So I decided to write my own Pali textbook

I can’t believe I’m getting close to the 2 month anniversary since I started learning Pali.

So now I have reached the stage of a fully developed Dunning Kruger syndrome - I am now conceited (and foolish enough) to believe I can write my own textbook on learning Pali.

Why am I bothering? I have been steadily getting frustrated with all the other textbooks - they don’t seem to teach me the way I would like to be taught (which is Everything, Everywhere, All At Once). Some textbooks take a whole book to cover the grammar - I prefer to get an overview right up front so I can start reading almost immediately. Textbooks also focus on the vocabulary and introducing lots of words I am not sure I’ll need. These days it’s easy to lookup an electronic dictionary so I prefer just to learn as I go along. Lastly a lot of textbooks either cover too much detail, or not enough detail. I want a textbook that just teaches enough to get me started, but provide a comprehensive overview, not just bits and pieces.

And lastly, there is an adage “Those who can, do. Those who can’t, learn. Those who can’t learn, teach.”

I started reading Kaccāyana about three weeks ago and decided that it contained all the information I need, but it was very hard to navigate and full of complex rules. So I initially started writing “An Idiot’s Guide to Kaccāyana” but quickly realised it was morphing into something quite different. I am still using Kaccāyana as my primary source of reference, but at the end of the day the structure and most of the explanatory text is mine. I also developed a notation style for marking up Pali sentences with grammatical tags. All the hard and complex stuff will be moved to appendices. The textbook is also focused on learning how to read Pali, and much less on being able to write or speak it, so my focus isn’t on teaching the finer points of grammar or how to do all the right transformations to conjugate a verb or decline a noun.

This currently represents about two week’s worth of effort. Actually, about 50% of it was written in the first 2-3 days in a bout of inspiration, and 80% in a week. The rest of the time has been spent getting the Introduction and Chapter 1 to a reasonable state that I can share with others.

The work is open source, so if you wish to contribute, by all means view the source code and create a pull request. I aim to release weekly updates until I am happy with it (or die, whichever comes sooner).

What would I like from you? At present, just read Chapters 0 and 1 and let me know what you think. The other chapters and appendices are mostly skeleton. If you have feedback on the structure, I will also be interested. Be constructive, and honest.

Some questions you may already have which I have prepared answers:

You are just a beginner! What would you know? How can I trust a single thing you say?

Yes, I am just a beginner. That’s why I need your feedback on the content. If you don’t like it, that’s fine, there are plenty of other textbooks. This one probably wasn’t written for you.

You use strange terms I can’t relate to and don’t correspond to other textbooks.

Yes, that is by intention. Read the Introduction. Again, if you don’t like it, move along.

Did you use ChatGPT to write the book?

No. I am not a big fan of modern technology, and AI in particular. I have grave concerns about the ethics and the bias of AI models, as well as the accuracy of any content produced. So, for better or worse, any mistakes or omissions you will find is my fault.

Do you actually have a life? Do you get any sleep?

Surprisingly, yes to both. I am not working on the book every day. Since I started writing, I have attended three birthday parties (including my own) plus other events, I still do my regular cycling (indoors and outdoors), and I have watched about a dozen movies and listened to lots of vinyl records (one of my vices - I bought a few dozen at a record fair last weekend). I try and get at least 8 hours of sleep every night.

Have you attained enlightenment? Or were you a Pali teacher in your past life?

Ha ha, not yet (I think I would know if I did). I can’t recall many past lives, or being a teacher (of any subject). I do have lucid dreams and have partial control over my subconscious, so I get it to learn Pali on my behalf when I sleep. Often I wake up, I have good ideas on what to include in the textbook, so the day is mostly spent typing rather than thinking.


As you are writing a new Pali grammar and relying strongly on Kaccāyana, I would encourage you if haven’t done already to look at Aleix Ruiz Falquez’s article “The Role of Pali Grammar in Burmese Buddhism”.

The article has a good discussion of Kaccāyana and the vutti, along with a discussion of modern day concepts of grammar vs. vyākarana, nirutti, and sadda. (All of which, of course, pre date Kaccāyana.)

The style of the ancient grammars are very different than what we call ‘grammars’ today, it will be interesting to see how you fuse the two.

I’m sure you’re familiar with Collins’s Grammar based on Saddanīti to a certain extent, will yours be a very different style of presentation?

1 Like

Dr. Aleix Ruiz-Falques’s Intro to Kaccāyana (in 6 parts)


Thanks stephen.

At this stage, I am open to how the book will evolve. It started from a desire to learn Pali in Pali, using Pali notions of grammar rather than relying on linguistic terminology.

It was the same with Japanese - I was very confused with English textbooks on Japanese which used linguistic terms. It was not until I started reading Japanese textbooks (and Japanese-Japanese dictionaries) that I learnt to appreciate Japanese on it’s terms and not based on foreign linguistic frameworks.

