Introduction

The three texts

This site translates three related grammars of Pāḷi, the works of one school, into modern English.

Text	Author and date	Extent	Translation
Kaccāyana (Kaccāyanabyākaraṇaṃ)	ascribed to Mahākaccāyana; compiled around the sixth or seventh century CE, probably in India	673 suttas in eight chapters, each with its vutti and examples	Kaccāyana
Padarūpasiddhi	Buddhappiya, a Coḻa monk of the twelfth or thirteenth century who calls himself a pupil of Ānanda	684 suttas in seven sections, with Buddhappiya’s own commentary	Padarūpasiddhi
Bālāvatāra	Dhammakitti, Sri Lanka, fourteenth century	269 paragraphs in four sections	Bālāvatāra

Kaccāyana is the earliest Pāḷi grammar that survives: terse rules (sutta) with a short gloss (vutti) and examples, arranged by topic. Buddhappiya’s Padarūpasiddhi takes over practically every sutta of Kaccāyana but replaces the old vutti with his own commentary and rearranges the rules into a teaching order: all the masculine lemmas in a together, with their full inflection table and the derivation of every form, then the other lemmas, and so on through the seven sections, so that a rule may be quoted more than once. It became the standard textbook of the Kaccāyana school. Dhammakitti’s Bālāvatāra abridges the two for beginners and follows the Rūpasiddhi’s arrangement; it stops at the secondary suffixes and has no chapter on the verb or on primary derivation.

The three translations are made in one style and cross-referenced rule by rule. Every Rūpasiddhi sutta and every Bālāvatāra rule carries a card to the Kaccāyana sutta it is; the Kaccāyana headings give the Rūpasiddhi number of each rule in parentheses; and diagrams on the Rūpasiddhi and Bālāvatāra pages map the later texts onto the Kaccāyana chapters.

Status

All three translations are essentially complete (September 2026): Kaccāyana §1–§673, the Padarūpasiddhi Rūp 1–684 and Bālāvatāra ¶1–¶269.

The aim throughout is a literal, word-for-word translation: each sutta and its commentary are rendered as the Pāḷi states them, straightforwardly and accurately, with nothing added by way of explanation and nothing left out. Analysis is given only by exception — where an option marker, a reading of the sutta or a carried word needs a sentence — and commentary is confined to the footnotes; the translation is not a commentary on the text.

The translations were produced through a mix of hand translation, the MITRA-QWEN Pāḷi–English model of the Dharmamitra project, run locally as an independent literal crib, and Claude Fable 5.1 (Anthropic), which reworked the drafts into the house style of the hand-translated chapters and re-read every page against the Pāḷi. Their accuracy has been checked rigorously: every sutta was read against the Pāḷi, cross-checked with the five earlier human translations reproduced on this site, and checked again with both models; the differences that survive are footnoted.

In detail, they have been validated against a range of reference sources:

• the Chaṭṭha Saṅgāyana text itself: a reading is corrected only where the rule requires it, and every correction is footnoted with the edition’s reading;
• the Tipiṭaka: every example and counter-example that could be traced is cited and linked;
• the five earlier English translations reproduced on this site, each cross-checked sutta by sutta (table below), together with the readings of the Thai, Burmese and Sinhalese editions and of the commentaries — the Nyāsa, the Padarūpasiddhi-ṭīkā, the Saddanīti — that they report; where a reading or a construal differs, a footnote records it;
• the studies of Aleix Ruiz-Falqués on optionality, which the Analysis of every sutta follows, and Pāṇini’s Aṣṭādhyāyī where a rule descends from it.

