Some translation issues in the Nivāpa Sutta

MN 25 has the Buddha telling how to escape from the clutches of Māra by means of the absorptions. The teaching is illustrated by an extensive parable of bait being laid down to trap animals, and the successively more wily ways they use to escape it. This essay is not so much an argument about the meaning of the sutta or particular interpretations, but simply a series of notes on translation issues that I encountered. The message of the sutta is simple, but it raises a number of surprisingly difficult translation issues.

The feeder feeds food

It’s common in Pali to find phrases with duplicated terms from the same root. We sometimes find comparable usages in English, eg. “to gather together”, or “she planted a plant”. But it’s much more common in Pali. It is somewhat rare, however, to find three related terms in close conjunction. But here we have:

nevāpiko nivāpaṃ nivapati
the feeder feeds food

While in translation nivāpa is often rendered as “bait”, it is actually just a general term for “harvest, crop” and hence “fodder” for animals (or, rarely, humans, see Ja 540#50) generally. It’s often used in a positive or neutral context (see, eg., Ja 385#2), but here, since it is being used to trap animals, it serves as bait. Note, however, that the sense of “bait” is not attested in Sanskrit, so it must be derived specifically from how the Buddha used it.

Though I create the echo above by rendering nivapati as “feed” it actually means to “sow” or “strew”. Thus the above phrase is rendered by Ven Bodhi as:

a deer-trapper lays down bait

Compare Horner’s much more literal translation:

the sower sows the crop

So the phrase, in and of itself, doesn’t specify that it’s a trap. However, it said that the crop is sown “for the animals”, thus it is clear that the intention is, in fact to set a trap.

If we wonder why the same word is used for sowing a crop and setting bait, presumably it is because the bait consists of grain. This comes straight from the root meaning. Scattering grain for crops or for bait is essentially the same act.

A further difficulty arises because nivāpa means both “grain” and “place where grain is laid out”, in which case the commentary glosses it as nivāpaṭṭhāna. In such contexts it is often rendered as “feeding ground”, as in the well known “squirrel’s feeding ground” where several discourses are set. It’s not always easy to disambiguate these senses, and translators and commentary sometimes disagree.

Moreover, later in the sutta we hear of “Māra’s nivāpa”, translated as “Māra’s bait”. However, it appears in a series of terms (MN 106) where it clearly has the sense of a place:

māradheyyaṃ, mārassesa visayo, mārassesa nivāpo, mārassesa gocaro
Māra’s domain, Māra’s range, Māra’s feeding ground, Māra’s territory

Kinds of animals

The subject of the parable is said to be migajātā. Most translators render this as “deer” or “a herd of deer”. However, the normal word for a herd of deer is migasaṅgha. The suffix -jāta signifies rather “kind, species”.

Now, the word miga is ambiguous. It may be used either of deer-like creatures, or of wild animals in general. However, the term migajātā only occurs in one other place in the canon, at Ja 172, where it unambiguously refers to a jackal as a “species of animal”.

Obviously the term includes deer, and it’s probably a convenient translation choice just to use deer here, but the text as it stands is probably not specifying deer. So far as I aware, only Sister Uppalavanna’s translation accurately specifies “wild animals” here.

Amid the bait

The trapper’s wish to catch the animals is expressed thus:

imaṃ me nivāpaṃ nivuttaṃ migajātā anupakhajja mucchitā bhojanāni bhuñjissanti, anupakhajja mucchitā bhojanāni bhuñjamānā madaṃ āpajjissanti, mattā samānā pamādaṃ āpajjissanti, pamattā samānā yathākāmakaraṇīyā bhavissanti imasmiṃ nivāpe

There are a couple of issues here. The term anupakhajja is a little unusual, and is rendered by Ven Bodhi as “going right in amongst”, and by Horner and Uppalavanna as “encroaching”. The term is found a few times in the Vinaya, where it always has the specific meaning of “encroaching” or “intruding”, i.e. going into a place you’re not supposed to be. It’s probably best to make sure that sense is retained here.

The more difficult question is the final phrase, imasmiṁ nivape. Literally, it just means “in this feeding ground” or “among the bait”. And that’s how it’s rendered by Chalmers, Horner, and Uppalavanna. This reading is supported by the commentary, imasmiṃ nivāpaṭṭhāne.

However, the phrase is perhaps better treated as a locative absolute, where it may have a variety of oblique senses. The most obvious would be a temporal locative absolute: “so long as they were in that feeding ground”. But Nynamoli/Bodhi take it in a causal sense: “on account of that bait”.

