[Image] Patañjali
sculpture by Natalia Rosenfeld Buddhist texts are preserved in a wide variety of languages. However in India for more than 1000 years, from just before the common era, most texts were composed and preserved in a variety Sanskrit. Sanskrit was not only the language of the Buddhist texts during this period, but was the literary language of all India, so that Buddhist exegesis was composed in Sanskrit also. Sanskrit was not restricted to Brahmins for most of the common era. Despite grammars dating back to ca. 4th century BCE, which served as prescriptive models, Sanskrit has always existed in a variety of dialects and there was influence in both directions with Prakrits (or vernacular languages that derive from one or other Sanskrit dialect).

Of course in Sri Lanka and countries under the influence of the Sri Lankan Buddhists, the North Indian mixed Prakrit we now call Pāli was important. And we are beginning to understand the importance of the language of Gandhāra (usually called Gāndhārī) and it's influence particularly in Central Asia and China.
Even in India there are some variations in pronunciation. Buddhists outside of India have struggled with Sanskrit pronunciation. Sanskrit contains sounds that are not part of the Tibetan, Chinese, Japanese, or English sound palette. The first three nations settled on standard pronunciations a long time ago. In the individualistic English speaking world the problem of pronunciation continues, with a wide variety of mispronunciations being common. So we may here the syllable saṁ (as in saṁgha) being pronounced like the English words sang, sung, sum, and (the name) Sam. Sometimes all four by one person. 
I have written about pronunciation before in 2009. I've given up trying to correct people's pronunciation in writing. English speakers have a strong tendency to pronounce written words as though they were English - with all the vagaries of English pronunciation. And because we tend to learn our Buddhist vocabulary from written sources we seem to be stuck with the morasse of mispronunciation. In the short term I don't think anything I can say will change things. Certainly nothing I write will change things. Although a few of us make forays into Buddhist canonical languages, I don't know anyone who is familiar with the International Phonetic Alphabet that phoneticists use to disambiguate spoken sounds.
But pronunciation is important, and not just for aesthetic reasons. An ancient commentary on the grammar of Sanskrit, the Vyākaraṇa-mahābhāṣya or Major Commentary on Grammar of Patañjali (also credited with composing the Yoga Sūtras), criticises bad pronunciation (mithyā prayukta) in this way: duṣtaḥ śabdaḥ svarato varṇato vā mithyā prayukta na tam artham āha; sa vāgvajro yajamānaṃ hinasti yathendraśatruḥ svarato 'parādhāt. A faulty word, badly pronounced due to a misplaced accent or badly articulated sound does not convey the intended meaning. A word-lightning bolt kills the sponsor of the sacrifice, just as a misplaced accent in 'indraśatru'. I've not attempted a verse translation here and have crammed in some extra information because much is alluded to this little verse that won't be obvious to readers.
Firstly there are two main ways to make mistakes: in accent (svarata) and in articulation (varṇata). Accent refers mainly to the Vedic pitch accent. Although the use of the pitch accent was changed to a stress accent in Classical Sanskrit, the placing of the accent still provided important information about how to conjugate any particular verb or render a compound as we will see, since this is the important aspect of the story. In English great use of the change in stress was made by the Two Ronnies in their hardware shop sketch: it's not clear whether Ronnie Barker is asking Ronnie Corbett for "four candles" or "fork 'andles" (the London accent drops the h of handles which accentuates the ambiguity). The pronunciation is the same and all that distinguishes them is the stress.
The mis-articulation of a word is a more obvious mistake. Articulation refers to the way the parts of the mouth move in order to create the sounds of speech. A single mispronounced letter can completely change the meaning of a word. This is used to great comic effect in Monty Ponty's Life of Brian in the character of Pontius Pilate who has a speech impediment (though of course we ought not to laugh at other's afflictions). When his friend, "Biggus Dickus", steps in to try and calm the situation, things get out of hand. Pilot remonstrates with the crowd, pointing out that his friend Biggus Dickus "commands a quack legion. He wanks as high as anyone in Wome." The crowd is already rolling on the floor laughing and any hope of regaining control or dignity is now lost.
