A History of Old English Meter.
Once upon a time, we thought we knew quite a bit about the chronology of the surviving poetry of Anglo-saxon England and about the regional dialects in which the poems were originally written. This knowledge depended heavily (though not exclusively) on the minutiae of Germanic philology -- on such arcane matters as etymological reconstructions, sequential sound changes, orthographical distributions, and metrical analyses. But, in the waning decades of the twentieth century, it has come to seem that the edifice erected by the old-fashioned philology was constructed of circular arguments, mistaken assumptions, and erroneous data.
R. D. Fulk challenges our suspicions of arguments built upon philological evidence. The primary target of A History of Old English Meter is an influential book, published in 1980 -- Ashley Crandell Amos's Linguistic Means of Determining the Dates of Old English Literary Texts. Amos passed in review the principal phonological, metrical, syntactic, lexical, and stylistic tests for dating Old English texts. Again and again, she concluded that the test under scrutiny failed to provide reliable evidence for this purpose.
The cost of the new skepticism, which began some years earlier with the spread of the notion that Anglo-Saxon poets composed in, or at any rate that their poems were copied in, a poetic koine that incorporated heterogeneous and therefore non-indicative dialect features, was high. No longer could we speak confidently of Beowulf as a work of the seventh or early eighth century, or of the Cynewulf canon as coming necessarily later in time than the "Caedmonian" poems. Old English poetry floated free of the Anglo-Saxon society that gave it birth. For critics who take an Olympian view of art, it may not greatly matter whether Beowulf is a seventh- or an eleventh-century poem. But most of us, living as we do in a time-bound human world, care about such things.
Fulk's riposte to Amos comes in two parts. The first is a vindication of the inductive method and of the role that probability plays in evaluating evidence that is both limited and necessarily flawed. The fact that alternative explanations can be offered for many or even all of the separate parts of a broad hypothesis does not suffice to render the hypothesis invalid. Only a more elegant hypothesis, one that will account for more of the data or for a wider variety of data, can do that. The second part rests upon an exhaustive review not just of the data supplied by earlier investigators (to which Amos largely confined herself) but of all of the evidence revealed by the new concordances of Old English prose and poetry.
The hypothesis that Fulk wishes to defend is the claim that the chronology established by the older philology can better explain the historical, phonological, metrical, syntactical, and dialectical facts than any later revisionist chronologies or refusals of chronology. Fulk employs, as representative of the established chronology, a dated list of twenty poems, from Caedmon's Hymn (657-80) to Durham (1110), drawn up by Thomas Cable in 1981.
At the heart of HOEM is Fulk's review of the arguments for dating and dialectal origins that were derived from such evidence as West Germanic parasiting, loss of intervocalic h, contraction of adjacent vowels in don and gan, compensatory lengthening after loss of h, changes in the metrical treatment of tertiary stress, and syncopation in long-stemmed verbs. To generalize very broadly, he finds that, while no one of these arguments is conclusive, nevertheless, whether taken individually or as a group, they either fail to contradict or positively reinforce the established chronology, as they do no other.
The credibility of Fulk's arguments depends crucially on the accuracy of his data. To test this, I have re-checked all the data for Beowulf. The results may be summarized as follows: Bwf 156b should be omitted from the list in [sections] 167 (p. 150), since fea is not the adjective but the dative singular of the noun feoh. In n63 (p. 201), Bwf 3007b should read 3007a. In n76 (p. 208), Bwf 2860a should read 2860b. In n4 (p. 239), in the list of verses in Beowulf ending in Pone, 2588b should read 2588a, 3018b should read 3081b, and 2959a should be added. In Appendix A (p. 408) in the paragraph concerning the diphthong in feore, Fulk says: "the verse type with a short second full lift after a resolved lift ... is exceedingly rare, with just two examples in Beowulf (2096b, 2796b ... ) . . ." The two verses cited from Beowulf are examples of a resolved second lift after an unresolved one in type C. They do not belong here.