I don’t know that it’s possible to resolve this unequivocally. Personally I feel the idiom reads best as a temporal locative absolute. However, in the third example, the trappers evidently succeed by seeing how the animals get in, and following them back to their lair. In other words, the danger does not apply only as long as they remain inside. In this case, it really only makes sense to read the locative absolute in a causal sense.

Unnatural spirits

When the animals succeed in getting in and out without being seen, the trappers are flummoxed, and say:

saṭhāssunāmime tatiyā migajātā ketabino, iddhimantāssunāmime tatiyā migajātā parajanā

The compounds break down:


The phrase is rendered by Ven Bodhi:

These deer of the third herd are as cunning and crafty as wizards and sorcerors

Note that the ‘as’ here is supplied from the commentary (viya); the text itself is stronger. And the repeated indeclinables assu and nāma show a strong, emphatic exclamation: “Crikey!”, “Wow!” Remember, this is an emotional reaction by a group of hunters, not a philosophical statement.

The first three main terms are straightforward. Saṭha means “devious, sneaky”. While ketubi is somewhat obscure, it always occurs with saṭha and māyāvī, so it must mean something similar, probably “cheat”. Iddhimanta is also clear: it means “possessing psychic powers”.

The last term, parajana, however, is apparently unique in this sense. It breaks down to “other-man”, and sometimes it is used in that simple sense (Mil 6.3.1). Here it clearly refers to something rather more astonishing. It’s tempting to render it literally as “alien”:

Strewth, those deer must be aliens!

The term does, however, turn up in MN 31 as the name of a yakkha, and the commentary says that’s what it means here. Compare with the common term amanussa, which means “non-human, spirit”.

I’m currently rendering the phrase as:

Wow, this third lot of animals is so sneaky and devious, they must be some kind of unnatural spirits with psychic powers!

Nets and hideouts

The next phrase offers some more puzzles. The trappers propose to foil the animals by setting up daṇḍavākara all around. Daṇḍa means “stick”, but vākara is more difficult. This is not helped by the fact that the same word is sometimes spelled as vāgura. It occurs only a few times in the canon, and evidently means some form of trap, perhaps a net.

It seems the etymology is traceable to English “wick, wicker”. Whether for this reason I cannot say, but Nyanamoli/Bodhi render the phrase as “wicker hurdles”, which sounds odd to my ears. Chalmers had “stake-nets”, Horner “stakes and snares”, while Uppalavanna just has “sticks”.

In this case, I think the oldest translation, Chalmers, is probably the best. I think what is intended is a net strung out on sticks or stakes. This either keeps the animals out, or forces the animals to leap over it, so they can be spotted and followed back to their lair.


The lair itself has another translation problem. The word gāhaṁ is unusual in this context, and, once more, is rendered very differently. It’s used when the deer flee back to their lair. Chalmers has where they “disappeared” to; Horner has where they “take” the bait; Uppalavanna has where the trappers “got at’ them; while Nyanamoli/Bodhi have where they go “to hide”.

Cone connects the word with gādha2 “(mouse) hole, hiding place”, which supports Chalmers and Bodhi/Nyanamoli. So in this case it seems the translations have cycled through every possible meaning, until, like the land-seeking bird, they returned to where they set out from!

And a ghost sentence

The Mahāsaṅgīti edition has a ghost sentence, which appears in the application of the parable to the third kind of ascetics and brahmins.

Te sabbaso nivāpabhojanā lokāmisā paṭiviramiṃsu, bhayabhogā paṭiviratā araññāyatanāni ajjhogāhetvā vihareyyāmāti.

This combines the aorist of the subsequent sentence with the optative of the previous sentence. It occurs in the VRI edition on which the Mahāsaṅgīti was based, and appears to be an uncorrected error. It doesn’t appear, even as a variant, in the PTS edition. It’s not unusual to find occasional errors, but an entire extra sentence is unusual.


Is this your latest for jhāna?

Yes, it is.


That’s not cool. What’s cool is how you managed to game Discourse’s five-character minimum for posts.

I had no idea. Just naturally cool I guess …


It must be a burden.

We should have more monastics with sīti in their name, like sīti-bhūto, “the cool one”.


While I’ve got you here, do you have any thoughts about door knockers? There’s the idiom avāpuraṇaṁ ādāya, which is usually translated, “taking a key”, spoken as part of making an announcement. I’ve never seen an explanation of why you’d use a “key” to make an announcement, or indeed, any reason to think that anyone, especially monks, would have a sophisticated thing like a lock and key.

