KN:Ud1.8 discrepancy in Pali



I’ve only very recently began undertaking Pali studies and came across a discrepancy in the abovementioned sutta:

In SuttaCentral the phrases are:
“khuddaputtañhi, samaṇa, posa man”ti.
“eso te, samaṇa, putto; posa nan”ti.
… brāhmaṇan”ti.

However, in and also in the notes provided in class, they are:
‘‘khuddaputtañhi, samaṇa, posa ma’’nti.
‘‘eso te, samaṇa, putto; posa na’’nti.
… brāhmaṇa’’nti.
(Also in .mobi from
and also sort of in

Do you know which one is the correct one? :slight_smile: :wink:

Suttacentral looked correct but when I searched for nti (& hoping to find none - hence sort of confirming that it is an error) there are 2 suttas with them in the Vinaya that have the same form: “…” nti.


Further to this, while examining MN82 Ratthapala sutta, I found something that probably doesn’t belong in the pali text: ()
Na tvaṃ, tāta raṭṭhapāla, kassaci dukkhassa jānāsi. () Maraṇenapi te mayaṃ akāmakā vinā bhavissāma.


Hi Waiyin,

I discussed this difference and other details recently in this post

This is a mark indicating a variant reading. Usually such variants are marked per word or phrase, but in some cases this is not convenient, so this convention is used to indicate that something has been omitted. Click “textual Information” in the sidebar, and you can hover and see the variant.

In this case, which is quite unusual, an entire sentence has been omitted by the editors of our text. The note says that the text is found in all manuscripts, however it is missing from the parallel passage in the Vinaya and also from the commentary. In their judgement the relevant sentence was a mistake, however they recorded it as a variant.


to follow up on Ven Sujato’s response, more specifically


Hello Bhante Sujato, How are you? Hope that you are well.

Thank you so much for the copious explanation (couldn’t go through them all), but thank’s to @LXNDR, thank you so much for specifically zooming in for me.

The two instances in the with "nti, according to SC search engine are:
13 occurences in
11 occurences in
e.g. pubbakaraṇa”nti vuccati.

Regards, WY


-nti is a form of -ti used when the final word of the quote ends in -a (short a!).

This can become very clear when you read Abhidhamma where there are many occurrences where single words are quoted.

For example in Vibhaṅgapāḷi:

  1. ‘‘Nepakka’’nti. Yā paññā pajānanā
  2. ‘‘Pītisukha’’nti atthi pīti, atthi sukhaṃ

Another hint about the -n- not being part of the word is that there is no grammatical case ending -n for -a nouns. There’s -ṃ but that’s a very different sound/letter.

-nti can be a conjugated verb ending, of course this has nothing to do with the mentioned texts.

So I think 'nti is correct and -n’ti is not, although in particular cases the later might be.

My personal feeling is that saying -ti after short -a might inadvertently produce another phoneme, maybe -adi or -athi. Whatever the reason was we should recall that for a very long time Pāḷi only was an oral language and introducing things like -ti -nti for quotating stuff was a necessity.


Sorry, I don’t think that’s correct.

If you look at the kinds of quoted forms used in such Abhidhamma passages, the defined term is the same form that it appears in the quoted text. Thus sampajāno or bhikkhu are in nominative, while sikkhāpadesu remains in locative plural, and so on. Thus in the cases that you mention, they cannot be the vocative or stem forms like nepakka, as such forms do not appear in the quoted text.

Rather, these are the accusative forms ending in anusvāra. The usual representation of anusvāra in Pali as is replaced in such instances by the class nasal, which here is n, due to the close conjunction with the following t. Pali manuscripts are full of such variations.

Interestingly, we are finding that in the oldest Pali manuscripts, the class nasal is usually neglected in favor of anusvāra, which is much easier to write. So we have, for example, paṁḍita rather than paṇḍita.


To me the most important thing was the use of -n- which is a very different sound from the pure nasal anusvāra. If it were 'ṃti I wouldn’t think otherwise. Also because many examples of -ṃ’'ti exist in the same texts where 'nti exist.

However what I didn’t note is that in my own examples the nominative case of both “nepakka” and “sukha” (both of them being neuter) - ends with ṃ. More importantly both are used in a nominative construct (a is b).

So now I agree with your theory, it seems neuter nouns’ nominative being the same as their accusative and moreover ending in -ṃ made me think otherwise. Because -ṃ is very often a marker of singular accusative indeed, regardless of gender (of course also a case marker for other cases and numbers).

You shouldn’t be sorry, I don’t mind being wrong, it’s easier to learn.

Just like you (?) I would first think about the texts being written in devanāgarī however maybe it’s better to think of them being written in Sinhala. I don’t know how to read/write, but I found that often the ‘nti’ thing is rendered as 'న్తి while -ṃ’ti is written as ං’ති

Using the ‘long’ way writing ṇaṁ’’ti would be ණං’’ති while using the ‘short’ way it would be ණ’’న్తి.
I think the explanation might be as simple as Sinhala not having pure vowel nasals (at least 2000 years ago or even today?) but has a very typical ‘nt’ consonant cluster so it might have been more ‘natural’ or easier when writing the texts down.

I wonder if there’s an online resource with clear pictures of the ola leaves with pāḷi so we see how those cases were written. In modern Sinhala there’s also ‘න්‌ති’ which is n + ti.

Simple question but I’ve learnt so much in 2 days. If there are Sri Lankan natives please verify my findings.


They would have been first written down in Brahmi, roughly 2,000 years ago. Modern Sinhala script didn’t come into use until much later, probably a little over 1,000 years ago.

I have hoped to find an article tracing the evolution of letterforms from Brahmi to Sinhala, but so far have not found anything.

The edited texts that are available today, as well as almost all the manuscripts in Sri Lanka, are based on manuscripts that were taken from Sri Lanka to Myanmar or Thailand, copied there in Burmese or Khom (Khmer) script, then returned to Sri Lanka and recopied in Sinhalese characters. So the influence of all these scripts must be borne in mind.

The 13th century Cullavagga that SC is working on will, when published, be the only substantial published Pali text that did not go through a process of copying in S-E Asia. Hopefully it will have things to teach us about this process.

At no stage have Pali manuscripts been transmitted in Devanagari. Probably the Nalanda edition of the 1950s was the first time Devanagari was used for the Pali Tipitaka.


Do you mean that there’s a 13th century MS of the Cullavagga? Or is this a typo for Cūḷavaṃsa?


No, the Cullavagga. It is the Dutiyaparakkamabahu Cullavagga housed in the Colombo Museum. We have images of the manuscript, and a team to type it up and publish it.


Hi Sujato, you are sharing very interesting information.

My idea was that we should try to find out what the source texts for the English transliterations were. However since no nasal vowels exist in Sinhala to a Sinhalese nibbanaṃti sounds exactly like nibbananti, that’s why I think we find those 'nti things, because it would have been natural to a Sinhalese to think of ‘nti’ as single phoneme rather than for example -aṃ + -ti (or -naṃ + -ti).

We could probably apply the same thought on texts being written in Thai or other languages, especially if they aren’t Indo-European and may lack a lot of the sounds of Pāḷi. If many of them came from Sri Lanka then maybe the original ‘error’ was there.

Where can I find more information on Cullavagga? Are the images you have freely accessible somewhere? If not - maybe you could check relevant texts where the 'nti thing would occur to see how it’s written.


Itipi so bhagavā arahaṃ…
Itipi = (iti + api) And thus

Puttā matthi dhanammatthi , iti bālo vihaññati;
The translation can be done without the word iti.
Here it means thus (like that or like what mentioned by “Puttā matthi dhanammatthi

When something is qouted in the texts iti is often used to point that part out “this is how it was said” with the same meaning of thus.

When it ends with “ti” the passage qouted with Quotation Marks.
So the meaning comes with the word thus (iti).
More often “i” of iti is elided and only “ti” remains.
(A junction: pubbasaralopa).

When it comes to a junction like
bhagavā’ti - bhagavā + iti (i lopped) : bhagavā’ti
(devamanussānaṃ buddho bhagavā’ti.)

Iti as deictic adverb; “thus, in this way”; pointing
to something either just mentioned or about to be mentioned.

A qouted sentense from a onversation ends with iti
“Na kho, raṭṭhapāla, tathāgatā ananuññātaṃ mātāpitūhi puttaṃ pabbājentīti.

pabbājenti + iti = pabbājenti + iti = pabbājentīti.
Here the end vawol become a longer sound making ī.

Iti as emphatic part. pointing out or marking off a statement either as not one’s own (reported) or as the definite contents of (one’s own or other’s) thoughts.

Imasmiñca pana veyyākaraṇasmiṃ bhaññamāne āyasmato koṇḍaññassa virajaṃ vītamalaṃ dhammacakkhuṃ udapādi – Yaṃ kiñci samudayadhammaṃ, sabbaṃ taṃ nirodhadhammanti.

nirodhadhammaṃ + iti = nirodhadhammanti.
junction: āgama (ṃ + i = n)

I do not understand what you are pointing here. Are you talking about the junction. Or not mentioning iti in the translation? Normally iti is not translated into other languages; the meaning remains unharmed.


Hi Amatabhani,

Using your example this is mostly about two things (for me):

  • what is the correct pronunciation of "nirodhadhammaṃ + iti, specifically the -ṃ + iti
  • what should the correct way of writing transliterated text in the Roman script be: nirodhadhamma”nti or nirodhadhammaṃ”ti

If indeed sandhi rules apply (and I’m reading now there are special sandhi rules only regarding niggahīta then it probably is best to stick with "nti.

Since Sujato pointed out it’s better to use -ṃ"ti I thought the -"nti spelling is due to error, but if niggahīta sandhi apply in this case why would we break the sandhi in written texts?

By the way, what would be the best source on Pāḷi grammar?

