Chapter XIX: Section II: The Dialect, Syntax, and Metre of "beowulf" as Evidence of Its

LITERARY HISTORY.

A full discussion of the dialect, metre and syntax of _Beowulf_ forms no part of the scheme of this study. It is only intended in this section to see how far such investigations throw light upon the literary history of the poem.

_Dialect._

_Beowulf_ is written in the late West Saxon dialect. Imbedded in the poem, however, are a large number of forms, concerning which this at least can be said--that they are not normal late West Saxon. Critics have classified these forms, and have drawn conclusions from them as to the history of the poem: arguing from sporadic "Mercian" and "Kentish" forms that _Beowulf_ is of Mercian origin and has passed through the hands of a Kentish transcriber.

But, in fact, the evidence as to Old English dialects is more scanty and more conflicting than philologists have always been willing to admit. It is exceedingly difficult to say with any certainty what forms are "Mercian" and what "Kentish." Having run such forms to earth, it is still more difficult to say what arguments are to be drawn from their _occasional_ {105} appearance in any text. Men from widely different parts of the country would be working together in the scriptorium of one and the same monastery, and this fact alone may have often led to confusion in the dialectal forms of works transcribed.

A thorough investigation of the significance of all the abnormal forms in _Beowulf_ has still to be made. Whether it would repay the labour of the investigator may well be questioned. In the meantime we may accept the view that the poem was in all probability originally written in some non-West-Saxon dialect, and most probably in an Anglian dialect, since this is confirmed by the way in which the Anglian hero Offa is dragged into the story.

Ten Brink's attempt to decide the dialect and transmission of _Beowulf_
will be found in his _Beowulf_, pp. 237-241: he notes the difficulty
that the "Kentish" forms from which he argues are nearly all such as
occur also sporadically in West Saxon texts. A classification of the
forms by P. G. Thomas will be found in the _Modern Language Review_, I,
202 _etc._ How difficult and uncertain all classification must be has
been shown by Frederick Tupper (_Pub. Mod. Lang. Assoc. Amer._ XXVI,
235 _etc._; _J.E.G.P._ XI, 82-9).

"_Lichtenheld's Test._"

Somewhat more definite results can be drawn from certain syntactical usages. There can be no doubt that as time went on, the use of _se_, _s[=e]o_, _thaet_ became more and more common in O.E. verse. This is largely due to the fact that in the older poems the _weak adjective_ + _noun_ appears frequently where we should now use the definite article: _w[=i]sa fengel_--"the wise prince"; _se w[=i]sa fengel_ is used where some demonstrative is needed--"that wise prince." Later, however, _se_, _s[=e]o_, _thaet_ comes to be used in the common and vague sense in which the definite article is used in Modern English.

We consequently get with increasing frequency the use of the _definite article_ + _weak adjective_ + _noun_: whilst the usage _weak adjective_ + _noun_ decreases. Some rough criterion of date can thus be obtained by an examination of a poet's usage in this particular. Of course it would be absurd--as has been done--to group Old English poems in a strict chronological order according to the proportion of forms with and without the article. Individual usage must count for a good deal: {106} also the scribes in copying and recopying our text must to a considerable extent have obliterated the earlier practice. Metre and syntax combine to make it probable that, in line 9 of our poem, the scribe has inserted the unnecessary article _th[=a]ra_ before _ymbsittendra_: and in the rare cases where we have an O.E. poem preserved in two texts, a comparison proves that the scribe has occasionally interpolated an article. But this later tendency to level out the peculiarity only makes it the more remarkable that we should find such great differences between O.E. poems, all of them extant in copies transcribed about the year 1000.

How great is the difference between the usage of _Beowulf_ and that of the great body of Old English poetry will be clear from the following statistics.

The proportion of phrases containing the weak adjective + noun with and without the definite article in the certain works of Cynewulf is as follows[234]:

With article Without article
_Juliana_ 27 3
_Christ (II)_ 28 3
_Elene_ 66 9

In _Guthlac_ (A) (c. 750) the proportions are:

With article Without article
_Guthlac_ (A) 42 6

Contrast this with the proportion in our poem:

With article Without article
_Beowulf_ 13 65

The nearest approach to the proportions of _Beowulf_ is in the (certainly very archaic)

With article Without article
_Exodus_ 10 14

On the other hand, certain late texts show how fallible this criterion is. Anyone dating _Maldon_ solely by "Lichtenheld's Test" would assuredly place it much earlier than 991.

{107}

It is easy to make a false use of grammatical statistics: and this test should only be applied with the greatest caution. But the difference between _Beowulf_ and the works of Cynewulf is too striking to be overlooked. In _Beowulf_, to every five examples without the article (e.g. _headho-st[=e]apa helm_) we have _one_ with the article (e.g. _se hearda helm_): in Cynewulf to every five examples without the article we have _forty_ with it.

