Beowulf 501b and the authority of Old English poetical manuscripts

Abstract

In his Old English grammar, A. Campbell put forward his theory of Old English accentuation, according to which disyllabic words like Bēowulf and ǣniġ receive a half-stress only if made trisyllabic by the addition of an inflection (as in Bēowulfes or ǣniġra), provided the middle syllable is heavy. A. J. Bliss tacitly rejected Campbell’s analysis when he postulated the existence of metrical type 3B2, a rare rhythmical pattern in which trisyllabic forms like Bēowulfes and ǣniġra must exceptionally be assumed to lose their otherwise conventional half-stress. Bliss based his analysis on the evidence afforded by four readings from the text of Beowulf: ll. 501b, 932b, 949b, and 1830b. Even though Bliss expressed reservations about his own analysis, he still preferred it to the alternative possibility of emending these four exceptional verses (a possibility that he did not even consider in his book). Non-metrical arguments in support of the emendation of at least one of these verses (1830b), however, had been advanced well before The metre of Beowulf was first published. Since then, ll. 932b and 949b have also been emended on grounds other than metre. This article offers new linguistic reasons for the emendation of l. 501b, the one remaining reading for which no alternative explanation had yet been proposed. It concludes that Bliss was unnecessarily cautious in his treatment of these four aberrant verses, and that, as Kenneth Sisam memorably stated in 1946, conjectural emendation is, and still remains, a useful tool for the study of Old English poetical manuscripts.

Personal names like Hrōðgār and Gūðlāf, which consist of two different elements or themes (for which reason they are known as dithematic), abound in Old English heroic verse. In terms of accentuation, these names can be seen to have affinities to both simplices and true compounds. If a second name element (or deuterotheme) is monosyllabic, then the personal name behaves like a regular disyllabic simplex. A verse like Beowulf 356b,¹þǣr Hrōðgār sæt, for example, is metrically equivalent to l. 42a, on flōdes ǣht (both are type B verses: x / x /). If the name is made trisyllabic by the addition of an inflection, however, the name is then metrically identical to a compound. Beowulf 1066b, Hrōðgāres scop, for instance, is a type E verse, like Genesis A 1990a, wælgāra wrixl (/ \ x /). Words that consist of heavy derivative suffixes, such as -iġ, -ing, and -isc, evince a behaviour in verse analogous to that of dithematic names, provided that, when the word is inflected, the middle syllable is heavy. In Beowulf 1099b, þæt ðǣr ǣniġ mon, which is a type B verse (xx / x /), -niġ occupies a drop because ǣniġ is disyllabic. When the word is made trisyllabic through the addition of an inflectional ending, the word then scans as a compound, as in l. 3127b, ǣniġne dǣl, a type E verse rhythmically equivalent to wælgāra wrixl and Hrōðgāres scop. On the basis of these differences in metrical behaviour between disyllabic and trisyllabic forms, which are well-attested in the poetry, A. Campbell formulated the grammatical rule that dithematic names and words containing heavy derivative suffixes receive a half-stress only when the addition of an inflection renders them trisyllabic (1959: §§88–89).² Thus, trisyllabic Hrōðgāres and ǣniġne scan / \ x (not / xx), but disyllabic Hrōðgār and ǣniġ scan / x (not / \).

Campbell’s account, though based on a very substantial number of unambiguous examples from verse, was tacitly rejected by A. J. Bliss. In his The metre of Beowulf, Bliss analysed as having an unstressed middle syllable a total of four words that, according to Campbell’s rule, ought to receive a half-stress by virtue of their trisyllabic forms (1967: §59). This group of four words consists of two dithematic names (Bēowulfes in l. 501b, wæs him Bēowulfes sīð, and Hiġelāce in l. 1830b, Iċ on Hiġelāce wāt) and two instances of (n)ǣniġ followed by an inflection (ǣniġra in l. 932b, þæt iċ ǣniġra mē, and [n]ǣniġre in 949b, Ne bið þē [n]ǣniġre gād). The rationale behind Bliss’s rejection of Campbell’s analysis is that, if -wulfes, -lāce and -iġ- received a half-stress in these four half-lines, as Campbell’s rule demands, then they would scan as type E verses with disyllabic anacrusis—a highly improbable interpretation, given that disyllabic anacrusis is virtually non-existent in stylistically conservative works like Beowulf (Bliss 1967: §50). Instead, he took these verses as the only instances in the poem of type 3B2 (xx / xx | /), a rare variety of metrical type B whose internal unstressed position is occupied by an unusually prominent element. Implicit in Bliss’s analysis is the assumption that, on these four occasions alone, the poet exceptionally ignored the otherwise conventional half-stress on inflected deuterothemes and heavy derivative suffixes. Bliss thus managed to avoid scanning these half-lines as type E verses with disyllabic anacrusis, to be sure, but he still expressed reservations about his own alternative analysis. As he put it, ‘even so the metrical pattern [of these four verses] seems anomalous’ (1967: 53).

