Dating Deor, Beowulf and other Theedish poems: a naming problem

To my mind, the question of how to date not only Bēowulf, but also Dēor and other short poems (most notably Widsiþ) depends as much as anything on how strongly one credits the argument (most magisterially articulated by Roberta Frank) that Germanic legend essentially originated — or only became popular in England — during the Carolingian period. If one accepts Frank's premise in toto, then it might follow that there is "no good reason" to regard Deor or Widsiþ as early compositions. It is part of what emerged since 1980 as the case for a Late Beowulf, which is in turn intertwined with how one dates the shorter "lore" poems like Widsith, Deor and possibly also Wulf and Eadwacer.

Now, I've already stated where I stand on the date of Beowulf. My position is a controversial one only among literary historians and theorists. It is pretty standard among linguists who have worked on the text. I've also already given vent to my general sense that  literary historians' arguments against an Early Beowulf strike me as unsound. The same objections raised against an Early Beowulf could just as easily — and just as misguidedly — be raised against an early date for a lot of Old Irish poetry that only survives in late manuscripts. It's worth noting that even among Anglo-Saxonists of a more literary persuasion, there are those such as Andy Orchard who have started to accept the linguistic evidence for an early date.

But Frank's argument, in a very modified and modest form, is worth considering.

But before I get to that, I should clear the air by saying that if taken without qualification, I find the hypothesis implausible. Leonard Neidorf and Tom Shippey in a series of articles give very good reason to think that figures like Theodric, Eormanric and others were known in 7th and 8th century England. Names like Ætla, Offa, Hama, Ingeld, Theodric and Wyrmhere deviate from statistically demonstrable name-formation trends in Anglo-Saxon England. (As does the name Beowulf.) These names, which could not have been generated by chance from native Anglo-Saxon name-formation practices, are commonly found as names born by historical Anglo-Saxons in the 7th-8th centuries, after which point they wane in popularity. Neidorf's most recent case on this point seems fairly secure. The straightforward interpretation is that 7th-8th century Anglo-Saxons were naming their children after these legendary figures.

If "Germanic legend" was new to the Anglo-Saxons in the 9th century it would imply a considerable ability on their part to take an unfamiliar name (or unfamiliar common noun) in a different Germanic language and confect an English pseudo-cognate. Some, such as Niles, have argued against an Early Beowulf by trying to dismiss this evidence:
"[it is a] time-honored dictum that the poem must be early because the names of the Scandinavians who figure in it are recorded in English-language forms (e.g., “Scyld” rather than “Skjöld”). There is no significance in that fact, however, since a poet working in the traditional Old English verse medium would have used native or nativized forms of those names as a matter of course, much as a Danish jarl is an Old English eorl"
To be sure, the ability to recognize and even create diasystemic cognates existed at some level at a late period. That much is suggested by a famous passage in the Old English Orosius which seems to imply that an English speaker in the Alfredian period, hearing Norse hreinn "reindeer" (or rather Viking-age *hræinʀ), would have recognized the diphthong corresponding to English ā and thus produced a pseudo-cognate hrān. (Scholarly discussion of Anglo-Norse language contact during the Viking Age has been extensive. Townend plausibly suggests that some sound-correspondences would have been so transparent that speakers may have been able to negotiate them diasystemically without even thinking about it.) It stands to reason that late OE speakers would have been able to Anglicize some continental Germanic names. All in all, an OE speaker of any period faced with the Frankish name *Dagahramn could probably have turned it into the attested Dæghræfn of Beowulf faster than you can say Dayraven. For somebody working off of Latin sources mentioning Ermanaricus to have produced the attested Old English form Eormanric in the 9th century would be possible if he recognized both cognates of this dithematic name. (If he did not, then the <eo> form would be unexpected in a late loan. Breaking of non-high short vowels before /rC/ was no longer productive, and is absent from words like berstan where the cluster is a late development due to r-metathesis.) But it is worth noting that when proper names were encountered in later periods, they could often be taken over into the recipient language as ordinary loanwords, nor diasystematic cognates. Ermanric is turned into Erminrekr in the Þidriks saga af Bern, a text with clear Low German connections, which suggests that the form Jǫrmunrekkr (attested e.g. in the Eddic Hamðismál) was either unavailable or unsuitable to the compiler. 

