On the reliability of Dīgha Nikāya translations, with reference to the Janavasabha Sutta

While the other three nikāyas have been graced with excellent translations by Ven Bodhi, when it comes to the Digha, we are not so lucky. The two main complete translations are the PTS translation by TW Rhys Davids, and the Wisdom edition by Maurice Walshe. (Note that on SC we use use more reliable translations where available, and the Rhys Davids translation as a backup. The Walshe translation is restricted by copyright; I use my own digital copy, with gratitude to those who ignored copyright and made it available.)

Both of these have their strong points. The Rhys Davids translation was a groundbreaking work, and many of his notes and introductions are quite brilliant. Groundbreaking and influential as he is when it comes to matters of Indian society and culture, he is less reliable when it comes to Dhamma and meditation, about which we have learned so much in the ensuing years. Walshe’s translation is a modern and very readable update. In many ways it captures the goal to which I aspire, of making a fluent and idiomatic text for readers who are not experts.

It would be easy to assume, as I used to, that, as a later translation, Walshe’s would be more accurate. However, there are plenty of cases where it perpetuates mistakes in Rhys Davids’ work, and even adds new mistakes where Rhys Davids got it right. Of course any translation, including mine, is subject to such errors. But the DN translations have many more of them than we find in Ven Bodhi’s work.

I haven’t done a detailed survey, but after noticing many such cases I thought it might be worthwhile to note a few instances, as a service for students wishing to understand the relevant issues.

It so happened that I was translating DN 18 Janavasabha when I had this idea, so I jotted down some examples as I encountered them. It’s not meant to be comprehensive, just a few cases that I noticed.

None of these cases are, in themselves, particularly important. That’s the point. Craftsmanship shows in details. If we want to understand the nature of the translations on which we rely for the big things, we can learn a lot by examining the care and precision with which the little things are handled.

antarahito yakkho
RD: invisible spirit
MW: a yakkha who had passed over

MW is doubly inaccurate. Antarahita means “disappeared, vanished”, typically by psychic powers, so RD is correct. Moreover, yakkha in the EBTs normally means “spirit”, and it is better to translate it as such to avoid conveying a misleading impression based on later conceptions of yakkhas.

Both: “one who bears the name”

Actually this just means name (borne by something or someone), as can be seen from MN 50.

Āsā ca pana me santiṭṭhati
RD: Now is there desire in me
MW: now the desire arises in me

The meaning is: “but I still have hope”, where santiṭṭhati has the sense of “remains”, not “arises”. The text makes clear later that these two things came about together. Here they are similar, but MW’s is a little less accurate.

RD: it was only the moment before
MW: since I had only just … it is no wonder that I thought

Here RD misunderstands the idiom, which means “it comes as no surprise”, and takes it as “just now, immediately before”. MW manages to be both more accurate—in that he correctly gets the sense of the idiom as “no wonder”—and less accurate, in that he also retains the sense of “immediately”, and moreover applies the idiom to the wrong part of the phrase. The correct sense is:

But it comes as no surprise that I have heard and learned the fate of the Magadhan devotees directly from the great king Vessavaṇa.

Purimāni, bhante, divasāni purimatarāni
RD: In days gone by, lord, in days long long gone by
MW: in earlier days, long ago

While these give the impression of a mythic time span, in fact the idiom purimāni divasāni means “a few days ago”, and with the addition of purimatarāni, “more than a few days ago”. It’s referring to the start of the vassa, not to some bygone age.

RD: feast
MW: fast-day

Whether intentionally or not, they give exactly the opposite meaning, both of which are wrong. The uposatha is a day of religious observance, but it has nothing to do with eating as such (although some will take the eight precepts for the day). It should be translated “observance” or “sabbath”, or left untranslated.