The five reference translations and how they were cross-checked:

Reference translation	Covers	Cross-check
U Nandisena, Kaccāyanabyākaraṇaṃ (2005)	Kaccāyana	done; the footnotes of the Kaccāyana chapters record where the two readings differ
Phramaha Thiab Malai, Kaccāyana-Vyākaraṇa: A Critical Study (1997)	Kaccāyana, with a study of the tradition	done; the footnotes of the Kaccāyana chapters record his edition’s readings and where his translation differs
James D’Alwis, An Introduction to Kachchāyana’s Grammar of the Pāli Language (1863)	the Ākhyātakappa (§406–§523), the earliest English translation of any part of Kaccāyana	done; the footnotes of that chapter cite him, and his text proves to be the Sinhalese recension Senart printed in 1871
Phramaha Sriporn Rachiwong, Rūpasiddhi: A Study of Some Aspects (1995)	the Padarūpasiddhi’s Nāmakaṇḍa, Samāsakaṇḍa and Ākhyātakaṇḍa	done; the footnotes of those three sections record her Thai, Burmese and Sinhalese readings and where her construal differs, and a “Reading Rachiwong” note heads each
Vidyabhusana, Punnananda Swami and Mitra, Bālāvatāra (1935)	Bālāvatāra	done; the footnotes of the Bālāvatāra chapters record where the 1935 translation differs

Corrections and suggestions are welcome on GitHub.

Approach

Optionality

Kaccāyana hedges its rules with option markers, and a translation stands or falls by how it reads them. Following Aleix Ruiz-Falqués’ studies of the question, Two levels of optionality in the Kaccāyana Vyākaraṇa and Boundaries and Domains, the translation distinguishes two levels of option:

• vā marks either an open option (vikappa), where both forms are good Pāḷi, or a restricted option (vavatthitavibhāsā), where the tradition fixes the domain in which the rule is obligatory (nicca), optional (anicca) or excluded (asanta);
• kvaci marks an exception to a wider rule: the form occurs only in the words the examples name;
• navā is glossed by the Rūpasiddhi as kvaci, and vibhāsā as vā; ca, api, tu and niccaṃ control options too.

In the sutta headings vā is rendered “optionally”, kvaci “sometimes” or “occasionally” and navā “rarely”; the Analysis under every sutta then states how the marker is read and, where Buddhappiya or the commentaries fix a domain, what that domain is.

Terminology

The grammar is described in its own terms rather than through the categories of Latin grammar. An inflection is a vibhatti, and its seven forms are numbered ① to ⑦ in the traditional order (paṭhamā to sattamī), with ⓪ for the vocative, ⨀ for the singular and ⨂ for the plural; the words “case”, “declension” and “indeclinable” are not used, and aliṅga is glossed as “without gender”. In the glosses of the examples gender is marked 🚹 🚺 🚻 and person 👆 🤘 🤟. Lemmas are liṅga, suffixes paccaya, roots dhātu; Pāḷi technical terms are kept, set in backticks, and glossed on first use.

Examples, derivations and diagrams

Every example the vutti gives is derived step by step, with + for the parts joined, → for a substitution and ~~x~~ for an elision, each step naming the rule that makes it; where an example is a quotation from the Tipiṭaka it is cited and linked to the World Tipiṭaka Edition. Counter-examples are explained as well as translated. Diagrams are used for structure only, for the kinds of rule, the map of each text, the sandhi decision charts and the taxonomies of compounds, suffixes and model nouns, never for an individual derivation, which the notation already handles.

Sources

The Pāḷi of all three texts is the Chaṭṭha Saṅgāyana Tipiṭaka edition (CST4) published by the Vipassana Research Institute at tipitaka.org, also available on GitHub: Kaccāyana with its vutti, Buddhappiya’s Padarūpasiddhi and Bālāvatāra, whose complete text is reproduced on the Bālāvatāra (CSCD4) page. The edition has misprints, and its Bālāvatāra and Rūpasiddhi files have a few paragraphs and numbers out of place; a reading is corrected where the rule itself requires it, every correction is footnoted with the edition’s reading, and the suttas are numbered by their position in the text.

