Chapter 6

[Skip to Content]


General Navigation

This chapter

[Beginning of Content]


§ 6.1    Witnesses to Cædmon’s Hymn can be divided into two main dialectal groups: Northumbrian and West-Saxon. Each of the poem’s five recensions belongs to one or the other of these two main dialectal divisions, although, as Dobbie 1937 demonstrated, the distinction is textually insignificant: key recensional readings cut across dialectal lines, and copies of the Hymn showing similar dialectal features need not show any other evidence of close textual affinity. This is particularly true of the two Northumbrian recensions: as demonstrated above (§§ 5.7-5.9), the Northumbrian eordu recension is closer textually to the West-Saxon eorðe and West-Saxon ylda recensions than it is to the Northumbrian aelda. The dialectal similarity among the three West-Saxon recensions is similarly incidental: while two of the three appear to be descended from the same broad tradition, various minor and major differences suggest that each was translated independently from the Northumbrian (see further, §§ 6.52-6.54).

§ 6.2    The following sections discuss aspects of the language of the Hymn as represented in the surviving witnesses. The emphasis throughout is on features that are significant for the determination of dialect or date. As in Chapter 4: Manuscripts, above, references in this section are intended to be current and useful rather than historically exhaustive. More detailed studies of the language of the Hymn exist. The West-Saxon eorðan recension is often discussed in general terms in studies of the dialect of the Old English Historia ecclesiastica (see Grant 1989 and Whitelock 1962 for bibliography). The aelda and eordu recensions are important sources for our knowledge of Northumbrian and, as a result, are frequently cited in dialectal studies and the standard grammars (e.g. Ross 1937; Campbell 1991; Hogg 1992; Bourns 2001). Witnesses to these recensions are also the subject of several self-standing studies and monographs, many of which deal in whole or in part with linguistic matters (e.g. Cavill 2000; O’Donnell 1996b; Wuest 1906). Readers interested in a more detailed account of the language of the Hymn are referred to these sources.

Northumbrian features

§ 6.3    Five witnesses to the Hymn, belonging to two distinct recensions, show Northumbrian features. Two of the witnesses, P and M, are found in early manuscripts of Bede’s Latin Historia ecclesiastica and can be dated with relative certainty on palaeographic and internal grounds to the eighth century (see below, Note A: The dating of M and P).[1] Both contain copies of the Northumbrian aelda recension. The other three witnesses, P1 Br Di, contain copies of the Northumbrian eordu recension. These manuscripts are in continental hands ranging in date from s. xiiex through s. xvex. As O’Donnell 1996b demonstrated, all three derive from a single insular exemplar. While none represents the work of a scribe familiar with Old English, their evidence can be used collectively to help isolate readings from the original manuscript with which this version of the poem travelled to the continent.

§ 6.4    In the material that follows, M and P are cited independently as appropriate. As the forms in Di, P1, and Br are of no independent linguistic value, readings from the Northumbrian eordu recension are usually cited (under the siglum Eordu) from the reconstructed text rather than from individual witnesses.[2]


§ 6.5    The Northumbrian witnesses have been copied, with some early or regional features, using the same Latin-based alphabet used in most surviving Old English manuscripts.

Native modifications of the Latin alphabet

§ 6.6    The native characters <ð>, <þ>, and <ƿ> are not used in M or P. M uses <th> initially and <d> medially for dental fricatives (i.e. /θ/ and /ð/): e.g. tha, 7a; modgidanc, 2b. In P these are represented by <th> in all positions: e.g. tha, 7a; modgithanc, 2b. Both manuscripts use <u> for the back approximant /w/: e.g. M P uard, 1b, and M P uuldurfadur, 3a. These are early features in insular manuscripts (Hogg 1992, §§ 2.59, n. 1, and 2.77).

