19

derivation.

This confirmed visscherij, bakkerij, weverij, posterij, olieslagerij, by many Dutch words of French spotternij, jokkernij, schilderij, tooverij, hoovardij, gasterij, voogdij, &c.; and others there are which are simply French words which seem as if they had been embalmed in Dutch with their antique sound expressly to corroborate Palsgrave's statement which might otherwise seem incredible to us moderns. Such are Marij, poezij, copij, harpij, galerij, tirannij, besides others which have a consonant after the vowel, as Latijn, Martijn (like the Austin and the Gamelin of, or attributed to, Chaucer), and patrijs = perdrix, prijs = prix, paradijs. I do not lay equal stress on all of these words because of the obvious possibility (I do not admit more) that some of them may have come directly from the Latin.

But the majority of grammarians seem to Mr. Ellis to confirm his opinion that the symbol i stood for (ii) at least as late as the early part of the 17th century. A few words only on this point.

'I' words.

Salesbury's Next to Palsgrave comes Salesbury, who, statements as to writing for Welsh readers, represents I, vine, wine by ei, vein, wein; and Mr. Ellis himself admits that "in modern Welsh the sound of ei seems to me as (əi)," nor is there a shadow of proof that the Welsh orthography has altered as to the value of ei since Salesbury's time. Yet Mr. Ellis immediately after the above admission proceeds with curious inconsistency (p. 111): "I think however that his letters ei justify me in considering, or rather leave me no option but to consider, that the English diphthong sounded (ei)* to Salesbury):" words which might with exactly equal force of reasoning be applied to Adelung or Grimm's pronunciation of the modern German mein and wein.

20 The true theory

as to these (əi)

words.

There is surely room for another theory, based not on symbols but on spoken words, as follows: almost universal tradition fixes the words (mǝin wǝin) for many long centuries in the Germanic races; and when the Roman alphabet came to be employed to repre* Which Mr. Ellis explains as "Scotch time, Portuguese ei."

sent sounds not recognized (I do not say unknown) in the Latin language, at least as then spoken, it came somehow to be customary in this island to represent this sound by i or í, and among the continental High Germans by i, ii, or î, or at a later time, the Moeso-Goths setting the example, by the digraph ei: (see below § 102.) It had to be represented one way or another, and these were the ways adopted. The digraph which the Germans chose stood for another sound (ee) in Old French and in the English of Chaucer and his contemporaries, as I shall show below.

(But the accents, it will be said, merely indicated the long vowel. Such is Dr. Bosworth's view: see his Orosius, Pref. p. lxiii. The Teutons borrowed their letters from the Romans, and therefore a was (a), á (aa), i was (i), í (ii), and so on. Plausible as this view is, I cannot accept it. The Romans needed no marks of quantity, and made no distinction in writing between incidit and incidit, refert and refert, confugit and confugit, and so on; nor therefore did the Teutons learn from them any mode of marking mere quantity. Nor have the moderns found any necessity for so marking the length: is there any book other, or later, than the Ormulum in which indications of quantity are given? And where in Icelandic an accent—or mark, as Rask calls it—is put over a vowel, it in most cases, if not all, indicates a considerable modification of the sound, According to Mr. Ellis, whose accuracy of ear may well be trusted, a = (a), á = (aau), e = (e), è = (ìee), i = (i) or (ii), è = (ii), o = (00), ó - (0ou), u = (3), ú = (uu). In no instance does the accent indicate, according to modern Icelandic pronunciation, a simple prolongation of the vowel.)

written ei in Eng

German.

21 The sound of At a later time the great learning and volu(əi) sometimes minous writings of Erasmus, Lipsius, and others, lish as well as in when they had adopted the ei, caused their mode of representing the sound to become familiar to English readers also. Hence we find Hart writing reid bei for ride by, and Gil writing ei for oculus, which Smith tells us was sounded like I = ego, and I or aye = etiam (Ellis, p. 112). But the modern pronunciation

C

of ei as (ǝi) in certain words-either, neither, and one or is probably due to court influence after the accession of the Hanoverian dynasty.

two more

the Latin I de

sius.

22 The English What Smith wrote about ei, as quoted by pronunciation of Mr. Ellis on p. 121, concerned the English ei fended by Lip- (ee), which was not by any means what Justus Lipsius (1586) intended-Lipsius was a Dutchman, it will be remembered-when he wrote: "Pronunciant etiam nunc (ita accepi) recte soli pæne omnium Europæorum Britanni: quorum est Regeina, Ameicus, Veita. Recte dico, quia non aliud insonuit hæc longa quam EI diphthongum." De recta Pron. Lat. Ling., p. 23. So we had in Lipsius's time-and rightly he affirms-a different pronunciation of regina, &c. from almost* all the other nations of Europe; and Gil emphatically declares: "retinebimus antiquum illum et masculinum sonum, atque unà etiam laudem quam Justissimus Lipsus [sic] nobis detulit in Reginâ, in amicâ vitâ, &c." All of this becomes instantly intelligible and lucid on the simple supposition that both the Dutchman and the Englishman spoke of the same sound (əi) that tradition has handed down to us.

