A better Pali Dictionary

Tags: #<Tag:0x00007f7893e26978> #<Tag:0x00007f7893e26770>


Stephen’s interpretation here is based on later texts, and doesn’t apply to the Suttas. Nimitta in the EBTs doesn’t mean “perceptual form”. In the context of meditation it means, like I said, “cause”. Even a word like “object”, which Stephen is used to, based on later Buddhism, has no place in the EBTs.

I would also be cautious about claiming that abstract terms become concretized. While this happens occasionally, it is far more common for concrete terms to assume abstract meanings. In fact, all abstract terms originate from concrete meanings: where else?

This raises a more general question, which is this: if we are to do a “doctrinal” dictionary of early Buddhism, then we need to have a broad knowledge base of understanding what that is.

And I am far from confident that this is the case. I have been talking about these things for years now—I discussed the meaning of nimitta in my first book, A Swift Pair of Messengers, written 15 years ago—but there are not many others.

If we were to crowdsource it, how would we ensure even a basic level of confidence that we are discussing meanings as found in the EBTs? I guess we could require references and argumentation for any citation.

How to deal with the commentaries, though? For many, probably most, Theravadins these are effectively canonical.

But even if we reject the authority of the commentaries, they remain useful, even indispensable, for countless terms. And in many cases there is no objection to using them, it is only in cases where they impose their own doctrines that it becomes problematic.

But what are those cases? To understand this, you have to not only have a good knowledge of the EBTs, but also of text-critical methods, and an understanding of the commentaries themselves, so you can see where their ideas intrude. When I have been reading, for example, the footnotes to Ven Bodhi’s Numerical Discourses, I immediately recognize where the commentarial explanations occur, and which ones should be rejected out of hand. But it has taken many years to have the familiarity with the material to do this. And many years to do the unlearning of habitual phrasings and words: I still trip over myself, falling into bad habits!

I am more and more of the opinion that including such commentarial explanations is, on the whole, a bad thing, even if you make it clear that that is what you are doing. It makes it actively harder to understand the text, and introduces a whole range of concepts and terms that are just unnecessary.

Anyway, i don’t want to get too sidetracked, just to say that the notion of having detailed doctrinal explanations of early Buddhism is not unproblematic. This is not necessarily a bad thing: perhaps the very process of discussing and debating issues can help clarify the many issues. But it is not as if we can simply draw on an established, well-understood body of knowledge.

Commentaries and other Old Texts

although not without some quirks like inability to search by spelling with diacritics, SPED 1.0 is a great tool for offline use


Yes, this whole area requires further research, with references and argumentation. For example, I can give such a gloss for nimitta as 'representation, perceptual image":

‘‘Katamañca, bhikkhave, anurakkhaṇāppadhānaṃ? Idha, bhikkhave, bhikkhu
uppannaṃ bhaddakaṃ samādhinimittaṃ anurakkhati aṭṭhikasaññaṃ
puḷavakasaññaṃ vinīlakasaññaṃ vicchiddakasaññaṃ uddhumātakasaññaṃ." (AN 2.16 = DN 3.225)

That’s what I like about Margaret Cone’s dictionary - it gives glosses for every meaning of the word, and you can explore these glosses further yourself. An electronic dictionary may well have a feature of adding glosses as well.

As Charles Muller writes, “it is desirable to have a locus classicus wherever possible”
( )
and a personal responsibility with clear authorship
( )


(Just one point to start, may I ask that you use the SC method of quoting texts, it makes everything work smoothly. You’ve used vol/page, we use the section and sutta number. In this case the passage is at AN 4.14)

Well, let’s have a look at this. First, to establish that one sense of nimitta is cause, we have the following at SN 48.40:

uppannaṃ kho me idaṃ sukhindriyaṃ, tañca kho sanimittaṃ sanidānaṃ sasaṅkhāraṃ sappaccayaṃ

This is nice clear explicit text, and nimitta is listed as a synonym of the standard words for “cause” or “reason”.