§ 6.7    Eordu used <d> (medially) and <ð> (probably initially and perhaps medially) for /θ/ and /ð/.[3] It did not use <þ>. Examples include eordu, 5b, and perhaps modgedanc (P1 Br ıııodgedaııc Di modgedeanc, 2b[4]) vs. ða (P1 Br da̽ Di da), 7a. /w/ is represented initially by <ƿ> and <ƿu> and medially by <u>: ƿueard (P1 Br Di pueard), 1b, ƿuldurfadur (Di puldurfudur Br P1 fadur), 3a, and gihuaes (Di P1 gıhuaes Br gihnaes), 3b. <ƿ> is invariably misinterpreted as <p> in the surviving witnesses to this recension.

§ 6.8    The use of <d> for insular <d> and <ð> is common in continental manuscripts and may represent a misunderstanding of the Anglo-Saxon character. The use of <ƿu> for /w/ is unusual although it is found occasionally in the Lindisfarne Gospels (Hogg 1992, § 2.77, n. 3; Cavill 2000, 521-522).


§ 6.9    The low front unrounded vowel /æ/ is spelled <ae> and <æ> in M, <æ> only in P, and probably <ae> only in Eordu. Examples include M hefaenricaes, metudæs, æfter; P hefenricæs, metudæs, æfter; Eordu hefunricaes, metudaes, aefter, 1b, 2a, and 8b. The witnesses to the eordu recension show one example of difficulty with this character: Di drıntınc ef|ter, 8a-b, where the faulty word division presumably goes back to the misinterpretation of some combination of <a> and <e> in aefter; P1 and Br, however, correctly read <ae> in the same word (Hr, a s. xiex/s. xiiin witness to the West-Saxon eorðe recension, shows evidence of a similar mistake in the same line: drıhtent efter|).


§ 6.10    The voiceless velar fricative /χ/ is written <h> initially in all Northumbrian versions of the text but represented before /t/ by <c> in M, by <c> and <h> in P, and by <ch> in Eordu: M maecti, dryctin; P mehti, dryctin; Eordu mechti, drichtin, 2a and 4a. <ch> is found in Northumbrian manuscripts through the ninth century (Hogg 1992, § 2.60, n. 1; Cavill 2000, 521).

Germanic *[β]

§ 6.11    The Germanic bilabial voiced fricative *[β] (later Old English [v], an allophone of /f/) is represented as <f> in all copies of the Northumbrian texts of the Hymn: e.g. M hrofe P hrofæ Eordu hrofe, 6a. In M the form also appears once as <b>: M heben (vs. P hefen Eordu hefen and the first element in M hefaenricaes P hefenricæs Eordu hefunricaes, 1b), 6a. Sporadic spellings with <b> are found in other early Northumbrian texts, especially the Leiden Riddle (e.g. ob, “from,” twice against ofaer, “against”) and in proper names in Bede (see Anderson 1941, 83-84; Ström 1939). It is especially characteristic of the Mercian glossaries and is also found, probably as a deliberate archaism, in Kentish charters of s. ixmed (Hogg 1992, § 2.54).


§ 6.12    There is one example of the use of double letters to indicate length: Eordu scoop (vs. M P scop), 5a. This is usually described as an early feature (Hogg 1992, § 2.4), though it is frequent in the mid- to late- tenth-century Lindisfarne Gospels (Cavill 2000, 520). There is one example of a similar use of double letters in a West-Saxon manuscript, O oór\d/, 4b.

Analogic and historical forms

§ 6.13    A number of unusual forms in individual witnesses to the Northumbrian recensions have their origins in analogic spellings or variation between alternate historical forms.

end : and

§ 6.14    M reads end for P Eordu and, 2b. The M reading is derived from Germanic *andi (> *ændi > *endi); the reading in the other witnesses is from Germanic *anda (Hogg 1992, § 5.4, n. 2).


§ 6.15    All witnesses to the eordu and aelda recensions read (-)fadur for more common Old English (-)fæder in 3a. Smith traces the form to Germanic *faðurz (cf. Old Norse foͅður) beside *faðraz (Smith 1978, 27, § 3). Campbell adds that the variant forms are derived from different Indo-European genitives: cf. Latin patris with Sanskrit pitur (Campbell 1991, § 630; Bourns 2001, 22-23).


