Mettā for Plants

sweet love gerbera

Some years ago, while learning Pāli, I made a translation of the Mettasutta, or ‘Discourse on Kindness’, one of the best-known early Buddhist discourses.[i] The verses of this discourse describe how a practitioner should develop the quality of mettā, or ‘kindness’.[ii] Having established oneself in an ethical lifestyle, one develops, imaginatively and emotionally, the quality of kindness to all beings, as part of the training of mind and heart that culminates in liberation and awakening. Among the verses that describe the development of mettā are these:[iii]

ye keci pāṇabhūt’atthi
tasā vā thāvarā vā anavasesā
dīghā vā ye mahantā vā
majjhimā rassakā aṇukathūlā

diṭṭhā vā ye vā addiṭṭhā
ye ca dūre vasanti avidūre
bhūtā vā sambhavesī vā
sabbe sattā bhavantu sukhitattā

Which I translated like this:

Whatever living beings there are,
whether plant or animal, without exception,
whether they are very long or large,
or middling in size, or short, great or small,

whether they are visible or unseen,
whether living nearby or far away,
whether they are born, or not yet come to be:
may all living beings have happiness.

When I translated these verses, it seemed to me obvious and uncontroversial that the class of ‘living beings’ (pāṇābhūta) should include living beings that are both ‘moving’ (tasa) and ‘still’ (thāvara), and that these Pāli words referred to animals and plants. However, last year, in conversation with Buddhist friends, I discovered that in fact almost all other translators of the Mettasutta translate the words tasa and thāvara as ‘weak or strong’, or words to that effect, with the implication that plants are not included.[iv]

This discovery surprised me. Surely, I thought, the class of living beings towards which Buddhists should develop mettā, or kindness, should include plants as well as animals. But in fact, as I found out, the Theravādin Buddhist tradition excludes plants from the category of sentient beings; it takes the Mettasutta to teach that one should develop kindness towards sentient beings, hence not towards plants. In this post I will argue two things: first, that the original intention of the Mettasutta was to recommend the development of mettā towards all living beings, including plants; and second, that the development of mettā towards plants ought to be an important part of the practice of developing mettā. But before that, some background on the traditional interpretation.

In Bhikkhu Bodhi’s recent translation of the Sutta-nipāta, the verses of the Mettasutta in question are translated like this:[v]

Whatever living beings there are
whether frail or firm, without omission,
those that are long, or those that are large
middling, short, fine, or gross.

Bhikkhu Bodhi also translates the traditional commentary on the Suttanipāta, giving the contemporary reader easy access to the way the Theravadin tradition understood the discourses. The section discussing ‘frail or firm’ reads like this:[vi]

In this way, with the expression “whatever living beings there are” having shown all beings collectively, classified into pairs and triads, now, with the expression “whether frail [tasa] or firm [thāvara], without omission,” he [i.e., the Buddha] shows all these classified by way of this pair. Here the frail [tasa] are “those that tremble (or thirst)”; this is a designation for those with craving and with fear. The firm [thāvara] are those that stand firm; this is a designation for arahants, who have abandoned craving and fear.

In an interesting long note, Bodhi explains how the commentary invokes a word-play on the two meanings of tasa, ‘trembling’ and ‘thirsty’.[vii] The commentary evidently connects ‘trembling/thirsty’ (which Bhikkhu Bodhi and others render into English as ‘frail’, ‘weak’) with living beings that are unawakened and experience craving and fear. By contrast, the commentary connects living beings who are ‘still’ or ‘firm’ with awakened beings who no longer experience craving and fear. Bhikkhu Bodhi admits that this commentarial interpretation feels forced. Not only that, one might add, but the English rendering ‘frail or firm’ does not even get across the forced commentarial explanation. In English, to speak of ‘frail or firm’ living beings tells the reader nothing at all about their craving or awakening. It tells the reader only about their physical and mental strength. In his long note, Bodhi continues:[viii]

Norman 2004, 81, takes the expression [tasā vā thāvarā vā] in its original sense [of ‘moving or still’]…, but since, on this interpretation, thāvara signifies vegetation or inanimate objects, this would mean that mettā would be developed towards non-sentient objects, which is contrary to the intent of the practice [my italics]. While the commentarial explanation may be forced, I would surmise that even during the Buddha’s time tasathāvara had lost its original sense and had come to serve as a conventional expression applicable solely to the domain of sentient beings.

We see, therefore, that Bhikkhu Bodhi translates the verses,  ye keci pāṇabhūt’atthi |tasā vā thāvarā vā anavasesā, as ‘Whatever living beings there are / whether frail or firm, without omission’, following the commentary, and with the surmise that even in the Buddha’s time, the phrase tasathāvara already meant ‘sentient beings’, excluding plants. Without implying any criticism of Bhikkhu Bodhi, since he has translated the Sutta-nipāta as it is understood in the Theravādin tradition, I would like to offer an alternative interpretation of the original meaning of tasathāvara. This is based on the remarkable in-depth scholarship of Lambert Schmithausen. In his unlikely-sounding book, The Problem of the Sentience of Plants in Earlier Buddhism,[ix] Schmithausen addresses the question of whether tasathāvara includes plants in early Buddhism. This phrase, meaning ‘moving or still’, is a common expression in ancient Hindu and Jain texts from the time of the Buddha for the class of living beings. The Jains, indeed, have not at all changed their conception of what counts as a living being, or jīva. As Paul Dundas puts it, according to Jain belief:[x]

Embodied jīvas are divided into two types, those which are stationary (sthāvara) such as plants, and those which are moving (trasa) such as insects, gods, hellbeings, animals and human beings.