Second, to consider why it’s a word for cause. The more basic meaning is, of course, “sign”. How do we get from a “sign” of something to a “cause” of something?

Well, consider in what sense it’s a sign. One common use is to say that the dawn is the pubbanimitta of the sunrise. In other words, when we see the sky starting to light up, we know that the sun will soon appear. So a nimitta is something, when it appears, you know that something else is about to appear. Of course, we know the difference between correlation and causation, and so did the Buddha, and we know that the dawn doesn’t “cause” the sunrise. But it is very natural and normal for the mind to make these kinds of connections, and thus it’s easy for the language to slip from one meaning to another.

(Notice also, that in this context, even though the actual meaning is that of “sign”, the connotation is of an emerging brightness. This is, I believe, one of the contexts that gave rise, much later, to the idea that nimitta somehow means a light or perceptual image.)

Okay, so what of meditation? Nimitta is used quite commonly in meditation, and it has the general meaning of “something that you pay attention to that will develop or give rise to some other quality”. So simplistically, cause or basis, but more subtly something like aspect or characteristic.

Consider the phrase samādhinimitta. The normal, expected meaning of this, based on the general use of nimitta in the suttas (and energetically purging our minds of the taints of the commentaries!) is “something that you pay attention to in order to give rise to samādhi”.

Thus we have at MN 44, cattāro satipaṭṭhānā samādhinimittā, “the four kinds of mindfulness meditation are the basis for samādhi”, i.e. you do satipaṭṭhāna in order to get into samādhi. This is just restating the normal function of the seventh path factor, right mindfulness, which is to lead to the eighth factor, right samādhi.

It is, of course, quite impossible for nimitta here to mean “perceptual image” or any kind of visual image, as this would not apply to most of the satipaṭṭhāna practices.

Now, in the passage you quoted above, the kinds of corpse contemplation are, of course, part of the first kind of mindfulness meditation, the observation of the body. So there’s no reason to think that any other meaning is relevant here. I’ve translated the passage you quote from AN 4.14 as:

a mendicant preserves a meditation that’s a fine basis for convergence: the perception of a skeleton, a worm-infested corpse, a livid corpse, a split open corpse, or a bloated corpse.

This is clear and straightforward. It’s talking about cause and effect: what to practice in order to develop the path factors.

In this case, since the meditation can include an element of visualization, it is possible to read nimitta as implying a perceptual image. This might be plausible if that meaning was well established elsewhere in the suttas. But it isn’t, which is why this passage is cited. But you can’t establish the meaning of a word by relying on an ambiguous context.

There’s no problem with reading this passage in terms of the normal, well-established meaning of nimitta. So there is no compelling reason to introduce a new meaning, which is otherwise unattested until many hundreds of years later.

Sure, these things are good. There should be no problem to set up a “wiki”-style system that preserves authorship. Regarding glosses, perhaps we can also include this as part of a “gradual enhancement”. Set up the basic meanings first, then create the ability to add glosses.



Dear Bhante @Sujato,

Thank you for this:

:heart_eyes:Excellent translation. Now I am beginning to understand the exact and correct purpose of the corpse meditation! It’s not to be repulsed but to understand nature and anicca!

with respect, reverence, and gratitude,



I don’t think we’ll get to agreement here, but let this be an example of textual criticism.

There’s no need to get from a “sign” to “cause”. These are just two different meanings of the word “nimitta”, well-attested in the early enough Sanskrit texts:

nimitta n. (possibly connected with ni- √mā above ) a butt , mark , target MBh.

  • sign , omen Mn. Ya1jn5. MBh. &c (cf. dur-n°)
  • cause , motive , ground , reason Up. Kap. Var. Mn. MBh. Ka1v. &c (in all oblique cases = because of , on account of cf. Pa1n2. 2-3 , 23 Pat. ; mfn. ifc. caused or occasioned by ; °ttaṃ √yā , to be the cause of anything Ka1d. )
  • (in phil.) instrumental or efficient cause (opp. to upā*dāna , the operative or material cause) Veda7ntas. Bha1sha1p.