Despite his reservations about his own analysis, Bliss still preferred it to the possibility that these four anomalous verses might be the product of defective textual transmission (an alternative that he did not consider in his book). This is no doubt a sign of the great authority that he accorded Old English poetical manuscripts. Still, compelling non-metrical arguments in support of emendation had been advanced for at least one of these verses well before the revised edition of The metre of Beowulf was published. Frederick Klaeber (1950: 193), in the notes to his third edition of the poem, mentioned the possible emendation of dative Hiġelāce to accusative Hiġelāc at l. 1830b, an emendation that was subsequently adopted in the text of the fourth edition of Klaeber’s Beowulf³:

Iċ on Hiġelāc wāt,

Ġēata dryhten (1830b–1831a)

[I trust that Hygelac, lord of the Geats.]

L. 2650b, God wāt on meċ, shows that the construction witan + on requires its object to be in the accusative case, not the dative. This emendation also remedies the lack of grammatical congruence between manuscript Hiġelāce and its appositive dryhten in the next line (an accusative form). The emended verse, Iċ on Hiġelāc wāt, is a conventional type B verse, like þǣr Hrōðgār sæt or on flōdes ǣht (with resolution of Hiġe-).⁴ It is easy to see how manuscript Hiġelāce came to substitute for Hiġelāc. Unlike in 2650b (God wāt on meċ), the verb wāt in 1830b comes after the preposition on. As a result, the scribe probably failed to construe the preposition in conjunction with witan, thereby writing mechanically the dative form of the name that he expected after on.⁵

In A history of Old English meter, R. D. Fulk argued that there are additional syntactic reasons to doubt the authenticity of ll. 932b and 949b (1992: §§242–243). These are the contexts in which these two half-lines occur:

Þæt wæs unġeāra     þæt iċ ǣniġra mē

wēana ne wēnde     tō wīdan feore

bōte ġebīdan (932–934a)

[It was not long ago that I did not ever expect to experience remedy for any of my troubles.]

Ne bið þē [n]ǣniġre gād.

worolde wilna,     þe iċ ġeweald hæbbe. (949b–950)

[There will not be for you lack of any worldly pleasures that are at my disposal.]

These two instances are parallel to the Modern English construction to have need of any help. Old English verse, however, has a preference for synonymous phrases of the type to have any need of help, in which any depends on the head rather than on the modifier of the phrase, as the following examples from across the corpus show:

þone synscaðan

ǣniġ ofer eorþan     īrenna cyst,

gūðbilla nān,     grētan nolde. (Beowulf 801b–803)

[Any paragon of irons over the earth, of war-swords, would not harm the criminal attacker.]

ah hē þāra wundra ā

dōmāgende,     dǣl nǣniġne

frǣtre þēode     beforan cȳðde. (Andreas 569b–571)

[But he, possessor of glory, did not ever reveal any portion of those wonders in the presence of that obstinate people.]

nē þē ǣniġ nēdþearf     næs ǣfre ġīet

ealra þāra weorca     þe þū ġeworht hafast (Meters of Boethius 20.20–21)

[Nor did you ever yet have any need for all the things that you have created.]

nysses þū wēan ǣniġne dǣl (Christ III 1384b)

[You did not know any measure of sorrow.]

As can be seen, forms of (n)ǣniġ agree grammatically with the heads of their phrases (cyst, dǣl, nēdþearf) rather than with the dependent genitives (īrenna, þāra wundra, ealra þāra weorca, wēan). Emendation of genitival ǣniġra and [n]ǣniġre to accusative ǣnġe and nominative [n]ǣniġ at Beowulf 932b and 949b results not only in more idiomatic syntax, but also in more conventional metre:

Þæt wæs unġeāra     þæt iċ ǣnġe mē

wēana ne wēnde     tō wīdan feore

bōte ġebīdan (932–934a)

[It was not long ago that I did not ever expect to experience any remedy for my troubles.]

