An Introduction to Pāḷi

Note

The following is a concise overview of the Pāḷi language covering key features. It is an abridged version of the introductory chapter from my upcoming Pāḷi textbook.

Note

The information in this introduction is heavily based on “Kaccāyana” - the earliest known Pāḷi Grammar, traditionally associated with the Buddha’s disciple Mahā-Kaccāyana (6th-7th century C.E.). All references to Kaccāyana are denoted as numbers in curly braces, e.g. {8} (representing the 8th sutta or “grammatical rule”). Altogether there are 675 suttas, which shows how compact Kaccāyana is.

For further information please view my [work in progress] translation of Kaccāyana into English at kaccayana.github.io.

Background

According to (Norman, 1988):

Pāli is a dialect of Middle Indo-Aryan (MIA), i.e. one of the dialects which lie chronologically between Old Indo-Aryan (=Sanskrit) and New Indo-Aryan (=the modern languages of North India and Sinhalese). The Indo-Aryan languages of India belong to the Indo-European family of languages, and the name Indo-Aryan stands for “the Indo-European languages of India”.

Norman further explains that the multiple dialects of MIA are due to successive migrations of Indo-Aryans into India. This occurred over several waves and multiple generations. They met with the indigenous inhabitants of North India, and in particular the inhabitants of the Indus Valley, and over time assimilated part of their culture and languages, such as Dravidian and other languages. Each migration wave would have resulted in a different MIA dialect resulting from different amounts of Dravidian loanwords mixed in with them and also a different mix of cultural and religious backgrounds between the invaders and the indigenous inhabitants. Over time these MIA dialects also evolved away from Sanskrit, developed unique morphological and phonological characteristics and became referred to as “Prakrit” languages.

By the time of the Buddha, there were at least three MIA dialects in use in the Greater Magadha region and beyond: a “Western” dialect, an “Eastern” dialect and a “North Western” dialect (Gāndhārī). The Buddha may have known all three dialects, and they were very similar to each other. (Karpik, 2019) regards them as “mutually intelligible” by native speakers of any of the dialects, although this is contested by (Levman, 2019).

According to (Levman, 2014), the Sakya tribe which the Buddha belonged to had a separate socio-political organisation, religious and cultural values from the Indo-Aryans, so it is unclear what language the Buddha would have spoken or taught in. It is also unclear to what extent the Buddha was exposed to Brahmanism and therefore his knowledge of Sanskrit would have been questionable. He may have used a mixture of MIA dialects and indigenous languages, depending on his audience.

However, given the Khandhaka shows the Buddha interacting with royal and wealthy converts and benefactors, the Buddha may probably have spoken in the Western MIA dialect, which is regarded as more prestigious. (Karpik, 2019) In any case, what we know of as Pāḷi today would seem to be closest to the Western dialect, but mixed in with words and inflection forms from the other dialects.

There are debates about the origin of Pāḷi, whether it was in fact a real language actually spoken by people, as opposed to an “invented” or “ecclesiastical” language specifically for the purposes of recording Buddhist doctrine. (Levman, 2016) wrote:

As is well known, Buddhaghosa equated Pāli (P) with Māgadhī but we know that Pāli is a composite dialect, and although it contains elements of what is probably an eastern dialect that the Buddha may have spoken, it is nevertheless not an “original language of Buddhism” but a translation of something earlier. It is usually characterized as a western dialect, but in fact, if closely analysed, it contains elements of both eastern, western, and northwestern dialects – it is a mixed language created by monks, normalized for religious purposes.

Pāḷi also contains a broad spectrum of loanwords from autochthonous languages such as Dravidian and Munda. This is especially evident with respect to plants, animals, customs and practices, slang words, and proper names which were foreign to the IA immigrants. Even some inflection forms (such as the absolutive or -tvā suffix) have been considered to be Dravidian in origin and imported into Sanskrit and MIA dialects. (Levman, 2014)

(Levman, 2014) adds that many grammatical features of Pāḷi and MIA dialects reflect conventions from autochthonous languages:

The above is only a sampling of some of the more conspicuous features of the Indic linguistic area; there are many others: causative verb structure, basic subject-object-verb word order; postpositions; goals of verb of motion, adverbial and infinitive complements which go in the object position; adjectives, genitive phrases, demonstratives and numerals which precede the noun they modify; qualifiers which precede adjectives; use of genitive for the verb ‘to have’; use of dative to express internal states of mind; caste system terminology similarities, and more.

Although Pāḷi is most similar to the “Western” MIA dialect, there are signs that it has been transformed and “Sanskritised” over the years (by well-meaning Sanskrit-trained editors and scribes no doubt trying to “correct” what they perceived to be spelling and grammatical “mistakes”) so it is no longer a pure version of the dialect that the Buddha would have spoken, and it is at least partially an “artificial” language. The morphology and grammar of Pāḷi were formalised by various grammarians in the 6th-12th centuries (with Kaccāyana being the first such work) and it is likely the texts were altered to conform to the “normative” grammar as described by grammarians.

As (Norman, 1997) points out:

… if we set out to understand what the earliest texts say, i.e. those ascribed to the Buddha himself, or his followers during his lifetime, we have to consider the fact that the language which we find in such texts is not necessarily, and almost certainly is not, the language of the Buddha himself, i.e. the language has been changed both synchronically - it has been translated or transformed into other languages as the need arose, perhaps as Buddhism spread into neighbouring areas - and also diachronically, i.e. as the language of the readers or recensionists developed in the course of time, this had an effect upon the language of the texts.

(Norman, 1988) adds:

The early works of Buddhism had been written in various Prakrits, but the growing prestige of Sanskrit led to an attempt to re-write these texts in Sanskrit. The success of such a translation process varied from sect to sect, and we find a variety of Sanskritised Prakrits which are generally referred to as Buddhist Hybrid Sanskrit, although it must be made clear that there is no one form of that language. The writing down of the Pāli canon in the first century B.C. probably put an end to the wholesale restoration of Sanskrit forms into the language of the Theravādin canon.

More recently, (Gombrich, 2018) suggests “Pāli reflects the idiosyncratic language used by the Buddha as he toured northeast India”. (Karpik, 2019) offers a similar hypothesis that Pāḷi was in fact originally the Western MIA dialect, which was actually spoken by the Buddha. (Levman, 2019) disagrees, and argues what the Buddha spoke may have been a “koine”, or an inter-dialect “lingua franca” that was commonly used for administration, trade and commerce. (Levman, 2016) This view has also been previously articulated by (Cousins, 2013). (Hinüber, 1996) raises the interesting possibility that the “lingua franca” may have been created after the Buddha died, by compilers of his teachings in order to standardise the texts. Regardless, Pāḷi texts probably represent the closest we will get to what might have been the Buddha’s original words.

The term “Pāḷi” actually just means “text”, and today “Pāḷi” is used primarily for Buddhist literature.

Pāḷi, like most other Indo-Aryan languages, is an oral language spoken as vākya (sentences) made up of vacana (words) which are a combination of akkhara (“sounds”, or “units from a writing system”). {1}

Pāḷi Writing System and pronunciation {2}-{9}

The Buddha may have spoken a language similar to Pāḷi, and when he died his sayings were collected and systematised so they could be transmitted orally generation to generation using memorisation techniques such as chanting. The Pāḷi canon was not written down until hundreds of years after the Buddha had passed away.

As such, there is no “official” Pāḷi writing system. Pāḷi can be transcribed fairly faithfully in a number of scripts, including Roman letters (with the addition of a few diacritical marks or accented characters similar to other European languages).

The Pāḷi Abugida

Strictly speaking, Pāḷi does not use an “alphabet”; it uses an “abugida”.

Note

An “abugida” is a segmental writing system in which consonant–vowel sequences are written as units; each unit is based on a consonant letter, and vowel notation is secondary, similar to a diacritical mark. This contrasts with a full alphabet, in which vowels have status equal to consonants.

Pāḷi is generally written in various Brahmic scripts (also known as Indic scripts), which are abugida writing systems. Brahmic scripts are used throughout the Indian subcontinent, Southeast Asia and parts of East Asia. They are descended from the Brahmi script of ancient India. Brahmi is clearly attested from the 3rd century BCE during the reign of Aśoka, who used the script for imperial edicts.

In Roman script, Pāḷi consists of 8 vowels (sara) and 33 consonants (vyañjana), arranged in the following order (the letters in round brackets are not commonly encountered):

sara: a, ā, i, ī, u, ū
kaṇṭhatāluja: e
kaṇṭhoṭṭhaja: o
niggahīta: (ṃ¹)
kavagga: k, kh, g, gh, (ṅ)
cavagga: c, ch, j, jh, ñ
ṭavagga: ṭ, ṭh, ḍ, (ḍh), (ṇ)
tavagga: t, th, d, dh, n
pavagga: p, ph, b, bh, m
antaṭṭha: y, r, l, (ḷ), (ḷh)
dantoṭṭhaja: v
sakāra: s
hakāra: h \

Groups

Articulation Place	rassa	dīgha	kaṇṭhatāluja kaṇṭhoṭṭhaja	anunāsika	aghosa sithila	aghosa dhanita	ghosa sithila	ghosa dhanita	nāsika	antaṭṭha	hakāra	sakāra
kaṇṭhaja	a	ā			k	kh	g	gh	ṅ		h
tāluja	i	ī	e		c	ch	j	jh	ñ	y
muddhaja			o	ṃ	ṭ	ṭh	ḍ	ḍh	ṇ	r ḷ
dantaja					t	th	d	dh	n	v l		s
oṭṭhaja	u	ū			p	ph	b	bh	m

How to form sounds

vagga	Group	Sound
kaṇṭhaja	Gutturals	pronounced in the throat
tāluja	Palatals	pressing the tongue on the front-palate
muddhaja	Linguals (Cerebrals, Retroflex)	bringing the up-turned tip of the tongue in contact with the back of the palate
dantaja	Dentals	pronounced with the aid of the teeth
oṭṭhaja	Labials	formed by means of the lips
hakāra	Spirant	a strong aspirated breathing
sakāra	Sibilant	has a hissing sound
anunāsika	Nasal	nasal breathing found only after the short vowels: aṃ, iṃ, uṃ
aghosa	Voiceless (Surds, stops)	hard, flat, and toneless
ghosa	Voiced (Sonants)	soft and uttered with a checked tone
dhanita	Aspirates	pronounced with a strong breathing or h sound added to them
sithila	Non-aspirates	pronounced naturally, without effort and without the h sound
nāsika	Nasals	sounded through the nose
antaṭṭha	Liquids (Resonants, semi-vowels)	readily combine with other consonants (except, perhaps, ḷ)
-	Mutes (Non-surds)	not being readily pronounced without the aid of a vowel

Pronunciation

No one knows for sure how Pāḷi words would have been pronounced, and how Pāḷi sentences would have been spoken, particularly with regard to tonality. Modern-day Pāḷi pronunciation is based on how the Buddhist Pāḷi canon is chanted in various countries including Sri Lanka, Myanmar and Thailand.

Letter	Pronunciation
a	but, cut
ā	art, father
i	mill, pin
ī	bee, machine
u	put
ū	cool, rule
e	ten, fate
o	hot, note
k	key, kite
g	get, good
ṅ	ring, singer
c	church, rich
j	jam, jug
ñ	signor
ṭ	hat, not
ḍ	good, hid
ṇ	hint, now
t	thumb
d	they
n	now
p	lip, put
b	but, rib
m	him, mind
y	yard, yes
r	rat, right
l	light, sell
v	vile, vine
s	sing, sit
h	hot, hut
ḷ	felt, light
ṃ	sing

Roman Script Transliteration vs IPA Transcription

The International Phonetic Alphabet for Pāḷi (IPA Pāḷi) 2008 has been proposed by Emeritus Professor Vichin Phanupong, which is used for audio versions of the World Tipiṭaka Edition 2009.

