Impermanence and Emptiness: a Reversal in Perspective?

Image courtesy of photomyheart at

Image courtesy of photomyheart at

The Buddhist notion of emptiness (suññatā/śunyatā) is famously difficult to get one’s head around. In a presentation this past Saturday Sharon Salzberg described it as a combination of impermanence (anicca) and interconnectedness. This is a good first go at understanding emptiness, although the simple concept “interconnectedness” doesn’t really do justice to the recondite complexity of dependent origination (paṭiccasamuppāda/pratītyasamutpāda).

Salzberg’s aim however is not to dig deeply into Buddhist theory, but rather to bring the dhamma to a contemporary Western audience. This is a task she does very well. If we are interested in turning a bit to the history, we may also reach some modest insights.

In my last post “A World of Impermanence, the Three Marks” I dwelled on the Buddha’s approach to the three marks of existence (anicca, dukkha, anatta, or impermanence, unsatisfactoriness, and non-self) as found in the Pāli Nikāyas.

Although the Buddha describes three distinct marks of existence, it may be correct to regard them as essentially a single mark of existence, impermanence, and its effects under conditions of ignorance. Dukkha comes about because of our misapprehension of the impermanence of reality. Non-self arises through ignorance as well, because a self of the sort that the Buddha was concerned with could not exist under conditions of impermanence: that is, a permanent ground of being, a true place of refuge, and locus of perfect control.

We will now look to something of a sketch or suggestion about another branch of the later Buddhist tradition. I call it a sketch or suggestion because I am unsure as to how deeply it can be defended, however perhaps at least it bears consideration.

Here we will leave dukkha to one side, and focus on the later interrelation between impermanence (anicca) and non-self.

The claim is this: while the early, Nikāya-based tradition tended to interpret non-self in terms of impermanence, the later tradition, in particular in the work of 2nd-3rd c. CE philosopher Nāgārjuna and his followers, did the reverse. That is, they took non-self as primary, in the guise of “emptiness” (śūnyatā).

The Primacy of Emptiness

From the earliest tradition, the word “emptiness” was taken to mean “emptiness of” something, usually “self”, in the context of meditative practice* (Viz., Saṃyutta Nikāya 35.85, 43.4; Majjhima Nikāya 43.33, 106.7; Sutta Nipāta 1119). There emptiness is basically a synonym for anatta. Since all things are impermanent and dependently arisen, there is no place for an everlasting, independent self.

However as Jay Garfield notes in a commentary to his translation of Nāgārjuna’s Mūlamadhyamakakārikā (MMK): “Without viewing the world as empty, we can make no sense of impermanence or dependent origination and hence no sense of change.” (p. 317).

Or in Richard Robinson’s terms, “[Nāgārjuna] asserts that when emptiness holds good all the Buddha’s teachings hold good, and that when emptiness does not hold good, nothing is valid” (p. 326).

Emptiness is seen as primary, a pervasive feature of reality, indeed the most pervasive feature there is. It is through emptiness that we make sense of impermanence.

The work of Nāgārjuna is notoriously recondite, and hence one hesitates to make any substantive claims about it whatsoever. I would prefer to cite directly from the source, but doing so would only raise issues of proper interpretation, since his verses are so cryptic. At any rate if we can take the scholarly interpretation as adequate, this appears to be a very key and far-reaching shift in philosophical emphasis. It is a shift from impermanence, something that is seen moment to moment in daily lives, to anatta, a difficult meditative realization and abstruse philosophical claim about ontology.

Why might this shift have taken place? My sense is that much of the elaboration one finds in the later tradition stems from an attempt to come to terms with anatta, which was the Buddha’s key insight. More than anything, anatta differentiated Buddhist dhamma from its cultural and philosophical rivals, and no doubt it received a good deal of argumentative attention.

Another related example from the later tradition involves the distinction between “conventional reality” and “ultimate reality”. This distinction is not found in the Nikāyas; perhaps it as well became necessary to make sense of the difference between everyday “conventional” discourse involving selves and the purportedly “ultimate” discourse which did not. So for example we find the Buddha saying in the Poṭṭhapāda Sutta that when using terms related to the self, “… these are popular expressions, popular ways of speaking, popular terms by which the Tathāgata, without being attached, does business.” (Dīgha Nikāya 9.53. Trans. Hayes, “Self” p. 9).

Arguably the early tradition had trouble making complete sense of anatta, since kamma was supposed to have its effect over lifetimes. To what or whom does it have its effect? So long as there is an answer to this question, there is a persistent self. So long as there is no answer to this question, kamma provides no solution to the problem of just reward. Even if that self is changing, or contains changing elements, it must retain some form of identity through time, or persistence. (It may also be said more appropriately to “perdure” in the contemporary philosophical sense).

The Buddha specifically declined to answer the question as to whether the one who performed the action was identical to the one who received the benefit or harm, or whether they were two different persons (SN 12.46). This refusal to elaborate a system of personal identity over time, with its concomitant refusal to elaborate any meaningful substitute for such identity, doubtless caused problems for those of a philosophical mindset. It does so to this very day.

That said, the tradition in the wake of Nāgārjuna has trouble making complete sense of anicca, since if all things are empty of self-identity there is no way of distinguishing one thing from another, except on a conventional level. As Garfield says, “[W]hen existence is understood in terms of emptiness and when entities are regarded as purely relational in character, identity and difference can only be understood conventionally” (p. 195).

Ultimately it seems there can be no separable things. But for there to be change, a thing must be followed by something else: this is what change means. Hence it sounds as though ultimately there can be no change, and anicca is false.

The rejoinder, hinted at by Garfield’s first quote above, is that one cannot make sense of change without emptiness: if things were not empty then change would be impossible. If things were not empty then there would be causally independent, permanently existing entities.**

However I believe Nāgārjuna’s argument for this conclusion depends upon various questionable moves. Certain of these were outlined by Richard Robinson, and crucial fallacies of equivocation were outlined by Richard Hayes. In short, Nāgārjuna equivocated between the concepts “dependence/independence” and “identity/difference” in several key Sanskrit terms of his arguments. That is to say, Nāgārjuna’s arguments are not good. Hayes shows one cannot demonstrate that something lacks self-identity, and thereby that it lacks a kind of self, simply by showing it is dependently arisen based on causes and conditions. Alternately, one cannot show that something with self-identity, with a kind of self, cannot arise dependent on such causes and conditions.

Dependent origination may argue for a thing’s being impermanent (indeed, dependent origination requires impermanence), however something’s being impermanent does not require that it lack self-identity. After all, for something to be impermanent, it must arise and pass away. Anything that arises and passes away will have an impermanent self-identity which is self identity of a passing sort.

Impermanent things can have self-identity. Self need not be permanent.

The Buddha’s Rejoinder

Or maybe not, at least for the Buddha. We are forced once again to come to terms with the Buddha’s odd understanding of anatta. As we saw above, many of the Buddha’s key arguments for non-self depended on taking self to be permanent. For example, see his argument here: “Is what is impermanent … fit to be regarded [as] my self’?” “No.” (MN 109.15). This is unpersuasive insofar as we are willing to accept impermanent selves.

For the Buddha however, an “impermanent self” was anathema. It would be a form of “annihilationism” of the kind rejected in the Brahmajāla Sutta (DN 1.3.10-16).

That said, as I have noted elsewhere the Buddha did make copious use of the term “self” (attā) as a “popular expression” and rejected the view that there is “no self”. This leaves us at something of a quandary when trying to reconstruct the Buddha’s view of self. It seems he was willing to make use of the term, but not to provide much theoretical framework for understanding how and when the hoi polloi made appropriate use of the term. We will come back to this problem, below.

Existence and Nonexistence

Nāgārjuna’s treatment of “emptiness” is more far reaching than use of the term “self” (atman/attā) would imply. As we have seen, his interest revolves around terms like “being”, “identity”, “dependence”: forms of the Sanskrit term “bhāva”.

The only Nikāya text that Nāgārjuna cites in the MMK is the Kaccānagotta Sutta, or Kātyāyanāvavāda. Perhaps we can look for clues to resolving our various quandaries there. It is where the Buddha says to Kaccānagotta,

This world, Kaccāna, for the most part depends upon a duality — upon the notion of existence (atthitā) and the notion of nonexistence (natthitā). But for one who sees the origin of the world as it really is with correct wisdom, there is no notion of nonexistence in regard to the world. And for one who sees the cessation of the world as it really is with correct wisdom, there is no notion of existence in regard to the world.

… [T]his one [with right view] does not become engaged and cling through that engagement and clinging, mental standpoint, adherence, underlying tendency; he does not take a stand about ‘my self’. …

‘All exists’ (sabbam atthi): Kaccāna, this is one extreme. ‘All does not exist’ (sabbaṃ natthi): this is the second extreme. Without veering towards either of these extremes, the Tathāgata teaches the Dhamma by the middle: [through the formula of dependent arising] (SN 12.15).

Nāgārjuna glosses this passage by saying (translation adapted from Hayes),

They who perceive identity (svabhāva), difference (parabhāva), being (bhāva), and nonbeing (abhāva) do not perceive the truth in the Buddha’s instruction.

In the Kātyāyanāvavāda the Lord, who clearly saw being (bhāva) and nonbeing (abhāva) denied both the view that one exists (asti) and the view that one does not exist (nasti)” (MMK 15.6-7).

As Bhikkhu Bodhi points out in a note to the Kaccānagotta Sutta,

… [I]t would be misleading to translate the two terms, atthitā and natthitā, simply as “existence” and “nonexistence” and then maintain (as is sometimes done) that the Buddha rejects all ontological notions as inherently invalid. (n. 29 p. 734).

Bodhi points to another sutta as evidence, where the Buddha says, “Of that which the wise in the world agree upon as not existing, I too say that it does not exist. And of that which the wise in the world agree upon as existing, I too say that it exists” (SN 22.94). The Buddha goes on to say that what does not exist is anything permanent or unchanging, and that what does exist are the five aggregates (form, feeling, perception, volitions, and consciousness), which are “impermanent, suffering, and subject to change”, that is, anicca and dukkha.

The Buddha’s Equivocation

To return to the issue of the Buddha’s fraught relation to the self, in this list of existing things one does not also find the word “self”. That’s no surprise. But it therefore seems the Buddha treats the notion of a self differently from that of the five aggregates. The former he is not willing to embrace as existing (nor as not existing!), the latter five he is. That is, he treats the concept “self” differently from that of other ordinary (one might say) “popular expressions”, like “form”, “feeling”, and so on.

Why exactly is this an equivocation? The Buddha says that he does not countenance a self as existing because its purported constituents are impermanent, not under our perfect control, and so on (e.g., MN 109.15). But of course the five aggregates are impermanent and not under our perfect control as well, and yet he is willing to countenance their existence.

It seems the later tradition tended to smooth over this equivocation, and take selflessness as a global feature of reality, through the concept of emptiness. That is to say, they tended to take “self” not as a different and unique concept, but as a concept somehow paradigmatic of all reality.

As Bodhi claims, in the Kaccānagotta the Buddha was not arguing against existence and nonexistence per se. He was arguing against a global reification of a sort: the kind that held that all things are in some way permanently existent, or alternately that all is a kind of illusion. (Cf. SN n.185 p. 1085). That is, his point was more about anicca (impermanence) than bhava (being).

In the Kaccānagotta, the Buddha used the more everyday terms “atthi/natthi” which correspond roughly to our “is/isn’t”. He wasn’t using the more philosophically freighted term “bhava/abhava” (roughly, “being/nonbeing”) which Nāgārjuna did in the MMK passage. It may be that Nāgārjuna was intentionally changing the subject somewhat in order to rebut what he felt were excesses of the philosophical schools of his day, particularly the abhidhamma/abhidharma. It may also be an indication that the Buddha was not intending to make a deep ontological point, but that Nāgārjuna took him as doing so.

In interpreting the Kaccānagotta, Garfield claims that “the Buddha argues that to assert that things exist inherently is to fall into the extreme of reification” (p. 223). But the Buddha makes no claim about so-called “inherent existence” (svhabāva) in that sutta; indeed so far as I know the term “svabhāva” doesn’t occur anywhere in the Nikāyas with the connotation “inherent existence”.

It does occur in the abhidhamma/abhidharma literature to which Nāgārjuna was most likely responding. Though as Robinson points out (and Hayes concurs), Nāgārjuna defines the term in a way that none of his opponents would have accepted. As Hayes says, “What Nāgārjuna is saying is that no being has a fixed and permanent nature. What the abhidarmikas maintained was that everything has features that distinguish it from other things.” (Nāgārjuna: Master of Paradox? p. 10).

That is, the abhidharmikas were trying to explain impermanence by grounding the separability of things in their identities and differences. Nāgārjuna was saying that nothing has the nature of permanence. In this sense perhaps the two were talking past one another.

Now, the Buddha’s aim in teaching anatta was to end clinging, by ending our continual identification with things in the world. Given the facts of impermanence, he was focused on anatta as it referred specifically to one’s own person: “All phenomena are non-self” (e.g., AN 3.136) meant that one should not take any item of experience as I, mine, myself. Here “self” is being used in the sense of a person who is immune to the vicissitudes of impermanence, independent of causes and conditions, the “locus of perfect control” that constituted the aim of early Indian spiritual practice.

The later tradition expanded the scope of non-self outside that of personhood to all of reality in the guise of “emptiness”: “All phenomena are non-self” began to mean that no phenomenon, no object, no dhamma, had its own self. But whereas formerly “self” was a concept that occurred in the context of personal experience of the five aggregates, now “self” became a general feature of ontology, basically a form of the problematic “svabhāva” (inherent existence) we saw in Nāgārjuna.

In a sense of course Nāgārjuna’s moves stemmed from the Buddha’s own equivocation on what he was willing to countenance as existing, as well as his implicit attribution of permanence to self. Ironing out what may have seemed a philosophical mess led to a radical shift in perspective, as the later tradition came to its own terms with these complexities in the dhamma.


We have seen, at least as a plausible speculation, that the later tradition tended towards a shift in emphasis, moving from the primacy of impermanence to the primacy of non-self in the guise of “emptiness” or śūnyatā. While we saw in the Nikāyas the suggestion that non-self required impermanence, in Nāgārjuna, as Garfield put it, impermanence (indeed, the entire Buddha dhamma, according to Robinson) required non-self.

For most of us the reality of impermanence is much more readily available than that of non-self, not to say “emptiness”. If we take Sharon Salzberg’s formulation to heart, understanding emptiness means understanding dependent origination, a theory of some formidable complexity.

Historically it may well be the case that Nāgārjuna’s aim was to dispell the attachment he saw in certain contemporary followers of the abhidhamma/abhidharma, and that his efforts were focused to that end. If so it may be that his words are useful in the context of a kind of rarified scholastic audience.

However I believe the rest of us are better aimed towards the facts of impermanence, particularly in a secular, lay audience. I think it also bears consideration that Nāgārjuna’s key arguments appear to rely on various fallacies of equivocation. Some of the confusion may go back to the Buddha’s own usage of terms like “self”, and his own subtle equivocation between senses of permanence, independence, and identity. Nevertheless I think enough remains of the genius in the Buddha’s approach that can be brought home to our everyday attachments that the early material on non-self bears learning and consideration. Doubtless its refinement will never be complete.


* Indeed Anālayo wrote regarding two of the largest treatments of emptiness in the Nikāyas that, “It is remarkable that the Cūḷasuññata Sutta and the Mahāsuññata Sutta, whose titles indicate that their main topic is emptiness, both stress the importance of impermanence.” (p. 640)

** Independence and permanence are often treated as near synonyms in Buddhism, as though it were as good to show one as to show the other, and in particular as though establishing something as independent would entail its permanence. But permanence does not follow from independence unless we assume a very strong metaphysical claim such as the Principle of Sufficient Reason. Without such a principle, it might be that something were independent and also impermanent.


Bhikkhu Anālayo, Excursions into the Thought-World of the Pāli Discourses. Pariyatti, 2012.

Jay Garfield, The Fundamental Wisdom of the Middle Way. (Oxford U. Press, 1995).

Pāli Nikāyas trans. Bhikkhu Bodhi, Maurice Walshe. (Wisdom)

Richard Hayes, “Nāgārjuna’s Appeal” Journal of Indian Philosophy 22, 1994.

Richard Hayes, “Nāgārjuna: Master of Paradox, Mystic or Perpetrator of Fallacies?”

Richard Hayes, “Self: delusion, fiction, myth or prerequisite?”

Richard Robinson, “Did Nagarjuna Really Refute All Philosophical Views?” Philosophy East & West, V. 22 No. 3 (July 1972).