{v}
§ 1. The English version of the Lay of Havelok, now here reprinted, is one of the few poems that have happily been recovered, after having long been given up as lost. Tyrwhitt, in his Essay on the Language and Versification of Chaucer, has a footnote (no. 51) deploring the loss of the Rime concerning Gryme the Fisher, the founder of Grymesby, Hanelok [read Havelok] the Dane, and his wife Goldborough; and Ritson, in his Dissertation on Romance and Minstrelsy (vol. i. p. lxxxviii of his Metrical Romanceës), makes remarks to the same effect. It was at length, however, discovered by accident in a manuscript belonging to the Bodleian Library, which had been described in the old Catalogue merely as Vitae Sanctorum, a large portion of it being occupied by metrical legends of the Saints. In 1828, it was edited for the Roxburghe Club by Sir F. Madden, the title-page of the edition being as follows:—‘The Ancient English Romance of Havelok the Dane, accompanied by the French Text: with an introduction, notes, and a glossary, by Frederick Madden, Esq., F.A.S., F.R.S.L., Sub-Keeper of the MSS. in the British Museum. Printed for the Roxburghe Club, London. W. Nicol, Shakspeare Press, mdcccxxviii.’ This volume contains a very complete Introduction, pp. i–lvi; the English version of Havelok, pp. 1–104; the French text of the Romance of Havelok, from a MS. in the Heralds’ College, pp. 105–46; the French Romance of Havelok, as abridged {vi} and altered by Geffrei Gaimar, pp. 147–80; notes to the English text, pp. 181–207; notes to the French text, pp. 208–10; and a glossary, &c., pp. 211–63. But there are sometimes bound up with it two pamphlets, viz. ‘Remarks on the Glossary to Havelok,’ by S. W. Singer, and an ‘Examination of the Remarks,’ &c., by the Editor of Havelok. In explanation of this, it may suffice to say that the former contains some criticisms by Mr. Singer, of which a few are correct, but others are ludicrously false; whilst the latter is a vindication of the general correctness of the explanations given, and contains, incidentally, some useful contributions to etymology.
§ 2. Of this first edition in 1828 but few copies were printed; and, as the work was seldom to be met with, a new edition was printed by myself for the Early English Text Society in 1868, with the permission and kind assistance of the first editor. A later edition was issued from the stereotyped plates, with few corrections atud additions, in 1889. An edition by F. Holthausen, with a carefully revised text, was published in London, New York, and Heidelberg in 1901, which I have consulted with much advantage.
§ 3. Description of the MS. The unique text of the poem is extant in MS. Laud Misc. 108, in the Bodleian Library, Oxford. It begins on fol. 204., and is written in double columns, each of which contains forty-live lines. It ends on the back of fol. 219, and is immediately followed, in the same handwriting, by the Geste of King Horn, re-edited by Mr. Joseph Hall in 1901, who gives a description of the MS. in his Preface, pp. viii–x. It appears to be a composite MS., made up of three distinct parts. Part I is chiefly occupied by Lives of the Saints, for which see Horstmann’s Early South-English Legendary, printed by the Early English Text Society in 1887; and the date of the handwriting in {vii} this section is supposed by Mr. Hall to be about a.d. 1290. Part II, containing Havelok and King Horn, is perhaps a little later, and may be dated about a.d. 1310. Part III is of a much later date, and contains the Lives of St. Cecilia and St. Blaise (also printed by Horstmann); St. Alexius, printed by Dr. Furnivall in 1878 (E.E.T.S., no. 69); and a poem called Somer Soneday, printed in Reliquiae Antiquae, vol. ii. pp. 7–9; followed by a few other scraps. In the note to l. 2933, at p. 125, I mention Zupitza’s conjecture that an older MS., from which the existing copy was made, contained only twenty lines to the page; and Mr. Hall has noted that twenty lines in the copy of King Horn are found to be out of place, which furnishes strong evidence as to the correctness of the suggestion. If so, the MS. must have been made with small pages for the purpose of portability, and would have been well suited for use by a wandering minstrel or reciter of poems. I have an Anglo-French MS., containing some Statutes, in my own possession, having about twenty lines to the page, and the pages measure only 4 inches by 3½. The two romances, extending to less than 4,600 lines, would not occupy 120 leaves; and if bound similarly to that just mentioned, would only be 1½ inch in thickness. A folio is lost between fol. 211 and 212, but no notice of this has been taken in numbering the folios; see p. 53, where fol. 212 should have been fol. 213. A facsimile of fol. 207, back, is given as Plate VII in my ‘Twelve Facsimiles of Old English MSS.,’ published in 1892; and a portion of the same Plate is reproduced as an accompaniment to the present volume, containing ll. 632–53, which gives a fair idea of the character of the handwriting.
