@diachronically

According to the Western tradition of analyzing language, languages always evolve from more ancient languages according to definite rules. As for instance, the modern Spanish words

estoy

estas

esta

estamos

estais

estan

have evolved from the Latin words

sto

stas

stat

stamus

statis

stant

according to definite rules of evolution like 'wordfinal t is lost', 'word-initial "st" turns into "est"', "unaccented u turns into o". So when explaining how Spanish verbs work nowadays, they will often mention how Latin verbs used to work.

Indian grammarians however seem to have no interest on how Sanskrit historically evolved. According to the old mythological views, Sanskrit as we know it (the /laukika) is a slightly corrupt version of the language of the vedas, which is the source of all other languages.

So if the Latin word for "is" is "est", the Greek word is "esti", the Sanskrit word is "asti" and the German word is "ist", Western grammarians infer logically that all these forms they must come from some common source, and they call it PIE. Indian grammarians however take that as proof that all other languages evolved from Vedic Sanskrit.

The rules of /pANini do not try to reconstruct a supposed history of the language. As long as the rules manage to build the accepted /laukika and vedic forms, they are good rules.