ABSTRACT A STUDY IN PHYLOGENETIC CHANGE: OLD ENGLISH /d/ >> MODERN ENGLISH /a/ REEXAMINED by Hans Frederick Fetting The sound change from OE /d/ to Mde /a/ before /vr/ as, for example, Egggg >> father is a.point that has concerned many writers of the handbooks of English phonological history. The established traditional doctrine recognizes two major stages in the develop- ment of /d/ > /a/ after stressed vowels and resonants and before /(V)r/: (1) CE /8/ > /d/ in the environment /ve-rv7; (2) during the Middle English period, this /d/ along with all others in the environment /V¥-(V)r/ became /a/. This explanation fails to account for the development of such forms as fodder << OE féggg, ladder << OE hlédgdzer, and the like. In an attempt to discover the mechanism of this sound change, as well as the reason for the lack of change, the present study_undertakes several different approaches to the problem. First a structural study, utilizing diachronic formssound charts in order to determine the distinctive phonetic features, reveals Hans F. Petting that a front vowel conditioning factor determines whether or not an 0E /Vd(V)r/ form'undergoes change to ModE /§aVr/. The same analysis, however, also reveals the possibility of CE /d/ >> ModE /a/ after long vowels. Neither analysis, however, is without exception. The next approach, a historical survey of the phone-graph characteristics of _d_ and 3!}, suggests a possible phonetic reinterpretation of the graph 51 as the fricative [a] or the aspirated dental stop [51h] when it is in medial position before /r/ in the next syllable. This conclusion is based on orthographic, orthoepic , and language contact evidence. Finally, the study turns to modern theories of sound change (those particularly of Chomsky, Halls, and Postal), which propose that nonphonetic conditioning factors determine sound change in many instances. Some possible nonphonetic and/ or extra-phonemic answers to the question of the mechanism of the sound change stem from such considerations as : homophonic aversion (e.g., /foder/ 'food' :: /foaar/ 'load'); functional shift (e.g., substantive > verb as £232 (n.) > to father, but not ‘53 adder or ‘39 ather < adder (n.)); Hans F. Petting or the stress readjustment factor concomitant with functional shift (e.g., fader (n.) > to fhther) under sentence stress. Apparently no single methodological or theo- retical system can cleanly cut through the complexities of this sound change. However, new insights into these complexities can be found and some satisfaction gained even in.partial answers and interesting pos- sibilities. A STUDY IN PHYLOGENETIC CHANGE: OLD ENGLISH /d/ >> MODERN ENGLISH /8/ REEXAMINED BY Hans Frederick Fetting A THESIS submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of English 1970 TABLE OF CONTENTS Page LIST OF SYMBOLS AND ABBREVIATIONS . . . . . . . . iii Chapter I. INTRODUCTION AND PROBLEM . . . . . . . . 1 II. THE TRADITION . . . . . . . . . . . . . . 9 III. THE EVIDENCE AND TRADITIONAL ANALYSIS . . 25 Iv. ON THE PHONETIC MANNER OFIQ . . . . . . . 41 v. SOME NONPHONETIC FACTORS . . . . . . . . 68 VI. SUMMARY AND CONCLUSION . . . . . . . . . 89 Appendix I. THE CORPUS . . . . . . . . . . . . . . . 94 II. SCHEMATIC DIALECTAL—DIACHRONIC FORM CHARTS O O O O C O O C O O C O O O O 0 100 III. SOME MIDDLE ENGLISH DIALECTAL DI~ TH FORmOOCOOOOOOOOCOvooOll-3 BIBLIOGRAPHY OF WORKS CONSULTED AND CITED . . . . 118 ii LIST OF SYMBOLS AND ABBREVIATIONS Symbols are used conventionally as, for example: OE Old English ME Middle English EMOdE Early Modern English Sc Scottish N Northern WM West Midland EM East Midland S Southern 5 A.macron or a breve over an Old English vowel is a quantity mark. [--] Brackets indicate a phonetic transcription. /¥-/ Virgules indicate a phonemic transcription. V‘C‘ Any vowel, any consonant. Underscore indicates graph or grapheme. Raised dot indicates phonic length. /V/ Acute accent indicates the suprasegmental phoneme of primary stress. > Becomes. < Comes from. >> or << As above but with intervening stages of development. (-—) Parentheses indicate an option or a variant. - Alternates with. Erratic underscoring has been normalized for con- sistency. iii I INTRODUCTION AND PROBLEM I shall be concerned here with what is noted by some