A Brief Introduction to Buddhist Hybrid Sanskrit

Last year, I celebrated my 40th birthday (whoo!). Thanks to some kind gifts from friends and family I bought some books that I had long kept on my wishlist.  This includes books on Buddhist Hybrid Sanskrit:


This book is titled Buddhist Hybrid Sanskrit Grammar and Dictionary by Professor Franklin Edgerton. It is an older book, first published in 1953 but there doesn’t seem to be a lot of books published since then on the subject. I was surprised to see this book included both a grammar text and dictionary, which was great. The drawback is that this book assumes you already have a good grasp of “standard” Sanskrit and can be pretty dense.

What is Buddhist Hybrid Sanskrit?

Well, the history of northern Indian languages somewhat mirrors the history of Western-European languages in that there is a “mother” language that either directly spawned or influenced other languages. For Western Europe, this is Latin. For northern India,1 this was Sanskrit.

But the historical Buddha, Shakyamuni, didn’t speak Sanskrit as a first language. He spoke an ancient form of Magadhi (still spoken today) and many of his followers did too.

Once the Buddha passed away, much of his sermons were recited and preserved among early disciples in various vernacular or “prakrit” languages throughout northern India. These local vernaculars all descended from Sanskrit and were close enough to one another that disciples could translate from one to another pretty effortlessly. Of these, Pali was probably the most common because it was a prakrit that was widely used for trade and commerce.

This was not an accident either, as according to the book, there are Buddhist sutras in which the Buddha encouraged the use of local languages because they were more accessible. In the Pali Canon, particularly the Cullavagga 5.33, the Buddha chastises a couple high-caste followers who wanted to put his teachings into the liturgical form (i.e. “Vedic” or Sanskrit prose) to avoid “corruption” by various students:

The Lord Buddha rebuked them: deluded men, how can you say this? This will not lead to the conversation of the unconverted…And he delivered a sermon and commanded (all) the monks: You are not to put the Buddha’s words into Vedic. Who does so would commit a sin. I authorize you, monks, to learn the Buddha’s words each in his own dialect. (trans. Edgerton)

Similar passages appear in the Dharmagupta sects copy of the sutras, and also in Chinese translations of lost texts in India. The point being: the Buddha wanted the teachings to be as accessible as possible, so it was encouraged to use everyday language and avoid liturgical language.

Where does Buddhist Hybrid Sanskrit come in?

For reasons not well-understood, the Buddhist community centuries later eventually changed their mind and started using Sanskrit more and more. At first this started by using Sanskrit terms in the teachings they had passed down generation after generation, but eventually the terms and language became more “Sanskritized” to the point that they almost looked like standard Sanskrit (as defined by Pāṇini).

Thus, different generations of Buddhist texts looked more and more like Sanskrit, but a discerning eye can still see the prakrit origins of the words and grammar. Thus it is not “standard” Sanskrit, but “Buddhist Hybrid” Sanskrit.

One of the earliest examples of Buddhist Hybrid Sanskrit is a little-known text called the Mahavastu composed probably in the 2nd century B.C., which, like all such texts, contains both verse and prose form.2 Edgerton points out that the verse sections definitely look like they are composed in a Prakrit language, but with some Sanskrit “polish” while the prose sections look more like Sanskrit. But, as noted by Edgerton, the Mahavastu has a lot of grammatical corruptions in it too as a result of this sometimes clumsy conversion from Prakrit to Sanskrit.

Additionally, the same Buddhist text, if multiple versions were composed, looks more and more like Sanskrit with each new version.

What are some examples of Buddhist Hybrid Sanskrit texts?

A number of texts that are now a core part of the Mahayana Buddhist canon were composed in various stages of Buddhist Hybrid Sanskrit. Loosely, the three stages are:

  1. Early conversion from Prakrit to Sanskrit (often with grammatical mistakes or corruptions): Mahavastu
  2. Middle stage: The Lotus Sutra, The Gandhavyuha Sutra, The Amitabha Sutra and The Sutra on the Buddha of Immeasurable Life.
  3. Latter stage (texts are mostly converted to Sanskrit, only some natively Prakrit grammar remains): The Diamond Sutra3, The Lankavatara Sutra

Which Prakrit was used?