Puratthimāya disāya dhataraṭṭho mahārājā pacchimābhimukho nisinno hoti deve purakkhatvā
RD: There was Dhataraṭṭha, king of the East, seated facing the west, presiding over his host
MW: There was the Great King Dhatarattha from the east at the head of his followers, facing west

Both these are misleading. The text doesn’t say that Dhatarattha is from the east, or king of the east, nor that he presides over the gods. It is strictly a description of his physical position in the assembly, modeled after similar descriptions of the Buddha with the Sangha, eg. in DN 2. The text doesn’t explicitly say that the gods he is in front of are his own, but it seems like a reasonable inference.

The Great King Dhataraṭṭha was seated to the east, facing west, in front of his gods.

Idaṃ sutvā devā tāvatiṃsā ekaggā samāpajjiṃsu
RD: And when they had heard this, the gods of the Three-and-Thirty were all together agreed
MW: Thus they were all agreed.

This is part of an obscure passage, where the 33 gods are anticipating Brahma’s imminent arrival. They agree to find out what the light is and go to it (or at least, that seems to be the sense of an obscure passage). However, when Brahma does appear, they end up staying in their seats. It seems to be an inversion (ironic?) of the five monks agreeing to not go out to meet the Buddha, then ending up doing so anyway.

However, what concerns us here is the account of heavenly governance. This sutta gives a number of interesting details, which are clearly meant as a model for good governance. The leaders of the gods have just been praised, not for their bully and bluster, but for their patience in listening to good advice and instruction.

Here, too, the procedure is meant as an example of good governance, but this is obscured in Rhys Davids, and almost lost in Walshe. The point is that initially, each god sitting in their own seat thought about going to greet the source of the light. Then the Four Great Kings had the same thought. It was only then, after the leaders had reflected and endorsed the will of the people, that the assembly of the gods achieved unity. The lesson is that Kings need to listen to their people, give voice to their aspirations, and lead them to unity based on that. Thus our phrase needs to have an active sense: it is the gods achieving unity (ekaggā samāpajjiṃsu).

Hearing (what the Four Great Kings said), the gods of the Thirty-Three agreed in unison:

Yo kho pana, bhante, brahmuno pakativaṇṇo anabhisambhavanīyo so devānaṃ tāvatiṃsānaṃ cakkhupathasmiṃ.
RD: For Brahmas usual appearance is not sufficiently materialized to impress the vision of the Thirty-Three Gods.
MW: his [Brahmā Sanankumāra’s] natural appearance is not such as to be perceptible to their eyes

It’s a slightly tricky idiom, although the sense is clear enough. Literally:

A Brahmā’s normal appearance is imperceptible in the visual field of the gods of the Thirty-Three.

Or more simply:

The gods of the Thirty-Three aren’t able to see a Brahmā’s normal appearance.

While RD’s translation is quirky, he correctly captures the detail that this sentence refers to all Brahmās, not just the Brahmā who happens to appear in this case. This nuance is lost in Walshe’s abbreviation. While abbreviating text is fine, and essential in many cases, it’s important to ensure that no details, no matter how trivial they may seem, are lost. The aim should be lossless compression, so that the full text can be reconstituted without error or ambiguity from the abbreviated version.

“yassadāni devassa pallaṅkaṃ icchissati brahmā sanaṅkumāro, tassa devassa pallaṅke nisīdissatī”ti.
RD: Of whichever god Brahma Sanaṅkumāra now desires anything, he will sit down on that god’s divan
MW: thinking he will sit down on the couch of that god from whom he wants something

Here Walshe appears to have paraphrased RD, without noticing that his translation was mistaken. It means:

“Now Brahma Sanaṅkumāra will sit on the couch of whatever god he wants.”