The story alluded to in the verse relates to Indra and his mortal enemy Vṛtra. Indra is the god of storms and rain, a counterpart of Thor in Germanic myth. Vṛtra is synonymous with 'drought', though literally the name means 'restrainer', i.e. the one who holds back the rains. The two characters of this story seem to be personifications of the annual anxiety over the arrival of the monsoon. Leading up to the monsoon rains, which last for three months, there is no rain at all for nine months. If the monsoon fails there is, even now, widespread famine in India as crops fail for lack of water. Also, as in other river valleys, the annual floods ensure the continued fertility of the soil despite heavy cropping. Since the regular arrival of the monsoon is the natural order, when it fails something must have held it back (vṛta). So each year Indra must do battle with Vṛtra (sometimes envisioned as a dragon) in order to release the rains from Vṛtra's restraint. This story has been assimilated to a much older narrative about a warrior slaying a demon which is found in mythology across Eurasia (associated with the so-called Nostratic proto-language). The story is found, for example, in the Old English story of Beowulf and in the Epic of Gilgamesh.
In one of the stories surrounding this mythic pair, Vṛtra prepares a magical rite in which he will kill Indra. Within the rite he pronounces the following mantra: indraśatrur vardhasva! He had intended to say "may Indra's killer (i.e. himself) prosper. However he mispronounces the word indraśatrur and was himself killed. Why?
The word indraśatru is a compound. And as with most compounds it can be read a number of ways. The most obvious reading, where the pitch accent falls on the final syllable of śatru, is that the compound is a tatpuruṣa compound and means 'the killer (śatru) of Indra. However, when he gets the accent wrong Vṛtra makes the compound a bahuvrīhi '[he whose] killer is Indra'.
The magic of the mantra is infallible and so Vṛtra effective ensures his own demise by effectively naming himself 'killed by Indra'. By merely uttering the words in the ritual, which is to say the sacrificial, context, Vṛtra becomes the one who is killed by Indra. The utterance becomes a word-lightning bolt (vāg-vajra) which strikes and kills the sponsor of the sacrifice! The last little cultural detail here is that the Vedic sacrifice is sponsored by a wealthy community member,  who is known as the sacrificer (yajamāna) but by Classical times carried out by a group of priests who are experts in the sacrificial ritual (yājñika). If the priests go wrong it is not they, but the sponsor upon whom the mistake rebounds. Of the four priests taking a central role in the ritual, one, the brāhmaṇa, has the role of silently following the proceedings and repairing any mistakes by chanting special mantras. Presumably it would have been incredibly bad for the priestly business to have one's mistakes killing one's benefactors and sponsors.
Buddhism was initially, and is once again in the modern times, a religion without intercessors. In between the early and most recent periods the Buddhist clergy did (and sometimes still do) act as intercessory priests, performing rituals and prayers on our behalf, but fortunately we have sidelined them to a great extent in modern Buddhism. For modern Buddhists it is our own deeds of body, speech and mind which are important. The onus is on us to practice effectively, which is a very onerous duty. It means we are back to the consequences of our actions rebounding directly onto us. Thus we are like Vṛtra - both the sponsor and performer of deeds and mistakes rebound on us directly.
Coming back to the verse, the problem is that mispronounced words do not convey the intended meaning (na tam artham āha). People who complain about "poor" English grammar, and there are lots of them, make the same point. Part of the difficulty is that spoken language and written language are quite different.

Ideas about when human's acquired spoken language vary but it's generally agreed that we had the vocal apparatus by about 200,000 years before the present, when anatomically modern humans begin to be found in the fossil record. Apart from a very few individuals with severe learning difficulties, more or less every human acquires some form of spoken language. Even the famous Helen Keller acquired language despite being both deaf and blind (and eventually became an author whose books inspired me as a child). Most of us learn our mother tongue effortlessly merely by hearing it spoken and we understand a great deal well before we ourselves can speak. No difficult grammar lessons are required. This has lead some linguists to propose a "language instinct", the idea that our brains are pre-prepared by genetics to absorb whatever language we hear spoken around us. (See Stephen Pinker's book The Language Instinct[Image]).