The examples and counter-examples are usually taken from the Tipiṭaka and its commentaries. Where they could be identified they are cited from the World Tipiṭaka Edition; some could not be found and may come from texts outside the canon.

Kaccāyana in context

Kaccāyana is the oldest extant Pāḷi grammar, and was composed around the sixth century CE, most probably in India. It is partly based on Pāṇini and Kātantra:

• Pāṇini, Aṣṭādhyāyī “Eight Lessons” 5th–4th c. B.C.E. Śalātura Northwest India
• Śarvavarman, Kātantravyākaraṇa 2nd c. C.E. South India

The author of Kaccāyana is attributed to Mahā Kaccāyana, but modern scholars believe it is authored by multiple compilers and gradually evolved over time. According to James D’Alwis in An introduction to Kachchāyana’s grammar of the Pāli language:

• the grammatical aphorisms (sutta), are ascribed to a certain Kaccāyana,
• the “gloss” or explanation (vutti) is ascribed to a certain Saṅghanandin
• the examples (payoga) is ascribed to a certain Brahmadatta

According to the Thai monk Dhammakitti in Saddhammasaṅgaha (Saddhamma-s IX, 18-20, 35):

kaccāyano kato yogo saṃghanandi pavuttikā, ṭīkā vimalabodhī ca brahmaputto ca kārakā.
(SADDHĀNANDA, 1890: 63)
Kaccāyana composed the yoga [“rule”], Saṅghanandin the pavuttikā [“notes”], and Vimalabodhī the ṭīkā [“commentary”], and Brahmaputto the kāraka [“section on factor of action”].

The structure of Kaccāyana

There are 673 rules (sutta), numbered §1–§673 throughout this site, spread across the following chapters (kappa) and sections (kaṇḍa). The Chaṭṭha Saṅgāyana edition counts the Kāraka, Samāsa and Taddhita chapters as the sixth to eighth sections of the noun chapter, and the Uṇādi chapter as the sixth section of the Kibbidhāna chapter.

1. Sandhikappa (§1–§51)
   1. Definitions (§1–§11)
   2. Vowel sandhi (§12–§22)
   3. Consonant sandhi (§23–§29)
   4. Niggahīta (ṃ) sandhi (§30–§41)
   5. Miscellanea (§42–§51)
2. Nāmakappa (§52–§270)
   1. Definitions, internal morphology, default inflections (§52–§119)
   2. amha and tumha, the ntu lemmas, numerals, rāja, puma, kamma (§120–§160)
   3. The pronouns, the mano group, gacchanta, brahma, atta, sakha, rāja, satthu, pitu (§161–§210)
   4. atta, the neuter, the feminine lemmas, kiṃ, ima and eta, the feminine suffixes, the vocative (§211–§246)
   5. The suffixes that stand for inflections (to, tra, dā …) and the comparatives (§247–§270)
3. Kārakakappa (§271–§315): the seven inflection forms and the relations they express
4. Samāsakappa (§316–§343): compounds
5. Taddhitakappa (§344–§405): nouns made from nouns
6. Ākhyātakappa (§406–§523)
   1. The endings and their times (§406–§431)
   2. Derived verbs, the passive, the eight classes of root (§432–§457)
   3. Reduplication, special lemmas (§458–§481)
   4. Special suffixes and lemmas (§482–§523)
7. Kibbidhānakappa (§524–§623)
   1. The kicca and kita suffixes (§524–§549)
   2. The suffixes by time: the past, the infinitive and absolutive, the participles (§550–§570)
   3. The shapes of ta after each root (§571–§589)
   4. Changes of the root before ṇ, tave and tuna (§590–§606)
   5. Further changes of root and suffix (§607–§623)
8. Uṇādikappa (§624–§673): the remaining primary suffixes

The structure of a sutta

Each sutta (rule, or grammatical aphorism) is described in a series of subsections or “fields”. The vutti is always present, the other fields may be present.

field	function	tags
vutti	gloss, or explanation	repeats and expands the words from the sutta
udāharaṇa	examples	taṃyathā
kimudāharaṇa	counter-examples	kasmā / kimatthaṃ
payojana	application/purpose/scope of the rule (“where is it used”)	kvattho / kimpayojanaṃ
gahaṇaphala	result of mentioned words, comments on anuvṛtti, meaning of ca or vā, etc.	(xxx)(sadda)ggahaṇena… eg. rule 35: casaddaggahaṇena iheva makārassa pakāro hoti.

