We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 02 December 2020
Farman's Mercian glosses to certain parts of the Rushworth Gospels have occupied the attention of scholars for more than a century. It would be difficult to find an aspect of the priest or his work which has not, at some time, been investigated. Yet there remain numerous controversial questions, some of them of a very fundamental nature. One of the most difficult concerns Farman's apparently irrational uses of the symbols e and æ. It will be found that many of the larger conflicts over the position of Rushworth in the general scheme of English phonology are essentially disagreements as to just what values these two letters had in Farman's orthography.
1 Glosses to Matthew, Mark 1-2: 15, and John 18: 1-3, in MS. Auct. D. ii. 19 of the Bodleian Library, Oxford. Farman's Mercian glosses are often referred to as Rushworth1.
2 For systematic recapitulation of the scholarship on Rushworth1, see E. M. Brown's dissertation, Die Sprache der Rushworth Glossen zum Evangelium Matthäus, i (Göttingen, 1891), 3-10, 79-83; and R. J. Menner's “Farman Vindicatus,” Anglia, lviii (1934), 1-4.
3 The general plan and methods of analysis owe much to Karl Bülbring's “E and Æ in the Vespasian Psalter,” in An English Miscellany. Presented to Dr. Furnivall in honor of his 75th birthday (Oxford, 1901), pp. 34-45. My organization must differ from his, however, since Farman's spelling lacks the comparative regularity and consistency of the Vespasian Psalter. In assembling the evidence, I have made use of Brown's dissertation and Ernst Schulte's Glossar zu Farmans Anteil an der Rushworth-Glosse (Bonn, 1904); but, wherever either of these differs from W. W. Skeat, The Holy Gospels in Anglo-Saxon, Northumbrian, and Old Mercian Versions (Cambridge, 1871-1887), I have followed Skeat.
4 The form el appears only once, Ðet or del 10 times, Ðette or dette 4 times. There are three examples of cwæÐ, beside 45 of cweÐ, and two of Ðæs (genitive singular of se—Ðæt), beside 30 of Ðes or des.
5 Bülbring supposed that the further raising and fronting from [æ] to [æe] in the dialect of the Vespasian Psalter followed immediately after the shift of [a] to [æ], without any “interval of stability”—An English Miscellany, p. 35. For my own views on this point, see PMLA, liv (1939), 11-19.
6 Forms from Mark and John are labeled; all others are from Matthew.
7 Cf. huœÐre, hweÐer, hweÐre, huoeÐer in the Lindisfarne Gospels, Gothic boaþar, Old Saxon hweÐar, Old High German hwedar.
8 I.e. ∗fratwjan rather than ∗fratwōjan.
9 The -ern of compounds like berern and carcern contains a definitely weakened vowel, which appears as e in West Saxon. These forms are treated in section 17.
10 Cf. þæme in Farman's glosses and in late West Saxon.
11 Cf. WS. mæsse, Nthb. mæsse, meassa. See also Karl Luick, Historische Grammatik der englischen Sprache (Leipzig, 1921), p. 194.
12 Cf. the six cweþ- and cweÐ-forms listed above.
13 Repetitions, and a few inflectional or spelling variants have been omitted in this list.
14 I count exactly 806.
15 A Glossary of the Old Northumbrian Gospels (Halle, 1894).
16 Eigentümlichkeiten des anglischen Wortschatzes (Heidelberg, 1906), pp. 16-17.
17 New English Dictionary.
18 Loosely, of course. The proportion in the latter is e:æ: : 1:5.
19 The writer, op. cit.
20 Compare the following words from the Katherine-Group with the words listed in this section: efter, aheven, bed, ber, brec, dei, feder, fedres, feier, hefde, kefden, hwet, hweÐer, lei, mei, meiden, mein, meinful, neiles, islein, steÐelfest, togeines, seide, set, spec, creft, þene, wes, westum, wettres. A much longer list might have been compiled from the published editions of the three legends: Eugen Einenkel (ed.), The Life of Saint Katherine, EETS, OS, lxxx (London, 1884); Frances Mack (ed.), Seinte Marherete þe Meiden ant Martyr, EETS, OS, cxciii (London, 1934); 0. Cockayne and Edmund Brock (eds.), The Life of St. Juliana, EETS, OS, li (London, 1872). The following forms appear in Sir Frederic Madden's edition of Lagamons Brut (London, 1847): iber, iqueÐ, deie, feier, hefde, whet, wheper, lei, meiden, togeines, seide, soÐfeste, spec, crefte, pene (acc. sg. masc.), wes. The following may be found in the entries after 1100 in the Peterborough Chronicle, Benjamin Thorpe (ed.), The Anglo-Saxon Chronicle (London, 1861), I, 369-385: bed (pt. sg.), beiet (pt. sg.), deg, dei, efter, her efter, þerefter, Ðerefter, frett (pt. sg.), hedde, hefde, hefden, messe, messedei, Candelmesse, sei, seide, seidon, soÐfeste, þet, þes (gen. sg.), Ðes, wes (pt. sg.), westme, westmas. Occasional examples of the same sort occur in the Ormulum and other early Midland texts.
21 Die Sprache, i, 14.
22 Æft occurs 17 times, eft 11.
23 “Farman Vindicatus,” p. 16.
24 Angelsächsische Grammatik (2nd ed., Halle, 1886), p. 2; An Old English Grammar. Translated and edited by A. S. Cook (Boston, 1903), p. 3.
25 Altenglisches Elementarbuch (Heidelberg, 1902), p. 10.
26 Angelsaksisch Handboek. (Haarlem, 1931), p. 8. Girvan's views are evidently based in part on H. M. Chadwick, Studies in Old English, Transactions of the Cambridge Philological Society, iv (Cambridge, 1899), 250 ff.
27 Untersuchung der Beziehung der allenglischen Matthäusglosse im Rushworth-Manuskript zu dem lateinischen Text der Handschrift (Bonn, 1903), pp. 22-30.
28 Op. cit.
29 Hist. Gram., pp. 165-166. Luick does, however, admit the possibility of West Saxon influence—pp. 33-34, 129.
30 Rudolf Zeuner, Die Sprache des kentischen Psalters (Halle, 1881), p. 20, cites Ðæs 7, 18 and 74, 11 as examples of æ for WG. e. This is a mistake, for both of the pronouns cited are normal genitive singulars of se. In 7, 18, Ðæs hestan glosses altissimi, and in 74, 11, Ðæs rehtwisan glosses iusti; cf. se hehsta (altissimus) 17, 149 Ðæm hestan (altissimo) 49, 14, etc.
31 There are 105 by my count, 102 according to Schulte's Glossar.
32 The ordinary uses of e for WG. e need not be dwelt upon since they do not differ from the usage of either the Psalter or strict West Saxon. But the following second and third singular forms of Strong iii, iv, and v verbs are worth noting: agefes 5, 24; agefep 17, 11. bereþ 1, 21; 1, 23; 3, 10; 7, 17; 7, 17; 7, 19; 12, 35; 12, 35; 13, 23; bereÐ 13, 52. breceþ 27, 40; tobreceþ 12, 20. cweþest 12, 23; cweþestu 7, 4; cweþ 7, 21; 8, 4; 8, 7; 8, 8; 8, 19; 8, 22; 8, 26; 8, 32; 12, 39; 13, 28; 13, 29; 13, 52; 14, 2; 14, 18; 15, 24; 15, 28; Mk. 2,14; cweÐ 14, 8; cweþaþ (pr. 3 sg.) 12, 32; 12, 32. eteþ 9, 11. geldeþ 6, 4; 6, 6; 6, 18; 16, 27. helpeÐ 16, 26. ongeteÐ 13, 23. spreces 13,10; sprecaþ (pr. 3 sg.) 10, 20; Mk. 2, 7. There are three examples of e substituted for WG. i in other words: smere (imper. sg.) 6, 17; smerennis 2, 11; smerenisse 26, 12; cf. smirenisse, smirede, etc., in the Psalter.
33 Hist. Gram., p. 264.
34 Ae. Elementarbuch, p. 36.
35 Die Sprache, i, 81.
36 “Farman Vindicatus,” p. 16.
37 Handboek, p. 81.
38 I refer especially to the Epinal and Corpus Glossaries, the Vespasian Psalter, and the early Mercian charters.
39 See footnote 20.
40 British Museum MS. Cotton Tiberius B. IV, sometimes called “MS. D.” of the Anglo-Saxon Chronicle.
41 During the tenth century, after the Danelaw had been reconquered by the kings of Wessex, West Saxon became the standard for all England. It was the dialect of the royal officials, who were appointed or sent out from Wessex by West Saxon kings. It was the speech of many of the church dignitaries, most of whom were either Wessex men or the pupils of Wessex men. As a result, most of the tenth- and eleventh-century texts written in Anglian or Kentish territory are mixed with West Saxon, even when they are not written entirely in that dialect. Karl Wildhagen went so far as to refer to “der westsächsischen 
im 10., 11. Jahrhundert”—“Studien zum Psalterium Romanum in England,” Festschrift für Lorenz Morsbach (Halle, 1913), p. 437. H. C. Wyld writes: “The fact is that all O. E. documents of the later period, with very few exceptions, are written in a common form which in all essential features is W. Saxon … so much so that it is now commonly assumed that after Ælfred's time the prestige of Wessex in Government, Arms, and Letters, was such that the dialect of that area became a literary 
in universal use in written documents.”—A History of Modern Colloquial English (Oxford, 1936), p. 49. W. F. Bryan's examination of the Kentish charters reveals that the early ones are mixed with the then-dominant Mercian dialect, the later ones with West Saxon—Studies in the Dialects of the Kentish Charters of the Old English Period (Menasha, 1915).
42 Some other features, common in WS. but not in the Psalter, which appear in Farman's glosses are the following: 1. a for WG. a before nasal—gangan 8, 21, gelamp 13, 53, etc.; 2. ea for WG. a before h—geseah 2, 16, etc.; 3. ea for WG. a before l plus consonant—ealde 9, 17, healfe 20, 21, etc., 4. ea for WG. or Lat. a after a palatal—geatt 7, 13, ceastre 21, 10, etc.; 5. a for WG. a before palatal plus back vowel—dagum 2, 1, nacud 25, 36, etc.; 6. a for WG. a before non-palatal consonant plus back vowel—faran 8, 28, getalu 24, 30, etc.; 7. eo for WG. e before palatal—feoh 10, 9, weorc 5, 16, etc.; 8. e or i for WG. e or i before dental or nasal plus back vowel—metaþ 7, 2, nimap 19, 11, etc.; 9. e or y in seolf—selfe 1, 21, sylf Jn. 18, 1, etc.; 10. y for i-umlaut of breaking-eo—awyrpeþ 12, 27, gewyrÐ 13, 32, etc.; 11. WS. unstable y—sylle 19, 21; 12. 
for WG. ā—cwædon 2, 5, wæron 12, 4, etc.; 13. eā for WG. ā after palatal—asceadeþ 13, 49, sceadenne 10, 35, etc.; 14. ēa for WG. au before palatal—eagan 18, 9, þeah 24, 23, etc.; 15. ēo for WG. eo before palatal—fleoh 2, 13, -seoke 12, 22, etc.; 16. 
for i-umlaut of ēa—gehyrde 22, 7, etc. For further examples in each category, see Brown's dissertation.
43 By the way, if Schulte's hypothesis needed a coup de grâce, I think Farman's æ for WG. e would provide it. From what kind of West Saxon exemplar could Farman have copied these æ's? I at one time considered the possibility that Farman was, not a Mercian copying WS., but a West Saxon copying an old Mercian gloss; of course the evidence in this section would refute that view as thoroughly as it refutes Schulte's.
44 For discussion of the rare exceptions, see Bülbring's article in An English Miscellany, p. 37.
45 For Sievers' explanation of WS. gemæccea, wæccende, weccan, and similar forms, see An Old English Grammar, p. 56.
46 Ængel 34, 6 and menn 36, 37.
47 Ae. Elementarbuch, p. 71.
48 See section 11.
49 An Old English Grammar, pp. 82-83.
50 The g in mængu, mengu, etc. (glossing turba), and mængistu (plurimae) represents an earlier -ig-; if the vowel of this suffix was lost in pronunciation before the time of lengthening, these words might also be included above.
51 Except, of course, when it was smoothed by a following palatal consonant.
52 Perhaps 
swærigaþ 25, 37 should have been included in this list, for it is written elsewhere in Rushworth1 with a: 
25, 44, andswarade 16, 2, etc. The Psalter has a different, though related, verb: ondsweoriu, ondsweorede, etc.
53 Bist. Gram., pp. 165, 211.
54 Ibid., p. 34.
55 Ibid., pp. 331-333.
56 Richard Jordan, Handbuch der mittelenglischen Grammatik (Heidelberg, 1925), p. 79; H. C. Wyld, A Short History of English (New York, 1927), p. 64.
57 Apparently there is no i-umlaut in any of these forms, cf. þearfe 14, 16, beÐearfeþ 15, 5, etc. See also F. Holthausen, Altenglisches etymologisches Wörterbuch (Heidelberg 1932).
58 In fearende, the ea is probably not the direct result of velar-umlaut anyhow, but the result of analogy with other forms of the same verb, e.g. fearu, fearaÐ.
59 If it were otherwise, Farman would hardly have substituted one vowel for another with such frequency in his inflectional endings. Note these substitutions of a, o, and u for e: strong nouns, aldorsacerdos (gen. sg. masc.) 26, 51, deaÐa (dat. sg. masc.) 15, 4, geofu (acc. sg. fem.) 23, 19, heofunas (gen. sg. masc.) 13, 24, mængu (nom. pl. fem.) 4, 25, restedægas (gen. sg. masc.) 28, 1, tintergu (acc. sg. neut.) 25, 46, etc.; present third singulars, awerdaÐ 5, 13, cerraþ 24, 18, cweþaþ 12, 32, cymaþ 23, 35, sellaþ 13, 44, sprecaþ 10, 20, weorþaþ (of weorþan) 24, 21, etc.; present optative plurals, biddan 6, 8, cuman 27, 64, cweoþan 23, 39, etan 6, 25, forstælan 27, 64, flugan (preterite?) 3, 7, geheran 13, 15, etc. Farman also substituted u for a and a for u: weak nouns and adjectives, ærestu (nom. pl. masc.) 20, 10, ældru (nom. pl. masc.) 26, 57, begengu (nom. pl. masc.) 21, 38, eorÐu (gen. sg. fem.) 5, 13, forþmestu (acc. pl. masc.) 23, 6, godu (voc. sg. masc.) 25, 23, næhstu (acc. sg. masc.) 5, 26, etc.; preterite indicative plurals, blewan 7, 25, coman 21, 1, cwoman 17, 42, eodan 21, 9, feallan 15, 30, feollan 17, 6, foeddan 25, 37, etc. For substitution of e and æ for the back vowels, see section 17.
60 These are awerp (imper. sg. of aweorpan) 5, 29, cwern 18, 6, werþe (pr. opt. of weorþan) 24, 20, and werþeÐ (pr. 3 sg.) 9, 16. The last may be compared with geweorþaÐ 21, 21 and forweorÐeÐ in the Psalter, both present third singulars. WG. i broke to io before r plus consonant, io fell together with eo, and then e was substituted for eo to produce werþeÐ. Note also the one substitution of æ for e in gewærþe, section 2.
61 P. J. Cosijn, Altwestsächsische Grammatik (Haag, 1888), p. 18.
62 Palatal-diphthongization of short e to ie is the only important WS. feature which Farman's glosses lack.
63 Cf. also sedle 19, 28; 23, 22; 25, 31; sedlum 19, 28; and the form seld in the Vespasian Psalter.
64 Ae. Elementarbuch, p. 130.
65 Bülbring attributes neht, mehtigra, etc., of the Psalter to i-umlaut. Ibid., p. 76
66 Such sporadic examples of unsmoothed diphthongs before palatal plus back vowel hardly indicate Kentish influence. In addition to the eo's in Rushworth1 and the Psalter, I note hreacan in the latter, heage in the former, onseacan and bordÐeaca in the Corpus Glossary. Girvan apparently considers such sporadic spellings evidence of Kentish or Southeastern origin, however—Handboek, pp. 87-89.
67 Luick looks upon these as characteristic of Farman's sub-dialect of Mercian—Hist. Gram., p. 215.
68 See Wörterbuch.
69 The proportions of e and æ are not so important here as in some other sections, yet they may be of interest. The proportions are approximately æ:e: :7:1 for smoothing of ea, and æ : e: : 1:7 for smoothing of eo.
70 Ae. Elementarbuch, pp. 85-86.
71 For explanation of the latter, see Jordan's Eigentümlichkeiten, pp. 51-53, or Holthausen's Wörterbuch.
72 All repetitions and some inflectional variants are omitted.
73 Glossar.
74 Op. cit.
75 These forms are doubtful, but probably contain WG. ā. They appear to be related to the third principal part of geþweran “to mix, stir,” Strong iv. See Fr. Klaeber's Beowulf (Boston, 1936), p. 343.
76 Hist. Gram., p. 129.
77 See section 5.
78 “Farman Vindicatus,” pp. 15-16.
79 Perhaps I am misrepresenting Menner's view. Sometimes he seems to use the term dialect in the sense of the speech of a tolerably homogeneous group of people who share certain special speech habits differing from those of other groups. At other times he broadens the meaning of the term until it includes all the personal speech habits of an individual as well, e.g. “The whole history of modern dialectology, with its doctrine that every individual may, in a sense, have a dialect of his own, points to the possibility that Farman's variations in Matthew may be natural to himself.”—op. cit., p. 18. If this is what Menner has in mind, there is nothing to be gained by arguing about definitions. We can agree that Farman's inconsistent spellings are variants of his own “dialect,” in the sense in which that word is sometimes used by modern dialectologists, but of course Farman's “dialect” was a mixture of the Mercian dialect (in the older sense of the term) and the West Saxon and Northumbrian dialects.
80 History of English Sounds (Oxford, 1888), p. 126.
81 Die Sprache, i, 80.
82 Sievers, An Old English Grammar, p. 38; Bülbring, Ae. Elementarbuch, pp. 52-53; Luick, Hist. Gram., p. 155; Girvan, Handboek, pp. 49-50.
83 Examples are very few: monÐuaeran and a few others in the Vespasian Psalter, naep from Latin napus in the Corpus Glossary, sporadic examples in other Mercian texts. Examples from the Life of St. Chad, the Tanner MS. of Bede, and the Worcester Chronicle prove nothing of course, for these texts are known to contain a strong West Saxon element.
84 Ae. Elementarbuch, p. 38.
85 The best discussion of the boundaries of Mercia is probably in J. Brownbill's article “The Tribal Hidage,” English Historical Review, xxvii (1912), 625-648.
86 Skeat, Holy Gospels, iv, xii-xiii.
87 The influence of Nthb. was not merely negative. The list of features in Farman's glosses which are common in Nthb. but not in the Psalter is rather long, although some of them are represented by only a few examples: 1. a for WG. a followed by r plus consonant—arþu 11, 3, forwarp Mk. 1, 39, etc.; 2. eo for WG. a before r plus consonant—eorfeþe 7, 14; 3. ea for WG. e before r plus consonant—bearma 16, 11, wearþe (pr. opt.) 5, 30, etc.; 4. ea for WG. a after palatal—geatt 7, 13, ceastre 21, 10, etc.; 5. ā for contraction of a plus a—ofslan 21, 38, thuaÐ 15, 2, etc.; 6. ī for Anglian ē after palatal—scipa (oves) 18, 12, etc.; 7. ēo for WG. au—bebeod (pt. sg.) Mk. 1, 25; 8. ēa for OE. ēo—eade 24,38, hread 12, 20, etc.; 9. ei-diphthongs—geceigde Mk. 1, 20, Ðreiga 16, 22, etc.; 10. loss of final inflectional n—lichoma (gen. sg.) 6, 22, gehera (infin.) 13, 17, etc. In addition, Nthb. is like WS., in that it has æ for the isolative development of WG. a and seldom shows velar-umlaut of WG. a.
88 The heht-forms may have had short vowel.
89 I do not include certain forms. The comparative adverb mae or mæ (cf. mae in the Psalter) occurs ten times. In these, WG. ai has not become 
æ, but the adverbial ending e has been added to the adjective ma. The form wé 23, 15 glosses Latin væ; it is a scribal error, influenced no doubt by the æ of the Latin, for Farman wrote wa everywhere else. Perhaps lære (dat. sg. of lar) Mk. 1, 22 has æ by analogy with lærende in the same verse.
90 Corpus contains gelestunne, scultheta, stictenel, and uuegiÐ.
91 Hist. Gram., p. 169.
92 The numerals twegen, or twægen, and begen are obscure but probably belong here. According to Girvan, Handboek, pp. 41-42, WG. ∗twaii received an analogical ō from the neuter ∗twō, the ō umlauted to 
, and then unrounded to ē. The Psalter has one example each of twoega and twegen. Farman has sixteen examples of ē: twegen 4, 18; 4, 21; 8, 28; 10, 29; 14, 17; 18, 16; 18, 16; 18, 19; 24, 40; 24, 41; 25, 15; 25, 22; twege 5, 41; 18, 20; twegra 21, 31; 27, 21; and sixteen in which æ has been substituted for ē: twægen 11, 2; 14, 19; 19, 5; 19, 6; 20, 21; 20, 30; 21, 1; 21, 28; 25, 17; 25, 17; 25, 22; 25, 22; 26, 37; 26, 60; 27, 38; 27, 51. There is a remote possibility that the æ-forms are directly descended from ∗twaii and contain i-umlaut of WG. ai, but Farman's tendency to use æ for ē makes the other alternative seem more probable. Begen 13, 30 and 15, 14 may be explained in the same manner as twegen—Girvan, op. cit., p. 267.
93 Not to be confused with 
meaning “renown,” which is related to blāwan. This 
, or rather blēd, is related to blōwan and contains i-umlaut of ō. See J. R. Hulbert, Bright's Anglo-Saxon Reader (New York, 1936), p. 258.
94 Schulte associates this form with 
“reach.” That is an error, for the verb is clearly rēcan “heed, care for, i.e. rule.”
95 Hwæne glosses pusillum, cf. modern Scottish wheen. Schulte, in his Glossar, mistakenly treats it as a form of hwa.
96 The two sound changes were very similar although one took place considerably later than the other. In each case, a following palatal sound caused the velar second element of the diphthong to disappear.
97 H in the superlatives is analogical, cf. nesta in the Psalter. The h of the positive neh found its way into nehsta, etc., but apparently the analogy did not affect the comparative narra.
98 In late WS., the ēa was frequently reduced to ē by palatal-umlaut. This ē also appears, through analogy, in late WS. superlatives, e.g. nehsta, etc., in the West Saxon Gospels.
99 For Bülbring's explanation of this word as it appears in the Psalter, see An English Miscellany, pp. 43-45.
100 Gesænæ and gesene correspond to WS. gesiene, which is usually classified as an adjective. Girvan, Handboek, p. 296, states that the word is used for the past participle of seon in the Lindisfarne Gospels and some other northern texts. At any rate, the connection with geseon is clear.
101 False gemination is extremely common in tenth-century Anglian texts.
102 Ae. Elementarbuch, p. 41.
103 Ibid., p. 42.
104 Brown, op. cit., Part II—The Language of the Rushworth Gloss (Göttingen, 1892), pp. 3-8.
105 For his substitution of back vowels for e, see footnote 59.
106 Cf. treatment of the i-umlaut of WG. ai in both Rushworth1 and the Psalter.
107 See section 3.
108 For explanation of the effect of these consonants, see Bülbring, Ae. Elementarbuch, p. 69; Luick, Hist. Gram., p. 172.
109 The monophthongization of breaking-ea may have begun in Farman's time, cf. þærf, etc., in section 5.
110 For other non-Mercian usages, see footnotes 42 and 87.
111 One would expect to find local variations in speech within an area so large as Mercia' and I would not deny for an instant the possibility that such variations existed. But they will be found, if at all, through study of the minor features of the extant texts, rather than in the broad, general features with which Luick has dealt.
112 Die Sprache, p. 81.
113 Ibid.