Another idiom is where someone knocks on the door of a hut, where we find, aggaḷaṃ ākoṭesi. I wondered why they were said to knock specifically on the latch (aggaḷa), but then it occurred to me that they were probably knocking with the latch. If it was a sizeable bolt-type thing, it seems to have doubled service as a latch and a door-knocker.

In Ss2, we have (your translation):

avāpuraṇaṃ ādāya ghaṭikaṃ ugghāṭetvā kavāṭaṃ paṇāmetvā vihāraṃ pāvisi.
Udāyī took the key, unlatched the bolt, opened the door, and entered the dwelling.

I wonder whether avāpuraṇa and ghaṭika are parts of the aggaḷa? Or are they just different things? Or different terms for the same thing?

In any case, combining the two cases suggests that, just as the aggaḷa was used as a knocker, the avāpuraṇa was used in the same way, except it was obviously detached. Maybe it was just a bolt that was knocked against a block of wood or bell of some sort.

1 Like

Yes, it is a bit curious. The commentary seems to consistently gloss it with kuñcika, which also seems to mean key. But we are at the most talking about some sort of proto-key, and for this reason this translation may be misleading regardless.

This does merit more research, but unfortunately I am in Bangkok right now, and I don’t really have the required tools.


What’s up in Bangkok?

Doing a retreat.

Lovely. Well, have fun.

I seem to recall Ven Anālayo talking about this idea of a ‘proto-key’ somewhere. But I’m afraid I can’t remember if it was in a lecture or a paper (so I’m not much help), but if i’m remembering correctly, he had tracked down some research as to what this was probably about.

Thanks, Linda. That was actually very helpful; I have been able to track down the reference. It’s in his Comparative Study of the Majjhima: footnote 148 to the Bakkula Sutta. The footnote says the following:

MN 124 at MN III 127,24 adds that he had made a tour of the dwelling places of the other monks to make this announcement. Regarding the reference to his taking a “door-opener”, apāpuraṇa ādāya (Be-MN III 168,16: _apāpuraṇa_), in order to announce his impending final Nirvāṇa to the other monks, helpful information can be found in von Hinüber 1992: 14-18, who explains that this expression probably refers to a simple type of lock, found at the inside of a door, whose bolt can be closed from the outside with the help of a rope that is connected to this bolt and then put through a small hole in the door so as to be accessible from the outside. The same rope does, however, not allow opening the door from the outside, for which an apāpuraṇa is required, a simple type of “key” that allows opening any lock of such a type. Hence, with the help of an apāpuraṇa any door with this kind of lock in a monastery could be opened in order to enter and convey some message to the monk(s) staying inside.

So Ven. Anālayo renders it as “door-opener”. But this seems a bit imprecise to me. The point is that the device functions as a kind of key; it is not merely a device to open a door, such as a door-handle. Moreover, if von Hinüber is right, the avāpuraṇa seems to have been a universal kind of key, something like a “universal bolt-opener.”

That is helpful. If correct it explains how it functioned in announcements.

@Brahmali @sujato

Great, I’m gald you were able to find the reference. Interestingly this description fits the type of latch we have on a gate outside. Amazing it’s the same simple design.

Really? Can you send us a photo?

Sorry, not sure how to reduce the size of the photos. This lock is a little more sophisticated in that it has a little chain instead of a rope and the bolt inside is metal not wood. Also, with this one, it is possible to open the door from the outside by pulling on the chain which lifts the latch inside. That is, unless someone has carelessly or roughly closed the gate previously which causes the chain to get tangled in the bolt on the inside thus preventing opening the door next time you need to get in from the otuside. Many times I’ve had to get a stool and climb up to reach over the top of the gate and untanlge the little chain so I could get inside!


Okay, cool, thanks. But this is different, in that the chain is there so that you can open from the outside, whereas according to the description, a rope is used to close it from the outside, whereas a “key” is needed to open it. I’m still not entirely sure I can visualize exactly what’s happening.

The problem is that if you can pull a string to close the latch from the outside, why can’t you also open it? All the designs I’ve seen allow this. I can only guess that the string pulled the bolt shut horizontally, rather than lifting it up. Once the bolt was pulled shut, it couldn’t be pushed open by the same string. For that, a key is required.

But I can’t find any design like this, and I wonder where Von Hinuber got the idea of a rope from; I can’t see it in the Pali texts. But this is all getting too complicated, i will open up a new thread.