US WURK XXX (1981), p. 49

[0580] AND DIALECTS is a peculiar fact each of the three Old English dialects, Anglian, West Saxon and Kentish, has, by different scholars, been considered more closely connected with old Frisian than has either of the remaining dialects.

Ferd. Wachter, e.g., finds the resemblance between Anglian and Frisian especially striking, something which attributes to Frisian settlement in Northern (1850:417, 420-21). But while Wachter has mainly the Northumbrian US WURK XXX (1981), p. 50 branch of Anglian in mind, .W. Skeat (1887:32, 77; 1896: 221) believes that Mercian is the OE dialect or subdialect to show the greatest similarity to Frisian. Both Mercian and Frisian take up intermediate positions, the "between the Anglian and the Saxon," the other "between and Holland." 1 As Mercian/Frisian agreements Skeat lists the absence of breaking in al(l), bald and half, the undiphthongized e of Merc. werc , werpan, Fris. werk, werpa, and the ē of dēd(e) ( dād) and of Merc. geh ēran, Fris. hēra (Goth. hausjan), cf. below. Unlike Wachter, .. Clement thinks that the special links of Anglian to Frisian date back to a time when and were neighbours in Schleswig (1862:17-19), and another scholar to advocate special Friso-Anglian relations is Otto Bremer (1900:843), cites the shared absence of front diphthongization after c, g, sc (cf. below) in support of his view. Carl Karstien (1939:12) tries to illustrate a similar by suggesting a further parallel, viz. West Germanic ā > OFris./Angl. ē. Finally, Wolfgang Jungandreas sees a particularly close connection between Anglian and Old Frisian, but does not provide any evidence ( 1949:60). However, Jungandreas (III 1949:24) also links Frisian (esp. Old West Frisian) to Kentish, and among the five shared innovations cited, three have been taken over from Theodor Siebs (1901:1157-8; 1930:65, 71-2): ū + i, j > ē, eu > ia and ēg > ēi. The inference made by both Jungandreas and Siebs is that such parallels must have arisen prior to the departure of the from the Continent. 2 In his lexical investigation of Anglian Richard Jordan does not find any special links connecting that dialect with Frisian.3 What he does find, though, is a number of correspondences between West Saxon on the one hand and OFris. and Old Saxon on the other (cf. 1906:120, 122). In Otto Bremer's view there is an especially close relationship between WS and (Old) Insular North Frisian 4 in that both dialect areas appear to share front diphthongization after palatal consonants (1900:848), but cf. below. Having thus, at least in part, uncovered the diversity of scholarly opinion, shall proceed to survey the parallels between Frisian and Angl., Kt. or WS. The evidence presented repeats items found in Nielsen 1981:. III.ix 5, though sometimes in an abbreviated or altered form. The reasons for basing the survey on morphological and phonological points are given in Nielsen 1981:73-4. 6 1. The Angl. nom.pl.masc. n-stem forms oexen, exen closely correspond to OFris. ixen and West Norse yxn , øxn, seeing that all the forms listed have undergone i-mutation, cf. Nielsen 1975:7. In Brunner's view (1965:§276 Anm.1, §277Anm.1), the Indo-European *-en -grade (Germanic *-in(iz) ) of the thematic element may be held responsible for this. However, there is not much support for US WURK XXX (1981), p. 51

IE *-en - in the paradigm, but another explanation is possible: the vanishing grade of the thematic element, which is to be found in gp. oxna , yxna , Gothic aúhsn ē and OE oxna , cf. Nielsen 1979: No. 5, may have been extended to the nominative, resulting in np. * oXsniz (* uXsniz ). The fact that a similar np. form is not attested in Goth., is likely to be a coincidence; the vanishing grade is well established in the n-stem paradigm of this (Krahe 1967:90-91). 2. A vernerized superlative form of ‘little’ is attested in OFris. ( lērest ) as an alternative to lēst . Among the OE dialects a similar form is recorded only in Kt.: : læ resta (once). 3. It is a characteristic feature of WS/Kt.and OFris. (OS, ) that have uniform accusative/dative forms in 1/2. personal : WS/Kt. mē, ūs ; þē, ēow ; OFris. (/OS) mī, ūs th ī, iu (MDu. mi , ons ; di , (j)u ). As pointed out in Nielsen 1981:225-7, early Gmc. had uniform a/dp. forms, while the two cases were differentiated in the singular. The ousting of acc. sg. forms in -k by the datives is therefore a common innovation. 4. None of the OE dialects exhibit forms with an initial s- in n/asf. and n/ap. of the 3rd pers.pron. In the south-eastern dialect of such forms do occur, however, perhaps in consequence of contacts between and the Continent, cf. OFris. se , MDu. si (Brunner I 1960:80-81, Steller 1928:§83 and Franck 1910:§210). See also Århammar 1966 (1968b):62-3, 72. 5. In some Nhm. texts asm.dem.pron. þene occurs beside the normal form, þone . Streitberg (1974:269b) thinks that the -e-form reflects the suffix * -in-ōn (i- mutation), cf. OE ænne *ainin ōn. Another explanation is possible, however. Very likely, the asm. was originally -a- in all Gmc. dialects, cf. Nielsen 1979:No.10. On the analogy of gen.sg., -e- was extended into asm. in ( den ), and from here it may have spread northwards, cf. OS thena , thana ; OFris. thene and MDu. dien . What was later to become Nhm., may have been affected by this development. 6. WS/Kt. and OFris. exhibit -e in the 1st pres.sg.ind. of the (strong) verbs instead of the regular reflex of IE *-ō, which we find in Goth. baíra , ON ber , Runic w r i t u, OS/OHG biru and Angl. beoru , bindo. WS/Kt. -e has been explained as (1) an early weakening of –u/-o ( Krahe II 1969:§69), (2) an optative ending (Hirt II 1932:§112) and (3) a reflex of Gmc. *-ōm, (-m added on the analogy of secondary endings, cf. Campbell 1959:§731). Any of these explanations may apply to OFris as well. 7 However, it is very doubtful whether this correspondence between the US WURK XXX (1981), p. 52 southern OE dialects and OFris. developed prior to the -Saxon emigration: in early WS (once) and Kt. -o (-u) forms occur (Campbell: 1959:§735 a). 7. Syncopated forms in the 2/3.pres.sg.ind. of the strong verbs (and Class I weak verbs) are characteristic of OFris. (binst, bint) and, with some vacillation, of WS and Kt., whereas unsyncopated forms are retained in Angl.: bindes, binde . According to Walde (1900:125Anm.), the origin of the shortened forms is to be sought in the inverted expressions bindis þu, bindi he; the generalization in WS/Kt. did not take place until perhaps the late ninth century, and uncontracted endings must be assumed for all OE dialects in the preceding centuries, cf. Sisam 1953:125-6, Brunner 1965:§358. 8. In the pt.pl.. of ‘do’ Angl. has dedon besides dydon, and in the inscription on Codex Aureus deodon crops up, the accented vowel of which Campbell (1959:§768b) to posit short vowel quantity (back mutation). (For a different view, see Brunner 1965:§429 Anm. l.) Kt. dede, dedon may represent exact counterparts of WS dyde, dydon. On the Continent -e-forms are attested in the sg. of OS/OHG (deda/teta) and in the OS pl. (d ēdun). If the first vowel in Angl. dedon is long, it would correspond to ā in OS d ādun, OHG tātun. The vowel quantity of e in OFris. dede, deden has been subjected to different interpretations, cf. v. Helten 1890:§310 and Steller 1928:§107. Siebs (1901:1333) thinks that the vowel is long (= Gmc. ē1) because modern dialectal forms presuppose length. No matter how the accented vowel of Angl. dedon is interpreted, it links up with the vowel of at least some of the OS/OHG/OFris. forms. In , the WS --forms (dyde, dydon) stand completely apart, and are due to innovation (Beck 1963:16). 9. Reflexes of *waljan in the paradigms of the verb ‘will’ are expecially frequent in OHG, but Angl., OFris. and OS also exhibit quite a few instances (-e-forms), cf. Nielsen 1979: No.18. In WS and Kt., on the other hand, -i-forms are practically universal. 10. The reflex of Gmc. ēl (IE ē) in Goth., OFris. and Angl.Kt. is ē (l ēta(n)), in WS æ (læ tan) and ā in OHG (lāζζ an ), OS ( lātan ) and ON ( lāta ). Despite the front in OE and OFris. most scholars think that only Goth. ē constitutes a retention, the innovation of ēl > ā taking place in all North and West Gmc. . There are at least three reasons for this assumption. First of all, the development of Gmc. *-ēn, -ēm > OE/OFris. -ōn, -ōm could hardly have taken place except by way of *-ān, -ām (Brunner 1965:§62 Anm.1). Secondly, the borrowing of str āta as str āζζ a in OHG and str āta in OS and as str ēt(e) in Angl./Kt./OFris. and stræ t in WS suggests that the US WURK XXX (1981), p. 53 forbears of OE/OFris. had an , which was subsequently fronted. And thirdly, the expansion of ē1 to ā was a direct consequence of the appearance of ē2 in the long/tense subsystem of late Gmc. (NG/WG), cf. Krupatkin 1970:66-7 and Antonsen 1972:131, 140. In Goth. there was no /ē2/. A different interpretation of the dialectal reflexes of Gmc. ē1 is given by Bennett (1950:232-5), who thinks that the innovation ēl > ā took place only in the central Gmc. dialects, leaving Goth. and OE/OFris. unaffected. There were no instances of ā (< ē) in Franconian until the early sixth century; according to Gysseling (1962:7-8, 1975:27-30) ā invaded the Low area () from the south, but was unable to oust ē along the coast, in Zealand, Holland, , Groningen and Ostfriesland until the eleventh and twelfth centuries. Even today some words with ē (< Gmc. ē1) have been retained in some of the coastal provinces. Most Dutch dialectologists have been of the opinion that the fronted reflexes of ē1 in the Gmc. West have not developed by way of ā, one reason being that northern Dutch exhibits æ even in front of nasals, cf. Vleeskruyer (1948:182) who thinks that -æn, -æm, a stage unrecorded in OE and OFris., must have preceded -ōn, -ōm. l Van Wijk established that the reflex of Gmc. ē was in fact æ in West Flanders, South Holland, Zealand, the south of North Holland and in a few of the inland districts, which should be seen in contrast to ē in the remainder of North Holland and in OFris. Schönfeld/v.Loey (1970:§80), who apparently see ē as a further development of æ, observe that this dialectal distribution corresponds to the English pattern (WS æ, Angl,/Kt. ē). A line can be drawn to Campbell (1959:§257) who thinks that the “Germanic invaders of Britain already most probably possessed one clear dialect distinction: the dialects from which W-S was to descend had æ from Prim.Gmc. æ, but those from which are descended all other known OE dialects had ē.” Wolff (1934:141-3) agrees with his Dutch fellow scholars in considering the 1 OE/OFris. (Dutch) reflex of Gmc. ē a common retention, but thinks that WS æ reflects the intermediate position of this dialect in relation to Angl./Kt./OFris. (ē) and OS ( ā). From a strictly linguistic point of view Wolff’s hypothesis cannot be accepted, however, since (pre-)WS might equally well be regarded as transitional in relation to any other Gmc. ā-dialect (ON, OLF, OHG). A compromise between the two principal views of the of ē1 in the Gmc. West is provided by Hans Kuhn (1955b: 44, 1955a: 32-3): he imagines that the old homelands of the Anglian and Saxon tribes, Schleswig and Northern , had become ā-dialect areas long before the Anglo-Saxon emigration. Very likely the Frisians, on the other hand, had retained the 8 in view of their US WURK XXX (1981), p. 54 proximity to the most conservative part of the Franconian area. On their way to England the Anglo-, according to Kuhn, went through Frisian , with the linguistic result that their ā reverted to æ/ē (except before nasals). By way of summing up, Angl./Kt. ē, WS æ can be said to agree with OFris. ē 9 1 (Dutch dial. æ) in having fronted reflexes of Gmc. ē ; these may represent a shared innovation, but a case can also be made for considering them a direct inheritance from Gmc. - or perhaps a combination of both, cf. Kuhn. Anyhow, the WS and the Angl./Kt. vowels point to approximately the same continential region: the Low Countries and . As for the seeming parallel between Goth. ē and Angl./Kt./ OFris. ē, it may, at best, be called a common retention. 11. Gmc. a continues as such in Goth. (fadar ), ON (faþer ), OS (fader ) and OHG (fater ), but is fronted (except before nasals) in OE (fæder ) and OFris. ( feder ). There is dialectal variation in OE (Kt./Merc. (Vespasian Psalter) e; æ elsewhere), but no special contact between OFris. and Kt./Merc. need be assumed on account of the -e-. Campbell (1959:§132) has given strong arguments for regarding the of Gmc. a in England and Frisia respectively as separate developments, arguments which I shall not repeat here. Instead, I shall draw attention to Krupatkin’s view of the evolution of the OE and OFris. vowel systems, according to which the subsystem of short vowels was restructured on the pattern of the long vowels, i.e. the split of a into a fronted and a nasalized vowel took place subsequent to a similar split within the long subsystem (1970:55). Whether the fronting process eventually resulted in æ or e seems, on the whole, to have been dependent on the corresponding long vowel: OFris. ē/e, Kt. ē/e, Merc. (VP) ē/e, WS æ/ æ , cf. Brunner 1965:§62 Anm. 2 and § 52 Anm.1. Note that VP, in which e prevails, has no instances of æ for ē, and that the I Rusworth , in which æ is frequent, has practically no e’s. In early Kt. there is vacillation between not only æ and ē, but also æ and e. However, in Nhm. there exists no such correlation of vowel quality between the long and the short reflexes of Gmc. ēl and a.

12. The preference for o instead of a before nasals in OEFris. and Angl./Kt. (mon(n) ) and the predilection for a in OWFris. and WS ( (n) ) in such cases have often been noted, most recently by Ramat (1976:72). But already Siebs (1901:1180-81, 1369) assumed that OWFris. a + nasal constituted a back formation from o + nasal, an assumption which has been substantiated by later investigations: o-forms occur sporadically in early OWFris. texts, and conversely, some instances of a before nasals crop up in US WURK XXX (1981), p. 55 late OEFris. (Sjölin 1966:30-31 - Hofmann (1970:87-90) is critical of some aspects of Sjölin’s view). As for the OE dialects o (Gmc. a before nasals) appears most consistently in the Angl. dialects of the ninth and tenth centuries. But in early Angl. there are many examples of a + nasal (Epinal Gl., Erfurt Gl.). The heyday of o-forms in Kt. as well as WS is the ninth century where they outnumber a-forms; early WS has o-forms besides a-forms. When scholars assert that WS is an a-dialect, such a statement can apply only to late WS. In Kt. texts there is a reversion to a simultaneous with that of WS, and a becomes prevalent before nasals in Angl. (except for the West Midlands) in the eleventh century (Brunner 1965:§79 and Luick I,1 1921:§110, §367). As in OFris., then, the distinction between a and o before nasals in OE is chronological rather than dialectal. 13. Bøgholm (1939:11) connects (northern) Nhm. ēa (< Gmc. eu ) and mutated īo (īu) with Ofris. ia and iu respectively (Nhm. þēaf , līode; OFris. thiaf , liude). Kt. īa (< Gmc. eu) has also been seen as a parallel to the OFris. form, cf. Kt./OEFris. bīade/biade (Siebs 1901:1158), but it should be noted that īa in Kt. can reflect the unmutated as well as the mutated form, cf. ahr īas đ ‘(he) falls’. Also, Kt. īa-forms crop up not in the earliest texts, but in ninth-century charters. Campbell characterizes the developments of eu (iu ) in OFris. and Kt. as “radically dissimilar” (1939: 85-6). On the other hand, he agrees that the resemblance between the OFris. and Nhm. reflexes is remarkable, though he sees no need to infer a close connection between OFris. and Nhm. from it. But Nhm. ēa (< eu) must have developed at a relatively early stage in view of its presence in the Ruthwell inscription (Luick I,1 1921: §127 Anm. la.2). Gysseling (1962:20) assigns the Frisian development to the eighth century (early ninth century at the latest), basing his chronology on the evolution of eu in the personal elements Gmc. *þeud ō- and *leudi-. 14. Brunner (I 1960:80) observes that Angl. and OFris. have both redeveloped a before ll and l + consonant, in contradistinction to the WS/Kt. fracture of æ > ea . The parallel goes even further in that Nhm. and OFris. exhibit retraction to a before r + cons. (Campbell 1959:§144), in Nhm. especially when a ( f, p, , m, w) precedes the vowel (or follows r-), and in OFris. without exception when w- precedes: Nhm. warþ , arm ; OFris. wartha , swart , but erm . The restoration of a in open syllabes when a, o, u followed is a development shared by all OE dialects, and is therefore not relevant to the present discussion, even though OFris. displays a retraction under similar circumstances: OE faran , macian ; OFris. fara , makia . It is unlikely that a was restored in these cases 10 prior to the Anglo-Saxon emigration seeing that the US WURK XXX (1981), p. 56 development presupposes the fronting of Gmc. a, cf . above, No. 11. 15. As already pointed out, Bremer (1900:848) saw a significant parallel between the WS front diphthongization of e, æ and æ after c, g and sc and certain diphthongizations that took place in the North Frisian dialects, and which he attributed to the influence of preceding palatal consonants, cf. WS gearn ‘yarn’, Amrum jûarn, Helgoland ju arn (Old Insular North Frisian *jæ rn); WS gēar ‘year’, Amrum/Föhr jûar, Helgoland jôar (Old Insular North Frisian *jæ r). Siebs, however, has repeatedly (1901:1214, 1931:66) rejected the idea of basing any special relationship on this similarity, his main reason being that the North Frisian diphthongizations took place irrespective of the quality of the preceding consonants. The chief evidence adduced by Siebs, viz. Helgoland mûárn, môárn, 11 is nevertheless erroneous. According to Nils Århammar (personal communication) the word was spelt maaren in the 19th century, and since about 1870 the pronunciation has been [mo: n]. Unlike the remaining North Frisian forms Helg. moorn derives from O(W)Fris. morn (= WFris. moarn ). But Löfstedt (1963/65:286) does give two instances of front diphthongization in Insular North Frisian which are not due to the influence of preceding palatal consonants: ju ālumm ‘ewe lamb’ (< * æ-lōmb ) and jūart ‘hither’ (< * hæ rd ), i.e. initially and following an h-. 16. ē is the i-mutated reflex of Gmc au not only in OFris. (nēd, hēra ), but also in Angl./Kt. ( nēd, hēra(n)), 12 cf. WS nīed, hīeran. But very likely the parallel is just a graphic one. In his investigation of the accented vowels in the Schiermonnikoog dialect Arne Spenter presupposes an opposition between æ and ē for OWFris., au + i, j resulting in æ (1968:190). Traditionally, Ofris ē is supposed to derive from Gmc. ē1, ē2, a/u + nasal + + i, j, ai , ai/au /ō/ū + i, j, a very heavy burden, indeed, and as Spenter points out, it is „äusserst problematisch, ob es jemals ein afries. ē-Phonem so vielfachen Ursprungs gegeben hat” (Spenter 1970:312, cf. also Hofmann 1964:160-85). Similarly Gmc. ai , whether i-mutated or not, may well have developed into OFris. / æ/, written ē, cf. OFris. dēla, h ēla, hw ēte; Schierm. dēl, vēt (Spenter 1968:191-3). Thus the OFris. forms may not correspond to Kt. dēlan, hēlan at all, but rather to WS/Angl. dæ la(n), hwæ te. Most scholars follow Luick (I,l 1921:§201) in dating the OE i-mutation to the sixth century, and in other Gmc. languages the change manifests itself even later. The late appearance of this and other phenomena in NG/WG dialect constitutes a difficult problem in respect of Gmc. dialect grouping. 17. In Kuhn’s view (1955a:28, 1955b:39) the development of y(< u + i, j >e in Kt. and OFris. can hardly be a co- US WURK XXX (1981), p. 57 incidental parallel: Kt. fellan, bre c (WS fyllan, bry c), OFris. fella, h ēd 13 (WS hy d). The unrounding in Kt. probably took place about 900, and such a dating does not contradict the Frisian evidence, - in fact, OFris. k does not undergo assibilation in front of ē (ū + i, j ), cf. kest (WS cyst). If Kuhn is right, then, this correspondence between OFris. and Kt. must be due to late contact across the . In Angl. and to a large extent in WS y was unrounded to i in the tenth and eleventh centuries: fillan, br īd (Luick I,1 1921:§287). The same development also occurs in insular North Frisian (*filla, *br īde), cf. Amrum/Föhr fal ‘fill’ (Sylt/Helgoland fel) and brid ‘bride’ (Sylt brir, Helgoland brid), whereas the vowels in OEFris. (Rüstring) ekimen, kining, sinne probably reflect an earlier e (< y), which has been raised in front of nasals. 14 Interesting, too, are the occurences of unrounded i (and e) < y in MDu, cf. OFlem. (OZeal., Holl.) brig(ghe), pit, ric and hil(le) ; OWFlem./OZeal. hīde (OE hy); North Holl . pet, reg(ghe) (Schönfeld/v.Loey 1970:§46, §33 Opm.3 and Franck 1910:§35).The modern Dutch and Frisian dialects show a similar pattern: the word for ‘ridge, back’, rug (AB), is rik in West Flanders and Zealand, ri(e)g in Brabant and West , reg in North Holland and rêch, rich in Friesland (Kloeke 1939vv.: Afl.4, No.3). See also Miedema 1980:180-84. The OFris./Kt. change of y > e thus has a counterpart in the unrounding of y > i in especially the southern Dutch coastal regions, in the North Frisian island dialects and in large parts of the late Angl. and late WS dialect areas. 18. In the past scholars (cf. Siebs and Jungandreas) have seen the development of final ēg(æ g to ēi as a shared Kt./OFris. innovation, cf. Kt. dei, wæi ‘weighed’, wei ‘way’, wr ēi ‘accuse!’; OFris. dei, wei ‘way’ , kēi ‘key’. The first examples of the vocalization of -g, however, are found in early glosses (Epinal Gl. gr ēi, bodæi; Corpus Gl. gr ēi, popei, cf. Campbell 1959:§266) which are not normally considered to be Kt. Spellings with -i become frequent in the Kt. charters of the ninth century; in tenthcentury WS, -eg- represents ON -ei- in Stegen, Swegen (ON Steinn, Sveinn), cf. Luick I,1 1921:§257 2 and Kuhn 1955a:28, and in later WS we often find -ig for -g: weig, mæ ig. Similarly, instances of -ig for -g are found in Nhm. and also in the Kentish Glosses to Proverbs, as a matter of fact. In ME -i and -y have replaced OE -g everywhere. According to Kuhn (1955a:28) the first attestations of eg > ei in OFris. are the eleventh-century in -dei in the Werden Cadastral Register I,II , cf. Adaldei, Amuldei , Birdei, Erdei, etc. In MDu., personal names in -dei, -dey also occur (Schönfeld/v.Loey 1970:§64 Opm.2), and spellings like peghel, peil and seghel, seil alternate (Franck 1910:§118, §26); as a matter of fact, place-name evidence seems to indicate that there was a relatively US WURK XXX (1981), p. 58 early change of egi > ei in Du., perhaps as early as 800, cf. Leyens (South Holland) < Legihan (Schönfeld/v.Loey 1970:§64 Opm.2). In OGhent. the personal name elements Eil-, and Mein-, Rein- alternate with Egel- and Megen-, and precisely the same sort of vacillation is seen in OS, where Eil- ,Mein-, Rein- frequently replace Egil -, Megin-, Regin- (Gallee 1910: §62 Anm., § 251 3). In the Alemannic dialect of the tenth and eleventh centuries egi develops to ei, and spellings like Ein-, Mein-, Rein- crop up here (instead of Egin-, Megin-, Regin-), cf. Braune/Eggers 1975:§149 Anm. 5a. Phonetically, the vocalization of g to i after e (æ) is not difficult to comprehend, and spread by contact need not be assumed, especially not where occurrences are irregular as in e.g. OS and OHG (Alem.). The change is first attested in England, and the fact that on the Continent it is most widespread immediately across the sea, suggests that contact influence may after all have been exerted by OE on Dutch/Frisian. But even though the -i spelling is most consistently used in OKt., we are hardly permitted to interpret this alone as a sign of special Kt./OFris./Dutch contacts: the difference between Kt. on the one hand, and WS/Angl. on the other, may well have been an orthographic one. 19. The presence of rising in MKt., OFris. and WN is the of an from 1953 by W.H. Bennett, who points out the marked parallelism of e.g. Mkt. dyep (Gmc. eu , see above, No. 13), OFris. diap, WN djúpr; MKt. yerþe, WN jo r (OE back mutation/ON breaking of the accented vowel in Gmc. *erþ ō); MKt. sye-, OFris. sia, WN sjá (contracted forms of Gmc. *se χwan); MKt. dyevel, OFris. diovel, WN djofull (loan words, cf. Lat. diabolus ). In Kt. the development from falling to rising diphthongs is usually assigned to the ME period (Luick I,1 1921:§359), but Bennett (1953:71-80) attempts to push the shift back into OKt. As for OFris., Bennett (1953:76) notes that the first records exhibit the change in a well-advanced stage. The WN shift dates back to before 1200, and may have begun as early as in the late ninth century. According to Bennett it is possible that the parallel outlined above is the result of coincidence, but it may also be interpreted as “the result of tribal inter- associations.” The way in which Bennett sees these, appears from the following lines: 15 “How much contact there was between the Jutes of and the West of is of course hard to determine, but it is at least interesting to note that the speech of both groups was characterized by rising [j]-diphthongs. If the Jutes later spread south along the coast of Friesland, as indicated by the Finn Episode, they entered the lower area and still retained vestiges of its culture when they finally settled in Kent. And the close association between the Jutes and the Frisians ... US WURK XXX (1981), p. 59

is also paralleled by their common use of rising [j]-diphthongs.” It is evident that the argumentation put forward here is circular: the rising diphthongs are taken as evidence for tribal interassociations even though Bennett set out to show the probability of such connections in order to explain the presence of these diphthongs in Kt./OFris./WN! To the theory outlined above it should be added that in Bennett’s view the rising diphthongs may have first arisen in WN owing to the influence of an early Finno-Ugric substratum. It is difficult for me to see any great significance for the pre-invasion period in this parallel (cf. Samuels 1971:7); there can hardly be any doubt that the Kt./ OFris./WN rising diphthongs arose long after the Anglo-Saxon emigration, cf. their various origins, and the only question that in my opinion is left open for discussion is whether the resemblance could be due to late contacts. There are several instances of late linguistic exchanges between Kent and Friesland, and the presence of rising diphthongs in both places may be another example. With a reference to Hammerich’s article from 1937 Gysseling (1962:20) asks whether ninth-century Danish influence may be held responsible for the Frisian rising diphthongs. Of course, I am in no better position to answer that question than is Gysseling, but it is worth pointing out that - despite Bennett (1953:76) - the shift of accent took place not only in WN, but also in East Norse, cf. Old Danish di ūpær, iorth, fi āndi, diavul (Brøndum-Nielsen I 1950:§176-80). contacts may there have resulted in an expansion of the innovation from Southern Scandinavia to Friesland, 16 from where it spread to Kent. 20. In WS g is often lost before d, þ, n , and (occasionally) l, cf. -br ōden (pp. of br ēden ), tīian, ong ēan and snæ l. The other OE dialects nearly always retain g in such cases, but early non-WS forms like Epinal Gl. sn ēl, str ēl (also Corpus Gl.) should be noted. On the Continent a similar loss is seen only in Frisian, cf. OFris. pp. br ūden, j ēn, j ēnst (OS gegin). The normal OFris. development of g before d, n, l (and s) is i, however, cf. inf. breida (OS bregdan), wein (WS wæn, OS wagan). The resemblance between WS and OFris.as seen here is hardly of great significance: (1) in WS there was much vacillation, esp. in early WS; (2) geographically, the loss was probably more widespread during the early period; and (3) in OFris. the loss was the exception rather than the rule. 21. The loss of final -n is a feature common to Nhm., OFris. and ON, although ON exhibits a larger number of -n losses than either of the two other dialects (-an , the asm. ending of the strong , where -n was re- US WURK XXX (1981), p. 60 tained in ON, was originally covered by a final vowel, cf. Noreen 1970:§299 5 Anm.4). Early (400 A.D.) in Norway and retain -n where it is later lost, cf. n-stems gsm. þ r a w i j a n (Kalleby) and dsm. h a 1 a i b a n (Tune), but the Norwegian Eggjum inscription (700 A.D.) appears to have dropped final -n. In Nhm. and OFris. -n is lost mainly in , in a/g/ds. of weak and in adverbs, cf. Nhm. cuma, hearta, ūta; OFris. kuma, kempa, b ūta (; ON koma, hana, þá), but usually not in pt.plur.ind.: Nhm. br ēcon, OFris. kōmon (ON 3rd pt.plur.ind. tóku) and pp.: Nhm. ārisen, OFris. faren (ON hlaupinn ). Forms without -n crop up in the earliest Nhm. sources, but there is a good deal of vacillation, cf. the Ruthwell Cross. Luick (I,2 1940:§654) may therefore be right in dating the loss to the period immediately preceding the earliest written texts, i.e. the seventh century. In Campbell’s opinion, which is that of a dialect geographer, the loss of final -n in ON, OFris. and Nhm. is not due to “descent from a common type of Germanic,” but should rather be seen as an innovation “which, cutting across the old linguistic grouping, linked North Germanic to Frisian and Northumbrian” (Campbell 1959:§415). We agree that the loss is a post-invasion development, and repeat that it is more general in ON than elsewhere. It appears from the individual discussions of the items listed above that some of the material must be rejected as evidence of special OE dialectal connections with Frisian. This holds true of Nos. 11 (Gmc. a > Kt./Merc./OFris. e), 12 (Gmc. a > o before nasals in OEFris. and Angl./Kt.), 15 (palatal diphthongization in WS and Insular North Frisian), 16 ( i-mutated reflexes of au and ai) and 18 ( eg (æ g) > ei). Ten of the remaining correspondences clearly belong to the post-invasion period, and Kt. participates in five of these parallels, either alone among the OE dialects, cf. Nos. 17 (Kt./OFris. ū + i, j >e), 4 (Kt./OFris./MDu. n/asf., n/ap. 3rd pers.pron. s-) and 19 (Kt./OFris./WN rising diphthongs) or in combination with WS, cf. Nos. 6 (Kt./WS/OFris. 1st pres.sg.ind. strong verbs -e) and 7 (Kt./WS/OFris. syncopated 2/3. pres.sg.ind. strong and Class I weak verbs). The only exclusive parallel between WS and Ofris. is No. 20 (loss of g before d, n, l ), while No. 17 ( ū + i, j > y> i) is common to not only WS and North Frisian (and MDu.), but also Angl. As for other post-invasion parallels in which Angl. participates, two are exclusively shared by Angl. and OFris., viz. Nos. 13 (Gmc. eu > ēa) and 14 (restoration of a before ll/l/r + cons.) and one is shared by Angl. (Nhm.), OFris. and ON, cf. No. 21 (loss of final (-n). The number of pre-invasion correspondences is relatively small, and the fact that Angl. participates in most of them. should not be overstressed, seeing that in four US WURK XXX (1981), p. 61 instances Ofris. is not the only continental language to be involved and that the fifth (and final) item (No. 10) has Angl. and Kt. participation. As shown in No. 1, the npm. n-stem forms with i-mutation (reflecting either an extension of the vanishing grade from gp. or, less plausibly, the IE *-en -grade of the thematic element) constitute a parallel between Angl., OFris. and WN, whereas No. 8 (accented vowel ē in pt. forms of ‘did’) and No. 9 (Gmc. * waljan ) connect Angl. with OFris., OS and OHG, No. 8 being a retention and No. 9 a common choice of the same Gmc. variant. The presence of -e- in asm. of the dem. pron. (no. 5) is open to much doubt (and alternative explanations), but may link Angl. to OFris., OS, OHG and MDu. Finally, the ē of Angl., Kt. and OFris. (No. 10) could be interpreted as a reflex of NG/WG ā (< Gmc. ē1), i.e. a common innovation; an alternative explanation would be that ē represents a retention shared by Angl., Kt., OFris. and Gothic. - There appears to be only one early exclusive parallel between Kt. and OFris., viz. No. 2 (Kt. læ resta , OFris. lērest ), and WS takes part in only item No. 3 ( in a/ds. 1/2. pers.prons.), an innovation which is attested also in Kt. and OS (in addition to OFris.). It is clear that none of the OE dialects can be regarded as the direct successor of a (pre-)Frisian tribal dialect of the fifth and sixth centuries on basis of the early evidence. This is not to say, of course, that no OE/OFris. parallels arose prior to the Anglo-Saxon emigration, in fact OE as a whole exhibits far more (early) cor- respondences with OFris. than with any other Gmc. language, cf. Nielsen 1981:ch.III.i and IV, but the pre-invasion material just does not warrant conclusions in the way of early Frisian tribal settlement in any particular OE dialect area. 17 Rather, the (absence of) evidence seems to point towards a genesis of the OE dialects in England as advocated by e.g. Sweet (1875-6:561-2), Chadwick (1907: 57-69) and Campbell (1959:§5). As for the post-invasion parallels cited the most striking feature is that seven of them have southern English (esp. Kt.) participation. This may at least in part be ascribed to late contact between Frisia and SE England across the North Sea.

Odense University Hans F. Nielsen

Acknowledgement: I should like to thank Nils Århammar whose valuable comments and suggestions have been of much benefit to this paper. US WURK XXX (1981), p. 62

NOTES 1. And Skeat (1896:221) goes on: “It was only natural that in crossing the sea the Scandinavians should make for the north of England, the Saxons (from the coast of Holland) should southwards, while groups of Frisians or East Saxons, would make for .” 2. According to , the Germanic invasion of Britain was undertaken by three Continental tribes, the Angles, the Saxons and the Jutes, the -mentioned tribe settling in Kent, the Isle of and the coastal districts of . 3. A recent attempt to identify a specifically Angl. vocabulary within OE has been made by Franz Wenisch (1979). 4. For the provenance of (Insular) North Frisian, see Århammar 1968a:313-14. 5. The section takes into account parallels between the OE dialects and all the continental Gmc. languages. 6. For lexical parallels between OE and OFris., see Löfstedt 1963/5: 281-345, 1966:39-65, 1967:11-61 and esp. 1969:25-39. Munske (1970:40-52, 1973:208-9) is concerned with OE/OFris. correspondences in the sphere of legal terminology. 7. According to Siebs (1901:1334) -e reflects earlier -u, -o. Nils Århammar has pointed out to me that such an inference is unwarranted in that no 1.pres.sg.ind. forms of strong verbs with short are attested in the Rüstring MSS, the only OFris. texts to retain -u, -o in unaccented final position, cf. v.Helten 1890:§57. 8. According to Thomas Markey, A Reader (München, 1976), p. xi, the “oldest Frisian runic inscriptions (Britsum and Wijnaldum) show ā in the 6th century.” Markey may well be right in assigning ā to Britsum, but it is probably not ā < ē1 as he seems to think, but ā < ai, cf. BUMA 1951:315 and Miedema 1974: 106-9, 1977:469. Of the Wijnaldum inscription H. Arntz and H. Zeiss in Die einheimischen Runendenkmäler des Festlandes (Leipzig, 1939), p.417, say: “Eine sinnvolle Ausdeutung erscheint unmöglich.” Possibly, a misinterpretation of Kuhn 1955a:33 underlies Markey’s statement. l 9. æ is the regular reflex of Gmc. ē in Old Insular North Frisian as well, cf. Jørgensen 1946:109-13, Hofmann 1964:160-85. 10. Århammar finds it more probable that a was in fact retained in OFris. and not restored. 11. These as well as the preceding North Frisian spellings in this stem from Siebs 1901:1214-15. 12. See Siebs 1901:1159 and 1233. 13. Brunner 1965:§31 Anm. 1 and Luick I,1 1921:§183. Campbell (1959: §290) argues in favour of a somewhat earlier change. 14. For a discussion of i-mutated (Gmc. ū in North Frisian, see Jørgensen 1946:95- 119, esp. 115-16, and Hofmann 1960:72-4. 15. This tallies with the view of Gysseling, who thinks that the rising diphthongs had become well established in Friesland in the eleventh century. In Ostfriesland the innovation seems to have been on its way in the tenth century (1962:20-21). US WURK XXX (1981), p. 63

16. The problem of linguistic exchanges between Friesland and Scandinavia is discussed by Feitsma 1962:97-121. 17. This subject is dealt with in a recently published paper by Bremmer (1981).

References: Antonsen, E.H., ”The Proto-Germanic Syllabics (Vowels).” Toward a Grammar of Proto-Germanic. Ed. F.v.Coetsem/H.L Kufner. Tübingen, 1972. Pp. 117-40. Århammar, N. „Friesische Dialektologie.“ Germanische Dialektologie (Festschrift für Walther Mitzka), I. Wiesbaden, 1968a. Pp. 264-317. - „Die Herkunft des Inselnordfriesischen im Lichte der Wortgeographie.“ Philologia Frisica Anno 1966. Grins, 1968b. Pp. 49-76. Bech, G. Die Entstehung des schwachen Präteritums. Historisk-filosofiske Meddelelser udg.af Det Kongelige Danske Videnskabernes Selskab, 40,4. København, 1963. Bennett, W.H. ”The Germanic Development of Indo-European ē.“ Language, 26 (1950), 232-5. - „A West Norse-Frisian-Kentish Parallel.“ = International Anthropological and Linguistic Review , 1 (1953), 71-80 Bøgholm, N. English speech from an Historical Point of View. Copenhagen-, 1939. Braune, W. / H. Eggers. Althochdeutsche Grammatik. 13.Auflage. Tübingen, 1975. Bremer, O. „Ethnographie der germanischen Stämme.“ Grundriss der germanischen Philologie, III . 2.Auflage. Strassburg, 1900. Pp.735-950. US WURK XXX (1981), p. 64

Bremmer, R.H. “Frisians in Anglo-Saxon England; a historical and toponymical investigation.” Fryske Nammen, 3 (1981), 45-94. Brøndum-Nielsen, J. Gammeldansk Grammatik, I. 2. Udgave. København, 1950. Brook, G. L. English Dialects. 2nd edition. London, 1965. Brunner, K. Die englische Sprache, I. 2. Auflage. Tübingen, 1960. - Altenglische Grammatik. 3. Auflage. Tübingen, 1965. Buma, W.J. „Das Runenstäbchen von Britsum.“ Beiträge zur Geschichte der deutschen Sprache und Literatur, 73 (1951), 306-16. Campbell, A. ”Some Old Frisian -Changes.” Transactions of the Philological Society, 1939, pp. 78-107. - Old . Oxford, 1959. Chadwick, H. M. The Origin of the English Nation. Cambridge. 1907. Clement, K.J. Schleswig , das urheimische des nicht dänischen Volks der Angeln und Frisen und Englands Mutterland wie es und ward. Hamburg, 1862. Feitsma, T. ”Sproglige berøringer mellem Frisland og Skandinavien.“ Sprog og Kultur, 23 (1962), 97-121. Franck, J. Mittelniederländische Grammatik. 2.Auflage. Leipzig, 1910. Gallée, J. H. Altsächsische Grammatik. 2.Auflage. Halle-Leiden, 1910. Gysseling, M. “Het oudste Fries.“ It Beaken, 24 (1962), 1-26. - “Hoofdlijnen in de evolutie van het Nederlandse vocalensysteem.“ Toponymie & Dialectologie, 49 (1975), 25-59. Hammerich, L.L. ”Über das Friesische.“ Mélanges Linguistiques offerts à Holger Pedersen. Aarhus, 1937. Pp. 351-8. Helten, W. L. v. Altostfriesische Grammatik. Leeuwarden, 1890. Hirt, H. Handbuch des Urgermanischen, II. Heidelberg, 1932. Hofmann, D. „Die Sprache der Nordfriesen und ihre alten Wohnsitze.“ Philologia Frisica Anno 1959 . Grins, 1960. Pp. 71-8. - „‘Germanisch‘ ē2 im Friesischen.“ Festschrift für Jost Trier zum 70.Geburtstag. Hrsg.W.Foerste / K.H. Borck. Köln-Graz, 1964. Pp. 160-85. - „Die osterlauwerssche Urkundenüberlieferung als Quelle für das Altfriesische.“ Flecht op ‘e koai, stúdzjes oanbean oan Prof. Dr. W.J. Buma ta syn sechstichste jierdei. Grins, 1970. Pp. 83-94. Jordan, R. Eigentümlichkeiten des Anglischen Wortschatzes. Anglistische Forschungen, 17. Heidelberg, 1906. Jørgensen, P. Über die Herkunft der Nordfriesen. Det Kgl. Danske Videnskabernes Selskab, Historisk-filologiske Meddelelser, 30,5. København, 1946. Jungandreas, W. Geschichte der deutschen und der englischen Sprache, I and III. Göttingen, 1949. Karstien, C. Historische deutsche Grammatik, I. Heidelberg, 1939. Kloeke, G.G. (ed.) Taalatlas van Noord- en Zuid-Nederland. Nieuwe Noord- en Zuid- Nederlandsche dialectbibliotheek, deel 1. Leiden, 1939vv. Krahe, H. Historische Laut- und Formenlehre des Gotischen. 2. Auflage bearb. von E. Seebold. Heidelberg, 1967. - Germanische Sprachwissenschaft, I-II. 7. Auflage bearb. von W. Meid. Berlin, 1969. Krupatkin, Y. B. ”From Germanic to English and Frisian.” Us Wurk , 19 (1970), 49- 71; 20 (1971), 60. Kuhn, H. „Zur Gliederung der germanischen Sprachen.“ Zeitschrift für deutsches Altertum und deutsche Literatur, 86 (1955a) , 1-47. US WURK XXX (1981), p. 65

Kuhn, H. „Friesisch und Nordseegermanisch.“ Us Wurk, 4 (1955b), 37-46. Löfstedt, E. „Beiträge zur nordseegermanischen und nordseegermanisch-nordischen Lexikographie.“ Niederdeutsche Mitteilungen, 19/21 (1963/5), 281-345; 22 (1966), 39-65; 23 (1967), 11-61; 25 (1969), 25-39. Luick, K. Historische Grammatik der englischen Sprache, I,1-I,2. Leipzig, 1921-40. Miedema, H.T.J. “Dialect en runen van Britsum en de oudste anglofriese runeninscripties.” Taal en Tongval, 26 (1974), 101-28. - “Noordzeegermaans (Ingweoons), Fries en Nederlands.” De Nieuwe Taalgids, 70 (1977), 465-72. - “’Breg, reg and is Old English and Old Friese.‘” Liber Amicorum Weijnen. Assen, 1980. Pp. 180-87. Munske, H.H. „Angelsächsisch-altfriesische Beziehungen in der Rechtsterminologie für Missetaten.: Flecht op ‘e koai, stúdzjes oanbean oan Prof. Dr. W.J. Buma ta syn sechstichste jierdei. Grins, 1970. Pp. 40-52. - Der germanische Rechtswortschatz im Bereich der Missetaten. Philologische und sprachgeographische Untersuchungen.I. Die Terminologie der älteren westgerma- nischen Rechtsquellen. Studia Linguistica Germanica, 8,1. Berlin-New , 1973. Nielsen, H. F. “Morphological and Phonological Parallels between Old Norse and Old English.” Arkiv för Nordisk Filologi , 90 (1975), 1-18. - “The Earliest Grouping of the Germanic Dialects.” Arkiv för Nordisk Filologi , 94 (1979), 1-9. - Old English and the Continental . Innsbruck, 1981. Noreen, A. Altnordische Grammatik, I.5., unveränderte Auflage. Tübingen, 1970. Ramat, P. Das Friesische. Innsbruck, 1976. Samuels, M.L. “Kent and the Low Countries: som linguistic evidence.” Edinburgh Studies in English and Scots. Ed. A.J. Aitken / A. McIntosh / H. Pálsson. London, 1971. Pp. 3-19. Schönfeld’s Historische Grammatica van het Nederlands . 8e druk, verzorgd door A. v. Loey. Zutphen, 1970. Siebs, Th. „Geschichte der friesischen Sprache.“ Grundriss der germanischen Philologie, I. 2. Auflage. Strassburg, 1901. Pp. 1152-1464. - “Die Friesen und die nächstverwandten Stämme.” Mitteilungen der schlesischen Gesellschaft für Volkskunde, 31 (1930), 44-84. - “Die Friesen und ihre Sprache.“ Die Friesen. Hrsg. C. Borchling / R. Muuss. Breslau, 1931. Pp. 56-69. Sisam, K. Studies in the History of Old . Oxford, 1953. Sjölin, B. „Zur Gliederung des Altfriesischen.“ Us Wurk 15 (1966), 25-38. Skeat, W.W. Principles of English . First Series. Oxford, 1887. - “The Frisian Origin of the Mercian Dialect.” The Academy, 1245 (1896), 221. Spenter, A. Der Vokalismus der akzentuierten Silben in der Schiermonnikooger Mundart. Køpenhagen, 1968. - Review of B. Sjölin, Einführung in das Friesische in Beiträge zur Geschichte der deutschen Sprache und Literatur 92 (1970), 310-14. US WURK XXX (1981), p. 66

Steller, W. Abriss der Altfriesischen Grammatik. Halle, 1928. Streitberg, W. Urgermanische Grammatik. 4., unveränderte Auflage. Heidelberg, 1974. Sweet, H. “Dialects and Prehistoric Forms of Old English.” Transactions of the Philological Society, 1875-6 (London, 1877), pp. 543-69. Vleeskruyer, R. „A. Campbell’s Views on Inguaeonic.“ Neophilologus, 1948, pp. 173- 83. Wachter, F. „Friesische Sprache.“ Allgemeine Encyklopädie der Wissenschaften und Künste, I, 51 (1850), 411-22. Walde, A. Die germanischen Auslautsgesetze. Halle, 1900. Wenisch, F. Spezifisch anglisches Wortgut in den nordhumbrischen Interlinearglos- sierungen des Lukasevangeliums. Anglistische Forschungen, 132. Heidelberg, 1979. Wolff, L. „Die Stellung des Altsächsischen.“ Zeitschrift für deutsches Altertum und deutsche Literatur, 71 (1934), 129-54.