25 November 2022

On the Cognitive Linguistics of Emptiness

This essay applies an analytical method developed by George Lakoff and Mark Johnson, especially as it occurs in the book Metaphors We Live By, originally published in 1981, with a revised edition 2003. I will also draw on their other published works, notably Lakoff (1987) and Johnson (1987). Lakoff and Johnson tell us that "cognitive metaphors" are ubiquitous in human language. These metaphors involve treating a target domain as if it were a member of the same category as the source domain. In these metaphors the source domain is usually some form of physical interaction that humans have with the objective world, and the target domain is some feature of cognition. In this way, cognitive metaphors are what enable us to think about the world in abstract terms. 

This is a modern form of philosophical analysis not available to the ancient world. So this type of analysis offers the possibility of new insights when applied to old discourses. This method has occasionally been applied to Buddhism in the past, though the application has been patchy and the methods involved have not become mainstream. In this essay, I am going to use the methods developed by Lakoff and Johnson to critique the abstract concept of "emptiness" as we mainly meet it in accounts of Buddhism. In this case, I'm not criticising any particular usage, but want to make some general points about the concept. 

Cognitive Metaphors

A metaphor involves treating one things as if it were another. In a series of five blog posts in 2016, I outline John Searle's account of social reality in which "as if" plays a major role (see Social Reality). In that account of social reality I noted that language is an institutional fact:

Language itself only works because of collective intentionality, i.e. we all agree that certain verbal sounds count as words; that certain words count as representing concepts; that certain combinations of words count as sentences, and so on. (Institutional Facts & Language: Social Reality. II).

What this means is that language relies on us all agreeing that a given word means what it means.  As Wittgenstein famously said, 

“For a large class of cases—though not for all—in which we employ the word “meaning” it can be defined thus: the meaning of a word is its use in the language.” (Wittgenstein 1967, section 43)

This is often abbreviated to "meaning is use". Individualism has a role to play in the evolution of language, especially where the individual is influential.  But, generally speaking, language relies on our collective agreement on what words mean (semantics) or do (pragmatics). Cognitive metaphors are no different; other people must understand our use of cognitive metaphors in order for us to communicate about abstractions. 

The metaphor relation is not arbitrary. It is not that anything can be anything. The relation requires that the target domain has some properties that make it a good candidate for metaphorical projection. I won't go more deeply into this since it involves invoking the image schema and explaining this is too involved for an essay like this one. The standard work on image schemas is still (as far as I know) Mark Johnson's book The Body in the Mind (1987). Suffice it to say that the target domain for the metaphor must be a good fit. 

For example, we may state a commonly used cognitive metaphor: IDEAS ARE OBJECTS. (I use Lakoff and Johnson's convention of citing metaphors in small caps). In this metaphor, the source domain is our physical interactions with objects, while the target domain is a subjective experience of thought. If we accept the metaphor IDEAS ARE OBJECTS, then any operation we can physically perform on an object we can perform mentally on an idea. If I can grasp an object, then under this metaphor I can grasp an idea, as if it were an object. I can turn an idea over and look at it from another angle. I can look at an idea from different angles. If I have more than one idea, I can juggle them. I can throw an idea out, toss it around, and kick it into the long grass. Virtually anything I can physically do with an object finds a metaphorical application to an idea under the cognitive metaphor, IDEAS ARE OBJECTS.

A poor metaphor might be IDEAS ARE COWBOYS. Cowboys ride, bait, and subdue semi-wild animals for entertainment. It's not clear in what way an idea is like a cowboy. This metaphor is not intuitive. Another one might be FISH ARE BICYCLES. Note that these propositions are not forbidden by the rules of English grammar. Still, they don't make for obvious metaphorical usage. The metaphor IDEAS ARE OBJECTS works because ideas have a limited scope, they can often expressible in a succinct way that makes each idea seem discreet from other ideas. Expressing the idea leads to a transfer of that discreet piece of knowledge to another person. It's not that an idea is an object, but that an idea is sufficiently like an object in specific ways. The similarity occurs at the level of "image schemas", which I'm trying to avoid for reasons of brevity. 

It may seem simplistic to labour the point, but I think it's worth saying that ideas are not real objects. In making the metaphor, we are not reifying the abstraction. Moreover, contrary to the prevailing view of humans amongst Buddhists, people are not easily fooled into reifying cognitive metaphors. It would be odd for a person to claim that ideas are objects in a substantial sense. We know this is not true. No one ever literally held an idea in the palm of their hand, for example. We know it's a metaphor and we intuitively deal with thousands of such metaphors every day. If we had to stop to analyse each one, abstract thought would not be possible.

Unlike a computing language I don't have to "declare" the metaphor before using it. We effortlessly decode hundreds and thousands of these cognitive metaphors on the fly without even noticing that we are doing it. When people are sitting around a table at a meeting and someone says, "we need to move on", and they change the subject rather than getting up and leaving the room, no one is surprised by this.

In this case, it is because we can form a cognitive metaphor: A CONVERSATION IS A JOURNEY. For example, we might be having a conversation and it "takes a turn" (perhaps a strange or unexpected turn, or a turn for the worse). Someone might wish to "return" to what was said earlier. If it's going badly, we might say "Let's start over". If the conservation was difficult but productive, we may say: "we got there in the end". When a conversation is at an impasse, we might say that we have to move on and leave the impasse unresolved. And a conversation may reach a natural conclusion: "let's stop there".

These cognitive metaphors are not incidental but rather they form an integral part of language use. The richness of our metaphorical use of language is part of what makes us human. Our ability to talk about one thing as if it were a member of a completely different class of thing is what distinguishes human communication from all other animals. Clearly, some animals and birds are capable of abstract thought to some extent. But they don't communicate in metaphors. We do. 

