Bhante and dear Sabbamitta,
after reviewing the relevant passages, I came to the conclusion that it cannot be determined with much exactitude, it is too vague. There is still the possibility that nirutti in this context means language and that this refers, with the evidence from the aṭṭhakatha, to the “language of Magadha” (māgadhabhāsā). What has been adduced against the validity of this explanation is, in the main, that the Buddha did speak against insisting on fixed expressions, but this passage is not unambiguous itself (see below).
One way of creating a congruent narrative is this: There was at the Buddha’s time a widespread, transregional koine in use (a theory now generally favored by scholars), or even several, depending on the main geographical areas, and it is possible that it (or one of them) was called Māgadhabhāsa, the language of Magadha, which the Buddha made also use of. Now the two Brahmins, still attached to things Vedic, were offended by the way the Buddhist teachings were transmitted in the peoples’ own mundane language (sakāya niruttiya) since the appropriate medium for such lore would be Vedic. They go to the Buddha and complain, but he refuses, instead declaring that this practice is unproblematic.
We also have to take into consideration the Vedic background with which infused the brahmins made this request; what did they understand by nirutti and what did they mean when they wanted to render it chandaso? I asked the renowned Sanskrit scholar Patrick Olivelle and this is his reply:
Thank you for your message and the query. With regard to “chandas”, which is your main question, the term has several related meanings. First, it is one of the Vedāṅgas, the limbs of the Veda, and in that context it refers to meter. It also had a related meaning of chant (probably metrical texts were chanted), and was connected especially with the Brahmins of the Sāma Veda. So we have the “chāndogya upaniṣad” belonging to the Chāndoga brahmins.
In grammatical literature chandas is used with reference to the language of the Veda (as Pollock has pointed out), especially in contradistinction to “bhāṣā” which was the spoken Sanskrit of the time. The distinction in grammar between the two is often pointed out. This distinction parallels the other distinction you find in grammatical works between “Veda” and “loka” — that is what is found in the Veda, including its language, and what is found in the world, the normal discourse and speech patterns.
So, the Buddhist reference [i.e. that of *sakāya niruttiyā*] clearly parallels the latter meaning, and must refer to the way Vedic texts are composed, especially the metrical part, and the language in which it is composed — which is supposed to be eternal and fixed. This also facilitated the memorization of the texts — after all there were no written texts then; the Veda was all in memory.
We can thus see that they probably intended to put his teaching into Vedic (perhaps alongside its distinctive rhythm) and that contrasts best with another language with which the Buddha’s teaching were perceived by them to be spoiled, possibly the common speech pattern of the transregional koine. That is exactly how the aṭṭhakathā and ṭīkā traditions took what the Brahmins intended, for example the Vimativinodanīṭīkā:
They ruin (dūsenti) the word of the Buddha with their own language (sakāya niruttiyā) as it relates to the canon (pāḷi): ‘Surely, those of inferior birth who learn [memorize; the buddhavacana] are ruining [it] with the language of Magadha (māgadhabhāsāya) to be spoken by all with ease (sabbesaṃ vattuṃ sukaratāya)’ – this is the meaning.
Also possible that the connotation of “grammar” for nirutti played into it as well, as understood in the context of the Vedaṅgas.
Bhante, I find that the occurrences of nirutti are quite variegated and that there are passages where it can mean language, for example:
In the context of the four analytical discriminations (paṭisambhidā) nirutti has the same implications bearing a significance of “term” but can also be understood as “grammar” – “language” as a whole not being impossible too.
The meaning of language has also found entry into Cone. The demarcation line between “way of speaking”, “nomenclature” etc. and “language” is also a very fine one.
In my opinion, I find this passage has generally been invested with too much linguistic significance. It does not rule out that there was a standard canon in, for example, the Māgadhabhāsā, or whatever it may have been, alongside profane regional speech differences. What this text, to me, is saying is that one simply shouldn’t fuss about what term to use in everyday interactions, which is also corroborated by the fact that very mundane items are given as examples, nothing necessarily to do with what the reciters were handling parallel during chanting. Just recently I had a conversation with a monk from the U.S. and I, with my background in British English, used the term “flat”, meaning “apartment”. He understood, but he perceived it as odd. I think it is about such things. So, when one is in the U.S. as a Brit, for example, one doesn’t insist on the term “flat” instead of “apartment”. 