the answer is, that they are long as being under the accent, so that us, -um, would be (-uus, -uum). Applying this test to English we should treat the us in English thus (C. T. v. 13384) and the -us in ignotius, which rhymes with it, as both long, and (uus). If then the Norman u was= (uu), as most of the authorities allow, though some of them speak of exceptions which they do not cite, adventure, quoted on p. 298, would have been (adventuur e) and lure, with which it rhymes, (luure), and nature (natuure). (See nature written nature in Alliterative Poems, p. 59, and salue rhyming to remwe in Sir Gawayne, p. 47). There appears indeed no proof whatever that the French (yy) was known in 13. and 14., but there are many proofs that u was consistently (uu).3 But as it is generally allowed that the English or Anglosaxon long u of those times, with which the Norman is continually found rhyming, was (uu), proofs are scarcely necessary. The greater difficulty lies in proving that the short u, or unaccented u, was not (u, u), but a different sound, approaching, if not identical with the obscure sound heard in the atonic a in a mán, e in the man, o in to-day, and represented generally in palaeotype by (e) or (ǝ, e, e, i). It is highly probable that this sound scarcely, if at all, differed from the atonic e of the French le in le livre, and that, in time, it generated the proper French eu. The development of this doctrine is essentially connected with a true conception of French, or, as far as we are concerned, the Norman system of accentuation. The Norman dialect, -and the remark applies equally to 1 That the accent lengthens the vowel on which it falls, is a phonetic theory which has been long since abandoned. See suprà p. 556, n. 1.-A.J.E. 2 But see suprà p. 424, and especially the latter part of note 3.-A.J.E. 8 That English u in 14. was (uu) and not (yy) seems inconsistent with the double orthography u, ou. See suprà pp. 298, 303, and infrà Chap. VII. 1, near the beginning. See also p. 583. n. 8.-A.J.E. 4 It seems to result from my investigations in Chap. V. that u ceased to represent (uu) in English during the period 1280 to 1310, when ou was gradually introduced as the representative of that sound. See especially p. 471, n. 2.-I don't know to what other writers Mr. Payne alludes.-A.J.E. Direct proof would be necessary to establish this remarkable difference between -seems to the actual Norman patois,have been characterised by an extremely strong and emphatic delivery of the accented syllable. The general principle of the accentuation consisted in singling out for the tonic accent the syllable which was accented in the Latin original, so that, for instance, Norman raisun from ratión-em was accented raisún, honor or honur from honór-em honur, etc., with a very forcible impact of the voice upon the last syllable. The effect of this predominant influence of the accented syllable would necessarily be, the transformation of the atonic syllables. We see evidence of this result in the not unfrequent appearance of henór, enór, and annór in the place of honór honúr. An instance, however, perhaps bearing more directly on our present purpose, is afforded by the derivatives of the old French or Norman coer or cuer (cœur). There is little doubt that this was originally pronounced (kuur). When, however, by the addition of -age, there resulted coráge, curage, and courage, all 13. forms, both the quantity and quality of the original (uu) was affected, and almost of necessity the atonic cor, cur, cour, would become (ker), and the entire word (keraadzhe). In the process of development coráge next receives the syllable -os or us, and becomes coragós, coragús curagós, or curagús, all of which are admissible Norman forms. The lately long vowel a is now changed both in quantity and quality, and has become (e, i, o) or (a, e), it is not easy to say which, and the result may be probably considered as (kǝreguus). Similarly it might be shewn that curt cour= (kuurt), becomes the old Norman system of accentuation, and that evidently adopted by Chaucer, which agrees with classical French, suprà p. 331. A.J.E. 6 Admitting that this obscuration of unaccented vowels often occurs, and has been especially active in many languages, I must deny it to be a necessity of pronunciation, any more than the prolongation of a vowel by the accent, witness the clear unaccented but extremely short a, and the decidedly short but accented o in the Italian amo (amo). See infrà p. 585, n. 4.