Panat in postcardThe Ranums'

Panat Times

Volume 1, redone Dec. 2014


Volume 1


Orest's Pages

Patricia's Musings



Musical Rhetoric

Transcribed Sources


How was the vowel-sound 'OI' pronounced in 17th-century France?

See also my cranky comments about la Musique Francoise, as for a girl's name??;
and see also how
Mme de Montmartre  oi in 1680. See also the evidence provided by Scipion Dupleix.

A few more important clues that I just came upon in my notes,
re how OI was pronounced in the early seventeenth century:

1: Bibliothèque Mazarine, ms. 3836, "Table de Rithmes [rimes] francoises," written before 1608, p. 431:  one pronounces aparoistre as if it were written parestre. and it rhymes with estre.
2: In Le Jeune's Airs, 1608 (ed. Walker, 1:90-91) the lyrics go: "Ve-cy l'Amour" for Voicy l'Amour.

3. In François Bassompierre's Memoirs, written prior to his death in 1646 but not published until 1665, one finds the following spellings for words normally written with an oi: "Jainville" for Joinville; "Narmoustiers for Noirmoutier. Bassompierre knew that dipthongues existed however, because he wrote "Suilly" for Sully.

Both examples agree with what  Lartigaut, Foigny and d'Aisy present (below)
 as the preferred pronunciation of 1670 and 1680s.

Some background on the pronunciation issue

One June evening in 1995, during the Brossard Days organized by the Centre de Musique Baroque of Versailles, I sat among the other participants in the Brossard colloquium; and, somewhat puzzled, I listened to a skit-like performance of Brossard's airs à boire. In both the airs and the linking comments invented for the show, the performers used an archaic pronunciation and turned every OI into a strong dipthong: "ou-wé." Although the tipsy- and quite rustic-sounding result was not unsettling for drinking songs, something about the performance rang false and I made a mental note to watch for clues to Parisian pronunciation in the 1670s, 80s and 90s. For, if that was what seventeenth-century French sounded like, it makes a big difference to my coaching and, even more, to my understanding of the mechanics of French declamation and how it was transferred to music paper.

Exactly a year later, a similar entertainment was presented at a conference devoted to Mme de Sévigné. In the gallery of the chateau of Grignan, before an audience of specialists in 17th-century French literature, two musicians sang airs by the one of the Marquise's relatives and read selections from Sévigné's letters. As the performance advanced, listeners' jaws dropped, and dropped, and dropped still further. Troubled frowns appeared. And although the post-performance comments of these specialists were polite, they could not conceal their shock.

Why so shocked? Well, throughout the entire performance, every time an OI appeared in the lyrics of a song or in one of the Marquise's letters, the performers pronounced it as a very broad dipthong: "ou-wé." Thus, for Ma fille, ce soir, quand j'allois voir mon ami François, j'avois soudain soif et, sans connoître où j'étois, je me suis arrêtée pour boire, the person playing the role of Mme de Sévigné would say: "Ma fille, ce sou-wér, quand j'allou-wé vou-wér mon ami Françou-wé, j'avou- wé soudain sou-wéf et, sans connou-wétre où j'étou-wé, je me suis arrêtée pour bou-wér." According to several literary scholars with whom I talked after the performance, this omnipresent "ou-wé" certainly was not the pronunciation used in the Paris or at the court during the 1680s.

During the year that followed, this "ou-wé" pronunciation began cropping up on France-Musique, via recordings by other French ensembles. In fact, these "ou-wé's" seem to have become the signature of performers who are striving for an authenticity in French vernacular song. This striving after authenticity is to be applauded: indeed, it goes hand in hand with the recent revival of a Latin pronunciation that is closer to 17th- and 18th-century French usage than is the italianate pronunciation that Rome forced upon the French in the 1930s (see my Méthode de la Prononciation latine, which summarizes Dom Jacques Le Clerc's rules of the 1660s). But — and this is a big "but" — a reconstruction of a lost pronunciation cannot, and must not be based upon oversimplifications that are elevated to blanket rules. In the early days of the Latin movement, I remember hearing singers pronounce every U like a "pointed" French one [y]: tuum = [tyym], Jesum = [zhezym] — although sources of the period make it very clear that U was sometimes pronounced [y] and sometimes "o", depending on the way it functions in the word as a whole: [tyom] and [jhezom]. (Sorry, I can't reproduce the international phonetic alphabet here, so I'll invent approximate pronunciations.)

And so, it is imprudent to assume that, for simplicity's sake, one can make it a general performance rule to pronounce every OI as "ou-wé," just because 17th-century grammarians say that the French dipthong OI is sometimes pronounced that way. Blanket rules of the sort are only acceptable when one can demonstrate that there are few or no exceptions. In addition, one must always consider the level of society for which the work was intended. Craftsmen did not speak like dukes; and members of parlement did not speak the same at home as they would at the bar or at court. That is a documented fact! (The same caution applies to the Latin of the period: we know that nuns tended to pronounce their Latin very well, but that the opera singers who performed for the Jesuits of Paris mispronounced a lot of words — probably saying "kee" for the Latin qui, and "kuh" for the Latin que, as if the words were French.) And so, referring back to the sources for guidance, one must examine the pronunciation that a given vowel or consonant is likely to have in a specific word, and in a specific social context. My contacts with the Arts Florissants and the Centre de Musique Baroque de Versailles as these two groups struggled to apply Dom Le Clerc's rules, then branch out to words that he had not discussed, demonstrated just how touchy the task can be. My subsequent coaching of singers in the Washington D.C. area has made this even clearer, for sometimes I can only make an educated guess. (See my page containing supplementary information about historical Latin pronunciation.)

To satisfy my curiosity about this supposedly historical way of pronouncing French, I began delving into grammar books. Here are the results of a brief excursion into the thickets surrounding the historical pronunciation of French. In the end, I admit to being skeptical about reviving this lost pronunciation. To be more precise: I am convinced of the merits of returning to a French pronunciation of Latin for religious texts because it markedly alters the placement of the voice and the color of the vowels; but I am less so about a similar revolution in French song and theater. If the audience must spend most of the performance trying to figure out what is being said, it is a great loss for both audience and actor/singer. Would it not be preferable, I ask myself, to take a more modest step backwards: for example, do more with the "feminine rhythms" created by the so-called "mute" E's, to pronounce now-silent final R's and final S's lightly as the singing books suggest, and to let nasals resound a bit more? None of these modifications would affect comprehension, but they would represent a move toward a more historical style of declamation based on long versus short syllables — albeit pronounced with modern vowel and consonant sounds. And so, although my investigation of the dipthong OI was impelled primarily by intellectual curiosity, it is beginning to affect my views on "authenticity" in performance. Even more, what I learned in the libraries conflicts so much with what I heard at Versailles, at Grignan and on French radio, that this article constitutes an intellectual protest against a reconstruction that apparently stands on feet of clay. Or was I unlucky enough to consult only those grammarians who present the reverse side of the pronunciation coin? Do an equal number of respected grammarians of the 1670s, 1680s and 1690s contradict the sources I am about to quote? If so, I hope those pioneers in performance practices whose approach I am questioning will send me their evidence, so that I can summarize it in these pages and make information about both sides of the coin available to scholars and performers.

Brief though my excursion into late 17th-century French pronunciation has been, my findings suggest that the supposedly authentic performance practice that troubled the specialists in French literature at Grignan is based upon an over-simplification. These performers seem to be using a pronunciation that was flourishing during the early decades of the 17th century but that in no way reflects how Brossard or Mme de Sévigné pronounced their native tongue in the 1670s, 1680s and 1690s.

One of the oldest handbooks I consulted is Le Gayanard, Pierre, L'aprenmolire François pour apprendre les jeunes enfans et les estrangers à lire (Paris, Jean Berjon, 1609), Mazarine, 49274 (7e piece).

This little treatise leaves no doubt that, circa 1600, the dipthong OI in roi, toi, moi, and so forth was indeed pronounced "oué." In fact, on p. 30 Le Gayanard places an acute accent on the E: "moé" — but so did a lot of printers of the day, even where the E probably was pronounced like an É or an Ê, that is, was longer than É. In other words, we can't be sure that roi was pronounced as if written "roé" (or "rou-wé") or whether the E-sound was closer to "roè" (or "rou-wè"). For example, on p. 29 we find, about the letters OI: "En l'impression reformee [which he is proposing], OY [that is, OI] a son Y alteré en E dont est faicte la forsonne [a term he coined to denote a "sound"] OE, qu'on voit es mots 'moe,' 'toe,' 'roe,' au lieu de moy, toy, roy" — where a Y technically equals an I and therefore should sound like "eeee." In other words, Le Gayanard proposes that authors use the spellings 'moe,' 'toe,' 'roe,' because that is more or less the way these words sound; but the Y of moy, toy, roy implies the pronunciation "mo-eeeee," and so forth. This pronunciation clearly continued into the late 1640s and beyond, with the E gradually moving from an É to an Ê. For example, Guillaume Tronson's manuscript about the Parisian militia, dated 1648-49 (which I am currently editing), leaves no doubt that this son of a respected judicial family pronounced some of his OI's as "wè": on fol. 18v he writes roys ("kings") for rouets ("little wheels"): [rweh]. This suggests that he gave roys and rouets more or less the same pronunciation: [rweh].  An anecdote proving that this pronunciation survived for more than a century among the lower classes of the captial is cited by Rebecca Posner, Linguistic Change in France (Oxford: Clarendon, 1997), p. 258: "A striking bit of anecdotal evidence comes from the Reign of Terror (1794): accused of saying 'Il faut un roi' (it would be a capital offense to argue 'We must have a king'), a lower-class Parisian woman protested that what she had really said was 'Il faut un rouet' ('I need a spinning wheel')."

But, continues Le Gayanard, in OI, the I sometimes sounds like an EI; but here he is talking about a nasalized OI, as in the words loin, foin, join, groin, soin, "pour lesquels on met 'loein,' 'foein,' 'joein,' 'groein,' 'soein.'" In other words, a nasalized OI was pronounced more or less as it is today: that is to say, his "loein" equals ours — although there may have been a slight dipthong or even a strongly sounded N followed by the G still heard in the South of France: [lowaing]. In short, be they nasalized or not, the letters OI equalled more or less "wè." (See below, Lartigault, p. 47, and Aisy, p. 290.)

Nowhere does Le Gayanard state, as later handbooks do, that OI sometimes is not a dipthong at all (that is, when it is preceded by a "w"-sound) and instead sounds like an Ê.

Lartigault, Antoine, Les Principes infaillibles et les regles assurées de la juste prononciation de nôtre langue, (Paris: J. d'Houry, 1670) Mazarine 44622

This handbook could, and should be one of the principal sources consulted by actors and singers who wish to reconstruct the "just" pronunciation of French for the period 1670-1700. Lartigault's entire book is written phonetically! For example, his presentation of the different sounds of the letters OI (note my use of the plural here: "sounds"), makes it clear that the adjective françois was pronounced "francês" — and not the "françou-wé" that is being presented as historical. Lartigault also writes "ortographe Francêze" for orthographe françoise (in the feminine) and "mês les véritables Francês" for mais les véritables François (masculine plural), thereby demonstrating that, as an adjective or as a noun, françois/françoise was no longer pronounced like the Christian name François [franswa]. Lartigault also writes: "si l'on écrivêt n'étênt pas ... il faudrêt" for si l'on écrivoit n'étoient pas... il faudroit. In other words, the final -OIT or -OIENT of verbs (here too the A that today has replaced the O was not pronounced "wé" but "ê." In other words, by 1670 OI was pronounced, in these specific contexts, like AI is today in français.

Here is what Lartigault says on p. 47, in the section entitled De l'-OI & -OY: "Partout où l'-OI, & l'-OY se prononcent comme un Ê ouvert, il faut les changer en Ê" — that is to say, he is urging the French, when they write their language, to replace OI by Ê, when it has the sound "eh". Then he continues, writing everything phonetically: "Eczanple: [pour] conoitre, il ètoit, droite, je croyois &c.... écrivez: 'conêtre,' 'il étêt,' 'drête,' 'crêyês.'" In other words, well before 1670, OI in the endings of imperfect verbs such as étoit, or croyois was pronounced as it is today: Ê, "eh." Note, however, the pronunciation of croyois: both the OY and the OI are pronounced Ê! The same logic impelled the very elegant Mlle de Guise to spell Joinville phonetically in the 1670s: "Ginville"— that is to say, she treated the OI as an Ê, then she nasalized it.

In some contexts the OI retains the "w"-sound that the international phonetic alphabet expresses as [wa]. For example, Lartigault writes doivent and remouvoir, and on p. 34 he specifies that OI can be a dipthong, as in the words roi, droit, resoit [reçoit], je dois, François, gloire, noire, victoire, etc. — all of which implies that he and his contemporaries pronounced these words more or less as we would today, although they probably ended the dipthong on a sound that is closer to an Ê than today's A, that is, [weh] rather than [wa]. But this sound surely was not the rather separate acute E (é) we have been hearing in reconstructions of seventeenth-century French: "ou-wé."

