Prior to Buddhist hybrid Sanskrit teachings used to be generally recorded in the Pali language.
[9] However, earlier works, mostly from the Mahāsāṃghika school, use a form of "mixed Sanskrit" in which the original Prakrit has been incompletely Sanskritised, with the phonetic forms being changed to the Sanskrit versions, but the grammar of Prakrit being retained.
Edgerton writes that a reader of a Buddhist Hybrid Sanskrit text "will rarely encounter forms or expressions which are definitely ungrammatical, or at least more ungrammatical than, say, the Sanskrit of the epics, which also violates the strict rules of Pāṇini.
"[11] Edgerton holds that nearly all Buddhist works in Sanskrit, at least until a late period, belong to a continuous and broadly unitary linguistic tradition.
The language of these works is separate from the tradition of Brahmanical Sanskrit, and goes back ultimately to a semi-Sanskritized form of the protocanonical Prakrit.
Pāli shares a large proportion of these words; in Edgerton's view, this seems to prove that most of them belong to the special vocabulary of the protocanonical Buddhist Prakrit.