Some inauthentic passages in the Early Buddhist Texts


This article is collaboratively authored by Bhikkhu Sujato and Bhikkhu Brahmali. It is a work in progress.

Those who come to the early Buddhist texts often ask how they are to know what passages can be regarded as authentic or inauthentic. By “authentic” we mean that texts that claim to be spoken by the Buddha or his direct disciples were in fact spoken by them. Authenticity does not imply literal word-for-word identity, but that the substance of the text, and much of the wording, is what it says it is: the teachings of the Buddha.

As per our book, The Authenticity of the Early Buddhist Texts, we believe that most of the texts included in what we call the early Buddhist Texts (EBTs) can be regarded as authentic. These texts are:

  1. The 4 main nikayas in Pali
  2. The six early books of the Khuddaka (Dhammapada, Udāna, Itivuttaka, Thera- and Therīgāthā, and Sutta Nipāta)
  3. The Vinaya (especially the patimokkha and portions of the Khandhakas; but excluding the Parivāra, a later addition)
  4. Such parallels to these texts as are found in Chinese, Sanskrit, Tibetan, etc.

All other Buddhist texts are later, and where they contain genuine words of the Buddha, these are quotes from the EBTs. In saying that these later texts are inauthentic, we are merely acknowledging the historical facts of their provenance. Whether such texts are true or beneficial expressions of the Dhamma is an entirely different matter.

Now, while all of our authentic sources are found in the EBTs, not everything in the EBTs is authentic. All scholars in the field admit that there are significant passages that must be regarded as later additions. Given the scope of the field, the paucity of scholarship, and the uncertainty of methodology, it is not possible to identify all authentic passages with any certainty.

It is, however, we believe possible to identify some passages as inauthentic with a fair degree of confidence. This is in cases where a series of independent criteria taken together all point in the same direction, with no counter-evidence. In this post we will assemble a list of such passages, together with a summary of why they are inauthentic. It is not exhaustive or authoritative. It is simply a convenient place to note significant discourses or passages that are likely to be inauthentic.

If you wish to suggest any passages to add to this list, please do so in the comments. If we both agree, we will add it.

###MN 111 Anupada

While including some early material, such as the jhana formula, the discourse as a whole appears to be a late composition.

  • Speaker: attributed to disciple
  • No Parallels
  • Vocabulary: abhidhamma style terminology
  • Abhidhamma: in addition, the overall manner of the discourse is very abhidhammic
  • Style: main list of dhammas is patched together as evidenced by two separate styles.

###DN 22 Mahāsatipaṭṭhāna

Most of the content is early, but the text as a whole is a late composite expanded from MN 10, which itself is somewhat late.

  • Structure: a duplicated discourse, made by combining MN 10 and the extended passage on the truths from MN 141
  • Form: hardly any discourses of this length are given as direct teachings, rather than Q&A.
  • Parallels: there are several parallels, and none of them have the extended noble truths section.
  • Abhidhamma: includes a long list of dhammas in Abhidhamma style.

###DN 30 Lakkhaṇa

An elaboration of the 32 marks of the Great Man, almost the entire discourse is very late.

  • Subject: the 32 marks of the great man, while included in a number of EBT texts, have dubious relevance for the Dhamma and claim a brahmanical provenance that is not supported by brahmanical texts.
  • Meter: Much of the text is verse, and the metrical styles are all late.
  • Parallels: There is no real parallel; the partial parallel at MA 59 merely gives the stock passage on the marks and has none of the extra explanation.
  • Doctrine: The detailed linking of sopecific deeds with specific results, while having some precedent in the EBTs, became a favorite doctrine later.
  • Style: The verses are unusually flowery.

###SA 604, SA 640, SA 641

These passages are not sutras at all, but rather passages from the Aśokavadāna. They were added by mistake to the Chinese text of the Saṁyuktāgama (SA = Taishō vol. 2, sutra 99), probably through a filing error.

How to teach metta in the sense of the early texts?
Studying the first five fetters

Okay, so @brahmali, here is a start. It’s very bare bones, obviously, but it’s something.

Please feel free to edit change or remove anything I write here, now or in the future.

I added a couple of texts as a template, but in future I think it would be best if we proposed texts in the comments, and added them when we both agree. Does that sound reasonable?

I think you should be able to edit the main post, but let me know if you can’t.


Wonder, how the drum will sound now after this attempt.
By the way, will the new drum be the old drum?


How marvellous! Thank you Bhante & Ajahn!

There are some other suttas I might have ordinarily asked about, but instead I wanted to put forward what seems a very obvious one: MN117. There are a whole bunch of instances on this forum where the very likely abhidhammic influence in that sutta has been mentioned and I likewise recall watching a talk in which Bhikkhu Bodhi suggested that this sutta is important for understanding the development of the Pali Canon and how abhidhamma crept into the suttas.

The reason I wanted to ask about it though, is because in many ways folk seem to find the sutta useful (gnlaera’s post here serves as a little indicator towards the point, and I’m sure it wouldn’t be hard to find more corroborative evidence). That’s absolutely no reason at all not to mark it out as inauthentic if it is most likely inauthentic, however, it does lead me to wonder about the kinds of approaches that might be advocated towards those suttas that can be confidently identified as inauthentic.


It would be great if it could be confirmed and made clear what exactly does not make MN117 an authentic EBT sutta.

As far as I understand, the argument for labelling it non EBT kosher is all about the Abhidhamma-influenced inclusion of the concept of mundane and supramundane factors of the path.

I really fear people may end up throwing the baby with the bath water and skip MN117, thus missing the practical value of the concept of a serial Noble Eightfold Path the MN117 is all about.



I question the authenticity of the entire Saṁyutta Nikāya 34: Jhānasaṁyutta

It seems to be an expansion and abhidhamma style matrix permutation contrived exercise on a couple of suttas in AN, 6.24 himalayas
an 6.72 which cover the essence of the matter in a short oral tradition style teaching as is typical of EBT.

What new information is SN 34 providing? Nothing I can see.


Yes, I see what you mean (was looking at a wrong text earlier and got muddled). They do seem inauthentic.

Would I be wrong in saying that though text is inauthentic that doesn’t imply the teachings contained within them are always out of line with the Dhamma @Sujato? It would be an easy assumption to make, IMO.

with metta


I didn’t say AN 6.24 and AN 6.27 are not authentic. I said, SN 34 may be inauthentic based on the observation that it uses abhidhamma matrix permutation technique that doesn’t match the simple oral tradition style of AN 6.24 and AN 6.27.

edit: removed comment that’s no longer applicable


You might consider adding some of the āgamas that I’m working with in my off-again on-again dharmakāya thread, if they are EBTs, they seem unable to be “proven” to be so by any of the more rigorous methodologies that can be applied to parallels in Indic languages, and seem to exhibit some of the rūpakāya-dharmakāya Buddhology that characterizes a later period of Buddhism than what one usually deals with in EBTs. We really need a handy discovery of a complete copy of the Sanskrit.

Its an ambiguous issue, because 2 of them “are” EBTs, they have parallels, they match their parallels reasonably, and yet anomalous language features in them (such as discussing dharmakāya) that sets them apart from “normal” EBTs.


Thanks everyone for their suggestions.

Regarding MN 117 and SN 34, my initial thought is that, while they may indeed have late features, the bulk of the content is pretty normal. Maybe they, or part of them, should be included, but they’re not at the top of my list.

I just checked, and I’m afraid that a substantial portion of that article is based on SA 604, which is not the Samyuktagama at all. It is, in fact, a portion of the Ashokavadana, a later biography of King Ashoka, which was mistakenly inserted in the Chinese text, probably through a filing error during the Chinese transmission. If it makes you feel better, Etienne Lamotte, one of the true greats of Buddhist studies, made the same mistake. So anyway, it is certainly inauthentic.

Meanwhile, @vimala, SA 604 has a display bug in the division table. The title appears as SA 604(六〇四) SA 604(六〇四), where the Chinese is just the number. Now, I think this is extracted from the text, and I have changed the title there, so hopefully this should be okay. However, we should also add a "Note’ to this entry, which can say

SA 604 is a portion of the Aśokavadāna. It appears to have been inserted in SA by a clerical error during the transmission in China.

A similar situation occurs with SA 640 and SA 641. We should add a similar note, changing the sutta number of course.

@brahmali, it should be a formality, but to confirm: add SA 604, SA 640, SA 641.


Splendid, thanks. But to give a tiny further push for the above point, taking whatever suttas you do think squarely belong on the inauthentic list, how is it that you (or others) feel they should be approached, is making this list an alternative way of saying “don’t bother read those ones”?


Yes, “inauthentic” essentially means “is not authored in the way that it claims”. I’ll clarify this in the OP.


For @brahmali: some low-hanging fruit:

  • DN 17 Mahasudassana
  • DN 20 Mahasamaya
  • DN 23 Payasi
  • DN 24 Pathika
  • DN 30 Lakkhana
  • DN 32 Atanatiya
  • DN 33 Sangiti
  • DN 34 Dasuttara
  • DN 16 Mahaparinibbana: closing verses and various other sections.
  • MN 116 Isigili

For fans of DN, sorry, but I may well want to add DN 18, DN 19, DN 21, and a few others to this. But I’ll be generous and leave it at this for now.