I am very interested in Saddanīti, but at this stage I am trying to keep the book “pure” by relying on Kaccāyana as my “primary” source of reference. Thitzana says it takes about 1-2 years to be comfortable with Kaccāyana, so I am hoping by this time next year I will start incorporating other sources.

My focus now that I’ve had an initial version of Chapter 1, is to simultaneously go deeper in some of the topics (like the kāraka endings and the taddhita and kita affixes) but also to start focusing on reading the Pali canon. I feel like I have enough grammar under my belt (so to speak) and it’s time to read, read, read.

Mastering the production rules is a distant goal (I’ve now learnt there are so many ways I can embarrass myself by misapplying the rules creating rather rancid results).

A vyākaraṇa is something to be learnt after you know the language really well (at least intermediate or advanced) - and is not learnt before the language is that familar.

I haven’t learnt Kaccāyana, but I have studied Pāṇini (for Sanskrit) - on which Kaccāyana is partly based (and from whom Kaccāyana borrows most of his rules, some nearly verbatim) - and therefore have a very good feel of how the Kaccāyana vyākaraṇa is organized and works.

Therefore my personal opinion is - learning Kaccāyana-vyākaraṇa before being able to read through the canon perhaps is not the optimal way leading to comprehending the Pāli canon - but it will give additional grammatical clarity about the word-forms used in the canon.

But I like the idea of using the native toolkit to study Pali - it is vastly better than Warder et al (who with their Latinesque terminology suited for Greek & Latin university students) unnecessarily complicate the subject in a lot of ways.


Here’s a detailed sūtra by sūtra explanation of (what appears to be) the entire Kaccāyana vyākaraṇa in English - although the accent (Burmese?) of the teacher is unfamiliar to me and therefore I find it a bit hard to follow.


I’m so very impressed by this Pali textbook. Thank you so much!!

1 Like

Agree - Kaccāyana is definitely not for the complete beginner, which is my main motivation for writing my textbook, which reorganises the information and provide additional information to enable a beginner to then tackle the book proper.

I do find its clarity refreshing though, and studying Pali from the lens of a native Pali speaker is infinitely preferable to studying it through the lens of Western linguistics. I found after studying Kaccāyana that I have to unlearn whatever terminology I’ve previously learnt, so I am hoping this approach saves a whole lot of pain for people not familiar with linguistics.

I do hope people will genuinely find my book useful as a bridge that takes someone to a level where they feel comfortable reading the Pali canon (with the aid of an electronic dictionary).

1 Like

Thank you for making your study public. I’ve got Thitzana’s edition of the text, and I recommend it.

1 Like

There are these days prakriyā (word-formation) processors programmed to dynamically identify and interpret Pāṇinian sūtras step by step to show how a word is derived from its fundamental roots.

For example see Vidyullekha for how the form bhavati is derived from the root bhū (you can change the script to IAST if you don’t read Devanāgarī, which it displays as default).

It would be cool to see a similar prakriyā engine for Pāli so we can throw words at it to determine to what extent Kaccāyana can explain their grammatical forms.

1 Like

I have been thinking about this. DPD has a sandhi splitter and compound word deconstructor, one day I’ll have to examine it’s algorithm to see how this is done.

In terms of Kaccāyana’s rules though, it’s not that simple. Although Kaccāyana gives the rules, he does not indicate the circumstances when they are applicable.

It would be possible to create an algorithm that applies all possible rules to a root and see what combinations pop up, but there will be so many it may not be useful. This is particularly the case with the kita and taddhita affixed nouns.

This is what led me to suspect the rules are somewhat flexible, and an educated Pali speaker will have to choose the specific combination of rules that results in a form that does not collide with existing words, and sounds pleasant. I am starting to get a sense of what is pleasant and what’s not, but I simply don’t have enough exposure yet to be able to apply this practically.

What might be possible is a reverse rule lookup “deconstructor” - ie. given a word, determine what rules have been applied to the root to achieve that form. This would be useful, and may possibly lead to a language model that can then generate new forms after being “trained” on existing forms.

In Pāṇini, the elaborate network of “paribhāṣā” meta-rules (and some implied meta-rules that Pāṇini doesnt mention explicitly but borrows implicitly from prior grammarians and therefore takes for granted that the reader knows, and which are clarified by his commentators to make sense of the reason the sūtras are ordered in a specific order) determine which rules take precedence over which other rules and in what circumstances. Many of these implicit paribhāṣās are listed and explained in works like the Paribhāṣenduśekhara of Nāgeśa (also called Nāgoji) bhaṭṭa.

Since Thitzana says Kaccāyana was primarily using the Sanskrit grammars of Sarvavarman (called Kātantra) and that of Pāṇini (the Aṣṭādhyāyī) as his models, perhaps he was following their method implicitly.

Intriguing thought. I can’t discern any specific pattern in the ordering of the suttas, but it is perhaps due to my lack of knowledge. There are sets of suttas where the order of application is important, but there are many suttas where Kaccāyana says “applies to all roots” so again it’s left to the reader to pick and choose which rule to apply.