Manner of Articulation (Karaṇa)	Voiceless Stop (Aghosa) Unaspirated (Sithila)	Voiceless Stop (Aghosa)Aspirated (Dhanita)	Voiced Stop (Ghosa) Unaspirated (Sithila)	Voiced Stop (Ghosa) Aspirated (Dhanita)	Nasal Stop	Approximant Non-lateral	Approximant Lateral	Fricative
Place of Articulation (Ṭhāna)
Glottal (Kaṇṭhaja)								h [h]
Velar (Kaṇṭhaja)	k [k]	kh [kʰ]	g [g]	gh [gʰ]	ṅ [ŋ]
Palatal (Tāluja)	c [c]	ch [cʰ]	j [j]	jh [jʰ]	ñ [ɲ]	y [j]
Retroflex (Muddhajā)	ṭ [ṭ]	ṭh [tʰ]	ḍ [ḍ]	ḍh [dʰ]	ṇ [ṇ]	r [r]	ḷ [ḷ]
Dental (Dantaja)	t [t]	th [tʰ]	d [d]	dh [dʰ]	n [n]		l [l]	s [s]
Bilabial (Oṭṭhaja)	p [p]	ph [pʰ]	b [b]	bh [bʰ]	m [m]
Labio-dental (Dantoṭṭhaja)							v [v]
Nasal Cavity (Nāsikā)					(a)ṃ [ā]
					(i)ṃ [ī]
					(u)ṃ [ū]

Pāḷi sentence structure

A Pāḷi sentence (vākya) or Pāḷi text (Pāḷi) is generally composed of multiple words (pada or vacana). Since Pāḷi was an oral language, punctuation mark and upper/lower case conventions are typically not used, although modern editors have sometimes added these to aid reading.

Words also do not need to be separated by spaces, although modern Pāḷi editions of text do add spaces, to aid reading. Pāḷi also has compound words, which consist of multiple words joined together to create a composite word.

To be able to distinguish the role of words in a sentence, Pāḷi relies heavily on vibhatti (inflection forms which are affixes to lemmas) to create distinctive word endings. Fluent Pāḷi listeners will no doubt listen to the sound of these endings to distinguish the words. These endings also give additional meaning to the lemmas and indicate how the words are used in a sentence.

Because of this, the word order in a Pāḷi sentence is less important than in many other languages; however, there are conventions for how to arrange words in a typical sentence. Speakers can unconventionally reorder words, typically to give emphasis to certain words or to augment or add nuance to a sentence’s emotional meaning.

Note

This is similar to our reaction when we hear “Yoda” speak in the “Star Wars” series of films. We appreciate who Yoda is as a personality, and we pay more attention to his words simply because of the different order.

dhātu (roots)

Fundamentally, most lemmas in Pāḷi are said to be derived from one or more dhātu (roots). These roots are similar to Sanskrit roots. Pāḷi can be regarded as a simplified version of Sanskrit; however, some of the roots may have different meanings, or nuances. There are some that say Sanskrit was a “sacred” or “divine” language reserved for priests and members of the holy caste, whereas Pāḷi, along with many other Indian languages, was spoken by the common people.

Some lemmas, however, are words borrowed from other languages, possibly indigenous to the region the Buddha grew up or taught in. These lemmas are obviously not derived from Sanskrit roots. Examples include names of villages, people names, animal names, farming words and slang words, etc. (Levman, 2014)

dhātu can have 1-3 syllables:

type	example	meaning
single-stem or one-syllable roots	vā	to move
	yā	to go
	pā	to drink
	ji	to conquer
	nī	to carry
	su	to flow
	bhū	to be
dual-syllable roots	gamu	to go
	paca	to cook
	vada	to speak
multiple-stem or three-syllable roots	vāyama	to attempt
	araha	to deserve
	kilisa	to torment

A series of transformations convert a root into a lemma by adding paccaya (affixes). Lemmas can also be formed by combining a root with one or more upasagga (prefixes). Multiple upasagga and different forms of paccaya can be applied to a root before it becomes a lemma, thus creating multi-syllabic lemmas. The result is a lemma that can be transformed into a word (through inflection forms - more later) used in a sentence.

Altogether there are around 1700 roots in Pāḷi. However, many, many lemmas can be made from them by attaching upasagga (prefixes) and various affixes (paccaya). In theory, fresh lemmas can be formed by new combinations of roots with different prefixes so the adventurous Pāḷi student can attempt to create Pāḷi lemmas for “modern” terms such as “computer” or “phone”.

It is not really necessary to memorise the 1700 roots, although it will be very useful if the reader chooses to do so. Knowing all the roots and the various prefixes and affixes will allow the reader to sometimes (not always) “guess” the meaning of a lemma not previously encountered, and will also help when applying the various transformation rules (see below).

nāmā (nouns) {52}

nāmā are the most common type of words in Pāḷi, and roughly (but not exactly) correspond to the term “noun” in English. nāmā also means “name”, so these words represent things that we can “name” - including people, places, things, numbers, feelings and many other objects or concepts.

Example {53}

Here is a simple sentence, consisting of 3 nāma:

eso no satthā

which translates as:

that person/being / our / teacher
That person [is] our teacher.

Instead of satthā other examples may be:

• brahmā (God)
• attā (self)
• sakhā (friend)
• rājā (King)

liṅga (gender) {53}

nāma can be classified into 3 different types or liṅga. The word liṅga actually means “gender”, and traditionally each nāma is assigned a gender. The liṅga represents an “intrinsic property” of the nāma.

The notion of a Pāḷi gender is similar to (but not quite the same as) genders in some European languages. However, there are such differences that it’s best to regard liṅga as a classification scheme or grammatical construct rather than what traditional definitions of “gender” suggest:

• pulliṅga - the “major” or “default” type. Most nouns fall into this category. Generally associated with masculinity or masculine persons (but not always)
• napuṁsakaliṅga - the “special” type. Tends to be used for nouns that don’t quite fit with the notions of masculinity or femininity, or to represent abstract concepts.
• itthiliṅga - the “minor” type, comparatively rare with respect to the other two types. Generally associated with femininity or feminine persons.

Note that for the remainder of this introduction, liṅga will be represented by the following symbols for brevity:

• 🚹 = pulliṅga
• 🚻 = napuṁsakaliṅga
• 🚺 = itthiliṅga

Ancient India was a patriarchal society, and men were regarded as the “major” component of society, doing activities considered important, hence the tendency is to associate important words with pulliṅga.

Multiplicity

Like in English, Pāḷi words are slightly different depending on whether they are used to refer to one thing vs many things.

• ⨀ = ekavacana (singular)
• ⨂ = bahuvacana (plural)

There used to be also a “dual” form but that is now largely obsolete and encountered in very rare instances. Examples include:

• to idh’āgato (these two having come)
• ubho (both)
• mātāpitu (father and mother)

vibhatti (inflection forms) {54}

Unlike English, words need to be “transformed” before they are used in a sentence. This is done by adding inflection forms or “suffixes” to the word (vibhatti).

Altogether there are seven types of vibhatti, and they are conveniently numbered 1 to 7 {55}:

symbol	⨀	⨂	vibhatti
①	si	yo	paṭhamā (“first”)
②	aṁ	yo	dutiyā (“second”)
③	nā	hi	tatiyā (“third”)
④	sa	naṁ	catutthī (“fourth”)
⑤	smā	hi	pañcamī (“fifth”)
⑥	sa	naṁ	chaṭṭhī (“sixth”)
⑦	smiṁ	su	sattamī (“seventh”)

Note the 4th and 6th vibhatti endings are the same, which means in practice which vibhatti is being used can only be determined by context (more later). Notice also the plural endings for some of the vibhattis are the same. Again, which vibhatti is being used is determined by context.

Let’s start with an example. Consider the lemma purisa (“man”, in the masculine gender or 🚹) and it needs to be inserted in a sentence in singular form representing itself. The 1st vibhatti (paṭhamāvibhatti) is used to represent the underlying meaning of nouns. By looking up the above table, the si ending is added (using rule of {55}):

purisa + si → puriso

Wait, what happened? Why does the transformation not result in purisasi?

This is where additional rules come into play. By rule of {83}, any lemma ending in a will be transformed into the 1st vibhatti by deleting the a and by rule of {104} replacing the si with o. Kaccāyana has a whole chapter of such rules.

So the transformation becomes:

purisa + si
→ purisa + si {83}
→ puris + si(o) {104}
→ puriso

The plural form undergoes a similar transformation:

purisa + yo
→ purisa + yo {83}
→ puris + yo(ā) {107}
→ purisā

This does make things interesting, because the transformation result for every word then depends on:

• kāra’nta (word ending)
• liṅga (gender)
• ekavacana (singular) vs. bahuvacana (plural)
• vibhatti (desired inflection form)

Here is a table of purisa transformed in all the 7 vibhattis and singular vs plural.

number	vibhatti	ending	transformation	sutta	result
⨀	①	si	purisa + sio	{104}	puriso
⨂	①	yo	purisa + yoā	{107}	purisā
⨀	②	aṁ	purisa + aṁ		purisaṁ
⨂	②	yo	purisa + yoe	{107}	purise
⨀	③	nā	purisa + ena	{103}	purisena
⨂	③	hi	purisae + hi	{101}	purisehi
⨂	③	hi	purisae + bhi	{99}-{101}	purisebhi
⨀	④	sa	purisa + ssa	{61}	purisassa
⨂	④	naṁ	purisaā + naṁ	{89}	purisānaṁ
⨀	⑤	smā	purisa + smā		purisasmā
⨀	⑤	smā	purisa + mhā	{99}	purisamhā
⨀	⑤	smā	purisaā	{108}	purisā
⨂	⑤	hi	purisae + hi	{101}	purisehi
⨂	⑤	hi	purisae + bhi	{99}-{101}	purisebhi
⨀	⑥	sa	purisa + ssa	{61}	purisassa
⨂	⑥	naṁ	purisaā + naṁ	{89}	purisānaṁ
⨀	⑦	smiṁ	purisa + smiṁ		purisasmiṁ
⨀	⑦	smiṁ	purisa + mhi	{99}	purisamhi
⨀	⑦	smiṁ	purisae	{108}	purise
⨂	⑦	su	purisae + su	{101}	purisesu

You will note that in some of the forms, there are multiple ways the lemma can be transformed. For example, plural purisa can be transformed in the 3rd vibhatti into either purisehi or purisebhi. Both are acceptable. Which should we use for a given sentence? It depends on the “sound” - a “native” Pāḷi speaker would have chosen the version that sounded the most pleasing, depending on surrounding words.

The multiple different variants of inflection forms can also be due to Pāḷi being an amalgamation of different Middle Indo-Aryan dialects. As (Norman, 1997) explains:

It is, for example, not always understood by non-specialists that an early Pāli canonical sutta is itself a translation, and forms which were left untranslated when the Pāli recension was made from some earlier version can sometimes be identified.

The rules can be quite complicated, so this introduction will not enumerate them. In practice, one can rely on looking up a relevant table to determine the appropriate transformation. For reading Pāḷi, with practice the various endings will quickly become familiar.