IMHO, defining the terms which don’t have explicit explanations in the Suttanta, on the basis of only Suttanta, necessarily implies a personal conjecture.

Thank you for the info. I’ve found a version of PTS dictionary which suits David Alfter well.


I had a look at the SPED code. It’s just a straight dump of the DPR PED XML, and uses a regular expression to perform a sans-diacritical search of the terms in the “Pali” search mode (if you use diacriticals it wont work) or a word match in “english” search mode (you can use diacriticals).


I’m actually more interested in the 83,593 words. So I think that would not really be a glossary or dictionary of main terms, but a dictionary of all actual words. Someone mentioned most common words first. Another possibility is to take a single Sutta and start with that.

Part of the reason I became interested in this actually has to do with MN 2. I have not found a single translation that does not consistently use the word “What” in places where by inference of the text that follows, the translation would much more likely be “How”, or “What about” or some other phrase.

For example,

What taints, bhikkhus, should be abandoned by using?

What would logically follow would be a list of taints, because the question (as translated) is “What taints,”

Instead, we get a list more like how taints are abandoned by using, or what things are used in order to abandon taints, but I do not see a list of taints, as would logically follow:

Here a bhikkhu, reflecting wisely, uses the robe only for protection from cold, for protection from heat, for protection from contact with gadflies, mosquitoes, wind, the sun, and creeping things, and only for the purpose of concealing the private parts.

There are other Suttas where I have run across similar issues, where it occurs to me that there is probably a translation challenge at hand.


What’s that?


Okay, sounds good.

Welcome to the club! I don’t mean the MN 2 club, I mean the “something seems a bit odd about this translation, I wonder what the text really says” club. Sometimes, it is true, knowing the underlying word is indeed helpful in sorting out translation strangeness, sometimes not.

In this case the word is katama, which does indeed mean “what”. Perhaps this is merely a textual corruption. How can we test this? If you go to our details page:

You’ll find there are a number of parallels for this sutta. Let’s see what they say. The Pali parallel at AN 6.58 also uses katama. We can check the Chinese texts, too, using our nifty Chinese lookup tool. All of them use 何 hé = what. So we can say that that the extant texts, so far as this survey guides us, are consistent in using “what”. So it’s unlikely that we’re dealing with a textual corruption, or, if it is, it will take some pretty sophisticated sleuthing to uncover it.

Now, given that the extant texts all agree on “what”, perhaps we need to re-read the text?

If you look at the structure, the first of the seven sections begins, “What defilements should be given up by seeing?” There follows a long section on various things, which doesn’t seem to answer the question exactly. But this section ends with:

Tassa evaṃ yoniso manasikaroto tīṇi saṃyojanāni pahīyanti—sakkāyadiṭṭhi, vicikicchā, sīlab­bata­parāmāso. Ime vuccanti, bhikkhave, āsavā dassanā pahātabbā.
Rightly focussing in this way, they give up three fetters: identity view, doubt, and mistaking morals and vows. These, mendicants, are the defilements that should be given up by seeing.

That is, the question “what” is indeed answered, but only at the end of a very long section. This is not an uncommon pattern in Pali, which tolerates longer logical structures than we are used to in modern English.

Nevertheless, it is possible that the very large gap between question and answer arose because of the interpolation of more material at a later date. I won’t pursue this further, but it’s worth bearing in mind.

Anyway, I hope that helps!


Bhante, I’ve read and learnt much over the years, but only properly studied the EBTs in the last 5 years. As an unexpected result I now find I have to remind myself; not of what the Buddha DID say but instead of what he did NOT say. Thank you for efforts in helping in this way. Here your clarification of nimitta is a good example. There is a lot said about nimittas that is not in the suttas. The same with many other things, like jhāna. For jhāna, the suttas generally just say something like:

“Quite secluded from sense pleasures, secluded from unwholesome states, he enters and dwells in the first jhāna, which is accompanied by applied and sustained thought and filled with the rapture and happiness born of seclusion."

