Fricatives and Semivowels

OE /G``/ and /j/ in Early Middle English

Note: unless specified otherwise, eME indicates the normalised early East Midland used in this site.

1. /ai/ and /ei/: OE initial g` /j/ (including g` following a prefix) is most frequently represented in 12th and 13th century texts by y% or y`. In Orm and most 12th century texts, OE g` following a front vowel is also represented by y%. This is reflected in the default eME spelling. That isn't the end of the story however. At some point, early in the ME period, there was a levelling of the four OE combinations: aeg`, ae_g`, eg` & e_g` to two diphthongs: /ai/ and /ei/. So do we use only ay% and ey% in the default eME spelling? How prevalent was this levelling in sound by 1150, and how did the various late 12th and early 13th century texts represent these sounds? We need answers to those questions, before we can decide on their spelling in eME. In particular, we need to sort out what happened to OE ae_g` - did it go to/ai/ or /ei/ early in the ME period? Lass <sup>1</sup> seems to indicate that OE ae_g` moved from /ae:j/ to /ai/. He gives the example OE grae_g` (> ModE gray/grey). Johannesson <sup>2</sup> on the other hand states that OE ae_ + g` developed into the diphthong ei, pointing to: OE clae_g` > clei and cae_g` > kei. Examples from 12th and 13th century texts are mixed but favour the simpler 2-diphthong system of ME (/ai/ & /ei/) over the 4-diphthong system of OE. eME sources also favour /ei/ over /ai/ as the reflex of OE ae_g`. Lmn has maei (< mae_g`) and faeie (< OE fae_g`e) while Owl has both aiware (< ae_g`hwae_r) and either (< ae_g`t`er). AW also has eid`er and PC2 has rachenteges (< racente_ag`es). Perhaps the clincher is Orm ey%y%t`er (< OE ae_g`t`er), which indicates that the first element of the new diphthong is not /E`:/, but is now indistinguishable from /e/, in both quality and length. Thus, in the Early Middle English Texts, the default eME spelling uses ay% and ey%, rather than <aey%> or <e__y%>. However I have preserved <ae_g`> and <e_g`> in the 'PC1' spelling options.
2. eh and oh: Lass ¹ estimates that the basic Middle English system of diphthongs was established in its final form by 1250. Presumably, this includes the development of OE oh and eh to ME ouh and eih respectively, (as outlined in his following paragraph). Accordingly, you will find forms such as douy`ter and feighten with the SO and Ch spelling options, but not in eME, since neither is represented in PC2 or Orm.
3. OE e_ag/e_og > eME e__g > lME eigh/ey/ei. It's tempting to conclude that OE <g> following a front vowel was pronounced /j/, as a rule. There are numerous examples to back that deduction. However, there were frequent exceptions as well, particularly before back vowels, but in some cases - before front vowels too. In Orm, (col 31) for example, we find niy%h`enn. This was a revelation for me. I had assumed that any <g> following a front vowel, would have been pronounced as a semivowel or continuant /j/ in OE. Obviously I was wrong. Here was one <g> following a front vowel, that was clearly a fricative. I looked around for more examples and quickly found in Barney <sup>3</sup> a handful of OE words in which a <g> following a front vowel was not marked as a semivowel /j/ - fle_ogan, dre_ogan, be_ag, -genga, wegan and wiga. In Orm I found another - rey%h`ell while in Lmn I found stihe from sti_gan. Baker threw up some more: si_gan, swi_gian and most surprising of all - e_age. There are certainly others. With the possible exception of regel, however, it is unlikely that the fricative in these words developed into a /w/ in later ME. It appears we're dealing with a different point of articulation here. Eventually, the <ig> of both wige and nigen became /i:/, to give wye and nine. But in the meantime we have wyy`e (Patience, c 