A much more important question to answer at this point is …

What is the purpose of vibhatti (inflection forms)?

In English, consider the following sentence:

The quick brown fox jumped over the lazy dog.

Words “quick” and “brown” refer to the “fox”, which is the “subject” of the sentence, and “lazy” refers to the dog, which is the “object” of the sentence. The verb “jumped over” separates the subject and the object. The brain automatically “collects” and “groups” the words, so the sentence can be read as:

(The quick brown fox) jumped over (the lazy dog).

In English, word order is important. If the words are “mixed”, the sentence becomes far harder to read:

Over the quick lazy brown fox the dog jumped.

This sentence would imply quite a different meaning.

In Pāḷi, all words with the same gender and the same vibhatti should be grouped together, and that’s how words belonging to entities can be distinguished. So the word order becomes far less important - the words can be “mixed” and the sentence would still have the same meaning!

In practice, the 1st vibhatti is usually reserved for the “most important” entity in a sentence, the 2nd for the “second most important” and so on.

So, in the above English sentence, were it to be translated to Pāḷi, “the quick brown fox” would all be in the 1st vibhatti, and “the lazy dog” would all be in the 2nd vibhatti. If there are more entities, they would take the 3rd, 4th vibhatti and so on.

In many sentences, the 1st vibhatti would usually refer to the entity or actor “performing” or “doing” the action (or, the “subject” of the sentence), and the 2nd vibhatti refers to the entity being impacted by the action (the “object”) of the sentence. Although this is a very common pattern, it is by no means rigid, and in theory (almost) any vibhatti ending can be used for any purpose. There are many examples in the Pāḷi canon where the vibhatti are used in unconventional ways.

It is also possible that a word in an inflection form from a different Middle Indo-Aryan dialect can be mistaken for a different inflection form in Pāḷi. For example, in what Norman calls the “Eastern” dialect the first vibhatti forms for nouns often end in -e whereas in the “Western” dialect they end with -o. As an example, the vocative bhikkhave typically encountered in Pāḷi texts should probably be bhikkhavo, although (Pind, 2021) disputes this and argues bhikkhavo represents an emphatic usage compared to bhikkhave.

Vibhatti usage typical meanings {271}-{315}

The following meanings are usually associated with each vibhatti. However, as noted above, these are conventions only, and the author of a sentence may choose to associate an unconventional vibhatti ending with a word (perhaps for emphatic or stylistic reasons).

vibhatti	number	role	usage	Pāḷi term	sutta
①	⨀/⨂	subject (active)	subject, doer	kattu/kattā	{284}
①	⨀/⨂	address	to address, to call	ālapana	{285}
②	⨀	to (object)	object	kamma	{297}
②	⨂	till	continuity for a certain period or distance	accantasaṁyoga	{298}
③	⨀	by (subject) (passive)	subject	kattu/kattā	{288}
③	⨂	by/with	supporting cause	karaṇa	{286}
③	⨂	with/together	company	sahayoga	{287}
③	⨂	by means of/in regard to	aspect or adjective	visesana	{292}
③	⨂	due to/for	reason	kāraṇa	{289}
④	⨀/⨂	receivership/purpose	recipient of action	sampadāna	{293}
⑤	⨀	from	source from which it occurs	apādāna	{295}
⑤	⨂	for/due to	reason	kāraṇa	{296}
⑥	⨀	(possessive) of	owner	sāmi	{301}
⑥	⨂	among/out of	selection for comparison	niddhāraṇa	{304}
⑥	⨂	while/despite	two concurrent actions	anādara	{305}
⑦	⨀	at, in, on	location/position	adhikaraṇa/okāsa	{302}
⑦	⨂	among/out of	selection for comparison	niddhāraṇa	{304}
⑦	⨂	for/due to	cause	nimitta	{310}
⑦	⨂	while/when	two concurrent actions	lakkhaṇa	{313}
⑦	⨂	despite/while	two concurrent actions	anādara	{305}

Note

In Pāḷi grammar books written in English the following nomenclature for the vibhatti types is often used (which these books will refer to as “cases”):

• ① = “nominative”
• ② = “accusative”
• ③ = “instrumental”
• ④ = “dative”
• ⑤ = “ablative”
• ⑥ = “genitive”
• ⑦ = “locative”

Note that these terms only partially reflect the possible meanings and roles associated with each vibhatti, so I would discourage using these terms. They can lead to confusion. However, these terms are common in Pāḷi-English dictionaries so I encourage you to substitute the vibhatti number when you see these terms.

⓪ ālapana (vocative) {57}

English Pāḷi textbooks often refer to a “vocative” “case”, but this is not recognised as a separate inflection form but a variant of the 1st vibhatti. The usual vibhatti ending si (for 1st vibhatti singular) is instead called ga but the transformation rules are similar to si.

Examples:

bhoti ayye! (Oh madam!)
bhoti kaññe! (Oh girl!)
bhoti kharādiye! (Oh the girl named Kharādiya!)

Contrast this to the normal 1st vibhatti form of ayyā (🚺⨀①):

sā ayyā (that lady)

In plural:

bhotiyo ayyāyo (Oh madams!)

*️⃣ sabbanāma (pronouns)

sabba means “all” so sabbanāma are “common” or “generic” nouns that can apply in “all” situations.

Examples of sabbanāma include the different vibhatti forms of ya (“who, what”), ta (“he/she/it”, “that”), eta (“this”), ima (“this”), amu (“that”), kiṁ (“what?”). One of the common usages of sabbanāma is as “pronouns”, where they can assume all genders to represent all things and persons of the different genders. But they can also function as adjectives, adverbs, honorifics and interrogatives.

Examples:

Sabbanāma	meaning	④	⑥
sabba	all	sabbesaṁ	sabbesānaṁ
ya	which	yesaṁ	yesānaṁ
ta	that	tesaṁ	tesānaṁ
ima	this	imesaṁ	imesānaṁ
ka	who	kesaṁ	kesānaṁ
itara	other	itaresaṁ (others)	itaresānaṁ (others)
katama	what	katamesaṁ (of which ones)	katamesānaṁ (of which ones)

⏏️ nipāta (particles)

nipāta means “to fall, i.e. to be placed anywhere”, and refers to words that can be found placed in the beginning, in the middle or at the end of sentences and words, but not usually regarded as prefixes (see below).

Generally, the vibhatti endings of nipāta words are elided as these words are not related to other words in the sentence and therefore do not need to be matched with them. {221}

However, in some instances, the vibhatti endings may be retained, if the nipāta word plays a role in the sentence meaning.

Examples

tvaṁ panā’vuso! tumhe panā’vuso!

you / (no specific meaning) / friend
you (plural) / (no specific meaning) / friend
How about you, my friend? And the rest of you (plural)?

padaso dhammaṁ vāceyya

by word / dhamma / teach (causative, optative)
(He) should teach and say Dhamma word by word.

vihāraṁ • sve upagaccheyya

to temple / tomorrow / should approach (optative)
(He) should come to the temple tomorrow.

The following is a brief descriptive list of some nipāta words. Some are plain nipātas without any affix, while some are with indeclinable affixes. Some may even have a sort of vibhatti-ending in them or may still have traces of vibhattis. Please note that the implied meaning of elided vibhatti still remains in effect for some nipāta words such as yathā, tathā and evaṁ etc.

nipāta	meaning	note
yathā, tathā	in a manner that, in such a way as	combination of sabbanāma ya ta with indeclinable affix thā. The affix itself is expressive of manner or mood.
evaṁ	thus, in this way, true as it is said, yes (many meanings)	dutiyāvibhatti, very frequently used in Pāḷi texts
khalu	used mostly in a reported narrative
kho	mostly meaningless, but sometimes it may mean “only, really”	a plain nipāta, found often in main Buddhist texts. Used together with atha evaṁ taṁ etc.
tatra	there	a combination of sabbanāma ta with indeclinable affix tra
atho	in addition, besides
atha	now, then, later on, if	Sometimes used in the beginning of a chapter or an episode, as an introductory or initiating word
hi	really, for, only	Sometimes a meaningless particle
tu	in addition, only	Also an expletive without meaning, sometimes added to be more emphatic in a statement
ca	also, too	It is used to express some additional meaning in grammatical texts. Sometimes meaningless. Sometimes it has a lot of meanings when used after other nipāta particles
vā	similar to either or, in other ways	In grammatical texts, sometimes it is used to express other option or method or inconsistency of a function. Sometimes with no particular meaning
vo	you	It is sometimes a meaningless expletive
haṁ, ahaṁ	I	the use of these two nipāta is seldom found in texts unless it means “I”
alaṁ	(a) enough, (b) suitable, (c) worthiness, (d) appropriateness, (e) ability	Also, (f) sometimes used as a prohibition or rejection, usually used in combination with (1) a noun in instrumental case, or (2) tuṁ-infinitive Kita verbs or (3) an āya-infinitive in dative case nouns
eva	equal to “only”	used next to the word it wants to modify or restrict or confirm in its implied meaning. A plain nipāta.
ho!, aho!		these are expressions of interjection or surprise. No vibhatti
he!, ahe!, re!, are!	“hey”	Vocative nipāta particles, similar to “hey” in English. Usually vocative case vibhatti is applicable, but to be elided

nipāta as prefixes

However, there are quite a few nipāta that can be placed before some verbs or kita-affixed nouns and used in a way resembling prefixes.

combined	nipāta	word	combined meaning
antaradhāyati	antara	dhāyati	(he) disappears
āvikaroti	āvi	karoti	(he) makes it openly, displays clearly
pātubhavati	pātu	bhavati	(it) arises visibly, i.e. appears
sacchikaroti	sacchi	karoti	(he) realises, i.e. attains
vinābhāvo	vinā	bhāvo	separation, being apart
sahacārī	saha	cārī	the one who used to be together, a friend
punabbhavo	puna	bhavo	being born again, rebirth

🔼 upasagga (prefixes)

There are 20 upasagga which are usually applied as prefixes to ākhyāta (verbs) and kita-affixed nouns (see below). Like nipāta, upasagga also have their vibhatti endings elided, but in some cases where they are functioning as independent words (not attached to a verb or noun), they may retain their vibhatti endings.

There are three kinds of upasagga:

1. dhātva’tthā’nuvattaka - those that follow the meaning of the root, without affecting its original meaning.
2. dhātva’tthabādhaka - those that absolutely mean the opposite of the root. These categorically change the original meaning of the root.
3. dhātva’tthavisesaka - those that enhance the original meaning of the root by adding more specific meaning and flavour to it.

Example:

parābhavo

🔼(parā) 🚹⨀①(bhava)
opposite / to be, being
loss, ruin

Here are various words that can be formed from the word hara (“carrying”) together with various upasagga prefixes and the ṇa-suffix:

example	upasagga	meaning	combined meaning
pahāro	pa	specially	carrying (instruments of harm) specially, i.e. beating, striking
nihāro	ni	out	taking out, evicting
nīhāro	nī	out	taking out, evicting (same as above)
uhāro	u	up	carrying up
duhāro	du	bad	carrying badly
saṁhāro	saṁ	well	carrying well
vihāro	vi	specially	carrying one’s body specially, staying, i.e. act of staying at a place using alternating bodily modes
avahāro	ava	lowly, contemptuous, mean	carrying in a mean manner, i.e. stealing
anuhāro	anu	following, subsequent	carrying accordingly
parihāro	pari	all around	carrying from all around
adhihāro	adhi	specially, nicely	carrying nicely
abhihāro	abhi	toward	carrying toward
patihāro	pati	again	carrying again
suhāro	su	well	carrying well
āhāro	ā	toward	carrying toward, food or meal
atihāro	ati	beyond, excessive	carrying beyond or excessively
apihāro	api	downward	carrying down
apahāro	apa	out	carrying outward, i.e. removing
upahāro	upa	near, close to	carrying toward proximity, bringing up closer, carrying complementary gift etc.