The words are often very close together. The initial letter of every line is written a little way apart from the rest, as in William of Palerne, and other MSS. Both the long and {viii} short s (ſ and s) are used. The long s is in general well distinguished from f, and on this account I have taken the liberty of printing both esses alike, as my experience in printing the Romans of Partenay proved that the difficulty of avoiding misprints is greater than the gain of representing the difference between them. The chief point of interest is that, as in early MSS., the long s is sometimes found at the end of a word, as in ‘uſ’ in l. 22, and ‘iſ’ in l. 23. The following are all the examples of the use of this letter in the first twenty-six lines; ſo (4), wicteſte (9), ſtede (10), criſt, ſchilde (i6), Kriſt, ſo (17), ſo (19), ſchal (21), Krift, uſ (22), iſ (23), ſtalworpi (24), ſtalworpeſte (25), ſtede (26). Some difficulty is caused by the use of the Saxon letter w (ƿ). This letter, the thorn-letter (þ), and y, are all three made very nearly alike. In general, the y is dotted, but the dot is occasionally omitted. Wherever the letter really appears to be a w, I have denoted it by printing the w as an italic letter. The following are, I believe, the only examples of it. Witdrow = withdrew, l. 502; wit, 997 (footnote); we, 1058; was, 1129 (cf. ‘him was ful wa,’ Sir Tristram, l. 2769); berwen, 1426 (written ‘berwen’ in l. 697); wat = said (?), 1674 (footnote); we, miswritten for wo = who, 1914. This evidence is interesting as showing that this letter was then fast going out of use, and I think that we may safely date the final disappearance of this letter from MSS. near the year 1300. As regards the th, we may remark that at the end of a word both þ and th are used, as in ‘norþ and suth,’ l. 434; sometimes th occurs in the middle of a word, as ‘sithen,’ l. 1238, which is commonly written ‘siþen,’ as in l. 399. The words þe, þat, þer, &c. are hardly ever written otherwise. But the reader may find many instances in which th final represents the A.S. ht (M.E. ght), as in brouth, 57, nouth, 58, lith, 534, þouth, 1190, &c.; cf. § 4. The letter t is sometimes shortened {ix} so as nearly to resemble c, and c is sometimes lengthened into t. The letters n and u are occasionally alike, but the difference between them is commonly well marked. The i has a long stroke over it when written next to m or n. On the whole, the writing is sufficiently distinct. The poem is marked out into paragraphs by the use of large letters; and I have introduced a slight space at the end of each paragraph, to show this more clearly. In printing the MS., I denote the expansions of marks of contraction in the usual way, by the use of italics; thus in l. 9, ‘man’ is printed instead of the form ‘mā,’ as in the MS.; and the curl denoting er is represented by printing ‘euere’ in l. 17. In l. 6, the dot below the second e in ‘yedẹ’ signifies that the e is mute.
§ 4. The Spelling. The spelling appears, at first sight, to be of a very irregular and lawless character, but is easily understood in the light of my discovery (in 1897) that many of our earlier MSS., especially those of the thirteenth century, abound with spellings which can only be understood rightly when we observe that the scribe was of Norman birth, and more accustomed to the spelling of Anglo-French than to that of the native language of the country, which he had acquired with some difficulty, and could not always correctly pronounce. This curious phenomenon, due to the resolute attempt on the part of the Norman to acquire English, is fully explained in my paper on ‘The Proverbs of Alfred,’ read on May 7, 1897, and printed in the Transactions of the Philological Society for that year (p. 399). I may note, by the way, that one permanent result of the reflex action of Norman upon the pronunciation of English is familiar to most people, viz, the total loss of the guttural gh in the pronunciation of standard English, though it is still conscientiously written down. Such a rhyme as that of light with bite was, for Chaucer, impossible; but the loss of the guttural was so {x} complete about the year 1400, that it came quite naturally to Lydgate, and to all his successors who employed the standard Midland dialect. With this clue, the spelling of our MS. becomes perfectly intelligible, and the English consonants are so easily recovered, that I have not hesitated to restore the usual Middle-English spelling in a large number of instances, relegating the Anglo-French spellings of the MS. to the bottom of the page, where every variation between the printed text and the MS. is carefully recorded, according to the notice at the bottom of p. 1. All words printed between square brackets are either supplied from conjecture to complete the sense (like the ‘and’ in l. 32), or denote corrections (like ‘wihtest’ in l. 25), where the MS. reading ‘stalworþeste’ ruins the metre, and was caught from the line above. The correspondences between the A.F. and M.E. spellings are easily tabulated, and are fully accounted for in the following sections. The comparison is one of no small moment, as it easily explains the numerous eccentricities of MSS. in the twelfth and thirteenth centuries. The matter has not yet received much attention; yet the deviation from the true M.E. spellings in such a text as the ‘Old English Homilies’ can be counted by the hundred. We have to remember that, when an Anglo-French scribe of the thirteenth century (or earlier) wrote out an English poem, he frequently spelt the words according to his own pronunciation, in a way which a native would hardly have employed. The chief peculiarities due to such habits have been enumerated, in the form of canons, in an Appendix to my Notes on English Etymology, at p. 471. I here repeat them for convenience, and show their application to Havelok afterwards.
§ 5. Initial Sounds in Anglo-French. To an Anglo-French scribe the difficult initial sounds were h, sh, th, wh, wu (or wo), and y (consonant). Of these, sh, th, wh, wu, y {xi} can hardly be said to exist in Old French, and h was very slight. I take them in order.
1. The French initial h was weak, the English h was strong. Hence arose a confusion, as in Avelok for Havelok, but Henglish for English; see § 7. Before 1400 the Norman had learnt his lesson, though he had unsettled the pronunciation of the lower classes.
2. Old French had no initial sh. The modern F. ch was pronounced by the Norman as ch in church. In trying to say sh, he merely said s. Hence he wrote sal for shal, even in Southern texts. The medial sh often appears as ss; in the Ayenbite of Inwyt, we even find ssss; as in esssse, an ash. But before 1400, the Norman had acquired the sound.
