The Buddha Knew Proto-SN Formulas: Play of Formulas and the Saṁyutta Collections

Recently, there has been some scholarly debate around the role of formulas in the formation of the early suttas and their development. Eviatar Shulman, a proponent of this hypothesis, has written about it in Visions of the Buddha: Creative Dimensions of Early Buddhist Scripture (2021) and in an article The Play of Formulas in the Early Buddhist Discourses (2021). Bhikkhu Anālayo has written a response: ‘Visions of the Buddha’: A Critical Reply. I would also like to shout out @knotty36 for having brought these questions to the front of my mind :slight_smile:

Here is a summary from Anālayo’s article:

The central theory proposed by Shulman (2021b: 227), referred to with
the phrase “play of formulas,” involves the following proposition:
The point here is that the formula, rather than the full discourse, is the main level of textual utterance … this theory suggests that the main texts of early Buddhism were the formulaic encapsulations of both narrative and doctrinal materials, and that full suttas are primarily legitimate combinations of such formulas. This notion allows us to understand how discourses were created from formulas bottom-up.
The term formula here refers to “fixed textual elements that are reproduced mainly across texts, often across many texts” (p. 171)

Now, obviously there is a “play of formulas” in the suttas. Prime examples of this which are well known are the two Mahā- and regular Satipaṭṭhāna Suttas. @sujato has done a comparative study of what are essentially the formulaic building blocks of these suttas and their parallels in A History of Mindfulness. What is even more interesting—and perhaps telling—though is that in this work, Bhante Sujato reconstructs a theoretical “Proto Satipaṭṭhāna Sutta” using formulaic building-blocks. That is, we still have modern examples of Buddhist monks making suttas out of formulas to access buddhavacana. Historically, many early schools of Buddhism understood ‘buddhavacana’ as not necessarily spoken by the historical Buddha, but in line with what he did say or re-arrangements of his statements. That is to say, composing new suttas out of older formulas to say the same types of things was not inauthentic buddhavacana for many, if not all, early communities.

The ‘debate’ surrounding the play of formulas has much more to do with the formative years. We know there is a consistent play of formulas in later parallels and already existing suttas. This was a mix of memory lapse and intentional addition, rearranging, swapping, etc. to make suttas more holistic or what have you. But what about the earliest forms of the suttas? What about what the Buddha and his disciples were memorizing and spreading? Etc.

In A History of Mindfulness, Bhante Sujato also spilt a lot of ink over the organization of the suttas and their early structure. This was building on scholarship from people such as Yinshun who also understood the Proto Saṁyutta Nikāya/Saṁyukta Āgama as the earliest base collection of doctrinal information. That is to say, originally there would have been sutas—short doctrinal statements and what have you—which were organized together in these saṁyutta-like bodies. These could be used for teaching, expanding on suttas, standardizing material in other narrative suttas, creating new ones, etc. Indeed, the formulas are the oldest prose in the canon that we have by their nature: If several suttas all contain the same formula, those formulas predate the full composition of those suttas or the full texts as a whole. The SN/SA collections are the most formulaic collections we have in that they have minimal narrative in comparison with the MN/DN and repeat formulas up the whazoo. Lots of suttas in the MN / DN also have shorter counterpart chunks in the SN (think DN 16, MN 49, etc.).

I think that there is a tiny speck of evidence for the early use of SN/SA-like formulas in the EBTs. I do not know if this has been discussed by Shulman or Anālayo, and so I apologize if it has. Either way, it’s good for it to be out there.