🆎 samāsa (compound nouns) {316}-{343}

Like German, Pāḷi often combines many nouns to form a compound noun.

The following table summarises the major types of compound nouns, typical liṅga (gender) and examples:

type	name	liṅga	example	rule
abyayībhāva	Adverbial	🚻	so napuṁsakaliṅgo	{320}
kammadhāraya	Appositional	variable, or follow last component	abhidheyavacano, paraliṅgo ca	{341}
digu	Numeral Appositional	🚻 (usually ⨀ but can be ⨂)	digusseʼkattaṁ	{321}
tappurisa	Determinative	variable	ayañʼca tappuriso abhidheyavacanaliṅgo	Rūpasiddhi {351}
dvanda	Copulative	Mostly 🚻 but variable in some instances per last component	samāhāre napuṁsakaṁ	Moggallāna {20} Chapter 3
bahubbīhi	Attributive	variable per the contextual meaning of the aññapada (the external word of attributive meaning)	bahubbīhi cāʼyaṁ abhidheyaliṅgavacano	Rūpasiddhi {352}

The rules for constructing a samāsa can be complex:

rāja + putta (the King’s son)
→ rāja + sa + putta (⨀⑥)
→ rāja + sa + putta + si (⨀①)
→ rāja(ññ) + sa(o) + putta + si {135}
→ rāñño + putta + si(o) {83}
→ rāñño + putto
→ rāñño(ja) + putto(a) {317}
→ rājaputta {318}
→ rājaputta + si (⨀①)
→ rājaputta + si(o) {104}
→ rājaputta + o {83}
→ rājaputto

Note

Fortunately, resources such as the Digital Pāḷi Dictionary can deconstruct compound words into their individual words, which helps the beginner to understand such words encountered. So there is less of a need to learn the complex rules associated with them.

➡️ taddhita (affixes) {344}-{405}

Various affixes can be added to a noun to create derived nouns and adjectives.

Compare and contrast with English use of affixes such as “-al”, “-ly”, “-er”, “-or” etc. For example, “topic” vs “topical”, “true” vs “truly”.

In English, “navigate” is a verb, whereas “navigator” (with an “-or” affix) is a profession. Similarly, in Pāḷi, nāvā is “boat”, nāvā with ṇika as an affix becomes nāviko (“traveller by boat”, or “sailor”). Note that the last vowel of the root and the ṇ are both elided, and then the noun is given the 🚹⨀① vibhatti ending.

The rules for constructing a taddhita affixed noun can be rather daunting:

vasiṭṭha + apacca (Vasiṭṭha’s son)
→ vasiṭṭha + sa (⨀⑥) + apacca
→ vasiṭṭha + s + sa + apacca {61}
→ vasiṭṭhassa + apacca + si (⨀①)
→ vasiṭṭhassa + apacca + si(aṁ) {219}
→ vasiṭṭhassa + apacca + aṁ {83}
→ vasiṭṭhassa + apaccaṁ (of Vasiṭṭha / son)
→ vasiṭṭhassa + ṇa + apaccaṁ {344}
→ vasiṭṭhassa + ṇa + apaccaṁ {317}
→ vasiṭṭha + ṇa {396}
→ vasiṭṭha + a {318}
→ va(ā)siṭṭha + a {400}
→ vāsiṭṭha + a {83}
→ vāsiṭṭha {601}
→ vāsiṭṭha + si (⨀①)
→ vāsiṭṭha + si(o) {104}
→ vāsiṭṭha + o {83}
→ vāsiṭṭho

type	meaning	affix	rule	liṅga
apacca	patronymic	ṇa	{344}	🚹 (but can be variable)
		ṇāyana, ṇāna	{345}	🚹 (but can be variable)
		ṇeya	{346}	🚹 (but can be variable)
		ṇi	{347}	🚹
		ṇika	by vā of {347}	🚹 (but can be variable)
		ṇava	{348}	🚹 (but can be variable)
		ṇera	{349}	🚹 (but can be variable)
aneʼkattha	various meanings	ṇika	{350}-{351}	🚹 (but can be variable)
		ṇa	{352}	🚹 (but can be variable)
		ima, iya	{353}	🚹 (but can be variable)
		ima, iya, ika	by ādi of {353}	🚹 (but can be variable)
		kiya	by ca of {353}	🚹 (but can be variable)
samūha	collective grouping	kaṇ, ṇa	{354}	🚹
		tā	{355}	🚺
taʼdassathāna	the source or the cause	iya	{356}	🚻 (but can be variable)
upamā	analogous similitude	āyitatta	{357}	🚻
tanʼnissitattha	dependent nature and source	la	{358}	🚻 (mostly)
bahula	prominent nature	ālu*	{359}	variable
bhāva	abstract conditions	ṇya, tta	{360}	🚻
		ttana	by tu of {340}	🚻
		tā	{360}	🚺
		ṇa	{361}	🚻
visesa	comparisons	tara, tama, isika, iya, iṭṭha	{363}	variable
taʼdassaʼtthi	possessed quality or inherent nature	vī	{364}	🚹 (but can be variable)
		so	by ca of {364}	🚹 (but can be variable)
		sī	{365}	🚹 (but can be variable)
		ika, ī	{366}	🚹 (but can be variable)
		ra	{367}	🚹 (but can be variable)
		vantu	{368}	🚹 (but can be variable)
		mantu	{369}	🚹 (but can be variable)
		ṇa	{370}	🚹 (but can be variable)
tappakati	made up of something	maya	{372}	variable

*ka is added sometimes. e.g. dayāluko

Lemmas with taddhita affixes are sometimes not recognised by Pāḷi-English dictionaries, since these dictionaries may not cover the full range of possible affixes and their meanings. So it is probably worthwhile learning how to recognise taddhita affixed lemmas and the associated meanings.

Numbers

Pāḷi numbers have their own set of taddhita affixes (to denote “first”, “second”, “third” etc.)

type	name	affix	rule	liṅga	notes
Saṅkhyā Taddhita	Numbers	ma	{373}	variable
		ī	{375}	🚺
		ti	{378},{389}	🚺	derivative morpheme
		tiya	{385}	variable
		ka	{392}	🚻

no	Cardinals	Ordinals
1	eka	paṭhama
2	dvi, di, du, dve	dutiya
3	ti or tri	tatiya
4	catu or catur (before a vowel)	catuttha, turīya
5	pañca	pañcatha, pañcama
6	cha	chaṭṭha, chaṭṭhama
7	satta	sattha, sattama
8	aṭṭha	aṭṭhama
9	nava	navama
10	dasa, rasa, lasa, ḷasa	dasama, dasī

Abyaya (indeclinable affixes)

These are indeclinable (and have no liṅga), and often used as adverbs or adjectives.

affix	rule
dhā	{397}
so	by ca of {397}
thā	{398}
thattā	by tu of {398}
thaṁ	{398}
khattuṁ	Uṇādisutta {646}

Abyaya affixes applied after Sabbanāma-nouns

These are often used as a substitute or replacement for vibhatti endings (particularly ⑤ and ⑦), since they are indeclinable. They also have no liṅga (gender).

use	affix	rule
To express reason or time	to	{248}
To express position or place	to	{248}
	tra, tha	{249}
	dhi	{250}
	va	{251}
	hiṁ, haṁ, haṁ	{252}
	ha, dha	{254}
To express time	dā, dācanaṁ	{254}
	hiṁ, haṁ, haṁ	{252}
To express day	jja, jju	{571}

ākhyāta (verbs) {406}-{523}

Like nouns, verbs also have vibhatti endings affixed to them when used in sentences, but the endings are different from those of nouns. These endings depend on:

• voice {406}-{407}
  • 🟢 = parassapada (active)
  • 🔵 = attanopada (middle) - impersonal
• number
  • ⨀ = ekavacana (singular)
  • ⨂ = bahuvacana (plural)
• person {408}-{412} (if more than one, the last applies)
  • 🤟 = paṭhama purisa (first person, equivalent to ‘third person’ in English)
  • 🤘 = majjhima purisa (middle person, equivalent to ‘second person’ in English)
  • 👆 = uttama purisa (primary person, equivalent to ‘first person’ in English)
• tense/mood {413}-{422}
  • ▶️ = vattamāna (present)*
  • ⏹ = pañcamī (imperative) - (used to express) āṇatti (command), āsīsa (wish) at a non-specific time-frame mode (anuttakāla)*
  • ⏯ = sattamī (potential/optative)*
  • 🔄 = parokkhā (past perfect)
  • ↩️ = hiyyattanī (past imperfect)*
  • ⏮ = ajjatanī (aorist)
  • ⏭ = bhavissanti (future)
  • 🔀 = kālātipatti (conditional)

* collectively referred to as sabbadhātuka vibhatti

Note

The above tense/mood categories have been labelled with the nearest linguistic term (e.g. “present”, “past perfect”, “aorist” etc.); these categories do not align precisely to those terms. I discourage the use of these terms and rely on the Pāḷi terms instead and the usage of each category will be described later in this chapter.

verb terminations {423}-{430}

	🟢	🟢	🟢	🟢	🟢	🟢	🔵	🔵	🔵	🔵	🔵	🔵
	🤟	🤟	🤘	🤘	👆	👆	🤟	🤟	🤘	🤘	👆	👆
	⨀	⨂	⨀	⨂	⨀	⨂	⨀	⨂	⨀	⨂	⨀	⨂
▶️	ti	anti	si	tha	mi	ma	te	ante	se	vhe	e	mhe
⏹	tu	antu	hi	tha	mi	ma	taṁ	antaṁ	ssu	vho	e	āmase
⏯	eyya	eyyuṁ	eyyāsi	eyyātha	eyyāmi	eyyāma	etha	eraṁ	etho	eyyāvho	eyyaṁ	eyyāmhe
🔄	a	u	e	ttha	aṁ	mha	ttha	re	ttho	vho	iṁ	mhe
↩️	ā	ū	o	ttha	aṁ	mhā	ttha	tthuṁ	se	vhaṁ	iṁ	mhase
⏮	ī	uṁ	o	ttha	iṁ	mhā	ā	ū	se	vhaṁ	aṁ	mhe
⏭	ssati	ssanti	ssasi	ssatha	ssāmi	ssāma	ssate	ssante	ssase	ssavhe	ssaṁ	ssāmhe
🔀	ssā	ssaṁsu	sse	ssatha	ssaṁ	ssāmhā	ssatha	ssiṁsu	ssase	ssavhe	ssaṁ	ssāmhase

Here is an example of the root gamu (to go) with all the verb endings applied. A number of transformations need to happen prior to affixing the vibhatti:

gamu
→ gamu (elision of final vowel per rule {521})
→ gam(cch) (m→cch per rule {476})
→ gacch

Here is an example for transforming into the parassapada (🟢), ekavacana (⨀), paṭhama purisa (🤟), vattamāna (▶️) form:

gamu + ti {414}
→ gamu + ti {521}
→ gam + a + ti {445}
→ gam(cch) + a + ti {476}
→ gacchati