3. Old French had no initial th. Hence the A.F. scribes adopted the A.S. thorn-letter (þ) as a new symbol. Some used the eth, or crossed d (ð); but this soon went out of use. The substitution of t for th, as in Torp (Domesday Book) for Thorp, is rare, except after d or t at the end of the preceding word, as in at te for at the. But the Normans soon acquired the initial sound; the final th took them longer to learn. See canons 14, 15 at p. xiii.
4. The English wh, as in modern Northern English, became a mere w, as in wat for what (hwat). Many Normans never acquired the sound, so that it has disappeared from Southern English.
5. The Norman could sound w before a, as in warant (warrant); or before e, as in werre (war). But not before u (sometimes written o). Hence we find ulf, wlf, for wulf, wolf. The w in this wlf was pronounced like the Welsh (vocalic) w; and this (I believe) accounts for the Welsh symbol. They unsettled the pronunciation of the lower orders, who still say ’ood for wood.
6. There was no common use of initial y (consonant) in {xii} Norman. Hence the spelling ou for you in many MSS. We all say itch, not yitch, though the M.E. form was yicchen.
7. Medial Sounds. The chief one to be noted is r. In spite of the strong trill of the r in modern French, the evidence suggests that the true English r was even stronger, as in Northern English. Burns has farls as a dissyllable in his Holy Fair, stanza 7. And this may explain the fact that, in Havelok, the scribe twice writes arum, harum, instead of arm and harm; whilst, conversely, the modern Southern English r has been so remarkably weakened, that it is commonly almost unheard. Compare the French final r in parler, aimer.
§ 6. Final Sounds. They had difficulties with final gh, ght, ld, lk, nd, ng, nk, t, and th.
8. The gh was a Norman symbol, to express the sound of A.S. medial or final h, sounded like the ch in German. But they failed to acquire the sound, so that in modern English we either drop it altogether (usually after a long vowel or in an unstressed syllable, as in dough, borough), or substitute an f after a short vowel (as in rough, tough); or we turn it into k (as in loch, or in elk for O. Mercian elh). We even find wealcstoda for wealhstoda in late A.S. of the latter half of the eleventh century (Napier’s Glosses, 2422); and Anglo-Norman habits reinforced this tendency.
9. Ght was a most difficult sound for them; for which st was sometimes substituted (but only at an early date). Hence it is sometimes written as wt or t.
10. Final ld was at first difficult; we find fel for feld (field).
11. Final lk was also at first difficult; we find il in some MSS., and ilek in others, in place of ilk.
12. Final nd became either nt or n. And is frequently spelt as ant or an.
{xiii}
13. Final ng, nk were new to them. We find a curious
confusion; as kinc for king, dring for
drink, bringhe for M.E. bringe. They
unsettled the true pronunciation, so that shillin for
shilling is still common.
14. It is surprising to find that the English t sounded differently. I suppose that it sounded to them stronger, with a sort of emphatic splutter when final. Hence we find neth for net. This final th is perfectly distinct from þ or the modern E. th, and signified a t followed by an explosive sound; like nett’. We even find thown for town in some MSS.; like the Irish Theddy for Teddy.
15. Final th, especially in an unaccented syllable, is turned into d or t. But the sound was fully acquired before 1400.
16. The A.F. z represented ts (as in German); and even c (before e) denoted the same sound. Hence assets is from A.F. asses; and early M.E. milce represents miltse.
§ 7. Restoration of the Middle English spelling. Most of the above peculiarities occur in Havelok, and they are easily put right. Thus the word neth in l. 752 obviously means ‘net,’ and would be spelt net in any M.E. work written out by an English scribe. Hence I print net in the text, but neth in the footnote and in the Glossary. The number of obvious corrections is large; and the A.F. spellings are curious and instructive. I now give numerous examples.
1. The initial h is dropped by the scribe in haueden, 163; Hauelok, 503; heþen, 690; his, 735; hosed, 971; Hauelok, 1395; &c. See the footnotes.
A needless h, not pronounced, is prefixed to er, 15; euere, 17, 88; olde, 30; ayse, 59; elde, 128; ele, 146; ore, 153; old, 192; &c. It is very common.
2. S is miswritten for sh in fleysh, 216; neysh, 217; Shal, 628; shame, 1941; Shule, 2419; shulde, 2835. The usual M.E. crusshe actually appears as cruhsse, 1992.
{xiv}
3. The case of th is discussed under no. 14; p. xv.
4. W is miswritten for hw (= wh) in Hwo, 4, 76; Hwil, 6; hwit, 48; hwat, 117; Hwom, 197; Hwan, 220; hweþer, 292; Hwider, 1139; &c. But the scribe often has hw also; writing Hweþer in 294, in place of weþer only two lines above; so that he knew the English spelling. Observe also qu for hw; as in qual, 753; quanne, 134, 204; qui, 1650; which I have left unaltered.
