Language and Discrimination.

Benjamin Lee Whorf

In this short essay I want to draw attention to some features of the Buddhist attitude to language as they appear in the Laṅkāvatāra Sūtra, and draw some parallels with the writings of the early 20th century linguist Benjamin Lee Whorf.

The Laṅka is an important text in East Asia, especially in the Zen schools. It is an unsystematic collection of teachings which draws on Yogacāra and Tathāgatagarbha streams of thought, and contains some very interesting passages on the subject of language and how it functions. Much of the material focusses on the process of (false) discrimination (vikalpa or parikalpa) which is a function of the discriminating consciousness (vijñāna) and results in erroneous conclusions about the world.

In the Laṅka the Buddha tells Mahāmati that words are neither the same nor different from discrimination "because words arise with discrimination as their cause". (XXXIII, p.76)(1) He goes on to confirm that words do not, indeed cannot, express the highest reality (paramārtha) since they are dependently arisen, and they are subject to decay and death. The Laṅka strikes a distinctively "Mind Only" note by adding:

Further, Mahāmati, word-discrimination cannot express the highest reality, for external objects with their multitudinous individual marks are non-existent, and only appear before us as something revealed out of mind itself." (XXXIII, p.77)

A few sections later the Laṅka comes back to this theme. Mahāmati enquires as to whether reality is a function of words. But the Buddha points out that some things which are not real are also denoted by words for example such as "hare's horns" or "tortoise hair" for instance. (XLII, p.91) This ability to name fantastic entities reinforces the notion that words do not express reality.

Laṅka also dwells in several places on the relationship of words and meaning (for instance in XLVI, LXI, LXV). There is a distinction here between relative and absolute similar to that in the Awakening of Faith in Mahāyāna. Words can express a relative meaning, but not the absolute. Words illuminate meaning as a lamp reveals things in the dark. (LXV, p.134) However the Buddha warns: "do not fall into the secret error of getting attached to the meaning as expressed in words". (p.160) This is because the process which produces words, ie discrimination, is the same as that which gives rise to attachments and to false speculations about the nature of experience.(2) "As varieties of objects are seen in Māyā [ie illusion] and are discriminated [as real], statements are erroneously made, discriminations erroneously go on",(LXV, p.134) and later "That the unintelligent declare words to be identical with meaning, is due to their ignorance as to the self-nature of words... words are dependent on letters, but meaning is not" (LXXVI, 166-7). And again "words are bound up with discrimination and are the carrier of transmigration". (LXXVI, p.169)

The Laṅka appears to be the source of the now famous aphorism about the finger and the moon: "For instance, Mahāmati, when a man with his finger-tip points at something to somebody, the finger-tip may be taken wrongly for the thing pointed at". (LXXVI, p.169).

The overall impression is that because of the process of discrimination, i.e. of dividing the world up into nameable entities, we make a categorical error and assume that because we can identify something and name it, that it must be real in some sense (and specifically we are thinking in Buddhist terms here of something lasting, substantial and not disappointing - c.f. my earlier essay The Apparatus of Experience). It is this categorical error that keeps us ignorant and prolongs our suffering. Language therefore is problematic because it is a product of the system of error. It is semiotic in that we are able to communicate and make some sense of things, but it is also asemiotic because it hides the ultimate meaning from us if we use it naively - and everyone one except a Buddha does this. So now let us look at some ideas from Benjamin Lee Whorf.

Whorf has been somewhat eclipsed by contemporary linguists such as Noam Chomsky. However for Buddhists I think his thought has many interesting features. In one his discussions of grammar whorf says:

We cut nature up, organize it into concepts, and ascribe significances as we do, largely because we are parties to an agreement to organise it in this way. [Whorf : 213]

By which he means we conceptualise nature in the way we do because of the grammar we have inherited as part of our culture. This operation is unconscious to the extent that we do not even perceive that we are doing anything of the sort: the way we think of the world, the way we divide it up, is completely natural and logical to us. Whorf's comparisons with North American Aboriginal languages, especially Hopi, show that this is far from the case. Whorf called this difference, which is so obvious between English and Hopi, Linguistic Relativism. This theory holds, he says:

that all observers are not led by the same physical evidence to the same picture of the universe unless their linguistic backgrounds are similar, or can in some way be calibrated. [Whorf : 214]

It is difficult to summarise the evidence for Whorf's conclusion - it rests in analysing Hopi and other languages in terms of their relationship to English,(3) and all of this has to be expressed in English which can only express things in its own terms. As an example Whorf tells us that the Hopi language has no words that map onto the Indo-European concept of time. They do not express multiples of units of time, and indeed do not appear to have units of time at all. Day for instance, is not a unit of time but a state. You can't have more than one day, day is simply day, more like our "daytime". Day is when the sun is out. It's hard to get across because time, and units of time, are built into the grammar of English and it's almost impossible to think outside that frame work.