For Jains, the practice of non-harming (ahiṃsa) extends to stationary (sthāvara) beings like plants as well as to moving ones like animals. In Schmithausen’s view, the Buddhists, like the Jains, used the word pāṇa (‘living being’) in a comprehensive sense, to include both tasa and thāvara, animals and plants. Schmithausen reviews early Buddhist literature and concludes that we should infer that the Buddhists used the phrase tasathāvara in just the same way as the Jains; practitioners should not harm or kill living beings, whether moving or still, but should protect them and suffuse them with mettā. The evidence that Schmithausen presents, despite being inferential rather than direct, very much undermines Bhikkhu Bodhi’s surmise that the meaning of tasathāvara had already in the Buddha’s time come to refer only to sentient beings.[xi]

However, Schmithausen also traces the way in which later Buddhists (such as the Pāli commentators) came to exclude plants from the category of sentient beings. He also ventures an opinion on how this change could have come about. Even in the Buddha’s day, plants were regarded by the Buddhists as borderline cases of sentient life; after all, harming plants was a necessity for obtaining food, without which no ascetic could eat and gain liberation. While wanton destruction of plants, based on an attitude of greed or hatred, was wrong, the careful use of plants for food did not incur any bad karma.[xii] This pragmatic attitude, so typical of Buddhists, was quite different to that of the more literalist Jains. In later times, the Buddhist attitude to plants shifted to exclude them altogether from the class of sentient beings, in a doctrinal shift that sorts out the ambiguities of Buddhist pragmatism.

Therefore, we should understand the original meaning of tasathāvara in the Mettasutta as ‘moving and still’, that is, ‘animals and plants’. I now turn to the idea that the intention of the mettā practice is to develop kindness only towards sentient beings (not plants). Bhikkhu Bodhi’s understanding of the mettā practice here no doubt reflects the practice as explained in Buddhaghosa’s Visuddhimagga.[xiii] Here, one is instructed to develop mettā towards human beings, in stages, beginning with oneself, then a good friend, a neutral person and a difficult person, and culminating in extending mettā to all sentient beings. But this constitutes, for practical purposes, a relatively narrow method of practising mettā. The early discourses, by contrast, teach the practice of mettā in terms of radiating boundless kindness in all directions, to all living beings, not specifically to human or sentient beings.[xiv] Again, this suggests that mettā should be developed towards plants.

Indeed, some contemporary meditation teachers recommend the development of mettā towards plants. Sharon Salzbergdraws on research that shows how elderly people in a care home who had been given a pot plant to care for became healthier and better connected to the world. Ajahn Brahm describes how one of his students began to develop the quality of mettā by bringing to mind the plants she had recently re-potted: she developed an attitude of appreciation, kindness and concern to those plants, and was subsequently able to extend this development of mettā towards humans and all beings. Such meditation teachers still teach the traditional five-stage practice of mettā-bhāvanā, but take a broad and creative approach to contacting the quality of mettā to start with.

Perhaps we should go further than this. In the modern world, many people are disconnected from nature and lack a sense of emotional appreciation of the living environment, upon which we depend for food, air, beauty, and more. As the mostly urban-dwelling humanity of the 21st c. heads towards the growing challenges of climate change, biodiversity loss, and over-population, the deliberate and systematic development of mettā for plants, trees and forests, in addition to animals, including humans, might be particularly valuable. An appreciation of our place in the wider world of life may awaken the heart to kindness, and then help us to formulate new attitudes and relationships to plants and insects, and to all the hidden interconnections between our lives, that we have for so long been able to take for granted, but which there is still time to learn to love. 

With this in mind, I propose that we should interpret the Mettasutta for our own times. We should translate tasā vā thāvarā vā as ‘whether plant or animal’, but we should understood these two kinds of living beings as representative of the whole world of life, including bacteria, plants, fungi and animals, and whatever other living beings are yet to be identified. And in our practice of mettā we should extend the quality of kindness towards the whole borderline-sentient world of plants, trees, forests, now at risk from human beings. Hence:

Whatever living beings there are,
whether plant or animal, without exception,
whether they are very long or large,
or middling in size, or short, great or small,

whether they are visible or unseen,
whether living nearby or far away,
whether they are born, or not yet come to be:
may all living beings have happiness.

[i] The Mettasutta can be found in the Suttanipāta, 1: 8; see for editions and translations. My translation can be found here

[ii] The Pāli word mettā is derived from the word mitta, ‘friend’, which suggests the meaning ‘friendliness’ (the Sanskrit equivalent maitrī is similarly derived from mitra). The word mettā can also be translated as ‘love’, ‘loving-kindess’ and ‘benevolence’. But I like the one-word translation ‘kindness’, as the English word ‘kindness’ means the quality of being friendly, generous and considerate, which is more specific than ‘love’, and suggests emotional open-heartedness. 

[iii] Suttanipāta, vv.146–7, taken here from the PTS edition.

[iv] For instance, H. Saddhatissa (The Sutta-Nipāta, London: Curzon, 1985, p.16) translates, ‘Whatever living beings there be: feeble or strong…’; Laurence Khantipalo Mills: ‘whether they be frail or strong’.The exception is K.R. Norman (The Group of Discourses, PTS, Oxford, 2001, p.19), who translates ‘Whatever living creatures there are, moving or still without exception…’.

[v] Bhikkhu Bodhi, 2017, The Suttanipāta: An Ancient Collection of the Buddha’s Discourses, Boston: Wisdom, p.179.

[vi] The Suttanipāta, p.577. I have included some of the Pāli in [square brackets]. Text in bold is quotation from the Mettasutta, the ‘lemma’, or text which the commentary comments on.

[vii] Bodhi, n.696, p.1407.

[viii] The reference to Norman is to K.R. Norman, ‘On Translating the Sutta-nipāta’, Buddhist Studies Review, 2004, 21: 1, pp.69–84. Bodhi’s reference should be to p.82 rather than p.81.

[ix] Lambert Schmithausen, The Problem of the Sentience of Plants in Earliest Buddhism, Tokyo: The International Institute for Buddhist Studies, 1991. This book is an accompaniment to Schmithausen, Buddhism and Nature, Tokyo: The Internation Institute for Buddhist Studies, 1991. Neither work is easy to get hold of, but I have created links here to downloadable versions.

[x] Paul Dundas, The Jains, Abingdon: Routledge, 1992, p.95

[xi] Here I summarise the detailed discussion in Schmithausen, Sentience, §§19–21, pp.58–65.

[xii] Schmithausen makes this argument in Sentience, §§22–7, pp.66–78.

[xiii] Buddhaghosa, Visuddhimagga, PTS: London, 1920, p.295ff; trans. Bhikkhu Ñāṇamoli, The Path of Purification, Kandy: BPS, 5th ed., 1991, p.288f.

[xiv] Discussed in Anālayo, Compassion and Emptiness in Early Buddhist Meditation, Cambridge: Windhorse, 2015, pp.20–6.