-A.J.E. 7 Not having sufficiently studied Norman orthography and pronunciation I am unable to speak on this point.-A.J.E. 8 It seems to me extremely doubtful that such a sound as (a) was known to the Normans, when regard is had to its very late introduction into England, suprà p. 172. curtéis (kertees'), and this again curteisie (kǝrtesii e), or perhaps, at least occasionally, (kortesee').1 The last word became, as is well known, in English curtesie, cortaysie, courtaysie, all of them, by the above theory, being pronounced (kortesii e) or (kertesee'e), or very nearly, accentu mutato, as the modern courtesy, that is (kor tesi).2 The spelling could not on this theory have affected the pronunciation,3 which was determined by the power of the tonic accent obscuring and transforming the independent value of the atonic syllables. It may further be observed, that the u in the former cur, being so close to the predominant accent, became positively eclipsed by it, and would therefore be exceedingly short and obscure, as (e) in English, while the u in the second cur, receiving a secondary accent, would probably have a clear and definite sound, equal to (kər). It is this sound which the English derivatives would receive when no longer under the influence of the Norman accentuation, but subjected to the entirely different system of the English. Hence the Norman: jurnée, trublér, colúr, cumfórt, suveráin, dozáine, covért, custúme, doblér, curtíne, hurtér, cumpainée, turnoiemént, sujurnér, sucúr, etc., when they became respectively: journey, trouble, colour, cómfort, sóvereign, dózen, cóvert, cústom, double, cúrtain, hurt, company, tóurnament, sójourn, súccour, etc. would naturally be pronounced very nearly as they now are, or very recently were. In the present sound then of these I do not feel satisfied that the above account of the successive formations of cœur, courage, courageux, is historically correct. -A.J.E. 1 If this termination were ever (-ee), it was only through the West Midland confusion of i, e, and rejection of final e, certainly not from reading ie as ei, and calling that (ee). It was dialectic, not literary.A.J.E. The absolute ignorance of the sound (ə) shown by all the authorities of 16., makes me inclined to reject at once the hypothesis that courtesy could have been called (kər"tesi) in 14. With regard to the second syllable of the word, more is said in Chap. VII. 1, near the beginning.-A.J.E. 8 Although after the invention of printing, spelling may have affected pronunciation, in 12. 13. and 14. we have no reason to assume anything but the converse, words, we see the Norman influence still persisting.5 Exceptions may no doubt be taken to this general assertion, but the main principle can hardly be affected by them. It may be further remarked, that the continual interchange in early English, of u, e, i, in such instances as: werk wirk, chirche cherche churche, kirtel kertel kurtle, erth urthe, sunne sinne, sturn stern, cherl churl. segge sigge sugge=say, in báthud, etc., compared with bathed, etc., in tellus for telles, lédus and lédys for lédes, and in such plurals as femálus, sýdus, cóupus, (see Anturs of Arther passim,) tends to shew that the short u had the same sound both in Norman and English. It is impossible to conceive that the unaccented us, which merely stands in these instances for -es, was pronounced (us). It must have had the same obscure sound as the u in curteis. When, however, this obscure unemphatic sound is required to take the accent, then it assumes the clear utterance of the u in curtesie. Hence the u in churche, urthe, sunne, sugge, was not unfrequently found interchanging with e and i short. The sound then of short u seems, in words of more than one syllable, to depend on the principal accent, and when atonic to be (ǝ), and this was also the sound in monosyllables naturally short, as church, churl, etc. The merits of the general theory, which I have here attempted to expound, can, however, hardly be fairly judged of by this brief and imperfect representation of it. namely, that pronunciation affected spelling.