At SN 12.45 and SN 35.113—parallel suttas in the canon—we read of the Buddha reciting a SN-like formula to himself on paṭiccasamuppāda. A monk overhears it, and the Buddha instructs the monk to follow the Buddha’s example and memorize/recite this formulaic passage. There are a couple of things to note:

  • It seems that the formulaic nature of what the Buddha was reciting is original; that is, in the proto-form of this sutta (if we can say there ever was one), the Buddha was reciting some kind of formulaic chunk of doctrinal content — a parallel to the current formula in the Pāḷi canon. I say this because he was reciting a “dhammapariyāya” — or a passage/exposition of theory/the teaching — to himself. This term seems to be a sutta-term for “discourse” or a specific, memorized teaching, rather than a theme of discourse or general discussion. “Sutta” originally meant some kind of short, doctrinal statement or discourse which Bhante Sujato and others have also discussed.
  • This passage is formulaic and it is doctrinal/technical, very much like the SN/SA formulas (in fact, it is an SN formula). This would support the idea that (a) the Proto-SN/SA likely existed in some form, and (b) the Buddha knew some of these and may have structured his teachings around or with them to aid people in memorizing them.

I’m not sure if there are any Chinese suttas which have this kind of narrative in them, i.e. one in which the Buddha recites a sutta to himself that is formulaic. If anybody knows of some, please comment them here! It seems that the Buddha knew the Aṭṭhakavagga and Pārāyanavagga, or some form of them/their discourses. He knew people memorized teachings, and seemingly he even knew formulaic ‘suttas’ that are the building blocks of so much of the canon.

I think this all adds support for the play of formulas, specifically Proto-SN/SA discourses which existed among the early community, along with early verse discourses and chunks of verse. Once there is this standard, it would be very easy to expand it into more narrative like suttas of the MN and DN in all kinds of different ways. From there, with established proto-narrative suttas, the formulas can continue to be played with or rearranged (intentionally and unintentionally).



These examples are great, and suggestive of the monastic practice of repeating doctrine to oneself in “mantra” style as a part of meditation, potentially lending support for certain vitakka vicara arguments.

However I think it is a stretch to use this as an evidence in favour of the SN/SA is the earliest collection argument.

For starters the formula in those suttas is absent from 2 of the Nikayas and so violates the principle enunciated vy Rhys Davids that the formulae most likely to be early are those recuring in all the books.

Secondly SN/SA in addition to formulaes, also has a marked feature of applying those formulaes to other formulaes in mechanical, repetitive ways, and in DN and the bulk of MN, especially those suttas in MN that are also in MA (as oppesed to parraleled in SA) we see the simple formulaes individually rather than combined, this is suggestive of the possibility that SN/SA, just like DN and MN, drew from a pre-exisiting stock of formulae that it used, rather than being the original source of said formulaes Otherwise we aught to see DN drawing on those combined formulae rather than only using them in their simpler singular form.

Third SN repeatedly qoutes DN suttas by name, while DN never qoutes ans SN suttas by name.

Fourth several formulaes that are given in DN as ending in “ti” are given in SN ending in “titi”.

Anyway, I am on my phone and not able to go through all the reasons to think SN reflects a later systemisation of dhamma rather than an earlier source of formulaes, but suffice it to say that the argument is not without issues.


The Buddha not only did not know porto-SN/SA formulas, but also did not know proto-MN/MA, DN/DA, AN/EA formulas. The four principal Nikayas/Agamas were not created by the Buddha.

You make some good points which I think should be addressed. I know you’ve said them elsewhere. I haven’t seen a thorough discussion of them all between people here yet. Here are a couple of thoughts.

T.W. Rhys Davids’ quote of course is that those are most likely to be early. This is a generalization and a statement of probability. There are cases of things which occur in all four nikāyas that are likely later, and there are formulas which occur in less nikāyas which are likely earlier. So while this is a factor to consider, it is not black or white in any regard.

But this is precisely part of the hypothesis: that this phenomenon of repeating and using a stock of formulas — short, doctrinal statements — to build upon one another was used not only by the early community but also perhaps by the Buddha himself. He would analyze things in terms of general feeling, but he might also analyze feeling at the six senses, and then the craving for the feelings at the six senses. That the SN/SA displays this behavior is precisely part of the reason why it may resemble a proto stock of formulae which were extended out.

These formulae pre-date the suttas by their nature as building blocks. If short formulae appear in longer narrative suttas in regular, standardized ways, then those formulae — as you mention — must have been part of an earlier stock in some sense, or at least were very likely so.