Example (the first number is the Kaccāyana rule number, the second is the equivalent Padarūpasiddhi rule number):

151, 250.Bahuvacanesu vono.
Sabbesaṃ tumhaamhasaddānaṃ savibhattīnaṃ yadā padasmā paresaṃ vonoādesā honti yathāsaṅkhyaṃ tatiyābahuvacanesu paresu. Kataṃ vo kammaṃ, kataṃ no kammaṃ. Padatoti kimatthaṃ? Tumhehi kataṃ, amhehi kataṃ. Bahuvacanaggahaṇena yomhi paṭhame vo noādesā honti. Gāmaṃ vo gaccheyyātha, gāmaṃ no gaccheyyāma.

field	text
no	151
Padarūpasiddhi	250
sutta	Bahuvacanesu vono.
vutti	Sabbesaṃ tumhaamhasaddānaṃ savibhattīnaṃ yadā padasmā paresaṃ vonoādesā honti yathāsaṅkhyaṃ tatiyābahuvacanesu paresu.
udāharaṇa	Kataṃ vo kammaṃ, kataṃ no kammaṃ.
kimudāharaṇa	Padatoti kimatthaṃ? Tumhehi kataṃ, amhehi kataṃ.
payojana
gahaṇaphala	Bahuvacanaggahaṇena yomhi paṭhame vo noādesā honti. Gāmaṃ vo gaccheyyātha, gāmaṃ no gaccheyyāma.

Other examples:

361, 388.Ṇavisamādīhi.
Ṇapaccayo hoti visamādīhi ‘‘tassa bhāvo’’iccetasmiṃ atthe. Visamassa bhāvo vesamaṃ, sucissa bhāvo socaṃ.

field	text
no	361
Padarūpasiddhi	388
sutta	Ṇavisamādīhi.
vutti	Ṇapaccayo hoti visamādīhi ‘‘tassa bhāvo’’iccetasmiṃ atthe.
udāharaṇa	Visamassa bhāvo vesamaṃ, sucissa bhāvo socaṃ.
kimudāharaṇa
payojana
gahaṇaphala

364, 398.Tadassatthīti vī ca.
“Tadassatthi”iccetasmiṃ atthe vīpaccayo hoti. Medhā yassa atthi, tasmiṃ vā vijjatītimedhāvī. Evaṃ māyāvī. Caggahaṇena sopaccayo hoti. Sumedhā yassa atthi, tasmiṃ vā vijjatīti sumedhaso.

field	text
no	364
Padarūpasiddhi	398
sutta	Tadassatthīti vī ca.
vutti	“Tadassatthi”iccetasmiṃ atthe vīpaccayo hoti.
udāharaṇa	Medhā yassa atthi, tasmiṃ vā vijjatītimedhāvī. Evaṃ māyāvī.
kimudāharaṇa
payojana
gahaṇaphala	Caggahaṇena sopaccayo hoti. Sumedhā yassa atthi, tasmiṃ vā vijjatīti sumedhaso.