§ 6.16    M P astelidæ, 4b, shows analogic insertion of /i/ between the root and dental of the past participle in a weak class I verb (Campbell 1991, § 753.9; cf. Bourns 2001, 33-34) with resulting first fronting (from */a/ to */æ/) and i-mutation (from */æ/ to /e/) of the root vowel. The Eordu form, astalde, shows the more common spelling without the intermediate /i/ and related changes. See also below, §§ 6.23-6.25.


§ 6.17    Etymological /u/ is preserved in Eordu hefunricaes, 1b. In M hefaenricaes P hefenricæs and M heben P hefen Eordu hefen, 6a, <ae> (confused with <e> in most cases) appears by analogy from oblique forms in *-/æn/- in prehistoric Old English (Hogg 1992, § 6.2, n. 3).

halig / haleg

§ 6.18    The derivational suffix -ig is spelled -<eg> and -<ig> in M, -<ig> only in P and Eordu. The failure of i-umlaut in the root syllable of all forms in all manuscripts suggests that both spellings are derived from prehistoric Old English */æj/ (cf. the common Old English spellings mihtig, syndrig, and, less common, mænig/menig where umlauted forms suggest a prehistoric suffix */ij/. See Hogg 1992, § 5.85.10.b and n. 8). M haleg, 6a, preserves the historical spelling of the suffix after first fronting (Hogg 1992, §§ 6.48, 6.51). The endings in P Eordu halig, 6b, and M allmectig P allmehtig Eordu allmechtig, 9b, all show the effect of fronting of æ to e and raising to i (Hogg 1992, §§ 6.2, 6.48, and 6.51-52).

swe / swæ

§ 6.19    Old English swe, swæ, swa appears as sue in M P, suae in Eordu (3b). The M P form is common in early Northumbrian (Hogg 1992, § 3.25, n. 3) and shows the expected development of West Germanic */aː/ (*[æː]) in Anglian. The Eordu form is the usual spelling in later Anglian (Hogg 1992, § 3.25, n. 3). The origins of the later form, and its West-Saxon equivalent swa, are obscure (see also Bourns 2001, 24-26).


§ 6.20    The following section traces features of M, P, and the reconstructed insular exemplar for Di, Br, and P1 that are most commonly associated with the northern (Northumbrian and more broadly Anglian) dialects.

Stressed syllables

Germanic */a/
Nasalisation and rounding

§ 6.21    Germanic */a/ is written <a> and <o> before nasal consonants and clusters. In later sources, there is a “clear-cut division between S[outhern, i.e. Kentish and West-Saxon] texts, where <a> predominates, and the Angl[ian] texts, where <o> predominates” (Hogg 1992, § 5.5); in most earlier Anglian texts, including the early Northumbrian poems, however, <a> and <o> are distributed more evenly.

§ 6.22    The Northumbrian recensions of Cædmon’s Hymn show both <a> and <o> for Germanic */a/ followed by a nasal. Spellings with <a> are found in the second element of M modgidanc P modgithanc Eordu modgedanc, 2b, and P Eordu and, 2b (M end, 2b, is derived from a different Germanic form; see above § 6.14). <o> is found in the first element of M P moncynnæs Eordu moncinnes, 7b.

First fronting, breaking, and retraction

§ 6.23    When not found before a nasal consonant or cluster, Germanic */a/ develops regularly into prehistoric Old English */æ/ (“First Fronting”; see Hogg 1992, § 5.10). In Anglian dialects, however, this sound change regularly fails to produce the expected developments before /w/, velarised /l/ (i.e. /l/ between back vowels or followed by another consonant), and, in Northumbrian, before clusters beginning in /r/ (see Hogg 1992, § 5.10, 5.13, and 5.29).[5] In non-Anglian dialects, the newly fronted sound participated in breaking, a process by which long and short front vowels were diphthongised through the addition of a back vowel, in all three of these contexts (Hogg 1992, § 5.16); in Northumbrian and (before /w/ and velarised /l/) in Anglian more generally, the first fronting either fails to occur or is “retracted” back to /a/ (Hogg 1992, §§ 5.10-5.13, and 5.29).