It makes much less sense that this stock of formulae would come from longer narrative passages where they are used sporadically. Especially when these narrative suttas themselves can often be easily deconstructed and taken apart similar to the SN/SA. Note too that when people would ask the Buddha for a teaching to practice alone on their own in seclusion, he would teach them some SN-like formula. This is further proof that the early community was memorizing these short stock passages we see in the SN and much less frequently in the MN/DN (ones on the senses, aggregates, etc.).

This is not to deny that there are formulae and passages from the MN/DN which are older, and it is not to say that the current SN/SA is older. Each nikāya is ~2600 years from the time of the Buddha. I am referring to a proto-nikāya which contained these stock formulae. Rearranging them in different ways and repeating them in different permutations would also be very helpful for memorizing them and then being able to use them later in larger narrative or other suttas.

This is not the case of these collections are specifically for longer, narrative suttas — which they are. They do not display this behavior because this behavior is found in the SN/SA. If we drop hypotheses and begin with a descriptive overview of the canon, we see that the SN is clearly used for storing not narratives — which are the exception — but doctrinal statements. There is less reason for a narrative sutta to combine the formulae in this way because they are telling narratives.

This is less relevant to the issue for two reasons. First because the primacy of the proto-SN/SA is based around their main function: doctrinal statements and teachings with the help of formulae. Commentary on other suttas is a minor occurrence in comparison to much of the collection’s meat. Second, the cases in which the SN quotes the DN are a couple. The SN clearly has later suttas, and so this is no surprise. That there would be later sections of it is only to be expected and is no news. The DN also has all kinds of later elements that pervade the majority of the collection: whole suttas on later mythologies; Abhidhammic collections of later material; later verse and spell-like chants, etc. This occurs more frequently in the DN than the SN, and so if we were to go off of material that must be later (one passage quoting another, later mythological narrative, etc.), statistically the DN wins the lateness competition.

Note too that the later references in the SN are not integral to its composition and structure. We can insert a sutta into the SN in which a passage from the DN is analyzed. But the later material in the DN is often integral to the core of dozens of suttas and passages. It is, again, not a minority but a majority of the material.

I would like to speak to the contrary though as well. As I mentioned, the DN contains portions of material which we find in the SN. DN 16 is a good example of this: much of it is spliced together from earlier material which can be found elsewhere in the canon. We also find these types of suttas in their shorter forms in the SN, and they don’t quote or reference being from the DN / MN (such as the parallels to MN 49, the intro at DN 15, etc. etc.). This means that the DN does contain formulae which are used in the SN in longer, spliced, and more embellished form.

The Nikāyas are organized by the type of suttas within them. This is uncontroversial; it is native to their naming system across schools. The DN is for longer suttas as the name suggests and as the suttas within it suggest even if it were not to be named as such. But long, narrative suttas cannot be the building blocks of Buddhism—the evidence for the majority of the Buddha’s dhamma talks contradicts, what people memorized, and the usage of formulae are some examples of evidence in support of this. The fact that we find things such as lengthy, detailed descriptions of the gradual training then can be no surprise at all: they belong in the DN. If they contain earlier building blocks, there is no evidence that they were originally in the DN because they would not be long enough. Anything else would be mere speculation. We do find shorter descriptions of them in the MN, so there is evidence that these teachings, as they get shorter, find themselves in other nikāyas. Their inclusion then cannot be assumed to be on account of the age of the DN itself but on the length of the exposition.

This question of length is relevant to the SN quoting passages from the DN as well. The Sakkapañha Sutta has very late material in it so far as we can tell. It is also very long, meaning it belongs in the DN and the DN alone. But what if it was not originally in the DN Nikāya but has been placed there due to its length and the way in which it was expanded upon? The same is true of the Brahmajāla Sutta: The elaborate discussion of monks’ rules is later than the doctrinal formulae which the Buddha began teaching with that we find in the SN. They are by necessity later than some things in the SN, as are responses to speculative views of a metaphysical ātman. It may not have originally been in the DN, and much of its defining material may be slightly later. The inclusion of quotations of these in the SN does not prove that the Dīgha Nikāya is older; it proves that the passages from these suttas which are now included in the Theravada DN are older than those few SN passages.