Foigny, L'usage royale de la Langue latine, 1675

This little book on Latin grammar and rhetoric provides several insights. Foigny's remarks suggest that our modern "vwa" is more open, and that the seventeenth-century OI was closer to "ouê" (but scarcely the broad and open "ou-é" heard during the concerts at Versailles and Grignan). On p. 14 he writes: "OI, quand il est final, se prononce comme OE, par exemple il voit, il croit, pouvoir, contoir, sçavoir, etc. Exceptez les imparfaits des verbes: où il le faut prononcer comme un E ouvert. par exemple, (p. 15) je parloy [sic], il parloit, je parleroi [sic], ils parleroient, prononcez 'je parlé,' 'i parlét,' 'je parleré,' 'i parleret,' exceptez encore endroit, étroit, droit [when it means dexter, "right-hand"], qu'il faut prononcer 'endret,' 'étret,''dret'; mais droit [when it means "the law"] se prononce 'droi.'" Here is further evidence that OI often was not pronounced "ou-è" at all, but simply "ê." (That Foigny renders il and ils as "i" comes as no surprise: that is the way the pronoun was pronounced — witness the numerous instances where, in dictated government documents, the pronoun ils becomes the adverb y or qu'il becomes qui.) "Quand OI n'est pas final," he continues, "il se prononce comme E ouvert, par exemple, connoître, 'connêtre.'"

D'Aisy, Le Génie de la langue françoise, new ed., vol. 1 (Paris: Laurent d'Houry, 1687)

This little grammar provides interesting corroborating evidence, for it cites throughout three of the leading literary figures-grammarians of the day: Vaugelas, Ménage and Bouhours.

The section entitled "On the pronunciation of certain letters or syllables in verbs" — i.e., the -OI and -OIENT at the end of verbs, and -OÎ in words such as connoître, paroître (p. 288) — it becomes clear that, in good society, these vowels were pronounced AI (ê), as they are today. Citing Vaugelas, Aisy writes: "On prononce OI comme AI à l'Imparfait de l'Indicatif, & à l'Imparfait Conditionnel du Subjonctif: Par exemple, je faisois, tu faisois, il faisoit, ils faisoient, je ferois, &c.... De même aux Temps des Verbes en OÎTRE, comme je connois, tu connois, il connoît, nous connoissons, je connoissois, &c." In short, Aisy confirms what Lartigault had written 17 years earlier.

By contrast, the OI in soit (in the sense of "whether or not") is pronounced "oi" — which could mean either "wé" "wè/wê" or "wa," it's not quite clear: "on dit ‘soit' pour sive, & pour esto. comme Soit qu'il vienne ou non: Soit, je le veux bien, en prononçant OI."

In verbs with -OI- in the infinitive, every attempt was being made to change the OI to an AI (a bit like Lartigault's pronunciation of OI as "ê" in the noun avoine: 'avène'). So one finds on p. 289: "On prononce OI comme il est écrit, aux Temps des Verbes en -OIR, & en OIRE. Exemples, Voir, Prévoir, Recevoir, Devoir, Boire, &c: Je voy [sic], je prévoy [sic], je reçois, je dois, je boy [sic]. &c. Croire & ses composés se prononce par OI, ou mieux par AI." The final observation is very important: OI = AI = Ê.

Especially important is a quotation from Ménage, which reveals that there were two sorts of pronunciation, one public, the other private. At home the pronunciation tended to be AI (that is, "eh"); in public it resembled the "vwa" of envoyer, as it is pronounced today: "On prononce 'craire, 'je cray,' dans le discours familier: & 'croire,' 'je croy,' &c, dans un discours public. On prononce envoyer, verdoyer, &c. par OI." A sound approaching today's "vwa" was therefore the sound these famed grammarians gave OI — for it is not improbable that envoyer could have been pronounced "en-vou-é-yé"?

Aisy sheds considerable light on the pronunciation of proper names, pronouns and adverbs. He paints the tableau of a courtly pronunciation that is close to today's. OI is pronounced "oi," that is, the diphtong began with a "w" as in today's phonetic rendering of OI [wa], and after that came a vowel that was moving rapidly away from the "ê" of private speech to today's "a." Citing Vaugelas, Aisy writes: "Dans tous les Monosyllabes on prononce OI, comme il est écrit" — that is, as two consecutive vowel sounds. Exemples include: roy, loy, bois, moy, toy, soy, mois, quoy, moins, neanmoins, &c. And he continues: "Il y en a fort peu d'exceptez, comme froid, droit, que l'on prononce par AI. Croy, soit & soient, dont nous venons de parler." (On p. 288 he had asserted that the vowel in the subjunctive forms of the verb être should be pronounced AI: so soit was pronounced "sê"!) He quotes Ménage as pointing out that: "Dans le discours familier, on dit,'le fraid,' 'il fait grand fraid': mais on dira mieux dans un discours public, 'le froid,' 'les froideurs' — which surely was not pronounced "frou-é-deurs"! And for the word droit, he continues. "il en est de même que de froid, lorsqu'il signifie dexter ["right-hand"]: mais en la signification de jus ["law"], il se prononce toûjours par OI. Le Droit civil. Il a bon droit."

On p. 290 Vaugelas is quoted as saying that OI is "pronounced as written" in nouns that end in OIR: "OI se prononce comme il s'écrit, dans tous les Noms terminez en OIR, comme miroir, mouchoir, espoir, &c. ... De même, dans les Noms terminez en OIRE, comme gloire, histoire, memoire, &c. Il faut prononcer 'voyage' & 'Royaume,' & non pas 'véage' & 'Réaume.' Ainsi il faut dire 'avoine' avec toute la Cour, & non pas 'aveine,' avec tout Paris." Ménage takes a somewhat different position: "On dit à la Cour & à Paris,'avoine' & 'aveine,' presque indifferemment. Il semble qu[e] 'avoine' est le meilleur dans le discours familier, & 'aveine' dans les compositions relevées, ou en Vers. On prononce, & on écrit, la 'Reine' & non 'Roine.' On prononce courtois, courtoisie & roide, roidir, roideur par AI. ... Il faut dire la 'Place Royale' & non pas 'la Place Réale.'"

p. 291: "Les Noms en OYE & en OYSE, se prononcent par OI; comme joye, toise, &c. Mais on prononce 'monnaye' ... mieux que 'monnoye'.... & 'turquoise' mieux que 'tirquaise.' ...On prononce par OI, Bourgeois, exploit, détroit: mais on dit mieux 'étrait' qu[e] 'étroit' ... De même voisin par OI: 'voicy,' 'voilà,' et non pas 'vecy,' 'velà.' 'Saint Benoist,' mais on dit, 'un grand Benaist.'"

Note that these discussions of pronunciation do not revolve around whether one should say "oué" versus "oua" — but whether one can continue saying "é" or "ê," thereby avoiding the "w"-sound altogether! This suggests that there was a gradual change during the 17th century that involved adding that "w" before the dipthong OI, and that trying to sound archaic by saying "ou-wé" — or "ou-wè" or "ou-wa"— is to miss one of the principal points in the evolution of French pronunciation.

Aisy also shows that the spelling François had two pronunciations, one for the name, and another when designating Frenchmen. He cites Vaugelas, pp. 291-292: "On prononce OI, comme AI à la fin des Noms Nationaux & Provinciaux, & des Habitans des Villes, comme les François, les Anglois, les Holandois, les Polonois, les Milanois, &c. On dit pourtant les Genois par OI: les Suedois, les Liegeois." (Ménage concurred.) And Aisy continues: "On dit François par OI dans la signification de Franciscus: comme S. François, le Roy François I. De même, Françoise pour une femme. Mais il faut prononcer par AI [for the adjective "French"] François: la Langue Françoise, l'Academie Françoise. On dit un Anglois par AI, en parlant d'un homme qui est d'Angleterre: mais il faut dire L'Anglois par OI, en parlant de ceux qui s'appellent de ce nom, 'monsieur Langlois.' Les Noms de pais sont aussi partagez sur cette prononciation: on dit par OI, le Gastinois, l'Orleanois, le Vendômois, l'Artois, le Retelois, le Danois, le Parthois, Vitry en Tardenois. Mais on dit par AI, le Lyonnois, le Bourbonnois, le Boulonnois, le Nevernois, le Crannois, le Chastelleraudois."

Olivier Bettens' book-in-process, entitled Chantez-vous français?, will make available to performers an informed tableau of French pronunciation across the centuries. (He has not yet presented the pronunciation of OI.) The book is, or used to be available at Bettens' website.