§ 6.24    M and P almost invariably show failure or retraction of first fronting in the expected contexts: e.g. M P uard, 1b and 7b; M P barnum, 5b; and the second element of P middingard, 7a. The sole exception is the second element of M middungeard, 7a, which shows breaking of fronted */a/ (<ea>). M P aelda, 5b, also shows retraction although the results are obscured by the effect of i-mutation (Smith 1978, 30, §§ 8.d and 9.a.iv). In M P astelidæ, 4b, the presence of an analogic /i/ before the past participle ending allowed fronting of the medial syllable (with subsequent i-mutation to /e/) to occur (see above, § 6.16).

§ 6.25    The evidence of Eordu is mixed. The only example of Germanic */a/ before velarised /l/, astalde, 4b, shows retraction or failure of first fronting. On the other hand, Prehistoric Old English */æ/ (i.e. the product of first fronting of Germanic */a/) is always broken before /rC/: ƿueard, 1b, ƿeard, 7b, bearnum, 5b, middungeard, 7a. Hogg notes that retraction/failure of first fronting is incomplete in all periods in the Anglian dialects, and that “there are many forms with diphthongization rather than retraction” (Hogg 1992 § 5.29).

Palatal diphthongisation

§ 6.26    Stressed vowels preceded by palatal consonants are frequently diphthongised in Old English. The change affects front vowels inherited from Germanic, vowels fronted by the operation of i-mutation, and back vowels. The operation of this sound change varies greatly by dialect, period, and context (Hogg 1992, § 5.48).

§ 6.27    The relevant forms in Cædmon’s Hymn are scylun and sceppend, 1a and 6b. Palatal diphthongisation fails in all Northumbrian witnesses in the latter of these two cases: P Eordu sceppend M scepen (see Hogg 1992, § 5.54). M scylun P scilun and Eordu sciulun, 1a, on the other hand, all show the effect of palatal diphthongisation of */u/ to */ju/ (Eordu sciulun) with subsequent monophthongisation to /y/ (M scylun), or /i/ (P scilun; see Hogg 1992, § 5.68; cf. Campbell 1991, § 176; Brunner and Sievers 1965, § 92 Anm. 1).

Anglian smoothing

§ 6.28    Long and short Old English front diphthongs are generally monophthongised (“smoothed”) in Anglian dialects when they occur before a velar consonant or a liquid and velar consonant cluster (Hogg 1992, § 5.93). The correspondences are as follows: /ĭu, iu/ to /i, iː/, /ĕo, eo/ to /e, eː/, /æ̆ɑ/ to /æ/ (later raised to /e/ before liquids), /æɑ/ to /eː/ (or /æː/). The change appears to have begun at or immediately before the time of the earliest texts.

§ 6.29    Smoothing is found regularly in both recensions of the Hymn, e.g. M maecti P mehti Eordu mechti, 2a; M P uerc Eordu ƿuerc, 3a; M P tiadæ Eordu tiade, 8b. In this last example, smoothing of original */iu/ has been obscured by lenition of */χ/ > [h] with subsequent loss of [h] and lengthening between voiced segments (see Hogg 1992, § 7.45).

Back umlaut

§ 6.30    The short front vowels /i, e, æ/ were diphthongised in early Old English to /ĭo, ĕo, æ̆ɑ/ (spelled <io, eo, ea>) when followed by a back vowel in the next unstressed syllable (Hogg 1992, § 5.103). All dialects show back umlaut before unstressed /u/; back umlaut before /ɑ/ is also common with all short front vowels in all dialects except West-Saxon, where it is largely restricted to /i/. In Anglian, back umlaut occurs regularly before all consonants except original velars. The sound change belongs to the early historic or late pre-historic periods; it is rare and sporadic in the earliest Northumbrian texts, but becomes increasingly regular in later texts (Hogg 1992, § 5.112).