Lateness of the Anguttara Nikaya

The exercise has been attempted here

This article offers comparisons between the textual material contained in the Mahācattārīsaka Sutta and texts of late origin, such as the Abhidhamma and some of the works found in the Khuddaka Nikāya. It will show that this sutta forms a patchwork of what is probably early material and inserted late formulas.

The article has been recently revised and the original tone softened.

Also this book contains a fair amount of research relevant to this thread:


I wonder whether we need to define authenticity in a bit more detail. Take DN22: as a whole the sutta is late, but there are substantial parts of it that are early. The same is true of MN 111. (By contrast some of the suttas you list below have hardly any early material at all, such DN 20, DN 30, and DN 32.) This is the sort of dilemma we will face throughout. I suppose we need to decide whether suttas that seems to contain any late editorial addition or change or deletion (apart from the narrative) should be included on the list. This would mean that MN 117, for instance, needs to be included. Perhaps we need different categories: entire suttas that are IA (inauthentic), suttas where the majority of the material in IA (e.g. DN 22), and suttas where a minority of the material is IA. We need to find a way of choosing suttas that has sufficient integrity and clarity. I haven’t made any changes to the OP, since I think we need to discuss this first of all.

@Coemgenu brings up the interesting point of whether the same exercise should to be done for the Āgamas in Chinese, and indeed other languages. My feeling is that we should focus on the Pali at present so as not to make the task unmanageable. If, however, Coemgenu and others wish to contribute towards the goal of including texts from other languages, I think that would be great. I suppose they can be added to the list when we feel the evidence is strong enough to warrant it. Whether we really have the expertise to make that call, however, is not clear to me.

Another thing we should perhaps do is reference secondary literature. The problems with the Satipaṭṭhāna Sutta(s) go far beyond the additions made to DN 22, as you have amply demonstrated in The History of Mindfulness. The summary you have made above is certainly useful, but it is a far cry from your thorough historical deconstruction. People who may not be persuaded by the former, might well find the latter compelling.


Sure. There is nothing very controversial here, but it is easier to make a firm decision once you see the actual criteria for lateness. Would you be able to give a list of hard-hitting and irrefutable criteria? (Yes, yes, I know there is no such thing.)

I think some of the issues I raised in my previous post are exemplified by the list you have presented here. Both DN 33 and DN 34 contain a lot of early material. Yes, the presentation is different and the amount of IA material is also pronounced. Yet I suspect both of these suttas belong to the category I would call “minority IA”.

I don’t have any basis for having an opinion on this, but I trust your judgement. So please go ahead.


Okay, well let’s go ahead and add these, and see about adding criteria as we go.

How about we expand the entries to include a description of the type of inauthenticity? Rather then trying to establish a typology, which would get complex fast. I’ve edited the OP to show what I mean, and added an entry for DN 30.

Yes, this would be a good enhancement.


I have just looked at your changes and it seems to work. This is probably the most accurate way of capturing the degree and the nature of the inauthenticity.


That looks a lot better. :anjal:


[quote=“sujato, post:11, topic:5795”]
I just checked, and I’m afraid that a substantial portion of that article is based on SA 604, which is not the Samyuktagama at all. It is, in fact, a portion of the Ashokavadana, a later biography of King Ashoka, which was mistakenly inserted in the Chinese text, probably through a filing error during the Chinese transmission. If it makes you feel better, Etienne Lamotte, one of the true greats of Buddhist studies, made the same mistake.
[/quote]Thank you, bhante. I figured when I started the inquiry, due to the nature of what I was inquiring about, that they may well all of them be later, so I put “and early Buddhist sectarianism” to cover myself :sweat_smile:. I’ll edit the article accordingly for the sections dealing with SA 604.

In response to Ven @brahmali: there is a small thread dedicated to collecting possible manuscript errors. We just need someone who is actually “professionally qualified”, if you will, to go over and confirm those that are not the observation of someone already qualified (and if such qualifications are already behind the decision to put something on that thread I will make sure that it is cited).

If the list grows big enough (and the corrections on it significant enough) to be an issue for people trying explore the Chinese Buddhavacana, perhaps that list of notes could be confirmed and then integrated with the site so that readers will have that information available to them, but as it is currently, it is moreso just a list of occasional curiosities than anything that would be vital for clarification and context when reading these texts.