One of the most fundamental grammatically based distinctions we make is dividing the world into objects and processes - nouns and verbs. Another American Indian language has no words which we would think of as nouns. Everything is a process. And how often do we Buddhists say this: "everything is a process"? But the word "everything" is a noun. It makes the statement false in a sense. Things are not processes in our grammar and therefore, according to Whorf, we cannot help but conceive of them as things. We also make statements like "it is raining" grammatically implying an agent which is doing the raining. The agent is fictional, we know this, but it is there in our grammar and it affects our world view. Our world appears to be made up of agents and actions. Whorf shows that in cultures where the grammar is radically different this need not be the way the world is made up.

So it seems to me that the dilemma which is being spelt out in the Laṅka is quite similar in many ways to Whorf's account of grammar and its effect on our views. We might say that the process of making sense of our world is underpinned by the grammar of our language: we cannot help but see things in terms of nouns and verbs, agents and actions, because that grammar has become integral to our thinking. The Laṅka highlights the dilemma that this creates in terms of epistemology - we make categorical errors when deciding what our experience of the world is telling us. We evaluate the data of our sense and minds, in terms of agents and actions - we come to feel that agents in particular, are real. Not real in a thought-out philosophical sense, but in a more gut level way - to most of us it only becomes apparent that we think of "things" as real when we are confronted with a text like the Laṅka which says things are not real, and we rebel against the idea.

Whorf's account of the influence of grammar on world view is interesting because it is a confirmation of the Buddhist approach, and it is logical and presented in rationalistic terms. It should therefore appeal to a Western Buddhist audience. He confirms Buddhist observations of how the mind makes sense of its input from the senses. There are also parallels with contemporary neuroscience that others have already begun to explore - I find the work of Antonio Damasio illuminating for instance, but that is a subject for another essay.

As Buddhists we recognise a central problem: how to correctly understand our experience of the world. We tend to project onto experience qualities which are the opposite of what it is really like. And this leads to constant disappointment. Both Whorf and the Laṅka show that the problem is not a trivial one. Our world view, the way we interpret our experience, is determined by deep grammatical structures (according to Whorf), but even more fundamentally by the very cognitive processes which give rise to language (according to the Laṅka). Making a change at this level of our psyche is never going to be easy. Even knowledge of the nature of the problem is conceived of within the system which is problematic, and this knowledge of itself cannot be enough to effect the major change required. This change is referred to as parāvṛitti and translated by Suzuki as "revulsion". However I think Sangharakshita comes closer to the spirit of the word when he renders it as "a turning about in the deepest seat of consciousness". (Sangharakshita 1994)


Notes.

1. In citing the Laṅka I will give the section no. assigned by Suzuki, and the page number in his translation. I have consulted the Sanskrit text to some extent mainly establishing key words as I don't know much Sanskrit. Suzuki for instance translates vāc as "words" throughout these passages. It would more usually mean "speech", and while there may be a reason for selecting "words" Suzuki doesn't give it. Return to text.

2. For those familiar with Ruichi Abe's The Weaving of Mantra, this is where I find his argument comes unstuck. He seems to be mapping the contemporary Western notion of the distinction between things as what makes them meaningful onto the Buddhist notion of discrimination. He wants this to be a "semiosis" or meaning making process, and for Kūkai in particular to have adopted this kind of view. However discrimination, while it does produce the identification of separateness and words to name things, is in Buddhism a source of illusion and falsification! We know that Kūkai was familiar with this kind of argument, and with the Laṅka in particular since he quotes from it. Abe seems not to take this into account in his argument and it is a fatal flaw. He further errs, in my opinion, by not taking into account the ancient episteme which must still have been functioning in Kūkai's time. And this says that for there to be knowledge there must be sameness, rather than difference. Indeed section LX (p.122 f.) of the Laṅka dwells on the importance of sameness in the Enlightened consciousness. I am currently also exploring the idea that some of these Laṅkāvatāra passages form part of the background for Kūkai's linguistic work: Shōji jissō gi. Return to text

3. Whorf reminds us that all Indo-European languages, the family of languages which includes both English and Sanskrit, are quite closely related when it comes to world view. Return to text


Bibliography

• Whorf, Benjamin Lee. 1956. Language, thought, and reality : selected writings of Benjamin Lee Whorf. (John B. Carroll. ed.) Cambridge Massachusetts : The MIT Press. See especially Science and linguistics (p.207-219)
• Sangharakshita. The meaning of conversion in Buddhism. Birmingham : Windhorse Publications, 1994.
  
• Suzuki, D. T.
  • The Lankavatara sutra : a Mahayana text. London : G. Routledge & Sons, Ltd., 1932.
  • Studies in the Lankavatara sutra, one of the most important texts of Mahayana Buddhism. London : G. Routledge, 1930.
  
  

image: Benjamin Lee Whorf. This image appears on more than a dozen websites and appears to be public domain.

Reading Buddhist texts

The average Buddhist reader has a naive approach to Buddhist texts. At worst we take them at face value as being what the texts themselves say they are: the actual words of the Buddha. At best we are slightly suspicious about translations, and might compare more than one when studying a text.

We all interpret texts. The study of the methods we employ when we read a text, the presuppositions and assumptions we bring to text, and the ideas that underlie how we understand a text, is known as Hermeneutics. The hermeneutic we employ in reading a text, or what we believe the text to represent even, determines to a large extent what we understand a text to be saying. As an exercise in this post I'm going to write about some of my working assumptions in reading a Buddhist text.

I am strongly influenced by contemporary Buddhology. For instance on the basis of scholarship - linguistic, historical, and text critical - I do not believe that the words preserved in Buddhist texts can have been literally spoken by the historical Buddha. The evidence is overwhelmingly that the Buddha did not speak Pāli or Sanskrit or any of the other languages in which texts are preserved. Therefore the texts have been translated at least once. Equally I do not accept that all of the teachings literally came out of the mouth of one person. The Buddhist doctrine evolved over time and new ideas and practices were added at each step along the way. Clearly some texts were written after the lifetime of the Buddha (I do accept the high likelihood of there having been a historical Buddha).

The Buddhist texts were preserved as an oral tradition for several centuries, and there is little or no evidence for the use of sophisticated mnemonic techniques as used by the Brahmins to preserve the integrity of the text. Also as you read the texts it's clear that some passages have been inserted rather crudely by some later editor. Presumably other more skilled editors were at work. Then there are differences between the Pāli texts and surviving parallels in other languages. Comparing the Pāli and Gāndhārī Dhammapadas for instance one can immediately see that they are far from identical. Some doctrines have been played down by some schools -0 such as the practice of mettā in Theravada. So Buddhist texts can not have the status of divine revelation.

One problem that emerges out of assuming that all the teachings came at once is that there are contradictions. In attempts to resolve these difficulties Buddhists have historically had to make some teachings provisional, and others ultimate. But this is problematic because each text proclaims itself to be the ultimate and final teaching of the Buddha, only to be superseded as something new emerges. The centuries have left us with many unstable towers of texts each one claiming to be the final teaching. Sangharakshita has said that there are no higher teachings, only deeper realisations. Following Sangharakshita I think it is time to dismantle the hierarchies of value and accept that texts emerged over time. If a text is profound, then it is profound. The fact that some later disciple wrote it on the basis of their own experience should not lessen it's value. in other words are we interested in truth or lineage?

Another important aspect of my hermeneutic is that I see texts as idealised. Texts are unreliable guides to history because they are what some people thought the ideal was at some point in time. We need an historical perspective which is not available from the texts themselves in order to fully appreciate that this is so. Greg Schopen for instance, in a long career of iconoclastic debunking of sacred cows, has emphasised the point that epigraphical and archaeological evidence often provides flat contradictions to the texts. Monks not only frequently handled money, for instance, but in at least one case actually printed it! Generally speaking we could say that a text represents the social and spiritual ideal as it was conceived by a particular community at the time it was written. The Pāli commentaries to the Suttas for instance tell us more about Buddhism in 5th century CE Sri Lanka, than they do about 5th century BCE India. They are still useful for understanding the texts they comment on, especially in terms of philology, but must be read cautiously for historical value. In fact some scholars have concluded that Buddhist texts have nothing useful to say about the history of Buddhism. This is going too far I think, but highlights the dilemma.