The plural form bahuvacana (⨂) undergoes a similar transformation:

gamu + anti {414}
→ gamu + anti {521}
→ gam + a + anti {445}
→ gam(cch) + a + anti {476}
→ gacch + a + anti {510}
→ gacchanti

voice	number	person	tense	ending	transformation	rule	result	irregular forms
🟢	🤟	⨀	▶️	ti	gacch + a + ti	{445}	gacchati	gacche
🟢	🤟	⨂	▶️	anti	gacch + anti		gacchanti	gacchare
🟢	🤘	⨀	▶️	si	gacch + a + si	{445}	gacchasi
🟢	🤘	⨂	▶️	tha	gacch + a + tha	{445}	gacchatha
🟢	👆	⨀	▶️	mi	gacch + ā + mi	{478}	gacchāmi	gacche
🟢	👆	⨂	▶️	ma	gacch + ā + ma	{478}	gacchāma
🔵	🤟	⨀	▶️	te	gacch + a + te	{445}	gacchate
🔵	🤟	⨂	▶️	ante	gacch + ante		gacchante	gacchare
🔵	🤘	⨀	▶️	se	gacch + a + se	{445}	gacchase
🔵	🤘	⨂	▶️	vhe	gacch + a + vhe	{445}	gacchavhe
🔵	👆	⨀	▶️	e	gacch + e		gacche
🔵	👆	⨂	▶️	mhe	gacch + ā + mhe	{478}	gacchāmhe
🟢	🤟	⨀	⏹	tu	gacch + a + tu	{445}	gacchatu	gacche
🟢	🤟	⨂	⏹	antu	gacch + antu		gacchantu
🟢	🤘	⨀	⏹	hi	gacch + ā + hi	{478}	gacchāhi	gaccha {479}, gacchassu {571}
🟢	🤘	⨂	⏹	tha	gacch + a + tha	{445}	gacchatha
🟢	👆	⨀	⏹	mi	gacch + ā + mi	{478}	gacchāmi	gacche
🟢	👆	⨂	⏹	ma	gacch + ā + ma	{478}	gacchāma
🔵	🤟	⨀	⏹	taṁ	gacch + a + taṁ	{445}	gacchataṁ
🔵	🤟	⨂	⏹	antaṁ	gacch + antaṁ		gacchantaṁ
🔵	🤘	⨀	⏹	ssu	gacch + a + ssu	{445}	gacchassu
🔵	🤘	⨂	⏹	vho	gacch + a + vho	{445}	gacchavho
🔵	👆	⨀	⏹	e	gacch + e		gacche
🔵	👆	⨂	⏹	āmase	gacch + āmase		gacchāmase
🟢	🤟	⨀	⏯	eyya	gacch + eyya		gaccheyya	gacche
🟢	🤟	⨂	⏯	eyyuṁ	gacch + eyyuṁ		gaccheyyuṁ
🟢	🤘	⨀	⏯	eyyāsi	gacch + eyyāsi		gaccheyyāsi	gacche
🟢	🤘	⨂	⏯	eyyātha	gacch + eyyātha		gaccheyyātha
🟢	👆	⨀	⏯	eyyāmi	gacch + eyyāmi		gaccheyyāmi	gacche
🟢	👆	⨂	⏯	eyyāma	gacch + eyyāma		gaccheyyāma
🔵	🤟	⨀	⏯	etha	gacch + etha		gacchetha
🔵	🤟	⨂	⏯	eraṁ	gacch + eraṁ		gaccheraṁ
🔵	🤘	⨀	⏯	etho	gacch + etho		gacchetho
🔵	🤘	⨂	⏯	eyyāvho	gacch + eyyāvho		gaccheyyāvho
🔵	👆	⨀	⏯	eyyaṁ	gacch + eyyaṁ		gaccheyyaṁ	gacche
🔵	👆	⨂	⏯	eyyāmhe	gacch + eyyāmhe		gaccheyyāmhe
🟢	🤟	⨀	🔄	a	jagamu + a	{458}	jagama	jagāma
🟢	🤟	⨂	🔄	u	jagamu + u	{458}	jagamu
🟢	🤘	⨀	🔄	e	jagamu + e	{458}	jagame
🟢	🤘	⨂	🔄	ttha	jagamu + i + ttha	{458},{516}	jagamittha
🟢	👆	⨀	🔄	aṁ	jagamu + aṁ	{458}	jagamaṁ
🟢	👆	⨂	🔄	mha	jagamu + ā + mha	{458}	jagamimha
🔵	🤟	⨀	🔄	ttha	jagamu + i + ttha	{458},{516}	jagamittha
🔵	🤟	⨂	🔄	re	jagamu + i + re	{458},{516}	jagamire
🔵	🤘	⨀	🔄	ttho	jagamu + i + ttho	{458},{516}	jagamittho
🔵	🤘	⨂	🔄	vho	jagamu + i + vho	{458},{516}	jagamivho
🔵	👆	⨀	🔄	iṁ	jagamu + iṁ	{458}	jagamiṁ
🔵	👆	⨂	🔄	mhe	jagamu + i + mhe	{458},{516}	jagamimhe
🟢	🤟	⨀	↩️	ā	a + gacch + ā	{519}	agacchā	gacchā, agaccha, gaccha
🟢	🤟	⨂	↩️	ū	a + gacch + ū	{519}	agacchū	gacchū, agacchu, gacchu
🟢	🤘	⨀	↩️	o	gacch + o		gaccho	agaccho, agaccha, gaccha, agacchi, gacchi
🟢	🤘	⨂	↩️	ttha	gacch + a + ttha	{445}	gacchattha	agacchattha, agacchatha, gacchatha
🟢	👆	⨀	↩️	aṁ	gacch + aṁ		gacchaṁ	agacchaṁ
🟢	👆	⨂	↩️	mhā	gacch + ā + mhā	{478}	gacchāmhā	agacchamhā
🔵	🤟	⨀	↩️	ttha	gacch + a + ttha	{445}	gacchattha	agacchattha
🔵	🤟	⨂	↩️	tthuṁ	gacch + a + tthuṁ	{445}	gacchatthuṁ	agacchatthuṁ
🔵	🤘	⨀	↩️	se	gacch + a + se	{445}	gacchase	agacchase
🔵	🤘	⨂	↩️	vhaṁ	gacch + a + vhaṁ	{445}	gacchavhaṁ	agacchavhaṁ
🔵	👆	⨀	↩️	iṁ	gacch + iṁ		gacchiṁ	agacchiṁ
🔵	👆	⨂	↩️	mhase	gacch + ā + mhase	{478}	gacchāmhase	agacchamhase
🟢	🤟	⨀	⏮	ī	a + gacch + ī	{519}	agacchī	gacchī, agacchi, gacchi
🟢	🤟	⨂	⏮	uṁ	gacch + uṁ		agacchuṁ	gacchuṁ, agacchiṁsu, gacchiṁsu
🟢	🤘	⨀	⏮	o	gacch + o		gaccho	agaccho, agaccha, gaccha, agacchi, gacchi
🟢	🤘	⨂	⏮	ttha	gacch + i + ttha	{445}	gacchittha	agacchittha
🟢	👆	⨀	⏮	iṁ	gacch + iṁ		gacchiṁ	agacchiṁ
🟢	👆	⨂	⏮	mhā	gacch + i + mhā	{516}	gacchimhā	agacchimhā, gacchimha, agacchimha
🔵	🤟	⨀	⏮	ā	a + gacch + ā	{519}	agacchā	gacchā, agacchittha, gacchittha
🔵	🤟	⨂	⏮	ū	a + gacch + ū	{519}	agacchū	gacchū
🔵	🤘	⨀	⏮	se	gacch + i + se	{516}	gacchise	agacchise
🔵	🤘	⨂	⏮	vhaṁ	gacch + i + vhaṁ	{516}	gacchivhaṁ	agacchivhaṁ
🔵	👆	⨀	⏮	aṁ	gacch + aṁ		gacchaṁ	agacchaṁ, agaccha, gaccha
🔵	👆	⨂	⏮	mhe	gacch + i + mhe	{516}	gacchimhe	agacchimhe
🟢	🤟	⨀	⏭	ssati	gacch + i + ssati	{516}	gacchissati
🟢	🤟	⨂	⏭	ssanti	gacch + i + ssanti	{516}	gacchissanti	gacchissare
🟢	🤘	⨀	⏭	ssasi	gacch + i + ssasi	{516}	gacchissasi
🟢	🤘	⨂	⏭	ssatha	gacch + i + ssatha	{516}	gacchissatha
🟢	👆	⨀	⏭	ssāmi	gacch + i + ssāmi	{516}	gacchissāmi
🟢	👆	⨂	⏭	ssāma	gacch + i + ssāma	{516}	gacchissāma
🔵	🤟	⨀	⏭	ssate	gacch + i + ssate	{516}	gacchissate	gacchissare
🔵	🤟	⨂	⏭	ssante	gacch + i + ssante	{516}	gacchissante
🔵	🤘	⨀	⏭	ssase	gacch + i + ssase	{516}	gacchissase
🔵	🤘	⨂	⏭	ssavhe	gacch + i + ssavhe	{516}	gacchissavhe
🔵	👆	⨀	⏭	ssaṁ	gacch + i + ssaṁ	{516}	gacchissaṁ
🔵	👆	⨂	⏭	ssāmhe	gacch + i + ssāmhe	{516}	gacchissāmhe
🟢	🤟	⨀	🔀	ssā	gacch + i + ssā	{516}	gacchissā	agacchissā, agacchissa, gacchissa
🟢	🤟	⨂	🔀	ssaṁsu	gacch + i + ssaṁsu	{516}	gacchissaṁsu	agacchissaṁsu
🟢	🤘	⨀	🔀	sse	gacch + i + ssase	{516}	gacchissase	agacchissase, agacchissasa, gacchissasa {517}
🟢	🤘	⨂	🔀	ssatha	gacch + i + ssatha	{516}	gacchissatha	agacchissatha
🟢	👆	⨀	🔀	ssaṁ	gacch + i + ssaṁ	{516}	gacchissaṁ	agacchissaṁ
🟢	👆	⨂	🔀	ssāmhā	gacch + i + ssāmhā	{516}	gacchissāmhā	agacchissāmhā, agacchissāmha, gacchissāmha
🔵	🤟	⨀	🔀	ssatha	gacch + i + ssatha	{516}	gacchissatha	agacchissatha
🔵	🤟	⨂	🔀	ssiṁsu	gacch + i + ssiṁsu	{516}	gacchissiṁsu	agacchissiṁsu
🔵	🤘	⨀	🔀	ssase	gacch + i + ssase	{516}	gacchissase	agacchissase
🔵	🤘	⨂	🔀	ssavhe	gacch + i + ssavhe	{516}	gacchissavhe	agacchissavhe
🔵	👆	⨀	🔀	ssaṁ	gacch + i + ssaṁ	{516}	gacchissaṁ	agacchissaṁ
🔵	👆	⨂	🔀	ssāmhase	gacch + i + ssāmhase	{516}	gacchissāmhase	agacchissāmhase

Note that some of the terminations are the same so in practice we will need to distinguish which tense is being used from context. Note also the irregular forms, particularly the aorist ones. The irregular aorists tend to be more commonly used than the regular forms. This is so prevalent that other grammar books such as Moggallāna Vyākaraṇa and Rūpasiddhi created rules to accommodate them.

