No publication that I’ve seen has analysed all canonical Pali vocabulary to determine what % of it relies mainly on Vedic grammar, what % of it is explainable through classical sanskrit, and what % can’t be explained by either of them - have you seen such a study published anywhere? None of those names that you quote above appear to have done such an exercise - as far as I know.
Academics usually cite (rely-on) one another’s opinions in their publications (but rarely do they have 100% consensus-ad-idem on any topic to make such a reliance meaningful) - so one isn’t sure who was the last person who actually applied his own mind independently and comprehensively on the topic - and who is simply bouncing off their conclusions based on on other people’s opinions or misconceptions. The scholars you have named don’t agree on what exactly is/was sanskrit, what exactly is prakrit, etc. They have fuzzy definitions that they change according to need.
I’ve recently worked on thousands of Pali words to trace their Sanskrit equivalents for the DPD - and my personal experience tells me the vast majority are directly traceable to classical sanskrit etymologically. Therefore I am basing my opinions not on other people’s opinions (whether published or not), but on my own (ongoing) work. Once I trace the etymology of many more thousands of Pali words, I will be able to say much more definitively on the issue.
Prakrits were not dialects - there were no distinct ethnicities of Prakrit speakers found anywhere in Indian history to my knowledge. There is no one-to-one systematic sound-shifts between any single prakrit and any modern Indo-Aryan language that anyone has proven conclusively. So what made Pischel think they were dialects?
The two brothers were not evidently complaining about the the buddha’s own language or about any specific regional common/standard language, when they said “etarahi, bhante, bhikkhū nānānāmā nānāgottā nānājaccā nānākulā pabbajitā. Te sakāya niruttiyā buddhavacanaṁ dūsenti."
The commentarial interpretation therefore appears incorrect when it opines that ‘sakāya niruttiyā’ refers to ‘vuttappakāro’ or ‘māgadhiko vohāro’ - as that would make the original complaint (of a diversity of bhikkhus linguistically and semantically distorting the buddhavacana in their own peculiar subjective ways) meaningless.
There is no recourse to ‘probability’ required here. All vedic mantras (of all vedas) are and were chanted - and this is known to all vedic scholars as it is an extremely common occurence even today. Chanting is not specific to the Sāmaveda. Olivelle is a great scholar - he should have known what he was talking about - not sure if he was trying to suggest something different here.
Chando-ga (< Chāndogya) refers not to mere chanting the metrical hymns but actually singing them - as grāmageya gānas, āraṇya gānas, ūhā gānas and ūhya gānas, (see page 26 here). The ga at the end of the word Chandoga is from √gai (to sing) -
as the Monier-Williams dictionary points out: chando-ga m. (√gai) ‘singer in metre’, chanter of the SV. , Udgātṛ priest, AitareyaBrāhmaṇa. iii, 32 ; ŚatapathaBrāhmaṇa. x ; ŚāṅkhāyanaŚrautaSūtra. &c.
The same is said by the Sanskrit-Sanskrit dictonary Vācaspatyam: chandoga puṁ - chandaḥ sāmavedaṃ gāyati , √gai.
Again the same is known also from the Śabdakalpadruma Skt-Skt dictionary: chandogaḥ, puṃ, chando vedaviśeṣaṃ sāmetyarthaḥ gāyatīti. √gai.
We don’t need Pollock to point out what is in fact practically universally known for millenia - that chandas in grammatical texts refers to vedic and that it is used in contradistinction to spoken sanskrit. The distinction between Veda and loka also are widely known for a very long time.
Scholars cannot “agree” that chandaso refers to classical sanskrit - because it rather refers to the language and hymns of early vedic metric poetry. I dont see where they agreed about it and on what basis.
As I have said in point 8 here - the word chandas does not mean prose (or versified) classical sanskrit. Chandas here means the metric style and language of the early vedic mantra hymns. So the distinction in the passage was not between Classical Sanskrit and Pali but between the spoken language (Classical Sanskrit) and versified Vedic.
You can read more about this in my discussion with Stefan Karpik here.