¹ Bülbring, Karl, Allenglisches Elementarbuch (Heidelberg, 1902), p. 8; H. M. Chadwick, Studies in Old English, Transactions of the Cambridge Philological Society, iv (Cambridge, 1899), pp. 249, 252–253; R. Girvan, Angelsaksisch Handboek (Haarlem, 1931), p. 7; Karl Luick, Historische Grammatik der englischen Sprache (Leipzig, 1921), p. 34; Henry Sweet, The Oldest English Texts, E.E.T.S., O.S. lxxxiii (London, 1885), pp. vii, 5; H. C. Wyld, A Short History of English (New York, 1927), p. 55.Google Scholar

² Keller, Wolfgang, Angelsächsische Palaeographie, Palaestra, xliii (Berlin, 1906), i, 18; J. H. Hessels, An Eighth-Century Latin-Anglo-Saxon Glossary (Cambridge, 1890), pp. ix-x; Sweet, p. 5.Google Scholar

³ Sweet, p. S; Eduard Sievers, An Old English Grammar. Translated and edited by A. S. Cook. (Boston, 1899), p. 244; Chadwick, p. 253; Bülbring, p. 8; Luick, p. 34.

⁴ For consonants, see Bülbring, p. 190; Chadwick, pp. 232–246; Girvan, pp. 13–14. For unaccented vowels, see Ferdinand Dieter, lieber Sprache und Mundart der ältesten englischen Denkmäler (Göttingen, 1891), i, 45–46; Bülbring, p. 146; Chadwick, p. 246; Girvan, pp. 71, 153–154; Luick, pp. 278, 297–304. For accented vowels, see Bülbring, pp. 42–43, 49–50, 71; Dieter, i, 9–10, 14, 42–43; Luick, pp. 122–123, 171, 216.

⁵ Chadwick, pp. 252–3; Bülbring, pp. 35, 63, 96, etc.; Luick, pp. 33–34, 165–166.

⁶ The term Mercian is loosely used by Luick to include not only Mercia but East Anglia and Essex. Generally, however, the term is applied to Mercia proper.

⁷ Lindsay, W. M., The Corpus, Epinal, Erfurt and Leiden Glossaries, Publications of the Philological Society, VIII (Oxford, 1912), pp. 1–5; Chadwick, pp. 189, 210.Google Scholar

⁸ For example, many of the items in Corpus are so similar to items in the Epinal and Erfurt Glossaries that they must have been taken from a common source, a lost English-Latin glossary which Chadwick calls “Archetype I.” That the compiler of Corpus was altering his materials is evidenced by such things as his elimination of intersonantal h in most of the items which he drew from Archetype I. Such correspondences as the following are numerous: Epinal Villosa—ryhae, Erfurt Villosa—ryhae, Corpus Villosa—rye; Ep. Villis—uulohum, Erf. Villis—uulohum, Cp. Villis—uuloum. See Chadwick, pp. 229–232.

⁹ The Vespasian Psalter has e; and the use of œ for West Germanic e in such words as rœgn, stœfn, pœgn, wœg, wœl (bene) in the Mercian portion of the Rushworth Glosses is hard to explain except on the assumption that WG a had fallen together with WG e.

¹⁰ The effects of a-o and of u are so nearly identical in Corpus that it does not seem worth while to distinguish here between a-o-umlaut and u-umlaut.

¹¹ Figures before the comma refer to the lines in Oldest English Texts, pp. 35–107; those after the comma refer to the edition by Hessels. All evidence has been carefully checked by W. M. Lindsay, The Corpus Glossary (Cambridge, 1921). The numbering in Lindsay is the same as that in Hessels.

¹² Possibly bearug 1284, M38 and pearuc 486, C488 belong here. But there is only one clear example of ea before r plus palatal, and both bearug and pearuc admit of different explanations. The latter is ordinarily traced to a West Germanic *parruk, whose a would become ea by breaking. There is one other possibility. I think it would not be unreasonable to assume that the scribe wrote both words as he pronounced them, and that the svarabhakti u was actually pronounced. With this in mind, I have listed the words as examples of velar-umlaut. The word healecas (hēa lēcas) 218, A773 contains ēa by contraction after loss of intervocalic h. The Psalter has hea, hean, in the oblique cases and in the weak forms, beside heh in the strong-nominative-singular-masculine. The form streagl A932 is printed by Sweet stregl 249. The manuscript has streal with g over the a. Evidently the scribe intended to substitute the g for the a but forgot to dot out the latter.

¹³ The form gierende 1986, T41 is not i-umlaut, but is probably from *gi-ērendian, cf. WS ***œtrendian, rather than from gerwan—gierwan. Theform ***iendi 1118, I260 is from pian (WG *pīhan), cf. similar contracted participles in the Psalter, e.g. gesiende from infinitive gesian or gesean.

¹⁴ Possibly some of the following should have been included: blaesbaelg 910, F305, glaeres 1958, S688, naep 1363, N40, spr$eLc (noun) 1852, S299, spraec 1769, S43, suu$eLr 633, D26. The first was evidently unfamiliar to the scribe or very corrupt in his sources, cf. bloestbaelg 28, Int. 308. In glaeres the vowel, according to NED, may have been short. In spr$eLc the vowel was probably short and probably related to the first or second principal part of sprecan rather than the third, cf. Psalter gespreocu (nominative-accusative plural). Another *** which is extremely doubtful appears in naep from Latin nāpus, for p plus back vowel altered the development of â, as in WS släpan. It is true that the Psalter has slipan, but this, like WS sl***pan, is analogical. Similarly, suu$eLr is analogical like WS swār, etc.)— See Bülbring, p. 53; Luick, p. 155.

¹⁵ Bryan, W. F., Studies in the Dialects of the Kentish Charters of the Old English Period (Menasha, 1915), pp. 28–29; Chadwick, pp. 182–183.Google Scholar

¹⁶ According to Sievers, pp. 114–115, early vocalization of g to i at the end of a syllable was a Kentish feature. The following examples appear in Corpus : romei 360 (Sweet, p. 667), C89, iserngrei 865, F153, grei 967, G91, grei 981, G117, omei 866, F154, popei 1516, P166, and perhaps streide 1910, S525, although Luick, pp. 219–220, explains the last as a contracted form. Chadwick, pp. 173–174, questions Sievers' opinion, presenting evidence of the change of endosyllabic g to i from the Northumbrian Liber Vitae and the Saxon Chronicles. The Psalter, also, contains guiu***u 24:7 and numerous examples of i for the suffix ig: hefie (for hefige Adj. Nom. Plu. Masc.) 34:13; 54:4; hwefie (for hefige Adv.) 4:3; syndrie 6:7; 7:12, etc.; weolie 21:30, etc. In view of the fact that most of Sievers' evidence was taken from Corpus and Epinal, texts which are no longer considered Kentish, and from ninth and tenth century Kentish texts, we may safely assume that Chadwick is correct. Bülbring, p. 204, seems to agree with Sievers, but he repeats Sievers' evidence with little alteration. Girvan, p. 202, has apparently discarded Sievers' view, for he says, “in het vr. kent. en in naburige dialecten werd de g aan het einde van een woord of letter greep gevo-caliseerd“. . . (Italics are mine.)

¹⁷ The double l in scell “shell” is probably the result of gemination, cf. Gothic skalja “tile.” Two words, ellaern 1775, S55 and elm 2149, U237, have e, but here the e is common Old English and lacks dialectal significance.

¹⁸ Scultheta may be corrupt—See Dieter, i, 22. Werna 301, B136 should perhaps be added to the list, if it is wrœnna with metathesis.

¹⁹ Zeuner, Rudolf, Die Sprache des kentischen Psalters (Leipzig, 1881), i, 15–16, 41; Luick, pp. 172, 175; E. M. Brown, Die Sprache der Rushworth Glossen (Göttingen, 1891), i, 29, 69–70.Google Scholar

²⁰ Sweet, pp. 148–149.

²¹ The forms barriggae and gewarht are very hard to account for. The former is evidently corrupt and may not even be English. The latter would be as difficult to explain in a Northumbrian text as in a Mercian.

²² Sondgewearp may contain Germanic a before r plus consonant, cf. Icelandic moldvarpa. The influence of the related verb forms, weorpan—wearp, must also be reckoned with in connection with this word.

²³ The Psalter has examples of ea for eo, eo for ea, ēa for ēo, and ēo for ēa—Zeuner, i, 22–23, 48–50. Both the Psalter and the glossary show a tendency to confuse the diphthongs resulting from contraction.

²⁴ The Psalter has geceigo 55:10.

²⁵ The evidence would be stronger, of course, if we included the use of œ instead of e, a instead of o, and the other features of Corpus noted in paragraph one of this section. These phenomena are not admissible as evidence because, as we shall see later, all of them have been explained or can be explained satisfactorily as owing to the early date at which Corpus was compiled.

²⁶ Luick, p. 217, seems to hold this view.

²⁷ In the Mercian portion of the Rushworth Glosses, there are numerous examples of both of these phenomena. These have sometimes been attributed to West Saxon influence.

²⁸ It seemed advisable to omit from consideration the consonants and also the vowels of the unaccented syllables. Such characteristics of Corpus as the frequent use of b instead of intervocalic f and the frequent use of œ instead of e in the inflectional endings are commonly accepted as evidence of the early date of the text. No one has suggested, as far as I know, that they indicate any difference between the dialect of Corpus and that of the Psalter.

²⁹ Studies, pp. 189, 203, 250.

³⁰ Bülbring, pp. 49–50; Luick, pp. 122–123; Girvan, p. 60.

³¹ Chadwick, pp. 204–206, 250.

³² Sweet and Lindsay print doppaenid; Hesseh has doppa. enid in the text but dopp-aenid in the index. For maenoe, see Sweet, p. 667. Wodhae is corrupt. It glosses coturno, which Sweet connects with coturnix, Lindsay with cothurnus. Sweet considers the word a corruption of wuduhenn, or wuduhœn, cf. Epinal edischaen 714. Lindsay, however, thinks wodhae is the instrumental singular of w***. It is true that cothurnus may have a figurative meaning of “eloquence,” or something of that sort, but Lindsay's conjecture seems far-fetched. Is wōdhenn a possiblity?

³³ Bülbring, p. 71; Luick, pp. 170–171; Girvan, p. 73.

³⁴ Bülbring, p. 81; Luick, p. 216; Girvan, pp. 84–85, apparently agrees.

³⁵ Ibid.

³⁶ Studies, pp. 96–98, 117.

³⁷ Luick, p. 130; Girvan, p. 49.

³⁸ For further examples of ēo for īo and īo for ēo in Corpus, see Sievers' article in P.B.B., xviii (1894), 415–416.

³⁹ Bülbring, pp. 43, 59; Luick, pp. 135, 236.

⁴⁰ A fairly complete list of the diphthongs from velar-umlaut of e and i will be found in Sievers' article, P.B.B., xviii, 415 ff.

⁴¹ Luick, pp. 34, 266; Bülbring, p. 92.

⁴² Angelsaksisch Handboek, pp. 92–93.

⁴³ Numbers refer to lines in Sweet's edition, Oldest English Texts, pp. 36–106.

⁴⁴ Studies, p. 226. Notice that Chadwick's explanation would place velar-umlaut in this class of words earlier than i-umlaut, since the -ende comes from *-andjaz, the e having resulted from i-umlaut of Q.

⁴⁵ All examples of a, œ and ea before c and g have been omitted. Examples of a before l have not been included because there are a number of examples of a before l plus back vowel in the Psalter—See Zeuner, i, 9. Naturally, a before w is not included—see Bülbring, p. 52; Luick, p. 145. The words bearug and pearuc are doubtful examples—See footnote 12.

⁴⁶ Hist. Gram., pp. 165, 211.

⁴⁷ Ibid., p. 266.

⁴⁸ There are a number of œ's before c or g plus back vowel. And -laeppan 873, F169 may contain an æ which has been raised and fronted but not velar-umlauted, cf. Psalter leappan 132:2.

⁴⁹ The smoothed vowels which appear before c and g form a group of exceptions. These two consonants were not velarized by following back vowels. Hence the [a^œ] did not undergo umlaut but developed into œ, as in dœgas, etc.

⁵⁰ Some of these examples are doubtful. Sweet prints gebellicum as though it were Latin. Chadwick, p. 195, considers it Old English. Hessels calls it Latin in his edition of the Corpus Glossary, but he has evidently changed his mind on that point—See A Late Eighth-Century Latin-Anglo-Saxon Glossary (Cambridge, 1906), p. 228. Lindsay, in his edition of Corpus, calls gebellicum Latin, with a question mark. Authority is heavy on both sides, but common sense seems to be on the side of calling the word Old English. Fiscalis reda—gebellicum (from *gafel-lïc?) wœgnfearu is a more reasonable reading of the item than fiscalis—reda gebellicum—wœgnfearu. Sweet considers the vowel of gebrec long, but see N.E.D., article “Brake v⁶.” Old Icelandic snigill, Old High German snegil might lead us to suppose that Corpus snegl contains WG e. Perhaps it does, but West Saxon snœgel, early Middle English snail, snayl, and modern snail point to a Prim. OE. *snagl. The word tebleth belongs here only if it is a relatively late formation based upon tebl or tœfi (from Late Latin *tab'la—Latin tabula).

⁵¹ Hist. Gram., pp. 165–166.

⁵² Brown, i, 31–32.

⁵³ Altenglisches Elementarbuch, p. 36.

⁵⁴ Luick, p. 266.

⁵⁵ Angelsaksisch Handboek, p. 83.

⁵⁶ Pp. 427–433. The importance of the scarcity of the e in these charters is somewhat decreased, of course, by the scarcity of words in which the e could appear.

⁵⁷ Studies, p. 183.

⁵⁸ Hist. Gram., p. 176.

⁵⁹ Wylde, p. 65.

⁶⁰ The Psalter has festenne 34:13, etc., efestgan Hymn 12:13, efestigne Hy. 11:7.

⁶¹ Hist. Gram., p. 166.

⁶² Ibid., p. 184.

⁶³ However, weagas (for woegas), which appears three times in the Psalter, indicates that ea before g plus back vowel was not unknown to the glossator.

⁶⁴ Here also probably fraecu*** 695, D331.

⁶⁵ Hist. Gram., p. 208.