It’s not always clear which of the many Middle-Indic or “Prakrit” languages was used as the source material. It may not always be the same one. For example Pāli language terms like those in the Pali Canon were often used, but the grammar might be a different Prakrit. Another prakrit used was Gandhari language, especially in the Immeasurable Life Sutra and Amitabha Sutra.

Oftentimes, it is not distinguishable to Edgerton which Prakrit was used because the grammar used was common to most Prakrit languages at the time.

What are some grammatical examples of Buddhist Hybrid Sanskrit?

This is by no means an exhaustive list. The Edgerton book above is quite huge, and even between texts, the grammar and degree of Sanskritization varies. Also, I am a beginning student of Sanskrit, and definitely not an expert, but I tried to find examples that at least I understood and could convey here.

Grammatical Case Standard Sanskrit ending Buddhist Hybrid Sanskrit
ā-stem Accusative Sing. ām aṃ
ā-stem Instrumental Pl. ābhiḥ āhi
ā-stem Ablative Sing. āyāḥ ātas (with variations)
ā-stem Genitive Pl. ānām ānā

Other noteworthy examples:

  • N-stem words were frequently converted to Sanskrit a-stem endings. For example mūrdhan might become mūrdhaḥ/mūrdho.
  • Dual-endings in Sanskrit, were evidentially not used much (if at all) in the original Prakrit, so when the text was converted to Sanskrit, a dual-ending was awkwardly “bolted on” where applicable.
  • Unlike Sanskrit where verbs have a special “root” (which doesn’t always look like verb in present form), the original Prakrit just used present stem.
  • Sandhi rules for the Prakrit were not the same as Sanskrit. For example, when the first word ends in a vowel, while the second word starts with the same vowel (a and a for example), rather than combining them into a longer vowel (ā for example), one would just get dropped.
  • Another example of differences with Sandhi: unlike Sanskrit, final endings like as and ar might sometimes just drop the “s” and “r” rather than altering the sound.

Anyhow, this page has been an amateur’s attempt to make sense of and distill a complex subject to a wider audience. Hope you enjoyed and if you happen to know about the subject, please post comments below! Thanks!

1 Southern Indian languages are an entirely different subject, but fascinating in its own right. I just don’t know much about it.

2 Even in the earliest Buddhist sutras, sometimes you see both prose and verse forms, but this “mixture” becomes more and more common in later Buddhist texts.

3 I’d venture that all the existing Perfection of Wisdom sutras are similarly written in Buddhist Hybrid Sanskrit depending on when they were composed. The subject of when each of the Perfection of Wisdom sutras was composed is still a matter of debate though.


Devanagari and Siddham: a brief comparison


My studies of Sanskrit continue and I am starting to get more familiar with the Devanagari writing system used in the textbook, and for modern Sanskrit studies.  However, I also branched out a little bit into learning Siddham as well.

As mentioned in an earlier post, Sanskrit language has no “native” script so various writing systems have been used for time.  Devanagari is widely-used script across many northern-Indian languages (Marathi, Hindi, etc) and is great for reading/writing Sanskrit, however, many ancient Buddhist texts were written before Devanagari’s invention, so many texts used older writing systems.

Siddham on the other hand was frequently used for ancient Buddhist writings and is still used in esoteric Buddhist sects, particularly Shingon Buddhism in Japan.  So, I’ve dabbled in Siddham a little along with learning Devanagari.

In one of my Japanese language books1 on Sanskrit and Siddham, はじめての「梵字の読み書き」入門 (hajimete no “bonji no yomikata” nyūmon) contains a great chart comparing the Siddham alphabet with Chinese characters, Japanese pronunciation, etc. The chart is very useful, but obviously I can’t post it here since it belongs to the author.  So, instead, I am constructing my own chart using information gleaned from the book and my own studies, plus additional information. This chart is not complete, but I will continue updating as time goes on.

Devanagari Siddham Romanization Japanese Katakana
Siddham a a as in “uh”
Siddham aa ā as in “ah-ha!” アー
Siddham i i as in “ee”
Siddham ii Ī as in “ee” but longer イー
Siddham u u as in “oo”
Siddham uu ū as in “oo”, but longer ウー
Siddham e e as in “ay” エイ
Siddham ai ai as in “aye” アイ
Siddham o o as in “oh”
Siddham au au as in “ow” アウ
Siddham ri
Siddham rii リー

Since Siddham and Devanagari are “genetically” related scripts (though, one did not descend from the other), you can see how some letters look similar, while others look quite different.

Note: Siddham is available in Unicode, but it is difficult to view without the right fonts, so I am (shamelessly) posting image files from Wikimedia Commons.


1 Another book I highly recommend if you’re serious about the subject is 梵字必携―書写と解読 (bonji hittei: shakyo to kaidoku)

Where Does Namu Come From?

The term “namu” shows up a lot in East Asian Buddhism, for example in devotional chants such as:

  • Namu myōhō renge kyō – Praise to the Lotus Sutra which is a Japanese chant for Nichiren Buddhism.
  • Namu Amita Bul – “Praise to the Buddha of Infinite Light” in Korean.
  • Nam mô A Di Đà Phật – same as above, but Vietnamese.
  • Namu Amida Butsu – same as above, but Japanese.
  • Nā mó guān shì yīn pú sà (南无观世音菩萨||南無觀世音菩薩) – “Praise to Avalokitesvara Bodhisattva” in Chinese, with both Simplified and Traditional characters shown.

You can see how each language has a variation on the word “namu”: “nam mo”, “na mo”, etc.  What the heck is this?

This is actually from Sanskrit language, which I’ve been studying lately.  The original, vanilla term is namas (नमस्) which according to the Sanskrit dictionary means “bow, obeisance, reverential salutation, adoration” etc, etc.  This term is not limited to Buddhism either.  It shows up a lot in Indian culture, and even in Yoga when you say to one another “namaste”.¹

Now, here’s the funny part.  Sanskrit words frequently undergo sound changes called “sandhi”, which I’ve talked about here and here and here among other places.  This means that people don’t always say “namas this” and “namas that”.  Sandhi rules in Sanskrit allow for a sound to get ready for the next sound by changing a little to more accurately fit the position of the tongue in the mouth.

The word namas changes like so, depending on the following sound:

  • namo (नमी) if the following sound is a vowel, or a voiced consonant (j, jh, g, gh, d, dh, b, bh) or by nasal sounds (r, l, h, n, m).
  • namaś (नमश्) if the following sound is a c or ch
  • namaṣ (नमष्) if the following sound is a ṭ or ṭh
  • namas (नमस्) if the following sound is a t or th
  • namaḥ (नमः) if the following sound is a k, kh, p, ph, ś, s, ṣ or it’s the last word in the sentence.

So, for example in Buddhist liturgy to say “praise to the Buddha of Infinite Light”, the words before Sandhi would be namas amitābhāya but due to Sandhi rules become namo’mitābhāya.²  This is probably what got imported into China as Buddhism spread there.

From there, the “namo” was translated using suitable Chinese characters that phonetically sound the same.  In this case: 南無 which even today in Chinese is pronounced nāmó.

But also since Chinese and Chinese Buddhist liturgy were imported into other neighboring countries and given more local pronunciation. The Chinese characters would have been the same, but every country would read/pronounce them slightly different.

Thus “namo” became “namu” in places like Korea and Japan, but still “namo” in Vietnamese.

¹ Technically, namaste is a Hindi word, not Sanskrit, but Hindi is clearly derived from Sanskrit. The easiest way to understand this is that Sanskrit is to northern-Indian languages (Hindi, Gujarati, Marathi, Bengali, etc) what Latin is to Western-European languages (French, German, English, Italian, etc).

² The apostrophe is because the “a” of the second word gets dropped.  This is a special rule in Sanskrit where aḥ/as + a changes to o ‘ .  Why?  It just does.

Sandhi Rules Are Kicking Me In The Pants

I am currently on Lesson 10 of my Sanskrit language textbook, and getting pretty immersed in the Sandhi rules that govern how one word blends into another. The rules are actually very sensible changes that allow the last sound of a word to “get ready” for the first sound of the next word. Not all words undergo sandhi changes, but there are a lot of rules that govern the ones that do. The textbook I use explores these rules over several lessons to give students time to master one rule at a time. Here’s a simple, but painful example.

Anyhow, imagine I want to say “you are beautiful”. The word for “beautiful” is sandaraḥ (सन्दरः) and the singular nominative of “you” is tvam (त्वम्). Thus, grammatically, you can either say “you beautiful are”:

त्वम् सन्दरः असि
tvam sandaraḥ asi

Or, in Sanskrit, you can shorten this to “beautiful you”:

सन्दरः त्वम्
sandaraḥ tvam

The second form is often used because the “to be” verb is often understood. But now I have to apply sandhi rules to make it more natural sounding.

First, the “aḥ” at the end of the first word runs up against the “t” of the second word. According to sandhi rules, the “aḥ” followed by a “t” becomes an “s” in this case:

सन्दरस् त्वम्
sandaras tvam

But as part of Sandhi, words often combine too, unless the first word ends in a vowel (including ṃ and ḥ) and the second is a consonant.1 So, now we have to combine the words into:


Thus if you want to flatter a lady in Sanskrit, this is what you would say, and how you would write it.

1 I am still not 100% clear when words combine and when they don’t, but the use-cases seem to be:

  • Consonant + consonant = combine
  • Vowel + vowel = combine
  • Consonant + vowel = combine
  • Vowel + consonant = not combine

Amitabha Buddha in Early Indian Buddhism

While continuing my read of the book Bones, Stones, And Buddhist Monks: Collected Papers On The Archaeology, Epigraphy, And Texts Of Monastic Buddhism In India, I was struck by a certain passage regarding an inscription where the Buddha Amitabha is mentioned.  The book explains that the inscription was made in the 26th year of King Huveṣka (Huvishka), which is roughly 166 C.E. by a layperson of some wealth. The inscription is said to be the “earliest, indeed the only, reference to the Buddha Amitābha in Indian inscriptions and is, therefore, one of the few hard facts we have concerning this Buddha and his cult in India proper”. (pg. 39)

The inscription reads (translations by author above):

bhagavato buddha amitābhasya pratimā pratiṣṭhapita buddha pūjāye

“[an] image of the Blessed One, the Buddha Amitābha, was set up for the worship of the Buddha”.

Then ends with:

imena kuśalamulena sarva(satana)anuttarajñānaṃ prātp(i)m (bha) (va) (tu)

“through this root of merit may there be the attainment of supreme knowledge by all beings.”

I find this really fascinating for a few reasons:

  • It definitely reaffirms that the existence of Amitabha Buddha in India was very sparse until the 2nd century C.E.
  • The fact that northern India at this time was ruled by the Kushan Empire out of Pakistan (and not a native Indian empire) suggests that this could have been imported from outside of India.  The layperson above is described as belonging to a family of merchants, caravan traders and bankers.  So, it’s likely had travelled to and from other parts of the Kushan Empire.
  • Also, the fact that inscription was made in reference to the reigning king at that time further suggests influence from the Kushan culture.

Further, the Kushan Empire was originally Zoroastrian, though it later became Buddhist under King Kanishka (Huvishka’s predecessor).  This definitely lends credence to the old theory that Amitabha had origins outside of India, or at least borrowed from Zoroastrian culture.  Perhaps local traditions mixed with imported iconography from Persia, but the Princeton Dictionary of Buddhism¹ suggests that iconography of Amitabha Buddha in the Gandhara region, the heart of the Kushan Empire, predates anything in India proper.

Interesting stuff.

¹ I am the lucky owner of a copy.

Sandhi Rules and Sanskrit

My studies of Sanskrit language have continued and I am now, as of writing, have reached chapter 9 of my textbook.  However, things are definitely getting more uphill now as we’ve reached the dreaded “sandhi” (संधि) rules of Sanskrit.

Sanskrit, like nearly every language in the world, has sound changes that naturally happen when certain sounds come together.  This is a natural human phenomena to reduce friction in spoken language, but also influences written language as well.  In the case of Sanskrit, this joining of sounds, and changing is called “sandhi”, but the term sandhi has been expanded in linguistics to apply to any sound changes in any language.

Anyhow, sandhi rules in Sanskrit are particularly complicated.  They make a lot of sense when you get the hang of them, but getting used to all the myriad rules takes a lot of time.

Some rules are kind of straightforward, such as vowel endings.

For example, when certain vowels come together, they either merge or just morph into a different vowel.  For example:

अत्र अश्वः atra aśvaḥ (here horse)

The “a” at the end of the first word, followed by the “a” in the second word, would merge to become ā:

अत्राशवः atrāśvaḥ

Notice how the word gets merged in the process. This commonly happens in Sanskrit, hence the words look super long, but in actuality it’s a multiple words combined together.

The vowel sandhi rule above is not too bad, though.  It gets more complicated with consonant endings and such.  For example:

रामः गच्छति rāmaḥ gacchati (Rama goes)

changes to:

रामो गच्छति rāmo gacchati


रामः तिष्ठति rāmaḥ tiṣṭhati (Rama stands)


रामस्तिष्ठति ramastiṣṭhati

Further, sandhi rules aren’t limited to word endings. For example “n” will sometimes become ṇ if preceded by an “r” or ṛ as in:

गजः (gajaḥ, the elephant) ->
गजेन सह (gajena saha, “with the elephant”)

पुत्रः (putraḥ, the son) ->
पत्रेण सह (putreṇa saha, “with the son”)

…but wait! There’s more!

Anyhow, the key to learning Sandhi rules is to see them in practice. Memorizing the rules is nearly impossible because they’re so complicated, but if you see enough examples of sandhi rules in action, then things tend to make intuitive sense.

It takes some patience to get used to it, but over time, you’ll see how sandhi rules smooth out awkward sound combinations in Sanskrit and make it such a lovely language to read and speak.

Sanskrit Misuse in Buddhist Teachings

This is why I do it. This is why I study Sanskrit:

Buddhist books and their authors sometimes like to give themselves an air of legitimacy by sprinkling in ancient Sanskrit phrases but the usage of sanskrit in this paragraph is unfortunately incorrect

Case in point, the original Sanskrit phrase of the nembutsu (namo amida butsu) is actually namo’mitābhāya (नमाेऽमिताभाय) not Namo-Amita-Buddha. Second Namo does not mean “I put my trust in”, etc. The word is variation on namaṣ (नमष्) meaning “praise” or “hail”.2 That’s why people say “namaste” in Yoga classes in such: you’re greeting/praising the divinity in the other person.

Also, for clarity, the phrase for “to trust or take refuge in” is śaraṇaṃ gam (शरणम् गम्).1 There are probably other words too but I am not aware of them.

So the actual translation of namo’mitābhāya is “Praise to the Buddha of Infinite Light” not “I entrust everything to …”. The translation provided has an overtly sectarian bias that ordinary readers would miss. In any case this translation is simply incorrect.

I would encourage anyone who is spiritual or religious to learn the ancient language of your tradition whether that be Koine Greek for Christianity, Hebrew for Judaism, or Arabic whatever. Don’t blindly believe what religious leaders tell you. Be skeptical, and validate for yourself.

1 Hence when taking refuge in the Three Treasures people say buddhaṃ śaraṇaṃ gacchāmi, etc.

2 Sandhi rules, or sound-shifts in Sanskrit, is a topic for another post.