The meaning of the Pāli word ‘sutta’

well said

Many Buddhists are familiar with the Pāli word sutta: it is equivalent to the Sanskrit word sūtra and it means ‘discourse’. It is used in the sense of a discourse of the Buddha, one of the many discourses which generally begin evaṃ me sutaṃ, ‘thus have I heard’, and which are traditionally regarded as having been remembered by Ānanda, the Buddha’s friend and attendant.[1] At the same time, from the point of view of the word itself, we often read that the word sutta does not literally mean ‘discourse’, but that it means ‘string’ or ‘thread’, and that the meaning ‘discourse’ is an applied meaning. However, in this essay I will show how some recent as well as traditional scholarship does not support the idea that sutta means ‘string’ or ‘thread’, but that the word was always understood to mean ‘discourse’.

Let us begin with the Pali-English Dictionary (PED). There are in fact two entries for sutta, in that sutta1 means ‘asleep’, being the past participle of supati ‘sleeps’. We can leave this meaning of sutta aside. The other meaning is as follows:

sutta2 (nt.) [Vedic sūtra, fr. sīv to sew] 1. a thread, string… 2. the (discursive, narrational) part of the Buddhist Scriptures containing the suttas or dialogues, later called Sutta-piṭaka… 3. one of the divisions of the Scriptures (see navanga)… 4. a rule, a clause (of the Pātimokkha)… 5. a chapter, division, dialogue (of a Buddh. text), text, discourse… 6. an ancient verse, quotation… 7. book of rules, lore, text book…[2]

PED thus relates sutta to Sanskrit sūtra and both words to sīv ‘to sew’, and gives its primary meaning as ‘thread’ as well as other meanings including ‘discourse’. Following PED, Buddhist commentators have tried to explain why a word meaning ‘string’ or ‘thread’ should also be used as the word for Buddhist discourse. Sangharakshita, for instance, explains that:

meaning literally a thread, the word [sūtra, also sutta] suggests a series of topics strung on a common thread of argument or exhortation. By implication, therefore, a sūtra is of considerable length, systematic in form and substantial in content.[3]

However, there is a puzzle associated with this kind of explanation.

There certainly is a Sanskrit word sūtra meaning ‘string’ or ‘thread’, and there certainly is a Pāli word sutta with the same meaning.[4] There also certainly is a genre of Sanskrit literature called sūtra. This genre, perhaps the best-known example of which is the Yoga-sūtra of Patañjali, consists of a number of brief aphoristic sayings in verse (each called sūtra). The genre long pre-dates Buddhism, being first used around 800 BCE in the Śrauta-sūtras, concerned with Vedic ritual, and the genre remained important in philosophy and literature for many centuries. The aphorisms of the genre can certainly be said to have been strung together or to have a common thread, and perhaps were so-called for this reason. However, neither the Pāli suttas nor the Sanskrit Buddhist sūtras are like this at all. The Buddhist discourses are not in the least aphoristic and neither do they consist in sayings of the Buddha strung together. It is therefore a puzzle to read even in an up-to-date Dictionary of Buddhism under the entry sūtra:

In Sanskrit, lit. “aphorism”, but in a Buddhist context translated as “discourse”, “sermon”, or “scripture”; a sermon said to be delivered by the Buddha or delivered with his sanction. A term probably used originally to refer to sayings of the Buddha that were preserved orally by his followers (and hence called “aphorisms”), the sūtra developed into its own genre of Buddhist literature, with a fairly standard set of literary conventions…[5]

Reading this entry, one might reasonably ask why a word meaning “aphorism” would have been used to describe the oral record of the Buddha’s teaching, and why this word later came to refer to a genre of Buddhist literature which was not in the least aphoristic.

Scholars have proposed a pleasing and elegant answer to this puzzle. It is that we have been misled by the Sanskrit word sūtra into supposing that the Pāli word sutta means ‘thread’ and therefore ‘aphorism’. The Indian Buddhists who used the Sanskrit word sūtra were incorrect to use it as an equivalent to the older Middle-Indo-Aryan word sutta, and this earlier word should actually be derived from sūkta, meaning ‘well-spoken’, hence ‘discourse’ of the Buddha. As Prof. K.R. Norman puts it:

Many Buddhist Sanskrit texts are entitled sūtra. To anyone who comes to Buddhist studies from classical Sanskrit studies, this name comes as a surprise, because, in Sanskrit, sūtra literature is a specific genre of literature, composed in prose, usually of a very abbreviated and concise nature, while Buddhist sūtras have an entirely different character. This difference is due to the fact that the word sūtra in Buddhist Sanskrit is a Sanskritisation of the Middle Indo-Aryan word sutta, which is probably to be derived from Sanskrit sūkta, a compound of su and ukta, literally “well-Spoken”. It would be a synonym for subhāṣita, which is the word used of the Buddhavacana [sayings of the Buddha] by the emperor Aśoka… when he said: “All that was spoken by the Lord Buddha was well-spoken”.[6]

According to this explanation, the word sutta means ‘well-spoken’ and hence ‘discourse’ of the Buddha, from the verb vac ‘to speak’ (the past participle of which is ukta) with the prefix su meaning ‘well’, ‘good’ or ‘excellent’. If this is true, the early Buddhists who used the word sutta to mean ‘discourse’ did so with good reason, and did not do so thinking that sutta meant ‘string’ or ‘thread’. This meaning of sutta is to be understood as distinct from the meaning of sutta as ‘thread’, just as sutta also means ‘asleep’. Hence, sutta1 ‘asleep’ (past participle of sup), sutta2 ‘thread’ (from sīv), sutta3 ‘discourse’ (from su+ukta).

Inevitably, however, other scholars have found fault with the details of this explanation. Prof. Oscar von Hinüber thinks that this proposed etymology of sutta from sūkta is unnecessary. He writes:

In der Theravāda-Überlieferung findet die Annahme, daß sutta eigentlich sūkta- entspräche, nirgends eine Stütze, wie die lange Erörterung zu sutta-, As 19, 15–26, mit aller Deutlichkeit zeigt.[7]

In the oral tradition of the Theravāda, the assumption that sutta really corresponds to sūkta nowhere finds a support, as the long discussion on sutta in As 19, 15–26, quite distinctly shows.