But the name Ætla, the Old English version of Attila, shows Anglo-Frisian brightening of *ă to æ, as well as medial syncope of a high vowel following an originally heavy syllable. Attila would be quite etymologically opaque to an Old English speaker, and the form Ætla can hardly have been a newcomer to English even in the 8th century. Like its High German congener Etzel, it can only be explained as an early acquisition that underwent pre-historic sound changes. Over a hundred years ago, Alois Brandl observed that if the transmission of Migration-Era heroic material to England was much later than the Migration Era itself, one would not expect the Old English forms of all relevant proper names to affected by early pre-historic sound shifts. It would be possible to turn an OE eorl into a Danish jarl, as Niles' review notes. It might even be possible to turn Ermanaricus into Eormanric. But it strains credulity to think that an OE speaker of the historical era would have been able to turn Attila into Ætla by accident. It would require someone to read Attila in a Latin text and then somehow be willing and able to create the pseudo-cognate Ætla. It would, in other words, require not only a diasystemic awareness of processes like Anglo-Frisian brightening and high vowel deletion but also a willingness to impose those processes on the the word Attila. (In fact Attila may not even be Germanic at all. If it is, is based on a Gothic root att- with no cognate in attested Old English.)

This point seems to have been quite lost on many who argued for a late Beowulf. Eric Stanley, for example, sought quite eye-popping levels of special pleading in his rejection of all onomastic evidence for an early date by proposing that — in essence — someone "who was unusually good at comparative Germanic philology" transposed all of these names. It boggles me a bit that one would reject the possibility that these names are found in their attested forms because they have existed in Old English long enough to go through OE-specific sound changes, only to embrace an implausible level of philological sophistication on the part of a medieval poet. Evidence that English speakers in the Viking age were nowhere near that good at transposing Norse proper names is so common as to approach staggering banality. The transcriptions of Norse names in the Anglo-Saxon chronicle, for example, show that attempts to nativize them were seldom fully successful and often spangled with misconstrued morphemes. The demonstrable bizarrity of this kind of argument is one reason, I think, why some Early-Beowulfers like Michael Drout have sensed that something was "not quite on the level" about how the Late Beowulf position was argued for from 1980 onward. When thoughtful, intelligent scholars publicly commit themselves to propositions that collapse into absurdity upon the slightest examination, it is not unreasonable to ask what's really going on....Alright I'm leaving all that alone now.

Anyway, another proper name that pretty certainly a Migration-era survival is the inherited Old English word for Scandinavia, Scedenig, which (along with the variation Scedeland) occurs only in Beowulf and is derivable from Common Germanic *Skadinawjō whence also Old Norse Skáney.

It is quite impossible to convincingly explain the inherited OE form as anything other than an old lexical item dating to the Migration Period. English-speaker hearing even the most conservative pronunciation of Viking-age Norse would not have been able to turn Scáney into Scedenig. And in fact they did not. Norse Skáney was loaned into English as Sconeg/Scani which replaced the native OE word in the 9th-10th centuries. Those partial to a late date for Beowulf, particularly those who wish to substantiate one or another theory that demands a late Beowulf as a corollary, have had to find ways to dismiss Scedenig as something other than a Migration-era survival. Dismissals have not been wanting. Roberta Frank suggested that Sconeg is the native English term whereas Scedenig is the Scandinavian loan, but this has generally — and rightly — been seen as implausible. A common dismissal tactic is to suggest that Scedenig is in Rix' words "borrowed from the classical geographic name for 'Scandinavia'". Historical linguistic developments, again, make this implausible. Right off the bat, Scandia is out of the picture due to the absence of <n> in the OE form. Scadinavia superficially seems more promising, but even if one accepts it for the sake of argument, one is still stuck with the vowel in the initial syllable which would require that the loan be effected before i-umlaut had taken effect. In which case the form is still quite early. A connection to Scadanan mentioned in the Lombardic Origo Gentis is thinkable, and it is completely possible that connection of Scede- to that Scada- was on the poet's mind. But it seems a weird thing to propose, and again the vocalic changes found in scede- would require the loan to be very early indeed in any case.