§ 6.31    There are no examples of back umlaut in either Northumbrian recension. The operation fails in M P metudæs Eordu metudaes, 2a, and Eordu hefunricaes, M hefaenricaes; P hefenricæs, 1b.

Unstressed syllables

Loss of final /n/

§ 6.32    Loss of word-final /n/ in specific morphosyntactic categories is a characteristic feature of the Northumbrian dialect. Loss is common in the oblique cases of weak nouns, infinitive endings, the subjunctive plural, and adverbs ending in -an. It is less common in the past participles of strong verbs, nouns ending in -en, and the past indicative plural. In the earliest texts, grammatically identical forms with and without final /n/ often coexist; in later Northumbrian, forms without /n/ are the norm (Hogg 1992, §§ 7.98-7.100).

§ 6.33    Word-final /n/ is lost consistently in the expected places in P and Eordu: e.g. P Eordu herga, 1a; Eordu, eordu, 5b; P Eordu foldu, 9a; M shows forms with and without final /n/: e.g. M herge\a/n (i.e. with a corrected from e), 1a, against fold\u/ (i.e. with u added in superscript), 9a.

Preservation of unstressed /u/

§ 6.34    Prehistoric Old English unstressed */u/ was lowered to /o/ in most dialects in unstressed syllables. Spellings in <u> are the norm in early Northumbrian and common in early Mercian, however. They remain common in later Anglian dialects (Hogg 1992, § 6.55-6.56).

§ 6.35    Unstressed /u/ is regularly preserved in inflectional endings, medial syllables, and derivative suffixes in both Northumbrian recensions: e.g. M scylun P scilun Eordu sciulun, 1a; Eordu hefunricaes (cf. M hefaenricaes; P hefenricæs), 1b; M P metudæs Eordu metudaes, 2a; M P uuldurfadur Eordu ƿuldurfadur, 3a; Eordu eordu, 5b (see Smith 1978, 35 § 18.a [note]); M middungeard (P middingard), 7a; M P Eordu foldu, 9a.

Preservation of unstressed /æ / /e/ and /i/

§ 6.36    Prehistoric Old English */æ, e, i/ fell together early in the historic period to a sound usually represented by <e> (see Hogg 1992, §§ 6.46-6.47, and, on individual sounds §§ 6.48-6.54), except when followed by a palatal consonant, in which circumstances <i> is common (see Hogg 1992, §§ 6.51-6.52). The sounds are represented with historical accuracy only in the earliest texts (Hogg 1992, §§ 6.48-6.49, 6.53). In later Northumbrian texts, etymological <i> is frequent; <æ> is the result of subsequent changes (see Hogg 1992 § 649, n. 3 and 6.62; cf. Cavill 2000, 520).

§ 6.37    Prehistoric */æ/ is generally preserved in both Northumbrian recensions of Cædmon’s Hymn, although more frequently in M and P than Eordu. Examples include the inflectional endings in M hefaenricaes P hefenricæs Eordu hefunricaes, 1b; M P metudæs Eordu metudaes, 2a; M Eordu gihuaes P gihuæs, 3b; M P astelidæ (cf. Eordu astalde), 4b; P hrofæ (cf. M Eordu hrofe) 6a; M P moncynnæs (cf. Eordu moncinnes) 7b; M P tiadæ (cf. Eordu tiade), 8b.

§ 6.38    Prehistoric Old English */i/ shows more variation in Northumbrian witnesses. It is preserved in both recensions in forms such as the gi prefix in 2b and 3b (a form also relatively common in later Northumbrian; see Hogg 1992, § 6.54) and M P Eordu eci, 4a and 8a. M maecti P mehti Eordu mechti, 2a, all show retention of /i/, though the failure of i-mutation in the M form, together with variation in the endings of M haleg (see also above, § 6.18), and M allmectig may suggest that the M scribe had trouble with the sound (see Hogg 1992, § 6.53, n. 1). /i/ is retained in the corresponding forms in P and Eordu: P allmehtig Eordu allmechtig, 9b. The sound is spelled <e> in the medial syllable of Eordu modgedanc, 2b.

