used with it, as in this example: John 9. 4 qimiþ nahts, panei ni manna mag waurkjan. In ON. đa without the relative particle er is uncommon.1

As to meaning and use, da de does not differ from the simple da. For data as to its distribution, see the tables and index-list.2

In this example from BH. it might be thought that the temporal particle is after pon... þe, and that the da is a demonstrative, and correlative with the da in the main clause, but the sense requires when rather than after: BH. 306. 18 Æfter þon þa þe Ceadwola was gemægenad þa geeode he... Wiht þæt ealond. In this example we have a somewhat similar case: Dial. 302. 9 ac mid by pa pe his lic was gegyred to þweanne, swa hit þeaw was æfter his fordfoðre ær man hine bebyrgde, his fet 7 his handa weron swa hale gewordene, swylce hi næfre næfdon nan dolh ænigre wunde læssan odde maran. In M. 18. 24 pe is probably a scribal error for he: pa þe þæt gerad sette, him was an broht se him sceolde tyn þusend punda. Instead of the pa þe of this example: Chron. 208. 3 pa pe se cyng Willelm p geaxode, þa bead he ut scip fyrde, MS. D. has þa þa þe. This, however, is the only instance of the kind I have noted. Thorpe's edition of the Chronicle (Rolls Series 1861) has pa be in this example, which occurs in the entry for 1013 in MS. Cott. Tiber. B. IV: pa þe he to pære byrig com, pa nolde seo buruhwaru bugan. All other MSS. have the simple da.

NOTE 1. I have found no example of da de in OE. poetry. NOTE 2. In Layamon's Brut, da de appears a few times, I quote one instance: 2. 385. 16 pa þe he wes old mon: þa com him ufel on. See also Brut 2. 50. 12. The Middle English whan that is somewhat analogous to da đe. It oc1 See Note 3, under da, p. 17.

2 p. 25.

curs somewhat rarely, and is even yet met with in archaic style; compare the translation of the passages referred to above. I quote the well-known opening lines of Chaucer's Prologue to the Canterbury Tales: Whan that Aprille with his showres soote The droghte of Marche hath perced to the roote, &c.

NOTE 3. As was stated above, the use of pa in ON. without the relative particle is unusual. This ON. pa er is precisely parallel to the OE. da de, and this, taken with the evidence from the Gothic panei, gives considerable support, I think, to the theory that the use of da as a temporal connective has been developed from da de. I quote one example from the ON. by way of illustration: En þa er Pall biskup hafði setið at stoli tolf vetr i Skalaholti, þa vard sa atburðr, at Herdis for heiman varit eptir Paska.1

4. donne de.

Only three examples of this connective have come under my notice, all of which I quote at length: Dial. 206. 26 ac forþon þe se ealda feond ponne getiho to wite 7 to yfle þa forhwyrfdan man þurh andan 7 æfæste, bonne be he sceawaþ þa godan fremian 7 weaxan to Godes wuldre; EH. 1. 48. 12 Se wisa Augustinus spræc ymbe das rædinge, and smeade hwi se halga cydere Stephanus cwæde þæt he gesawe mannes bearn standan æt Godes swyðran, and nolde cweðan Godes bearn; ponne de is geþuht wurolicor be Criste to cwedenne Godes Bearn donne mannes Bearn; BH. 350. 23 Ond he þa, þæne þe þæt unmæte sar nom 7 hefigode his scylde gewitnes,... cwæo he to pam mæssepreoste.

Donne is from the root of the demonstrative pronoun se, and the use of the relative particle with it is not, in itself, anything extraordinary. Probably donne was originally the accusative singular of the demonstrative, and probable its use as a temporal conjunction 1 Vigfusson and Powell, Icelandic Prose Reader, p. 225.

was developed through this appended de. If this is so, then we should regard such cases as a survival of an earlier syntactical usage, or as the result of a lingering feeling of the primitive force of donne. In Gothic the cognate form þan is occasionally found with the relative particle ei.1

The meaning and use of donne de are parallel to those of donne, and call for no special discussion here. I have not found any examples of this particle in the poetry, or in Middle English. The Middle English and archaic when that offers some analogy.2

NOTE 1. I quote two of the Gothic examples of the use of ðanei: M. 25. 40 jah þanei tawidedup ainamma þize minnistane bropre meinaize, mis tawidedup; Greek, iq'orov. Also M. 25. 45. These are rather causal than temporal in force. John 9.4 Ik skal waurkjan waurstwa pis sandjandins mik, unte dags ist; qimiþ nahts, panei ni manna mag waurkjan; Greek, ore. 3

5. Ŏe.

There are ten1 cases in OE. prose in which de is used as a temporal connective. Because this use is so rare, I shall quote the examples rather freely, especially since sometimes a particular case could be considered as causal.

That de is not a weakened form of da is indicated by the fact that in two of the examples it occurs with the present tense. The particle is occasionally used to introduce a causal clause or even a resultclause. I have not been able to find parallel uses of the corresponding relative particles in other Germanic languages, save that er is frequently used to introduce temporal clauses in ON. O. 2.6 Hu Sar1 Cf. under đa đe, p. 23.

2 Cf. under da de, note 2, p. 24. Cf. p. 186, 105.

8 Cf. under da đe, note 3, p. 25. Cf. Benham, The Expression of Result in Old English Prose (still unpublished).

danapalus was se sipemesta cyning in Asiria, ond hu hiene beswac Arabatus his ealdormon; 7 hu þa wifmen bysmredan hiora weras, be hie fleon woldon. On this Wülfing remarks: 'Oder sollte de relatives Fürwort sein, das durch hie ergänzt und verdeutlicht wird?' If this were the only example of de used to introduce temporal clauses in OE., it might be easier to accept this explanation; but since there are some nine other instances of a similar use, it seems unnecessary to explain it in any such way. If we include the instances in which de is used to introduce non-temporal subordinate clauses, the necessity for explaining away its use here is still further diminished. In all the cases except two, de is to be translated when. In the Laws there in found one case in which it seems to demand the rendering after and is used to translate the Latin postquam.1

O. 148. 31 to don þat he wolde þæt þa folc him by swipor to buge pe he hæfde hiera ealdhlafordes sunu on his gewealde. Here de has a causal coloring, as da often does, but the primary notion is that of time. Thorpe translates: Because he would that the people should the more readily submit to him, when he had their old lord's son in his power.

BH. 240.6 Ono pe da preo winter gefylled wæron, æfte Pendan slege þæs cyninges, þa wunnon 7 fuhton wio Osweo Mercna heretogan.

Sol. 8.5 Du þe oferswiðdest donne deað, þe þu sylf arise, and æac dest þæt ealle men arisað. Here the temporal meaning seems clear, and Hargrove translates: Thou who didst overcome death when Thou thyself didst arise, and also wilt make all men arise. CP. 73.9 Donan cymeð sio mettrymnes dæm healedum, de se wæta dara innoða astigo to dæm lime, 1 1 p. 104.

donne asuild hit and ahefegað & unwlitegað. Sweet translates Hydrocele is caused by the humours of the body collecting in the member, so that it swells and becomes heavy and disfigured; which certainly does not help us to understand the syntax of the sentence. I should punctuate differently, placing a colon after healdum, and construing donne as a demonstrative adverb correlative to de, the temporal conjunction.

CP. 85. 21 we magon beon getrymede mid Johannes cuide des godspelleres, de he cuæð. Sweet does not translate the words at all. Here the particle might be construed as being a relative, which of course it most often is. Or it might be considered a weakened form of da, especially since the expression da he cwed occurs almost hundreds of times in this text. But the fact that the Cotton MS. also has de makes against this hypothesis. I prefer to consider it a temporal conjunction, used as da so often is in this connection.

Dial. 273. 17 ic pa gyt was wuniende ealling in þam mynstre, be he me pis cydde, þæt, &c. This is one of the clearest examples I have seen, and there seems to be no doubt as to the temporal character of the de-clause.

BIH. 129. 25 ge efne eac manige hæpne men ungeleafsume oft þurh þæt to godes geleafan gecyrraþ, be hie geseod hu God pa stowe weorpap. This example also has a causal coloring; but it is secondary, and the temporal notion is the chief one.

To sum up, then, the evidence of these ten cases establishes, beyond doubt, the fact that de is used as a temporal conjunction in OE. prose. Wülfing notes only the one case in O., which he considers doubtful.