It is a common complaint in the Late Beowulf camp that we simply do not know when the word Scedenig or Scedeland fell out of general use or knowledge. This is in a strict sense correct. (Though its antiquity does suggest that — one way or another — the word for Scandinavia remained relevant to enough OE speakers to warrant its survival. If the word for Scandinavia is very old, so too is the habit of talking about it. This must give some modest support to the idea that stories set in Scandinavia were not new in the 9th century.) Since we have next to nothing in the same genre as Beowulf to compare the text with, we don't know the extent to which lexical choices were determined by genre. The word scedenig may never have been needed, or thought appropriate, in anything other than a poem dealing with the Heroic Age. And a toponym of course may remain in memory long after it has stopped being commonly used (compare English Cathay.) But we do have some evidence that the specific word scedenig presented a problem for the copyist responsible for our only surviving instance of the word. One of the scribes copying Beowulf originally wrote <scedeninge> at 1686a before correcting himself by imposing a second g upon the second n. The self-corrected instinct to write -ing suggests that the scribe took this for an ethnonym rather than a toponym, and ethnonyms themselves clearly gave the scribes a lot of trouble. The scribe, in other words, looked at the word in his exemplar and thought it might belong to a class of words which had already proven confusing to him on multiple occasions. This does not by itself prove that the word <scedenig> was unfamiliar. It may just be that a word <scedening> also existed, and the scribe simply lapsed into a malapropism which he later corrected. But the balance of probability does not seem to lie there.

Now onto another name.

Alcuin's famous interrogative tantrum "Quid enim Hinieldus cum Christo?" is interesting in this regard too. This reference to Hinieldus is not found in all witnesses to Alcuin's letter, but as far as I have been able to determine the spelling <Ingeldus> does not occur at all. I suggest that, if the reference is original to Alcuin, then the spelling <Hinieldus> is likely original to him. If it is an interpolation, then it is original to the interpolator. It seems unlikely that a spelling <Ingeldus> would have been transformed into <Hinieldus>. Either way, it is significant that we have here a spelling of Ingeld's name with <i> rather than <g>. Palatalization of velars had ceased to be a productive process in Old English before the onset of i-umlaut. The onomastic theme -g(i)eld is not productively used in Anglo-Saxon naming. The spelling <Hinieldus> thus tells us two things. The first is that the name Inġeld existed in English early enough to be subject to a sound-change that pre-dates even the earliest texts. This is not a huge deal, as it merely confirms what we know from the onomastic evidence found in the early 7th-8th century core of the Northumbrian Liber Vitae, where no less than sixteen people named Ingildus or Ingeldus are listedand the PASE database which also lists three people by the name of Ingeldus from the early-to-mid 8th century. The second point is that Alcuin did not take the name of the legendary Ingeld from a continental or Latin source. He is remembering a name he had clearly heard, but may well have never seen at all in written form. This name was spelled in Latin — as in English — with <g>, and Alcuin's letter is the only attested exception.

Now, about dynastic names.

Frank's position:
In England, the devising of elaborate royal genealogies was a fairly late, antiquarian exercise. The several backward expansions of the Anglo- Saxon king-lists testify to a growing, and constantly changing, need to establish legitimacy through illustrious continental ancestors. In the age of Bede, Woden was the stopping point. But the Anglian collection of royal pedigrees, compiled around 796, gives Woden a progenitor; and then – for the kingdoms of Lindsey and (in a narrative part of the Historia Brittonum, c. 830) Kent – additional ancestors going back several generations to Geat (Primitive Germanic *Gautaz, probably the Gothic eponym). A pedigree going back to Geat apparently had propaganda value for English kings around 800, when the Carolingians were rediscovering their Gothic roots. The genealogy of King Alfred’s father Æthelwulf, added around 892 to the Anglo-Saxon Chronicle, gives Geat a number of northern ancestors, among whom five – Scyld, Scef, Beaw, Heremod and Hwala – appear as legendary figures in Old English poetry. The expansion backwards to Scyld, eponymous ancestor of the Danish Scyldings, marks what looks like a new social reality, the integration of Dane and Englishman in one kingdom. Royal houses acquired not a little mythological depth and perhaps even some political legitimacy by claiming descent from the gods and rulers of the heartland of northern Europe. And what was of interest to kings was of practical and immediate interest to their subjects.
This argument doesn't convince me. Sam Newton has made a good case that elaborate versified genealogies were by no means late, and did not become antiquarian until fairly late. There is strong evidence for composition of early Anglian dynastic verse. It is true that the Anglian genealogies are preserved in a Mercian collection compiled ca. 780–796. But these appear in turn derived from an earlier Northumbrian collection compiled several decades earlier. Several factors suggest that — whether or not they existed in their attested form during the actual reigns of the respective kings — these genealogies draw on much older dynastic lists one way or another. And as Neidorf points out, Æthelwulf's genealogy in particular contains forms like <sceldwa> and <beowi> with morphology that cannot have been the product of either dictation or free composition. Some form of somebody's genealogy containing these names must have already existed in a written document which was already quite old. Moreover, as Sisam originally noted, the late iterations of the Æþelwulf genealogy contain so many variant forms and scribal corruptions that I have a hard time believing that these were part of a coherent, robustly-supported political program. If they were, one might have expected scribes to take a good deal more care than they did. Corrupting Scef into Seth, or conflating him with Scyld, would be quite a whoopsie moment for a scribe padding out a genealogy at the behest of a royal house in need of a beefier dynastic legitimacy. These names issue from the scribes' quills in a way that suggests that they were hardly on the tip of their tongues.