West-Saxon features

§ 6.39    Sixteen witnesses belonging to three distinct recensions of the Hymn show late West-Saxon features. In eleven witnesses, the Old English poem has been copied alongside or within the main text of Bede’s Latin Historia ecclesiastica. In five manuscripts, the poem appears as part of the main text of the Historia’s Old English translation. The manuscripts range in date from s. x to s. xv (C, a s. xmed-s. xi1 copy of the Hymn was largely destroyed in the eighteenth century; it is cited from N, a 1562 transcription by Laurence Nowell). Since Cædmon himself was Northumbrian, all West-Saxon versions of the poem presumably have been translated from Northumbrian; major and minor linguistic and textual differences among these recensions, however, suggest that the extant versions stem from more than one translation (see Chapter 5: Filiation and transmission; §§ 6.52-6.54; Orton 1998). In the case of the West-Saxon eorðan recension, this translation was, like the rest of the Old English Historia, presumably via a Mercian intermediary (see the discussion and references in Whitelock 1962 and Grant 1989).

§ 6.40    As in the previous section, the following account provides a review of the most characteristic and significant features of the surviving witnesses and their exemplars rather than a comprehensive linguistic discussion. Even within this limited scope, however, no attempt has been made to detail all exceptional forms in individual witnesses.


§ 6.41    The West-Saxon witnesses have all been copied using the expected Latin-based alphabet. The native characters <ð>, <þ>, and <ƿ> are found in all witnesses except the late Tr1 and SanM, where <p> is substituted consistently for <ƿ>. CArms uses <u> for expected <ƿ> once (sua, 3b), and once reads <s> (seorc [expected ƿeorc], 3a).

Analogic and historical forms


§ 6.42    Old English swe, swæ, swa appears as sƿa (CArms sua) in all West-Saxon witnesses of the poem on its sole occurrence in 3b. The expected West-Saxon development is swæ, found in early West-Saxon. As in most dialects, however, swa develops irregularly in the later period. Swa is the usual late West-Saxon reflex (see Hogg 1992, § 3.25, n. 3).


Stressed syllables

Germanic */a/
Nasalisation and rounding

§ 6.43    Germanic */a/ is usually written <a> before nasal consonants and nasal clusters in later texts from southern dialects, although <o> is found sporadically (Hogg 1992, § 5.5).

§ 6.44    The West-Saxon witnesses show some variation in the spelling of this sound. Most witnesses to the West-Saxon recensions show <a> in modgeðanc, 2b. The only exceptions are the early eorðan witnesses N and O, both of which read modgeþonc. In the case of mancynnes, 7b, <a> and <o> are more evenly distributed: four of six manuscripts (T1 N O Ca) from the West-Saxon eorðan recension and all three representatives of the West-Saxon eorðe recension read mon-; all six witnesses to the West-Saxon ylda recension and two witnesses to the West-Saxon eorðan recension (B1 and To) read man(n)-. N reads Ond in 2b, although this may be the work of the witness’s s. xvi transcriber. All other witnesses to the West-Saxon recensions except SanM use the symbol for the conjunction. SanM has the nonsense form æ, presumably a misinterpretation of &, the Latin <e+t> ligature.

First fronting, breaking

§ 6.45    When not found before a nasal consonant or cluster, Germanic */a/ develops regularly into prehistoric Old English */æ/ (Hogg 1992, § 5.10). In West-Saxon, this newly fronted sound participated in breaking before /χ, lC, rC, w/ (see Hogg 1992, § 5.16).