Von Hinüber’s point is that, while it is theoretically possible that sutta is derived from sūkta, and that this would elegantly explain its usage, there is no traditional support for such a derivation. He cites the Atthasālinī, the Theravādin commentary on the Dhammasaṅganī, the first book of the Abhidhamma-piṭaka.[8] This commentary gives the following explanation of the word sutta:

atthānaṃ sūcanato, suvuttato savanatotha sūdanato;

suttāṇā suttasabhāgato ca ‘suttan’ti akkhātaṃ.

tañhi attatthaparatthādibhede atthe sūceti. suvuttā cettha atthā veneyyajjhāsayānulomena vuttattā. savati cetaṃ atthe, sassamiva phalaṃ, pasavatīti vuttaṃ hoti. sūdati cetaṃ, dhenu viya khīraṃ, paggharatīti vuttaṃ hoti. Suṭṭhu ca ne tāyati rakkhatīti vuttaṃ hoti. suttasabhāgañcetaṃ. yathā hi tacchakānaṃ suttaṃ pamāṇaṃ hoti evametampi viññūnaṃ. yathā ca suttena saṅgahitāni pupphāni na vikiriyanti na viddhaṃsiyanti evametena saṅgahitā atthā. tenetametassa vacanatthakosallatthaṃ vuttaṃ –

atthānaṃ sūcanato, suvuttato savanatotha sūdanato;

suttāṇā suttasabhāgato ca suttanti akkhātan’ti.[9]

From showing (sūcana) the good, from having been well spoken (suvutta), from begetting (savana) and from giving out (sūdana);

Through being an excellent shelter (suttāṇa), and from being like thread (sutta), sutta is called ‘sutta’.

For it shows the good (attha) consisting of the good for one’s self, the good for others, and so on. And meaning (attha) has been well spoken in this respect through being spoken in conformity with the dispositions of those ready for the teaching. And it begets the good (attha), like crops do fruit, so it is said that it brings forth. And it gives it [the good] out, like a cow does milk, so it is said that it flows out. And it excellently shelters and protects it [the good]. And it is similar to thread, for as the carpenter’s thread is a measure, so it is too for the wise, and as flowers tied together with thread are not scattered and damaged, so by it good things are tied together. Therefore this has been said about it for the sake of knowledge about the meaning of the word: [repeat of stanzas].’[10]

This traditional discussion of the meaning of sutta is revealing, in that although Von Hinüber is correct in saying that it does not definitively support the derivation of sutta from sūkta, neither does it support the derivation of sutta from the word sutta meaning ‘thread’. Let us look more closely at this traditional explanation.

The Atthasālinī explains the meaning of sutta (as in sutta-piṭaka, the ‘discourse collection) in six distinct ways:

  1. It means sūcana (‘showing’, ‘indicating’), as it shows the good. The word sūcana comes from sūcī (‘needle’) via the denominative root sūc. Monier-Williams’ Sanskrit Dictionary (MW p.1241) links sūcī with sīv (‘to sew’), the root of sūtra, but the PED doubts this connection (p.721).
  1. It means suvutta (‘well spoken’, ‘well said’) from su (‘well’, ‘excellent’) and vac (‘to speak’). This explanation amounts to the derivation proposed by Norman, that sutta = Sanskrit sūkta. However, the commentary is not claiming that sutta is the same word as suvutta or that it derives from su+vac, only that sutta can be explained as suvutta (‘what has been well spoken’).
  1. It means savana (‘begetting’), which can be derived from the Sanskrit root su4 (= 2) (‘to generate’) (MW p.1219). This explanation gains strength from the fact that the past participle of su is suta, literally meaning ‘issue’, hence ‘son’ (PED p.717). There is hence an edifying background word-play between sutta and savana via suta.[11]
  1. It means sūdana (‘gives out’), which is cognate with the Sanskrit root sūd, which, according to MW p.1242 can have the meaning ‘eject’ (nikṣepana).
  1. It means su+(t)tāṇa (‘excellent shelter’), from the prefix su together with the word tāṇa (‘shelter’), cognative with Sanskrit trāṇa from the root trai (‘shelter’, MW p.457). This explanation is an example of explanation through edifying word-play, since the commentator would not have supposed that the word sutta was etymologically connected with suttāṇa, only that the resemblance of sounds between the words could be exploited to explain the meaning of sutta. 
  1. The final explanation is in the form of a comparison. Sutta is said to be suttasabhāga (‘like or similar to sutta’) where sutta in this case means ‘string’ or ‘thread’, which is derived from the Sanskrit root sīv (‘to sew’).

From these six explanations of the meaning of the word sutta, we can see how the commentators primarily took the word to mean ‘discourse’, and then they explained this meaning in various ways, relating sutta to other words that were either homonyms (sutta meaning ‘thread’), or were edifyingly similar in sound (suttāṇa, sūdana, sūcana), or were both similar in sound and related in meaning (suvutta), or were related in meaning (savana). The impression one gets is that the commentator does not have a single view about the derivation of sutta.

However, from a historical perspective the commentator’s explanation of the meaning of sutta is from a later period, and does not tell us much about how the early Buddhists who first used the word sutta understood it. We can also only wonder whether the commentator was familiar with the Buddhist Sanskrit word sūtra meaning ‘discourse’ as the equivalent of the Pāli sutta. If he was, which seems likely, then two interesting conclusions seem to follow. Firstly, the Pāli commentator does not seem to relate the words sutta or sūtra to the genre of Indian literature called sūtra or ‘aphorism’. Rather, the words sutta or sūtra are explained as comparable to a string or thread only as an edifying metaphor. Secondly, the Pāli commentarial explanation of sutta seems to allow that this word may be the equivalent either of Sanskrit sūkta or of sūtra.

In conclusion, then, the Pāli word sutta, when used to refer to Buddhist literature, need not be taken literally to mean ‘thread’. It is equally possible to derive sutta from su+ukta as from the root sīv (‘to sew’), and the former derivation would support the meaning of sutta as ‘discourse’, in the sense of ‘what has been well spoken (by the Buddha)’. While the Pāli commentary does not give any direct support to this derivation, it does support the meaning of sutta as ‘discourse’ and does not appear to support any connection of sutta to the Sanskrit word sūtra meaning ‘aphorism’, derived from the meaning of sūtra as ‘thread’. In short, despite our not knowing for certain the derivation of sutta, it is consistently used to mean ‘discourse’ in a way that supports its derivation from sūkta, ‘well-spoken’.

[1] The situation is in fact more complicated, in that the early Buddhist scriptures record a nine-fold analysis of Buddhist literature, the first sort being sutta, meaning ‘discourse’, the second being gāthā, ‘verse’, the third geyya, ‘mixed prose and verse’, and so on. However, this nine-fold analysis appears to have been superseded by the more now-familiar three-fold division of the scriptures into three piṭakas or collections, including the sutta-piṭaka or ‘discourse collection’.

[2] Rhys Davids and Stede, Pali–English Dictionary, PTS: London, 1925, p.178.

[3] Sangharakshita, The Eternal Legacy, Tharpa: London, 1985, p.14. Cf. A Survey of Buddhism, 6th ed., Tharpa: London, 1987, p.17.

[4] This and the following information from Brian Levman, Linguistic Ambiguities, the Transmissional Process, and the Earliest Recoverable Language of Buddhism, unpublished PhD thesis, 2014, pp.228–30.

[5] Robert Buswell and Donald Lopez, The Princeton Dictionary of Buddhism, Princeton University Press, 2014, p.875.

[6] K.R. Norman, A Philological Approach to Buddhism, PTS: Lancaster, 2006, p.135. His explanation was first suggested by Walleser in 1914. Norman’s suggestion has been taken up by Richard Gombrich, ‘How Mahāyāna Began’, in Journal of Pāli and Buddhist Studies, 1988, 29–46, p.32; also by Rupert Gethin, Sayings of the Buddha, Oxford University Press, 2008, p.13, n.1.

[7] Oscar von Hinüber, ‘Die neun Aṅgas: ein früher Versuch zur Einteilung buddhistischer Texte’, WZKS 38, 1994, 121–35, p.132. Von Hinüber’s view is followed by Johannes Bronkhorst in Buddhist Teaching in India, Wisdom: Boston, 2009, p.xi n.4.

[8] As is an abbreviation for Atthasālinī.

[9] Atthasālinī ed. Edward Müller, PTS: London, 1897, p.19.

[10] My translation; there is also a PTS trans. by Pe Maung Tin and Mrs Rhys Davids, The Expositor, vol.1, London: PTS, 1920, p.24. The explanation of sutta is also found in a slightly different form in the commentary to the Sutta-nipāta, the Paramatthajotika II, vol.1, ed. Helmer Smith, PTS: London, 1916, p.1.

[11] Thanks to Bryan Levman for his advice on savana. This word may be related to several different Sanskrit roots: ‘impel’, su ‘press out’ as well as su ‘generate’. It is possible that the Pāli commentators had several meanings in mind.

‘Going for Refuge’ as Idiom and Metaphor


Anyone familiar with Buddhism will know that Buddhists say that they ‘go for refuge’ to the Buddha, the Dharma (the Buddha’s teaching), and the Sangha (the Buddhist community). This usage goes back to the time of the Buddha, so it is rightly hallowed. Moreover, in the Triratna Buddhist movement, in which I practise, going for refuge is seen as especially important, because not only is it the one thing that all Buddhists have in common but, as Sangharakshita puts it, it is the central and defining act of the Buddhist life.[1] However, it seems to me that the English expression ‘going for refuge’ is also a piece of jargon, and a good example of ‘Buddhist Hybrid English’, that is, an expression which nobody else uses except Buddhists.[2] In this essay I want to explore how this expression ‘going for refuge’ is (i) an inexact translation of (ii) an ancient Indian idiom, which (iii) even in its day was a metaphor in Buddhism. I want to open up some imaginative space in which to explore how to give better expression in English to the central and defining act of the Buddhist life.

‘Going for refuge’ is the usual English translation of the Pāli expression saraṇa-gamana, and it is also used as a title for the traditional form of words by which one declares that one is a Buddhist:[3]

buddha saraa gacchāmi
dhamma saraa gacchāmi
sagha saraa gacchāmi

I go for refuge to the Buddha.
I go for refuge to the Dhamma.
I go for refuge to the Saṅgha.

However, a look at the grammar of the Pāli expression quickly reveals a difficulty for the translator. The verb gacchāmi means ‘I go’ and is related to the verbal root gam (with which ‘go’ is cognate). In Pāli and Sanskrit, the goal of verbs of motion like gam is expressed by the accusative case, whereas in English the goal of verbs of motion is indicated by the preposition ‘to’. Hence here buddhaṃ is in the accusative, and buddhaṃ gacchāmi means ‘I go to the Buddha’. But the word saraṇam, which means ‘refuge’, ‘shelter’ or ‘protection’, is also in the accusative, and is thus in apposition to the word buddhaṃ. We should therefore properly translate the expression buddhaṃ saraṇaṃ gacchāmi as ‘I go to the Buddha as refuge’, and saraṇa-gamana as ‘refuge-going’.[4]

This may not seem so important since ‘I go to the Buddha as refuge’ means more or less the same as ‘I go for refuge to the Buddha’. However, what do we mean by ‘go’ here? We are not actually going anywhere, so what do we mean when we ‘go to the Buddha as refuge’? This question takes us to another issue of translation. The root gam has some meanings that the word ‘go’ does not, such as ‘reaches, obtains’,[5] and for this reason it is perfectly correct to translate buddhaṃ saraṇaṃ gacchāmi as ‘I take the Buddha as refuge’, as long as it is clear that there is no implication that one is taking the refuges from anybody else.

But now, whether we go to the Buddha as refuge, or take the Buddha as refuge, the question arises of what we mean by ‘refuge’. The word saraṇa is used to mean ‘house’ in Pāli,[6] so in a figurative sense we could say saraṇa-gamana means ‘coming home’ to the Buddha, Dharma and Sangha. But really the Buddha is not a building but a ‘refuge’ in the sense of offering shelter and protection. Chogyam Trungpa picks up on a different figurative sense of saraṇa when he explains that anyone who ‘goes for refuge’ must therefore be a ‘refugee’, so that Buddhists are ‘refugees from conditioned existence’.[7] But this figurative language does not explain how the Buddha, the Dharma and the Sangha were considered saraṇas. To understand this we have to go back to the social and linguistic context of ancient India in which the expression saraṇa-gamana arose.

In the Pāli canon it occurs mainly in the context of people declaring their conversion to the Buddha’s teaching, so it evidently became a formulaic expression in the Buddhist tradition early on. Yet this Buddhist use relates to an idiom in which saraṇa-gamana means placing oneself into a reciprocal relationship with a protective power. This is apparent from a later commentarial story illustrating a verse about the unreliability of friends. A king falls ill, and his faithless ministers decide to leave him to seek employment in the court of another king. They think, ‘“This king is not going to live for long; why don’t we seek protection (saraṇaṃ) for ourselves elsewhere?”, and coming before another king they asked to enter his service.’[8] The idiom of saraṇa-gamana is therefore similar to ‘swearing allegiance’ in feudal Europe. In the feudal contract the vassal swears allegiance to the Lord, in return for which he receives the right to land and to protection from the Lord, though the Lord could call on him to bear arms on his behalf. Even today there is a tradition of ‘pledging allegiance’ to the flag in the USA, which has the connotation of a reciprocal relationship, of having a right to the benefits of living under the US constitution by by pledging to serve it.

However, the idiom of saraṇa-gamana had already been transferred from a social to a religious context before the Buddhists took it up. We can see this in the well-known stanzas from the Dhammapada:[9]

Impelled by fear many people go mountains
as refuge (saraṇa), to woods, to park- and tree-shrines.

But this is no secure refuge, not the highest refuge.
Going to this as refuge one is not entirely released from pain.

It is not clear what the first stanza means, and the commentary is not very helpful, but it suggests a religious context in which it was believed that making offerings to supernatural beings, perhaps yakṣas, in the mountains and woods, or at their local shrines, would induce those beings to protect the devotee in return. Certainly there are stories in the great Indian poem, the Mahābhārata, in which people go to trees and their guardian spirits for refuge (śaraṇa) from storms.[10]

Combining the political and religious uses of the idiom of saraṇa-gamana is its use in the context of devotion to Lord Kṛṣṇa. The true devotee surrenders to Kṛṣṇa as śaraṇa, as refuge.[11] This devotional sense of saraṇa-gamana is probably later than the time of the Buddha, but illustrates how, in the Indian religious context, saraṇa-gamana means seeking divine protection. But alongside such religious superstition and devotionalism there was a use of the idiom of saraṇa-gamana in relation to more sophisticated, though still essentially magical, form of thinking, by which the act of giving to a worthy ascetic or samaṇa was believed to produce merit or puñña, which was understood as existential or metaphysical shelter or protection in the sense of bringing a good rebirth. We can see this kind of thinking in a conversation between the Buddha and a young Brahman called Māgha, who approaches the Buddha, wishing to know to whom to give so as to gain the most merit. The Buddha gives a series of stanzas explaining how renunciants like himself are the most worthy of gifts, including the stanza:[12]

One who has knowledge, loves meditation, is mindful,
has attained awakening, is a refuge (saraṇa) for many –
a brahman seeking merit should sacrifice to him
should make an offering at the right time.

The implication here is that, according to the religious belief of the young Brahman, the person worthy of gifts is a saraṇa because giving to him produces merit.

In conclusion, the saraṇa-gamana or taking-refuge was an ancient Indian idiom indicating a movement of assent on the part of the inferior or suppliant or devotee, to a social or religious power, in a reciprocal relation of protection in return for service or devotion or religious giving. This is the cultural context in which the phrase saraṇa-gamana was used. ‘Taking refuge’ meant expressing one’s assent to a superior power that offered protection – even protection from a bad rebirth; and it also signified agreeing to serve that power – for instance, by giving food. In the Buddhist context, it is clear from the early discourses that people expressed their new-found faith in and devotion to the Buddha, his Dharma and the spiritual community, by just such a saraṇa-gamana. This is expressed by a stock paragraph by which people declare themselves converts having heard the Buddha teach:

‘Wonderful, Lord, wonderful! It is as if someone were to set upright something overturned, or explain something that was obscure, or show the way to someone lost, or carry a lamp into the darkness so that those with eyes can see. In the same way, Venerable Gotama has made the Dharma known in many ways. I go for refuge to the Blessed One, to the Dharma and to the Sangha of monks. May the Venerable Gotama remember me as a lay-follower who has taken refuge (saraṇaṃ gataṃ) from this day onward for as long as life lasts.’

It is as if this saraṇa-gamana was simply the accepted form of words by which one expressed one’s conversion to a teaching. In this sense, the saraṇa-gamana simply signifies an act of faith, and what makes one specifically a Buddhist is saraṇa-gamana to the Buddha, the Dharma and the Sangha. We should add that, in the Indian cultural context, a saraṇa-gamana to the Buddha, Dharma and Sangha meant giving food and requisites to the monastic community with the expectation of merit in return.[13]

However, although the Buddha obviously accepted the cultural idiom of saraṇa-gamana and merit-making, he gave it a new and metaphorical twist. Without trying to dissuade people from giving to ascetics like himself, he re-interpreted the meaning of doing so, from a quasi-magical religious act to an explicitly ethical one.[14] He explained that the best kind of saraṇa is acting ethically, and that it is by keeping the precepts that one makes merit and gains existential or metaphysical saraṇa or protection. This is made explicit in a story of two old brahmans, who approach the Buddha to say that they have not done anything good or wholesome, and they have not made a shelter from what is fearful.[15] The Buddha reassures them by explaining that restraint of body, speech and mind will be their shelter, their protection, their island in the flood and their refuge, after their deaths.[16]

Just as the saraṇa-gamana was a reciprocal relation, with the inferior person going and the superior protecting, so the Buddha’s metaphorical re-interpretation of the idiom is reciprocal in that he offers protection by teaching people how to practise ethics. One discourse explains that the Buddha is one who honours and reveres the Dharma, and who provides protection, shelter and defence by setting in motion the wheel of the Dharma, and by making clear what are good actions of body, speech and mind.[17] In the background of the Buddha’s metaphorical re-interpretation of the saraṇa-gamana is his insight into how our character and destiny are shaped by our own actions: our deeds are our refuge. This is explicit in the last of the ‘five topics for frequent recollection’:

I am the owner of my actions, heir to my actions, born of my actions, joined to my actions, and actions are my refuge (paṭisaraṇa). Whatever actions I might do, good or bad, of these I will be the heir.[18]

In this sense, the saraṇa-gamana, the taking-refuge in the Buddha, the Dharma and the Sangha, has the metaphorical implication that we alone are responsible for our destinies, and that the most reliable source of safety in the cosmos is our own accumulation of goodness.

In this brief survey of saraṇa-gamana as idiom and metaphor, I hope to have shown that the English expression ‘going for refuge’ does not communicate the rich network of connections between social protection, religious power and merit-making that the Indian idiom of saraṇa-gamana implies. And in consequence it does not allow us to appreciate the Buddha’s metaphorical twist of the cultural idiom. So why do we use an inexact translation of a idiom which is so culturally remote from us? I would suggest that we do so because we suppose that the expression ‘going for refuge’ is somehow in itself ‘Buddhist’. However, this is not the case, since one is not a Buddhist because of taking refuge, but because of taking refuge in the Buddha, the Dharma and the Sangha. Taking refuge is simply the ancient Indian idiom for conversion, whereas the important thing from the Buddhist point of view is the metaphorical twist to the idiom, by which taking refuge in the Buddha means ‘obtaining shelter’ in virtue.

We will no doubt continue to use the English expression ‘going for refuge’ since it has already become so well established, albeit in Buddhist Hybrid English. But how do we understand for ourselves this jargon, and how can we explain it to others? How, then, to re-imagine the saraṇa-gamana in good, native English? Movement has to be involved, for, just as the old idiom implies going under a tree or out to give alms, so the saraṇa-gamana implies assent and resolve – a change of life, and not just of belief. I myself have been experimenting with the expression ‘placing trust’ as a translation of saraṇa-gamana. When I say, ‘I go for refuge to the Buddha, to the Dharma, and to the Sangha’, I feel myself to be using a kind of code or jargon that I have to interpret to myself; but when I say ‘I place my trust in the Buddha, the Dharma, and the Sangha’, I directly feel the devotional, grateful tenor of the implied existential commitment, while the very words leave an open aura of cultural familiarity in which to hear the deeper implication: I take responsibility for my thoughts, words and deeds; I wrest my destiny from both the cosmic powers and the crowd.

So when I say buddhaṃ saraṇaṃ gacchāmi, dhammaṃ ca saṅghaṃ ca, I translate the expression as ‘I place my trust in the Buddha, the Dharma, and the Sangha’. For me ‘placing trust’ is a better translation of saraṇa-gamana than ‘going for refuge’.

[1] This is explored by Sangharakshita in The History of My Going For Refuge, Windhorse Publications, Glasgow, 1988.

[2] See Paul J. Griffiths, ‘Buddhist Hybrid English: Some Notes on Philology and Hermeneutics for Buddhologists’, Journal of the International Association of Buddhist Studies, 4 (1981) 2:17–32. Buddhist Hybrid English is ‘a dialect comprehensible only to the initiate’ (p.17).

[3] For instance, in Khuddakapatha 1.

[4] The Pāli texts also use the verbs ‘go’ and upa-i ‘come’ with saraṇa to express the same thing, and these verbs also take two accusatives, e.g. saraṇaṃ taṃ upema cakkhuma, ‘We come to you as refuge, man of vision’, Sutta-nipāta v.31; ime dasasatā yakkhā iddhimanto yasassino / sabbe taṃ saraṇaṃ yanti tvaṃ no satthā anuttaro, ‘These one thousand yakṣas, powerful and famous, / all go to you as refuge; you are our supreme teacher’, Sutta-nipāta v.179.

[5] See Margaret Cone, Dictionary of Pāli, vol.II, s.v. gacchati. The old PED, s.v. gacchati, gives the figurative meanings ‘to come to know, to experience, to realize’.

[6] See PED, s.v. saraṇa, and Sutta-nipāta v.591.

[7] ‘By taking refuge, in some sense we become homeless refugees… The point of becoming a refugee is to give up our attachment to basic security. We have to give up our sense of home ground, which is illusory anyway… because we don’t have any home ground, we are lost souls, so to speak.’ Chogyam Trungpa, ‘Taking Refuge: the Decision to Become a Buddhist’, on

[8] From the commentary on Sutta-nipāta v.75, Pj II pts p.129. The faithless ministers came to no good, for their new king did not pay them.

[9] Dhammapada vv.188–9: bahū ve saraṇaṃ yanti pabbatāni vanāni ca / ārāmarukkhacetyāni manussā bhayatajjitā // netaṃ kho saraṇaṃ khemaṃ netaṃ saraṇamuttamaṃ / netaṃ saraṇam āgamma sabbadukkhā pamuccati //

[10] For instance, in Mahābhārata 12:141. There is a summary of the incident in The Mahābhārata, trans. John D. Smith, Penguin, 2009, p.623.

[11] Mahābhārata 13:145; also Śrimad Bhagavatām 11:11:29–32.

[12] Sutta-nipāta v.503: yo vedagū jhānarato satīmā / sambodhipatto saraṇaṃ bahunnaṃ / kālena tamhi havyaṃ pavecche / yo brāhmaṇo puññapekho yajetha

[13] These days we might set up a standing order with the bank to support our teacher or Buddhist Centre, but we would probably not suppose that this would affect our destiny.

[14] The Buddha similarly gave a metaphorical twist to the Brahmanical concept of karma, redefining karma in terms of ethical instead of sacrificial action: see Richard Gombrich, What the Buddha Thought, Equinoxe, London, 2009.

[15] They are akatabhīruttāṇā, for they have not made (akata) a shelter (tāṇa) from fear (bhīru).

[16] This story is from Aṅguttara-nikāya 3:51 and recurs at 3:52.

[17] This is from Aṅguttara-nikāya 3:14.

[18] This is from Aṅguttara-nikāya 5:57.

Pāli Safari


This review recently appeared in the Western Buddhist Review, at

Bhikkhu Anālayo, Excursions into the Thought-World of the Pāli Discourses, Pariyatti Press, Onalaska WA, USA, 2012, 326pp., $19.95 pback

The idea of an ‘excursion’ into the thought-world of the Pāli discourses suggests a guided tour, a day trip, or, figuratively, an author’s venture into unfamiliar territory. However, Bhikkhu Anālayo’s book is more like a safari into the heart of Pāli country, with a capable guide who is not afraid to show you some unexpected features of the less well-known areas. Anālayo explores 24 different words used in the Pāli Buddhist discourses, beginning with Craving (taṇhā), and ending with Liberation (vimutti). The entries vary in length, depending on the relative importance of each topic and the difficulties encountered in exploring it, but they follow a similar pattern. An introduction to the word’s meaning and implication is followed by a survey of its usage in the discourses (and sometimes in the vinaya and in some post-canonical works). This is followed by a discussion of the implications of this term for Buddhist practice, concluding with an evocation of the successful result of such practice. But this summary of the formula for each essay does not do justice to the effect. Anālayo has a wonderfully broad and discerning knowledge of the Pāli discourses, and his essays present a formidably learned but nevertheless tremendously inspiring basic ground of clarity about early Buddhist concepts.