Ne bið þē [n]ǣniġ gād

worolde wilna,     þe iċ ġeweald hæbbe. (949b–950)

[There will not be for you any lack of worldly pleasures that are at my disposal.]

Forms of (n)ǣniġ now modify the heads of their noun phrases (bōte, gād) rather than their dependent genitives (wēana, worolde wilna), as in the four examples above. The two verses, moreover, are now regular instances of type B with their two medial positions occupied by disyllabic words scanning / x (ǣnġe, nǣniġ): (x)xx / x /. The editors of the fourth edition of Klaeber’s Beowulf underpoint the re of [n]ǣniġre at l. 949b, thus signalling that [n]ǣniġ is their preferred reading here. Although they leave ǣniġra at 932b unmarked in their text, their belief that this verse has suffered some sort of scribal tampering is registered both in the commentary and in the appendix on textual criticism (Fulk et al. 2008: 173, 331).

It is only natural that a scribe substituted genitive plural ǣniġra for accusative singular ǣnġe at l. 932b. The word is in close proximity to genitive plural wēana, in l. 933a, while accusative singular bōte (on which ǣnġe originally depended) does not show up until 934a. In this instance, the flexible word order of Old English verse, which allows closely related words to stand quite apart from each other, confounded the scribe, who misconstrued ǣnġe as an erroneous modifier of wēana in need of correction. The form [n]ǣniġre cannot be similarly explained, to be sure, as the word stands next to the head of the phrase, nominative gād. That the scribe spoiled the line’s alliteration by writing down a form of ǣniġ rather than of nǣniġ, however, suggests that his exemplar was in some way defective at this point. According to Fulk, the scribe decided to alter the word, originally in the nominative, in an attempt to make it agree with masculine genitive plural wilna, but ended up making it agree with feminine genitive singular worolde (both in l. 950a).⁶ Regardless of how the erroneous form might have crept into the text, syntactic considerations suggest that ǣnġe and [n]ǣniġ are likely to have been the readings at 932b and 949b in the scribe’s exemplar. Thus, a variety of non-metrical arguments has been advanced suggesting that three of the four verses that seemingly contradict Campbell’s analysis (1850b, 932b and 949b) are in actuality regular type B verses that comply with his grammatical rule. L. 501b, wæs him Bēowulfes sīð, is therefore the only one whose authenticity has never been called into question on non-metrical grounds, for which reason the editors of the fourth edition of Klaeber’s Beowulf consider this verse the only secure attestation of metrical type 3B2 in the poem (Fulk et al. 2008: 331).

There are, however, linguistic reasons to suspect that the exceptional metrical pattern of l. 501b is likewise the result of scribal interference rather than of poetic practice. This is the context in which this verse occurs:

Ūnferð maþelode,     Ecglāfes bearn,

þe æt fōtum sæt     frēan Scyldinga,

onband beadurūne.     Wæs him Bēowulfes sīð,

mōdġes merefaran,     miċel æfþunca,

forþon þe hē ne ūþe     þæt ǣniġ ōðer man

ǣfre mǣrða þon mā     middanġeardes

ġehēdde under heofenum     þonne hē sylfa (499–505)

[Unferth spoke, the son of Ecglaf, who sat at the feet of the Scylding lord, unbound secret hostility. The expedition of Beowulf, that brave seafarer, was for him miċel æfþunca, since he would not allow that any other man in the world should care for glorious deeds more than he himself.]⁷

I have left l. 502b, miċel æfþunca, untranslated because it is upon consideration of its debatable meaning that the problematic character of this passage will become recognisable. The Dictionary of Old English (DOE) classifies the 25 occurrences of the word æfþunca into two semantic groups, ‘a. spite, disdain; insult, offence’ and ‘b. envy, cause of envy’, and gives Beowulf 501–502 under the second definition. This is probably where this instance belongs, since all the occurrences listed under the first definition refer to actions perpetrated with the intention to insult or harm other people. The following examples will suffice to illustrate this point:

Eorl ōðerne     mid æfþancum

ond mid tēonwordum     tǣleð behindan,

spreceð fæġere beforan,     ond þæt fācen swā þēah

hafað in his heortan,     hord unclǣne. (Homiletic Fragment I 3–6)

[A man abuses another with insults and calumnies behind his back, yet speaks fairly to his face, holds all that malice within his heart, an impure hoard.]