Voices

Like Sanskrit, Pāḷi supports 3 different ways of expressing sentences, based on the “voice”:

1. the active voice (kattu) {444}-{452}
2. the passive voice (kamma)
3. the impersonal voice (bhāva)

However, in Pāḷi literature, bhāva (middle, or impersonal, voice) is seldom used, probably because most of the literature is based on the Buddha (or a disciple) “speaking” to a group of monks, which has then been transformed into a structure suitable for memorisation and chanting.

🔴 Passive verb endings Vattamānā Vibhatti {440}-{443}

A passive verb is created by using the attanopada (🔵) endings together with an infix of ya in between the root and the ending.

Note that passive verbs can only be formed based on roots that can be expressed in a “passive” (kamma) sense - some roots can only result in “active” verbs and cannot be transformed this way.

Example transformation (based on the root paca (to cook)):

paca + ya + te {440}
→ paca + y(cc)a + te {441}
→ paccate (“is cooked”)

The full set of attanopada (🔵) endings in the present tense vattamāna (▶️) across the 3 persons and plurality:

voice	number	person	tense	ending	transformation	result
🔴	🤟	⨀	▶️	te	paca + y(cc)a + te	paccate
🔴	🤟	⨂	▶️	ante	paca + y(cc)a + ante	paccante
🔴	🤘	⨀	▶️	se	paca + y(cc)a + se	paccase
🔴	🤘	⨂	▶️	vhe	paca + y(cc)a + vhe	paccavhe
🔴	👆	⨀	▶️	e	paca + y(cc)a + e	pacce
🔴	👆	⨂	▶️	mhe	paca + y(cc)a + ā + mhe	paccāmhe

Because the middle voice is seldom used, rule {518} allows these verbs to be retransformed into “reversed” parassapada (🟢) endings:

voice	number	person	tense	ending	transformation	result
🔴	🤟	⨀	▶️	ti	paccate + ti	paccati
🔴	🤟	⨂	▶️	anti	paccante + anti	paccanti
🔴	🤘	⨀	▶️	si	paccase + si	paccasi
🔴	🤘	⨂	▶️	tha	paccavhe + tha	paccatha
🔴	👆	⨀	▶️	mi	pacce + ā + mi	paccāmi
🔴	👆	⨂	▶️	ma	paccāmhe + ā + ma	paccāma

Typically we would use the parassapada (🟢) endings for passive verbs.

Other examples (using different rules):

kara (“to do”)
→ kara + ya + te {440}
→ kara + ī + ya + te {442}
→ kara + ī + ya + te {521}
→ karīyate (“is done”)

vaḍha (“to grow”)
→ vaḍha + ya + te {440}
→ vaḍha + ya(ḍḍha) + te {443}
→ vuḍḍhate (“is grown”)

🟠 Causative verbs

There are two types of causative verbs:

1. Causative verb in the active voice
2. Causative verb in the passive voice

There are 4 typical causative affixes (kārita) used to construct causative verbs {438}:

• ṇe
• ṇaya
• ṇāpe
• ṇāpaya

Just like for taddhita noun affixes, the ṇ is typically elided when joined to a root, leaving the remaining letters. {523}

Various vowel gradation transformations may also take place when joining the affix to the noun as per rule {483}.

Example transformation (based on the root paca (to cook)) and the causative affix ṇe:

paca + ṇe + ti {438}
→ paca + ṇe + ti {521}
→ pac + ṇe + ti {523}
→ pa(ā)c + e + ti {483}
→ pāceti (“(he) causes to cook”)

🟠 Causative (active) verbs {438}

The full set of causative versions of the vattamāna (▶️) forms of the root paca (to cook) using parassapada (🟢) endings:

voice	number	person	tense	ending	causative	transformation	result	irregular forms
🟠	🤟	⨀	▶️	ti	ṇe	pa(ā)ca + ṇe + ti	pāceti
🟠	🤟	⨀	▶️	ti	ṇaya	pa(ā)ca + ṇaya + ti	pācayati
🟠	🤟	⨀	▶️	ti	ṇāpe	pa(ā)ca + ṇāpe + ti	pācāpeti	pacāpeti
🟠	🤟	⨀	▶️	ti	ṇāpaya	pa(ā)ca + ṇāpaya + ti	pācāpayati	pacāpayati
🟠	🤟	⨂	▶️	anti	ṇe	pa(ā)ca + ṇe + anti	pācenti
🟠	🤟	⨂	▶️	anti	ṇaya	pa(ā)ca + ṇaya + anti	pācayanti
🟠	🤟	⨂	▶️	anti	ṇāpe	pa(ā)ca + ṇāpe + anti	pācāpenti	pacāpenti
🟠	🤟	⨂	▶️	anti	ṇāpaya	pa(ā)ca + ṇāpaya + anti	pācāpayanti	pacāpayanti
🟠	🤘	⨀	▶️	si	ṇe	pa(ā)ca + ṇe + si	pācesi
🟠	🤘	⨀	▶️	si	ṇaya	pa(ā)ca + ṇaya + si	pācayasi
🟠	🤘	⨀	▶️	si	ṇāpe	pa(ā)ca + ṇāpe + si	pācāpesi	pacāpesi
🟠	🤘	⨀	▶️	si	ṇāpaya	pa(ā)ca + ṇāpaya + si	pācāpayasi	pacāpayasi
🟠	🤘	⨂	▶️	tha	ṇe	pa(ā)ca + ṇe + tha	pācetha
🟠	🤘	⨂	▶️	tha	ṇaya	pa(ā)ca + ṇaya + tha	pācayatha
🟠	🤘	⨂	▶️	tha	ṇāpe	pa(ā)ca + ṇāpe + tha	pācāpetha	pacāpetha
🟠	🤘	⨂	▶️	tha	ṇāpaya	pa(ā)ca + ṇāpaya + tha	pācāpayatha	pacāpayatha
🟠	👆	⨀	▶️	mi	ṇe	pa(ā)ca + ṇe + mi	pācemi
🟠	👆	⨀	▶️	mi	ṇaya	pa(ā)ca + ṇaya + mi	pācayami
🟠	👆	⨀	▶️	mi	ṇāpe	pa(ā)ca + ṇāpe + mi	pācāpemi	pacāpemi
🟠	👆	⨀	▶️	mi	ṇāpaya	pa(ā)ca + ṇāpaya + mi	pācāpayami	pacāpayāmi
🟠	👆	⨂	▶️	ma	ṇe	pa(ā)ca + ṇe + ma	pācema
🟠	👆	⨂	▶️	ma	ṇaya	pa(ā)ca + ṇaya + ma	pācayama
🟠	👆	⨂	▶️	ma	ṇāpe	pa(ā)ca + ṇāpe + ma	pācāpema	pacāpema
🟠	👆	⨂	▶️	ma	ṇāpaya	pa(ā)ca + ṇāpaya + ma	pācāpayama	pacāpayāma

Note: not all forms of the above are encountered in Pāḷi literature - they represent the full theoretical set of causative forms that can be generated.

🟠🔴 Causative passive verbs

It is also possible to construct causative verbs in the passive voice, just by adding the passive ya affix after the causative affix. The ṇaya and ṇāpaya causative affixes are rarely used in the passive voice.

Example transformation (based on the root paca (to cook)) and the causative affix ṇe and the passive affix ya:

paca + ṇe + ya + ti {438,440}
→ paca + ṇe + ya + ti {521}
→ pac + ṇe + ya + ti {523}
→ pa(ā)c + ya + ti {483}
→ pāciyati (“is caused to be cooked”)

voice	number	person	tense	ending	causative	transformation	result
🟠🔴	🤟	⨀	▶️	ti	ṇe	pa(ā)ca + ṇe + i + ya + ti	pāciyati
🟠🔴	🤟	⨀	▶️	ti	ṇāpe	pa(ā)ca + ṇāpe + i + ya + ti	pācāpiyati
🟠🔴	🤟	⨂	▶️	anti	ṇe	pa(ā)ca + ṇe + i + ya + anti	pāciyanti
🟠🔴	🤟	⨂	▶️	anti	ṇāpe	pa(ā)ca + ṇāpe + i + ya + anti	pācāpiyanti
🟠🔴	🤘	⨀	▶️	si	ṇe	pa(ā)ca + ṇe + i + ya + si	pāciyasi
🟠🔴	🤘	⨀	▶️	si	ṇāpe	pa(ā)ca + ṇāpe + i + ya + si	pācāpiyasi
🟠🔴	🤘	⨂	▶️	tha	ṇe	pa(ā)ca + ṇe + i + ya + tha	pāciyatha
🟠🔴	🤘	⨂	▶️	tha	ṇāpe	pa(ā)ca + ṇāpe + i + ya + tha	pācāpiyatha
🟠🔴	👆	⨀	▶️	mi	ṇe	pa(ā)ca + ṇe + i + ya + mi	pāciyāmi
🟠🔴	👆	⨀	▶️	mi	ṇāpe	pa(ā)ca + ṇāpe + i + ya + mi	pācāpiyāmi
🟠🔴	👆	⨂	▶️	ma	ṇe	pa(ā)ca + ṇe + i + ya + ma	pāciyāma
🟠🔴	👆	⨂	▶️	ma	ṇāpe	pa(ā)ca + ṇāpe + i + ya + ma	pācāpiyāma

ākhyāta examples

▶️ vattamāna {414}

(used to express actions or events which occur) at the present

pāṭaliputtaṁ gacchati

Pāṭaliputta (city) / he goes
He goes to Pāṭaliputta.

sāvatthiṁ pavisati

Sāvatthi (city) / he enters
He enters Sāvatthi.

⏹ pañcamī {415}

(used to express)

• āṇatti (command)
• āsīsa (wish)

at a non-specific time-frame mode (anuttakāla)

karotu kusalaṁ

do / meritorious deed
(Please) do meritorious deed.

sukhaṁ te hotu

happiness / for you / be
May you be happy.

⏯ sattamī {416}

(to express)

• anumati (permission)
• parikappa (thought)

at a non-specific time (anuttakāla).

tvaṁ gaccheyyāsi

you / can go
You can go.

kima’haṁ kareyyāmi

what I (kiṁ + ahaṁ) / should do
What should I do?

🔄 parokkhā {417}

Note

paro’kkhā = para + akkha
beyond (i.e. past) / physical senses

(to be applied to express things happened in one’s absence, in a situation where the speaker is not a witness to) at a past time-frame mode (atīta - past).

supine kila’māha

in the dream / apparently / (he) said (kila + āha)
(He is supposed to have) said.

evaṁ kila porāṇā’hu

thus / apparently / ancient sages / said (porāṇā + āhu)
Ancient sages are supposed to have said thus.

Note

kila is a nipāta (particle) used in indirect hearsay where the speaker is uncertain of the truthfulness of what he heard or not a witness to the actual event that occurred. Sometimes it means an asseveration or an emphasis.

↩️ hiyyattanī {418}

(used to express events that took place) in past time (yesterday or before), either being witnessed or not witnessed (by the speaker)

so agamā maggaṁ

he / went / road
He went along the road.

te agamū maggaṁ

they / went / road
They went along the road.

Note

mā is a nipāta (particle) used to indicate negation, and is typically associated with verbs with hiyyattanī endings, but the verb is not associated with the past (or any other) tense. {420}

mā gamā

do not / gone
Do not go.

mā vacā

do not / said
Do not say.

⏮ ajjatanī {419}

(used to express things happened) in the near past starting from today, either being witnessed or not witnessed (by the speaker)

so maggaṁ agamī

he / road / gone
He has (just) gone along the road.

te maggaṁ agamuṁ

they / road / gone
They have (just) gone along the road.