Such a handful of mainly typographical errors is trivial in a work of this scope. There is good reason for confidence in the facts that Fulk presents. (The application of different metrical criteria [Fulk employs those of A. J. Bliss] would disambiguate four verses listed in [sections] 103 [pp. 97-99]. B w 1003a is uncontracted [a clause non-initial verse cannot be a type [A.sup.3]; 919b, 1016b, and 1984b are contracted [no type Al with anacrusis]. This would change the proportion of uncontracted to contracted forms in the table in [sections] 104 [p. 103] from 15:9 to 16:12 and the incidence from .38 to .44. Similarly, in [sections] 138 [p. 128] Bwf 706b and 967b should be listed as contracted [affecting the discussion in [sections] 145].)
An outline of one of the arguments in HOEM will serve to illustrate the rest. In chapter 6, Fulk reconsiders the implications of Kaluza's law. Briefly, in 1896 Max Kaluza observed that disyllabic words consisting of a short stem syllable followed by a vowel were treated differently by the Beowulf poet, depending on whether the final vowel was, in its reconstructed pre-historical form, long or short. The difference often shows up, for example, when the disyllabic word forms the second element of a compound. Thus, sundwudu sohte (208a) is an acceptable verse, but *sundwuda sohte would not be. Kaluza added that only a perceptible distinction between the two types of vowels in the Old English period could account for the practice of the Beowulf poet.
Bliss, who was inclined to believe that the distinction had become morphological by the historical Old English period (although he is very guarded in his judgment about this), suggested that the degree to which the distinction was observed in other poems might offer "a new criterion of the relative chronology of Old English verse." Amos would have none of it. Taking her point of departure from Bliss rather than Kaluza, she assumed that there could be no actual difference in length between "long" and "short" final vowels by the historical Old English period and thus that the distinction must have been preserved as a vestige of a decaying poetical tradition. Furthermore, since Bliss demonstrated that the distinction was observed in Beowulf in some metrical contexts but not in others, thereby lending more weight to exceptions that Kaluza had previously found, she argued that it was as likely to be a correlate of poetic style as of date, and thus worthless as an indicator of relative chronology.
Fulk reviews and corrects the data. He shows, conclusively I think, that Kaluza's law does apply in Beowulf and that neither morphological structures nor formulaic patterns could have preserved the distinction. (The "rule of the coda" [[sections] 221-54] and the 116 exceptions to that rule listed in n5 [p. 240], which are intended to cover certain classes of exceptions to Kaluza's law, are less persuasive. They mask, I suspect, an underlying regularity which remains to be discovered. In any case, they do not affect Fulk's basic argument.) The distinction must have been phonological and must have been perceptible to the Beowulf poet. This leads to a new insight. Kaluza's law provides a criterion for absolute rather than relative chronology. That is, a poet might choose to ignore the difference in vocalic length at any period of time; but any poet who used the difference as a basis for metrical patterning must have done so prior to the loss in the spoken language of that difference. In a complex chain of reasoning, Fulk argues that the loss must have occurred at least a generation before the centralization of unstressed -ae to -e. Since the latter sound change can be shown to have begun in the South and the Midlands by the middle of the eighth century and in Northumbria by the middle of the nineth century, the terminus ad quem for Beowulf is ca. 725 if the poem is Mercian in origin or ca. 825 if it is Northumbrian. On the basis of other evidence, Fulk argues that there is "a higher probability for a Mercian origin than for any other" (p. 390) for the poem, and for a probable terminus a quo of ca, 685.
This argument and the conclusion drawn from it are of first importance. They give us a point of reference about which to situate other Old English poems, and more importantly a circumscribed place and time within which to look for the culture that nourished the Beowulf poet.
HOEM is a work of immense erudition. Fulk's arguments and the data he has selected to support them are certain to be vigorously challenged. If they hold up, as I believe they are likely to, the book will stand as a landmark of Old English scholarship.
|Printer friendly Cite/link Email Feedback|
|Author:||Kendall, Calvin B.|
|Article Type:||Book Review|
|Date:||Jun 22, 1993|
|Previous Article:||'The House of Mirth' and Edith Wharton's "Beyond!"|
|Next Article:||Adam, Eve, and the 'tacen' in 'Genesis B.' (Old English 'sign')|