A further test of antiquity is in the use of the weak adjective with the instrumental--a use which rapidly diminishes.

There are eighteen such instrumental phrases in _Beowulf_ (3182 lines)[235]. In _Exodus_ (589 lines) there are six examples[236]--proportionally more than in _Beowulf_. In Cynewulf's undoubted works (c. 2478 lines) there is one example only, _beorhtan reorde_[237].

This criterion of the absence of the definite article before the weak
adjective is often referred to as Lichtenheld's Test (see article by
him in _Z.f.d.A._ XVI, 325 _etc._). It has been applied to the whole
body of O.E. poetry by Barnouw (_Textcritische Untersuchungen_, 1902).
The data collected by Barnouw are most valuable, but we must be
cautious in the conclusions we draw, as is shown by Sarrazin (_Eng.
Stud._ XXXVIII, 145 _etc._), and Tupper (_Pub. Mod. Lang. Assoc._ XXVI,
274).

Exact enumeration of instances is difficult. For example, Lichtenheld
gave 22 instances of definite article + weak adjective + noun in
_Beowulf_[238]. But eight of these are not quite certain; _se g[=o]da
m[=ae]g Hygel[=a]ces_ may be not "the good kinsman of Hygelac," but
"the good one--the kinsman of Hygelac," for there is the half line
pause after _g[=o]da_. These eight examples therefore should be
deducted[239]. One instance, though practically certain, is the result
of conjectural emendation[240]. Of the remaining thirteen[241] three
are variations of the same phrase.

The statistics given above are those of Brandl (_Sitzungsberichte d. k.
Preuss. Akad. d. Wissenschaften_, 1905, p. 719) which are based upon
those of Barnouw.

"_Morsbach's Test._"

Sievers' theories as to O.E. metre have not been accepted by all scholars in their entirety. But the statistics which he {108} collected enable us to say, with absolute certainty, that some given types of verse were not acceptable to the ear of an Old English bard.

Sceptics may emphasize the fact that Old English texts are uncertain, that nearly all poems are extant in one MS only, that the MS in each case was written down long after the poems were composed, and that precise verbal accuracy is therefore not to be expected[242]. All the more remarkable then becomes the fact, for it is a fact, that there are certain types of line which never occur in _Beowulf_, and that there are other types which are exceedingly rare. Again, there are certain types of line which _do_ occur in _Beowulf_ as we have it, though they seem contrary to the principles of O.E. scansion. When we find that such lines consistently contain some word which had a different metrical value when our extant MS of _Beowulf_ was transcribed, from that which it had at the earlier date when _Beowulf_ was composed, and that the earlier value makes the line metrical, the conclusion is obvious. _Beowulf_ must have been composed at a time or in a dialect when the earlier metrical values held good.

But we reach a certain date beyond which, if we put the language back into its older form, it will no longer fit into the metrical structure. For example, words like _fl[=o]d_, _feld_, _eard_ were originally "u-nouns": with nom. and acc. sing. _fl[=o]du_, etc. But the half-line _ofer fealone fl[=o]d_ (1950) becomes exceedingly difficult if we put it in the form _ofer fealone fl[=o]du_[243]: the half-line _f[=i]felcynnes eard_ becomes absolutely impossible in the form _f[=i]felcynnes eardu_[244].

It can, consequently, with some certainty be argued that these half-lines were composed after the time when _fl[=o]du_, _eardu_ had become _fl[=o]d_, _eard_. Therefore, it has been further argued, _Beowulf_ was composed after that date. But are we justified in this further step--in assuming that because a certain number of half-lines in _Beowulf_ must have been composed after a certain date, therefore _Beowulf_ itself must have been composed after that date?

{109}

From what we know of the mechanical way in which the Old English _scribe_ worked, we have no reason to suppose that he would have consistently altered what he found in an older copy, so as to make it metrical according to the later speech into which he was transcribing it. But if we go back to a time when poems were committed to memory by a _scop_, skilled in the laws of O.E. metre, the matter is very different. A written poem may be copied word for word, even though the spelling is at the same time modernized, but it is obvious that a poem preserved orally will be altered slightly from time to time, if the language in which it is written is undergoing changes which make the poem no longer metrically correct.