Many scholars now point to influences in the Buddhadharma. It seems pretty clear that many early Buddhist teachings are in response to Brahminical and Jain religious discourse for instance, and that Tantric Buddhism was in dialogue with Śaivism. I have found one or two examples of this myself, but Prof. Richard Gombrich has lead the scholarly investigation of the sources of the Buddha's ideas. Some concepts can't really be understood without reference to Vedic or Vedantic doctrine. The clearest example of this ātman (Pāli attā) which has it's roots in the general religious discourse of the day, a discourse which is seldom if ever replicated in the modern West. This should not be confused with the early 20th century idea, also popular with Hindu scholars, that Buddhism is a kind of reformed or even unreformed Hinduism. The point is that the Buddha responded to his audience which was frequently deeply versed in Vedic lore. He had to communicate in terms which would be understood by his audience - though he frequently redefines words as he goes. But there are also examples of Buddhist and Hindu texts (for example the Dhammapada and the Mahābharata) which draw on a common pool of religious or moral stories. Aspects of the Pāli canon, the obvious worship of yakṣas for instance do not seem especially Buddhist, but relate to what the people around the Buddhists believed.

We must also remember that Indian religious traditions interacted in a way that it quite foreign to the West. In an earlier post [Religion in India and the West] I compared the Indian religious attitude to the Microsoft business model. Each Indian tradition has adopted and adapted ideas from the others. Where one tradition begins to dominate, the minority religions will borrow their ideas, symbols, and practices. During the time when the Tantra emerged amongst the ruins of the Gupta Empire in the 6th century, many different traditions were synthesised into a grand new pattern incorporating various strands of Buddhism, Vedic ritual, Śaiva ritual and iconography, Vaiṣṇava devotional practices and so on. Westerners have tended to characterise this as a degeneration. Contemporary scholarship has exposed the origins of this attitude in Protestant criticisms on the Catholic church based on the model of the Roman Empire. In fact Tantric Buddhism was a much needed revitalisation of Buddhism, but one which may have had too little influence on mainstream Buddhism to save it from collapse. The decline of Buddhism is often attributed to the invasion of Islamic forces from Persia, but they were only the final nail in the coffin of a moribund institution. The vigour of Tantric Buddhism is obvious in Tibet. In Japan Tantric Buddhism dominated for 400 years, but was eclipsed to some extent by indigenous forms such as Zen and Jodo Shin Shu - although not many people know that Shingon is still the largest Buddhist sect in Japan.

Finally I would say that over recent times I have become increasingly suspicious of translations. Translations often hide a multitude of sins. There are huge problems with representing the intellectual content of Buddhism in English. English language and culture is vastly different from Indian. Often the connotations of an English word used as a translation of a Buddhist technical term are entirely unhelpful. For instance it is quite misleading to translated Dharma as "Law", but this frequently happens. The translator must make a large number of discriminations and decisions in translating a text. Just as in English a word in Pāli or Sanskrit may have a range of meanings (polysemic). Some words are highly polysemic and deciding which of sometimes a dozen or more potential senses are indicate depends on how well the translator knows the language (and English), how deeply they understand the Dharma, and

We need also to be aware that what is available in English translation is not the entire corpus of Buddhist texts. While the Pāli canon has been translated more or less entirely, the older translations are unreliable. In the case of Māhayāna Sūtras we have only a small proportion in English. It is most likely that this has skewed our understanding of the development and emphases of the Mahāyāna. [see also Which Mahāyāna Texts?] Many scholars think for instance that the influence of the White Lotus Sūtra in India was negligible, and that it is a rather idiosyncratic text.

I hope that all of this skepticism does not add up to cynicism. But it's clear that I express a lot more doubt about the provenance of texts and the uses that they are put to, than most of my peers and colleagues. I find that I continue to be fascinated by Buddhist texts and textual studies. I have also found that my studies inform and enrich my practice of Buddhism. I suppose I want to call for an intelligent and informed approach to texts. I am far more intellectual than most of my peers and don't expect them to adopt my rather intellectual approach, but I do hope that if you've read this then you might spend a bit of time examining your beliefs about the texts. Where do your beliefs come from, and on what are they based? Are their elements of blind faith in there for instance? It's good to be aware of biases, of likes and dislikes, and to see how they operate to shape your experience of the world.

Often we are looking for certainty and there is a lingering desire for the texts to be the Absolute Truth (paramartha satya). This is both generally true of humans, but especially true in the post-christian west where ideas of absolutes, and concrete answers remain in our psyches. Since Absolute Truth cannot be summed up or put into words we should be at least a bit suspcious of texts. What words are good for is giving us the recipe. They cannot give us the experience of eating the cake.


image: Sinhalese Buddhist text. Jayarava on Flickr.