I did not bring in Frank's argument merely in order to have something to disagree with. She is onto something, even if I disagree with her about what, precisely, she is onto. There is, to be sure, something interesting about the fact that anybody would even bother to go rummaging for a document where these archaic mythical figures lay mummified in scraps of archaic Anglian morphology. My takeaway is that some people certainly were interested in reeling out an old and Scylding-bejeweled lineage, and that interest may have expressed itself in a flashpan resurrection of arcane material. But those charged with transmitting this undead filiation were quite literally not on the same page.

Antiquarianism is a common motivating explanation put forth in arguments for a Late Beowulf (and by extension a late Widsiþ, Waldere etc.) and while I am unconvinced by the argument itself, I do think antiquarianism is at play in a different way. It seems to be a recurrent theme in the vestigial survival of archaic Migration-era figures and legends in Anglo-Saxon material. With the Æþelwulfian genealogy as with Beowulf, the evidence suggests that a document with extremely archaic material had to be disinterred. Anglo-Saxons' gradual re-discovery for the very first time that they were indeed "Germanic" may have eventually given a few antiquarians the notion that such archaic heroic material as was still around to be re-copied might indeed be worth preserving. The existence of the Waldere fragment suggests, if nothing else, that the copying of Beowulf was not a one-off fluke.

 Frank puts it thus in the Cambridge Companion to Old English Literature:
Goths were not seen as chic or German during the long period stretching from the death of Theodoric to the coronation of Charlemagne. Isidore, writing in seventh-century Spain, could see no family relationship between Goths and Franks; he believed that the former were descended from the Scythians. Fredegar, a Frank writing around 660, portrayed Theodoric the Ostrogoth as a Macedonian, reared in Constantinople; he, like the author of the Liber historiae Francorum (c. 727), honoured the Franks with Trojan, not Germanic, ancestry....
...Between 805 and 860, we can trace, decade by decade, a growing interest in the Goths and their language: shortly before 800, in one early Carolingian text, the term theodisca lingua had expanded to include Old English and Langobardic as well as Frankish; by 805 Gothic had joined; by 830 all nationes theotiscae (‘Germanic peoples’), Franks included, were, like Jordanes’s Goths, given Scandinavian ancestry; and finally, around 860, a theologian could speak of a gens teudisca, a community of German-speaking people. Stories about Ermanaric were recorded by scholars, and, on at least one occasion, used by a cleric to restrain a king. Towards the end of the century, Archbishop Fulk of Rheims (883–900) asked Arnolf of Carinthia, the East Frankish king and emperor (887–99), to show mercy to his kinsman Charles, exhorting him ‘not to follow evil counsels, but to have pity on his people and strengthen a declining royal race, keeping in mind the example found in German books (ex libris teutonicis) of King Hermenricus [Ermanaric].."
I would add that all of the examples Frank has adduced for this surging interest in Germanic legend come from Continental Europe. With the exception of scattered references to Wayland the Smith, who survived in folklore, Germanic legendary figures are conspicuous by their near-total absence from texts actually produced (rather than merely copied) by Anglo-Saxons after ca. 850.