5. The A.F. w, considered as a vowel, really meant uu, as its name imports. This sound (A.S. ū) was written u only in early MSS.; the later ones have the regular French ou. Hence Hw in l. 93 really means Hū; whilst in l. 120 we have simply W as a spelling of the same word (canon 1). So also wman, 174, 281, represents mod. E. ’ooman, i.e. woman, and should be written wuman in English; yw in l. 453 represents you, as it rhymes with nou. Wurthe is written wrþe, 434; so also wulf becomes wlf, 573; swungen, wunden appear as swngen, wnden, 226, 546. In l. 464, the spelling þs means that þ has been miswritten for ƿ (A.S. w), and the latter represents the A.F. w, so that the word meant is the A.S. ūs, often written ous, but usually us with the short u; see l. 461, only three lines above. In l. 2992, we actually find hwou (= hūū), in which the symbol for ū appears twice over, and in different forms!
6. Canon 6 is not illustrated by Havelok.
7. The English r (in Havelok) was very strongly trilled, and sometimes counts for a syllable in the scansion; as in þor(e)nbake, 759, 831; for(e)þ, 810, 821; Cor(e)nwaile, 884, 2908; nor(e)þ, 1255; bor(e)d, 1722; car(e)l, 1789; er(e)l, 2861; þ’er(e)ldom, 2923; hence I have allowed the spellings boren, koren, to stand in ll. 1878, 1879; as well as arum, 1982, 2408; harum, 1983, 2409. So also sembling, 1018, was pronounced as sembeling (in three syllables).
{xv}
8, 9. The guttural h (gh) is often wholly ignored; as in
browt for brouht, 58; nowt for nowht, 123;
dowter, 258; knit, 2427. For further remarks, see canon 14
(below).
10, 11, 12, 13. The scribe occasionally reduces final ld to l; as in hel for held, 109; gol for gold, 357; shel for sheld, 489; bi-hel for bi-held, 1645.
Also lk sometimes appears as l; as in il for ilk, 218, 1644; Hwilgat for Hwilkgat, 836. Hence the absurd word kilþing, 1736, is merely a bungled form of the very common ilk þing, with the k in a position where it could more easily be uttered.
Even lt is reduced to l in shal for shalt before a t, 1161 (with shalt in full, before a vowel, in the very next line); cf. shal for shalt before th, 1273. And rd is reduced to r once; as in forthwar, 731.
Final nd appears as nt in ant, 36; but the d is usually lost, as in an, 29, 58, 131, 151, &c.; cf. an’ for and in modern English. So also lon for lond, 340; spen for spend(e), 1819.
In a similar way, final st once appears as s; see bes for best, 354.
The A.S. ng was really sounded as ngg, or as the ng in E. finger, and was only reduced (in M.E.) to the ng in ring when final. The scribe attempts to show this in a few instances only, writing bringhe, þinghe for bringge, þingge in ll. 65, 66. I substitute ng for uniformity. In l. 2561 he writes rang for rank, but we have þank in the previous line. Cf. bringge, 1381; puttingge, 1042.
14. The most characteristic spelling is the frequent use of th to express a strongly pronounced final t. This th is quite distinct from the modern E. th, and is therefore never written þ. That the F. t was weaker or less distinct than the English one is rendered probable by the evidence of modern French; cf. E. met with F. met.
{xvi} Examples occur in with for wit, for hwit, white, 48; þuruth for þurut, for þurh-ut, throughout, 52; nouth for nout, for nouht, 58; nicth for nict, for niht, 143; woth for wot, 213; leth for let, 252; neth for net, 808, 1026; greth for gret, 1025; weddeth for weddet, more correctly wedded, and beddeth for beddet, more correctly bedded, 1127, 1128.
But the most extraordinary variations occur when the scribe has to express ht (ght). As he denotes the guttural not only by h (the most usual symbol), but also by c or ch, or sometimes suppresses it altogether—whilst at the same time he expresses final t either by t or th—we obtain as the possible varieties of ht the following, viz. ct, cht, t, cth, chth, and th, of which all but the fifth form actually occur. Hence we find thoucte, 197 ; bitaucte, 206; awcte, 207; mowcte, 210; mouchte, 147; þouchte (actually miswritten þouthte, by the common confusion of c with t), 1073; browt, 58; nowt, 123; knit, 2427; micth, 35; knicth, 77, 80; ricth, 78; micthe, 88, 199; brouth, 84; knith, 87, 90; mouthe, 145; nouth, 149; &c. In every case I replace this Protean symbol by ht, as in A.S. and in the Harleian MS. of King Horn and many other M.E. poems.
15. The final th sometimes appears as d (for ð), as in haued for haueth, 1372 (cf. l. 1373); spared for spareth, 2813; but much more frequently as t, as in Herknet, 1; wit for with, 19, 52, 113, 144, &c.; this is very common. Conversely we find quodh for quod in l. 1800; unless indeed quoth is intended.
