Since at least 1610 the English for Sanskrit has been Sanskrit. The red letters above are in modern academic transcription. This represents without ambiguity the sounds as encoded in the South Asian scripts.
The South Asian scripts have always been remarkably insistent on representing sound accurately. This comes from the foundational impulse of Sanskrit itself — that the potency of the ancient Vedic hymns lay in their actual sounds (like a form of magic, or prayer as a type of software) which must be preserved exactly if it is to work, even as the ordinary spoken language(s) of the priests continued to change.
Sanskrit arose from an anxiety that an unbridgeable gulf was opening between sacred speech and everyday speech. It was a deliberate fixing. By the time that step was taken it was still a living, spoken language, but it was by then nobody’s first or only language.
The name can properly really only apply once that stage had been reached (perhaps by about 500 BCE), but the term is often used for the whole reach of the language (sometimes Vedic or Vedic Sanskrit for the very early stratum and Classical Sanskrit for the general later standard).
sam + (s)kṛta
sam- together; concentration, intensity
kṛta made, done
The general meaning of the word is ‘perfected.’ Per-fected is itself a close parallel: in the Latin word building system ‘thoroughly made, completely done’. Used as a neuter noun (saṃskṛtam) it gives the name of the language—and makes a claim about its nature.
Unlike most languages then (Japanese, Welsh, Klingon) Sanskrit is not named after a region or a people, but after itself as a standard.
Sanskrit belongs to the Indo-Aryan subfamily of Indo-European. The subfamily has three broad phases:
Old Indo-Aryan (1500 BCE on) e.g. Vedic/Classical Sanskrit
Middle Indo-Aryan (500 BCE on) e.g. Pali, Gandhari
Modern Indo-Aryan (1200 CE on) e.g. Hindi, Bengali
The dates are only rough. Setting out the stages like this gives an impression of neat succession, but it’s better to think of parallel streams: the changes from Old to Middle, quite similar to the changes from Latin into the Romance languages, take place at various times and places. Most importantly all the stages give rise to fixed civilisational languages that may endure side by side for centuries. For the pre-modern languages only literary standards are available for study.
Saṃskṛta/‘perfected, elaborated’ is the opposite of prākṛta/ ‘general, common.’ Prakrit is Sanskrit’s term for the Middle Indo-Aryan languages. Sanskrit grammar was in effect a machine for producing and certifying ‘correct’ forms. In the north of India where most of the Indo-Aryan languages are found, Sanskrit and the later vernaculars exist side by side. In the south of India (where Sanskrit arrived much later) the vernacular languages had a completely separate origin. Forms counted for more than usage or style: in phrasing and style Sanskrit shows the influence of the later stages of Indo-Aryan its speakers also lived in.
In the first millennium of the common era Sanskrit underwent a massive expansion on at least two axes. Firstly its range of interests enlarged: from being the relatively closed language of a brahmanical priesthood centred on the Vedic poems and on the various outgrowths from these (ritual, linguistics, philosophical speculation etc) it became a language of political power and a very broad high culture (plays, epic poems, love poetry etc). Secondly it exploded geographically: from its base in north India it and the culture it carried spread south into the rest of the subcontinent and beyond across south-east Asia where it arrived as part of a package of new state formation. At that point it became a cosmopolitan language like classical Arabic, Chinese or Latin, reaching across a vast swathe of the world.
The case of Buddhism is instructive. The earliest Buddhist texts are in Middle Indo-Aryan languages like Pali or Gandhari. Sanskrit at that point was like a closed cultural space that they had no wish to enter: in fact they rejected that world. But by the mid-first millennium Buddhism had taken up Sanskrit as the prestige language with the greatest reach.
By the time the haṭha-yogic texts were composed (from perhaps 1200 onwards) Sanskrit had new rivals in the early modern Indo-Aryan languages that set up their own national literary languages—just as in western Europe the vernaculars arose as rivals to Latin. By the 18th century even yoga texts were shifting away from Sanskrit.
We’re left with the names of the poses in Sanskrit. Iyengar in his 1966 Light on Yoga writes in English but follows the practice of the haṭhayogic texts (whether written in Sanskrit or a modern South Asian language) of giving the poses Sanskrit names. Many of the poses in LoY are modern formations (for instance, almost all of the standing poses), but these too are given Sanskrit names.
This has two related effects: the repertoire declares itself to be a single coherent system, which is a good and reasonable ambition; and the history of the poses is erased. There are no ancient poses or freshly-minted poses, no indigenous or borrowed poses, just poses. Effectively the whole repertoire is assumed to be ancient. It’s a kind of fabrication by omission.
This was not inevitable: in competitive mallakhamb (wrestling pole gymnastics) the trainers call out the moves to the performers using terms derived from three languages — Sanskrit for the positions from yoga, Marathi for wrestling moves and English for elements drawn from gymnastics. Although very different in scope, mallakhamb has like yoga expanded and changed in its encounter with modernity.
There is no need of course to have any interest in the history of yoga, just as there is no need for chess-players or knitters or doctors to be interested in the history of those practices. But in the recent past yoga has in effect been making false historical claims about itself. That’s no longer tenable.