production”, he merely takes the term for granted. The full causal connotation of the term is admittedly hard if not impossible to define (no one, to my knowledge, has so far succeeded in doing so), but I submit that no concept can definitely be proved or disproved without some definition, so we can doubt Nagarjuna’s “refutation” of production on this ground alone. But let us, like him, take the term as understood (I do suggest a working definition further on), and consider his reasoning anyway.
The first two premises typically rely on a possible confusion in the reader between the present tense (“the act of producing”) and the past and future tenses (“already” and “yet to be” produced). Of course, if we artificially freeze the present tense in the present, as he does, we cannot find (“establish”) it in the past or future tenses. But if we consider the past as having once been the present or the future as the eventual location of the present, there is no difficulty in saying that “the act of producing” was in what is now and since then classed as “already produced” or will be in what is now and until then classed as “yet to be produced”. The reasons he gives in his two premises, “it is “finished” and it is “not yet”, beg the question and do not constitute proof that production cannot be “established” outside the present.
His third premise, that the “act of producing” can be “established” in the present only by being so in the past and future, is gratuitous, and only serves to again demonstrate that his conclusions are foregone. Why would we need to refer to past or future, to infer the present situation? We can well find the “act of producing” directly in the present, by empirical means. Watching someone go through certain motions, which are exclusively and invariably followed by certain perceived changes in his environment, we name the someone “producer”, his motions “production”, and the changes “products” (this sentence, by the way, can serve as an inchoate definition). Clearly, when I say this is empirical, I mean empirically-based. The statistical reasoning involved, and many other underlying presuppositions such as memory of past instances and comparison between instances, are conceptual and logical. We cannot establish production by means of a single present perception, but have to appeal to past perceived instances and ensure that future perceived instances keep concurring. Nevertheless, all that is more direct than what Nagarjuna proposes.
In sum, Nagarjuna’s argument is merely an attempt to delude us, and in no way justifies his conclusion against the concept of production. As for the concept of producer, further discussion is required. I have already, higher up, discussed one issue involved, that of the existence of a substratum to motion – for the producer is conceived as an abiding entity. But of course, we also have to here point to the implied issue of causality. The term production is colloquially applied even to a machine, but ultimately it signifies human invention, initiation and supervision of a process – that is, consciousness and will. A machine is merely a tool of production, not a producer – the latter term only properly applies to human beings (or entities with similar powers). But we need not try to deal with this more difficult issue here, as it does not arise in the context of Nagarjuna’s above argument.
(b) Another argument of Nagarjuna’s relates to whether an effect is “real in” or “unreal in” a cause (“or an assemblage of causes and conditions”). The meaning of this question will become apparent as we develop his answer. For him, the question is fourfold, not just twofold – the effect might also be “both real and unreal in” or “neither real nor unreal in” the cause (see discussion of the tetralemma, above). Cheng relates his argument as follows. First premise: “if an effect is real in a cause,” it does not need to be “produced again” – “there cannot be causal production” since “nothing new is produced” in such case. Second premise: “if an effect is at the outset unreal in a cause and yet is produced by a cause, then in principle anything should be capable of being produced from anything else” – “there would be no particular or distinct relation between the two, and hence one would not be the cause of the other.” Third premise: an effect cannot be “both real and unreal in” a cause, “because real and unreal are contradictory in nature” and such things cannot be “together”. Fourth premise: “to say that an effect is neither real nor unreal in a cause is tantamount to accepting that there is no causal relation between cause and effect.” Conclusion: “none of these can be established, and thus theories of causality should be renounced.”
Many objections can be raised to this argument. For a start, we can again point out that Nagarjuna (or perhaps only Cheng) does not define causality precisely – so how can he succeed in disproving it? Similarly, the expressions “real in” and “unreal in” are left very vague. Nevertheless, we can rephrase his question as, more clearly: is the effect already present or not in the cause? His first two premises are then: if yes, it did not need to be caused; if no, how could it have been caused? My answer is this: Nagarjuna is ignoring or obscuring the (very Aristotelian) distinction between merely potential presence and actual presence. We can say that the effect is to some degree present, in the sense of potential, in the cause; it only becomes fully present, in the sense of actual, under appropriate conditions.
For example: a healthy woman has in her womb, potentially but not actually, sons, grandsons, etc. Concretely, that potential has actual expression in her physiological characteristics (womb, eggs, genetic make-up, etc.); but her descendants are still not actual; they are actualized only when and if certain existential and biological conditions are met (she remains alive long enough, she has intercourse and is fertilized, she bears a son, then her son in turn finds a woman, and so forth). The woman is not the complete “cause”, in any case (other factors come into play, positively or negatively); the son, grandson, etc. are not to the same degree her “effects” (her son is more of an effect of hers than her grandson, etc.) since more and more conditions have to be met.
These concepts are quite reasonable, so Nagarjuna’s attempted denial cannot be upheld. Furthermore, note that the things we call the “cause” and the “effect” do not merit this label until and unless causality takes place (and is known to have done so); it is only after such event that the terms become applicable, so that it is absurd to apply them to things while denying such event, as Nagarjuna tries. Also, it is important to clarify what we mean by “the” cause or “the” effect. Nagarjuna focuses on one of the conditions involved (in the example given, a woman), and ignores the others (her healthy fertilization; successive generations of women and their fertilizations, in turn); likewise, he does not distinguish between direct effects (her son) and indirect ones (her grandsons, etc.).
Thus, his first two premises are misleading – for an effect is potentially present in a cause, in the sense that certain conditions are actual in it; but the effect is not actually present in that cause, because certain additional conditions are not actual in it. When the latter conditions are actualized, they – together with the already actual former conditions – cause actualization of the effect. The sum of the earlier and later conditions may be referred to as “the” cause; whereas each of these sets of conditions can only properly be referred to as “one of the causes”; each is only a potential cause without the other. Similarly, we have to distinguish between an effect of this total cause, and an effect of an effect of it, and an effect of an effect of an effect of it, etc. In each case, additional conditions come into play, and what was admitted the cause of the earlier effect, may only be regarded as a cause (among others) of the later effect. Nagarjuna uses the terms cause and effect without any effort at precision; is it no wonder then that he formulates false premises.
With regard to his claim in the first premise that “nothing new is produced”, then, we would reply that there is novelty in the intensification of existence from a inchoate, potential degree to an overt, actual degree. As for the reasoning he uses in his second premise, the following may be offered as rebuttal.
He argues that if an effect were not present at the outset in a cause, then any other effect might “in principle” emerge from the cause, so that in fact no relation would exist between the cause and any such effect. To make the issue clearer, let us remove the terms cause and effect from the sentence, since as already stated they may not be used before a causal relation is established. Indeed, “in principle” a thing might be accompanied or followed by just anything. This only means that, at first glance, before any data has been gathered, we must have an open mind and look at the facts without prejudice, without anticipation – this is the epistemological principle he seems to be referring to, and the one we would admit. He cannot be taken to appeal to an ontological principle, that everything occurs by happenstance, without connection to anything else – for that would be begging the question, surely. But of course, Nagarjuna cunningly obscures the wide and deep gulf between these two senses of the term “in principle” to make his point.
But this is in itself interesting, because it shows that he is quite aware of the reason why we formulate a concept of causality and believe in it, and of the inductive process through which such a relation is established. If we lived in a world (or field of appearances, to be more precise) where, indeed, anything could happen, i.e. a world without any regularity, then we would have no basis for a concept of causality, and no such concept would even occur to us. But we notice in our experience that some things are constantly conjoined (and so forth – this is just one of the many types of regularity), and therefore – in order to record and explain such uniformities in our experience – we form a concept of causality. Furthermore, we use these very same observed regularities to determine whether or not the label of causality may be applied in particular cases. There is thus nothing arbitrary in the concept, nor in its applications; it is experience that suggests it, and experience that establishes it.
There is nothing circular in the concept, either. Nagarjuna denies that causation can be established with reference to experience, saying that this “begs the question”. Even though “we have seen sesame produce sesame oil, but have never seen sand produce sesame oil,” we are not justified to “seek sesame oil in sesame but not in sand” because “causation or production has not yet been established, and so one cannot legitimately make that claim.” But as just explained, this is not the order of things in knowledge. Both the concept and its particular applications come from the same experiences. We have a concept of causation because we observe regularities and these same regularities tell us where to apply it. There is no basis for a demand that the concept be known independently of experience. The reason why the concept seems to exist “independently” of any one of its empirical instances, is that it is the common feature of many and indeed all instances of regularity, and does not merely refer to the one regularity under scrutiny at one particular time.
Thus, Nagarjuna should have said the following: if some thing and another thing are always apparent together and never apparent apart, then we may call the one “cause” and the other “effect” and their relation to each other “causality” (this is only the strongest form of causation, to repeat, but I do not want to needlessly complicate the issue here); but if no such regularity (nor a lesser degree of regularity) occurs in their appearances (i.e. our experience of them, in a first phase, and by generalization, their existences), then they cannot be called thus. Had he formulated the sentence in this way, he would have had no argument. He chose to confuse the issue or was himself confused, by anticipating application of causal terminology, and by failing to distinguish between epistemological and ontological aspects.
Let us turn now to his last two premises. They logically add nothing to the present argument about causality, but they are interesting as denials by Nagarjuna himself of his previously expressed or implied views about the tetralemma (see higher up). Here, he admits that contradictories (like “real in” and “unreal in”) cannot “be together”. Likewise, he admits that negative contradictories in conjunction (like “not real in” and “not unreal in”) are not a further kind of meaningful relation (in the present case, a “causal relation”). This shows that he understands the Laws of Non-contradiction and the Law of the Excluded Middle, and appeals to them when he finds it convenient for his ends. It makes us, once again, doubt his sincerity.
(c) Cheng lists another five issues concerning causal relations raised by Nagarjuna, but does not present the latter’s arguments about them, other than to say that he “criticized each relation more or less the same way.” We may infer from this that Nagarjuna tried to show, using his usual methods, these various questions about causality unanswerable or absurd in some way. The questions he asked were the following.
· Is a cause “identical” or “different” to an effect?
· Do a cause and an effect “appear simultaneously” or not?
· Does a cause “become” an effect or not?
· Does a cause “before it ceases to be, give a causal nature to” an effect or not?
· Is a cause “within” an effect or is an effect “within” a cause?
Let us consider what his arguments might be, and how we would answer his questions and preempt his skeptical conclusions.
With regard to the first issue, my guess is that Nagarjuna would argue, as he did in similar circumstances, that if cause and effect are “identical” there is no point in naming them distinctively, and if they are “different” there can be no connection between them. But we can easily reply that they are not identical, and that difference does not imply disconnection.
With regard to the second issue, knowing Nagarjuna, he would presumably complain that if two things appear simultaneously we cannot regard one as causing the other; and if they do not appear simultaneously, how can we establish that they are linked? Philosophers who insist that causality requires that the effect temporally follow the cause, are focusing on one mode of causal relation, that in dynamic process (as for instance, in natural causation); but in truth, we also commonly acknowledge static causal relations (in the extensional mode), and even simultaneous events in dynamic processes may be causally ordered. It is only after we have established that two things, features or events are regularly together and/or apart to some degree, and therefore causally related, that the decision as to which to regard as cause and which as effect arises. This issue may in some cases be quickly resolved with reference to time’s arrow: the cause is the temporally earlier, the effect is the temporally later. But this is only one resolution, the simplest case. If the two items are simultaneous, we can still order them on other, more conceptual grounds. For instance, we will consider the more generic item as cause, the more specific as effect, judging the former as a ‘deeper’ (more widely present) aspect of nature than the latter.
With regard to the third issue, I am not sure what Nagarjuna means by a cause becoming an effect (or not). Perhaps he is referring to the frequently used Buddhist image of a seed becoming a plant? I would guess in such case that he plays on the ambiguity of the word “plant”, i.e. on whether it refers to any of its stages (including as a seed) or to a developed plant (excluding the seed). If so, we can forewarn that the word ‘becoming’ has two colloquial senses: a stronger sense of mutation (for which let us reserve the same word) and a weaker sense of alteration (for which let us prefer the expression ‘getting to be’). In mutation, the change is from ‘X and not Y’ to ‘Y and not X’; whereas in alteration, it is from ‘X and not Y’ to ‘X and Y’. Thus, in our example, the plant is initially a seed, then ‘gets to be’ a developed plant; or, in other words, the seed (undeveloped plant) ‘becomes’ a developed plant.
The fourth question is likewise unclear – what does he mean by the cause “giving causal nature to” the effect? I presume the pronoun in “before it ceases to be” refers to the cause, and that he is asking whether some sort of power of causation is transferred from cause to effect in causal chains. If that is his concept, or the concept he attributes to ordinary thinking, I would beg to differ. We do not regard that, in all causal chains, cause A gives its effect B the power to be cause B of effect C, i.e. that A does not merely cause B, but also causes B to cause C. This may be true when both the successions A-B and B-C are invariable, i.e. in the one case of complete causation; for in such case, we may say that A is invariably followed by both B and C, i.e. that A causes C as well as B. But this is one special case of regular succession – when the causations involved are of mixed determination, the syllogism is not always possible. Furthermore, in some instances (where A is not a necessary cause of B) it is still possible for “B causes C” to occur in the absence of A, when B is caused by something other than A.
One might anyway wonder at the legitimacy of an iterative causal concept “causing to cause”, in the above implied sense, for we regard “B causes C” as a relation and event that just “is”, a fact of (“caused by”, if you like) “Nature” – not as something that something else (viz. “A”), itself within nature, might “cause”. The iterative concept occurs somewhat legitimately in volitional situations, where we may say that an agent A (a person) influences or forces another, B, to perform some action, C. But in such cases, the causality involved is very different, clearly. ‘A influences B’ is a sort of causality, but it does not mean A causes B in the sense of causation. Nor is ‘B does C’ a causation, but a volition. So in this case, the iterative concept is quite different in meaning.