23

I have not found in Lipsius's writings any statement of the reasons on which his opinion is based, but they were probably such as these: Ist, that the traditional sound in certain localities was (əi), (see quotation from Sir Thomas Smith in footnote); 2nd, that Greek words with « generally have the simple i in Latin; and 3rd,—for which however in many cases itacism will sufficiently account-that Latin words in i are not infrequently found in Greek with «, as Πεῖσαι, Ὠστεία, Λείγηρ, Σειρῖτις.

A ray of light But all southern Europe, it may be said, is against Lipsius. It unanimously affirms that

from ancient

Greece.

* Could the Lombards have been an exception? Sir Thomas Smith writes: "Quis Anglus Gallum Latinè loquentem, nisi assuetus intelliget? certe ego non potui: at Italum statim, quia nos ab Italis cùm Latinè sonamus, nisi in valde paucis, a Longobardis autem Italiæ propemodum in nulla re dissidemus : at à Gallis infinitum quantum dissentimus, quamvis nostri sint vicini." De Ling. Gr. Pron. (1568), p. 14. I must leave this nut for some student of early Italian pronunciation to crack.

the juice of the grape-to take one typical example—was not called (wǝin) but (wiin) in the ancient Classical Languages. No doubt it is easy to assume that the Italians, Spaniards, Portuguese, &c., have preserved the true Latin sound of this word; but what of Greek? Some scholars believe that in the o of olvos the o is merely a variant of the digamma, and that Fios is the old form and points to (wiin). But ancient inscriptions show us the F and the o both used in such words. In Boeckh's Corp. Inscr. Gr., No. 4, we have TAN FOIKIAN: which, being confirmed also by other inscriptions, conclusively shows that in that word at least-very probably therefore in others like it-the F was not followed by the pure sound of (ii).

24 Conclusion as And such is the conclusion at which I arrive,* to 'I' words. from the evidence of Palsgrave and Mons. Le Héricher, of Salesbury and Lipsius, from that of modern High German and Dutch, and above all from that of our southern English dialects, both literary and provincial; that Chaucer pronounced the class of words which we have been discussing with precisely the same long i (əi) as we now give to most of them; and that in Southern AngloSaxon "the long i with an accent, as in win, wif, tím, rim, was," as Mr. E. A. Freeman has affirmed in the preface to his recently published work,† "certainly sounded as it is now."

25 Mr. J. A. H. Murray has called my attention to two facts of considerable importance in reference to Northern English. The first is that all Gaelic proper names that contain (ii) are written with y or i in Lowland Scottish, in

* There is yet one argument which I defer till after discussing some of the E words: see § 101.

+ Old English History for Children, p. xvii.—It is pleasant to be able to quote the name of any scholar who is a brother barbarian, if the system of pronunciation for which I contend is indeed so "barbarous as Mr. Sweet pronounces it in the Academy for Oct. 22nd, 1870, p. 27. Why (mǝin wǝin) should be a correct and classical pronunciation now in the mouths of a hundred millions of mankind, and yet deserve to be stigmatized as "barbarous," supposing it to have been used by their ancestors five or ten centuries ago, is not easy to discern. But the question is not to be settled by a random epithet.

26

which they are now pronounced with (əi). Thus Cantire or Kintyre, with (əi), is Ceanntìr (Kaa'ntiir) in Gaelic, Fife is Fíbh (fiiv), Skye is Sgiath (skjiie), Dalry is Dailrighe (daljriij), and so on. The second is that numerous words with that have been borrowed from Lowland Scottish into Gaelic are pronounced with (ii), as trìabh (triiv) = tribe, prìom (priim) = prime, spiorad (spiiradt) = spirit, prìs (priisj, priish) = price, Criosd (Kriisdt) = Christ, sgrìobh (skriiv) write, fìon (fiin) = wine, lìon (liin) = flax, dìsinn (diisinj, diishinj) = dicing, rìdir (riitjer, riitsher) = eques, mìle (miilə) mile, tìm (tiim) = time, pìan (piin) = O. E. pyne, pìob (piip) = pipe, and ìarunn (ii-rən) iron. These facts constitute a double argument which seems to me incontrovertible. It concerns however Northern English only, that is to say the dialects from the Humber to the Moray Frith, whose affinity with Old Norse, and partial derivation from it, quite prepare us to expect (ii) where the southern dialects had (əi).

==

dence derivable

=

But Welsh, it may be said, is the language of a people adjacent not to the Northern but the Southern English, No such evi- and there are instances of Welsh words which rom Welsh. when transferred to English underwent just the same change as Cantire, the original sound having been with (ii). A good example is ap Rhys, which has yielded us the proper names Rice, Price, and Brice. We know that the original sound was, as it still is in Welsh, (riis); and therefore these names were at first (priis) &c.: the English i, so it is argued, stood for (ii). But there is not the slightest difficulty in dealing with such cases. A Welshman bearing the name of Rhys or ap Rhys migrates into England, and spelling his name as hitherto with a y or an i, still calls himself (riis) or (apriis), and doubtless endeavours to get his neighbours to follow his example; but the name being similar to the familiar rys or prys, they pronounce accordingly, and he becomes, in spite of himself, (rǝis) or (prǝis). Another Welshman of the same name, anxious to maintain the sound, changes the spelling, and calling himself Rees or Reece succeeds in making his neighbours sound the name