(a) ai before a nasal was in Anglo-Norman writing very commonly represented by ei. This is merely a question of spelling apparently, the sound designated being the same in either case. Our author (or his scribe) had a certain preference for uniformity of appearance in each set of rhymes. Thus he gives us first solein, plein, soverein, certein, mein, Evein, in Mir. 73 ff., then vain, grain, main, gain, pain, vilain, 2199 ff.; or again haltaines, paines, acompaines, compaines, restraines, certaines, 603 ff., but peine, constreine, vileine, peine (verb), aleine, procheine, 2029 ff. Sometimes however the two forms of spelling are intermixed, as vein, pain, main, &c., 16467 ff., or meine, humeine, capitaine, 759 ff. Some of the words in the ai series, as pain, gain, compaine, are spelt with ai only, but there are rhyme-sequences in -ain without any of these words included, as 6591 ff., main, prochain, vilain, certain, vain, sain; also words with original French ei, such as peine, constreine, restreines, enseigne, plein (plenus), veine (vena), meinz (minus), atteins, feinte, exteinte, enter into the same class. Thus we must conclude that before a nasal these two diphthongs were completely confused. It must be noted that the liquid sound of the nasal in such words as enseigne, plaigne, had been completely lost, but the letter g with which it was associated in French continued to be very generally written, and by the influence of these words g was often introduced without justification into others. Thus we have the rhymes ordeigne, meine, semeigne (= semaine), desdeigne, peine, 2318 ff.; peigne (= peine), compleigne, pleine, meine, halteigne, atteigne, in Bal. iii; while in gaign, bargaign, rhyming with grain, prochain, &c., g is omitted at pleasure. Evidently in the Anglo-Norman of this period it had no phonetic value.
(b) When not before a nasal, ai and ei do not interchange freely in this manner. Before l, ll, it is true, ei has a tendency to become ai, as in conseil consail (also consal), consei(l)ler consail(l)er, merveille mervaille; also we have contrefeite, souffreite, 6305 ff., eie for aie (avoir), eir for air 13867, gleyve 14072, meistre 24714, eide (eyde) for aide in the rubric headings, paleis (palois) for palais, and vois (representing veis) sometimes for vais (vado); also in ante-tonic syllables, cheitif, eiant, eysil, leiter, meisoun, meistrie, oreisoun, peisible, pleisir, seisine, veneisoun, beside chaitif, allaiter, maisoun, maistrie, paisible, plaisir, saisine. This change is much less frequent, especially in tonic syllables, than in some earlier texts, e.g. the Vie de S. Grégoire.
The Anglo-Norman reduction of the diphthong ai and sometimes ei to e, especially before r and s, still subsists in certain words, though the Continental French spelling is found by its side. Thus we have fere, affere, forsfere, mesfere, plere, trere, attrere, retrere, tere, debonere, contrere, rhyming with terre, guerre, quer(r)e, &c.; also mestre, nestre, pestre, rhyming with estre, prestre; and pes, fes (fascem), fetz, mes, jammes, reles(s), in rhyme with ades, pres, apres, deces(s), Moÿses, dess, mess, confess. (This series of rhymes, which has ę, is of course kept distinct from that which includes the terminations -és (-ez) in participles, &c., and such words as ées, dées, lées, prées, asses, malfés, &c., which all have ẹ.) We find also ese (with the alternative forms aese, ease, as well as aise), frel, ele, megre, plee (plai, plait), trete, vinegre, and in ante-tonic syllables appeser, enchesoun, esance, feture, lesser, mesoun, mestrie, phesant, pleder, plesance, plesir, sesoun, tresoun, treter. In the case of many of these words the form with ai is also used by our author, but the two modes of spelling are kept apart in rhymes (except l. 18349 ff., where we have tere, terre, aquerre, faire, mesfaire), so that affere, attrere, rhyme with terre, but affaire, attraire, with haire, esclaire, adversaire, and, while jammes is linked with apres, ades, pes, we find jammais written when the rhyme is with essais, lais, paix. This may be only due to the desire for uniformity in spelling, but there is some reason to think that it indicates in these words an alternative pronunciation.
It is to be observed that on the neutral ground of e some words with original ei meet those of which we have been speaking, in which ai was reduced to e in rather early Anglo-Norman times. Thus we have crere rhyming with terre, affere, &c.; crestre, acrestre, descrestre, with estre, nestre; and encres, descres, malves, with apres, pes. These forms, which have descended to our author from his predecessors, are used by him side by side with the (later) French forms croire, croistre, acroistre, descroistre, encrois, descrois, and these alternative forms must undoubtedly be separated from the others in sound as well as in spelling. This being so, it is not unreasonable to suppose that the case was the same with the ai words, and that in adopting the Continental French forms side by side with the others the writer was bringing in also the French diphthong sound, retaining however the traditional Anglo-Norman pronunciation in both these classes of words where it happened to be more convenient or to suit his taste better.
(c) The French terminations -aire and -oire, from Lat. -arius, -oria, -orius, are employed by Gower both in his French and English works in their Continental forms, the older Anglo-Norman -arie, -orie, which passed into English, being hardly found in his writings. The following are some of the words in question, most of which occur in the Confessio Amantis in the same form: adversaire, contraire (contrere), doaire, essamplaire, lettuaire, necessaire, saintuaire; consistoire, Gregoire, histoire, memoire, purgatoire, victoire. We have however exceptionally rectorie 16136, accented to rhyme with simonye, and also (from Lat. -erium) misterie (by the side of misteire) accented on the ante-penultimate.
iii. French ei not before a nasal.
This diphthong, which appears usually as ei in the Anglo-Norman texts of the thirteenth century, is here regularly represented by oi and levelled, as in the French of the Continent, with original French oi. In its relations to e and ai it has already been spoken of; at present we merely note that the later French form is adopted by our author with some few exceptions both in stems and flexion. Isolated exceptions are deis (debes) for dois, heir by the side of hoir, lampreie, malveis (also malvois, malves), teille, and vei (vide) from veoir; also in verbs of the -ceivre class and in derivatives from them it is often retained, as resceivre (but reçoit, resçoivre), receipte, conceipt (also conçoit), conceive, deceite, &c. Under the influence of rhyme we have in 6301 ff. espleite, estreite, coveite, rhyming with deceite, contrefeite, souffreite, and 10117 ff. pareies (parietes), veies, preies, moneies rhyming with pareies and journeies (for parées, journées); but elsewhere the forms are exploite, estroite, covoite, voie, proie, monoie, and, in general, Anglo-Norman forms such as mei, rei, fei, treis, Engleis, have disappeared before the French moi, roi, foy, trois, &c.
The terminations of infinitives in -eir have become -oir, except where the form has been reduced to that of the first conjugation; and those of imperfects and conditionals (imperfects reduced all to one form) have regularly oi instead of ei. There is no intermixture of ei and oi inflexions, such as we find in Angier, in the Vie de S. Auban, and in Bozon. In a few isolated instances we have ai for this oi of inflexion, as poait in Mir. 795, solait 10605 &c. (which last seems to be sometimes present rather than imperf.), and volait 13763. Also occasionally in other cases, as curtais, 5568, in rhyme with mais, mesfais, &c., elsewhere curtois, array, 18964, rhyming with nay, essay, usually arroy, and desplaie, manaie, Bal. xxvii. 2, elsewhere desploie, manoie. There is however nothing like that wholesale use of ai for ei (oi) which is especially characteristic of Langtoft, who besides the inflexion in -ait has (for example) may, cray, ray, for moi, croy, roi.
In ante-tonic syllables we may note the ei of beneiçoun, freidure, leisir (usually loisir), Malveisie, peitrine (also poitrine), veisin (beside voisin), veisdye, &c., and ai in arraier, braier.
iv. The diphthong oe (ue) is written in a good many words, but it may be doubted whether it had really the pronunciation of a diphthong. The following list contains most of the words in which it is found in the tonic syllable: avoec, boef, coecs (coquus), coer, controeve, demoert, doel, joefne, moeble, moel, moet moeve (from movoir), moers moert moerge (from morir), noeces, noef, noet, oef, oel, oeps, oevre, poeple, poes poet, proesme, soe, soeffre, soen, troeffe, troeve, voegle, voes (also voels), voet (also voelt). In the case of many of these there are variations of form to o, u, ue, or ui; thus we have cuer (the usual form in the Mirour), controve, jofne, noces, owes (dissyll. as plur. of oef, also oefs, oes), ovre, pueple, pus (also puiss), puet (also poot), prosme, sue, truffe, trove, volt, and (before an original guttural) nuit, oill (oculum). Two of these words, cuer and oel, occur in rhyme, and they both rhyme with ẹ: mortiel, oel, fraternel, viel, 3733 ff., and cuer, curer, primer, 13129 ff., by which it would appear that in them at least the diphthong sound had been lost: cp. suef in rhyme with chief, relief, Bal. L. 2. The same rhyming of cuer (quer) occurs in the Vie de S. Auban, in Langtoft and in Bozon (see M. Meyer’s introduction to Bozon’s Contes Moralizés). With avoec we also find aveoc and avec, veot occurs once for voet, and illeoc, illeoque(s), are the forms used from Lat. illuc.