71

suspicion that their not occurring more frequently is simply due to the fact that a word which is seeking a mate to rhyme with naturally looks among those of exactly the same form.* Still this is only a suspicion, and we may not tread on such thin ice with safety.

Here, however, are facts that may help us. Even before the close of the 16th century we find Smith, Hart, and Bullokar (like Gil only a little later) clearly distinguishing the a in far, mark, allow, grammar, manner, half, after, &c., from the at in another large class of words-blame, name, tame, same, bacon, capon, able, table, stable, declare, cradle, made, lady, make, take, &c. &c. And yet all these orthoepists have a third quite distinct class of words, though they now are (e) words and would rhyme with the list last given. Such are remain, say, great, plain, swear, their, besides many more, which in modern times have changed (ee) or (ee) into (ii)-receive, either, breathe, please, &c.

A had in some

Elizabeth's time,

(a) and (ee).

Since then blame, name, &c., had lost the words, in Queen sound of (aa), and had not yet acquired that of a sound between (ee), and yet were on the road to it; the conclusion seems unavoidable that in the time of Queen Elizabeth they had some intermediate (see § 28) sound. Most probably it was (æ), the sound of the a in mat or man, or (ææ), the same sound prolonged. Moreover Giles du Wes charges Englishmen in learning French to pronounce "your e almost as brode as ye pronounce your a in englysshe;" which points to the conclusion that it was the established habit of the English in Henry the Eighth's reign to sound their a, at least in many words, almost like e, that is probably (ææ). Palsgrave also (1530) clearly recognizes two as in English, one of them the same

* Although in this 19th century -ail and -ale, -air and -are, &c., are beyond question, pronounced without the slightest distinction, yet in Byron's rhymes of (e) words, setting aside the final ay, I find, in nearly seven instances out of eleven, the words are spelt alike; so also in Tom Moore's. Sir Walter Scott, on the contrary, seems to have completely emancipated himself from such bondage, and to rhyme according to the ear alone.

+ Written a by Hart, ä by Smith and Gil, á by Bullokar.

72

as the French and Italian, the other different. It is therefore clearly not Scottish influence that commenced the change from (aa) to (ee) in these words, though it not improbably gave the (aa) its coup-de-grâce.

Classes of 'A' But what of Chaucer, from whom Henry words in Chau- VIII. is distant more than a century? Answer, as in other cases,—distinguendum est.

cer: some had

(a).

Some words there are which in their earliest stage in the language had almost certainly (a), as certainly had (a) in the 16th century, and still have (a) in the 19th century. It is therefore scarcely questionable that they always have had that vowel in English. Such are large, charge, bar, spar, from the French large, charge, barre, Italian sbarra, and other such, including the interjection a! which takes Emelya with it (C.T. 1080), and therefore also probably the Latin termination in omnia, and the name of the vowel A itself (C.T. 161). In these words all the evidence is in favour of (a).

73 Others had (A).

A second class, so far back as the orthoepists will carry us, was distinct from these, being written, or described as equal to, aw by Cooper (1685), au by Butler (1633), â by Gil (1621), aụ by Bullokar (1580), and au by Hart (1569). It includes all, call, royal, several, dance, command, &c.; many of which still retain the sound of (AA): that sound we shall probably be right in assigning to them in Chaucer's time also, though in so many of these words as are of Anglo-Saxon origin there is no difference in the mode of writing these and the class preceding.

It may be added that Butler expressly states* that in his time a before l, nc, and nd was sounded as au; and it is exactly in these words that the oldest and best MSS. of

*

"A is in English, as in all other languages, the first vowel, and first letter of the Alphabet: the which, like i and u, hath two sounds: one when it is short, an other, when long: as in man and mane, hat and hate. And before / it is sounded like au: as in also, palsi, fals, altar, alter, halter: except f, v, k, l, or m, for then al hath the sound of au: before ng for ai, as in change, range, danger, stranger; before nc like au, as in chance, dance, france, lance; and also before nd, as in demand, command." P 5.

74

Chaucer seem to write almost indifferently -ance and -aunce, -and and -aund; more commonly with au. (I assume for the present, what I shall endeavour to prove further on, that au in Chaucer's time stood for (AA), as it does now.) Such words in Chaucer are al, bal, calle, Malle, halle, schal, falle, wal, thral, general, and other words of Latin derivation in -al; penaunce, pitaunce, chance, meschance, daunce, daliaunce, remembrance, suffisaunce, countenaunce, plesaunce, comaunde, &c. To these must be added land, hand, stand, and strand, which occasionally rhyme with comaunde, and show a sound other than (a) by their being not uncommonly spelt with o; and words in -ant (servant, marchant, covenant, &c.) are at least as frequently written with -aunt. Butler adds that chaunge, straunge, daunger, &c., in the North of England still retained the old pronunciation; and Chaucer's orthography indicates the same. Lastly the spelling of ensample also as ensaumple, and its rhyming with temple, suggest the French sound of the vowel in both of these. I take all of these words to have had (AA). But words in -ale do not rhyme with those in -al or alle, even when both have the final e. Such are tale, pale, ale, male (adj. and subst.), dale, nightingale, &c. I find, on running through over 6000 lines of the Canterbury Tales, twenty-two rhymes formed by these words with one another, and sixty of words in -alle with one another: only in three other instances does a word in -ale rhyme with one in alle, and in each case it is smale, the pronunciation of which is thus seen to have been at that time unsettled.

