This earlier was a "factlet" dated April 4, 2008
For several years now, I have been protesting the pronunciation of the dipthongue oi that has been adopted by a number of Baroque ensembles. (See my Musings: How was the vowel-sound 'OI' pronounced in 17th-century France?; and "La Musique Françoise", plus my newest bit of evidence: Scipion Dupleix and the pronunciation of "OI," 1651)
I return to the attack, because in January 2008 I copied out a collection of letters written in the fall of 1680 by two Highnesses (see A trip through Champagne, 1680.) One Highness was the Abbess of Montmartre, the other was her sister, the Duchess of Guise. Throughout the Abbess's letters I caught her writing -ais -ait, and so forth, instead of the more common -ois, -oit. In other words, she pronounced avoit as if it were written avait ― using the very same vowel she used when saying the name of the Guise estate she was visiting: Marchais. In my transcription of this correspondence, I have highlighted in in bold type all the instances where she either uses an e (which, of course, has the same sound as -ais, or -ait), or else writes ai instead of the more common oi. She, of course, uses ai elsewhere to convey the phonetic sound "eh." For example, just as we do today, she writes "souhait, plaît, fais, and plaisir." In other instances, instead of ai or oi, she uses the letter e. In short, in her spelling of the sound "eh," the Abbess was well over a century in advance: not until the Revolution did -oi yield to -ai in published works and in most correspondence.
Some of these very revealing spellings are: serais, when most people in her day wrote serois; avet for avois; estet for estoit; auret for auroit; and j'yrais for j'irois. Especially interesting is fesblesse in lieu of foiblesse and crayés, for croyez. The latter phonetic rendering parallels the preferred pronunciation of avoine as avène.
There are fewer instances of the phonetic rendering "ai" in the letters of the Duchess of Guise, for whom spelling was always a challenge. But she too offers us a few instructive insights into the pronunciation that Lully ― and all his contemporaries who hoped to have a work performed at court or before a high noble of her standing ― would have envisaged for their vocal music. For example, instead of reconnoissance she wrote reconnessance: that is, she pronounced reconnaissance as we do today.
As for the third correspondent, M. Du Bois, a future member of the French Academy, although he consistently wrote -oi, it would be plain obstinacy for us to imagine that he would have embarrassed Their Highnesses, and himself, by pronouncing that vowel "oueh" or "ouah."
PS: Orest has been working on the death inventory of Cardinal Richelieu, It's clear that notaries in the early 1640s also pronounced many oi's as "E": for example, this particular notary, referring to the weight of an item, spelled pesant as "poisant"!! I would infer from this that he pronounced poids as "pouwa," or "pouwhe", more or less as we do today; but that he knew when to treat the oi as an E: poisant/pesant. In short, he wrote two very different sounds in the very same way. As one of the Musings at the top of this factoid reveals, the notary's contemporaries spelled the woman's name Françoise but pronounced it "Fransouwaze," as we do today; but when talking about a Frenchwoman, a "Française" (which almost everyone in those days spelled "Françoise," just like the christening name) they said "Franceze," as they do today.