I published an essay about five and a half years ago on an alternative reading of Nāgārjuna’s MMK 24:18 (PEW 60:1) which was first disputed by Jay Garfield and Jan Westerhoff a year later (PEW 61:1) and then by Mattia Salvini (JIP 39:3). Though I certainly accept some of Salvini’s grammatical correctives regarding my original rendering of the verse, I would like to defend my foregoing interpretation and then suggest a slightly different reading from my original one here.
In the original essay, I questioned the earlier renderings of MMK 24:18 that appeared prior to 2010, and I will add one below that has appeared in a major rendition since.
“It is dependent arising that we term emptiness; this is a designation overlaid (on emptiness); it alone is the middle path.” (Robinson, 1967, 40)
“We declare that whatever is relational origination is śūnyatā. It is a provisional name (i.e. thought construction for the mutuality of being), and indeed, it is the middle path.” (Inada, 1970, 148)
We interpret the dependent arising of all things as the absence of being in them. Absence of being is a guiding, not a cognitive notion, presupposing the everyday. it is itself the middle way.” (Sprung, 1979, 238)
“We state that, whatever is dependent arising, that is emptiness. That is dependent upon convention. That itself is the middle path.” (Kalupahana, 1986, 339)
“Whatever is dependently co-arisen, that is explained to be emptiness. That, being a dependent designation, is itself the middle way.” (Garfield, 1995, 304)
“Dependent origination we declare to be emptiness. It (emptiness) is a dependent concept; just that is the middle path.” (Siderits and Katsura, 2013, 277)
These are closely, though variously interpreted, renderings of the Sanskrit verse:
yaḥ pratītyasamutpādaḥ śūnyatāṃ tāṃ pracakṣmahe / sā prajṅaptir upādyāya pratipat saiva madhamā //
No difficulty is created by the first line of this stanza, which simply equates the meaning of śūnyatā or “emptiness with that of pratītya samutpāda or “casually conditioned co-arising.” The translators, in rendering the second line into English, rather straightforwardly take it to be saying that “emptiness” is further being called a “prajṅapti upādāya,” a standard Buddhist technical term meaning a notion or label (prajṅapti) that is “derived from” or “dependent on” (upādāya) something else. Their understanding follows Candrakīrti’s explication in Prasanapadā, where the exegete harkens back to the Milindapaṅṅha and Kathāvattu in pointing out that the label (prajṅapti) “chariot” is analytically reducible to the parts that make up the chariot, and so does not denote an independent entity. Candrakīrti is then suggesting that “emptiness” is, like “chariot,” just a “derived notion” that should be understood in terms of its constituent parts rather than understood as a term denoting some discrete metaphysical reality. As I have stated previously, given the support of this reading in the commentarial tradition and what seems at first blush to be an uproblematic chain of reasoning, this is indeed a legitimately possible reading of the verse. Indeed, it is the dominant interpretation found not only in contemporary commentators, but also traditional Indian, Chinese and Tibetan ones. Given the seeming plausibility of the Prasańgika interpretation and the credentials of those who support it, why even doubt it? I got in some hot water a few years ago not merely for making some errors in my original alternative translation that I’m happy to admit and correct, but even for questioning the Prasańgika reading to begin with. I’m no sage, I’m just a scholar, so how I dare I? Why dredge it all up again.
I dredge it all up again because a number of factors continue to make this reading uncompelling for me. Some of these factors are exegetical and some philosophical considerations. And they lead me to suggest a I a possible alternative grammatical construal of MMK 24:18.
First of all, the term prajńapti is predominately used in the MMK, either in nominal or verbal form (prajńapyate) is its most generic senses of “notion” or “label,” as in 9:3, 18:6-7, 19:5 and 23:10-11. On some occasions, such “notions” or “labels” are certainly false ones, but they can also legitimately make something known or reveal something to us. The one set of stanzas where upādāyaprajṅapti is invoked in a technical Abhidharma sense as a “derived notion” is 22:1-11. But here, it is argued that tathāgata, the Buddha after his final liberation from rebirth, an “empty” being in this state, cannot be understood as being dependent on the skandhas or “personality aggregates.” That is to say, the empty Buddha should not be understood in terms of “derived notions” (upādānena sa kathaṃ prajṅapyate tathāgataḥ). So, if an empty Buddha cannot be reduced to the aggregates in the manner in which other “derived notions” can, then how could Nāgārjuna consistently say in 24:18 that the notion of “emptiness” itself is just such a “derived notion?” Furthermore, the cluster of verses surrounding MMK 24:18, such as verses 16-20, give no indication whatsoever that Abhidharma “derived notions” are relevant to the discussion at all; they merely emphasize that all empty phenomena are conditionally co-arisen phenomena. There seems, within the entirety of the text of the Kārikā, to be very little corroboration of the suggestion that Nāgārjuna considered “emptiness,” a synonym for “causally conditioned co-arising,” to be a “derived” notion reducible to its constituent parts, like a “chariot” is reducible to its wheels and axles or a “person” is reducible to her pyschophysical aggregates.
And yet, this is precisely Candrakīrti’s reading of the stanza. He insists that “emptiness,” meaning “causally conditioned co-arising” is a “derived notion” (upādāyaprajṅapti) in the same sense as the idea or label “chariot” is. The idea and word “chariot” is “derived from the wheels and the other parts of the chariot” (cakrādinyupādāya rathāṅgāni rathaḥ prajńapyate). But if, like the notion of a “chariot,” “emptiness” is reducible to “parts,” (ańgāni), what would these parts be? Candrakīrti himself does not say. But we can probably safely deduce that, if “emptiness” means the same thing as “conditioned co-arising,” then the “twelve limbs” (dvādaśāṅga) of conditioned co-arising, such as ignorance, craving, volitional habits, clinging, birth-and-death and so on are the “parts” of emptiness. However, the early Buddhist philosophers, merely because they identified the “limbs” or processes of conditioned co-arising, did not classify “conditioned co-arising” as a mere conceptual construction that did not point to the way the world really is or that things come to be and cease. On the contrary, for them, pratītya samutpāda was the “right view” of the causal relations between both natural and mental phenomena–it was not considered a fiction that needed to be analytically “deconstructed” lest it mislead us. And yet, Candrakīrti posits that “emptiness” is indeed merely just such a “derived notion.” The exegesis does bear the stamp of Prasaṅgika thought. And yet, there is, I believe, something deeply philosophically odd about this interpretation. After all, Candrakīrti himself exerts a great deal of effort to show how the MMK dismantles the Abhidharma ontological scheme to its core. That dismantling cuts through even the most basic Abhidharma distinctions between a dravyasat or an irreducible “elemental entity” and a saṃskṛtasat or a “compound entity,” the latter of which the Abhidharma would have us believe are “derived notions” to be analytically reduced to their fundamental constituents. So why would a text that undermines the very ontological dichotomy between irreducible and reducible entities than invoke that very dichotomy to explain the relationship between “emptiness” and its supposed parts? The reasons Candrakīrti offers, therefore, for reading MMK 24:18 in this way are for me, on a philosophical level, quite mysterious, to say the least.
But let us return to other textual considerations. Candrakırti claims that MMK 24:18 equates the meanings of four “identifying terms” (viśeṣasaṃjṅā), namely “conditioned co-arising,” “emptiness,” “the middle path” and “derived notion.” That is to say, for Candrakīrti, “emptiness” means “conditioned co-arising,” and the two together make up the “middle path” of Buddhist understanding, but all of these terms are nothing but “derived notions,” reducible to their ingredients. One does not find a “four-term equation” of this sort in other works attributed to Nāgārjuna, however. In the “auto-commentary” to Vigrahavyāvartani 70, Nāgārjuna proclaims that “conditioned co-arising,” “emptiness” and “the middle path” are synonymous, but no mention at all is made of “derived notions.” Likewise in Lokatitastava 22 and Acintyastava 40, crucial verses that announce the equivalence in meaning of “conditioned co-arising” and “emptiness,” “derived notion” is not invoked; the verses merely underline that “conditioned co-arising” is the “true principle” (sadharma) of the Buddha’s teachings. If we are to believe that a bedrock of Nāgārjuna’s thought is found in the warning that “emptiness” is only a “derived notion” to be analyzed away, should we not expect to find that point emphasized in similar verses in other of his writings identifying the equivalence of “conditioned co-arising” and “emptiness,” not to mention in other verses inside the MMK itself?
However, philosophical inconsistencies and textual evidence aside, are we not forced to grammatically read MMK 24:18 just as Candrakīrti has? In view of all the foregoing difficulties with the standard Prasaṅga interpretation, I suggest that there is a grammatical ambiguity in the crucial second line of the text of MMK 24:18 which makes an alternative reading possible. That ambiguity lies in the fact that the syntax of the line may force us to read it as two separate sentences rather than one compound sentence.
sā prajńaptir upādāya pratipat saiva madhyamā /
Were this a single compound sentence to be understood as saying something like: “this (emptiness) is a derived notion and is itself the middle path,” we should expect the two instances of “sā” in the sentence to be connected by the conjunctive particle “ca.” Instead, the second “sa” is accompanied by the emphatic particle “eva,” and this clearly indicates that the second “sa” stands for a different subject term than the on that begins the line. If then we do have two separate sentences here, both possessing the often omitted indicative verb “asti,” then we can parse the line like this.
sā prajṅaptir (asti) / (tāṃ) upādāya pratipat saiva madhyamā (asti) /
In this version, the line reads: “This (“emptiness”) is the idea (prajṅapti). After apprehending (upādāya) that (“emptiness”), this is truly the middle path to be walked upon.” Here the verse does nothing more than underscore that “emptiness” means, not “non-existence” (abhāva), but “causally conditioned co-arising,” and this accords with the Buddha’s teaching and puts one on the “middle path.” This would make the verse utterly consistent with the immediately surrounding verses, with the larger framework of what Nāgārjuna asserts about “emptiness” in the rest of the treatise, and with the stated relation between “emptiness” and “conditioned co-arising” in other works attributed to him.
I freely concede that the translation of 24:18 immediately above is somewhat forced. After all, we could observe the two-sentence formulation and still translate the second line of the stanza like this: “sā prajṅaptir upādāya (asti) / pratipat saiva madhyamā (asti)” and we would still have Candrakīrti’s parsing. The grammatical ambiguity of the second like of 24:18 hardly decides the interpretive question by itself. But I believe this fact reinforces my argument rather than undercutting it. Given all the problems I have found with the standard Prasańga reading explained above, I think an alternative way to understand MMK 24:18 is at least justified and plausible. And for me, it’s a philosophically preferable reading of Nāgārjuna. After all, if “conditioned co-arising” is a conceptual fiction just as much as, and just as dangerous as, the fiction of an eternal, fixed “self,” than recommending Buddhist “derived notions” over those of other traditions will always lack a vindicating explanation. And if we must, as Nāgārjuna says in MMK 24:10, rely on conventions to attain our highest aim, then good grounds for distinguishing between reliable and unreliable conventions will always be needed. If we eliminate all grounds for making such distinctions, then “right view” as the first step on the eightfold path is deprived of its importance, and all the work Nāgārjuna has put into refuting false theoretical constructions is rendered unnecessary.
Amod had left a question at the end of an earlier draft of this post, regarding what I thought of Dan Arnold’s reading of MMK 24:18 in his Buddhists, Brahmins and Belief (pp. 162-174). I generally agree with what I take to be Arnold’s inclination to read Nāgārjuna as espousing a kind of conventionalist realism, in which words and things exist in dynamic dependence on one another, like everything else. But Arnold reads Candrakīrti, insofar as he says he wants to preserve and not eliminate worldly conventions, as more in support of this conventionally realist position than I take him to be. I tend to see Candrakīrti as closer to Vijṅānavāda, particularly given what he says about skandhas constituting their own “world” (loka) in P 24 and given many other things he writes. I also still have great difficulty understanding why Nāgārjuna would take all the trouble to reject the Abhidharma characterization of what an upādāyaprajṅapti is in the first place, namely a “label” reducible to parts having svabhāva, and then say that “emptiness” is just that–a label reducible to parts, a term that is merely “derived” from its parts. But I don’t think this reading of Nāgārjuna is necessary, though Candrakīrti insists that it is. In short, I’m sympathetic to Arnold’s take as an understanding of Nāgārjuna, but I’m less convinced that Candrakīrti is arguing for a kind of conventional realism.
Thanks, Doug. I find the idea of Candrakīrti as conventional realist incredibly hard to swallow given the first chapters of the Catuḥśatakaṭīkā, which Arnold treats as an authentic work of his. The fact that our everyday conventional perceptions are misperceptions seems to me absolutely crucial to Candrakīrti’s philosophical outlook. The ultimate truth is inexpressible, but it’s still vital that we get to it, and our everyday conventions are in some respect a barrier to our doing so.
Yes. Lots of things Candrakīrti writes make it quite doubtful that he is a conventional realist, including what he writes on perception, on the relation between characteristic and what is characterized, on the nature of saṃvṛti to “conceal” reality, and so on. Actually one short, but interesting discussion on MMK 24:18 is in Joseph Walser’s Nāgārjuna in Context (258-60). Walser there suggests that Nāgārjuna is with this verse addressing the Pudgalavādin conception of prajṅapti upādāya such as is found in the Saṃmitiya Nikāya Śāstra, where the “person” is neither identical no nor different from the skandhas. Walser argues that, while Nāgārjuna does not endorse the Pudgalavādin view of the person, his use of upādāya prajṅapti is functionally similar to theirs in that Nāgārjuna wants to suggest in 24:18 that all prajṅapti represent the identities of their concepts or terms, when closely examined, as revealing mutually interdependent relations. Walser’s explication strikes me as a possibly fruitful way to make sense of the verse too.
Douglas, thanks for this important post (and apologies for joining the discussion just now). I see (and like) the way you try to construct a consistent view out of Nāgārjuna’s assertions and I understand that this leads you to a problem in the case of the dependent-ness of śūnyatā. But what about strophes like MMK 13.8, which seem (also independently of Candrakīrti’s well-known commentary thereon) to point to the vacuity and śūnyatā itself?
Thanks, Elisa. 13:8 says that śūnyatā should not be developed a dṛṣṭi. I take to mean that it should not be understood as pointing to some unconditioned dharma exists (bhavet), which is what the foregoing verses are focused on. That sort of reification of śūnyatā itself would be the flip-side of the error that śūnyatā points to “non-existence.” That strikes me as pretty standard fare for Nāgārjuna and all Mādhyamikas and I have no difficulties with that.
What I do have difficulties with is the notion that a “central tenet” of Nāgārjuna’s thought is that “emptiness” is an upādāyaprajṅapti that can be analyzed away into its constituent parts, just like the terms “chariot” or “self.” This is Candrakāirti’s reading, and most readers follow it. But even if Nāgārjuna even says such a thing in 24:18 (which is at least debatable), he only says it once in all of his works, never elaborates on it and never mentions it in the cluster of the same concepts (causally conditioned co-arising, emptiness itself and the middle path) again in other works. Some “central tenet!”
On top of that, with regard to the issue of śūnyatā nat being a dṛṣṭi, Nāgārjuna is defending the four noble truths and the eightfold path in MMK 24, and the first limb of the eightfold path is “right view.” I have a feeling that Wilhelm Halbfass, in an essay on “darśana” from the late ’80’s, was right when he suggested that, for Nāgārjuna, the dṛṣṭi that were to be rejected were all the “views” accept the one he identified with the Buddha’s own, and the Buddha’s own view makes dependent origination central to Buddhist insight and praxis. So, why Nāgārjuna would want to reduce it in the same context to the class of rejected concepts like “self” is beyond me. He devotes whole chapter of the MMK to logically refuting false views of the self and other Abhidharma categories–but he certainly doesn’t devote a chapter to refuting the notion of dependent origination, which is precisely what “emptiness,” he tells us, means. That itself should show that Nāgārjuna doesn’t hold emptiness to be on par as a concept with other concepts like “self.” So why should we take 24:18 to be claiming that it is like them?
So, to sum up, my difficulty isn’t really with the notion that “emptiness is itself empty,” my problem is with the notion that it, as a concept, is for Nāgārjuna just as dangerously false, when rightly understood as “dependent arising,” as concepts like the “self” or svabhāva. To me, that doesn’t make sense, and it’s not a position one can actually find in Nāgārjuna’s works at all except for a certain reading of MMK 24:18.
Dear Doug,
nice to see that the upādāya prajñaptiḥ discussion goes on!
Some clarifications:
in your discussion, you have not taken into account my own understanding of the syntax of the verse, which I had expressed in the JIP article you quoted. I do not agree with other translators, who read Nāgārjuna (and, perhaps, Candrakīrti) as saying that emptiness is *a* dependent designation, i.e. a notion. Importantly, I consider that to be a misreading of Candrakīrti. My own translation (in the JIP article) is as follows:
Whatever is dependent arising, we call that ‘emptiness’,
the latter is the act of designating after relying (upon something), and is itself the middle path.
I also offered a longer discussion of how I read Nāgārjuna and Candrakīrti, and how it differs from Jay’s reading, in the following section:
“I remarked that one of the most faithful translations may actually have been based on the Tibetan. I was referring to Garfield’s rendering:
Whatever is dependently co-arisen, that is explained as emptiness.
That, being a dependent designation, is itself the middle way.24
There are two points, though, where I would differ from Garfield: not so much in the grammatical understanding, but in the emphasis. The first is that pratītyasamutpādaḥ may be better rendered with ‘dependent arising’ rather than ‘dependently co-arisen’ (which would rather suggest pratītyasamutpannaḥ). Furthermore, the second half could be (misleadingly) read as implying that the equation is between the explanation of dependent arising as being emptiness, and the dependent designation.
In other words, Garfield’s wording sounds as ‘We explain what is dependently co-arisen as being emptiness, and such explanation is a dependent designation’. I do not understand the equation in this manner: although, due to a recurrent usage of pronouns as referring to entire clauses, this option is not impossible (as pointed out by Oetke), it seems like a more roundabout way to understand the syntax, and may not be the favored option. Rather, in accordance with Candrakīrti, I believe that the equation implies four terms: dependent arising=emptiness=dependent designation=the middle way. I would translate the verse as follows:
Whatever is dependent arising, we call that ‘emptiness’,
the latter is the act of designating after relying (upon something), and is itself the middle path.
We can also of course rephrase ‘designating after relying’ simply as ‘dependent designation’, or even ‘designating in dependence upon taking up as a reference’ especially if we take Candragomin’s comment about paraapek.saa as authoritative (see above).
This is in line with Candrakīrti’s remark, that emptiness, dependent designation and the middle path are specific names for dependent arising […]”
Your newer attempt does not really engage with my own reading of the verse (and of Candrakīrti’s commentary) as presented in the JIP article you mention. This is of course perfectly fine, but I wished to clarify that my own reading is *not* that “emptiness is a notion”, nor do I think that this is a consistent way to read Candrakīrti’s commentary. (The mention of my article at least suggests that your reconstruction of an opponent position includes my own reading of the verse, and I wished to emphatically clarify that I do not read the verse that way). Rather, my own reading of both root verse and commentary is: dependent arising is emptiness, and it is the same as notional dependence (the dependence of a notion upon its basis of designation), and it is the Middle Path. Philosophically, this is very different from Jay’s reading (although it is also different from your own reading).
Hope this may be of some interest,
Best,
Mattia
Thank you for this helpful clarification, Mattia.
I am actually quite sympathetic with the idea that what is being discussed in MMK 24:18 is what you (and Westerhoff) term “notional dependence” and in that respect, prajṅapti should be understood as an “act of designation.” One of the most helpful things about Walser’s account, in his Nāgārjuna in Context, is that there were three senses of upādāya prajṅapti current in Buddhist philosophical discussions in Nāgārjuna’s time, one the familiar reductionist type that explains chariots and persons on the basis of their parts, a second, Puggalavādin understanding that refers to ultimately inexplicable types of interdependence, and a third which talks about conceptual or logical dependence of one word or concept on another (anyonya prajṅapti) If I understand your explication of the verse correctly, you sense that 24:18 is invoking the last kind of designational dependence. If that is the case, that I think your reading of Nāgārjuna himself is quite plausible.
But, when Candrakīrti writes about this verse in PP, he starts right off by illustrating how the type of upādāya prajṅapti he has in mind is the first kind enumerated above, the kind that reduces a putative whole into its constituent parts. It is clear that he maintains, you are right to say, that four separate terms are being explained by the verse, dependent arising, emptiness, dependent designation and the middle path. But it seems to me that, with respect to upādāya prajńapti, he maintains that terms like “dependent arising” and “emptiness” are like terms such as “chariot” or “person” in that they are all reducible to their parts. Based on all else Nāgārjuna seems to say about such matters, particularly in MMK 22, I’m skeptical that this was the kind of upādāya prajṅapti he claimed “emptiness” is. It’s much more likely, I think, that upādāya prajṅapti, if this is an independent “term” in 24:18, indicates some kind of “notional dependence” between “emptiness” and “dependent arising,” such that, for example, saying the word and understanding the concept of “emptiness” must be based on, or relies on, the expression or concept “dependent arising.”
Apologies if I did not properly engage with your understanding of the verse in your JIP piece. I referred to that piece in my original post primarily to demonstrate how my initial reading of the verse had been critiqued and forced me to reconsider my own reading. I had not intended to align your reading with that of other scholars by implication.
Thanks again for your clarifications.