I found this shedding some more light on niggahīta:

In all cases we were discussing the anusvāra / niggahīta is “true” since it’s the case ending at the end of a word. Still apparently Sandhi rules should apply.


First one is correct.
nirodhadhammaṃ + iti = nirodhadhammanti
It is a “n” āgamasandhi.

Lets see how it came.

Grammar rule


a, ā, i, ī, u, ū, e, o


ka vagga: k, kh, g, gh, ṅ;
ca vagga: c, ch, j, jh, ñ;
ṭa vagga: ṭ, ṭh, ḍ, ḍh, ṇ;
ta vagga: t, th, d, dh, n;
pa vagga: p, ph, b, bh, m;
Avagga: y, r, l, v, s, h, ḷ, ṃ

Vaggantaṃ vā vagge

When there is a consonent after the niggahīta it is replaced by the fifth consonent of the particular vagga.

1. kiṃ + karo = kiṅkaro
2. dhammaṃ + care = Dhammañcare
3. taṃ + ṭhānaṃ = taṇṭhānaṃ
4. jutiṃ + dharo = jutindharo
So on…


The reason the romanised version appears as it does is simply because it was digitally converted from Burmese Pali.

Consider, for example, the phrase “pāpaṃ me katan’ti tappati” (‘Evil have I done,’ torments him) in Dhammapada 17. In the Mahāsaṅgīti romanised edition it appears as:

‘‘Pāpaṃ me kata’’nti tappati.

But why not in the more usual PTS form: ‘‘Pāpaṃ me katan’’ ti tappati.” ?

Because in Burmese it appears as

‘‘ပာပံ မေ ကတ’’န္တိ တပ္ပတိ၊

As you can see, the components of န္တိ (= nti) form a graphetic unit in which there’s no place to insert speech marks. If the Burmese had inserted the speech marks after the န္တိ, then it would have made iti part of the speech rather than a marker of the speech. And so it had to be inserted before န္တိ.


This was my initial thought, a similar thing might happen when the text is written in Sinhalese. However as later Amatabhani pointed out there’s more complexity to this because of sandhi. But it’s very interesting to learn texts were digitally converted from Burmese. What’s the source of this information?


It was Lance Cousins’ explanation to me when I asked him about the nti thing many years ago.


“At no stage have Pali manuscripts been transmitted in Devanagari. Probably the Nalanda edition of the 1950s was the first time Devanagari was used for the Pali Tipitaka.”
Bhante, interestingly, there is one old nagari (not modern Devanagari) manuscript discovered in Nepal which contains fragments of Pali Vinaya; Dated based on the alphabets between 8th-10th Centuries AD.
I am sure you are already aware of this.

P. V. Bapat
Annals of the Bhandarkar Oriental Research Institute
Vol. 33, No. 1/4 (1952), pp. 197-210

The article also has photos of the manuscript but it’s not too legible; The transcription in Devanagari has “nti” in a couple of places. But per the article it has a variety of ways dealing with anusvara and nasals.

This is a very old article, but don’t think any other manuscript of Pali in Indian script has been found subsequently and perhaps none were made - till the Nalanda edition as you mention…


Thanks so much for that explanation, I never knew why that was so!

The ultimate source of the digital text on SC is the edition created by the Vipassana Research Institute, which was based on the Burmese edition of the Chatthasangayana text. I believe their primary source was the second edition. The digital files were proofread and corrected by the Dhamma Society of Bangkok, who checked it against multiple printed editions.

Oops, I stand corrected. Let me rephrase: “The Pali texts on which modern editions are based were never transmitted in Devanagari.”

The Nepalese Pali manuscript is a fascinating sidebar into the complexities of Indic texts. We actually created a digital edition of that last year:

It seems likely that this text was copied in Nepal based on an original. Further details are pretty much guesswork: it was perhaps taken from Bodhgaya, and the script may have been Sinhala. But it could have been something else entirely. In any case, it seems clear that the copyist in Nepal was not familiar with Pali.

Incidentally, this text also shows an inconsistent treatment of anusvara:

  • suttamayaṁ”ti
  • pāsakan”ti


Hello again, everyone!

To add to the discussion… While reading “A new Course in Reading Pāḷi” by James. W. Gair / W. S. Karunatillake I found more information on sandhi concerning niggahīta and also texts indicating their preferred way of spelling such sandhi, particularly with iti/-ti. Not sure if this was meant as a guide to students or is a generic spelling convention, but I thought it’s good to mention it.

So, what’s important in the first place, of course is that cannot exist between vowels, it would either become ‘m’ (etaṃ+attho in accusative becomes etamatthaṃ) or ‘n’ (viññāṇaṃ /nom./acc./ + ti becomes viññāṇanti).

In the text that follows in the book they use quotation marks before ti but spell the case marker as ‘n’, for example: "Bhante Nāgasena, kiṃlakkhaṇaṃ viññāṇan"ti?

I think this might be the best way to spell this in the Roman script.