After about age 12 learning a new language becomes a laborious process of rote learning, though each additional language is said to come easier by those who go in for polyglotism. The change is related to changes in the brain around puberty which involve pruning brain connections to optimise for the local conditions. In others words we spend our first 12 years being a generalist and learning whatever we can and then we settle down to specialise in the most common events, actions, etc. Of course the brain remains plastic throughout our lives and learning certainly takes place, but it doesn't have the effortless quality with which we learn our mother tongue.
Writing is a quite distinct process. Left to themselves humans learn to speak, but not to write. Writing emerged only about 3000 years ago. Learning to write is laborious even for children, and most of us never excel at it either in terms of graphic form or content. Indeed those who write well are celebrated in literate cultures precisely because the skill is rare. Writing is not a skill we have evolved directly. It is one that employs a variety of general skills, mostly optimised for other tasks. Many cultures never develop writing. Amongst the thousands of languages in Australia, New Guinea and Melanesia for example, not one was written before contact with Europeans.
To be fair, most English speaking Buddhists come across their first Sanskrit words when they are already adults. And we all tend to fall back on what we know to interpret new stimuli. Presented with new words, most of us rely on habits to try to pronounce them. And even when correction is offered it is ignored. So despite repeated reminders an erstwhile housemate of mine pronounces the tri in tri-ratna as English 'try'. He knows it's wrong because I repeatedly told him so, but he prefers the incorrect-but-familiar to the correct-but-unfamiliar when it comes to language. And most of us are like this. Indeed this is a microcosm of a human problem on a macro scale. Most of us want most things to be settled and stable, with a modicum of novelty to keep it interesting. We think in well worn grooves that are not always helpful.

Very few Buddhists make the effort to learn canonical languages or even make much effort with accurate pronunciation. People I know just shrug and say they don't really care. And making them care is beyond me.
In my book Visible Mantra I've argued that one ought to be concerned with good pronunciation on various grounds (p.15f) , but my book is hardly on the 'Best Sellers' list and I'm not a person whose words have influence. In Malcolm Gladwell's taxonomy of players in change, from The Tipping Point[Image], I am a maven, but not a connector or a persuader. Though of course you, the reader, are reading this and will perhaps be influenced by it. And perhaps you are a connector or persuader?
The general disinterest in our canonical texts and languages is not a new thing. For most of the history of Buddhism most texts were known only to a few cognoscenti. The average Buddhist probably did not know any sutras, or at best might have chanted one or two as a magical charm. Most Buddhists in the past did what Buddhist do now and repeated edifying stories about their teachers and figures of the past (perhaps, but not necessarily, including the Buddha). We like  to tell and hear stories in which principles are personified. Even amongst the monastic institutions the education focusses on commentaries in the vernacular. Texts like the Heart Sūtra might be memorised in Sanskrit, but Sanskrit itself is not studied, so the text is not understood in Sanskrit and the recitation relies on the idea that Sanskrit has magical qualities. Sanskrit has no magical properties - it's just a language like any other. There is ample evidence of Buddhist texts being garbled by scribes who mechanically copied without understanding, or indeed by teachers who were unable to speak the language of the texts. My work on the Vajrasattva Mantra is a good example of the latter problem.

It is sad that the increasing popularity of Buddhism in the West has coincided with a slide towards the new form of libertarianism (sometimes called Neoliberalism, though it is extremely illiberal) and the decrease in funding to the liberal arts (including the study of religion and ancient languages). The decline of Buddhist Studies and Sanskrit and Pāli Studies in the UK has been marked since the 1970s. Pāli and Buddhism have almost entirely disappeared from Cambridge University for example. But given that most Buddhists don't care about Buddhist Studies, and that many Buddhist leaders are openly hostile to academia, it can be no surprise. The down side is that the study of our texts, history and culture is largely in the hands of those with no interest in the practice of Buddhism and even for them funding and opportunities are dwindling.
~~oOo~~
Should anyone be interested in following up the reference to the Vyākaraṇa-mahābhāṣya an English translation, accompanied by a translation of a standard Indian sub-commentary, can be found here.