This resemblance of the names contained in the two versions might lead us to believe the Hanover MS. of Fierabras to be the original of the second part of the Sowdan, just as the Destruction, found in the same MS., is the original of the first part. But as, according to Gaston Paris, the Hanoverian version “is the same as the printed text, differing only in slight variations of readings,”68 we may suppose it likely that in all passages where the Sowdan differs from the printed Fierabras, it also differs from the Hanover MS. Nevertheless, as the differences between the Sowdan and the printed Fierabras are, on the whole, not very significant; for the several instances of omission in the Sowdan, being easily accounted for by the general plan of the poet, cannot be regarded as real variations; and as some names, the spelling of which differs in S and F, are found to be identical in S and H, we might, perhaps, be entitled to think the second part of the Sowdan to be founded on a MS. similar to the Hanover one.
It still remains for us to compare the Sowdan with the Provençal version.
In most cases where S differs from F, it also differs from P, therefore S cannot have taken those variations of readings from the Provençal poem.
The account of the knights sent on a mission to Laban, in S 1663–1738, considerably differs from the corresponding passage in P 2211 ss.
In P the scene of the whole poem is placed in Spain, there is no mention of the combat before Rome,69 as in the first part of the Sowdan.
The game of blowing a coal, S 1996 ss., is not mentioned in the Provençal version.
From these variations, taken at random out of a greater number, ‹xxxiii› it becomes evident that the Provençal poem has not been the original of the Sowdan.
If now we compare the Sowdan with Caxton’s version, which we know to be simply a translation of the French prose romance of Fierabras;70 the few following instances of differences between C and S will show at once, that also that version from which the prose romance was copied or compiled71 cannot have been the original of the Sowdan.
There are several variations in the names contained in the two versions. Thus we find Ballant in C for Laban in S; Fyerabras in C for Ferumbras in S; Garin, C 55/3 = Generyse, S 1135; Amyotte, C 176/26 = Barrokk, S 1135, &c. The game of blowing a coal is told with more details in S 1998, and somewhat differently from C 118/24; the incident of Laban’s seizing the image of Mahound and smashing it, which is related in S 2507, is omitted in C, &c.
Looking back now to our investigation concerning the original of the Sowdan, we sum up what results from it, in the following resumé:
Most probably the Destruction de Rome is the original of the first part of the Sowdan. As to the second part, we are unable to identify it with any of the extant versions. The French Fierabras, as edited by MM. Krœber and Servois, is not the original, but the differences between the two poems are not significant; apparently a version similar to the Hanover MS. may be thought to be the original.
The Sowdan is no translation, but a free reproduction of its originals; the author of the Sowdan following his sources only as far as concerns the course of the principal events, but going his own independent way in arranging the subject-matter as well as in many minor points.
The Sowdan differs from the poem of Syr Ferumbras in two principal points:
(1) In being an original work, not in the conception, but in the treatment of the subject-matter, whereas the Ashmole Ferumbras is little more than a mere translation. ‹xxxiv›
(2) In representing, in its first portion, the first part of the old Balan romance, whereas Syr Ferumbras contains only the second. But as that second part of the old Balan romance appears to be considerably modified and greatly amplified in the Ashmole Ferumbras, so the first part of the Sowdan contains a likewise modified, but much shortened, narration of the first part of the old Balan poem, so that the Sowdan has arrived to become quite a different work from the original Balan or Fierabras romance, and that a reconstruction of the contents of that old poem would be impossible from the Sowdan.
LANGUAGE AND SUMMARY OF GRAMMATICAL FORMS. ◊
AS regards the language of the Sowdan, the first point is the dialect. Looking at the plurals of the present indicative in -en or -n, we at once detect the Midland peculiarities of the poem. Thus we find, l. 1331, gone rhyming with one, l. 1010, goon : camalyon, l. 506, gone : than, l. 1762, lyven : gyfen, l. 1816, byleven : even.
The verbal forms of the singular present indicative and of the second person sing. preterite of weak verbs lead us to assign this poem to an East-Midland writer. The 2nd and 3rd person singular present indicative end in -est, -eth; and the 2nd person sing. preterite of weak verbs exhibits the inflection -est: l. 1202, goist : moost; 1314, 1715, knowest; 1344, trowest; 1154, blowest; 1153, saiest; 2292, forgetist; 560, doist; 1193, doistowe;—1093, goth : wroth, 1609 : loth, 1620 : doth; 1728, sleith : deth; 561, sholdest; 1244, shuldist; 603, madist; 563, hadist; 2219, askapedist, &c.—Twice we find the 2nd person preterite without -est (made, wroght); but see the note to l. 2.
If, now, we examine the phonological and inflectional peculiarities of the Sowdan, we find them thoroughly agreeing with those of other East-Midland works,72 which still further confirms the supposition of the East-Midland origin of the poem. ‹xxxv›
I or y, the descendants of original u (which in Old English [Anglo-Saxon] had already become y or i in consequence of i- mutation or umlaut)—are found rhyming with original i:—ll. 449, 881, kyn : him, 2060 : wynne; 1657, fille : stille; 1973, fire : desire, &c. It must, however, be noted that the rhyme king : inne (l. 372) or king : thing (ll. 173, 236) cannot be regarded as an East-Midland peculiarity, because king, drihten, chikken, the i of which is a modification of original u, are to be met with in all Middle-English dialects, as has been shown by Professor Zupitza in the Anzeiger für deutsches Altertum, vol. vi. p. 6.
Old English short a, which is liable to change into o, appears in this poem—
(1) always as o, before n- combinations (nd, nt, ng):—531, stronge : istonge; 3166, bronte : fonte; 214, amonge : longe, &c.
(2) as a, before the single consonants m and n:—1120, name : shame, 935 : same, 1739 : grame; 785, 1773, man : Lavan; 3125, came : Lavan (cf. 2579, Lavan : tane); 2160, came : dame, &c.—The fact that com (ll. 547, 1395, 3095, &c.) is used as well as cam as sing. preterite indic. need occasion no difficulty if we remember that the original short a (or o) of cam (or com) had already been lengthened into ô in the O.E. period.73 Came and come as pret. sing. are employed indifferently in Chaucer as well as in