59, 182.Te itthikhyā po.
Te ivaṇṇuvaṇṇā yadā itthikhyā, tadā pasaññā honti. Rattiyā, itthiyā, dhenuyā, vadhuyā. Itthikhyāti kimatthaṃ? Isinā, bhikkhunā. Saiccanena kvattho? Pato yā.

field	text
no	59
Padarūpasiddhi	182
sutta	Te itthikhyā po.
vutti	Te ivaṇṇuvaṇṇā yadā itthikhyā, tadā pasaññā honti.
udāharaṇa	Rattiyā, itthiyā, dhenuyā, vadhuyā.
kimudāharaṇa
payojana	Saiccanena kvattho? Pato yā.
gahaṇaphala

57, 71.Ālapane si ga sañño.
Ālapanatthe si gasañño hoti. Bhoti ayye, bhoti kaññe, bhoti kharādiye. Ālapaneti kimatthaṃ? Sā ayyā. Sīti kimatthaṃ? Bhotiyo ayyāyo. Gaiccanena kvattho? Ghate ca.

field	text
no	57
Padarūpasiddhi	71
sutta	Ālapane si ga sañño.
vutti	Ālapanatthe si gasañño hoti.
udāharaṇa	Bhoti ayye, bhoti kaññe, bhoti kharādiye.
kimudāharaṇa	Ālapaneti kimatthaṃ? Sā ayyā. Sīti kimatthaṃ? Bhotiyo ayyāyo.
payojana	Gaiccanena kvattho? Ghate ca.
gahaṇaphala

Types of suttas

The rules can be classified into four categories, and additional subcategories, with examples given below:

• saññā (technical term)
  • anvattha (meaningful)

    8, 10.Aṃiti niggahitaṃ. anvattha
    Aṃ iti niggahitaṃ nāma hoti.
    Tena kvattho? Aṃ byañjane niggahitaṃ.

  • rūḷhī (artificial)

    58, 29.Ivaṇṇuvaṇṇā jhalā. rūḷhī
    Ivaṇṇuvaṇṇāiccete jhalasaññā honti yathāsaṅkhyaṃ.
    Isino , aggino, gahapatino, daṇḍino. Setuno, ketuno, bhikkhuno. Sayambhuno, abhibhuno.
    Jhalaiccanena kvattho? Jhalato sassa no vā.

• adhikāra (domain)
  • sīhagatika (comprehensive) \

    52, 60.Jinavacanayuttaṃhi. sīhagatika “Jinavacanayuttaṃ hi” iccetaṃ adhikāratthaṃ veditabbaṃ.

  • maṇḍūkagatika (selective)

    131, 0.Itthipumanapuṃsakasaṅkhyaṃ. maṇḍūkagatika
    “Itthipumanapuṃsakasaṅkhyaṃ” iccetaṃ adhikāratthaṃ veditabbaṃ.

  • yathānupubbika (progressive)

    413, 427.Kāle. yathānupubbika “Kāle” iccetaṃ adhikāratthaṃ veditabbaṃ.

• paribhāsā (metarule)
  • saññaṅga (definition metarule)

    9, 11.Parasamaññā payoge. saññaṅga
    Yā ca pana paresu sakkataganthesu samaññā ghosāti vā aghosāti vā, tā payoge sati etthāpi yujjante.
    Tattha ghosā nāma-ga gha ṅa, ja jha ña, ḍa ḍha ṇa, da dha na, ba bha ma, ya ra la va ha ḷa, iti ghosā nāma. Aghosā nāma-ka kha, ca cha, ṭa ṭha, ta tha, pa pha, sa, iti aghosā nāma.
    Tena kvattho? Vagge ghosāghosānaṃ tatiyapaṭhamā.

  • vidhyaṅga (operational metarule)

    10, 12.Pubbamadhoṭhita massaraṃ sarena viyojaye. vidhyaṅga
    Tattha sandhiṃ kattukāmo pubbabyañjanaṃ adhoṭhitaṃ assaraṃ katvā sarañca upari katvā sarena viyojaye. Tatrāyamādi.

  • aññaṅga (other metarule)

    54, 62.Tato ca vibhattiyo. aññaṅga
    Tato jinavacanayuttehi liṅgehi vibhattiyo parā honti.

