Sometimes an analysis of ordinary language -- such as Grice propounded -- is based not on ordinary language but LATIN -- although it surely should be allowed Latin to have been ordinary to the Romans!
Grice was an Englishman: if you see an Englishman using Latin (cfr. Shakespeare, and his 'better Latin' than 'Greek') you have to suspect that there may be a conceptual gap in the Englishman's lexicon.
Is this the case with 'onus probandi' -- and its conceptual analysis?
The fist record of the phrase "onus probandi" is 1722, when in an act for the encouragement of silk manufacture, published in the London Gazette, one reads:
i. The onus probandi shall lie on the exporter, claimer, or owner thereof.
Today's Gazette uses a less heavy lexicon -- and since, the sellings have increased!
The second use is dated 1793, when J. Smeaton, in his narration of Edystone Lighthouse notes that
ii. The onus probandi should be upon me.
This is deictic, and has to do with that infamous lighthouse.
In 1803, in the Philosophical Transactions of the Royal Society, we read:
iii. The onus probandi ought in justice to fall to the share of those who would deny the truth of what we may call a fact.
This more or less entails that the onus probandi should fall on Witters who was always, especially in "Philosophical Investigations" rejecting his earlier views where he would 'hardly' deny the truth of what he there he (and his followers, the Wittgensteinians) call a 'fact'.
In 1842, W. G. Simms in "Beauchampe" notes that
iv. The onus probandi lies with the state.
-- a concept analysed in 'political philosophy'.
In 1885, in the Law Rep: Chancery Div., one reads (if you are into that sort of thing):
v. The onus probandi that the lease was improperly drawn would lie upon him.
In the twentieth century, the old Roman phrase was still in use. Thus, in 1936, one reads in "The Philosophical Review" a section that goes:
vi. The fixing of the onus probandi on those who 'do not wish to believe' in the dogmas.
Why such a fixing is required may be a cause of Grice's 'defense of a dogma', a few years later ("Trust him to rally to the defense of the underdogma," Strawson said -- "and the sad thing is, he rallied ME with him -- their joint "defence" of the 'dogma" that there is an analytic-synthetic distinction was meant to refudiate Quine who had used 'dogma' freely and not Catholically).
In 1964 -- the swinging sixties -- "onus probandi" is still being used (which would delight Cicero). Thus we read in the American Mathematical Monthly:
vii. The onus probandi is not in the scientist who refuses [NOT 'refudiates' -- Speranza] to call a reputed hypothesis false to show that his hypothesis can be saved by some suitable A' ".
-- which is a case of LEGALESE extending to what we may dub the analytic philosophy of science -- as when Popper generally criticised Kuhn for not reconstructing what he sees as the 'scientist''s CRITICAL attitude.
Note that the source is the Mathematical Monthly, and we are never sure if Popper followed Lakatos in thinking of mathematics as a 'science' (Lakatos did his best by titling his essay, "Proofs and refutations: the logic of mathematical discovery" combining two essays by Popper, but Popper was unconvinced).
But LEGALESE is alive and kicking by 1987 and beyond. In 1987, in the American Journal of International [granted] Law, one reads:
viii. The court has long operated with a careful respect for the onus probandi of the Roman and civil law systems.
-- which is something that would endear the American Journal to at least the good ole Romans ("of yesteryear," Geary adds).