§ 6.46    Breaking appears with some exceptions in the expected contexts in all witnesses to West-Saxon recensions of the Hymn. Examples include ƿeard (all West Saxon witnesses except SanM pærd, 7b), 1b and 7b; ƿeorc (all witnesses except the erroneous forms N ƿeoroda and O Ca ƿera, and the readings H Mg ƿurc and SanM pure, which show subsequent developments), 3a; -stealde in all West-Saxon witnesses (except Mg astalde and SanM astald), 4b; bearnum (all West Saxon witnesses), 5b; and -geard (all West Saxon witnesses), 7a. ylda, the reading of all witnesses to the West-Saxon ylda recension in 5b, underwent the expected change (from */æ/ to */æɑ/); the results are obscured by the subsequent effect of i-mutation (see Hogg 1992, § 5.74, 5.82). H Mg ƿurc SanM pure (for ƿurc) show late West-Saxon raising and retraction of short vowels and diphthongs between /w/ and /r/ (Hogg 1992, § 5.183).

Palatal diphthongisation

§ 6.47    Palatal diphthongisation of original /æ, æː/ to /ɪ̆ə, ɪə/ (spelled <ea>) is common in early and late West-Saxon (Hogg 1992, § 5.50). The resulting diphthong fell together with other diphthongs spelled <ea> and underwent i-mutation, producing early West-Saxon <ie>, late West-Saxon <y> or <i> (Hogg 1992, §§ 5.74, 5.82-5.83, 5.163-5.175). Diphthongisation of original /u/ after /ʃ/ is sporadic in early and late West-Saxon, with forms in <u> appearing alongside forms in <eo> and <y> that show palatisation and subsequent changes (Hogg 1992, §§ 5.66 and 5.67).

§ 6.48    All West-Saxon witnesses (except To, which reads drihten) show evidence of palatal diphthongisation with subsequent i-mutation to /y/ in scyppend, 6b, and, in W scippend only, /i/. The evidence for scylun, 1a, is more mixed. Four manuscripts from the West-Saxon eorðan recension (T1 N O B1) and five from the West-Saxon ylda recension (H W Mg SanM Ln) show unmutated forms with original <u>; one witness to the West-Saxon ylda recension (Tr1), two to the West-Saxon eorðan recension (To Ca), and all three witnesses to the West-Saxon eorðe recension, on the other hand, show forms with the post-palatal diphthong <eo>.

Back umlaut

§ 6.49    Back umlaut is more constrained in West-Saxon than in other dialects. It is common only before liquids, and labial and labiovelar consonants (i.e. /l, r, p, f, m, w/). When the back vowel is /u/, umlaut occurs with all front vowels; before /ɑ/ it is usually restricted to the high front vowel /i/ (see Hogg 1992, § 5.103).

§ 6.50    All West-Saxon witnesses to the Hymn show back mutation in heofon (-an, en), 1b and 6a. Back umlaut fails in metudes/metodes (except T1 meotodes), 2a.

Later changes to Old English diphthongs

§ 6.51    When not subject to subsequent changes, long and short Old English front diphthongs (/ĕo, eo, æ̆ɑ, æɑ/) are generally preserved in West-Saxon witnesses to the Hymn: e.g. ƿeorc (all witnesses except H Mg SanM, which show subsequent changes, and N O Ca where corrupt forms N ƿeoroda and O Ca ƿera may show signs of Northumbrian influence [see below § 6.53]), 3a; teode (West-Saxon eorðan and West-Saxon eorðe recensions only), 8b (cf. the late changes discussed in Hogg 1992, §§ 5.206-5.212, 5.214). In H Mg SanM, the form ƿurc shows late West-Saxon raising and retraction of short vowels and diphthongs between /w/ and /r/ (see Hogg 1992, § 5.183). The West-Saxon ylda recension form tida is best explained as a misunderstanding of an original exemplar showing the Northumbrian spelling -<ia>- (see below, § 6.53).