So if Anglo-Saxons knew about Theodric, Attila, Eormanric, Hroðulf and others long before the 9th century, how is it that a generalized "Germanic" sensibility in Europe only emerges in the mid 800s, just as interest in the old "Germanic" lore seems to have turned moribund Anglo-Saxon England?

An answer, or at least the silhouette of one, presents itself if we abandon some traditional assumptions about what "Germanic legend" and "Germanic identity" are, or rather what they were. Migration-era Germanic languages were more than mutually intelligible enough to facilitate easy borrowing and the spread of sound-changes in geographic waves. Stories about increasingly legendary heroes could easily spread among Germanic groups, especially as long as Anglo-Saxon migration from northern Europe to Britain was still ongoing. But this by itself tells us nothing about how members of different Germanic groups — at war as much with each other as with "foreign" peoples — conceived of themselves or each other, let alone how this changed over time. That Anglo-Saxons in the 700s already knew of legendary Migration-era figures does not necessarily mean that the legends were always felt to be "Germanic" in an ethno-linguistic sense of the term. Goths may have been very much a part of cultural imagination in Germanic-speaking Britain well before the year 800. But there is, of course, no reason why an 8th century Anglo-Saxon named Þēodrīc should necessarily have regarded his Gothic namesake as ethnic kin of the same kind as continental Saxons, even assuming that his 6th century ancestor might have.

Anglo-Saxons never forgot their connection to the lands across the North Sea. But it is well to remember that whatever attitudes may have continued to prevail toward the inhabitants of ancient Scedenige in the 7th-8th centuries, they would not necessarily translate to Goths or Franks.

"Germanic-ness" itself as we understand it may indeed be a partly Carolingian invention. It may be a flawed concept when dealing with the way Anglo-Saxons saw themselves in the 7th or early 8th centuries. But how did they see themselves? Certainly not as "Anglo-Saxons" yet either.  Words that could translate as "Anglo-Saxon" are not attested before the 9th century. Bede and others can already be found referring to a "Germania" with a sense of connection, and Aldhelm even uses the adjective "Germanicus" in direct reference to the English "gens" (neminem nostrae stirpis prosapia genitum et Germanicae gentis cunabulis confotum). Before 800 the adjective Englisc may well have meant something like "Insular Germanic" (though for a value of "Germanic" that did not necessarily include every group we would mean by that term.)

In fact, if the earliest references to "theudisc" identity come from the Carolingian continent, the same is true for "Anglo-Saxon" identity. Paul the Deacon, whom Frank characterizes as "the first 'German' historian with a sense of ethnic solidarity" is also one of the first attestations of the term Anglisaxones. 

"Pan-Germanic" identity — of a different character than whatever complex of tribal, supra-tribal and linguistic affiliations may have obtained during the Migration Era — seems to me to go hand-in-hand with a more forcefully articulated concept of "Englishness."

If 7th-8th century Anglo-Saxon identity was not "Germanic" in the simple sense we associate with that word, then neither was it yet anything we could call English.

In any case it will not do to insist and infer from this that Anglo-Saxons had no idea who Ingeld, Theodoric or the Scyldings were until the Carolingians started tooting their þeodisc trumpet.

I suspect that at nearly the same time that tales of "Theudisc" heroes were growing popular (at least among monkish sorts) in Carolingian Europe, Viking invaders inspired anxiety about the heathenness of certain English customs and traditions in Anglo-Saxon England. Alcuin makes clear that his problem with Ingeld is not that he is from the Northlands, but that he is heathen. Germanic legendary matter would not pose the same kind of problem for continental Carolingians. Alcuin elsewhere puts the gears to King Ethelred and his court for wearing the same hairstyles as heathen Northmen. A letter-fragment plausibly attributed to Ælfric puts the matter quite bluntly: "Ic secge eac ðe, broðor Eadweard, nu ðu me þyses bæde, þæt ge doð unrihtlice þæt ge ða Engliscan þeawas forlætað þe eowre fæderas heoldon and hæðenra manna þeawas lufiað, þe eow ðæs lifes ne unnon." That author has in mind Danish fashions in clothing. But, in a context like this, one should not be surprised that the sort of Migration-era lore prized by the Danes might fall out of favor with the types of people and institutions that produced most of our surviving texts.  

1 comment: