On the simile of the raft, and letting go of what is not the teachings

MN 22 is a fascinating text for many reasons. It tells the story of the truly monumental smack-down the Buddha delivers to Ariṭṭha the vulture trapper, who misrepresented the teachings to argue that having sex was no obstacle to spiritual practice.

Perhaps the most widely known passage is the famous simile of the raft. Here, for your edification and enjoyment, is my translation:

“Suppose there was a person travelling along the road. They’d see a large deluge, whose near shore was dubious and perilous, while the far shore was a sanctuary free of peril. But there was no ferryboat or bridge for crossing over. They’d think:

‘Why don’t I gather grass, sticks, branches, and leaves and make a raft? Riding on the raft, and paddling with my hands and feet, I can safely reach the far shore.’

And so they’d do exactly that. And when they’d crossed over to the far shore, they’d think:

‘This raft was very helpful to me. Riding on the raft, and paddling with my hands and feet, I have safely crossed over to the far shore. Why don’t I hoist it on my head or pick it up on my shoulder and go wherever I want?’

What do you think, mendicants? Would that person be doing what should be done with that raft?”

“No, sir.”

“And what, mendicants, should that person do with the raft? When they’d crossed over they should think:

‘This raft was very helpful to me. … Why don’t I beach it on dry land or set it adrift on the water and go wherever I want?’

That’s what that person should do with the raft.

In the same way, I have taught how the teaching is similar to a raft: it’s for crossing over, not for holding on.

By understanding the simile of the raft, you will even give up the teachings, let alone what is against the teachings.

Most of this translation is straightforward, but the last sentence, which advocates giving up even “dhammas”, is not easy to make out. The problem is, of course, that dhamma has so many meanings. Taken in isolation, it’s impossible to say what exactly is meant here. Various translators approach this is different ways.

  • Ven Bodhi, citing the commentary, has “states”.
  • Ven Thanissaro avoids the issue by keeping “Dhammas”.
  • Chalmers has good and bad “things”.
  • Horner, referring to the commentary, has right and wrong “mental objects”.
  • Nyanamoli’s draft translation had “true ideas” and “untrue ideas”. He was endeavoring to use “true idea” everywhere for dhamma. The fact that here it is in lower case suggests he understood it in the sense of “mental phenomena or qualities” rather than “teachings”.

Thus the bulk of the translators have taken dhammā here in the sense of “mental qualities, mental phenomena”. This is following the commentary, which, however, defines it rather more precisely as regarding attachment to advanced meditation states of samatha and vipassanā. This is contrasted with the bad example of Ariṭṭha, who was attached to adhamma in the sense of vulgar practices, i.e. sex.

The commentary cites two passages as examples of this. The letting go of states of samatha is said to be taught in MN 66; that’s not quite what the sutta is getting at, but I’ll leave that for you to check out.

The letting go of vipassanā is found in MN 38. This is especially interesting, as it is apparently the only other place in the suttas where the simile of the raft is mentioned directly. Let’s detour to have a look at that first.

In MN 38, the Buddha interrogates the mendicants in the context of dependent origination, asking whether they have doubts and whether they have seen with right wisdom. These terms indicate that the mendicants are, at least, stream-enterers, so the commentary is incorrect in describing this as vipassanā. That’s not a great start. But even more cogently, the thing that the mendicants are said to not attach to is not “states of vipassanā meditation”, but diṭṭhi, view.

This aligns with the original context of MN 22, where the whole sutta is about views, and the bad consequences of acting out of wrong views.

There is, of course, somewhat of an ambiguity here. Normally, a “view” is merely a belief or an opinion, which may or may not be well grounded. But the view of a stream-enterer is the right view of the noble eightfold path, and it arises because of directly seeing the truth for what it is. That doesn’t mean it’s not a view. It means it’s not just a view: it’s a view that’s grounded on experience of reality.

The suttas has an ambivalence around this, as sometimes views are said to be developed, other times abandoned; but if we understand the context it’s not too confusing. As a “view”, it is essentially an opinion formed on the basis of the teaching. Luckily for us, the Buddha’s teaching happens to be correct, so when we truly see reality, the view is confirmed. Thus not only does “view” skirt the boundary between “opinion” and “vision”, dhamma skirts the boundary between “teaching” and “true principle, how the world really is”. (It was, in fact, to capture this that Nyanamoli adopted “true idea”.)

In any case, the take-home lesson is this: MN 38, when referring to the simile of the raft, specifically states it is for letting go of “views”, not meditation states.

This is supported by the use of dhamma and adhamma throughout the suttas. Given how variably dhamma is used, it seems most translators have been content to accept that any of these meanings could apply here. However, a search of the canon reveals that the pair of dhamma and adhamma is almost always used in the sense of “teachings” and “what is not the teachings” (or “what is against the teachings” or “wrong teachings”). (AN 10.37, AN 4.22, AN 2.104, etc.)

(Apart from this, the only significant context we find dhamma/adhamma is in phrases using adhammacariya, visamacariya, “unrighteous conduct, unethical conduct”, which is obviously not the case here.)

Of course, it’s always possible that this context is simply an unusual usage. But there’s no reason why it should be, as the sense of “teachings” makes perfect sense.

It is true that dhamma in the sense of “teaching” is usually singular, and I guess this may have contributed to the confusion here. However, this is by no means always the case. Indeed, earlier in this very sutta, dhammā in plural is repeatedly used in the sense of “teachings”, eg:

Take a foolish person who learns the teaching—statements, songs, discussions, verses, inspired sayings, legends, stories of past lives, amazing stories, and analyses. But they don’t examine the meaning of those teachings with wisdom, and so don’t come to a reflective acceptance of them.

Given that the sense is quite straightforward, the question must arise as to why did the commentaries go to such lengths to insert an unrelated meaning? In understanding exegesis, it is helpful to understand the perspective of the exegete.

The commentaries stem from a tradition that had apparently moved far from the practice of meditation in favor of preservation of the texts. Perhaps it was too controversial for them to say that the teachings must ultimately be let go of, and better to point out the fallacies implicit in meditation practice.

There is always a tension between doctrine and experience, and the guardians of doctrinal orthodoxy always nurture a certain suspicion of meditative experience. This is not unreasonable: the plain fact is that meditators frequently misunderstand or misrepresent their experience, and a degree of skepticism, tempering reports of experience with the understanding of the text, is essential.

While this is understandable, it seems that here the tendency went too far, deflecting attention from the limitations of textual knowledge to the limits of meditation experience. As so often, the commentarial views are embodied in the translations, and these always need to be carefully and critically examined. Hopefully this is one small step to rectifying this.


I am wondering if it was just too much for earlier translators to let go of the idea that the teachings (as Dhammas) were to be let go of, thus the stretch to use other terms, like “things,” or “states.” But it seems that the Buddha’s intent was to focus on the raft ( what separates man from loss vs. survival in a flood) as being equivalent to the lifesaving quality of the Dhamma. Once the Far Shore is reached, even this Dhamma, like the treasured lifesaving raft, must be let go of. It is only a vehicle, and not to be carried about on the Far Shore (eg Parayana Vagga themes of the Sutta Nipata). Once liberated, it’s just not needed anymore. This section segues nicely, as well, from the Simile of the Snake. I’m relating essentially what Bhante has taught us here, but it’s interesting to see how translations can be influenced this way.

“While people are still binding together a raft,
Intelligent people have crossed over.” Ud. 8.6

Cool essay. I put it on Facebook, so others can see how MN22 can be read with the implications that the Buddha intended.


This is a particularly important point to make with regard to this sutta, as well as to (e.g.) those in the Aṭṭhakavagga. I am currently reading through Gil Fronsdal’s new translation of the latter. In his intro he says that the text establishes the Buddha as teaching ‘no-views’, but later on he says that the text cannot actually be taken that way. For the Aṭṭhakavagga is replete with positive claims and castigations. As in MN 22, Ariṭṭha was one of wrong view. Indeed the Buddha says there is a right view contradicting him in 22.9, “Bhikkhus, that one can engage in sensual pleasures without sensual desires … is impossible.” This is quite a modally strong claim or “view”.

I think though the exegetical problem is how to skirt this boundary, as you put it. Personally I rather like Paul Fuller’s take, which is that abandoning the raft does not mean literally abandoning those views, but rather abandoning the affective attitude of clinging or identification towards those views. Thus “right view” is essentially affective rather than cognitive. (Or perhaps better, it is essentially affective as well as cognitive).


I haven’t read his translation, but this is unfortunate. In my introduction to the Sutta Nipata I said this:

I would add a word of caution against over-interpreting the text. Perhaps the most striking themes, sometimes taken as emblematic of the collection as a whole, are the first and last in the above list: the virtues of renunciation, and the dangers of disputatious views. If one assumes that the Snp is a uniquely archaic text, it is tempting to see these positions as more authentic to the Buddha’s orginal teachings than what is found in the bulk of the prose texts. I don’t think this is the case. The virtue of renunciation and a simple wandering life is mentioned in many places in the prose Suttas, as is the danger of getting involved in disputations based on theoretical views. Perhaps they are emphasized more in the Snp, but this is hardly a sign of any substantially different doctrine. It is, in my view, a mistake to develop a theory of Early Buddhism based on a few verses.

I got to know Paul while he was teaching at Sydney U, and we discussed this point. I agree, I think he has a good perspective on it. Do you have a link for it?

While we’re on the subject, I have promoted this before, but let me give one more plug for Paul’s blog. He’s one of the few modern Buddhist academics who directly engages through writing blog articles on interesting things happening in the Buddhist world.



No link I know of unfortunately, this is from his book, fruit of his PhD dissertation, The Notion of Ditthi in Theravāda Buddhism which deals with the topic in some detail. (Pardon inconsistent dīacrîtícals, I’m writing on an iPad).

Agreed with your take on the Sutta Nipāta. I think it should be said though that it is to some extent an empirical question whether the doctrines in any text reflect those in another. I am persuaded that there is nothing in the Atthakavagga that is different in anything other than tone or emphasis than the doctrines found in the bulk of the Nikāyas. Some interpretations therefore at times strike me as anomaly hunting.

I do think however that there may be lessons about early Buddhism we can glean from the Atthakavagga about how those doctrines may have developed from less to more mature forms, or even to a shift in the Buddha’s own sense of the value of dispute and debate. E.g., he disdains it in the Atthakavagga but engages in it frequently in the other Nikāyas. Surely this is a development worth noting. It makes me wonder if he grew tired of it as a young man but then re-engaged with it as a mature leader.

1 Like

Oh and thank you for the kind reply! Nice to meet you, sujato. I’ve enjoyed reading your material.

The classic interpretation would be that he’s responding to different situations. In this case, it seems like a sensible reading, so we’d need a good argument that something else is going on.

Maybe so, though the emphases seem to have shifted.

Again, I wouldn’t want to make too much of this, but it might bear thought.

If there is a historical shift, it would make sense in terms of the history of the Sangha.

In the early days, it seems likely that much of the Sangha was comprised of converts from other ascetic movements (the five monks, the thousand kassapas, Sariputta and Moggallana, and so on.) They may well have come from an educational background that emphasized such debates. This would mean that they already had philosophical training and were able to grasp the principles of the Buddha’s teaching readily, but they may have required special advice to get them to stop worrying about “winning”, especially in debates with their former colleagues of other religions.

Later converts came from a more general population base, so lacked such philosophical backgrounds, and needed more schooling in the substance of “right view”. However, as they did not have an extensive background in non-Buddhist religious practice, the whole idea of winning debates was not such a big deal.

This would explain a shift in emphasis during the Buddha’s life time, if indeed there is such a thing.


Very good, yes.

It is also I think possible that in order to get some of those first followers the Buddha had to engage (or was expected to engage) in formal debates and either disliked them from the start or came to dislike them over time. (Cf. the Pasūra Sutta’s talk of judged debates, Snp. 827).

Once the sangha was established there was no reason to do these sorts of debates. This might have changed his emphases.

I don’t see any evidence of this. The Buddha’s preferred method seem to have been to wait until people were receptive before engaging. He did debate with people, but not in a contest setting, only as a genuine discussion.


Yes, agreed. I only raise it as speculation in this context.

I’ve been digging on what is the meaning of sammā diṭṭhi?

It seems there are quite a few definitions in the suttas.

AN10.211: Right View is about kamma and rebirth

‘There is what is given, sacrificed, and offered; there is fruit and result of good and bad actions; there is this world and the other world; there is mother and father; there are beings spontaneously reborn; there are in the world ascetics and brahmins of right conduct and right practice who, having realized this world and the other world for themselves by direct knowledge, make them known to others.’

MN9: it is defined as knowing basic Buddhist morality, the 3 unwholesome roots, and the 3 wholesome roots as counterpart; then the Four Noble Truths and various other dhammas fitting the 4NT schema: a concept, it’s origin, it’s cessation, way leading to it’s cessation.

SN45.8: I like how straightforward and succinct this sutta is. Sammā diṭṭhi is knowing the 4NT’s. (sammā sati is the 4 satipaṭṭhānas, sammā samādhi is the 4 jhānas, etc)

AN10.96, snp1.8: Sounds like one should make no fixed views.

Is this the book you mention?

The Notion of Ditthi in Theravada Buddhism_Fuller.pdf (1.8 MB)


I, too, have occasionally come across passages where dhammā (plural) needed to be read as “teachings” or contexts in which “teachings” made better sense. I recall being stuck at first, precisely because of the distinction between the plural and singular usage made by acclaimed scholars. But at one point some years ago I took the leap.

It’s a curious dilemma. On the one hand you have to rely on previous translators, without which translation work on the scale you are doing now would be impossible and on a smaller scale much more time consuming. On the other hand the very fact that you rely on them means you end up repeating some of their mistakes and/or biases (as well as adding a few new one’s, of course). I am not sure if there is any way out of this, except awareness of the problem. The more you translate, the more you realise the erudition and the scale of the scholarship of previous translators. But you also recognise the limitation inherent in all such ventures. A kind of oxymoronic mixture of awe and irreverence may perhaps take us the furthest.


Is that a thing? I had a vague idea that the sense of “teaching” was usually singular, but can’t recall any actual argument about it.

1 Like

Ven. Bodhi makes this distinction in the introduction to his Majjhima translation:

In revising Ven. Ñāṇamoli’s translation I have retained the Pali word Dhamma only when it refers to the Buddha’s teaching, or in several cases to a rival teaching with which the Buddha’s is contrasted (as at MN 11.13 and MN 104.2). In its other uses the context has been allowed to decide the rendering. Thus when dhamma occurs in the plural as a general ontological reference term it has been rendered “things” (as at MN 1.2 and MN 2.5). When it acquires a more technical nuance, in the sense either of the phenomena of existence or of mental constituents, it has been rendered “states” (as at MN 64.9 and MN 111.4).

To be fair to Ven. Bodhi, I should add that his position seems more nuanced in his introduction to “The Connected Discourses”:

Used in the plural, dhammā can also mean teachings, and so I render it at III 225,9 foll., though the exact sense there is ambiguous and the word might also mean the things that are taught rather than the teachings about them.

1 Like

Okay, but this more of a general guide than a detailed argument. Checking his translation of MN 22, he also uses “those teachings” where appropriate.

Yes, that should be the book, with the caveat that I haven’t clicked the link.


By the nature of the teachings, this is obviously not true. I think one should always err on the side of caution when imagining a one-off-isolated-phrase sums up the entire Buddha-Dhamma.

Personally, I have always viewed the ‘raft simile’ as relating to the ‘snake simile’, i.e., do not use the good Dhamma (Teachings) for the purpose of arguments but only for crossing-over .

Have I missed something here?

[quote=“sujato, post:1, topic:3873”]
Idha, bhikkhave ekacce moghapurisā dhammaṃ pariyāpuṇanti—suttaṃ, geyyaṃ, veyyākaraṇaṃ, gāthaṃ, udānaṃ, itivuttakaṃ, jātakaṃ, abbhutadhammaṃ, vedallaṃ.

Take a foolish person who learns the teaching—statements, songs, discussions, verses, inspired sayings, legends, stories of past lives, amazing stories, and analyses.[/quote]

Does the above indicate anything about the authenticity of MN 22? The above gives the impression the systematic structure of the teachings was already established when the Buddha was alive, with suttaṃ, udānaṃ, itivuttakaṃ, etc . Also, jātaka (birth stories) do not seem prevalent in suttas, being only a handful (such as MN 123 that contradicts MN 64 about self-view in a new born child), giving the impression birth stories (jātaka) were later additions.

1 Like