⏭ bhavissanti {421}

(to be used) in the (expression of) future

so gacchissati, karissati

he / will go / will do
He will go, will do.

te gacchissanti, karissanti

they / will go / will do
They will go, will do.

🔀 kālātipatti {422}

(to be used) in (the expression of) an action that was past (without being materialised due to adverse conditions or lack of the supporting causes)

so ce taṁ yānaṁ alabhissā, agacchissā

he / if / that / vehicle / should have got / might have gone already
If he had gotten that vehicle, he might have gone already.

te ce taṁ yānaṁ alabhissaṁsu, agacchissaṁsu

they / if / that / vehicle / should have got / might have gone already
If they had gotten that vehicle, they might have gone already.

Example sentence constructions

kattu-vācaka vākya (active voice sentences)

An active voice sentence is a subject-principal sentence where Kattā (the agent, doer subject) is much more dominant.

1. kattu (subject) - Kathita Kattā or Vutta-kattā (predominant-subject) must be in the 1st vibhatti (①)
2. kamma (object) - Akathita-kamma or Avutta-kamma (non-principal object) must be in the 2nd vibhatti (②)
3. The verb must be either in the parassapada termination (🟢) in most cases or in the attanopada termination (🔵) with or without an affix of kattu sense. [Affixes of kamma sense such as ya, or kita affixes e.g. tabba, ta, etc. cannot be used]
4. The persons (subjects) and verbs should be concordant.

Subject (①)	Object (②)	Verb	Meaning
Puriso	odanaṁ	pacati	Man cooks the rice
Purisā	odanaṁ	pacanti	Men cook the rice
Buddhā	dhammaṁ	desenti	Buddhas teach the Dhamma
Tvaṁ	odanaṁ	pacasi	You cook the rice
Tumhe	odanaṁ	pacatha	You (plural) cook the rice
Ahaṁ	odanaṁ	pacāmi	I cook the rice
Mayaṁ	odanaṁ	pacāma	We cook the rice

Kamma-vācaka Vākya (passive voice sentences)

A passive voice sentence is an object-principal sentence where the Kamma (the thing being done, the object) is more visibly dominant by being in the 1st vibhatti (①).

The rules in a passive voice sentence are:

1. kamma (object) - Kathita Kamma or Vutta-kamma (predominant-object) must be in the 1st vibhatti (①)
2. kattā (subject) - Akathita-kattā or an Avutta-kattā (non-principal subject) must be in the 3rd vibhatti (③)
3. The verb should be mainly in the attanopada termination (🔵) or it can be in a reversed parassapada form (🟢) with an affix which has a Kamma-sense only. {440}-{443}.
4. The object (not the subject!) and verb should be in agreement.

Subject (③)	Object (①)	Verb	Meaning
Purisena	odano	pacīyate	The rice is cooked by man
Purisehi	odano	pacīyate	The rice is cooked by men
Buddhena	dhammo	desīyate	The Dhamma is taught by Buddha
Rājena	tvaṁ	dīyase	By king, you are given
Rājena	tumhe	dīyavhe	By king, you (plural) are given
Raññā	ahaṁ	dīye	By king, I am given
Raññā	mayaṁ	dīyāmhe	By king, we are given

bhāva (impersonal) middle voice sentences

It’s like a passive voice sentence except the focus is purely on the action and does not focus on either the subject or object. Therefore the verb tends to be in ⨀ (ekavacana) 🤟 (paṭhama purisa) form. If there is a subject, it can (but not always) be in the 3rd (③) or 6th vibhatti (⑥).

Subject (③)	Verb	Meaning
Devadattena	bhūyate	Devadatta’s being
Pabbatena	ṭhīyate	The mountain’s standing
Purisena	kathīyate	The man’s speaking

Causative sentences

There are four integral parts in a causative sentence:

1. Causative subject - Hetu-kattā (the prompter), or Payojaka-kattā (the mover)
2. Subordinate Object - Kārita-kamma (causative object)
3. Root-object - dhātu-kamma, the object of the root which has direct connection/meaning to the root
4. Causative verb

If the root is a dual-kamma-indicative root (dvikammaka) such as nī (to carry), duha (to milk), then there can be three objects.

Causative Subject (①)	Subordinate Object (②)	The root-object (②)	Causative Verb	Meaning
Puriso	purisaṁ	odanaṁ	pāceti	Man causes the other man to cook the rice
Purisā	purise	odanaṁ	pācenti	Men cause the other men to cook the rice

🔽 kibbidhāna (kita-affixed nouns) {524}-{623}

These are derivative nouns which are verbs transformed into “verb nouns” through the addition of suffixes.

These “verb nouns” are nouns, and have noun vibhatti endings applied to them, but can function like verbs in a sentence, or as auxiliary verbs such as the absolutive (also called gerund), the present participle and past participle.

verbal affixes

use	affix	rule	voice	liṅga
Verb or adjective	tabba	{540}	🔴 (sometimes 🔵)	variable
	anīya	{540}	🔴 (sometimes 🔵)	variable
	ṇya	{541}	🔴 (sometimes 🔵)	🚻 (sometimes variable)
	teya	by ca of {541}	🔴 (sometimes 🔵)	🚻 (sometimes variable)
	kha	{560}	🔴 (sometimes 🔵)	🚻 (sometimes variable)
	ta	{555}, {556}, {557}	🟢🔴🔵	variable

gerund or auxiliary verbs

use	affix	rule	voice	liṅga
to-infinitive	tave	{561}	-	-
	tuṁ	{561}, {562}, {563}	-	-
gerund	tuna, tvāna, tvā	{564}	no voice, but sense of “active”	-
Present participle or adjective or adverb	māna, anta	{565}	🟢 (or 🔴 with ya affix)	variable

Individual noun affixes

All of these can be variable gender, even when a gender is specified in the following table.

no	affix	rule	gender
1	ṇa	{524, 528, 529}	variable
2	a	{525, 526, 527, 553}	variable
3	ṇvu, tu, āvī	{527}	🚹
4	kvi	{530}	🚹
5	ramma	{531}	🚻🚹
6	ṇī, tu, āvī	{532}	🚹
7	yu	{533, 553}	variable
8	rū	{534, 535}	🚹
9	ṇuka	{536}	🚹
10	ra	{538}	variable
11	ricca	{542}	🚻
12	i	{551}	variable
13	ti	{552, 553}	🚺
14	ririya	{554}	🚺
15	tavantu, tāvī	{555}	🚹
16	ina	{558, 559}	🚻🚹
17	ratthu	{566}	🚹
18	ritu	{567}	🚹
19	rātu	{567}	🚺
20	tuka	{569}	🚹
21	ika	{570}	🚹

Example

Here is a worked-through example of how a root transforms into a kita-affixed noun (from caja, using the ṇa affix):

caja “to share generously”
→ caja {521}
→ caj + ṇa {529}
→ cajṇa (kārita) {621}
→ cajṇa {523}
→ ca(ā)ja (vuddhi) {483}
→ cāj(g)a {623}
→ cāga (noun) {601}
→ cāga + si (⨀①)
→ cāga + si(o) {104}
→ cāga + o {83}
→ cāgo (“charitable giving, sharing”)

Here is another example based on budha (to know):

budha “to know”
→ budha {521}
→ budh + ta {557}
→ budh + ta(dha) {576}
→ budh(d) + dha {611}
→ buddha (noun) {601}
→ buddha + si (⨀①)
→ buddha + si(o) {104}
→ buddha + o {83}
→ buddho (“The one who knows truth, who awakened”)

Note

Given the complexity of the above rules, it’s probably better just to recognise the completed forms whenever you come across them, but try and recognise the affix and associated meaning if you can.

⏬ uṇādi (uṇādi-affixed nouns) {624}-{673}

These are also derivative nouns that can behave like verbs or auxiliary verbs or future participles.

The majority of uṇādi-affixes, except some verbal affixes, are used mainly in the formation of individual nouns of varying genders.

Examples:

sama “to calm”
→ sama + tha {628}
→ samatha {601}
→ samatha + si (⨀①)
→ samatha + si(o) {104}
→ samatha + o {83}
→ samatho (“that which calms the mind, Samatha meditation”)

gaha “to take”
→ gaha {521}
→ ga(e)ha {629}
→ geha {601}
→ geha + si(aṁ) {104}
→ geha + aṁ {219}
→ gehaṁ (“home”)

sādhana

The process of attaching either kita or uṇādi affixes to create new words is called sādhana. Like kāraka (the syntax of case or vibhatti endings), sādhana-affixed words can be associated with specific usage meanings.

sādhana-affixed words can be associated with a voice (kattu, kamma, bhāva), in which case their usage can affect the “voice” of a sentence. This is especially true when a sādhana-affixed word is used as a verb-noun, i.e. it carries the “action” of a sentence.

Example sentences of kattu sādhana-affixed words

dānaṁ dinno devadatto

🚻⨀②(dānaṁ) 🚹⨀①(⏬(ta)(dinno) devadatto)
alms / gave (past participle) / Devadatta
Devadatta gave alms.

kumāro antepuraṁ gato

🚻⨀②(antepuraṁ) 🚹⨀①(kumāro ⏬(ta)(gato))
King’s palace / prince / went (past participle)
The prince went into the King’s palace.

so bhagavā sayam’pi cattāri saccāni buddho

🚹⨀①(so bhagavā ⏬(ta)(buddho)) ⏏️(sayam api) 🚻⨂②(cattāri saccāni)
that / Bhagavā / understood (past participle) / by oneself / just / four / truths
That Bhagavā (Buddha) understood the four truths just by himself.

Example sentences of kamma sādhana-affixed words

buddhena dhammo desito

🚹⨀③(buddhena) 🚹⨀①(dhammo ⏬(ta)(desito))
by Buddha / dhamma / taught (past participle)
The Dhamma is taught by the Buddha.

bhotabbo odano bhavatā

🚹⨀①(⏬(tabba)(bhotabbo) odano) 🚹⨀③(bhavatā)
eaten (future passive participle) / rice / Your Honourable
That rice is to be eaten by Your Honourable.

Example sentences of bhāva sādhana-affixed words

tassa gītaṁ

🚻⨀⑥(tassa) 🚻⨀①(🔽(ta)(gītaṁ))
of that person / sung (past participle)
Sung of that person.

sayitabbaṁ bhavatā

🚻⨀①(⏬(tabba)(sayitabbaṁ)) 🚹⨀③(bhavatā)
eating (future passive participle) / by Your Honourable
Eating by Your Honourable.

sādhana associated meanings and usage

The associated meanings are rather weak compared to case endings (vibhatti) and can sometimes have no significant role.

association	example	meaning	rule	usage
kattu	nāyako	the leader (who leads)	{527}	netī’ti nāyako
kamma	kammaṁ	work (which is being done)	{531}	karīyate tan’ti kammaṁ
bhāva	cāgo	generosity (act of giving, sharing)	{529}	cajate, cajanaṁ vā cāgo
karaṇa	vinayo	the discipline (by which one is trained)	{527}	vineti etenā’ti vinayo
sampadāna	dānīyo	the recipient monk (to whom alms should be given)		databbo assā’ti dānīyo
apādāna	bhemo	the fearsome object (from which people fear)	{627}	bhāyanti etasmā’ti bhemo
adhikaraṇa	ṭhānaṁ	the place (where people stand)	{548}	tiṭṭhanti tasmin’ti

sandhi (word transformations)

As Pāḷi is a spoken language, when words are combined to form a sentence, changes may occur between words (or sometimes even within a word) to make the sentence “flow” or “sound” better.