Imagine the state of things at the period when final _u_ was being lost after a long syllable. This loss of a syllable would make a large number of the half-lines and formulas in the old poetry unmetrical. Are we to suppose that the whole of O.E. poetry was at once scrapped, and entirely new poems composed to fit in with the new sound laws? Surely not; old formulas would be recast, old lines modified where they needed it, but the old poetry would go on[245], with these minor verbal changes adapting it to the new order of things. We can see this taking place, to a limited extent, in the transcripts of Middle English poems. In the transmission of poems by word of mouth it would surely take place to such an extent as to baffle later investigation[246].

Consequently I am inclined to agree that this test is hardly final except "on the assumption that the poems were written down from the very beginning[247]." And we are clearly not justified in making any such assumption. A small number of such lines would accordingly give, not so much a means of fixing a period before which _Beowulf_ cannot have been composed, as merely {110} one before which _Beowulf_ cannot have been fixed by writing in its present form.

If, however, more elaborate investigation were to show that the _percentage_ of such lines is _just as great_ in _Beowulf_ as it is in poems certainly written after the sound changes had taken place, it might be conceded that the test was a valid one, and that it proved _Beowulf_ to have been written after these sound changes occurred.

This would then bring us to our second difficulty. At what date exactly did these sound changes take place? The chief documents available are the proper names in Bede's History, and in certain Latin charters, the glosses, and a few early runic inscriptions. Most important, although very scanty, are the charters, since they bear a date. With these we proceed to investigate:

A. The dropping of the _u_ after a long accented syllable (_fl['=o]du_ becoming _fl['=o]d_), or semi-accented syllable (_St['=a]nf[`o]rdu_ becoming _St['=a]nf[`o]rd_).

There is evidence from an Essex charter that this was already lost in 692 or 693 (_uuidmundesfelt_)[248]. From this date on, examples without the _u_ are forthcoming in increasing number[249]. One certain example only has been claimed for the preservation of _u_. In the runic inscription on the "Franks casket" _flodu_ is found for _flod_. But the spelling of the Franks casket is erratic: for example _giutheasu_ is also found for _giutheas_, "the Jews." Now _u_ here is impossible[250], and we must conclude perhaps that the inscriber of the runes intended to write _giuthea su[mae]_[250] or _giuthea su[na]_[251], "some of the Jews," "the sons of the Jews," and that having reached the end of his line at _u_, he neglected to complete the word: or else perhaps that he wrote _giutheas_ and having some additional space added a _u_ at the end of his line, just for fun. Whichever explanation we {111} adopt, it will apply to _flodu_, which equally comes at the end of a line, and the _u_ of which may equally have been part of some following word which was never completed[252].

Other linguistic data of the Franks casket would lead us to place it somewhere in the first half of the eighth century, and we should hardly expect to find _u_ preserved as late as this[253]. For we have seen that by 693 the _u_ was already lost after a subordinate accent in the Essex charter. Yet it is arguable that the _u_ was retained later after a long accented syllable (_fl[=['o]]du_) than after a subordinate accent (_uu[=['i]]dm[`u]ndesf[`e]lt_); and, besides, the casket is Northumbrian, and the sound changes need not have been simultaneous all over the country.

We cannot but feel that the evidence is pitifully scanty. All we can say is that _perhaps_ the _flodu_ of the Franks casket shows that _u_ was still preserved after a fully accented syllable as late as 700. But the _u_ in _flodu_ may be a deliberate archaism on the part of the writer, may be a local dialectal survival, may be a mere miswriting.

B. The preservation of _h_ between consonant and vowel.

Here there is one clear example which we can date: the archaic spelling of the proper name _Welhisc_. _Signum manus uelhisci_ occurs in a Kentish charter of 679[254]. The same charter shows _h_ already lost between vowels: _uuestan ae_ (_ae_ dative of _[=e]a_, "river," cf. Gothic _ahwa_).

Not much can be argued from the proper name _Welhisc_, as to the current pronunciation in Kent in 679, for an old man may well have continued to spell his name as it was spelt when he was a child, even though the current pronunciation had changed[255]. But we have further evidence in the glosses, which show _h_ sometimes preserved and sometimes not. These glosses are mechanical copies of an original which was presumably compiled between 680 and 720. We are therefore justified in arguing that at that date _h_ was still preserved, at any rate occasionally.

{112}

Of "Morsbach's test" we can then say that it establishes something of an argument that _Beowulf_ was composed after the date when final _u_ after a long syllable, or _h_ between consonant and vowel, were lost, and that this date was probably within a generation or so of the year 700 A.D. But there are too many uncertain contingencies involved to make the test at all a conclusive one.

Morsbach's _Zur Datierung des Beowulf-epos_ will be found in the
Goettingen _Nachrichten_, 1906, pp. 252-77. These tests have been
worked out for the whole body of Old English poetry in the
_Chronologische Studien_ of Carl Richter, Halle, 1910.

* * * * *