On the contrary, the inclusion of formulas in the SN is part of the nature of the collection as a whole. The formulas are also demonstrably earlier than the sutta compositions themselves, including in the various nikāyas, and thus there is reason to suggest that the SN contains the footprints or basis for the organization of early material. There had to be a stock of these earlier verses and formulae somewhere as we’ve already mentioned — but where did it go? Lots of the verses in the SN are found repeated in the Snp or Thag/Thig, etc. So not only are there formulae in the SN, there are earlier chunks of verse which is expanded on or repeated elsewhere as well. This is something I did not discuss but which I find is relevant and interesting.

We can investigate further into the nature of the material in the Nikāyas. The DN consistently is closer to Brahmanism — especially more orthodox Brahmanism — and other sects’ views/doctrines. DN 1 and DN 2, for instance, are all about these different peoples’ views and debunking or discussing them. It contains verses praising Brahmanical gods/deities. It contains narratives showing the mythological significance of the Buddha as a cosmic figure and a destined identity. They contain demonstrations of the Buddha being more powerful than Brahmanical deities. All of this in more frequency than in the other collections and beyond the length. These types of narratives do appear in other collections, so it cannot be by virtue of length alone, and yet they appear much less frequently in these collections.

All in all, I think that there is obviously material in the DN which is earlier than other material in the SN, and vice versa. I’m sure that some of the narratives we find there have been around in some form or another for a while. I think there is a lot more evidence that would suggest that the proto (emphasis on proto) SN/SA-like collection is earlier in certain senses than the DN. Even if we go statistically, there is more late material in the DN than early. If something is a late narrative, it is more likely to continue to be expanded on, edited, and added on-to; if something is earlier and in a more proto-form (and is intentionally being preserved as such), it is less likely to be added onto and will instead appear sporadically in other material.

Sorry for the length of the response. I hope this is helpful in responding to some of these points. I’m open to other ideas, and I think the one on the DN is interesting, but I really don’t see much support for it.



@kaccayanagotta, you’re putting me to shame here: I’m supposed to be going to school for this, and you, just in your spare time, are delving into areas I haven’t even considered.

I am enjoying the little back-and-forth between you and @josephzizys here: it’s not about whose view is right and whose view is wrong, but these are important issues both of you have brought up, and I would love to see you has them out. (I’ll be travelling for the next 2-3 weeks, so I don’t know that I’d be able to participate, unfortunately.)

Two points came to mind when I was reading this thread; I’ll include them though they may be minor points which you both might already be cognizant of:

  • One thing which helped frame my thinking when I first came across Shulman’s ideas (and there’re are also two videos on YT [links here and here]) was a point that I believe Shulman pointed out in an article if not in his book. (Sorry, his book’s packed away right now, so I can’t pull it out and check.) What was said was that there was a distinction between sutta and suttanta that seems to have lost its significance. I’m sure there are many in the forum who are far more qualified to speak to this than I, but nevertheless the idea as I remember it was that the former was constitutive of the latter. The implication, in the context of the play of formulas, was that suttas represented formulas and suttantas represented discourses. I mention this because one of Mark Allon’s criticisms (I think it was him) was that, if there was a foundational period of somewhat free-floating formulas which preceded the finalizing and subsequent transmission of discourses as we know them, there should have been a record of it, there should be a term for these formulas. Well, perhaps there was, and perhaps there is.

  • Second, the main focus of Shulman’s work thus far has been on narrative formulae of the sort we find in the Dīgha or Majjhima Nikāya and their literary function. I know he has plans to expand into more strictly doctrinal texts, and he has started that already by looking at the Aṅguttara Nikāya and its role as a bridge between the discourse canon and early Abhidhammic literature viz formulas. Anyway, perhaps I’m missing something in your posts, but I didn’t see much distinction being made between narrative and doctrinal formulas. Shulman argues that the various nikāyas (at least the first three) weren’t “repositories” of ready-made texts in the sense of “Okay, you’ve got a long one? Put it here.” Rather, these were methodologies for creating texts: “Let’s make some longer narratives,” or “Since we’re already mixing and matching formulas, applying them to different categories, why don’t we arrange the resultant texts according to those categories?” (i.e., as opposed to according to the formulas themselves, though, at times, saṃyuttas are arranged just that way.) So, although there’s overlap and formulaic borrowing between the DN and SN, I would think that (in many, not all, cases, for the DN is quite an eclectic collection) most of this intersection is peripheral, almost coincidental due to their very different purposes. Anyway, I just want to say that I think any discussion (especially a historicist one) should take these different purposes as its staring point.

