changes made for the sake of avoiding awkwardness or increasing the smoothness of the line. What can possibly justify the omission of the second he in line 1423: For he waytez on wyde, his wenches he byholdes'? The metrical scheme of the second half-line is exactly the same as that in the line following- aboute bi pe wozes.'

Other emendations, due to fancied difficulties, are equally unnecessary. 'pe moste mountaynez on mor þenne watz no more dryze (385), Gollancz alters to 'pe moste mountaynez on mor þenne [on] more dry3e,' which he interprets the highest mountains on the moor then alone were more dry, i. e. less submerged than the rest,' the reason for the change being the fact that the next line declares & peron flokked pe folke.' But one may easily interpret the line as it stands to mean not that the mountains were completely covered, but simply that the waters of the flood were creeping upon even the highest of them; the floating comparative more dryze' of the emendation makes more difficulty than the manuscript reading.

6

8

In attempting to restore the alliteration by means of emendation, Gollancz has made many needless changes and suggested others, because he failed to take into account the poet's use of double alliteration, especially of the type a a b b, a common characteristic of Middle English alliterative verse, and his practise of permitting an unstressed syllable to take the alliteration.9

In line 345: "Now Noe," quop oure lorde, " art pou al redy?" Gollancz suggests lede for al, and similarly in line 1304: '& Nabugo de Nozar makes much joye,' he suggests nouthe for much; but the manuscript is justified in each case by the same form of alliteration (a a bb) in line 299: Sem soply þat on, þat oper hyst Cam,' and line 1573: 'out-taken bare two, & penne he pe prydde.' In view of these cases, another tempting emendation leue to beue

Here I should include vponande for vpon (318); þat for þer (432); wyth for þat (594); I for & (917); so inserted after for (1057); hatz for is (1143); hem inserted after spylt (1220); he inserted after hade (1336); per for þat (1532); he inserted before cluchches, bend' (1541); his for bat (1811).

The following changes are unnecessary for various reasons: forperde for forferde (1051); prystyly for pryuyly (1107); plit strange for plit stronge, 'great sin' (1494; see NED. for both meanings).

K. Schumacher, Studien über den Stabreim in der me. Alliterationsdichtung (Bonn, 1914), p. 27.

(1622) is unwise, since the alliteration may be: Baltazar vmbebrayde hym, & "Leue sir," he sayde.' Line 745: 'pen Abraham obeched hym & lo3ly him þonkkez,' suffers an extraordinary change at the editor's hands, becoming 'þen [þe burne] obeched hym & [b]03[som]ly him ponkkez.' But even if obeched is obviously the alliterating word,' as Gollancz says, the alliteration may be on the unstressed o, not on the b, and by changing lo3ly to hezly (compare Gawain 773), surely a much less violent emendation, we have a line with regular vowel-alliteration.10

Since Professor Gollancz's own suggestions for the emendation of the text are so numerous and so important, it is singularly unfortunate that his editorial method does not permit him to give proper acknowledgment to the many scholars who have preceded him in endeavoring to interpret the difficulties of the text, and to distinguish between those emendations which have been proposed in print by others, and those appearing for the first time in his edition. Aside from some thirty obvious corrections of the manuscript which every editor or commentator has made mechanically, there are fifty-six emendations adopted in the present text which had been previously proposed by other students of the poem. Of these only nineteen are attributed to those who first proposed them in print, and though they may have occurred to the present editor independently, this hardly excuses his failure at least to mention the fact that he has been anticipated."1

Similarly, even a desire to avoid controversy and condense as much as possible hardly excuses the careless and misleading way in which Professor Gollancz employs the phrase 'hitherto unexplained.' It is hardly fair, for example, to use the expression in connection with an interpretation of the difficult lines 433-4 which differs somewhat from four previously suggested (see my edition).

10 Vowels of course alliterate freely with h before vowels; cf. lines 11, 14. "For example, Morris changed stystez pat myz to styntez pat nyzt (359); Bateson proposed forþering for forering (3), and heryed for heyred (1527); Emerson proposed bekyr ande bol[l]e (1474); sanctorum [per] sopefast (1491). As my own edition came into Professor Gollancz's hands only after part of his was already in type, he can hardly be blamed for not mentioning the fact that I anticipated him in adopting heven for her even (50); so wer for sower (69); murnande (I read mornande) for wepande (778); þer for pen (926); nomen (Emerson, too, proposed nomon) for no mon (1002); smelle for synne (1019).

So, after Professor Emerson (and I, independently) had given an explanation of the word ungoderly (145), it is disconcerting to find what seems to me an extremely fantastic etymology proposed in this manner: This word, hitherto unexplained, seems an Englishing of "boner" (i. e. bonaire, a (i. e. bonaire, a common ME. form for deboner, debonaire) = well-bred, with the negative prefix="de mal aire," ill-mannered. The -ly suffix was due to analogy with "ungodly." Finally, it should be noted that for olipraunce (1349) of hitherto unknown origin,' the editor elaborately presents an etymology (the French name Olibrius) proposed as long ago as 1890 by Henry Bradley (Academy, January 11).

6

6

The notes contain many ingenious and valuable explanations of obscure words and phrases, though the etymologies suggested are sometimes far-fetched. Jumpred,? confusion' (491) is more likely to be connected with Chaucer's verb jumpre (Troilus 2. 1037) than with jumper, 'bore with a jumper,' a technical meaning which has every appearance of being modern. For þat schewe me schale in po schyre howsez' (553) is paraphrased 'Because it, i. e. a speck or spot, is shunned in those radiant mansions,' schewe being considered aphetic for eschewe, avoid.' But this is seeking trouble; schewe means appear' and taking me as one,' the line may be translated 'In order that one shall appear,' etc. Gollancz's defense of the manuscript hokyllen (1267, I emended to he kyllen), ashockle, cut down (like grass),' now seems to me right. 'Stepe stayred stones' (1396), usually translated brightly gleamed jewels,' cannot mean 'ascended the staired stones,' because step is not used in this sense without a preposition. Gollancz's connection of umbepour (1384) with umbethourid, which occurs twice in The Wars of Alexander (3857, 4806) is probably correct, though the line is still obscure. I suggest the possibility of putting a comma after prowen, and paraphrasing lines 1383-4 Pinnacled towers at intervals, the length of twenty spears apart, and (even) more thickly crowded, surrounded by a paling set crosswise.' That med (1391) is related to OE. gemet, measure,' or mode (1635) to ON. mōt, 'stamp,' or 'mark,' seems to me semantically improbable and phonologically impossible. In several instances Gollancz assumes that final y is the equivalent of e-skyly (62), mayny-molde (514); clyvy (1692), which he derives from OE. clife); but if this is true, it is possibly only a scribal error, as in

each case a y is found in the syllable preceding.12 One of the most ingenious suggestions in the book is p'ydres bowls, vessels,' Vulgate hydria, for the difficult pede (1717). Here, as elsewhere, Gollancz is careful to state precisely how the corruption of the text may have come about.

This review has naturally emphasized the points in which I dissent from Professor Gollancz, but I have not undertaken to discuss the differences, which are many, in the general plan and scope of his edition and my own. In spite of the fact that Professor Gollancz has solved, or at least brought us nearer the solution of, many cruces in Purity, fascinating problems of textual interpretation still remain. It is to be hoped that the appearance of two new editions of a poem long unduly neglected will direct attention not only to these problems, but to the importance of Purity in its relationship to the other poems of the alliterative school.

ROBERT J. MENNER.

Yale University.

Das dichterische Kunstwerk. Grundbegriffe der Urteilsbildung in der Literaturgeschichte von E. ERMATINGER. Leipzig und Berlin, B. G. Teubner, 1921. viii+ 405 pp.

What principles shall guide us in estimating the poet and his work? Such is the main question proposed in this book. As between two prevalent methods, the one historical, objective, and descriptive (tending to formalism), the other subjective and philosophical (tending to caprice), the author believes in der Mitten liegt holdes Bescheiden. While the critic should not be dominated by an ideal of abstract verity (unattainable anyway), he must have a sense of responsibility, appealing to his scientific and his social conscience.

The first distinction is between Welt and Ich. By Welt is meant not Ding an sich (excluded from the discussion as unknowable), but a sort of Gesamtich, a conventionalized ego, formed by tradi

12 But Emerson (Publ. Mod. Lang. Assoc., xxxvii, 58 f.) cites many examples from this manuscript which seem to show that this representation of final unstressed e reflects confusion in the language itself.

tion. The varying conflict between these two forces is termed Erleben, and is the source of the poet's dynamic and vital idea, his Weltanschauung, which expresses itself in symbolic forms, his particular works. Thus at the outset naturalism and impressionism are condemned on principle, and the creative sovereignty of the poet is asserted.

The interplay of these forces (Ich, as feeling, sensation, will; and convention, as understanding, roughly speaking) results in imagination, which is a creator of new values. And the intensity of the conflict measures the creative power of any individual. Also the varying participation and energy of these elements in the struggle seem to the author to furnish fundamental distinctions between epic, lyric, and dramatic poets. He too readily finds agreement of certain examples with his theoretical views, and does not avoid the seductive fallacy of reasoning from a particular to a general. Mörike was indeed passive (though his writing poetry in bed does not prove it!), but Goethe, greatest of lyric poets, was not. And certainly the author should be more sceptical of his own reasoning than to commit himself to the assertion (p. 25), that, compared with epic poets, dramatists, because of the intenser conflict in them between Ich and Welt, seldom grow old. Virgil lived to be 51, Dante 56, Racine 60, Aeschylus 69 (and died by accident), Euripides 74, Corneille 78, Ibsen 78, Calderon 81, Grillparzer 81, Sophocles 90. In the same way Keller as calm, cool materialist, is made too exclusively the type of the epic poet. What about Dante? Dostoevsky? Accidental personal qualities are here confounded with the characteristics of genre.

More convincing is the discussion of the creative experience (künstlerisches Erleben). Philosophy is considered an aid to the poet in finding himself, striking examples being Keller, Kleist, and Hölderlin. Rejecting the milieu theory as impersonal and uncreative, the author affirms the unique and dynamic force of the creative spirit, and proceeds to set up the reach of experience, its intensity, and its degree of faith (as opposed to Wissen) as proper tests of genius. Goethe's reach, for example, is large, Storm's narrow. Schiller's dynamic ideas are limited chiefly to the conflict between realism and idealism, and no idea at all is manifest in his later works, from Maria Stuart on (p. 116), these plays being condemned as technical feats. (This view, for Maria