Bhikkhu Anālayo’s scholarship began with his PhD exploring the Satipaṭṭhāna-sutta (published by Windhorse, Cambridge, 2003, as Satipaṭṭhāna: The Direct Path to Realization). More recently he has learned Chinese and has been writing about the relationships between early Buddhist discourses preserved in Pāli and those originally written in other Indian languages, which now survive only in Chinese translation. Anālayo combines scholarship in Buddhist scriptures with actual Buddhist practice based on those same texts, and in Excursions he writes very much as a scholar-practitioner, mainly about Pāli discourses, but with the occasional reference to the Chinese parallels to certain difficult passages. The 24 essays were originally written for the Encyclopaedia of Buddhism, a Sri Lankan project begun soon after the 1955 Buddha-jayanti, and only completed in 2009, in a total of eight substantial volumes. Many scholars have contributed over the years, with Anālayo evidently joining in at a late stage, since his contributions begin with rāga (among the entries under ‘r’) and run through to Yuganaddha Sutta.

Hence the apparently random selection of topics in this book, which are actually edited versions of the essays appearing in the Encyclopaedia, but here made to work together in an inter-related collection. Being of the nature of reference articles, the essays will be basically familiar to anyone who has studied Pāli Buddhism, or consulted Nyanatiloka’s Buddhist Dictionary. However, Anālayo’s work takes the genre of a reference article to new levels of thoroughness, and these provide a fascinating survey of usage and nuance. There are essays on each of the five hindrances – Passsion (rāga), Ill-will (vyāpāda), Sloth-and-torpor (thīnamiddha), Restlessness-and-worry (uddhaccakukkucca), Doubt (vicikicchā); essays on four of the twelve nidānas of paṭicca-samuppāda – Volitional Formations (saṅkhārā), Feeling (vedanā), Craving (taṇhā), Clinging (upadāna); and essays on four of the twelve factors of the path – Happiness (sukha), Concentration (samādhi), Knowledge and Vision according to Reality (yathābhūtañāṇadassana), Liberation (vimutti). There are also fascinating essays on Personality View (sakkāyadiṭṭhi) and Contemplation of feelings (vedanānupassanā), as well as on lesser-known concepts in Pāli Buddhism such as Seclusion (viveka), Letting Go (vossagga) and Emptiness (suññatā). Although encylopaedia articles are not usually the place to present new research, Anālayo nevertheless also manages to bring in some illuminating new interpretations of difficult issues, of which I will mention five.

First, in his discussion of Craving (taṇhā), he suggests a very interesting way of understanding the concept of vibhava-taṇhā, ‘craving for annihilation’. This concept is usually given as the third kind of taṇhā, after kāma-taṇhā, ‘craving for sensual pleasure’ and bhava-taṇhā, ‘craving for existence’ (i.e. craving to continue as the same person). Anālayo interprets vibhava-taṇhā as not only the desire to commit suicide but also, and much more importantly, ‘the aspiration for leaving behind the sense of selfhood through a mystic merger with an ultimate reality’ (p.16). Needless to say, this suggestion is accompanied by a discussion of various discourses and an admission of conjecture. Nevertheless, it is a proposal quite in line with the Buddha’s rejection of the kind of mysticism found in the Upaniṣads. Second, in his discussion of Right View (sammādiṭṭhi), Anālayo very neatly solves the old problem of how to reconcile the admonition found in some of the suttas of the Aṭṭhakavagga of the Sutta-Nipāta, to let go of all views, with the admonition found more widely in the Nikāyas, to have right view. He writes: ‘right view as the vision gained through deep insight is what ‘sees through’ any view’ (p.102), and hence the person of transcendent right view or perfect vision is also someone who has let go of views. Third, in his discussion of Tranquillity and Insight, samatha & vipassanā, Anālayo explains how, in the Pāli discourses, these two qualities are actually ‘two central qualities that are to be developed in conjunction with any type of meditation practice’ (p.232). There is no question, therefore, at least in the early Buddhist view, of a successful insight practice without tranquillity. Anālayo develops this theme further in his discussion of Concentration (samādhi), in which he concludes that: ‘the so-called “dry insight” approach, which dispenses with the formal development of mental tranquillity up to the level of the first absorption, may not be capable of leading to fully liberation, but might suffice only for stream-entry’ (p.256).  Such a stand makes clear the importance for early Buddhism of jhāna, and draws a clear distinction between early Buddhist doctrine and the dry insight approach of some modern Theravādins. Fifthly, Anālayo’s judicious use of Chinese parallels is apparent in his discussion of Seclusion (viveka). While in the Pāli discourses the Buddha recommends silence and seclusion, in one discourse he censures some monks who had decided to keep silence together for their rains retreat. A consideration of a parallel preserved in Chinese, however, reveals that these monks had decided ‘not to criticize each other even in the case of a breach of conduct’ (p.263), and it was evidently for this reason that the Buddha had censured them. In this way, Anālayo sheds light on Pāli discourses that are unclear, through his knowledge of parallel passages in Chinese translation.

These were simply five points that particularly attracted this reviewer’s attention, and other readers will no doubt find different points of interest. Overall, my response to Anālayo’s Excursions was delight and pleasure in the appearance of a new standard for reference articles on early Buddhist concepts. In this sense, the present book is highly recommended. It is, however, frustrating in that it covers only a few topics. One can only hope that, despite the Encylopaedia of Buddhism now being complete, Bhikkhu Anālayo will continue to write articles like the ones gathered in this book, and that these articles are eventually gathered into a more comprehensive reference work on the important terms and concepts of the Pāli discourses.

Dhīvan is the editor of the Western Buddhist Review, and author of This Being, That Becomes: The Buddha’s Teaching on Conditionality, Windhorse, Cambridge, 2011.