The Functions of Old English GeweorVan 265 the thing meets us in the phrase geweorfian latan, well known from Beow. 1996 f.: (pat fiu . . .) lete Sufi-Dene sylfe geweorfian / gufie wifi Grendel, ' that you should let the Danes themselves agree as to the fight . . .,' or '. . . decide about . . .,' and then translating decision into action 'attend to [settle] the fighting. ' 37 Strictly speaking, the accusative Dene sylfe should be considered to depend on geweorfian. If the emphasis was placed on the person (cf. sylfe, 'let them do it'), the meaning could easily shade off into 'leave it to them,' further by the omission of the object which was originally expressed by the genitive into 'do not interfere with them,' 'let them alone.' This development is so natural and has been carried out with such consistency that the failure to adopt the above explanation can be accounted for only by preconceived notions concerning certain modern phrasal combinations. Cosijn in his famous Aanteekeningen op den Beowulf, p. 30 cites as a parallel, the Middle Dutch phrase laten gewerden (geworden), in which gewerden (according to Verdam) denotes 'begaan, ' 'te werk gaan,' and at the same time pronounces emphatically against Bugge's illuminating interpretation (Beitr. xii, 97). Cosijn is followed by Hubbard, who, in addition, brings forward interest- ing examples of the analogous combination in Middle English. Middle High German, and Middle Dutch. However, neither Cosijn nor Hubbard commits himself very definitely on the construction of the Beowulfian passage. Should we translate (with Hubbard): 'that you let the Danes alone in their war against Grendel'? This looks like a make- shift, which is hardly in agreement with Old English usage. Or should we follow Shipley (The Genitive Case in Anglo-Saxon Poetry, p. 43, cf. also Hubbard): 'let the contest with Grendel be for the Danes'? In the light of the material presented in our second division it seems to me pretty clear that gufie is the normal genitive, and that Bugge's derivation of the expression 87 wiZ is perhaps to be connected with (Dene sylfe geweorfian) gude. It is quite possible, however, that wid (denoting association, dealing) is to be construed directly with geweorVan ('agree'); in that case, the last example under 'a.' and the first one under *f.' (hi our second division) could be referred to as partial parallels; cf. also footnote 24 and the M. Du. and Ger. examples with met, mil, below. Lines 424b-26a show a similarity of meaning, though
not of construction.
