The "Play of Formulas" theory of the development of the EBT's

I have started watching;

this video from Centre for Buddhist Studies Berkeley of Eviatar Shulman talking about a thesis that argues that the earliest strata of texts are the “one-liners” or short formulas that are repeated and recombined and juxtaposed all through the suttas. and that it is these formulas that are elaborated into discourses over time.

I am starting a new thread but was actually inspired by my conversation with @Pasanna in the topic.

I am only about 20 minutes in so will go watch the rest and come back to this thread but wanted to share with others and get peoples thoughts of this interesting take.


Ooh! @sujato you rate a mention! at 19:35, although you get mistaken for a “UK” scholar :slight_smile:

The book Eviatar Shulman mentions in his talk is available now. The preface is free to read online using the ‘look inside’ feature on amazon. I’d put a nice quote from it here, but I can’t easily copy the preview.


Fantastic! thanks so much!!

Having watched the video, a few quick impressions;

Firstly Eviatar speaks pali very fast, much like a person would their first language, this was interesting because it was a style of speaking pali i had not heard before and it was fun.

Secondly while he positions himself in opposition to scholars like Gombrich, Analayo (and our own Sujato!) He really does not enunciate a thesis that would be particularly shocking to those scholars, as I understand them, it is perhaps a matter of emphasis more than substance, and in the Q&A he more or less completely walks back his criticism of the “comparative project”.

Despite his straw-manning of comparativists, the talk is quite interesting, and I look forward to reading the book if it ever comes my way.

I have his earlier book: Rethinking the Buddha: Early Buddhist Philosophy as Meditative Perception in PDF and a few papers by him from Academia. If anyone wants them, let me know.

1 Like

not sure in what sense this is context? The discussion in the video focuses on EBT’s and nothing is mentioned about syncretism or judaism at all.

I’m glad for Shulman’s reflections on the transmission process for they might revive fruitful studies that have received too little attention, particularly

Gethin, R. M. L. (1992). The Mātikās: Memorization, Mindfulness and the List.

Allon, M. (1997). The Oral Composition and Transmission of Early Buddhist Texts.

In the chapter of my thesis on oral transmission (pp.13-41) I offer reflections that I think complement Shulman’s.


Can you summarize his thesis for us? Or at least, how it differs from the “comparativist” project?

Well, I guess the major contention is that the “formulas” are like Lego blocks and that the “discourses” where built up from the “formulas” by the reciters over time, combining them in many ways and that therefore the scholars involved in the “comparativist project” have placed too much reliance on whole discourses, or rather not given enough attention to the folkloric, narrative, literate and doctrinal aspects of the formulas themselves, honestly there wasn’t an amazing amount of detail, but one of the examples was how the motif of realising that it was too early for the alms round is always associated with visiting the wanderes of other sects - sounds like maybe more details are given in the book - sorry I’m really not doing the talk justice, it was quite entertaining and “provocative” - he calls into question how much we should rely on things like the location or the name of interlocutors- using the example of odd it is that Vacchagotta seems to need to ask the same questions of the Buddha, then Mogallana, then Amanda, then the Buddha again, and suggests that even the locations and the interlocutors and the frames of the discourses should be thought of as “formulas” that where probably used quite freely in recombinations by the early reciters as much for literary reasons as any kind of concern for historical accuracy or factuality.

Hope my word salad is of some use to you Bhante.


Somewhat off topic: thank you for sharing your dissertation. It’s very interesting work, and seems to reinforce Bronkhorst’s ideas about the religious milieu of ‘Greater Magadha’.


Thank you for posting this very interesting lecture.

It seems one of his main ideas is that the pursuit of some kind of essential ‘ur-text’ of what the Buddha himself actually said, through the method of comparing the Chinese, Gandhari, and Pali texts, is ultimately impossible and probably pointless. In other words, he rejects the idea of comparing these texts to get closer to the Buddha’s actual words.
He seems to view the Pali canon, and all early Buddhist texts, as a rich and artistic spinning out of essential ideas (formulas), that were probably heard from the Buddha originally.
How this ‘play of formulas’ eventually made it into various bodies of closed collections didn’t seem to be addressed here, I probably have to read his book.

1 Like

Here is a similar presentation:

From the abstract:
“The play of formulas is a new theory designed to explain the manner in which discourses (Suttas, Sūtras) were composed in the early Buddhist tradition, and especially so in the Dīgha- and Majjjhima- Nikāyas (the collections of the Buddha’s Long and Middle-length discourses). This theory combats the commonly accepted views in scholarship that texts are mainly an attempt to record and preserve the Buddha’s teachings and life events. Rather, driven by a variety of creative vectors, mainly literary and contemplative ones, the texts betray a rich engagement with the figure of the Buddha and a visualization of his figure. The play of formulas explains how formulas – the basic unit of Buddhist oral textual culture – combine in order to produce meaningful textual patterns and statements. Formulas connect according to set narrative designs, in which different types of audiences are represented not only with their unique formulas, but also with their specific narrative trajectories and complementing doctrinal emphases. It is not that the early authors were not at all trying to preserve the Buddha’s words. It is only that there was much more going on, and that thinking about the early Buddhist texts in this way misses all the fun, and ignores their beauty. These probably tell us more about what the texts actually were for the people that first produced and studied them than dry formulaic doctrine.”

In other words, he rejects the idea of comparing these texts to get closer to the Buddha’s actual words.

I don’t think he entirely rejects the approach, but just believes it’s heavily overemphasized. Moreover he believes that the assumption of ‘the discourse’ as the fundamental unit of the early texts often underlies comparative analysis, when instead careful reading and comparison within one of the existing EBT collections reveals the primacy of the formulas.

I recommend the book, it opens up a lot of new ways to relate to the discourses. In his book, he also emphasizes the importance of literary elements in the discourses. So the play of formulas idea is not the only thing discussed. He discusses some of the suttas in the DN as a form of literary Buddhanussati. Those sections make me think of Kastrup’s book More Than Allegory which discusses how myth conveys truths that cannot be directly stated. It seems he takes some suttas as attempting explanation of deep truths (about the nature of a Buddha for example) via myth. (Although this shouldn’t be confused with a rejection of rebirth.)


Could you expand on this? I’m not sure what you mean.
Is it that while comparison between entire discourses in different languages/ recensions is how analysis is traditionally carried out, with the aim of getting closer to buddha-vacana, comparing texts within one collection does this better?

Maybe what I’m wondering is if the goal of his analysis is rather different than, say, Ven. Anālayo.
Are they approaching the same thing from different angles, or are they very different?

"This theory combats the commonly accepted views in scholarship that texts are mainly an attempt to record and preserve the Buddha’s teachings and life events. Rather, driven by a variety of creative vectors, mainly literary and contemplative ones, the texts betray a rich engagement with the figure of the Buddha and a visualization of his figure. "
(from his Abstract)

Thanks, this is helpful.

It reminds me of what was called “redaction criticism” in Bible studies. The point being that the original text has passed through “redactors” and their hands left interesting marks on the texts as we have them.

I think the point of this criticism is that the “discourse” is assumed to be a given teaching stated at a given time to a given group of people. Differences are considered to be more-or-less trivial representations of that fundamental unit.

IIUC, he is suggesting that we pay more attention to the manner in which discourses are composed, since their composition reveals details of how the teachings were received and understood.

There’s a saying by someone in the field, unfortunately I forget who, to the effect that: no hypothesis of early Buddhism is less plausible than that the Buddha spent his whole life merely repeating the same few doctrinal formulas without variation.

Beware of academic contrarianism. It’s basically a requirement to get any work noticed that you have to say, “Everyone else is doing that. But I say we should do this.” Whether it’s true or not is beside the point, it’s a rhetorical strategy.

It’s certainly a valid approach to look at how formulas are used and to analyze their relationships and so on. I’ve done some of it myself. For example, in the Northern suttas on satipatthana, there is a commonly used short formula. Is this a genuine difference from the Pali version? I found that there’s a place in the Chinese samyukta where it gives the abbreviation, and then says that all abbreviated suttas should be expanded like this. So clearly the short form is not intended as a different content, merely a convenience.

As to formulas of setting and so on, it has been an accepted axiom of the field since the work of Ven Thich Minh Chao in the 60s (at least) that the narrative settings vary more and hence are likely to be later than the doctrinal teachings. So sure, the use of narrative formulas (pericopes) may reveal something of the way the texts were organized and the methods and values of those doing the organizing.

But there are plenty of things in the texts that are not formulas. Lots of narrative contexts are quite unique and personal. On the other hand, lots of suttas have no narrative, or if they do, they are so generic as to be useless (and often just abbreviated away). Paying attention to formulas might illuminate some patterns, but the choice of method has already focused the light of attention in a particular way that will illuminate particular kinds of truths and obscure others. This is acknowledged in Shulman’s choice of subject: by focusing on MN and DN he chooses to look precisely at that subset of EBTs where narrative invention is present. There’s nothing wrong with that, it’s just a way of seeing is all.

One general problem with the field of Buddhist studies is the dislocation between fundamentals and theory. Consider the state of fundamentals in the field:

  • we don’t have full translations of Chinese Agamas
  • we don’t have full translations of Sanskrit texts
  • we don’t have full translations of Chinese Vinayas
  • we don’t have a complete, up-to-date dictionary.
  • few scholars are fluent in all relevant languages

Things have got a lot better, to be sure. But there’s a long way to go. The field is large and complex, the workers are few and scattered, and funding is somewhere between thin and non-existent.

Meanwhile, academics are influenced by current trends in academic theory. But those theories evolved in other fields, such as Bible studies, where none of these problems apply. They have multiple fine translations, excellent and thorough study programs, and lots and lots of money. Like, lots of money. Hundreds of millions of dollars every year kind of money. Bible studies has evolved through generation after generation of theoretical approaches, all building on the fact that the Bible was already well-translated and the basics well understood since the age of textual criticism began well over a century ago. It’s a much smaller text, with many more people studying it critically for a longer time.

So there is a tendency to want to apply different theoretical models to Buddhist texts, influenced by developments in other fields (and of course Bible studies is just one of them). There’s nothing wrong with this—I’ve studied Bible criticism precisely to learn from what they have been doing. But for the field as a whole, what’s really needed is to lay the groundwork, to create a solid foundation. Perhaps insights from other fields will help illuminate Buddhist studies. Or perhaps we will find our own way.

It’s early days, is all I’m saying. Let’s listen to different theoretical approaches and see what we can learn. But just be a little cautious before hoisting your flag on the latest mast. (Does that metaphor work? Seems suss to me!)

Worth noting that TW Rhys Davids wrote some great literary analysis of Suttas; see for example his introductions to DN suttas, or his analysis of Jatakas.


The point isn’t that it works better or worse exactly nor that it’s based on within collection analysis (probably I shouldn’t have added italics there). Basically by comparing suttas (without needing to compare with agamas for example) which contain similar formulas we can see a process of ‘sutta-generation’ that involves combining them in creative ways. There are many suttas which are ‘generated’ like this based on a ‘seed formula’. So it seems rather than being fixed records of historical events, suttas like this at least are examples of exploring the meaning of certain formulas by the authors of the canon. (See @sujato’s comment as well which explains this better than me-the one after the first time he quotes me.)

He (unsurprisingly) references his work several times in the book.

1 Like

So to compare different versions of a sutta can help identifying what was probably the earlier/earliest version. We make out the purpose and agenda of the text and of the alteration.

But obviously only few suttas are textually independent and there are numerous trans-sutta connections. Conversion formulas, question formulas, politeness formulas, winning-a-debate formulas, hesitation formulas, criticism formulas, astonishment formulas, invitation formulas, enlightenment formulas, etc.

Beyond analyzing and understanding a single sutta the analysis of such formulas can show us the purpose and agenda of such text modules which, per definition, must be older than the individual suttas which include the modules/formulas.

I find this approach very valuable. And add the approach/question: What are examples of suttas which are free of trans-sutta structural elements? They could represent individual voices we could really treat as individual suttas, standing on their own. Maybe we find these only in verse suttas, but maybe also in some prose. We should approach the interpretation of such suttas differently.


It’s funny how people can’t seem to conceive of the Dhamma as having been discovered and expounded by one being.

I guess it’s a kind of compliment to the Buddha that people feel there must have been a committee of authors working behind the scenes to create the body of suttas …

1 Like

Please add substantial arguments against the formula approach. It’s good for you that you have faith in the texts as they are, but maybe that’s better expressed in a seperate topic.