Once we get attuned to this idea of cognitive metaphors, we begin to see them everywhere. When I talk about typing on my keyboard (a physical act) and words appearing "on my screen" this is two cognitive metaphors: WORDS ARE OBJECTS and SCREENS ARE SURFACES. Of course the screen is literally a surface, but the words are not on it in a physical sense. I can't physically interact with words on a screen. Even on a touch screen that's not what is happening. Rather the patterns of light and dark created by pixels make words seem to appear on the screen, and electrical interactions between surface and finger help to create the illusion of physical interaction. At the end of the day there is dust and fingerprints on my screen, but no physical objects called "words". Still, all the verbs that can be used to describe interacting with objects on a surface, can now be applied to "words on a screen".

In order to get at the underlying metaphors involved in talking about emptiness in a Buddhist context, we have to consider the use of container metaphors.

Container Metaphors

A very common cognitive metaphor involves likening something to a container. For example, in English we have the metaphor: A BOOK IS A CONTAINER. A book can, for example, be filled with ideas (here again: IDEAS ARE OBJECTS). With this combination we make a complex source domain: putting objects into a container maps onto putting ideas or words into a book. We use the same verb in each case, but use it substantively on one hand and metaphorically on the other.

A very common metaphor in English is MIND IS A CONTAINER and more specifically, mind is a container of experiences. In this view, experience happens in the mind; experience is the content of the mind qua container of experiences. Interestingly, however, Indian Buddhists do not seem to have used a specific container metaphor that we take for granted: i.e. sensory experience is contained in the mind. In Buddhism, the mind (manas) is more like a translator that turns (physical) sensory experience into (mental) perception. An ancient Buddhist could not, for example, say something like "empty your mind" because this relies on the container metaphor and they did not conceive of the mind as a container or sensory experience as the content of the mind. They are more likely to use a surface metaphor for the mind, and to talk about sensory experience as a disturbance of that surface. They may also talk about a sensory event in terms of the sense organ being struck by the appearance of an object. Keeping in mind that "appearance" (rūpa) is to the eye as sound is to the ear.

Despite the fact that ancient Buddhists did not use the container metaphor for the mind, it is so ingrained in us as English speakers that it's almost impossible to not think of the mind as a container and sensory experience as the content. 

Given all this, what can we say about how to understand the term "emptiness" (Skt. śūnyatā)

Emptiness and Experience

The adjective "empty" and the abstract noun "emptiness" are part of the broader cognitive metaphor involving containers. There is no abstract "emptiness" in the absence of a container that could potentially contain something. Moverover, emptiness in the dictionary sense boils down to "the absence of content". "Emptiness" is defined by the Online Etymology Dictionary as "the state of containing nothing". Similarly Merriam-Webster defines emptiness as "containing nothing, not occupied or inhabited" and "lacking reality, substance, meaning, or value."

These definitions are curiously opposed to Buddhist definitions of "emptiness" which specifically state that it does not mean "void" or "nothingness". As one writer seeks to clarify:

"Emptiness is not complete nothingness; it doesn't mean that nothing exists at all. This would be a nihilistic view contrary to common sense." - Lewis Richmond.

In other words, in a Buddhist the concept "emptiness" does not mean emptiness, at least in any general sense. Rather it means, we are told, that things are not as they appear to us. It is the difference between appearance and reality. In which case, "emptiness" is obviously the wrong term for this concept. Still I want to press on and consider the cognitive metaphors that apply to our English word and circle back to the doctrinal mismatch.

Any given container—physical or metaphorical—may contain something or not, but to be a container it must potentially contain something. If a container contains anything, then it is not empty. If it contains nothing, it is empty. 

Note that this is unrelated to the expected content of the container. I drink my morning coffee from a teacup I like. The rest is in a thermos and stays hot. One could say that my cup is empty of tea, for example, but by being specific one falls down a rabbit hole. My cup may well be empty of tea, water, lime-juice, cooking oil, kerosene, and every other kind of liquid, but it presently is filled with coffee and thus my cup is not empty at all. This gives emptiness an important parameter. Emptiness tends to be an absolute: if my cup has any kind of content, then it is not empty. My cup is only empty when there is no liquid in it; i.e. when there is emptiness.

So far, so logical. But this is not how Buddhists, especially Mādhyamikas, use the termin practice. Mādhyamikas use the abstract noun "emptiness" in a concrete sense. The classic example is the statement "form is emptiness". This is a valid English sentence, but there is something wrong with it. Even when we take "form" to be "form in the abstract" (or matter generally as many Mādhyamikas do), this sentence is not logically valid because it is trying to equate two different levels of abstraction. "Form" here is generally taken to mean "phenomena". If the metaphor is FORMS ARE CONTAINERS then we might validly state that form is empty. 

There are several problems here. The first is that rūpa is (in English at least) not the container of experience, it is the content of experience (or part of it). Rūpa is to eye what sound is the ear. And note that this applies across the senses. Importantly, rūpa is to the eye as tangibles (spraśtavya) are to the body (kāya). Rūpa is on the wrong side of the equation to be equated with body, even metaphorically. In Chinese, rūpa is routinely translated as 色 "hue (from original meanings "form, appearance, complexion"); visual surface quality." (definition from Kroll). In Sanskrit, rūpa is typically a property of a surface reflecting light, it is not a metaphorical container. 

That said, there is no doubt that some modern Buddhists do take rūpa to mean "substance", "matter", or "body". We can see that this is incoherent even at face value since the word is neither defined that way nor used that way in ancient texts. Even the translation "form" misleads most English-speakers into thinking in substantive terms about rūpa. Rūpa means "appearance". Moreover, even if we invoke the container metaphor, it can't be applied to rūpa because rūpa is an element of experience, this is to say that rūpa is content. Ancient Buddhists preferred to see rūpa as a disturbance on the surface of the mind, but even in this metaphor, rūpa is not substantive.

The second problem is that even if rūpa were a container we could go as far as saying that it is empty if it did not contain anything. We could not logically assert that it is "emptiness". If emptiness is the absence of content and rūpa is content, then the two are logical contraries. Despite a great deal of hand waving in modern Buddhist philosophy, "form is emptiness" simply does not make sense in English. But then it doesn't make any more sense to state this in Sanskrit; rūpameva śūnyatā is still equating two different levels of abstraction. This is an egregious wrong turn in Buddhist philosophy.

I might never have thought of this had I not discovered that the phrase was not originally rūpaṃ śūnyatā "form is emptiness", but rūpaṃ māyā "appearance is illusion" (Attwood 2017). This equation occurs in Aṣṭa and in a few places in Pañc as well. It is clearly translated into Chinese by Kumārajīva as 色不異幻、幻不異色,色即是幻、幻即是色。 (e.g. at T 223, 8.239c6-7). Here huàn 幻 translates māyā "illusion", though it originally meant "creation" or the creative power of the devas to keep the world in harmony (ṛta). Given the long history of Buddhists comparing sensory experience to an illusion this makes perfect sense. A classic example of this is the Pheṇapiṇḍūpama Sutta, which concludes with a well-known verse:

Pheṇapiṇḍūpamaṃ rūpaṃ, vedanā bubbuḷūpamā /
Marīcikūpamā saññā, saṅkhārā kadalūpamā;
Māyūpamañca viññāṇaṃ, desitādiccabandhunā
(SN iii.142).
Appearance is like a ball of foam, valence like a bubble.
Recognition is like a mirage, volition like a plantain.
Discrimination is like an illusion. So Ādiccabandhu taught.

Here, Ādiccabandhu means the Buddha, but it is a distinctively Brahmin name completely unconnected to any of the standard myths of the Buddha. A similar verse occurs at the end of the Vajracchedikā, where the simile becomes a metaphor:

tārakā timiraṃ dīpo māyāvaśyāya budbudaḥ |
supinaṃ vidyud abhraṃ ca evaṃ draṣṭavya saṃskṛtam ||Vaj 22 || (Harrison and Watanabe 2006)
We should see the conditioned as a star, a kind of blindness, a lamp;
An illusion, a dewdrop, a bubble, a dream, a lightning flash, a cloud.

We also find the simile in Aṣṭa, “appearance is like an illusion” (māyopamaṃ rūpam. Vaidya 1960: 9). And this is all quite straightforward: experience and reality are not the same thing; different rules apply. 

There is a popular rhetorical strategy for dealing with "form is emptiness" amongst Buddhists which can be illustrated with a random example from the Tricycle website:

Avalokita found the five skandhas empty. But, empty of what? The key word is empty. To be empty is to be empty of something.

If I am holding a cup of water and I ask you, “Is this cup empty?” you will say, “No, it is full of water.” But if I pour out the water and ask you again, you may say, “Yes, it is empty.” But, empty of what? Empty means empty of something. The cup cannot be empty of nothing. “Empty” doesn’t mean anything unless you know empty of what. My cup is empty of water, but it is not empty of air. To be empty is to be empty of something. This is quite a discovery. When Avalokita says that the five skandhas are equally empty, to help him be precise we must ask, “Mr. Avalokita, empty of what?”

What we see here is a fantastic distortion of reality, leading to a false conclusion. It is nonsensical for you to ask me what my cup is empty of, because to be empty in any sense, it has to be empty of everything. As I noted above, my cup could be and regularly is empty of tea (and all other liquids) but full of coffee: in which case my cup is not empty at all. The conclusion here—“Empty” doesn’t mean anything unless you know empty of what—is simply not true. This is a case of the tail wagging the dog. That is to say, we know what the answer had to be in order to legitimise Buddhist dogma on emptiness, and the question is phrased in such as way as to elicit only that answer. But in doing so, Buddhists blithely defy the conventions of language. 

We can never legitimately ask "empty of what?" The question is meaningless and the answer is simply a restatement of a dogma that doesn't make any sense. The idea that "empty of what" is a natural question is either extraordinarily naive or disingenuous. Either way, Buddhists propagate this falsehood in all sincerity. 

This invalid method and false conclusion are often parlayed into an even worse question using the abstract noun: "emptiness of what?" Such a thing is not allowed under English grammar. Emptiness is emptiness. "Of what" is an entirely meaningless question because if the answer is not "everything", then the vessel is not empty at all. 

We do sometimes suggest that emptiness might have degrees.  For example, we may say that a cup may be half full or half empty. Still, it's only from the point of view of being half full that we can ask "of what?" The "of what?" question only applies to the content of the container. An empty container has no content; a half empty container is half empty of all content. Even if we say the glass is half empty, no one in their right mind asks "Half empty of what?". This is simply not how the container metaphor works. 

We can see that the cognitive linguistic perspective is a powerful method for understanding utterances. But it also highlights how dogmatic the Buddhist discourse on emptiness is. This kind of invalid logic is de rigueur for Buddhist philosophy and is almost never questioned or critiqued: either from within or without. Rather such views are carefully curated by Buddhists, in the sense of being framed as deep truths, discovered by visionaries and mystics, and accompanied by frenzied hand waving so that they can be presented as something they are not, i.e. true. This is what we expect of a religious philosophy or theology. There are axioms that cannot be questioned or the whole thing would fall apart. The fabric of Madhyamaka is held together with unquestioned, religiously inspired, axioms. 

The same argument holds for Sanskrit which has identical cognitive metaphors. In Sanskrit it is nonsensical to say rūpaṃ śūṇyatā, but it is sensible to say rūpaṃ māyā, and even better to say rūpaṃ māyopamaṃ "appearance is like an illusion"

So my, rather awkward conclusion is this: Buddhists don't seem to understand the concept of empty, let alone the concept of emptiness. If they did understand, the question "empty of what?" would never occur to them. Worse, Buddhists routinely insist that this flawed concept of emptiness is what makes sense of Prajñāpāramitā. Two wrongs don't make a right. 

In this case, how should we understand the word emptiness?

Making Sense of Emptiness

The key here is to note that the first use of śūnyatā as a technical term is to refer to the state of meditative concentration in which all sensory experience has ceased due to the withdrawal of attention from the senses. This state is called suññatāvihāra or śūnyatāsamādhi. Since sensory experience is dependent on attention (manasikāra), by practising non-attention (amanasikāra), one prevents sensory experience from arising and causes arisen sensory experience to cease. 

Here, sensory experience can be seen as the content of experience or, in Buddhist terms as a distortion of the (naturally) smooth surface of the mind. As such, sensory experience may be present or absent and even admit degrees of these. Hence, between ordinary waking awareness and emptiness there are numerous stages (āyatana) of increasingly attenuated sensory experience. But here, too, absence is absolute; the presence of an any sensory experience means that emptiness doesn't apply. This point is made repeatedly in the Aṣṭasāhasrikā, for example. Emptiness in this case, is the complete absence of sensory experience. 

There are several Buddhist approaches to analysing the content of experience: a range of reductive ontologies into which experience is analysed. For example, the skandha-ontology, which focuses on the processes that give rise to experience, or the dhātu-ontology, which is focussed on the sense faculties and their objects. Mainstream Buddhism foregrounds this reductive, analytic approach of breaking experience down into simpler components in order to eliminate it as a source of absolute being. That complex objects disappear under analysis is not some great metaphysical truth, it is simply a consequence of methodological reductionism. 

If I dismantle my chariot, of course I no longer have a working chariot because I've just broken it on purpose. Who does that? Why would I want to dismantle a working chariot in the first place? And why would my destruction of the thing constitute proof that it never existed in the first place? This is the claim that many Buddhists make but, again, it is nonsensical.   

Prajñāpāramitā Buddhists, building on a tradition that is probably older than Buddhism itself, sought first to bring sensory experience to a halt. They didn't analyse sensory experience in any depth because the acme of their program was not an insight into sensory experience. What they sought, first and foremost, was an insight into death and rebirth. The whole fetish of emptiness was originally established on the analogy of emptiness with death. Mastery (vidyā) over sensory experience, in the form of the ability to voluntarily make it stop, equated to mastery over repeated death. This mastery was and is the driving force of Buddhism, even when it is buried in centuries of intellectual accretion. 

My current thinking is that the discovery of how to do this probably arose around the same time as major socio-political changes in India, reflected in, for example, the replacement of red and black pottery type by the painted grey ware style of pottery. Within a few centuries we see the emergence of walled city states which are stable for some 200-300 years before the Moriyan Dynasty of Magadha overwhelmed all the others, founding the first pan-Indian empire. One possible source of mind-training techniques that limit sensory experience is the "interiorisation" of Brahmanical rituals. In this development, some Brahmins began to perform their daily rituals in imagination rather than physically. This led to the discovery of radical changes in sensory experience, especially in the form of hallucinations due to sensory deprivation, and ultimately to the cessation and absence of sensory experience. By the time the Bṛhadāraṇyaka Upaniṣad was composed (in or around the Kingdom of Kosala) the correct performance of rituals was being linked to one's afterlife destination. Buddhists and Jains had similar ideas but focussed on actions more generally, with Buddhists refining this to just volitional actions. 

However it happened, it is apparent that in this milieu some religieux developed and shared the techniques that allowed them to bring sensory experience to a halt and to dwell in a state in which there is awareness but no content. Some Buddhists called this "emptiness" (śūnyatā). Other Buddhists called it "extinction"(nirvāṇa) and other names. This state is also known in modern times as "contentless awareness", "minimal phenomenal awareness", or "non-dual awareness". 

This is how I presently understand "emptiness" in Prajñāpāramitā.  I believe this is a better approach than anything based in later traditional interpretations based on the Madhyamaka telos (which sees Prajñāpāramitā merely as proto-Madhyamaka). 

Dharma as Container?

One of the key concepts in Madhyamaka is "the emptiness of dharmas". In this usage, dharmas have to be considered as metaphorical containers. The broader translation of dharmas as "phenomena" (as distinct from noumena; i.e. appearances rather than reality) seems to fit here, but what is the content of  a phenomenon? Is there really any phenomenon that is not sensory experience?

Nāgārjuna tells us that he expects that we will expect a dharma to have svabhāva in the sense of being autopoietic or self-creating. Nāgārjuna points out that this self-creating property of dharmas cannot exist in any changeable phenomenon. So far so good. The problem is that no one ever believed in self-creating dharmas. No one ever proposed this before Nāgārjuna. But he said that everyone believed it. Nāgārjuna appears to have lied about this. What puzzles me is that no one really cares about the lie. Many people seem to prefer this lie. 

The svabhāva of a dharma, according to Abhidharma lore, is the sui generis quality that gives us the ability to identify it. For example, it's important to all Buddhists to distinguish skillful (kuśala) motivations (cetanā) from unskillful (akuśala) ones. If I experience a moment of greed or generosity, I identify it as such by introspecting the content of the experience. The fact that I can identify an experience as motivated by greed or generosity doesn't imply anything like Nāgārjuna's autopoietic dharmas. As far as I can see, there is no way to even infer autopoietic dharmas from any early Buddhist doctrine. We have different kinds of experiences and these are identifiable by certain characteristics. No one disputes this, not even Nāgārjuna. 

However, Nāgārjuna also assumes that to be real a dharma must have svabhāva in his autopoietic sense. This axiom is incoherent, but is blindly accepted by all and sundry; even Graham Priest, the academic logician, seems to fail to see this basic logical error in Nāgārjuna's argument. Since he can (trivially) prove that no dharma can be autopoietic, he then deduces that dharmas are not real, that they don't exist. But this definition of "real" is completely incoherent. Not only did Buddhists never use this definition of real, they weren't even interested in the question of the reality or unreality of dharmas. They were interested in the arising and ceasing of dharmas; especially in the light of a state in which all dharmas cease except for the asaṃskṛtadharma, i.e. emptiness. Emptiness is asaṃskṛta because it does not occur due to the presence of a condition but rather occurs when all conditions for sensory experience are absent. 

In order to square the circle, Nāgārjuna has to introduce the nonsense idea of a "relative truth", which is not true at all. The ultimate truth, in this view, is that dharmas don't exist, because they are not self-creating. I can see no good reason to take Nāgārjuna seriously as a philosopher or even, frankly, as a Buddhist. He seems to have entirely missed the point of Buddhism and has gone off on a tangent. And still, he is routinely cited as "the most important Buddhist philosopher". 


The term "emptiness" is generally used in an incoherent way by Buddhists, especially in statements containing the idea "emptiness of...". We can never legitimately ask "empty of what?" let alone "emptiness of what?" because this is not how the container metaphor works. 

The idea that the proposition "form is emptiness" is meaningful now seems doubtful. Moreover, when we look at the kinds of post hoc arguments put forward to justify this proposition, they simply don't make sense. In addition, we know that it used to make sense when presented in the form: "appearance is an illusion." A sensory experience is like an illusion. I doubt anyone would argue with this.

It is also true that in the state called "emptiness" there are no dharmas because that state occurs only when all dharmas have ceased and no new dharmas are arising. This sense of the term is far more coherent than the general religious consensus that emptiness is reality. 

The incoherence reaches its apotheosis in Madhyamaka rhetoric about the emptiness of dharmas, by which Mādhyamikas mean that they think dharmas don't exist, since they tie existence to self-creation and it is trivial to show that dharmas cannot be self-creating. Nāgārjuna insists on an incoherent definition of what "real" means and uses that to argue that the concept of existence is incoherent. Prior to Nāgārjuna no one ever used this definition of real. Apart from his devotees, most Buddhists still don't use this definition. 

The standard ways we have of talking about this all seem to miss the point. Early Buddhists did not venture opinions on the existence or nonexistence of dharmas, except in the case of the sarvāsti doctrine. Even the Sarvāstivādins did not argue that the existence of dharmas was due to self-creation. The logic of sarvāsti is completely different but not difficult to follow. If a past dharma can be the cause of a present effect, then the doctrine of dependent arising itself says that it presently exists since imasmin sati, idaṃ hoti and imasmin asati, idaṃ na hoti. If the dharma doesn't exist now, then it cannot be a factor in the arising of a dharma in the present. This central argument is not even considered by Nāgārjuna, let alone refuted. 

The nature of dharmas is irrelevant in light of the fact that dharmas arise and cease, depending on where our attention goes. To say that dharmas lack svabhāva in Nāgārjuna's sense is trivial. To say that they have svabhāva in the Abhidharma sense is also trivial since we routinely recognise hundreds of different kinds of experience (for which we have thousands of words). The key to understanding Prajñāpāramitā lies in another direction entirely. The main point is that attention can be withdrawn from sensory experience. When we withdraw attention from sensory experience, it ceases, leaving us in a particular state characterised by some kind of basic awareness without any experiential content. That is, in a state of emptiness.

While it is not essential to my critique of Madhyamaka, it helps to understand the cognitive metaphors of emptiness and how cognitive metaphors function generally. This is so because "the emptiness of dharmas" is a cognitive metaphor: DHARMAS ARE CONTAINERS. But this is only true if dharmas exist and are capable of acting as metaphorical containers.

Still, it is only Madhyamakas who believe that in order to exist, to be real, a dharma must be self-creating. "Self-creation" is an odd choice for the content of that container. I can imagine a thing being self-creating, but I cannot imagining a thing containing self-creation. Self-creation doesn't fit the cognitive metaphor. 

So even if we could legitimately ask "empty of what?" the answer "empty of self-creation" is nonsense on several levels. For example, it would require us to relate to "self-creation" as content. To my mind this simply doesn't work. "Self-creation" is not a suitable target for the source domain of things we put in containers, except in the very broadest sense that IDEAS ARE OBJECTS. The idea of self-creation might be the content of a metaphorical container, the fact of self-creation cannot be.  

On the other hand, the emptiness of the mind, i.e. the concept of the absence of mental content in meditation, is not plagued by these inconsistencies and incoherences. In English it is natural to use the container metaphor for this. It is not so natural in scriptural languages, but, nevertheless, the absence of dharmas in meditation is the key concept here, not the absence of being self-creating. The whole idea of self-creating dharmas is a red herring. 

The metaphysical speculations that attract us as explanations for emptiness are largely based on prior indoctrination. In my reading, such speculations are absent from both early Buddhism and Prajñāpāramitā. This is not to say that metaphysics is generally absent from or irrelevant to Buddhism.  All ancient Buddhists believed in karma and rebirth, for example. These involve commitments to metaphysical views that we now know to be false, though few Buddhists will admit to this. 

The methods of cognitive linguistics are a powerful tool for thinking critically about Buddhist doctrines. That said, most existing applications of these methods have been in the service of tradition, i.e. used purely descriptively by scholars who have no interest in critiques of Buddhist philosophy. Whatever the reason for it, this side-stepping manoeuvre allows those people to continue evangelising for traditional Buddhism without ever confronting the inevitable antinomies between Iron Age or Medieval thought in India and present day science and philosophy. Many Buddhists seem attracted by the idea of subsuming all knowledge within Buddhism. This tends to involve a rather blasé form of dualism in which science is merely concerned with the "physical" and Buddhism is concerned something that we often see called "spiritual".

Unfortunately, this exceptionalist discourse appears to obscure and devalue the real contribution of Buddhists, i.e. the cultivation and exploration of states of contentless awareness. I see this as a lose-lose scenario. I see the neuroscience community studying this phenomenon and developing their own terminology for it, though at present we still see a proliferation of different terminologies. At some point, an objective account of the methods and consequences of meditation will eclipse the religious accounts. Those who insist on the religious accounts, with all their incoherence and misdirection, will be relegated out of the conversation and become irrelevant. I'd prefer to see experienced meditators staying in the game, but as long as they cling to outmoded forms of talking about emptiness, they will not be part of the conversation for much longer. 



Attwood, J. (2017). "Form is (Not) Emptiness: The Enigma at the Heart of the Heart Sutra." Journal of the Oxford Centre for Buddhist Studies, 13,52–80. http://jocbs.org/index.php/jocbs/issue/view/15/showToc.

Johnson, Mark. (1987). The Body in the Mind: The Bodily Basis of Meaning, Imagination and Reason. University of Chicago Press.

Lakoff, George. (1987). Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. University of Chicago Press.

Lakoff, G. & Johnson, M. (2003). Metaphors We Live By. New Ed. [Originally published 1981]. University of Chicago Press.

11 November 2022

On the Indo-Tibetan Commentaries and Methods in Buddhist Studies

I have almost no interest in popular translations, or commentaries, since these all repeat the same mistakes and result in cliches that I know to be untrue. I do try to be completist when it comes to academic publications on the Heart Sutra (at least in English). Being completist in this sense is seldom rewarding because the standard of work coming out in this field is typically quite poor. I've published a couple of critical reviews now (2020, 2022) as well as posting quite a few more on my blog (e.g. here, hereherehere, and here). 

The eight Indo-Tibetan commentaries on the Heart Sutra have received a relatively huge amount of attention in the form of two books by Donald Lopez; one a study and the other a complete translation with reflections on themes in the commentaries. Other commentaries, such as those by Kuījī and Woncheuk, have also been translated but, at least in English, they have not been studied with anything like the same level of attention. And what have we learned? This is summed up in the conclusion of recent article by, long-established scholar of Tibetan Buddhism, Klaus-Dieter Mathes:

We have seen how the quintessence of the Prajñāpāramitāsūtras, the formula “form is emptiness; emptiness is form” has been interpreted in eight Indian commentaries from nearly all possible Mahāyāna views and approaches. (Emphasis added)

This is no more that what Alex Wayman had observed in 1984, i.e.

“The writers seemed to be experiencing some difficulty in exposition, as though they were not writing through having inherited a tradition about the scripture going back to its original composition, but rather were simply arranging their particular learning in Buddhism to the terminology of the sūtra.” (1984: 309)

Or Malcolm David Eckel in the same decade:

“... to approach the Indian commentaries in the hope that they will somehow yield the ‘original’ meaning of the text is to invite disappointment... what they thought it meant was shaped as much by the preoccupations of their own time as it was by the words of the sūtra itself. (1987: 69-70)

Mathes cites neither Wayman nor Eckel. Nor does he cite my (2017) article: "Form is (Not) Emptiness" which was directly relevant to his topic. Nor does he cite Huifeng (2014) which is also relevant. Mathes does cite Jan Nattier, but it is the most bizarre reference to her work that I have ever seen:

"The Heart Sūtra lends support to a simultaneist realization of emptiness, and for that reason Jan Nattier has even argued that it was a Chinese composition and brought to India by Xuanzang."

Leaving aside the fact that I don't know what "a simultaneist realization of emptiness" means, the logic here is not valid. The reason we—Nattier, Huifeng/Matthew Orsborn, Jeffrey Kotyk, and I—conclude that the Heart Sutra was composed in Chinese has nothing to do with "a simultaneist realization of emptiness" (to my knowledge none of us has ever used such terminology). Our argument is philological. Nattier (1992) showed that the core passage in Hṛd was too different from that in Pañc for Hṛd to have copied directly from Pañc. On the other hand Xīn and Mōhè are more or less identical, with a few tweaks to include some of Xuanzang's preferred translations. Clearly, Xīn copied from Mōhè. In addition, it's apparent that the differences in Hṛd are the result of some kind of paraphrase that is consistent with being back-translated into Sanskrit from Chinese. Huifeng (2014) and I (in various papers) have extended this observation to other parts of the text and shown that the patterns Nattier observed in the core section generalise to the other half of the text also. Moreover, I (Attwood 2018b: 19-22) showed that tryadhvavyavasthitāḥ sarvabuddhāḥ is a calque of sānshì zhū fú 三世諸佛 “buddhas of the three times”, while Sanskrit Prajñāpāramitā literature has a strong preference for the unabbreviated “buddhas of the past, future, and present” (atītānāgatapratyutpannā buddhāḥ).

I wouldn't mind so much but, outside of our circle, Buddhist Studies scholars seem loath to give Nattier the basic respect of accurately describing her evidence, methods, and conclusions. The recent notable exception is Sarah Mattice's (2021) book which devotes fully 19 pages to this task. Mattice has devoted more space to this issue than all the naysayers and fence-sitters combined. Compared to this, Mathes' distorted account may be the worst example of this I have seen by an academic. Of course, his article was published too late to take into account my recent overview of this issue in JIABS (Attwood 2021), but most of the earlier works on this topic were available. And Nattier's article is thirty years old this year (2022). 

In a couple of polemical reviews (2020 and 2022 forthcoming) I take academics to task for not doing a proper literature review before conducting their research. This is not simply because they don't cite me, though of course this is an issue for me. I've published fourteen articles on this topic and they are all widely available. In Mathes case, half a dozen of my articles could have been cited. 

For example, Mathes, unlike the vast majority of Buddhist Studies scholars, has noticed a problem in Conze's 1967 revised edition of the Prajñāpāramitāhṛdaya. In the first sentence, pañca skandhās (nominative plural) is impossible to relate to the rest of the sentence and the transitive verb, vyavalokayati sma, has no object. In 2015, I published this observation along with the solution which was inspired by two variant manuscripts, i.e. we add an anusvāra to the dhā-akṣara to give us pañca skandhāṃs (accusative plural). With this slight change we solve both problems at once: pañca skandhāṃs is the missing object of vyavalokayati sma. Now, I don't expect a Nobel Prize for this observation but I do expect to be credited as the first scholar to publish it. This is the tacit agreement that we all make; if you get there first, other scholars will acknowledge this and give credit where it is due. Mathes doesn't do this and it's bad form. 

Mathes does not notice the other big mistake in Conze's text, though one can see from his translation that he struggled to know what to do with it. Conze's misplaced full stop after acittāvaraṇa leaves the end of the sentence hanging; it is a "sentence" with no verb and no subject, i.e. not a valid sentence. Simply removing the full stop allows one to parse the now combined passages as one sentence. Mathes' approach is to break the text apart until the garbled grammar ceases to be an issue: 

Therefore, Śāriputra, because bodhisattvas have no attainment, they rely on, and abide in, the perfection of insight. They have no mental hindrances. Because their minds are without hindrance, they have no fear. They pass completely beyond error and go to the fulfillment of nirvāna.
Mathes has curiously misconstrued the Sanskrit here. The absolutive in the first clause, āśritya is generally translated as a gerund "having relied on" or a present participle "relying on". The text clearly says something like "having relied on prajñāpāramitām", but viharati "he dwells" cannot also relate back to prajñāpāramitā in this sentence; whatever "dwelling" is being dwelled, it is subsequent to "relying on prajñāpāramitā". Contra Mathes, one thing this passage cannot say is, "[they] abide in the perfection of insight.". The bodhisatva is the agent of both verbs, but prajñāpāramitām only goes with āśritya here. It might have said something like that if āśritya were in the form of a finite verb such as āśrayati.

Most translators, Conze included, take acittāvaraṇaḥ to be the state in which the bodhisatva dwells, though I admit this has never made sense to me. The case is masculine nominative singular, meaning that acittāvaraṇaḥ ought to be an adjective of some other noun in the masculine nominative singular and there is only one in this sentence, i.e. bodhisatvaḥ. Note that this relation was obscured in Conze's (1948) original edition and in the popular (1958, 1975) edition (Buddhist Wisdom Books) because his text has bodhisattvasya (genitive singular). In the revised  (1967) edition, he repairs this blunder. 

In Conze's editions, what follows, after the erroneous full stop, is a conjunction and a series of adjectives of the bodhisatva who relies on Prajñāpāramitā. There is absolutely no need to make these into separate sentences, let alone into four separate sentences. The revised text and my translation read

tasmāc chāriputra aprāptitvād bodhisatvaḥ prajñāpāramitām āśritya viharaty acittāvaraṇaś cittāvaraṇanāstitvād atrasto viparyāsātikrānto niṣṭhānirvāṇaḥ | 

Therefore, Śāriputra, because of being in a state of non-attainment, the bodhisatva who is without mental hindrance dwells having relied on perfect paragnosis; because of the nonexistence of mental hindrance he is not afraid, transcends delusions, and his extinction is complete. 
This is not beautiful prose by any means, but it does at least translate the text as given. It's not until we dig into the Chinese text and the relations between the two that the Sanskrit emerges as a garbled version of a much more straightforward Chinese text:
Since the bodhisatva relies on perfect paragnosis their mind is not attached anywhere; being detached they are not afraid, transcend illusions and delusions, and attain final extinction.

Unfortunately the Chinese Buddhist monk who created the back-translation did a really terrible job of this part of the text; he got the verbs all wrong (and this much is clear from reading Huifeng 2014). Here is an alternative Sanskrit translation of the same Chinese passage showing how it might have been done better: 
yato bodhisatvaḥ prajñāpāramitām niśrayati tato 'sya cittaṃ na kvacit sajjati | tena ca atrasto viparyāsamāyāvivikto niṣṭhānirvāṇa || 
“Since the bodhisatva relies on perfect paragnosis, his mind does not adhere anywhere; and for this reason he is unafraid, isolated from delusions and illusions, and his extinction is complete.”
How many translators have looked at Conze's defective edition and rather than asking the obvious questions, simply fudged their translation? All of them. And this is an indictment of Buddhist Studies. If a Sanskrit sentence is not a properly formed sentence, then one can't simply fudge the translation so that it is a properly formed English sentence. At least not in an academic philological study.

Mathes apparently understands that everything in this weird sentence relates back to "the bodhisatva" (the subject of the correct sentence) and we can forgive his use of the plural here as translator's licence. Still, his text here is entirely in the singular, so why not translate it as given? This is a minor point compared to other faults and as a standalone fault might be overlooked. 

Mathes' literature review seems to have been perfunctory at best. I don't know how they teach research methods these days, but when I was learning about doing academic research I was taught that one could not skip this step. And yet I see this time and again: no proper literature review, and apparently no oversight of this failure from editors or reviewers, who are equally ignorant of the literature. At times it seems to me that no one in Buddhist Studies knows the literature of the Heart Sutra, but everyone recalls it as presented in some long distant undergraduate lecture and a handful of now dated sources. And it keeps happening, despite ten years of effort on my part to do better. Somehow Buddhist Studies scholars who are otherwise extremely competent, like Mathes obviously is, let all that go when they write about the Heart Sutra. I previously noted this phenomenon with Harimoto Kengo, for example, a highly competent Sanskritist who wrote yet another underwhelming article on the Heart Sutra.


Like the general public and novels, it seems that most Buddhist Studies academics have one Heart Sutra in them. Some manage to write that article, but few if any ever return to the text. To be fair, Nattier intended to write more about the Heart Sutra, with her husband John McRae (who also wrote one article), but he died before that could be completed. Weirdly, when they tackle the Heart Sutra, many academics abandon doing research and write as theologians. This is a puzzling phenomenon and I hope one day an anthropologist might study it. 

One thing we can say is that the expectation of nonsense appears to be self-fulfilling, in that people don't expect the Heart Sutra to make sense and don't seem too bothered if writing about the text also doesn't make sense. 

I see two main conclusions emerging from reading Mathes' article that are not part of his fairly prosaic written conclusions about the lack of coherence in the Indo-Tibetan commentaries:

Firstly, there was no Indian tradition of commentary on the Heart Sutra. Hence Alex Wayman's point that the commentaries attributed to Indian pandits all take a different approach that is based on the religious professions and presuppositions of the day. There is no unified tradition of understanding the Heart Sutra anywhere in the Buddhist world. We can now safely say that the Heart Sutra was unknown in India. Certainly, apart from the Tibetan texts attributed to India pandits, there is zero evidence of the text in India. If anyone has such evidence then I would urgently like to hear from them.

We now know that at least two of the Indo-Tibetan commentaries were based on a Tibetan Heart Sutra text (via Horiuchi 2021). As far as I know, no one has really investigated the plausibility of these attributions. And some of them cannot be investigated because the putative author is otherwise unknown. The idea that a canonical attribution is prima facie plausible seems doubtful at best. We know that, in Chinese at least, many of these are apocryphal: not least for the Heart Sutra itself, the whole standard history of which is a fiction.

This means that these Indo-Tibetan commentaries can only tell us about Buddhism in and around medieval Tibet. That is to say, they reflect what medieval pandits—possibly Indian pandits or, more likely, their Tibetan followers—made of the Heart Sutra when they encountered it in Tibet, often in the form of a Tibetan translation of the Heart Sutra. However, the sectarian approaches they adopt are all different and thus these commentaries tell us little or nothing about the Heart Sutra, per se. Rather, the Heart Sutra is shoehorned into various sectarian religious systems.

As such, these commentaries are of interest mainly to Tibetologists and contribute nothing to understanding Prajñāpāramitā as a form of Buddhism in its own right. A corollary of this, which is evident in Chinese commentaries as well, is that while one can read the Heart Sutra as a Madhyamaka text, one is not bound to do so. Some of the Indo-Tibetan and Tibetan commentaries see the Heart Sutra as a statement of, or consistent with, Yogācāra Buddhism. The connection between Madhyamaka and Prajñāpāramitā was not at all obvious to some ancient commentators. Indeed, Kuījī's commentary acknowledges that one can use a Madhyamaka approach, but a Yogācāra approach is superior. 

None of this interests me as much as reading the Heart Sutra as a Prajñāpāramitā text.

Secondly, I think we come back to a point I have made here and in various articles: the methodologies that academics employ when studying the Heart Sutra leave a great deal to be desired. In particular, it seems that almost no one bothers to do a proper literature review before sitting down to compose an article on the Heart Sutra. It looks suspiciously like academics simply grab whatever is to hand rather than making an effort at comprehensive coverage. Having published fourteen articles on this topic in peer-reviewed journals since 2015, I am getting heartily sick of academics being completely unaware of my existence or simply ignoring me. If they do a literature review at all, they are somehow excluding all my published work from consideration, even when it is directly relevant. 

And it's not just me. Sadly, for Matthew Orsborn, his seminal article on the text, i.e. Huifeng 2014 is routinely overlooked. I would argue that no one can begin to understand the Heart Sutra without being au fait with this article. To be fair, it took me a long while to come to terms with it too. Still, in order to be informed on Heart Sutra research, one must read Nattier 1992 and Huifeng 2014 at a minimum. I'm pleased to say that I have just read a draft article by a friend who is attending Dharma Drum College, Taiwan, that does give Orsborn his due. This ought to be published sometime next year (presuming China does not invade before then). 

Nattier has a different problem. Thirty years after her brilliant and insightful article appeared, academics like Mathes are still casually misrepresenting her evidence, methods, and conclusions. Mathes may be the worst example of this I have seen amongst academics. However, it is still the case that most academics refuse to acknowledge Nattier's work on the Heart Sutra, except in Japan where Nattier's work is acknowledged in the context of a series of shoddy polemical articles by older male academics who are also high up in Japanese ecclesiastical hierarchies. 

The final conclusion of Mathes (2022) is more or less the same as what Wayman and Eckels wrote in  the 1980s. And one wonders whether this is an example of publish or perish, since it doesn't add much to what we already know. 

I don't particular enjoy writing these critical responses. I'd prefer to have something meaningful to engage in; I'd much prefer to be learning something. That said, most of the publications about the Heart Sutra emerging in English are very poorly researched and written. Most barely qualify as "scholarship" since the normal methods of research are seemingly in abeyance in most cases, as with Mathes (2022). The situation is so bad that to not comment at this point would amount to complicity in an ongoing intellectual fraud. I want it to be clear to academics that if they publish these kinds of poorly researched and badly argued articles on the Heart Sutra, they can expect me to dissect them in public without fear or favour. 



Attwood, J. (2015). "Heart Murmurs: Some Problems with Conze’s Prajñāpāramitāhṛdaya." Journal of the Oxford Centre for Buddhist Studies 8, 28-48. http://jocbs.org/index.php/jocbs/article/view/104

Attwood,J. (2017). "Form is (Not) Emptiness: The Enigma at the Heart of the Heart Sutra." Journal of the Oxford Centre for Buddhist Studies, 13,52–80. http://jocbs.org/index.php/jocbs/issue/view/15/showToc.

Attwood, J. (2020). "Studying The Heart Sutra: Basic Sources And Methods (A Response To Ng And Ānando)." Buddhist Studies Review, 37 (1-2), 199–217. http://www.doi.org/10.1558/bsrv.41982.

Attwood, J. (2021) "The Chinese Origins of the Heart Sutra Revisited: A Comparative Analysis of the Chinese and Sanskrit Texts." Journal of the International Association of Buddhist Studies 44: 13-52. DOI 10.2143/JIABS.44.0.3290289

Attwood, J. (2022 forthcoming) "The Heart Sūtra Revisited: The Frontier of Prajñāpāramitāhṛdaya Studies. Acta Asiatica [No. 121]. 2021." Buddhist Studies Review, 39(2).

Horiuchi, Toshio. (2021). “Revisiting the ‘Indian’ Commentaries on the Prajñāpāramitāhṛdaya: Vimalamitra’s Interpretation of the ‘Eight Aspects’.” Acta Asiatica 121: 53-81.

Mathes, Klaus‑Dieter. (2021). "The Eight Indian Commentaries on the Heart Sūtra’s Famous Formula 'Form Is Emptiness; Emptiness Is Form'." In Gateways to Tibetan Studies: A Collection of Essays in Honour of David P. Jackson on the Occasion of his 70th Birthday. (2 vols. Indian and Tibetan Studies 12.1–2). Edited by Volker Caumanns, Jörg Heimbel, Kazuo Kano and Alexander Schiller, 659–84. Hamburg: Department of Indian and Tibetan Studies, Universität Hamburg.

Mattice, Sarah A. (2021). Exploring the Heart Sutra. Lanham: Lexington Books.

Wayman, A. (1984) Buddhist Insight: Essays. Motilal Banarsidass.

Eckel, M. D. (1987) "Indian Commentaries on the Heart Sūtra: The Politics of Interpretation." The Journal of the International Association of Buddhist Studies 10(2): 69-79.

Related Posts with Thumbnails