1370) and niy%h`enn (Orm). Similarly, the eME /e:J`/ <eg> became either <ee> or <eiy`/eigh>, giving modE dree, bee and weigh. Thus there appear to be distinct paths of development for OE fricative <g>, depending on whether it fell between two back vowels, or between a front vowel and either a back vowel or consonant. In eME I've represented the reflexes of both with <g> but in alternate eME with <y`>. In the SO and Ch spelling versions, I've represented this sound with <y`> and <gh> respectively. (These also happen to be the spellings for /x/ and /C``/ in the SO and Ch versions.)
   1. In other words, OE fricative g (/G``/), remained a fricative in the early ME period. However there were 2 allophones of this sound, determined by their context and these developed differently in later ME. IF the fricative <g> had followed a back vowel or <r> or <l> in OE, it passed to /w/ by the latter part of the 14th century: boga > boge/boy`h`e > bow. If on the other hand, it had followed a front vowel in OE, it passed to /j/ or was assimilated to the preceding front vowel in lME: OE wiga > eME wige/wiy`h`e > lME wye. Similarly: nigon > nigen/niy`h`en > *nyen > nine, e_go (Nth) > e__ge/e__y`h`e > eighe > eye.
   2. So, the reflex of OE e_age/e_ge should be e__ge in eME, rather than *e__y%e or *e_he, as some may have expected.
4. Note - sometimes after <n>, <g> was pronounced /dZ``/; the examples usually given are engel, ic` senge; in eME sources this sound is generally spelt gg but in the default eME spelling of this site it is cy%, thus - sency%e; however, I have stuck with engel, since Orm's spelling enngell gives no indication of a palatal sound here.
5. The eME reflex of an OE word containing <g> is not always easy to determine. In the absence of an indication of the pronunciation of <g> in a given OE word in Barney ³, I have assumed the following for the OE pronunciation of <g>:
   1. initially and before either a back vowel (<a>, <o>, <u>) or <ae> or <y> or a consonant - /g/; e.g. go_d, glad, gaers, gylt
   2. initially and before a front vowel (<i>, <e>) - /j/; e.g. g`e_ar, g`i_efan; but genga
   3. finally and following either <ae> or <i> or <e> - /j/; e.g. manig`, daeg`, weg`
   4. between two front vowels - /j/; e.g. sig`e, faeg`er; but e_age
   5. between <e> and a back vowel - /j/; e.g. wreg`an, swe_g`an; but wegan
   6. between two back vowels - /G``/; e.g. a_gan, boga
   7. after <l> or <r> - /G``/; e.g. beorgan, folgian, iergt`u, but not byrg`an herg`ian
   8. between the diphthong <e_o> and a back vowel - /G``/; e.g. fle_ogan, dre_ogan
   9. between <i> and a back vowel - /G``/; e.g. wiga, nigon
   10. between a consonant other than <l> or <r>, and any vowel- /G``/; e.g. myndgian
6. Both PC2 and Orm had a two-fold approach to the spelling of the reflex of the OE prefix <ge>. Orm uses <y%e> with verbs (e.g. y%ehatenn), but <i> with pronouns and adverbs (e.g. iwhilc, imaen). Similarly, PC2 has <ge> with its sole reflex of a <ge> prefixed OE verb - also g`ehaten, and strangely <o> to represent the reflex of <ge> in OE g`enoh - onoh. I have reflected this dichotomy in the PC2 and Orm spelling options. In these two spelling options, only verbs take the <y%e/g`e> prefix. Nouns, pronouns and adverbs take the <i> prefix instead.
7. byrg`an > birien, due to support for that form in PC2, Lmn, AW and Ch (similar to herien and ferien); while herg`ian > both hery%en and hergen due to OE derivatives with probable <g> (G``) - e.g. hergat`, as well as modE harrowing; coincidentally, both of these forms avoid confusion with herien - to praise. Note: the eME normalised forms ferien and herien are the two exceptions to the rule that OE ~ian > ~en; that rule would have produced feren and heren if not for the peristence of the ~ien ending in Ch, in both cases. See the section on verbs for for more detail.