Imamatthaṃ, bhante, brahmā sanaṅkumāro bhāsittha; imamatthaṃ, bhante, brahmuno sanaṅkumārassa bhāsato aṭṭhaṅgasamannāgato saro hoti vissaṭṭho ca viññeyyo ca mañju ca savanīyo ca bindu ca avisārī ca gambhīro ca ninnādī ca.
RD: This was the matter of Brahma Sanaṅkumāra’s speech. And he spoke it with a voice of eightfold characteristics—in a voice that was fluent, intelligible, sweet, audible, continuous, distinct, deep, and resonant.
MW: Now to the matter of Brahmā Sanankumāra’s speech, and as for the manner of his speech, his voice had eight qualities: it was distinct, intelligible, pleasant, attractive, compact, concise, deep and resonant.

The forms of bhāsati are a little obscure here: bhāsittha is third singular aorist middle (type 4); and bhāsato is present participle genitive third singular. I’m not sure where either translator recognized the first of these. In any case, the repeated idiom, in past and present (he spoke … while he spoke …), clarifies the otherwise obscure sense of the repeated imamatthaṁ. The first phrase refers back to the previous verses, while the second connects that with the new idea of the sound of his voice. Rhys Davids’ translation gets this fairly well, while Walshe’s misses it entirely, and is not really coherent. I would render it:

That is the topic on which Brahmā Sanaṅkumāra spoke. And while speaking on that topic, his voice has eight qualities: it is clear, comprehensible, charming, audible, rounded, undistorted, deep, and resonant.

cattāro iddhipādā paññattā iddhipahutāya iddhivisavitāya iddhivikubbanatāya.
RD: the Four Ways to Iddhi for the development, thereof, for proficiency therein, for the elaboration thereof?
MW: the four roads to power, and how to develop, perfect and practise them

This is a particularly knotty idiom, found nowhere else. Both translations treat the phrases iddhipahutāya iddhivisavitāya iddhivikubbanatāya as more or less general terms for development. But then, why not use the stock terms? Surely such a unique phrase must have a special sense in this context.

Readings of iddhipahutāya iddhivisavitāya iddhivikubbanatāya are uncertain. Comm and Tika go to unusually elaborate lengths to explain. However, I don’t find their arguments compelling in all cases, and will not attempt to unravel them here.

Starting with what we know, from Thag 20.1 we know that vikubbana means transformation of one’s corporeal form. This specific psychic power is obviously an elaboration of the stock power mentioned in descriptions of the iddhis:

ekopi hutvā bahudhā hoti, bahudhāpi hutvā eko hoti

And this is, in fact, precisely what Brahma Sanankumara has just done, so it must have the same sense here.

This gives us a suggestion for the reading of the first term, for which we have readings of pahūta (either “abundant” or “capable”) and bahulīkata (“made much of”). In the light of this, it is tempting to emend this to bahudha, “multiplies”.

Visavita is harder to parse, and it is not clear what the root is. The closest we have to a hint is in the Culaniddesa, in a gloss of the term “hero”:

Bhagavā vīriyavāti vīro, pahūti vīro, visavīti vīro, alamattoti vīro, vigata­loma­haṃ­so­tipi vīro
The Buddha is “hero” because he is energetic; because he is capable; because he is (powerful?); because he is competent; because he is dauntless.

Obviously the fact that pahūti and visavi occur next to each other, in a context not too dissimilar from our main text, is significant. Here, pahūti would seem to require the sense “capable”. The commentary on this text explains visavi as “one who arouses energy with respect to others” (parasantāne vīriyuppādako). This harks back to an idea expressed in the Suttas, where one may arouse energy through considering the benefit for oneself, or for others. The core of the idea seems to have something to do with spreading or extending. However, I’m not really sure what, if anything, this has to do with the meaning of visavitā in DN 18.

Leaving this aside for now, I think visavitā should be related to the Sanskrit savitṛ, “the sun, a generator, producer”. The vi carries the sense of “out”, thus, “producer of diversity”, i.e. one who generates multiple forms. This relates the term, not only specifically to the generation of multiple bodies which Brahma had just demonstrated, but to solar imagery—the orb of the sun, streaming out its rays—which is wholly apt in this context. Brahma has only just been described as one whose presence is intimated by the appearance of light, in language that closely echoes the appearance of dawn as described elsewhere in the suttas.

While the text remains an obscure one, I am leaning towards the following:

“What do you think, gentlemen? Haven’t the four bases of psychic power been clearly described by the Blessed One, who knows and sees, the perfected one, the fully awakened Buddha, for the multiplication, generation, and transformation [of corporeal forms] through psychic powers?

If we reject the emendation of the first term to bahudha, then the sense should be “empowerment, enhancement”, or perhaps “amplification”.

If it is objected that this is not a proper aim for Buddhist practice, perhaps this is the point. It is the first of the psychic powers, and it seems such things were used as a hook for getting people interested in cool displays, before drawing them deeper. It seems to me that Brahma is doing exactly that here. He shows up at a meeting of some lesser gods, and wows them all with his cool superpowers. Then he shows them how they too can get such powers, before leading them on to deeper matters.

sammāsambuddhena tayo okāsādhigamā anubuddhā
RD: Three Avenues for arriving at Bliss manifested by the Exalted One
MW: three gateways to the bliss proclaimed by the Lord

Anubuddha means neither “manifested’ nor “ proclaimed” but “understood”.

So aparena samayena ariyadhammaṃ suṇāti, yoniso manasi karoti, dhammānudhammaṃ paṭipajjati.
RD: He on a certain occasion hears the Aryan Truth, studies it and acquires both the main and the subsidiary doctrines.
MW: At some time he hears the Ariyan Dhamma, he pays close attention and practises in conformity with it.

In this case, Walshe’s translation is a welcome improvement.

RD: twenty-four lacs
MW: twenty-four hundred

RD is correct, it’s 2,400,000.

evarūpā uḷārā visesādhigamā paññāyissantī
RD: such glorious avenues to distinction should be made known
MW: such glorious paths to the sublime should be made known

Again, Walshe is paraphrasing Rhy Davids, without correcting his mistake. There’s nothing about “paths” in the text. It is referring back to the main topic of the discourse, i.e. the Buddha describing the various attainments realized by the devotees.

that such achievements of high distinction should be made known!

yāva devamanussehi
RD: among men
MW: among mankind

They evidently follow the PTS reading yāvadeva manussehi, but this is a stock passage, and should be read (as per comm.) “among gods and humans”.

That’s all I have to say about the Rhys Davids and Walshe translations, but in researching this I came across a related issue, which I will note here.

Ven Bodhi in his translations of this idiom at AN 8.70, AN 5.196, and SN 12.65, as well as Ven Anandajoti in his translation of DN 16, treat devamanussehi here in a locative sense. Despite the agreement of such esteemed scholars, it seems to me the matter is not so clear-cut.

At SN 12.65, Ven Bodhi notes:

I follow Spk in its explanation of yāva devamanussehi suppakāsitaṁ. The point is that, despite the use of the instrumental form -ehi, the Dhamma is not proclaimed by devas and humans, but “throughout the region (inhabited) by devas and humans in the ten-thousandfold galaxy, within this extent it is well proclaimed, well taught, by the Tathāgata” (yāva dasasahassacakkavāḷe devamanussehi paricchedo atthi, etasmiṁ antare suppakāsitaṁ sudesitaṁ tathāgatena). It is possible -ehi here is a vestigial Eastern locative plural; see Geiger, Pāli Grammar, §80.3.

Note that the passage in Geiger that is cited here is a general discussion of Magadhisms, and, while it mentions the ehi/esu substitution, it does not mention this passage. Geiger, in fact, cites virupakkhehi, backing it up with the unambiguously locative Sanskritic form virūpakṣeṣu. However, the Sanskrit form of devamanussehi in SF 245 Mahāparinirvāna is the instrumental, devamanuṣyebhyaḥ. Admittedly, this is reconstructed. However, the phrase also occurs in the Sanskrit parallel to SN 12.65, where it is not reconstructed, and there too we find devamanuṣyebhyaḥ.

Magadhan forms are extremely rare in Pali, and there doesn’t seem to be any reason why one might occur here.

Also relevant is that the commentaries do not all support the locative reading.

DN 16, AN 8.70, SN 51.10: yattakā viññujātikā devā ceva manussā ca atthi sabbehi suṭṭhu pakāsitanti attho
the meaning is, wherever there are intelligent gods and humans, by all them it is well proclaimed.

It is possible that the commentary is echoing the Magadhan locative form here, but this is usually understood to be a sign of an early vestigal form, and I think it’s unlikely to be used by the commentary in its gloss.

However, the commentary to DN 29 assumes the locative sense:

DN 29: devalokato yāva manussalokā suppakāsitaṃ
from the heaven realms to the realms of men it is well proclaimed.

This is confirmed in the subcommentary:

Yāva devamanussehīti vā yāva devamanussehi yattakā devā manussā ca
yāva devamanussehi means “wherever there are gods and men”

Given that the commentaries appear to be ambiguous, and the Sanskrit does not support the locative sense, we should fall back to the overall sense of the passage.

Typically it occurs at the end of a description of the great extent of the teaching, especially after saying that the teaching had been spread to, and indeed proclaimed by, the fourfold assembly.

In DN 18, of course, the context is explicitly about how the Dhamma is proclaimed by the gods, not to them. Indeed, it seems to me that this passage was included in the sutta precisely to illustrate this point. DN 18, remember, is one of those suttas that was spun out of a short passage in DN 16 Mahaparinibbana. The overall thrust of the sutta is to place the words of the Buddha in Brahma’s mouth, and hence claim his authority.

Thus it seems to me that the contexts on the whole support the instrumental sense.

AN 5.196 may be an exception to this. The image is of a creeper reaching from the ground to the sky, which fits well with the explanation, “from the human realm to the heavens”. Still, the instrumental sense is not impossible.

DN 29 seems to be another context where the locative makes more sense. In fact, the whole passage is about the Buddha teaching the Dhamma, and, in contrast with DN 16 etc., how the disciples do not learn it well. In such a case it seems odd to introduce the teaching by gods and humans.

Well, there’s good points on both sides, I think, and no wonder the commentaries are a little ambiguous. I am inclined, however, to accept the locative sense based mainly on DN 29.

This does raise an interesting question regarding the Sanskritic version. Given that it is instrumental, it must be translated from a similar form. Now, if it had been translated from an earlier MIA dialect that used a similar form as locative, i.e. Magadhan, then this would have been recognized and it would have been put into locative. So if our reading is correct, it must have been translated from something which, like Pali, normally uses such forms as instrumental, but which preserved an earlier form. I’m not really sure about research into this process of translation, so can’t do anything other than raise this question.

well proclaimed wherever there are gods and humans.

Note: I changed the conclusion of this essay, so some comments may be outdated.


Just the tiny remark and suggestion to avoid “sabbath” since it is based on the Hebrew word “Sheva” which means “seven”, i.e. the seventh day and thus not the fortnight that I think is usually meant with uposatha.


Actually no, it’s the 14th and 15th, but also 8th day, so pretty much once a week. The uposatha tradition is really quite similar to the sabbath. Given that the Abrahamic sabbath stems from Babylon, it is not inconceivable that the two traditions have the same origin.

Also, according to Wikipedia, the Hebrew of sabbath is actually “day of rest”; Babylonian Sapattum; Sumerian sa-bat in the same sense. Not that etymology should be the basis for such translations, but this is not so far removed in sense from “dwell-close”, upa-avasatha.


It obviously just a side issue, but the wikipedia entry does not fully reflect the research - to be honest I wasn’t even aware of the ‘rest’-hypothesis. For etymology buffs here a detailed derivation by a Cambridge Prof for Middle Eastern Studies that reflects both theories and concludes

…it is etymologically not connected with the Hebrew root š‑b‑t ‘to rest’… The Sabbath is in essence the seventh day. The linguistic link with its seventh-ness is via the Babylonian sibūtu, which was contaminated (as we say in linguistics) with the unrelated šapattu. The connection with ‘rest’ is secondary, and essentially theological.

Sorry about the mistake re. uposatha. I had the upavasatha of the Satapatha-Brahmana in mind, where it is mentioned to be done at new and full moon (SB, SB, and just transferred it to the uposatha.


Interesting, thanks!

If sabbath is connected with “seven”, is it just a coincidence that it resembles Sanskrit sapta = Pali satta = “seven”?

My Proto-Indo-European etymological dictionary has the same question - without an answer :slight_smile:
Sorry, I’m not deep enough into linguistics to quickly find a reliable reference. Superficially they (and in other languages) seem connects through the “s/sh” first consonant and the “p/b” second consonant, but language moves in mysterious ways… Maybe there is a proper linguist who can answer this?

1 Like

Searching ‘pahutā’ I found this page with another instance of it:

It was translated as ‘arising’. I thought this was a good place to start as these terms are often stated in a sequence of development of the ability in question. So this term may mean something about the ability to give rise to iddi abilities , which is a skill.

I came across this spelling: iddhivisevitāya (syā.) here.

‘Sevana’/‘sevitabbo’ has the meaning to use, practice etc. So in our sequence, once given rise to, this term might mean to use the iddis skillfully -master it.

This seems to mean transforming. Contextually, the Brahma performs a transformation miracle when he took on a grosser form. He seems to be saying ‘you can do it too, only if you practice’.

These gods are in the sensual realm and are highly impressed by luxuries and power. They may need an incentive in terms familiar to them and be impressed by a being who have what they don’t have, even if it is to listen to the dhamma. Brahmas maybe similarly impressed if they cannot find someone with their divine eye, -a display of power.

‘bhavita’- use; ‘bahulikata’- proficient?

I’m in no way a Pali scholar so feel free to ignore!

with metta

[quote=“sujato, post:1, topic:5819”]
However, the commentary to DN 29 assumes the locative sense:

DN 29: devalokato yāva manussalokā suppakāsitaṃ
from the heaven realms to the realms of men it is well proclaimed.[/quote]

Though it’s possible for a noun following yāvā to be nominative, I think manussalokā is more likely to be ablative singular here. There is no indication in the sutta that more than a single cakkavāḷa is involved and the commentarial cosmology would assume one human realm per cakkavāḷa.

Yes, the plural is unexpected; but then, how would you parse the phrase as an ablative? I am not familiar with commentarial idioms, so any help is appreciated.

I would translate, “To the human world.”

According to the grammarians yāva as a preposition can govern either the accusative or the ablative case. In both spatial and temporal phrases, the difference seems to be merely morphological, not semantic. That is, the difference in case doesn’t result in any difference in meaning and nor (as far as I can see) is there any rhyme or reason as to why one case is being used rather than another.

Yāva jīvitakālassa pariyosānaṃ
Until life’s end.

Tādisāya iriyāpathasampattiyā samannāgatassa me ayyassa yāva imamhā kālā niddāyanaṃ nāma natthi.
“How extraordinary that my reverend master, one of such perfect deportment, should sleep until this time.”

1 Like

Okay, thanks, I wasn’t familiar with this usage. I’ll bear it in mind.

Hello Bhante!

Since Ven. Bodhi has translated at least three suttas in the DN (DN1, DN2, DN15) and maybe others, I was wondering what you thought of those translations as they compare to Walshe and Rhys Davids.


It should go without saying, they’re excellent! Obviously his translation style has evolved since then, but his early translations are still very reliable and readable.

Also Anandajoti’s translation on DN 16, and Kelly et al’s translation of DN 31 are excellent. All these are available on SC.

1 Like