As for the fifth issue, I presume we have already dealt with the question as to whether the effect is “within” the cause when we discussed whether the effect is “real in” the cause. So only the question, is a cause “within” an effect? remains to be answered. Supposedly Nagarjuna has in mind here some kind of lingering existence of the cause in its effect. This could be conceived, and is indeed used as an explanatory hypothesis in some causal situations. Thus, Newton’s theory of motion of physical bodies after impact postulates that “energy” is the substratum of motion; the first body’s motion is an expression of the energy in it, and when it hits the other some or all of that energy is passed on to the second, which therefore moves or changes velocity and/or direction in accord with precise formulas. This theory involves calculated predictions, which are empirically confirmed. In this context, the second body is caused to move by the first, but the underlying cause of both their motions is “energy”.
In that case, we might say to Nagarjuna that we have a case in point, where the cause is “within” the effect. I do not, however, think that this is necessarily or always appropriate, in every causal relation, or even in every causation, to say that the cause is “within” the effect. For we establish causation primarily with reference to constant conjunction of presences and/or absences of two things, without prejudice as to whether some third thing is passed on from the one to the other. For us, causation is at first just a happenstance of regularity observed in the field of appearances. A theory may later be inductively established that some sort of transfer always occurs in it, but this would at best be a generalization warranted by confirmed predictions, not a deductive necessity. We should at least remain open-minded to either outcome in principle.
In sum, having with reference to his five vague questions anticipated Nagarjuna’s possible additional arguments against causality, we can safely say that his intended refutations of the concept are here again likely to be fallacious and inconclusive.
 These events may each be positive or negative; we shall clarify this further on. The point to note here is that cause or effect here may be motions or qualities, and their purported relation is “mechanistic”.
For this formula, see p. 84. The discussion of Nagarjuna’s
treatment of causality is found mainly in pp. 83-88. Cheng there refers to MT
XV:1a,2a,2b, XVII:1-33, XX:1-4,16-17, XXIV:18,40, and Hui-cheng-lun, 72, as
well as to TGT I-III.
Cheng there refers to MT XV:1a,2a,2b, XVII:1-33, XX:1-4,16-17, XXIV:18,40, and Hui-cheng-lun, 72, as well as to TGT I-III.
 I am thinking of Hume, who (as I recall) apparently only refers to constant conjunctions of positive events, say A and B, failing to consider the flip side of constant conjunction between their negations, non-A and non-B. He also ignores (as do Buddhists, in the said definition) “hindrance”, i.e. cases of constant conjunction between A and non-B, and between non-A and B. Of course, if all such cases of causation are considered as implicitly intended in the expression “constant conjunction”, then it is equally acceptable. J. S. Mill’s later treatment is much better, though also it has its faults. Note additionally, that “when this is, that is; when this is not, that is not” seem to logically imply “this arising, that arises; this ceasing, that ceases”, so that the latter is redundant as definition, though well to point out and remember.
 On p. 37.
 See Appendix 1: fallacy H.
 See Appendix 1: fallacy D.
 To repeat, “exclusively and invariably” (making possible and necessary) is only the strongest case; weaker causations include exclusively but not invariably (only making possible), invariably but not exclusively (only making necessary), and others (partial contingent causation, i.e. conditional causation).
 See Appendix 1: fallacy F.
 See Appendix 1: fallacy G.
 It is interesting to note that Cheng earlier (on p. 85) mentions, parenthetically, that a cause may be understood as “an assemblage of causes and conditions”. For it shows that Nagarjuna is aware that a cause is not necessarily monolithic, and indeed this awareness is found in Buddhist doctrine from its inception.
 See Appendix 1: fallacy E.
 See Appendix 1: fallacy F.
 See Appendix 1: fallacy G.
 Cheng, p. 87.
 With regard to the subject-predicate relation (ch. 4), in argument (a), and the object-characteristics relation (ch. 5), in argument (d).
 That this is acknowledged in Buddhism is evident in the traditional definition of causality earlier mentioned. The statements “when this is, that is; when this is not, that is not” refer to static causation; the statements “this arising, that arises; this ceasing, that ceases” refer to dynamic causation. These statements are (I seem to remember and presume) attributed to the Buddha himself in some sutra, and demonstrate commendable precision of thought. Static and dynamic causation may be viewed as two aspects of the same relation, or we may view the latter as a special case of the former (since given the former, we can infer the latter).
 See my Future Logic, ch. 17.
 Later, the energy transfer idea is found valuable in other contexts or domains. Later still, be it said in passing, the theory is corrected by Einstein, for greater empirical precision.