Compare this to some of the detailed explanations and commenatries about jhāna and you start to get the feeling that there is a proliferation of jhāna going on. The problem then becomes, how much should we rely on the explanations of teachers? Is jhāna something realtively accessible for most whose morality, sense restraint and mindfulness is consistent and who practice meditation regularly and experience qualities such as those given in the suttas? Or is jhāna something much deeper, much more sublime and something not to be taken so lightly? These are retorical questions which I often find myself asking. With metta - S.


Sorry, I should have been clearer. I wasn’t talking about SPED, but about Mizuno’s table of Pali suffixes.


I know, right? This is where wisdom comes in, and this is something that you can’t get from a text or a dictionary; but perhaps these things might help a little.


Thanks to everyone who has contributed to this thread, I think I, and hopefully we, have a better idea of where to go with this now.

May I propose that we proceed along the following general lines.

Phase 1

  1. Make corrections to CPED.
  2. Filter out words and meanings that are not in EBTs, and add missing words and meanings that are in EBTs.
  3. Match the entries with the list of tokens from the EBTs, aiming for 100% coverage.

Phase 2

  1. Create a wiki-style environment for collaborative enrichment of the entries, especially those of significance.

I think this is a workable approach, and it has the benefit that we will make a meaningful improvement at each stage. Even if all we do is correct the current digital CPED, this is worthwhile.

I won’t complicate this post by going into details, but let me know what you think about this. I’ll make a separate post where I discuss the first step in more detail.


It turned out that the PTS dictionary included in the Digital Pali Reader project seems to be marked up even further than David Alfter’s now lost version.


Are we looking at the same thing? Because I just checked the PED.xml files in the DPR, and they have no markup other than what was in the original source from U Chicago. In fact, the DPR files have less semantic information. The original, like the version on SC, clearly marks the headwords, whereas in the DPR the headwords have the same <b> tags as are used for Pali terms throughout the entries. So even the most basic semantics in a dictionary—headword and entry—are absent. The remainder of the markup is just some presentation stuff, which is itself carried over from the source.


I did not check the files. David Alfter wrote me that this version seems to be marked up even futher.


Yes, thank you!

I found some more info on this conundrum and am still thinking that it is a matter of translating that word. The PTS dictionary does include “how” at the very end of the definition.

Also, the printed translation that I have of MA 10 uses “how”.

So, I am wondering if there is a grammatical rule that I don’t know that would be a reason for not using “how”?

If not, then it occurs to me that as one is correcting and developing a dictionary, one would want to include all the possible EBT meanings of a word so as to be thorough. Would “how” be a possible EBT meaning of “katama”?

Many thanks for your thoughts and knowledge !


Only in one specific context. I have checked the different dictionaries.

  • PED acknowledges the meaning “how” as an adverb in a passage from the Milinda (see below)
  • CPD also acknowledges this.
  • DOP acknowledges the same meaning and passage, but doesn’t render it as “how”.

So basically we have one passage, Mil 3.3.7:

Taṃ kiṃ maññasi, mahārāja, deve vassante katamena udakaṃ gaccheyya?
What do you think, Your Majesty, when it rains, where (how, in what direction) does the water flow?
Yena, bhante, ninnaṃ, tena gaccheyya
It flows downhill, sir.

Here, how you render it is a matter of style, but using “how” is certainly not required.

I’d say not. The Milinda is not an EBT (it’s dated around 400 years after the Buddha), and it seems there is only this one idiomatic passage. In any case, at the end of the day, the only reason for writing a dictionary is to help people understand the text, and in this case, the passage is clear enough, so I don’t think it would make any difference whether “how” was included in the dictionary or not.

In the original MN passage you were looking at, it really does mean “what”, the problem was not in the word katama but in the structure of the passage.


Aww… :slight_smile:
There are a number of entries that are missing in the ' version of the CPED and they were also missing from SuttaCentral ‘CPED dictionary’ (130 from Anuvijjhati to Apagamma).
But… very glad to find that they are in the file (& also in the site).