Ongan æfþancum     āgendfrêan

halsfæst herian,     hiġeþrȳðe wæġ,

wæs lāðwendo,     lustum ne wolde

þēowdōm þolian,     ac hēo þrīste ongan

wið Sarran     swīðe winnan. (Genesis A 2239–2243)

[The slave (Hagar) started to denigrate her mistress with disdain, she carried herself with arrogance, she was hostile and did not endure servitude gladly, but she began to compete fiercely and boldly with Sarah.]

Mæġen nēalǣhte

folc Ebrea     fuhton þearle

heardum heoruwǣpnum,     hæfte guldon

hyra fyrnġeflitu     fāgum swyrdum

ealde æfðoncan (Judith 261b–265a)⁸

[The military force approached, the Hebrew people fought bitterly with hard martial weapons, they avenged their ancient struggles, the long-standing grievances, with blades, with ornamented swords.]

There is an evil intention behind the actions denoted by the æfþuncan of these passages: the term is used to refer to the hurtful words or cruel acts that an anonymous man (in Homiletic Fragment I), Abraham’s female slave (in Genesis A), and the Assyrians (in Judith) directed respectively against another man, Abraham’s wife, and the Hebrews. In the passage from Beowulf, on the other hand, æfþunca denotes Beowulf’s expedition to Denmark, which he undertook for the benefit of Hrothgar, not out of hostility towards Unferth. It could of course be argued that Beowulf’s presence at Heorot is perceived by Unferth as an aggression, after all, and that æfþunca at l. 502b should therefore be construed as an instance of the first definition, ‘spite, disdain; insult, offence’. The excessive sense of rivalry evinced by Unferth (who found it unacceptable that anyone else in the entire world should do better than him) suggests, however, that the word should indeed be interpreted as belonging to the semantic field of ‘envy’.⁹ Moreover, if Unferth regards Beowulf’s arrival as an insult, this would lend æfþunca a metaphorical dimension unparalleled among other instances of the word and uncharacteristic of Old English, in which metaphors other than kennings are, in general, exceedingly rare.

The DOE gives 11 instances of æfþunca under definition ‘b. envy, cause of envy’. In nine of these, different forms of the Old English word are used to gloss Latin inuidia (two glosses to Aldhem’s prose De uirginitate and one gloss to the Penitential of Pseudo-Theodore) or zelus (one gloss to Defensor’s Liber scintillarum and five Anglo-Saxon psalter glosses), both of which mean ‘envy, jealousy’ in their respective contexts.¹⁰ Another instance, also meaning ‘envy’, occurs in the following passage from Vainglory:

Bið þæt æfþonca     eal gefylled

fēondes fligepīlum,     fācensearwum. (26–27)

[That is entirely filled with the devil’s flying arrows of envy, (with his) treacherous wiles.]

Thus, according to the DOE, of the 11 occurrences of æfþunca under definition ‘b. envy, cause of envy’, the sense ‘cause of envy’ is attested by only one instance in the entire Old English corpus: Beowulf 502b, miċel æfþunca. It is easy to see why this occurrence alone was assigned such an odd meaning. In the clause Wæs him Bēowulfes sīð, mōdġes merefaran, miċel æfþunca (ll. 501b–502), two nominative nouns, sīð and æfþunca, are linked by the copula, wæs. Sīð therefore equals æfþunca in this passage, but ‘expedition’, a countable noun, cannot equal uncountable ‘envy’, and so an alternative sense for æfþunca had to be found. In the light of the ensuing causal subordinate clause introduced by forþon (ll. 503–505), ‘cause of envy’ probably presented itself as a reasonable alternative.

That such an unusual sense occurs in a clause that also happens to contain a metrically exceptional half-line (wæs him Bēowulfes sīð) suggests that something has gone awry with this passage in terms of textual transmission.¹¹ Of all the constituents of this clause, it is sīð that appears to be at the root of the problem. It forces æfþunca to mean something other than ‘envy’, it makes l. 501b feature an excessively heavy metrical pattern, and its contribution to the overall meaning of the passage is superfluous. Deletion of sīð results in an improvement on all these fronts¹²:

Wæs him Bēowulfes,

mōdġes merefaran,     miċel æfþunca,

forþon þe hē ne ūþe     þæt ǣniġ ōðer man

ǣfre mǣrða þon mā     middanġeardes

ġehēdde under heofenum     þonne hē sylfa (501b–505)

[He had great envy of Beowulf, of that brave seafarer, since he would not allow that any other man in the world should care for glorious deeds more than he himself.]

