¹ The article by Kuhn and Quirk containing their objections appeared in Lg. 29.143-56 under the title Some recent interpretations of Old English digraph spellings. All three-digit page references given by us refer, unless otherwise specified, to that article, which is abbreviated as SRI. Our article appeared as Occasional Papers No. 4, of Studies in linguistics (1951), to which all one- and two-digit page references refer, unless otherwise specified. In the emendations, numbers in parentheses are page and line references to our article, which is abbreviated as OP 4.

A word about the symbols used for phonetic transcription, and about the printing of symbols that represent manuscript graphs.—Our phonetic symbols for vowels are those used by George L. Trager and Henry Lee Smith Jr. in An outline of English structure 11 (Norman, Oklahoma, 1951). The assignment of phones to phonemes in Old English is as follows: [i] and [Ɨ] to /i/; [e] and [∂] to /e/; [æ] and [a] to /æ/; [u] to /u/; [ō] to /o/; [ō] to /ɔ/; [ü] to /ü/; to /ö/. [α] was assigned by us in OP 4 to /ɔ/, the phoneme which is spelled a in the OE manuscripts, but we have since found reason to believe it should have been indicated to be a rounded vowel in the lower back corner.

There is an inconsistency between this article and OP 4 in the representation of ae and æ. In the printing of OP 4, it was found that the lithoprinter had no italicized æ. He therefore used italic a and e placed fairly close together, but rather closer sometimes than others. On the title page, æ was printed as ae. Since it is well known that æ and ae alternate freely and indiscriminately in OE manuscripts, we did not expect this would cause any difficulty, but a need for the clarification of it was indicated in SRI.

² Kuhn and Quirk have themselves pointed this out in item 3 of their conclusion (155).

³ Op.cit. In MdE all the possible combinations of /V/ with /y w h/ are demonstrated to occur. In OE, not all can be demonstrated to occur at any one period, but granted this kind of structure it is probable that many of the blanks were filled by unrecorded alternants that occurred in actual usage. At any rate, the assignment of specific spellings to specific slots is a problem in diachronic analysis, carried out by examining (1) what the nucleus now is; (2) what it can be reconstructed to have been in Pre-OE; (3) what course of change can be charted for its passing from one of these end points to the other without crossing and without violating reasonable postulates about the nature of the spelling evidence. It is not pertinent to list here in detail the specific assignments we make of all the OE ‘long vowels’ and ‘long diphthongs’ to /Vy/, /Vw/, and /Vh/, since the immediate problem is the short digraph æ and the short diphthong ea. A sample item with its complete diachronic development will illustrate the utility of this kind of frame. Pre-OE (the first one ‘standard’, the second in the dialect of some Philadelphians —cf. Smith, Lg. 28.147).

⁴ Interesting corroboration of this interpretation appears in Kuhn's review of Henning Hallqvist, Studies in Old English fractured ea, Lg. 26.319-23 (1950): ‘In Chapter 1, Hallqvist argues that a raised pronunciation of the first element of OE “fractured” or “broken” ea altered the diphthong from [æa] to [ea], or even to [ea] in southern England. The evidence consists of three groups of spellings: (1) ia, ya, ie, etc., as in -biare < OE bearu “grove”, or fiern- < OE fearn “fern” ...’ (320). Kuhn discounts (2) and (3), saying ‘The best evidence is found in the spellings of group 1’ (320). He goes on: ‘It is significant that the first element was raised so far as to be heard occasionally as [i] or [i]’ (320). Then taking up the survival of these items in ME, he agrees that ‘In general the area in which evidence of a diphthongal pronunciation is strongest coincides with the area of raising ...’ (321). Then ‘Chapter 3 deals with stress-shifted diphthongs, OE [ǽa], which become [æá], e.g. in ME yald- or yold (< OE eald), yern- (< OE earn). The evidence of Devon is decisive ...’ (321). We do not see how it is possible to gather from this material, added to the evidence of SME forms like dyaf < OE dēaf cited above, any conclusion except that all these SME spellings ia, ya, etc. are of the same structure in ME, regardless of diversity in source, i.e. regardless of whether they are obviously from original OE ēa, from lengthened ěa before lengthening clusters, or from ěa lengthened by processes not now described in the traditional handbooks but nevertheless lengthened.

⁵ J. E. B. Gover, A. Mawer, and F. M. Stenton, The place-names of Devon 106 (London, 1931-2).

⁶ Ibid. 66.

⁷ Lg. 28.278 (1952).

⁸ Gover et al. 32, 66, 314, 310, 60.

⁹ Of the sources listed in fn. 19 of SRI, we have examined both Mack and Einenkel closely, and d'Ardenne cursively. All of them lead to the same result, that no pattern of contrasts emerges which separates out OE ěa reflexes as phonemically distinct. Like OE ĕa reflexes, they are spelled e, a, and ea, though more often ea than OE reflexes are because of the lengthening clusters (earlier breaking clusters) which caused them to fall in with ēa.

Mack has conveniently listed, with a careful etymological breakdown, a vast number of the correspondences between OE as, , ĕa, ēa source items, in all environments, and their spelling in Seinte Marherete (EETS, OS 193; London, 1934). We cite a number of these below, with page and line references, for the reader to examine. The examples have not been verified beyond Mack's text, and the etymologies are hers. Even if Mack is not accurate in every detail, the evidence still is quite clear.

OE (Vespasian Psalter e, Mercian e): feder 4.16, fedres 4.1, feader 18.29, efter 2.3, et 34.21, pet 20.2, berninde 12.3, bearninde 42.13, hwet 8.23, hweat 30.7, 32.2, ber 46.15, bree 12.18, wes 2.19, hefde 4.20, wecchinde 36.19, wesch 44.8, beað 44.4, feat 40.18, weater 44.2, wleatewile 28.24, stealewurðe 36.18, 38.6; Mack calls the following forms ‘retractions’ from OE : war 38.10, unwarre 32.16, pat 52.28, blackre 24.22, attri 32.15; OE after palatal consonants: scher 50.29, bi ӡet 4.4, for ӡet 22.5, ischepen 46.21, frumscheft 46.20, schefte 26.25, nebschafte 10.11, schape 10.10, schal 6.16 (Mercian retraction of æ to a before l); deriving from Mercian forms with ‘back umlaut’ of : heatele 14.30, eateliche 28.11, teaperes 42.13, nease 20.27; due, according to Mack, to analogies of various sorts: Þeauie 32.21, gleadien 48.24, glede 16.13, heatieð 38.35, geapede 20.35, dearie 38.25, fearen 44.22, forfeare 18.14.

OE ěa: OE ea + r + C: bearmes 52.12, dear 38.25, earme 28.14, hearm 20.5, wearð 10.2 (elsewhere warð), smertliche 50.28, scherpe 50.29; before lengthening groups: bearn 24.29, bearnes 52.17, eardið 22.16, bern 38.31, berd 20.23; a is found uniformly after w except in wearð 10.2; smeortliche 22.8 (perhaps through influence of smeorten); OE ea + r + c, k (with Anglian ‘smoothing'): sterke 36.18, sterclukest 32.33, sperclede 20.28, merke 12.12 (perhaps ON merki), stearc 20.33; OE ea + r + C + i/j (Anglian e; WS ie > i, y): merren 10.4, derue 28.15; ideruet 36.21, snercte 42.15, ӡerde 26.23 (before lengthening cluster); from Anglian æ (umlaut of unbroken a): charden 8.12, awariede 12.17, dearne 18.25, dearnliche 32.15; OE ea + h + C (Anglian æ by ‘smoothing'; LWS e): feaht 52.27, feht 4.3; mahte 16.18, 2.16, strahte 20.34, arahte 2.15, waxeð 24.32; Anglian a + l + C (WS ea): alle 4.7, fallen 32.2, halt 48.3; Anglian al + C + i/j (Anglian æ alternating with e): welle 14.6, smellen 10.26, auellet 26.7, melten 14.33, helde 4.20, welden 4.28, wealden 14.18, wealde 48.31, wealdest 22.15, afeallen 24.4, mealteð 36.9.

OE ēa: bileaue 24.32, deade 2.20, deað 2.3, earen 4.7, heaued 44.15, leat 30.3, beleue 6.25, beteð 12.16, deð 12.13, bred 20.18; OE ēa + ӡ, h, c, Anglian and LWS ē: deh 2.17, dreh 4.6, ehnen 20.2, heh 8.10, hec 34.14, steah 2.5, beah 50.21; OE ēa + i/j (non-WS ē): heren 36.2, nede 38.18, schene 10.11, leue 12.3.

OE ¹ (WGmc ā): dede 52.2, strete 40.29, forletest 14.16, drede 12.11, hwerto 38.21, wepnen 32.32, beren 52.11, breken 44.16, cweÞen 8.13, isehen 16.28, read 14.10, reade 14.22, Þear 14.10, fearlac 22.5, unmeaðliche 34.28, reasde 24.10; OE ² (OE ā + i/j): biteache 46.12, cleane 32.1, gleam 20.32, heale 14.21, healent 2.18, heaÞene 14.28, leade 32.24, leaf 8.10, ear 16.3, bitechen 12.7, glem 28.31, flesch 16.5, unwreste 32.11, lesteð 2.23, heÞene 4.16, ilened 2.23, lest 30.18, lef 28.8; with shortening of : ledden 46.5, Þreste 20.28, lefdi 12.4, leafdi 28.8 (8 times), wreððe 42.31, wreaððe 22.11, wraððe 16.2, fleschliche 34.22, wreastlin 32.26, totweamde 40.24, healewi 32.16, earst 32.14, hahte 4.30, lahte 20.29, bitahte 4.26, rahte 24.11, bitaht 6.25.

OE e remained e: astenche 28.24, besmen 12.17, biset 10.23, helde 4.20, etc.

None of these lists are complete. We make nothing of them except (1) that the regularity is not especially astonishing when one looks at them to see if ea can be isolated out as representing a separate and distinct structure point, and (2) that they certainly do not reveal a pattern of contrasting reflexes of OE and ěa.

¹⁰ Thomas Wright, Anglo-Saxon and Old English vocabularies 1.258-83 (London, 1884).

¹¹ Historische Grammatik der englischen Sprache §197 (1921).

¹² Miscellen zur angelsächsischen Grammatik, PBB 9.206 (1884).