The sound then of these -ale words seems to have been with (a) or (æ), but which of these, we will for the present leave undetermined. The repeated rhyme of talys, i.e. tales, with Alys, helps very little. Alys, now (ælis), may have been (alis) in Chaucer's time; or it may have been an inexact rhyme.

A third class

A third class had a in Chaucer, but not in had short (æ). Anglo-Saxon, and the modern pronunciation is various. The A.S. form has a, which was probably (a),*

* See § 117.

75

or, thinner still, e, i.e. (e). Such are hadde, A.S. hæfde; was, A.S. wæs; black, A.S. blæc; bak, A.S. bæc; bladde, A.S. blæd; glas, A.S. glæs; bras, A.S. bras; skathe, A.S. scedan; Bathe, A.S. bæ; hath, A.S. hæf. Some of these are now pronounced with (a), some with (æ), some with (A), some with (ee). The 16th century writers do not assist us, as they do not distinguish short (a) from short (æ); but as the majority had apparently the same sound in A.S. as in modern English, it is reasonable to conclude that they have had the same sound during the whole interval. Some words of French derivation go with these-cas, pas (subst.), solas, alas, &c. rhyming with bras and was, for instance; so that we must suppose them also to have assumed this peculiarly English vowel. Then these words will be (bræs), (glæs), (alæs), (hæth), (hæd), (blæk), (pæs), just as at present; and (solæs), (Bæth) or (Bææth),* (wæs), (cæs) (skæth), (spræd), (spæk), contrary to present use.

It is notable however that the verbs pace and solace, together with space, grace, face, place, embrace, manace, Trace (i.e. Thrace), purchace, of French derivation, and lasse and asse from the A.S. lassa and assa, refuse to rhyme with gras, bras, &c. The final e however sufficiently accounts for this.

or (ææ).

A fourth class But there yet remain others chiefly of had a longer (æ) French origin in -age, -able, -ame, -ate, &c., as to which, as well as those in -ale and -ace already referred to, and numerous English words in -are, -ake, &c., the imperfect evidence seems to leave it doubtful whether (a) or (a) was the pronunciation in Chaucer's age. But though the problem is difficult, a faint ray of light seems to fall on it from the Ormulum. Assuming that make,

*It is not easy in the case of several of these words to determine whether the vowel is long or short, and therefore whether (æ) or (ææ) is the right symbol. It is certain that few or perhaps no English speakers pronounce ass, glass, grass, pass with as short a vowel as that in the first syllable of astronomical, and yet they do not so prolong the sound as a Somersetshire peasant in naming Bath. We have in fact, as Mr. Melville Bell and others have pointed out, various degrees of length of our vowels, minute differences of quantity as well as quality in different words.

from the A.S. macian, and take, from the O.N. taka, were in their earliest forms sounded with (aa), we find that in Ormin's time they had undergone a change, at least in the imperatives, which he writes macc and tacc. This doubled consonant is Ormin's mode of indicating a short preceding vowel; and these imperatives in this form have the same vowel as that of annd, att, bacc, brass, chappmenn, &c. It is true, Ormin's short a, like his long one, may have stood for more sounds than one; so that before r, as in arrke, arrmess (i.e. arms), arrt, the vowel may have been the short (a); but the point that I call attention to is the fact that the a in macc and tacc has been shortened, and before the guttural tenuis it is more likely, as in the other instances, that the sound was (a). And it may be so that the very thing which Ormin intended by his ǎ (sic) was (ææ); for six out of the eight words which he so writesthe other two do not occur in Chaucer-are among the very words which we are discussing. They are, dăle, hătenn (also hatenn), late (also late), năme (also name), tåkenn, and tåle (also tale). And we are not at all bound to assume that the signified what we now use it to signify, especially as Ormin had another mode of showing the short vowel. At any rate this & indicates some other sound than the a of afell (Icel. afl), afledd, abidenn, abufenn, abutenn, adl, anig, &c.,*—in other words, some change; and since four of these words have forms in a also, the change itself seems to have been incomplete, and the pronunciation unsettled when Ormin wrote. If then English words had (aa) in Anglo-Saxon, but at least since the seventeenth century have had (ee), and there are even as soon as the early part of the thirteenth century signs of an incipient change; it is but reasonable to suppose that that change was somewhat advanced when Chaucer wrote, a century and a half later. And if we may so conclude for English words, it is highly probable

Very many of Ormin's a words, as ald, an, nan, gan, gast, mare, rap, were from A.S. words with a, and I therefore believe them to have been sounded with (00).