The emended version of l. 501b can now be seen to comply with Campbell’s grammatical rule of dithematic names: wæs him Bēowulfes is a regular type C verse with a half-stress on its inflected deuterotheme, -wulfes (cf. 856b, ðǣr wæs Bēowulfes, 1043a, ond ðā Bēowulfe). With sīð deleted, there is no need to assign a non-attested sense to æfþunca: its meaning here can now be seen to be precisely the same as that in the other ten occurrences of the word discussed above (‘envy’). Ll. 501b–502 are now similar to passages from the poem like the following:

Denum eallum wearð,

ċeastẹrbūendum,     cēnra ġehwylcum,

eorlum ealuscerwen. (767b–769a)

[All the Danes, the town-dwellers, every brave man experienced distress.]

Norð-Denum stōd

ateliċ eġesa,     ānra ġehwylcum

þāra þe of wealle     wōp ġehȳrdon (783b–785)

[The North-Danes, each one of those who heard the mourning from the wall, felt real terror.]

As in ll. 501b–502, the experience of an emotion is being described in each of these two passages. In all three of them, the emotion experienced appears in the nominative singular as the subject of the clause (æfþunca, ealuscerwen, eġesa) while the experiencer appears in the dative (him, Denum eallum, Norð-Denum).

Scholars of Beowulf will no doubt be reluctant to accept the emendation here proposed unless a plausible explanation for the scribal interpolation of sīð at l. 501b (upon which my argumentation depends) can be offered. The phrase Bēowulfes æfþunca, which I take to be the original subject of the clause at ll. 501b–502, consists of a noun of emotion modified by an objective genitive. This means that the noun in the genitive denotes the object rather than the subject of the emotion, as in Godes eġe (‘fear of God’, not the fear felt by God), mannes eġe (‘fear of a man’, not the fear experienced by a man), or þissere andwerdan worulde lufe (‘love for this transitory world’, and obviously not the love felt by this world).¹³ Thus, Bēowulfes æfþunca does not designate Beowulf’s envy, but rather Unferth’s envy for the Geatish hero. In her monograph on Genitives in early English, Cynthia L. Allen demonstrated that objective genitives with nouns of emotion underwent a process of gradual obsolescence throughout the history of the English language. To put it in her own words:

While ‘subjective’ genitives such as John’s love (for Mary) are still generally possible, and ‘objective’ genitives of some types, such as the president’s assassination are also still possible, some other types of objective genitives became impossible in ME or EModE, and I will focus here on one particular type: objective genitives in which the genitive refers to the target of an emotion. In PDE, an expression such as God’s love can only refer to the love which God has for someone, not the love which someone has for God (2008: 74).

Objective genitives with common nouns of emotion like lufu and eġe were still possible in late Old English, to be sure, but æfþunca is a rare word, disproportionately frequent in poetry.¹⁴ If this noun was no longer recognized as taking an objective genitive by the time when the Beowulf manuscript was written,¹⁵ the scribe might well have thought that the genitives at 501–502a (Bēowulfes, mōdġes merefaran) needed to depend on a noun other than æfþunca, and hence he supplied one.¹⁶

The metre of Beowulf is a major achievement in Old English studies. Published for the first time in 1958, the book promises to remain influential and serviceable to future generations of scholars in years to come—and justly so, given its extensive range of analysis and its careful attention to detail. In the view of the present author, however, Bliss’s refusal to consider the possibility of textual corruption for readings like wæs him Bēowulfes sīð is an unnecessarily cautious stance to adopt, especially in the light of his reservations about his own alternative interpretation. L. 501b and its immediate context are perfectly legible on fol. 141r of the Beowulf manuscript (Zupitza 1959: 24–25), to be sure, but the co-occurrence of metrical and linguistic problems at ll. 501b–502 renders conjectural emendation an appealing and justified course of action. As Kenneth Sisam argued, ‘there would be a real gain if conjecture, instead of being reserved for the useful but disheartening task of dealing with obvious or desperate faults, were restored to its true functions, which include probing as well as healing’ (1946: 268). It is to be hoped that emendation proposals for similarly aberrant manuscript readings will continue to be brought forward in the future, lest the very real possibility that they might be non-authorial in origin go unacknowledged.