I don’t know if these two points hold any value for your thread: maybe I misread your posts or missed something; feel free to disregard if I did.



A thought that I have on this is to look at how sutra is used in other Indian traditions. I believe among Jains and later Brahmanical texts, it refers to a short aphoristic passage about which larger commentaries are written. This may have been the case originally among early Buddhists, but the commentaries were canonized as suttas later, and storytelling at some point became an important element to make them more memorable as unique texts. I could certainly imagine new renunciates being given short lists of topics to memorize with oral explanations, which would gradually expand over time.

Generally speaking, it’s eye-opening to compare the way standardized commentaries vary between Buddhist canons, and also the way their placement in sutras differs, too. It points to a period of sutra development that happened after independent canons were created. The Theravada texts seem to have gone through a rigorous editing process to standardize many passages throughout, but the actual occurrences and frequency of those passages varies quite a bit at times when compared to MA or DA. And those collections have their own formulae that they deploy more often than Theravadins did. But I’ve only looked at this anecdotally while translating parallels. It’s quite a task to sit down and document it systematically.


I don’t know how historically accurate this particular terminological distinction is, but I agree this distinction may be relevant. I agree ‘sutta’ would have originally been something close to the doctrinal formulas what we see in the SN, and a discourse seems to have been called a ‘dhammapariyāya’ (like the Madhupindika discourse, the Ajitamānavapucchā, etc.) several times in the nikāyas.

Thank you for clarifying this. I wasn’t aware Shulman was focused on more strictly narrative formulae as of now. It makes sense to divide them. I think that the proto-SN/SA doctrinal formulae (suttas) are slightly different from the play of literary narrative formulae in constructing MN/DN type suttas as well. I am more focused strictly on the former here because I think there is evidence that they circulated in the early community and that they form half of the basis for different discourses (the other half being narrative material).

This is good to know. I have tried to intentionally set these questions aside here, mainly because to use it as a form of argumentation would be a kind of ‘begging the question.’ If we take a descriptive approach of the nikāyas, we can see that, for instance, the DN contains long material and often material relevant to converts/outside views. Were the discourses now in the DN intentionally composed apart from other length narratives and organized this way intentionally from the start? I’m cannot be sure as of now. But what I can say is that, however the formation of the discourses came to be, there is reason why parts of the DN now present aren’t found in the SN, like say the long sections on the gradual training (and vice versa).

I think your point on the collections’ composition being tangential makes sense here as well and fits with what I’ve said above. If the formulas on the detailed gradual training are inherently long, then they will inherently be included in other types of collections, whether made intentionally long (as Shulman posits) or just by chance and organized as such. This is good to consider though for the distribution of the formulas throughout the nikāyas; several MN ones come to mind.

I hope that makes sense. That was my attitude going into that I at least tried to maintain. Maybe you have criticisms of this approach, or think that there are angles of inquiry we’ve missed out on? Any input you might have would be very valuable as someone who has looked into this quite a bit.

Thank you so much for the feedback and contribution! Safe travels, and hopefully you get the time to participate at some point or another.

1 Like

Good to see these discussions about the formation of EBTs’ formulas.

Regarding SN or proto-SN formulas, it is better to note SN/SA contains three structures/forms, according to Ven. YinShun. That is, SN/SA is a synthesis of the three parts/aṅgas (sūtra, geya, vyākaraṇa). According to Ven. YinShun, without knowing the characteristics of these three parts and their connection with the formation of the other three Nikayas/Agamas (MN/MA, DN/DA, AN/EA), it is impossible to understand the gradual process of forming the four Nikayas/Agamas based on the SN/SA three parts.

Cf. :


Yeah this is a good point. It also reminds me of the Udāna/Dhammapada and Atthakavagga phenomena: in the Pāli, the commentary is in the former and not the latter; in the Chinese, it’s the opposite. The Jātakas have the canonical verse with commentarial prose narratives. The Pāli Dhammapada is also often taught with the commentary as are other suttas, which I imagine, as you said, goes back a good while.

We see this in the Pāli canon as well: a poem will re-appear throughout the canon, but attributed to different people/backstories. At 7.9, we have the story of Sundarika Bhāradvāja. This is repeated at Snp 1.4, but with some more details in the narrative.

So at least in principle we see that this type of thing has and does happen.

Do any particular examples of this come to mind? IIRC the MA features, say, the jhāna formulae more than the Theravādin parallels generally. Is this the type of thing you were thinking of?


1 Like

Yes. I’m also thinking of those series of standard passages that get assembled together in MN covering various topics that are chained together to form a larger progression. Or sometimes elements are swapped out in the parallels, like the introductory story. Cf. my essay from a couple years ago about MN 22/MA 200/EA 50.8: MN 22/MĀ 200 The Simile of the Snake: Parallels of a Patchwork Sutta

1 Like

Thanks for sharing that post. This is really interesting, and it’s quite relevant to this topic of formulas and creating long/middle-length discourses like @knotty36 pointed out from Shulman’s thesis. The types of comparisons you did on MN 22 are, as you said, time consuming. But that type of work is quite valuable for understanding the composition of discourses.

The Satipaṭṭhāna Suttas (in the DN and MN) are good examples. Another one that reminds me of MN 22 is MN 28 (Mahātaṇhāsaṅkhaya Sutta). You can clearly break it down into different sections, and Alexander Wynne has also quite convincingly discussed how you can see where/why/how different formulaic passages get added on. First there’s the section on consciousness and the impermanence of consciousness and consciousness as being dependently arisen on X nutriment. Then the sutta goes into the four nutriments. Then it goes into the conditionality of the four nutriments beyond craving, then into all of paṭiccasamuppāda. Then it goes forward and back. Etc. Etc. — All the way to the end, where the Buddha says the monks should remember this “brief teaching.” “Brief”??? It’s a very lengthy discourse, and it takes a couple twists and turns. When we break apart the formulaic strata though, it makes sense.



Aww, you guys are like Godfather III: “Just when I thought I was out, they pull me back in!” I really have things to do, but I see you are going to give me no rest @kaccayanagotta starting tempting and tantalizing threads like this!


This is what should be the primary mission in Buddhist textual studies for the next few decades at least; maybe then we could at least scratch the surface. Especially with Chinese texts, which are so eclectic, where it might be a medieval commentary preserving snippets that parallel EBTs. Dhammadinnā’s pretty good with these sorts of things, but the range of requisite background knowledge is so daunting. We probably need more collaborative scholarship like medical or scientific scholars do: papers with 35 authors and such.

The historical significance of these and related terms (which may differ from school to school) needs to be explored. Shulman’s real contribution, in my estimation, lies not in any conclusions he reaches, any reconstruction of the creation of the suttas, or any new tidbits of trivia concerning individual discourses he may turn up, but just in giving us a new outlook on how to approach the texts. From a bigger picture standpoint, I could see him being outlandishly wrong in every paper he writes and then, when reviewed 50 years from now, totally vindicated in his having given rise to a new approach which was subsequently refined and taken up by later generations of scholars.

I actually couldn’t follow you here. Sometimes, e-chat doesn’t work so well for me. Would you mind giving it another go?

Thank you @thomaslaw for bringing him up. I’m actually going to spend about six months in Taiwan at Dharma Drum starting about a month from now. Seeing how to tie Yinshun’s theories (findings?) in to Shulman’s approach (or maybe the other way around?) would be great. I really know very little about what he says about the formation of the early texts beyond basically what you just posted. Anything you (or @cdpatton) can add? I’m sure there are a bunch of threads at SC that have dealt with this which I could review. But even anything you might offer in the form of questions that come to mind would give me a good compass for what to do with my time there. Also, which Yinshun books I should be looking for for the necessary historical background. (I should probably just man up and grab the whole collection.)

I, too, was convinced by Wynne’s argument here. But, then, I was also just as convinced by Anālayo’s rebuttal. (Here, if you’ve never read it.) And that’s really a big problem (for me, at least) when it comes to the back-and-forth between Shulman and other’s regarding the formation of the texts, formulas, and so on: each one sounds really convincing. I think that’s why they’re each so convinced of their own theory. It has to do with, in my opinion, the fact that there’s just so little empirical data to go on. And these people are all the cream of the crop, so their theories are of course compelling. But it’s almost all speculation, truth be told. And, again, being as frank and honest as possible, none of us is really making very educated guesses here (at least, no more than anyone else), because there’s really so little to go on.

But, like I said, I feel Shulman’s greatest contribution is shaking up the field a little (a lot?). He’s at least gotten the conversation started. Debate is necessary here. Because, after a time, we should be able, on that basis, to at least start ruling out impossibilities. Perhaps then we’ll be a little closer to making “educated guesses.” And, if these Gandharan guys keep turning up stuff, who knows? We may even be able to make a categorical statement or two at some point.

1 Like

I had never read it, and thank you very much for sharing it! I am now in the same position as yourself more or less — I was convinced both ways.

I think it’s interesting that Anālayo discussed the hypothetical insertion of certain formulaic passages contra Wynne. In a way, it’s part of the common ground with Shulman: clearly the formula is the basic unit for the composition and structure of the early discourses. I very much agree with everything you’ve pointed out on speculation, Shulman’s contributions, etc. We’re getting somewhere, in that at least we are considering new perspectives for analyzing and understanding the texts.

Speaking of, have you seen this talk by Mark Allon, or read any of his research? He discusses some of this and is able to demonstrate some more categorical developments in the textual tradition — at least educated, informed hypotheses.

What I mean is that my goal here was to try and be descriptive. What differences are there between the current forms of the DN and SN that would explain why two types of texts are separated — rather than the origin or specific means in which that separation came about. For example, saying that the long descriptions of the gradual training with monks’ rules belong in the DN because of their length and would be out of place in the SN. Originally, they may have been placed into the DN because they were later arranged there or because they were specifically made long for a DN-type discourse, but either way we know descriptively that the current categorical difference in the types of formulae or texts can, to a certain extent, explain the lack of cross over between the two collections.
I was wondering if you think that this type of approach is flawed or could be altered/improved upon.


1 Like

This article by Choong Mun-keat may be useful:

“Ācāriya Buddhaghosa and Master Yinshun 印順 on the Three-aṅga Structure of Early Buddhist Texts”, Research on the Saṃyukta-āgama (Dharma Drum Institute of Liberal Arts, Research Series 8; edited by Dhammadinnā), Taiwan: Dharma Drum Corporation, August 2020, pp. 883-932.


1 Like

Hey, @knotty36 ! It’s been a bit since we’ve discussed some of this. I’ve been having some thoughts on it and on DN 15 (which we both seem to be fascinated with :smiley:). Wanted to run some things by you and initiate some more conversation. Feel free to take your time responding if you are busy.

I’ve looked some more into Shulman’s work and ideas, and I’ve understood better. I see where there confusion and misunderstanding can easily arise, because his point is extremely subtle and yet must be presented in certain ways for academic, social, and other reasons.

My main understanding of the “play of formulas” is that we must look at suttas, especially narrative suttas, as more than just “a vague record of a most-likely historical event with details that can be comparatively deduced.” Instead, we have to see that these texts are also performing literary, mythological, devotional, instructional, and other functions, while also trying to capture important aspects of the Buddha, his life, his dhammavinaya, the community, etc. When we look at these texts, we see that they are made up of smaller stock passages and formulaic structures that repeat again and again, and any particular discourse is really about the dynamic created by using these formulas to construct a scene. From these formulas and their dynamics, we can deduce broader facts about the Buddha, community, etc. And we can look at a text as a whole unit that tells a general narrative via varying formulas across schools/parallels. But all of that is still an aspect of the dynamics of how these smaller formulas work together with un-spoken narratives and settings to weave together a discourse.