§ 8. Further corrections. Beside the above variations, which can all be easily accounted for by considering the difficulties which the scribe had in acquiring or expressing the true English pronunciation, there are many others which are less intelligible. The scribe, for example, frequently omits a final l; as in a for al, 610, 936; we for wel, 115, 287, {xvii} 392, 772; mike for mikel, 960, 1744, 1761, 2336 (but mikel in l. 2352; and it rhymes with swikel, 1107, and with fikel, 1209, 2798). The common mistake of writing o for e occurs in wol, 18; hoslon (for hoslen), 362; hwor (for hwer=hweþer), 1119; horn, 1298; holed, 2039; conversely, e is written for o in eld, 546; here, 742. The slight curl which signifies er is constantly omitted before a final e; giving þe for þ(er)e, 142, 476, 639, &c., better spelt þer; so also oþe for oþ(er)e, better oþer, 861; he for h(er)e, better her, 1911; we for w(er)e or wer, 2055; unless, indeed, the scribe dropped a final r as he sometimes dropped a final l. This perhaps arose from carelessness, as we find Rirth for Ricth (=Riht), 37; wrobberes for robberes, 69 (because preceded by Wreieres); serf for self, 1667; maude for made, 436. In l. 31, his copy doubtless had tayn for thayn, the t being just like c; hence he misread it as cayn, and wrote it as kayn; which singular error is repeated at l. 1327. He also confuses the symbol þ (th) with the A.S. ƿ (w), writing þi for ƿi, better spelt hwi, 2578; and likewise the symbol for long s (ſ) with f, writing þhes (with ſ) for þhef, which again is an error for þef, 2289; and slawen (with ſ) for flawen, 2476. He frequently omits an essential final e, as in rim, rym, 2995, 2998; or a necessary final n, as in drawe, 2477; þeþe, 2629; wreke, 2849. We find a curious instance of anticipation, i.e. the too soon writing down of a coming letter, as in meme for neme, 2201; and again, of reminiscence, or the too late writing down of a letter that has already occurred, as in Skabbeb, 2505; togidede, 2972. But the commonest error is the careless omission, sometimes of single letters, sometimes of several, as in kaue for knaue, 481; bigge for brigge, 881; shres for sheres, 857; goldebow for Goldeborw, 1103, &c.; and even le for let þe, 1827; swe for swiþe, 2140; louen for loueden, 2198; &c. Sometimes there is a letter too much, as in anilepi for anlepi, {xviii} 2107; talevaces for talevas, 2323; hungred for hungre, 2454; leuin for leun, 2690. The omission of words (necessary both to the sense and the metre) is rather common; see several instances denoted by the use of square brackets. It will now be readily understood that many emendations have necessarily to be made; and many more have been proposed which I have not always noticed, as there is a tendency on the part of critics to increase their number to too liberal an extent. A few peculiarities were probably intentional; as, e.g., latus for lat us, 1772; bihetet for bihete it, 647; hauedet for hauede it, 714; yeuenet (miswritten youenet) for yeuen it, 1643; setes for set es, 784; dones for don es, 970; see note to l. 1174. The text is, in fact, in several respects faulty; but when the spelling has been partially reformed, and all the more obvious errors corrected, it becomes fairly readable, and the merits of the narrative can be perceived and esteemed.
§ 9. Other peculiarities. A few other peculiarities deserve notice. The vowel u answers to the modern ou in the words prud, 302; suth, 434; hus, 740; but, 1040; spusen, 1123; cf. hws in l. 1141. Mr. Ellis shows, in his Early English Pronunciation, chap. v, that in pure specimens of the thirteenth century there is no ou in such words, and in the fourteenth century no simple u. This furnishes a ready explanation of the otherwise difficult sure, in l. 2005; it is merely the adverb of sour, ‘sourly’ being used in the sense of ‘bitterly’; to bye it bitterly, or bye it bittre, is a common phrase in Piers Plowman. Other spellings worth notice occur in ouerga, 354; stra, 315; plawe, 950. There are several instances of words joined together, as biddi, 484; haui, 2002; wiltu, 905; wenestu, 1787; shaltu, 2186; wilte, 528; thenkeste, 578; shaltou, 1800; sawe, 338; latus, 5772; where the personal pronouns i, þu, we, us are added to the verb. Hence, in l. 745, it is very likely that calleth is written {xix} for callet, i.e. ‘call it’; though the rhyme requires it calle. In like manner goddot is contracted from God wot; and þerl from þe erl; see the Glossary.
§ 10. Nouns. As regards the nouns employed, I may remark that the final e is almost always sounded in the oblique cases, and especially in the dative case; as in nedè, stedè, &c. (see ll. 86–105); willè, 85; wisè, 1713; blissè, 2587; cricè, 2450; cf. the adjectives lessè, 1830; longè, 2299; also the nominatives rosè, 2919; newè, 2974. Frend is a pl. form; cf. hend, 505, 2069, 2444. In the plural, the final e is fully pronounced in the adjectives allè, 2; hardè, 143; bleikè, 470; starkè, 1015; fremdè, 2277; and in many others; cf. the full form boþen, 2223. Not only does the phrase none kines, of no kind, occur in ll. 861, 5540, but we find the unusual phrase neuere kines, of never a kind, in l. 2691; though neuere is here almost certainly an error for none. Among the numerals, we find not only þre (1399), but þrinne (four times).
Pronouns. The first personal pronoun occurs in many forms in the nominative, as i, y, hi, ich, ic, hic, and even ihc; the oblique cases take the form me. For the second person, we have þu, þou, in the nominative, and also tu, when preceded by pat, as in l. 2903. We may notice also hijs for his, 47; he for they (repeatedly); sho, 112, scho, 126, sche, 1721, for she; and, in particular, the dual form unker, of you two, 1882; and the pl. es, ‘them,’ for which see the Glossary. This es or is is possibly short for his, actually used in the accusative plural, though some equate it to the G. sie; see examples in Matzner, Glossary, ii. 449, col. 2, and the two articles on †His in the New Eng. Dict. p. 302, col. 1, of the letter H. The most noteworthy possessive pronouns are minè, pl. 1365; þinè, pl. 620; his or hise, pl. hisè, 34; urè, 606; youres, 2801; hirè, 84, 2918, with which cf. the dat. sing. {xx} hirè of the personal pronoun, 300. Þis is plural, and means these, in l. 1145; but in l. 606 it is short for þis is; see the note. As in other old English works, men is frequently an impersonal pronoun, answering to the French on, and is followed by a singular verb; as in men ringes, 390; men seyt and suereth, 647; men fetes, 2341; men nam, 900; men birþe, 2101; men dos, 2434; cf. folk sau, 2410; but there are a few instances of its use with a plural verb, as men haueden, 901, men shulen, 7471. The former is the more usual construction.