• vidhi (operation)
  • utsarga (general rule)

    12, 13.Sarā sare lopaṃ. utsarga
    Sarā kho sare pare lopaṃ papponti.
    Yassindriyāni samathaṅgatāni. No hetaṃ bhante sametāyasmā saṅghena.

  • apavāda (exception)

    13, 15.Vā paro asarūpā. apavāda
    Saramhā asarūpā paro saro lopaṃ pappoti vā.
    Cattāro’me bhikkhave dhammā, kinnu’ māvasamaṇiyo. Vāti kasmā? Pañcindriyāni, tayassu dhammā jahitā bhavanti.

According to Aleix Ruiz-Falqués in A Short Introduction to the Pāli Kaccāyana Grammar:

Several saññā “technical terms”,
within an adhikāra “domain”,
following paribhāsā “metarules”,
will be subject to vidhi “operations”
in order to derive words or sentences.

Megarules

mahāsutta or “megarules” are four rules that are invoked in order to derive forms that cannot be derived otherwise.

• 391, 423.Yadanupapannā nipātanā sijjhanti.
  Ye saddā aniddiṭṭhalakkhaṇā, akkharapadabyañjanato, itthipumanapuṃsakaliṅgato , nāmupasagganipātato, abyayībhāvasamāsataddhitākhyātato, gaṇanasaṅkhyākālakārakappayogasaññāto, sandhipakativuddhilopāgamavikāraviparitato, vibhattivibhajanato ca, te nipātanā sijjhanti.
  “Those [words] that are not formed by the suttas, should be formed as irregular forms.”
• 403, 354.Kvacādimajjhuttarānaṃ dīgharassā paccayesu ca.
  Kvaci ādimajjhauttaraiccetesaṃ dīgharassā honti paccayesu ca apaccayesu ca.
  “Sometimes, there is lengthening or shortening at the beginning, middle or end [vowels of a word], even before affixes.”
• 404, 370.Tesu vuddhilopāgamavikāraviparītādesā ca.
  Tesu ādimajjhuttaresu yathājinavacanānuparodhena kvaci vuddhi hoti, kvaci lopo hoti, kvaci āgamo hoti, kvaci vikāro hoti, kvaci viparīto hoti, kvaci ādeso hoti.
  “And in those cases [cf. rule 405], [sometimes,] there is full grade, elision, augment, modification, change, or replacement.”
• 517, 488.Kvaci dhātuvibhattipaccayānaṃ dīgha viparītādesalopāgamā ca.
  Idha ākhyāte aniddiṭṭhesu sādhanesu kvaci dhātuvibhattipaccayānaṃ dīghaviparītādesalopāgamaiccetāni kāriyāni jinavacanānurūpāni kātabbāni.
  “Sometimes [for verbal forms that cannot be derived otherwise, there is] lengthening, change, replacement or augment of the verbal root, the personal ending, or the affix(es).”

Rule 52 is halfway between a megarule and a normal rule:

52, 60.Jinavacanayuttaṃhi.
“Jinavacanayuttaṃ hi” iccetaṃ adhikāratthaṃ veditabbaṃ.

Ghost metarules

“Ghost Metarules” are paribhāsās that are consistently invoked in the commentaries, but that you will never see in the sūtra treatise.

Examples:

• ekamhi vattabbe ekavacanaṃ bahumhi vattabbe bahuvacanan [e.g. Mmd 28,15–16]
  “The singular is used when one [object] is to be stated, the plural is used when many [objects] are to be stated.”
  [NOTE: in Pāṇini, 1.4.21 bahuṣu bahuvacanam (“plural for many [objects]”) & 1.4.22 dvyekayor dvivacanaikavacane (“dual and singular for two or one [objects respectively]”) are considered saṃjñā (“definition”) rules.]
• vatticchānupubbikā saddapaṭipatti [e.g. Mmd 64,17–18]
  “The meaning of a word is determined by the intention of the speaker.”
• vuttatthānam appayogo [23 times in Mmd, e.g. 133,20]
  “Words whose meaning has already been stated are not used.”
• yogavibhāgā iṭṭhappasiddhi [Mmd, e.g. 252,29]
  “After the splitting up (vibhāgā) of the sutta (yoga) [there is] the expected result [only].”