Anglian and Northumbrian features in West-Saxon recensions of the Hymn

§ 6.52    As Wuest 1906, Dobbie 1937, Ross 1949-1950, and especially Orton 1998 have pointed out, anomalous forms in all three West-Saxon recensions of the Hymn can be explained as evidence of the influence of an underlying Anglian or Northumbrian original. In three cases, the evidence for Northumbrian influence lies in problems with oblique endings of -n stem nouns (cf. § 6.32 above) where the surviving Northumbrian recensions show expected loss of final -n. Thus the O text of the West-Saxon eorðan recension reads folda\n/ (i.e. with a corrected to an), 9a, for expected West-Saxon foldan. In the case of the West-Saxon ylda recension, all surviving witnesses read a nonsensical dative plural foldum (presumably the result of a misinterpretation of Northumbrian -u as -ū) for expected accusative singular foldan. In the West-Saxon eorðe recension, all three witnesses read nonsensical eorðe (nominative singular) for expected eorðan (accusative or genitive singular), presumably reflecting, once again, a misinterpretation of an underlying exemplar in -u or -a.

§ 6.53    Two further examples show more serious problems involving the likely misinterpretation of originally Anglian forms. The nonsensical West-Saxon ylda recension reading tida for expected teode, for example, is best explained, as Dobbie demonstrated, as a misinterpretation of Northumbrian “tiadæ, or a similar form,” in the recension’s original exemplar (Dobbie 1937, 39). Unexpected N ƿeoroda and O ƿero{a} (i.e. with a corrected from original o) for expected ƿeorc, 3a, in two early witnesses of the West-Saxon eorðan recension, likewise, are explained by Wuest and Orton as misinterpretations of original forms showing Anglian smoothing, as in M P uerc and Eordu ƿuerc (Wuest 1906, 223; Orton 1998, 155-156).

§ 6.54    Several other forms spread across the different recensions show sporadic non-West-Saxon spellings in individual witnesses. Germanic */a/ before a nasal consonant and cluster, for example, appears in the early and Anglian spelling <o> rather than expected West-Saxon <a> in the early West-Saxon eorðan recension witnesses N and O in modgeþonc, 2b, and more generally in moncynnes (T1 N O Ca and all witnesses to the West-Saxon eorðe recension). Expected back umlaut of /e/ in metudes, 2a, fails in all West-Saxon witnesses except T1 meotodes. O middongeard for expected middangeard, 7b, has been explained as a “modernized spelling of the form middungeard” (Orton 1998, 157), the reading of M, a witness to the textually distant Northumbrian aelda recension (Orton 1998, 157; see also Ross 1949-1950, 94, n. 38, who suggests the O form may also be an example of careless copying).


[1]Arguments against the authenticity of the Hymn in these early manuscripts in Kiernan 1990b are unconvincing; see below, Note A: The dating of M and P and Note B: Cædmon’s Hymn and Bede’s paraphrase.

[2]A detailed defence of the reconstruction can be found in O’Donnell 1996b and, in abridged form, in the textual introduction to this edition (Chapter 7: Editorial introduction, Northumbrian eordu recension).

[3]All surviving witnesses read <d> in all positions; evidence for the reconstruction of initial and medial <ð> is discussed in O’Donnell 1996b, Wuest 1906, and below, Chapter 7: Editorial introduction, Northumbrian eordu recension; see also Cavill 2000, 520.

[4]Wuest argues that -<dea>- in the Di form may represent an original -<ða>-, see Wuest 1906, 216 and O’Donnell 1996b, 159-160. The reconstruction in Dobbie 1937, 21, normalises the spelling of all medial dental fricatives to <ð>. For a discussion of Dobbie’s procedure, see O’Donnell 1996b, 160, n. 53, and below, Chapter 7: Editorial introduction, Northumbrian eordu recension.

[5]This discussion is based on Hogg 1992, §§ 5.10-5.40. Hogg’s account differs in some details from that found in standard grammars of the mid-twentieth century (e.g. Campbell 1991, §§ 131 and 139, n. 1; Brunner and Sievers 1965, § 50); see Hogg 1992, §§ 5.10-5.13 for a discussion of the issues involved.