When the Sangha wanted to create a discourse about things that transpired, or things they believed to have transpired, or things they believed were in line with the Buddha’s message (buddhavacana), they could combine basic historical information alongside literary narratives and use stock formulas to create an easily recitable discourse that contained some history, some mythology, etc. etc. — and the formulas themselves (+ the dynamics between them as they progress and present themselves) are the main unit for conveying these particular nuances of meaning, purpose, intention, etc. If these formulas in and of themselves were from an earlier time, it would give even more authority to discourses constructed out of them, because the entire discourse and the dynamics were the same message / word of the Buddha, but put together in different unique ways with some contextual variations. The texts are then able to retain authenticity and fluidity.

As for more strictly doctrinal formulas, this same type of usage and ‘play’ happens and is quite common. I actually think people are more aware of this than they might think, but it tends to be overlooked. A very easy example is the Kaccānagotta Sutta, SN 12.15. The Buddha answers a question by, at the end, just listing off the 12 links of dependent origination. But how would some other wanderer or new monastic have any idea what that means? Sometimes the Buddha responds with these 12 links to non-Buddhists with no exposure whatsoever, and the same is true for other doctrinal formulas (sense fields, aggregates, etc.) The point is clearly that these doctrinal formulas are a stand in for what is meant to be a philosophical statement. So dependent origination itself answers the question, and the 12-link formula stands in for that answer.

How is my understanding? Do you think I have grasped the heart of what Shulman is trying to get across and open our minds to considering? Do you perceive any misunderstandings on my part? You are much more well read and studied on this than I, so I wanted to run this by you.

I have some ideas related to the DN and DN 15, but I’d love to discuss them some with @josephzizys , so I’m going to leave that comment under another post. Joseph, you may find this discussion relevant too, though I know you’ve seen Shulman’s work already.

Mettā! :pray:

1 Like

I don’t know that I should be the go-to on this subject: it’s not that I’ve really studied it so much, and, actually, I doubt that I would have been to articulate some of the subtleties of the theory as well as you just did. In any case, I think you grasped it pretty well: everything you said may not appear explicitly in his book, papers, and talks, but it all sounds to me fully consonant with it. Again, the theory’s not really such a big thing in-and-of-itself; it’s just an approach. It’s like a prism effect, or maybe a kaleidoscope: it’s just a simple refraction, but there’s infinite potential.

Please be aware, though, of how Shulman paints himself into corners. (Someone mentioned this to me, and so I’m now I’m on the lookout for it.) For example, he hasn’t yet, but he’s seeking to extend the theory into vinaya and abhidhamma areas. I don’t know (particularly because I’m not well-read in those areas) if that’s fully appropriate. To whatever extent, these would seem to be products of a different era which came together under different circumstances and for different purposes. Also, they are constructed quite differently, from a formulaic standpoint. (This also extends to the AN collection to a certain degree.)

Another point to be aware of is the danger in his (or anyone’s) making overly ambitious historical statements about the construction of the suttas. This is where he and people like Anālayo get caught up in a dosey doe when, really, no one knows for sure what happened. While I fully concur with your assessment of the Kaccānagotta Sutta’s construction, we have to remain fully cognizant of the fact that we are treading in the realm of speculation–albeit perhaps wholly logical! I don’t see any problem with what you wrote per se, but “historical reconstruction” of this sort can be a slippery slope. If we get too over-zealous and project too much, it’s very easy, again, to paint oneself into a corner.



Thanks, @knotty36 ! And I definitely agree. Behind all the academic advertising and whatnot, it’s really just a reminder that the suttas are composed of smaller chunks that interact in interesting literary ways — and we should be paying attention to this, rather than overlooking it for what the Buddha really said more or less verbatim or something. At the end of the day, the suttas are the compilations and work of the Sangha, not the Buddha. This is easily forgotten in some circles. :slight_smile: