Etymology

The PGmc str. V v. *legjan- (wk. pres.) produced OE licgan (in which a palatalized /j/ in the pres. indic. 2 and 3 sg. liges(t), ligeð and some other parts of the paradigm would be expected), beside OFris lidza, OS liggian, OHG liggen and OIcel liggja. The usual form in the Gaw MS is the (str.) lyʒe (pres. 3 sg. <lys, lis>), which has generalized stem-final /j/ to the infin. But there is also a smattering of forms in <g> or <gg>, including the wk. past <lyggede> at Pat 184, which seem to indicate that a separate wk. paradigm in /dʒ/ or /g/ had developed in this dial. Spellings with <gg> are of course ambiguous, since <gg> could stand for /dʒ/, but <lygez> at Gaw (1179) seems a more likely representative of a /ɡ/ pronunciation, and we have included this and other forms with single <g> in the occurrences list in the present entry. As [ɡ] is non-initial, note however that the absence of palatalisation cannot be a certain test of loan and only some authorities invoke Norse input (TGD, Jordan-Crook §192, Thorson 35).

PGmc Ancestor

Proposed ON Etymon (OIcel representative)

liggja 'to lie'
(ONP liggja (vb.))

Other Scandinavian Reflexes

Far liggja, Icel liggja, Norw liggja, Dan ligge, Sw ligga

OE Cognate

licgan 'to lie'

Phonological and morphological markers

[absence of palatalization of */ɡ/] (may not be applicable)

Summary category

CC2c

Attestation

Only a handful of spellings in <g> are cited in MED (<gg> is widely attested, but these may represent /dʒ/), found mainly in N and E texts (e.g. a1425(?c1350) Ywain (Glb E.9); Cursor) but also elsewhere (viz. a1500(?a1450) GRom.(Hrl 7333)). EDD cites a variety of MnE dial forms apparently in /ɡ/ from Sc., N, N/EM, War. and Wor. 

Occurrences in the Gersum Corpus

Gaw 1179; WA 756*, 1570, 4775

Bibliography

MED līen (v.1) , OED lie (v.1) , HTOED , Dance lygez, de Vries liggja, Mag. liggja, Bj-L. ligge, Seebold leg-ja-, Orel *leʒjanan, Kroonen *leg(j)an-, AEW licgan