In English, words are often shortened, i.e. “can not” to “can’t”, or “I will” to “I’ll”. This happens very frequently in Pāḷi, especially when two vowels become adjacent to each other from two words next to each other, and there are complex rules governing how such transformations occur.

As an example {10}:

tatra + ayaṁ + ādi
= tatra + aā yaṁm + ādi
= tatrā’ya’mādi
= tatrāyamādi

When there are two adjacent vowels from joining two or more words together, sandhi typically involves deleting one of the vowels, and then either retaining, lengthening or changing the remaining vowel. Sometimes the consonant next to the vowel may also be changed.

The overall effect is to make the words easier to say as a group. Although tatrāyamādi may seem like a single word because of sandhi, a listener should still regard them as separate words when trying to understand a sentence. However, some sandhi combinations of common words are so often encountered they have acquired distinctive meaning as combined words.

Another example {11}:

tatra + abhiratiṁ + iccheyya
= tatra + aā bhiratiṁm + iccheyya
= tatrā’bhirati’miccheyya
= tatrābhiratimiccheyya
“… should prefer enjoying at that …”

This introduction will not go into detail on these rules but the reader should be aware that endings and beginning letters in words may change in a sentence. Eventually, as reading proficiency increases, these changes will become easier to recognise.

Conclusion

This introduction is compact and dense, and not every feature of Pāḷi is explained here. However, there is enough information in here to start reading the Tipiṭaka with the aid of a good dictionary.

The sentences in the Tipiṭaka are often repetitive and used in similar forms again and again. They are intentionally structured this way to assist in memorisation and chanting. So, once a complex sentence is parsed, it may be encountered again and again and reading will become easier and easier over time.

Other Opinions

• Norman - The origin of Pāli and its position among the Indo-European languages (1988) (Norman, 1988)
  Pāli is a Middle Indo-Aryan language that originated not as a specific regional tongue but as a composite literary dialect used for the Buddhist canon. Descending from the Indo-European language family, it developed from the various Prakrits spoken in North India following the migration of Indo-Aryan peoples and their interaction with indigenous Dravidian and Munda populations. The Buddha and his followers preached in these local vernaculars, and Pāli emerged from the subsequent effort to standardise these diverse teachings into a single language for the Theravādin tradition. While primarily based on a Western Prakrit, Pāli retains features from other dialects, such as Māgadhī, reflecting its mixed origins and the process of translation and compilation. The language was largely fixed when the canon was committed to writing in the first century B.C., after which it underwent further editorial standardisation, including significant Sanskritisation and, in later periods, minor influences from Sinhalese and the work of medieval Pāli grammarians.
• Cousins - The Early Development of Buddhist Literature and Language in India (2013) (Cousins, 2013)
  L.S. Cousins re-examines the early development of Buddhist literature, arguing that the Pali Canon was first systematically written down not in Ceylon, but on the Indian mainland around the first century B.C. before being imported to the island. This written canon was based on a largely non-sectarian oral tradition of chanted texts (bhāṇaka tradition) that had been developing since the Buddha’s time (c. 400 BCE), and its initial contents included the four main Nikāyas, Vinaya, Abhidhamma, and several Khuddakanikāya texts. The language used was not a spoken dialect but an evolving written lingua franca, which Cousins terms ‘Old Pali,’ derived from a post-Asokan epigraphic Prakrit. This written language subsequently underwent a process of Sanskritisation, becoming ‘Hybrid Pali,’ before being standardised into the classical Pali we know today by commentators around the fourth or fifth century A.D., who consistently referred to it as Māgadhī.
• Levman - Cultural Remnants of the Indigenous Peoples in the Buddhist Scriptures (2014) (Levman, 2014)
  Bryan Levman’s article argues that Buddhist scriptures contain significant cultural remnants from India’s indigenous Munda, Dravidian, and Tibeto-Burman peoples, which have been obscured by a later “brahmanisation” that historicised the Buddha within the dominant Indo-Aryan (IA) tradition. The author identifies this indigenous influence by examining the hostility between IA immigrants and eastern ethnic groups like the Buddha’s Sakya clan, whose distinct socio-political organisation (gaṇasaṅgha), rejection of the Brahmanical class system, and non-Aryan marriage customs are evident in the texts. Furthermore, key Buddhist concepts and practices are traced to autochthonous roots, including the idea of the Mahāpuruṣa (Great Man), the veneration of trees and serpents (nāgas), the culture of sacred groves, and the unique funeral rites described for the Buddha’s parinibbāna, all of which differ significantly from Vedic norms.
• Levman - The language of early Buddhism (2016) (Levman, 2016)
  The earliest recoverable language of Buddhism was not a specific regional dialect but a Middle Indo-Aryan (MIA) koine, an inter-dialectal common language created through simplification and the levelling of distinct features to facilitate communication. Evidence for this underlying language is found by comparing parallel Buddhist transmissions in Pāli, Gāndhārī, and other Prakrits, where translation errors and ambiguities (hyperforms) point to a shared, phonologically reduced source. Key characteristics of this koine include the weakening or elimination of intervocalic stops, the reduction of aspirated stops to simple aspirates, the assimilation of consonant clusters, and the merging of sibilants. The formation of this simplified language was driven not only by the need to harmonise different MIA dialects but was also significantly catalysed by contact with other languages in the ancient Indian “Sprachbund,” such as Dravidian, Munda, and Tocharian, whose speakers lacked features like phonemic voicing or aspirated stops, thereby accelerating the development of a more universally intelligible form of MIA.
• Karpik - The Buddha taught in Pali - A working hypothesis (2019) (Karpik, 2019)
  This paper challenges the scholarly consensus that the Buddha taught in multiple dialects like Māgadhī, a theory known as the Multiple Oral Transmission Theory (MOTT). Instead, it argues for a Single Oral Transmission Theory (SOTT), proposing that the Buddha used a single standard language, which is the language we now call Pali. The author contends that oral translation of the vast canon was impractical, unnecessary due to the mutual intelligibility of Indo-Aryan dialects, and explicitly discouraged by the suttas’ emphasis on word-for-word transmission. The traditional identification of the Buddha’s language as Māgadhī is based on a misreading of the term Māgadhabhāsā in the commentaries, which actually defined it as Ariyaka (Indo-Aryan), a pan-regional term. The paper posits that Pali was a natural, pre-existing standard language based on a western dialect, a precursor to the widespread Epigraphic Prakrit, and that the Buddha adopted this sociolect of the educated to ensure a fixed, albeit phonetically fluid, transmission of his teachings across different regions. This working hypothesis refutes the MOTT by explaining dialectal variants in Pali as natural artefacts of a single transmission rather than remnants of a translation, thereby aligning with the Theravada tradition’s claim.
• Levman - The language the Buddha spoke (2019) (Levman, 2019)
  Bryan Levman argues that the Buddha spoke not Pāli as we know it, but an earlier, related vernacular or koine that differed in word form, morphology, and sometimes meaning. Evidence for this earlier linguistic layer is uncovered by applying comparative historical linguistics to the various Pāli recensions and parallel transmissions in other Prakrits, such as Gāndhārī. This method reveals that many variant readings and semantic ambiguities in the canon can be traced back to a common ancestral form in the koine which had undergone significant phonetic simplification, such as the weakening of intervocalic consonants into glides or the reduction of various aspirated stops to a simple aspirate. Later redactors, transmitting the teachings into their own dialects, reinterpreted these simplified sounds, resulting in the multiple cognate forms found today, a process further complicated by influences from prestigious northwestern dialects and the assimilation of non-Aryan words. Levman concludes that this linguistic evolution is consistent with the Buddhist doctrine of impermanence and supports the Buddha’s historicity.
• Karpik - A Reply to Bryan Levman’s The Language the Buddha Spoke (2019) (Karpik, 2019a)
  In a comprehensive rebuttal to Bryan Levman’s theory, Stefan Karpik rejects the proposition of an early “pre-Pali” lingua franca that evolved into a later, mutually unintelligible Pali. Karpik argues that Levman’s reconstructions are undateable, do not prove a lingua franca over other possibilities like bilingualism, and fail to explain why an original transmission would be discarded. The core of Karpik’s critique focuses on Levman’s flawed linguistic methodology, particularly the incorrect assumption of directionality in sound changes, which leads to proposing numerous implausible and unattested “fortitions” (sound strengthening) where lenition, orthographic errors, or dialectal variations are more likely explanations. Karpik further dismisses the claim of rapid linguistic change and mutual unintelligibility due to a lack of syntactical evidence, accusing Levman of circular reasoning and misrepresenting other scholars, ultimately advocating for a simpler model of steady linguistic development where Pali itself is the pre-Aśokan language of the Buddha.
• Pind - Did the Buddha address the monks in Māgadhī? (2021) (Pind, 2021)
  This paper argues that the distribution of the Pāli vocatives bhikkhave and bhikkhavo in the canon is determined by prosody and syntax, not by historical or dialectal origins as is commonly assumed. Refuting the theory that bhikkhave is a preserved “Māgadhism” reflecting the Buddha’s speech, the author demonstrates a structural and complementary distribution between the two forms. Bhikkhavo functions as an emphatic vocative, used at the beginning of prose sentences—notably in stock introductory formulas—and exclusively in verse, a pattern that parallels the use of the emphatic vocative bhaddante. Conversely, bhikkhave is the non-emphatic, or clitic, form which never occurs initially and is always attached to a preceding word within a sentence, mirroring the usage of the clitic bhante. This distinction is therefore intrinsic to the syntactical and recitational structure of the canonical language itself, making explanations based on regional dialects or later scribal interpolations unnecessary.

Cousins, L. S. (2013). The Early Development of Buddhist Literature and Language in India. Journal of the Oxford Centre for Buddhist Studies, 5, 89–135.

Gombrich, R. F. (2018). Buddhism and Pali. Mud Pie Slices.

Hinüber, O. von. (1996). A Handbook of Pāli Literature (A. Wezler & M. Witzel, Eds.). Walter de Gruyter.

Karpik, S. (2019a). A Reply to Bryan Levman’s “The Language the Buddha Spoke.” Journal of the Oxford Centre for Buddhist Studies, 17, 106–116.

Karpik, S. (2019b). The Buddha taught in Pali: A working hypothesis. Journal of the Oxford Centre for Buddhist Studies, 16, 80–86.

Levman, B. (2014). Cultural Remnants of the Indigenous Peoples in the Buddhist Scriptures. Buddhist Studies Review, 30(2), 145–180. https://doi.org/10.1558/bsrv.v30i2.145

Levman, B. (2016). The language of early Buddhism. Journal of South Asian Languages and Linguistics, 3(1), 1--41.

Levman, B. (2019). The language the Buddha spoke. Journal of the Oxford Centre for Buddhist Studies, 17, 63–105.

Norman, K. R. (1988). The origin of Pāli and its position among the Indo-European languages. Journal of Pali and Buddhist Studies, 1, 1–27. https://doi.org/https://doi.org/10.20769/jpbs.1.0_1

Norman, K. R. (1997). A Philological Approach To Buddhism: The Bukkyō Dendō Kyōkai Lectures 1994: Vol. V.

Pind, O. H. (2021). Did the Buddha address the monks in Māgadhī? Journal of the Oxford Centre for Buddhist Studies, 20, 83–106.

Footnotes

1. ṃ (niggahīta) can be written η, ṁ or ṃ ↩