Usage of rūḷhī

In a sutta, a rūḷhī (artificial technical term) carries the inflection form that matches the role of the term within the rule. For example jha (used for masculine/neuter nominal bases ending with i or ī):

form	jha	usage
①	jho	i/ī (*replacement)
②	jhaṃ	i/ī (substitute)
③	jhena	not used
④	jhassa	not used
⑤	jhamhā/jhato	“after i/ī”
⑥	jhassa	“instead of i/ī”
⑦	jhasmiṃ	“before i/ī”

Types of vidhi (operation)

• ādesa “replacement”
• āgama “insertion”
• vikāra “modification”
• lopa “elision”
• viparīta “metathesis”

Structure of a vidhi (operation)

An operation converts from into to surrounded by before and after, in the scope of option and context.

option	before	from	operation	to	after	context

As an illustration, the sequence AXB is replaced with AYB:

option	before	from	operation	to	after	context
	A	X	becomes	Y	B

Option markers:

• ca
• api
• vā
• kvaci
• na
• niccaṃ
• …

Context: “in the sense of …”

Vibhatti endings are used to determine the various roles in the rule

option	before	from	operation	to	after	context
	⑤	⑥	ādesa	①	⑦	⑦
	⑤	①	ādesa	①	⑦	⑦
	⑤	①	ādesa	②	⑦	⑦

Example

§103 ato n’ ena: (After a, nā [becomes] ena):

option	before	from	operation	to	after	context
	a	nā	ādesa	ena
	⑤	①	ādesa	①

	buddha + nā	rule
→	buddha + (nā→ena)	§103
→	buddhena

Example

§70: jhalānam iyuvā sare vā: (Optionally, before a vowel: iya replaces i/ī, uva replaces u/ū):

option	before	from	operation	to	after	context
vā		i/ī	ādesa	iva	sare
vā		u/ū	ādesa	uva	sare
vā		⑥	ādesa	①	⑦

	bhikkhu + āsane	rule
→	bhikkh + (u→uva) + (āsane)	§70
→	bhikkh + uv + (a + ā) + sane	§12
→	bhikkhuvāsane

Example

§12 sarā sare lopaṃ: (Vowel, before vowel, is elided):

option	before	from	operation	to	after	context
		sarā	lopaṃ	sara	sare
		①	lopaṃ		⑦

	tena + upasaṅkami	rule
→	ten(a) + upasaṅkami	§12
→	tenupasaṅkami

Most vidhi “operational” rules teach paccaya “affixation” or ādesa “replacement” using previous definitions (saññā-s). It is all about adding particles to a base, and then replacing them if/when needed. To understand vidhi-s, we should keep in mind the following code of inflection forms:

• ⑤ - “after X”
• ⑥ - “instead of X”
• ① - “X”, ādesa
• ⑦ - “before X”
• ⑦ - “in the sense of X”

Motivation

I set out to translate Kaccāyana anew, and then the two grammars of its school that grew from it, for several reasons:

• I am not entirely satisfied with the existing translations. Some of them seem to lack clarity (primarily due to the writing style). Others don’t explain the examples well enough, and yet others are too verbose, inserting explanations not in the original text.
• There are some spelling errors and mistakes in the edition published on Chaṭṭha Saṅgāyana Tipiṭaka, some of which have been incorporated into existing translations. I have taken the opportunity to correct the mistakes, uncovered by scholars such as O. H. Pind and A. Ruiz-Falqués, and to footnote every correction with the edition’s reading.
• Unlike some translations, I am honouring the use of option markers in Kaccāyana, and reading them as the tradition read them, after Ruiz-Falqués’ two levels of optionality (see Optionality above).
• I describe the grammar in its own terms, with inflection forms rather than cases, and cross-reference the use of technical terms across the three classical Pāḷi grammars translated here.
• Translating the Rūpasiddhi and Bālāvatāra alongside Kaccāyana shows how the school itself taught the rules: Buddhappiya’s inflection tables and derivations make explicit what Kaccāyana leaves to the reader, and Dhammakitti’s abridgement shows what a beginner was expected to know.
• It is a good way to further my understanding of Pāḷi.
• It is also a good opportunity for me to introduce the Brāhmī script. As we know, there are no surviving Buddhist texts written in Brāhmī, so there is no need to learn this script, however, I like it due to its elegance and simplicity.
• I wanted to augment the translations with diagrams, and incorporate worked through grammatical analysis and rule derivations in the examples.
• In comparison with the other classical Pāḷi grammars, Kaccāyana is brief and succinct, minimising repetition. It was used as a standard textbook for teaching Pāḷi in countries such as Burma for many years, and often used in various monasteries as a Pāḷi reference text.
• I prefer Kaccāyana to most English Pāḷi textbooks and grammar books, as it describes Pāḷi without the distorting lens of Western linguistics.
• My translation is presented using a modern responsive web design, featuring Google’s Noto Sans font.

References

• James D’Alwis, An introduction to Kachchāyana’s grammar of the Pāli language, with an introduction, appendix, notes, etc., Williams and Norgate, London (1863); reproduced as D’Alwis: Kaccāyana (1863)
• Buddhappiya, Padarūpasiddhi, Chaṭṭha Saṅgāyana Tipiṭaka edition, Vipassana Research Institute; translated on this site as Padarūpasiddhi
• Dhammakitti, Dr. Satischandra Vidyabhusana; Punnananda Swami; Sailendranath Mitra, Bālāvatārā - an Elementary Pāli Grammar Abridged for the Undergraduate Course, University of Calcutta (1935); reproduced as Bālāvatāra (1935 Translation)
• U Nandisena, Kaccāyanabyākaraṇaṃ, translated into English, Myanmar (2005); first electronic edition, Instituto de Estudios Buddhistas Hispano (2017); reproduced as Nandisena: Kaccāyana (2005)
• Phramaha Thiab Malai, Kaccāyana-Vyākaraṇa: A Critical Study, Ph.D. thesis, University of Pune (1997), two volumes; reproduced as Thiab Malai (1997)
• Aleix Ruiz-Falqués, On the Authorship of Kaccāyana, the Oldest Pāli Grammar, Pariyatti: Studies In Pāli Language And Literature, Aditya Prakashan, New Delhi (2017); reproduced here
• Aleix Ruiz-Falqués, A Short Introduction to the Pāli Kaccāyana Grammar, Shan State Buddhist University, Taunggyi (Myanmar) (2020)
• Aleix Ruiz-Falqués, “Two levels of optionality in the Kaccāyana Vyākaraṇa”, in Peter M. Scharf (ed.), Śabdānugamaḥ: Indian linguistic studies in honor of George Cardona, vol. I, The Sanskrit Library, Providence (2021), pp. 431–465; reproduced here
• Aleix Ruiz-Falqués, “Boundaries and Domains: Understanding Optionality in Buddhappiya’s Rūpasiddhi”, Journal of the Pali Text Society XXXIV (2021), pp. 227–250; reproduced here
• Mahesh A. Deokar, Technical Terms and Technique of The Pali and Sanskrit Grammars, Central Institute of Higher Tibetan Studies, Sarnath, Varanasi (2008)

Further articles and reviews by Aleix Ruiz-Falqués, and the other editions and translations consulted, are listed on the Other editions and translations page.