-A.J.E. 4 But they were not so pronounced in 16., as we know by direct evidence, and they are not now so pronounced by the illiterate in our provinces. It was only the other day that I heard a porter at Clapham Junction shouting out many times in succession (Klap am Dzhuq shun). with pure (u) and not (ə), and without any obscuration of the unaccented vowels.-A.J.E. 5 The history of the introduction of (ə) being now on record, and the battle between (a, u) being still undecided, I do not see how this conclusion can be admitted. -A.J.E. 6 See suprà p. 299, and 300, n. 2, also p. 425, p. 507 and numerous instances in Chap. V. 1, No. 3. But there seems no reason for supposing this u to have been anything but (y, e, i).-A.J.E. OE, EO (IN ONE SYLLABLE), OI, IO (IN ONE SYLLABLE), UE, EU, UI, IU (IN ONE SYLLABLE), AND THE VARIANTS EOU, EOW, EOUW, EW, IEU, IW, IEW, W, UW, EACH = (uu). The illustrations and arguments by which the above proposition is supported, are given at some length in my paper. A brief summary, which understates the proof, is all that can be given here. Assuming that Norman long or tonic u(uu), it was ascertained that Norman ui, and inferred that the inverted iu, had the same sound as u alone, that is, that nuit = (nuut), fruit =(fruut), riule = (ruul'e). These conclusions depend on the light shed by Norman and English on each other. Thus in English texts frute rhymes with dedute, i.e. Norman deduit, and again frut with dedwt, whence ui = u = w = (uu). Again Norman suir, siur to follow, becomes siw in Layamon, suwe in Ancren Riwle, swe in the Alliterative Poems, and sewe in Chaucer, shewing ew, ui, iu, iw, uw=(uu), and therefore sewe of Chaucer (suure). The argument thus gained, applied to triw-e (Robert of Gloucester), trewe (Chaucer), truwe (Occleve), and treue (Audley), gives theoretic (truue), which is shewn to be correct by true in Alliterative Poems, p. 27, where due also rhymes with it, supported by Promptorium Parvulorum true. Thus, in addition to the digraph above given, ue and eu also appear to =(uu). If then the ags. treowe, which appears as treowe and treouwe in Layamon's earliest text, and as trewe in the later, had a sound different from trewe, = 1 The proof must be sought in the paper referred to, and having not seen it, I can only express my own doubts of its correctness founded upon my own small amount of observation, see p. 458.-A.J.E ? Apparently from the theory that an inversion of the order of the letters in a digraph does not affect its value, which is to me extremely doubtful.-A.J.É. 8 In nuit, fruit, the i, still pronounced, is as much a representative of the lost guttural, as the y in day, may.-A.J.E. 4 Which I doubt.-A.J.E. 5 An examination of the age and locality of MSS. is necessary before judging of the value of their orthography in determining sounds. The Alliterative Poems, Sir Gawayne, and Anturs of Arther are West Midland, in which part of the country a very peculiar pronunciation still prevails, so different from the South Eastern, that the ancient orthography of that district requires especial study. It is very probable that (uu) was unknown in those districts as a sound of u, w, but that it was always replaced by (yy, y) or some cognate sound. triwe, or trwe, it could only have been for a short time, and it may probably be assumed to have been the same.7 The supposition, then, that ew had one sound in words of Norman origin, and another in those of native growth (p. 302) is unnecessary, and indeed inconsistent with the fact that, though it may be true that Chaucer does not rhyme together words in ew of different origin, other writers do. As a case in point we find in Alliterative Poems, p. 13, trwe English, blwe probably Norman, grewe preterit. remwe Norman, and again knewe English, (which is also found written knwe) swe Norman due Norman, hwe English, untrue English and remwe Norman, all rhyming together. We note also in this text Chaucer's newe always spelled nw or We should, therefore, perhaps read such rhymes as those found in Lyrical Poetry, p. 37, viz: reowe, newe, heowe, kneowe; as (ruu e, nuu'e, Huu'e, knuu e). Many confirmatory instances might be cited from various texts, but the above may suffice to shew the great probability that Norman and English ue eu, ui iu, eou, etc. were in 13. and 14. (uu), and hence that the modern pronunciations of: rue, true, sue, suit, rule, pursuit, bruit, fruit, and the vulgar sound of nuisance (nuu), duty (duu), new (nuu), beautiful (buu), are but echoes of that of 13. and 14.9 nwe. 8 = On Layamon see p. 496, and on the Ancren Riwle, see p. 506. The orthography of these works offers so many points of difficulty that it cannot be safely appealed to for any proofs. The whole of our Western provincial pronunciation has first to be studied.-A.J.E. 6 In the last note it was conjectured that the w of the Alliterative Poems may have been (yy). As regards the Promptorium the author only knew the East Anglian pronunciation (suprà p. 23, note 2), and to this day the East Anglians use (yy) for (uu). The above inference is therefore in the highest degree hazardous.-A.J.E. 7 On treowe see p. 498, 1. 14. No Anglosaxon scholar would be likely to admit eo to have had the same value as u. See p. 511.-A.J.E. 8 Probably all these rhymed as (yy), as they still would in Devonshire. See suprà n. 5.-A.J.E. 9 This conclusion is directly opposed to all I have been able to learn on the subject. -A.J.E. OI, IO (IN ONE SYLLABLE), OE, EO It is remarkable that two sounds so remotely allied as (uu) and (ee) should frequently, both in Norman and English, be used one for the other. Nothing, however, is more probable than that or in early French generally, must have represented the sound (uu). Nothing at the same time is clearer than that in the Norman texts the oi of Central France is very generally to be read (ee). Thus the forms moi, toi, etc., which in proper Norman would be mei, lei, etc., are by no means excluded from Norman texts, but are constantly found rhyming with the Norman ei or ee. Thus tei rhymes with moi, moi with foi, voir with veer, roi with lei, etc., and are therefore to be pronounced (mee, veer, lee), etc. The concurrence, however, of such forms as genoil genou, genoul, genue; acoiller, acuiller, where ui = (uu); agoille aguille; angoisse, anguisse, angusse; noit, nuit; poi, pou peu; fusoyn (rhyming with corbiloun in De Biblesworth, Wright p. 158), seems to shew that oi, ui= (uu). This conjecture may be further confirmed by assuming oi =oe, and observing that oile oil of 12. becomes oele and uille in 13., and huile in 15., while buef, boef are bouf= (buuf) in De Biblesworth. This word he rhymes with ouf œuf, of which the variants were oef, uef. Again boe, moe, roe of 13., become later boue, moue, roue. But eo also (uu), as is seen in the numerous words of the form empereor, etc., which became emperour, etc. The most difficult case is that of io=oi=(uu). It is proved, however, by the formation of such words as mansion, which became by the loss of then and fusion of io into u, maisun. Raisun may be explained in the same way, as may also maçun mason, from low Latin macio. The word in its Normanised form machun occurs in Layamon, and is erroneously translated machine by Sir F. Madden. These views respecting Norman oi io, oe eo= (uu), are singularly confirmed by English examples of adopted Norman words. Mr. Ellis's inferences (p. 269) I should generally endorse, except that, as before = 1 In this further investigation respect would have to be paid to the principle of palatalization produced by an inserted i, familiar to those who have studied phonetic laws, and well illustrated by Prof. Halde = stated, I should pronounce boiste, for which buiste is also found (buust e) not (buist e), and perhaps Loi, coy, and boy (Luu, kuu, buu). Merour mirror of Chaucer, is directly taken from Norman mireor. It occurs as myroure in Political Songs, Wright, p. 213. Norman poeste also appears constantly in English as pouste. The case of io (uu) is not considered by Mr. Ellis. It is, however, rendered more than probable by our word warrior written werroure by Capgrave, and referable to Norman guerreur, which by analogy =guerrour. Analysing the ou = (uu) into oiio, we obtain the modern English warrior. Similarly we may trace carrion to Norman caroine. So the word riot, conjecturally referred by the editor of Ancren Riwle to route, may be really a variant of that word. It must be remembered, however, that the English riot came directly from Norman riote, and the variation, if variation it be, must have belonged to the original source. Diez, Ménage, Scheler and Burguy virtually give up the etymology altogether. It is only probable then, but not proved, that Norman caroine and English carrion, might have been (karuune), and that riot might have been sometimes (ruut). The subject requires further investigation. The fluctuations of Norman orthography suggested the enquiry that has been sketched, but the results lead us on still further, and render it probable that eo, oe, etc., when found in pure English words, had also the sound (uu). Heo she, therefore, with the variants hu and hue, was probably (Huu), as it still is in Lancashire. Heore their, too, and huere, interpret each other, and so do, duere and deor, beath and bueth, beon and buen, preost and pruest, glew and gleo. We infer, then, that in Layamon's beorn warrior, cheose, leode, leof, leose the eo= (uu). The subsequent forms burn (Piers Plowman), choose, luve, loose, etc., and the contemporary form lued for leod, (Pol. Songs, p. 155), render this hypothesis very strong, while such forms as goed good, compared with goud (Layamon,) man, in his article on Glottosis Analytic Orthography, pp. 67-71. So far as I can understand them, I entirely dissent from the views expressed in the text.-A.J.E. = toen town, proeve Norman preove English, doel and deol sorrow, shew that oe as well as eo (uu). The great difficulty in assigning the phonetic values of oi, eo, oe arises from the undoubted fact that they were represented both by (uu) and by (ee).1 Thus we find that nearly all the Norman and English words cited above appear to have both sounds. Thus heo appears as he, heore huere as here, deor duere as dere, beoth bueth as beth, beon buen as ben, preost pruest as prest, cheose as chese, leose as lese, etc.,2 also proeve preove as preve, caroine as careyne carayne, puple, pueple, people as peple, etc. This divarication in the case of Norman words, was more apparent than real, since the usual Norman sound of oi was (ee). Yet the numerous examples of oi also = § 2. The Expression of the Sounds. The list in the last section suggests its counterpart, how have the sounds of the English language been expressed by letters at different times? Up till the invention of printing at least, the object of writers seems to have been to represent their pronunciation, and the possibility of using the same symbols with altered values does not appear to have occurred to them, although each sound was not uniformly represented by the same sign, and some signs had more than one value." It is also not at all improbable that very provincial writers may have been accustomed to attach values to the letters corresponding to their local pronunciations, and have then used them consistently according to their lights. From these causes arose the occasional picturesqueness of scribal orthography, which was unchecked by any acknowledged 1 My own indicated explanation of the phenomena to which Mr. Payne refers are to be found on p. 269, and 131, note, col. 1, p. 138, note col. 1. The question seems to be one affecting the treatment of Latin e, o, in the Romance languages.-A.J-E. 2 These anomalies, occurring in MSS. not expressly named, seem readily explicable by the known interchanges of eo, e, p. 488, and of u, e, suprà p. 585, n. 6.-A.J.E. 8 Oe, eo are so rare in Chaucer, see p. 262, 1. 33, that I have not been able to judge of their origin or intentional use as distinct from (ee). But we must not forget the two modern forms reprove, reprieve.-A.J.E. 4 The Norman was an old Norse phonetic system modifying the langue d'oil, so that the latter had the main share in the result. The English was a pure Anglosaxon system, slightly modified by an old Norse element. There seems to be no connection between the two systems of sound. The orthographies were both derived from the Latin, but the Norman spelling came direct from Roman sources, and the Anglosaxon was only a priestly transcription of the preexistent runic. The whole application of the orthographies was therefore diverse. The Norman accidentally came into collision with the English, but the developments seem to have proceeded independently, and the share of Norman in 13. English was scarcely more than that of English in 13. Norman. Ultimately the whole character of our language, both in idiom and sound, became English, and Norman words were ruthlessly anglicised. Hence, I am not inclined to admit Mr. Payne's conclusion.-A.J.E. 5 See the table on p. 407, where in col. 2, " (ou) o oo oa" is a misprint for " (oo) o oo oa." |