1 But in l. 747, shal men would read better; and l. 901 should rather be: ‘Þan men him hauede holpen dune’; where hauede = hav’de.
Verbs. The infinitives of verbs rarely have y- prefixed; three examples are y-here, 11; y-lere, 12; y-se, 334. Nor is the same prefix common before past participles; yet we find i-gret, 163; i-groten, 285; and i-maked, 5, as well as maked, 23. Infinitives end commonly in -en or -e, as riden, 26, y-lere; also in -n, as don, 227, leyn, 718; and even in a vowel, as flo, 612; slo, 1364; fle, 1195. The present singular, 3rd person, of the indicative, ends both in -es or -s, and -eth or -th, the former being the more usual. Examples are longes, 396, 1443, haldes, 1382, fedes, 1693, bes, 1744, comes, 1767, leues, 1781, 2105, glides, 1851, þarnes, 1913, haues, 1952, etes, 2036, beres, 2323, fetes, 2341, bedes, 2392, ledes, 2573, strenes, 2983; dos, 1913; also eteth, 672, haueth, 804, bikenneth, 1269, suereth, dereth, 647, 648; liþ , 673, doth, 1876. The full form of the 2nd person is -est, as louest (before a vowel), 1663; but it is commonly cut down to -es, as haues, 688, etes, 907, getes, 908, slepes, 1283, weldes, 1359; cf. dos, 2390, slos, 2706, mis-gos, 2707; and this still more marked in rhyme, as wenes, 598. The same dropping of the t is observable in the past tense, as in dedes, 2393, reftes, 2394, feldes and claddes, 2907. The A.S. meaht, miht, {xxi} answers to maght, 1348; cf. ll. 689, 852, 1219. In the subjunctive mood the -st is wanting, as in Anglo-Saxon, and hence the forms bute þou gonge, 690, þat þu fonge, 856, &c.; cf. bede, 668. In the 3rd person, present tense, of the same mood, we have the -e fully pronounced, as in shildè, 16, yeuè, 22, lesè, 333, leuè, 334, redè, 687; and in l. 544, wreken should undoubtedly be wrekè, since the -en belongs to the plural, as in moten, 18. The plural of the indicative present commonly ends in -en, as, we hauen, 2798, ye witen, 2208, þei taken, 1833; or, very rarely, in -eth, as ye bringeth, 2425, he (they) strangleth, 2584. Sometimes the final -n is lost; note wone, 1325, to-deyle, 2099, binde, 2583 (in rhymes). The present tense has often a future signification, as in eteth, 672, etes, 907, getes, 908; and in beth, 1260, bes, 2744.
Note. The rhymes show that the third person singular in -es belongs to the original dialect of the poem (examine the examples). It was afterwards copied out in the south of England, by a scribe who frequently turned -es into -eth. The only examples at the end of a line are suereth, dereth, 647–8; instead of sweres, deres.
Past tense. Of the third person singular and plural of the past tense the following are selected examples. Weak Verbs: hauede, 770, sparede, 898, yemedè, 975, semedè, 976, sparkëde, 2144, þankede, 2289; pl. loueden, 955, leykeden, 954, woundeden, 2429, stareden, 1037, yemede, 2276, makeden, 554, sprauleden, 475. Also calde, 2225, gredde, 2427, herde, 2420, kepte, 879, fedde, 786, ledde, 785, spedde, 756, clapte, 1814, kiste, 2279; aute, 743, laute, 744, bitauhte, 2212; pl. herden, 150, brenden, 594, kisten, 2162, ledden, 1246. Compare the past participles osed, 971, mixed, 2533, parred, 2439, gadred, 2577; reft, 1367, wend, 2138, hyd, 1059; told, 1036, sold, 1638, wrouth=wrouht, 1352. There are also at least three past participles in -et, as spuset, 1266, slenget, 1923, grethet, {xxii} 2615; to which add weddeth, beddeth, 1127; but this -et or -eth is an A.F. form of -ed. In l. 2057, knawed seems to represent the modern ‘knowed’; see the note.
Strong Verbs: third person singular, past tense, bar, 815, bad, 1415, yaf or gaf, 218, 315, spak, 2389, kam, 766 (spelt cham, 1873), nam, 900, kneu, 2468, hew, 2729, lep, 1777, let (spelt leth), 2651, slep, 1280, wex, 281; drou, 705, for, 2943, low, 903, slow, 1807, hof, 2750, stod, 986, tok, 751, wok, 2093; pl. beden, 2774, youen or gouen, 164, comen, 1017 (spelt keme, 1208), nomen, 2790 (spelt neme, 1207), knewen, 2149, lopen, 1896, slepen, 2128; drowen, 1837, foren, 2380, lowen, 1056, slowen, 2414, &c. By way of further examples, I may instance the singular forms bigan, 1357, barw, 2022, karf, 471, swank, 788, warp, 1061, shon, 2144, clef, 2643, sau, 2410, grop, 1965, drof, 725, shof, 892, fauth (= fauht), 1990; pl. bigunnen, 1011, sowen, 1055, gripen, 1790, driue for driuen, 1966; also bunden, 2436, scuten, 2431 (spelt schoten, 1864, shoten, 1838), leyen, 2132, &c. Compare the past participles boren, 1878, youen or yeuen, 1643, cumen, 1436, nomen, 2265 (spelt numen, 2581), laten, 1925, waxen, 302, drawen, 1925, slawen, 2000. The two last become drawe, slawe in ll. 1802, 1803.
We should also observe the past tenses spen (i.e. spend’), 1819; stirt’, 812, citte, 942, bere (subj.), 974, kipte, 1050, flow, 2502; and the past participles demd for demed, 2488, giue for giuen, 2488, henged, 1429, keft, 2005, plat, 2755.
Imperative Mood. Examples of the imperative mood singular, 2nd person, are et, sit, 925, late, 1376, bringgè, 1381; in the plural, the usual ending is -es, as in liþes, 2204, comes, 1798, folwes, 1885, lokes, 2292, bes, 2246, to which set belong slos, 2596, dos, 2592; but there are instances of the ending -eth also, as in cometh, 1885, yeueþ, 911, to which add doth, 2037, goth, 1780; herknet (for herkneth), 1. Indeed, {xxiii} both forms occur in one line; as in Cometh swiþe, and folwes me (1885).
Of reflexive verbs, we meet with me dremede, 1284, me haueth met, 1285, me þinkes, 2169, him hungrede, 654, him semede, 1652, him stondes, 2983, him rewede, 503. The present participles end most commonly in -inde, as fastinde, 865, grotinde (? gretinde), 1390, lauhwinde, 946, plattinde, 2282, starinde, 508; but we also find gangande, 2283, driuende, 2702. Compare the noun tiþande, 2279, which is a Norse form, tiðindi (pl.) being the Icelandic for ‘tidings.’ The suffix -ing occurs as a noun-ending only, never (that I remember) in the present participle. Examples of it are greting, 166, dreping, i.e. slaughter, 2684, buttinge, skirming, wrestling, putting, harping, piping, reding; see ll. 2322–7; also coruning, 2948, ioying, 2949. Amongst the auxiliary verbs, may be noted the use of cone, 622, as the subjunctive form of canst; we mone, 840, answering to prov. E. mun, i.e. must. We should particularly observe the use of the comparatively rare verbs birþ, it behoves, pt. t. birde, it behoved, and þurte, he need; for which see the Glossary.
The prefix to- is employed in two senses, as explained in the Glossary, s.v. To-. In to-brised, to-deyle, &c., it is equivalent to the German zer- and Latin dis-; of its other and rarer use, wherein it answers to the German zu- and MoesoGothic du-, there is but one instance, viz. in the word to-yede, 765, which signifies ‘went to’; cf. Germ. zugehen, to go to, zugang (A.S. togang), access, approach. There are several instances of the peculiar syntax whereby the infinitive mood active partakes of a passive signification, as in he made him kesten in feteres, he caused him to be cast into fetters; l. 81. It may be considered as a phrase in which we should now supply the word men, and we may interpret it by ‘he caused [men] to cast him into fetters and to fasten him securely’; {xxiv} for in ll. 1784, 1785, the phrase is repeated in a less ambiguous form. See also l. 86. So also, in ll. 2611, 2612, we must consider keste, late, sette, to be in the infinitive mood. This construction is at once understood by comparing it with the German er liess ihn binden, he caused him to be bound. In l. 2352, appears the most unusual form ilker, written for ilk here, i.e. each of them. The word þrie, 730, answers to the M.E. adverb thrie, thrice, but it must be an error, probably for yete; liues, 509, is an adverb ending in -es, originally a genitive case. Þus-gate is, according to Dr. Morris, unknown to the Southern dialect; it occurs in ll. 785, 2419, 2586; cf. hwilgat, 836.
Phonology. The text is too corrupt, and the dialect too mixed, for satisfactory results; except as regards the consonants, which have been already considered. Thus ‘both’ occurs as baþe, rhyming with raþe, 2594, 2936; as beþe, rhyming with rede, 360, 1680, which also rhymes with bede, 2084; and as boþe, rhyming with wroþe, 2973 (cf. 2977)1. Assonances are common; as in 21, 172, 693, 1303, 1397,&c.
1 Holthausen has attempted the correction of the vowel-sounds, but admits that some uncertainty remains. See pp. x and xii of his edition.
I may add that Havelok contains as many as nine expressions, which seem to refer to proverbs current at the time of writing it. See ll. 307, 600, 648, 1338, 1352, 1693, 2036, 2461, 2983.
§ 11. Date of Composition. The present poem cannot easily be dated without considering the dialect (see § 12). But we must first of all look at the internal evidence.
Prof. Hales has pointed out (Folia Litteraria, p. 30) that the curious reference, in ll. 139, 265, to the extent of England as reaching ‘from Roxburgh to Dover,’ points to a date when Roxburgh had become a border fortress. But this was not the case till it was seized by Edward I in 1296. In {xxv} l. 1006, there is an allusion to a parliament summoned to meet at Lincoln; and a parliament actually met at that city early in 1301, at which the archbishop of York was present; cf. l. 1178, and see note to l. 1006. There may also be an allusion in the poem to a Friary of Black Monks, founded about 1280; see note to l. 2521. All these things point to a date as late as 1301.
On the other hand, the Lay cannot be later than 1303, for it is actually quoted at that date by Robert Manning of Brunne, in his poem entitled Handlyng Synne; see notes to ll. 679, 819. And Sir F. Madden has shown that it is expressly alluded to in 1310 by Meistre Rauf de Boun (MS. Harl. 902). This seems to limit the date very closely, to the year 1301 or 1302; and if we were to put the date at ‘about 1300,’ we should expect to be not far wrong. And such, accordingly, is the date usually assigned to it.
At the same time, I much doubt if this date is at all admissible, except as applied to the particular version of the story which has come down to us; for reasons which will be given in § 12.
§ 12. The dialect of the poem. As the poem is full of local interest as regards the city of Lincoln, it is generally agreed that the dialect in which it was originally written was that of Lincolnshire. But to discuss this question is at first sight difficult, owing to the astonishing jumble of dialectal forms which our sole MS. presents. Holthausen’s statement is that ‘the north-eastern Midland dialect of the original has passed through the hands of at least one northern and one southern scribe, who have mixed it up with their respective idioms’; and he adds that ‘the poet seems not to have used a uniform language, but to have inserted, especially in his rimes, the forms of neighbouring dialects.’ On the other hand, Mr. Hall (Introd. to King Horn, p. xlv) thinks {xxvi} that the scribe of the Havelok MS. was ‘a mechanical copyist who made no consistent attempt to substitute his own dialect for that of the original’; and he adds that ‘his own dialect … appears to be East Midland with much resemblance to that of Robert of Brunne.’
I have, however, already shown above that, if the latest scribe did not alter the dialect, he very largely altered the spelling under strong Anglo-French influences; unless, indeed, such alterations had already been made in the older copy which he had before him, a supposition which agrees with Mr. Hall’s suggestion, and is rendered probable by the careless errors noted in § 8. It is likely that our copy is, at least, the fourth in descent from its original; see § 28.
§ 13. If we were to accept the date as being about 1300, and the dialect as that of Lincolnshire, it would follow that the grammar of the Lay and that of the Handlyng Synne must be practically identical. But we are confronted by the obvious fact that they are nothing of the kind, nor could ever have been so. Compare, for example, ll. 1–100 with the ll. 5575-5674 of Handlyng Synne, as given in Specimens of English, ed. Morris and Skeat, pp. 50–53. I find that in ll. 1–100 of the Lay (omitting examples of final -en) there are at least 32 instances in which the scansion of the line is incomplete unless we suppose a final -e to be sounded (as e.g. in l. 10, we must read þurt-ë); and there are at least 66 lines with feminine rhymes, of which all but 10 involve a final -e. But in the 100 lines of Handlyng Synne, I can only find 18 cases (not at the end of a line) where the scansion requires a final -e; and hardly 40 lines with true feminine rhymes, 6 of which involve no final -e. In other words, the Lay has 88 examples in which the final -e constitutes a syllable where Manning has but 521. If we compare {xxvii} another 100 lines, we shall obtain similar results; and even if my calculations be somewhat inexact, the general conclusion is not much affected. The difference in grammatical usage is very clearly marked. I think it follows that, if the two poems were written in the same dialect, the Lay must have been originally written at a considerably earlier date; and that it acquired additions and alterations in the process of transmission from one reciter to another. Compare, for example, the following lines from Havelok and from Manning’s Handlyng Synne, as regards the treatment of final -e:—
1 In 100 lines of The Owl and Nightingale (in Morris, Specimens of English) I find about 150 examples of the final e. But this is a southern poem, and perhaps as early as 1250.
All-e gret-en swiþ-e sor-e; Hav. 236.
But son-e ded-e hir-e fet-e; 316.
Þin-e cherl-es, þin-e hin-e; 620.
Grim-es son-es all-e þre; 1399.
But to þe por’ boþ’ mek’ and kynde; H. S. 5692.
Þat þe por’ man of hym had; 5730.
And þoght’ gret wunder and seþen seyd; 5740.
Unto a cherch-e boþ’ þey yede; 5777.
I think we can only conclude that the extant copy shows the poem in quite a late stage, with just a few interpolations in it to bring it up to date1. The first draft of the poem must surely have been composed earlier than 1300; but how much earlier I am unable to say. That the dialect was, in the first instance, that of Lincolnshire, is consistent with the fact that we can still detect the characteristic suffix -es of the pres. s. indicative as occurring in fifteen unambiguous rhymes (306, 396, 597, 1359, 1443, 1693, 1781, 1851, 1913, 2105, 2323, 2341, 2392, 2573, 2983); and the pl. suffix -e at the end of ll. 1325, 2099, and 2583. The prefix y- occurs {xxviii} before verbs but rarely, as in Handlyng Synne (y-lore, 5788). See § 10.
1 Lines 138, 139, 264, 265, can be omitted without injury to the sense. And l. 1176 has to be emended, in order to make ll. 1177-1180 fit in.