linguists as a fifteenth century sound change in Middle English, which in its simplest form says: Old English postvocalic /d/ becomes Modern English /a/ before /vr/. The problem first came to my attention when it was presented as a student's exercise in phylogenetic change at the end of Chapter Forty-four of Charles Hockett's A Course in Modern Linguistics. In this problem, Hockett lists thirty Old English words, transcribed phonemically, with /d/ or /9/ after a stressed vowel, along with.their Modern English reflexes with either /d/ or /8/ in the corresponding position. Hockett states the problem as follows: "Describe what has happened to OE /d/ and /8/ in this position by NE [MOdern English] times, in such a way that if we were presented with an OE word not on the list, we would be able to assert definitely what would happen to any properly located /d/ or /8/ in it."1 The list reads: /°'dela/ Essie. /6'geer/ either /bodi§/ body 1Charles A. Hockett, A Course in MOdern Linguistics (New York, 1958), pp. 335:333. /bri'del/ bridle /br6'60r/ brother2 /féder/ father /féflm/ fathom /fé6er/ feather /f5r9ung/ farthing /ffir90r/ further /géderian/ gather /hé°8en/ heathen /hider/ hither /hré90r/ rather /hwé9er/ whether /léGer-/ leather /mé°dwe, -wa/ meadow /m6'dor/ mother /né'del/ needle /né6era/ nether /nOrGerne/ northern /6‘9er/ other /ré’dié/ ready /rfidié/ ruddy /Skédwe, -wa/ shadow /sfi'6erne/ southern /trédel/ treadle /9ider/ thither /wéder/ weather /widuwe/ widow 2Note that in Old English there is one (inter) dental fricative phoneme /9/, having two allophones: (l) [8% which occurs singly between voiced sounds, and (2) [9‘ which occurs otherwise (i.e., initially, finally, doubled, and next to a voiceless sound). The phoneme /a/ is a Late Middle English development. These words illustrate two different phonetic processes, the phonemic shift and the phonemic split. A.phonemic shift is the bifurcation of two phonemes out of the allophones of one original phoneme. Formulaically: /A'/ /A/ > /2. (5/, a process in which the phoneme in the ter- minal pattern is of a different type from the original and is not identical with any phoneme already in the language. 0n the basis of these words, the solution to Hockett's problem seems to be that OE /d/ after a stressed vowel and preceding OE /er/ or /Or/ becomes MOdE /a/ (a phonemic shift) but remains /d/ when followed by any other combination of phonemes; and that OE /6/ after a stressed vowel and before any vocoid becomes MOdE /a/ (a phonemic split). This latter process, the phonemic split, is an important type of change which results in a new phoneme. On the phonetic level, there is no observable change. In Hockett's corpus, each of the intervocalic /6/'s = [a], which is pho- netically identical with its Modern EngliSh reflex, intervocalic /a/. The same process is at work when OE /f/ and /S/ become ME /v/ and /z/, respectively; as, for example,/raven/ < /hrafn/ and /mazen/ < /masian/. Unfortunately for the student who puts his trust in this simple formulation of the conditioned phonemic shift, that the OE /d/ > MOdE /a/, a little search into the native stock of the Medern English lexicon will turn up words that have not behaved as he would have predicted, such as f6dder << OE fgdgd20r; ladder << OE hlid(d)er; madder << OE Eé§2£§3 fidggg << OE Egg. This point has occupied many philologists and historical linguists concerned with English phonological history. Some handbooks indicate that the change undergoes intermediate stages of develop- ment in Middle English and Early MOdern English and is affected by dialectal differences as well. Many suggest that the data which present themselves are very complicated, involving at least two sets of less 5 than perfectly understood combinatory changes at four hundred or five hundred years' distance from each other; namely, that in the tenth century. /9/ after vowels and resonants and before /(e)r/ became /d/, but that in the fifteenth(or beginning in the late fourteenth century) this /d/, along with all others similarly placed, became /8/. For example, OE féggr > ME fédgr > ModE feather to illustrate the late Old English change /9/ > /d/; and OE Eggs; > MElféggg > MOdE father to illustrate the late Middle English change /d/ > /a/. Only a few linguists have attempted to establish the sound change patterns involved here. My purpose in this study is first to make a historical inquiry into the scholarship concerning OE /vd(d)(V)r/ >> EModE /V8Vr/. I shall begin by looking at scholarly studies from the late nineteenth century, which is the latter part of the period when the notion of history began to be applied to language in the same way that it had been applied to other historical phenomena-eby establishing the relationship between events through time and space. It is the flowering of the period when scholars became interested in explaining language change in the narrowest sense of what we now call "linguistics." Secondly, I shall make a structural study of the relevant corpus in the manner and fashion handed down to us by the neogrammarian tradition of the nineteenth century, in an attempt to formulate this sound change in terms of the phonetic conditioning factors, which traditionally determine change. Then I shall consider the orthoepic evidence-~an increasingly popular type of inquiry-u—particularly those state- ments concerning the phonetic quality of the graph‘d. Next, language contact evidence will be considered; and finally, as a footnote to this study, I shall look at a recent phonological theory which suggests that the specifications of some phonological changes can be explained with non-phonetic information; that is, with reference to morphophonemic and/or superficial grammatical structure. This is the view suggested by phonological work done within the framework of generative grammar, a theory that is consistent with older theories (which are purely phonetic in appli- cation) but which adds more abstract environments to the phonetic character of sound change: namely, that part of the conditioning environment affecting sound change involves surface constituent Structure; changes happen only in particular form classes or stem types. Granting its heuristic quality, as one must in speculative inquiry, my hypothesis is that in Old and Middle English, the single grapheme g, postvocalic and before (1);, never stands in con- trastive distribution with any allograph representing the voiced (inter)dental fricative [a] in any one English dialectat any one time, and is therefore, in free graphic alternation with th and its variants in this position. The exact phonetic realization of,d in this environment is ambiguous, but it is probably not a voiced alveolar stop; its phonemic indistinction is clear. Therefore, I propose that the isolative change of OE /Vdvr/ >> EMOdE /var/ is more graphic than phonetic; it is probably owing to a long-standing sound/symbol variant interpretation of the graph Q, which has finally been resolved in Modern English spelling and pronunciation. As a result, now‘g a [d] in all environments; and the older, ambiguous, intervocalic g has been replaced with the digraph th = [a], which more nearly reflects the Late Old English and Early Middle English inter- vocalic /d/. I further suggest that a structural study (analogous patterning) cannot conclusively reveal a conditioned sound change. MOreover, the application of the broader concepts underlying modern theories (that extra-phonemic morphological and suprasegmental factors might well have affected this not—so—simple sound change) can Offer only some very interesting footnotes to a phenomenon which is complex and COD! tradictory. The sound change OE /d/ >> EMde /5/ before /Vr/ has not been and apparently cannot be conclusively explained with any methodological techp niques available to the historical linguist at this time. II THE TRADITION Henry Sweet was one of the first to deal with the problem of OE and ME /l?d(V)r/ becoming EMOdE /fiavr/. As a university undergraduate, Sweet wrote for the 1869 meeting of the Philological Society a paper on the question of OE /6/ and /d/, predicated chiefly on a physicalist principle of change: If we compare the two extremes, Latin pater and English father, an exami- nation of the various forms will soon COD? vince us that these changes are due to as- similation. The most abrupt transition possible is from a vowel to a voiceless stopped consonant, as in Latin ater, which has every right to be considered the ori- ginal form. In the Gothic fadar, the first stage of assimilation is entered upon; the voice runs on without interruption through the whole word. Finally, in the English father, the d is further approximated to the adjoining vowels; not only is the voice continuous, but the voiced breath flows out continuously. If the th were to undergo a further change into an I, the combination would almost amount to a regular diphthong.1 ———v 1Henry Sweet, "The History of the TB in English," Collected Pa ers of Henr Sweet, ed. Henry—C. Wyld, or m1, , p. fiBS. 10 If we forgive Sweet Obscurities such as the value he gives Gothic d or the reasoning behind his suggestion that a fricative becomes a lateral, he fairly well typifies one facet of the neogrammarian school of the late 1800's in his view that sound change proceeds from a striving toward an easier manner of articulation or that sound changes are always based on a decrease of effort. The theory is perhaps well exemplified in original /t/ > /d/ as IE ‘pgtgg >> OE £2933, a fortis > lenis weakening, followed by /d/ > /a/ as ME gags; > EModE father, a reduction step to spirantization which involves fewer muscular actions than a stop articulation. HOwever, analogous patterning suggests that the next step would be /5/ > 23;; as in Latin patrem >> French.p§£g, Provencal aire, Catalan.pa£g rather than /a/ > /1/ as Sweet suggests. This concept of simplification of pronun- ciation is, I think, held by no modern except in re- gard to unfamiliar sounds and the difficulties a language learner has regarding imitation of the unv familiar. The "ease" concept runs counter, further— more, to develOpment in the Opposite directionF-for example, /d/ > /t/ (OE tigg << IE fggzkg) and /t/ > */ts/ (ORG.E§EEE.<< Primec ftgggg). 11 In a later revision of that 1869 paper, Sweet moved to a behaviorist position when he stated that /d/ becoming /a/ " . . . resultfs] from relaxation of articulative energy modified by assimilation tendencies. Of all articulations the stopped con- sonants require the greatest exertion. [When it does not change, it is owing to] . . . the tendency to save trouble by continuing a given formative position unchanged, or with as little change as possible."2 Thirty-one years later, Sweet passed over the matter by writing, "In First MnE [1500-1600] (d) preceded by a vowel and followed by (r) was opened into (a) in many words, such as father, 39: ether, hither = OE £5933, Late Mqugggg, fgdgy, OE t6 edere, mpg.” As one moves through a historical survey of the change in question, one notes that modern scholars tend to EModE /EBVI/. wright states it thus: Intervocalic d followed by‘g in the next syllable became in the first instance 3 in all dialects . . . . This 5 from g (OE fader etc.) fell together with OE g in the same position (OE feaer etc.) and underwent all further changes in common with it. It has thus become (1) d beside Q; in . . . [Lake Country]. (2) g in . . . [the extreme Northern and Kentish areas]. At first it might seem as if forms like fada(r, ade(r, muda(r, etc. had retained OE‘Q and that forms like fadaa(r muda (r, etc. represent the intermediate stage of develOpment of OE‘Q toig in this position, but from the fact that words like brother, feather, leather, other, rather, weather, etc. have had the same development 13 in these dialects, it is clear that the'd, g; in the former class of words started out fromg.4 Wright's comments are typical of most twenr tieth century writers on this topic--terse statements that the phonetic change /d/ > /a/ did indeed occur, but without explanation of ghy in some words and ghy Egg in others. Of course, some descriptivists feel that ghy is not in their purview, but only'yhgt. I think, however, that ghgp, fully explained, can answer ghy. Early twentieth century German writers of philological monuments have shown deeper insights into the problem and have offered fairly full ex- planations. Richard Jordan restricts the fifteenth century change of /d/ to /a/ before /(V)r/ to the 4Jose h wright, En lish Dialect Grammar (Oxford, 1905 , sec. 297. Note Wrigfit's statement, " . . . it might seem as if forms . . . [with da] represent the intermediate stage . . . ." Apparently infelicitous readings have given rise to the suggestion that /dB/ is an intermediate stage of development (of. quotations—from Jordan and Schlauch, ff.). wright did not intend this, and the evidence suggests nothing more than that /d5/ forms are a late (seventeenth cen- tury) dialectal feature of the Lake Country. Andrew Ellis, 9g Earl En lish Pronunciation (London, 1869), p. 893, caIIs the E o? gph a literary man's diacritic for indicating voice. 14 short or single /d/, positing a doubled or lengthened /d'/ in words like adder, bladder, fodder, and udder, where a stop remains. While g before non-syllabic g (and other liquids) results in the stop g at an early stage, g, on the other hand, when preceded by a vowel and standing before syllabic g or 33, softened to g by way of gg from 1400 on. Thus fader > father (d), mOder > mOther, correspondingly gatherLen), together, wgther, wither(en), hither ShEther), thither, whither; in addition cosetherygg a considering . . . . The intermediate stage gg was preserved in northern dialects and partly reverted to g (Murray 121, wright sec. 297): cf. adther, hydther, etc. in Tyndale, which may, however, be compromise spellings. Lengthened g before 3 was pre- served in . . . adder, bladder (from addre, bladdre . . . OE dere, blsddre), fbdder, udder (OE deder, ‘udder); This shows that the beginning of the change of the g goes back to the time when there were still long consonants. Yet blather (1520), uther (1515), and corresponding forms in living dialects also occur with the softening of shortened 9.5 5Richard Jordan, Handbuch der mittelen lishen Grammatik, Part I: Phonolo (ReIdEIEerg, I925, p. 252, sec. 298. "Wfififend d vor unsilbischen g (und anderen.Liquiden) fruh den Verschlusslaut d ergab (sec. 206), wurde umgekehrt‘g, dem Vokal vorhergIng, vor 15 This analysis is cogent and clear except for the abstruse reference to compromise spelling. It is not clear what is being compromised. It is apparently Jordan's way of saying that Tyndale is aware of /d.-5/, so he puts down his graph for both phonemes. This does not, however, account for Wright's modernexamples. Karl Luick gives quite a full discussion of the change. He agrees with Jordan that it occurs only with single /d/, not doubled, and also not in combination with a preceding liquid or nasal. But Luick does not make the matter out quite so simply as Jordan and wright, for he admits the difficulties of chronology that remain. Most significant are silbischen Oder er von 1400 ab fiber gg zu‘g gelOst. So fader > ther (a) 'Vater', m6dor,> mUther 'Mutter', entsprechend gatHer(en) 'versammeIn', to etfier 'zu— sammen', wéther 'Wetter', witherQen) 'verwlttern', hither Shétfier} 'hierher', thither'Tdorthin', whither 'wOEln'; azu cosether = conSIderin . . . . DIe Zwischenstufe dd erEleIt sicE in nbrdIlchen Mundarten und kehrte teiIWeise zu'g zurfick (Murray 121, Wright sec. 297); vgl. gadther, hydther u.a. bei Tyndale, die aber auch Kompromisschreibungen sein kOnnen . . . . Bewarht blieb gelangtes d vor 3 . . . . . . . adder 'Natter', bladder 'BlaseT (aus addre, bladdre . . . ae. gddre,-BIIddEe) deder 'Futter', udder 'Euter' (ae. ffiddor, =udder); dies zeigt das der Anfang der Veranderung des d noch in die Zeit der erhaltenen langen Konsonanten zurfickgeht. Doch be egnen auch mit Erweichung von gekfirztem.g blapheg €1520),,g§hgg (1515) und entsprechende Formen in lebenden Mundarten. 16 his notes: Note 5. The phonetic process is not easy to completely clarify. It is very likely that an interdental or postdental Q, such as the northern English dialects still exhibit today, existed as a transitional stage be- tween the original dorsal d and the present-day interdental or postdental g. But the cause of the shifting of the g is not clear, for trilled 5 follows dorsal 9 without diffi- culty. Perhaps initially the entire group was fronted and g was moved to the foremost position as a result of a certain exaggeration which can frequently be observed in the case of newly appearing sound tendencies. The movement from interdental or postdental d to g is liable to be accompanied by friction, as can be readily observed, and this friction has eventually prevailed in many cases. Note 4. (Chronology) Judging from the time of the first appearance of written evidence, this change is most likely to have taken place in the fourteenth century, or in the fifteenth century in some parts of the country. It seems, in any case, to have been colloquial in Character to begin with, and has, consequently, prevailed only slowly in the literary language. 6Karl Luick, Historische Grammatik der eng- lischen S rache (Leiszg,1914—19EO), I, Part 2, p. IOI2, sec. 52. "Anm. 5. Der phonetische VOrgang 17 In a note, Karl Brunner refers the reader principally to Luick for discussion of the change. He does, however, find the change of /9/ to /d/ before /r/ regular, but either divided on the basis of origin (as before /l/) or not clear (as before /m/) when the /6/ stands before other resonants, though the change before any of these resonants seems to be part of a single phenomenon. Actually, Brunner gives only WSax 5dr beside Mercian éar as examples ist nicht leicht vOllig klarzulegen. Dass als Zwi- schenstufe zwischen dem ursprfinglichen dorsalen Q und dem heutigen interdentalen Oder post dentalen g ein interdentales Oder postdentales d, wie es die nordenglischen Mundarten noch heute Eieten gegolten hat, ist sehr wahrscheinlich. Aber die Ursache der Verscheibung des d ist nich klar. Denn an ein dor- sales d schliesst sich Zungenspitzenfig ohne weiteres an. VIelleicht wurde zunfichst die ganze Gruppe vorgeschoben und dabei mit einer gewissen Ubertrei- bung, wie sie bei neu auftretenden Lauttendenzen Ofter zu beobachten ist, das g bis an die vorderste Stelle gerfickt. Beim Ubergang eines interdentalen Oder postdentalen.d zu r stellt sich aber, wie der versuch zeigt, leicht eIn Reibegerfiusch ein und dieses hat schliesslich haufig die Oberhand erhalten. Amn.4. (Chronologie). Nach dem Zeitpunckt des ersten Auftauchens von Belegen muss sich dieser Wandel wohl im 14. Jahrhundert, in manchen Landes- teilen wohl erst im 15. Jahrhundert vollzogen haben. Allerdings scheint er zunachst umgangsprachlichen Charakter gehabt zu haben und ist daher in der Gemeinsprache nur langsam durchgedrungen." 18 of the alternation before /r/, and no examples with /V1c'/.'7 A present-day historical linguist, Alistair Campbell, discusses only the various changes of /9/ in Old English as " . . . a tendency to develop stops from spirants before liquids and nasals,"8 rather than as an invariable change; and he also distinguishes changes peculiar to individual dialects from those characteristic of Old English as a whole and tries to compare them with parallel changes in Low German and Frisian. He explains that the ten- dency for g; >Igg as in.§gg 'vein' is Kentish, while Anglian (which is the proper developmental forerunner of Middle English) kept the spirant, which ultimately became voiced after long vowels, as in Eggg. Too, Campbell notes that in Anglian and Mercian, g before 1 equals 3 [C]. Campbell relegates the change /6/ > /a/ > /d/ in words like fedm to West Saxon; preserved 7Karl Brunner, ed., Alten lische Grammatik by Edward Sievers, 2nd rev. ed. (EEIIe, 1951), p. 175, n. 7. 8 . . . Alistair Campbell Old English Grammar (Oxford, 1959), p. 171, seé.”EI9. " 19 forms with /5/ remain in Mercian, and Northern has /a/ beside /d/.9 Campbell does not give any examples involving /er/- /dr/, or indeed, any in which /8/ or /d/ precedes vowel plus resonant except heairlo 'kidney, rein' as an example of a retained spirant following a long vowel. E.J. Dobson alludes to Jordan and wright in his comment on the change of /d/ > /a/, mostly preliminary to the period with which he is concerned. He writes: " . . . about 1400 postvocalic [d] became [a] before [r] or [er] in father, mother, to ether, weather, hither, etc., perhaps through the interme- diate stage [d5] preserved in.NOrthern dialects (see Jordan, sec. 298, but contrast Wright, sec. 397)."11 Dobson adds in.a note: "The sound change —____ 9Campbell, secs. 423, 424, p. 172. 10This word is a good example of what is noted by Henry Sweet as an advanced, less clumsy form in which a is substituted for an ambiguous _d_; 1.6., d-== [87. The Oldest En lish Texts: Cogpus Gloss TE.E3.T.S., London, I885), OriginaI Series, 35, p: 3. ll . . . Eric J. Dobson En llsh PronunCIation lECK>—l700 (Oxford, 1957): II, p. 955, sec. 535. 20 [d] > [a] is restricted to the case where a vowel precedes."12 Dobson's remarks reflect the established doctrine of historical English linguistics. This doctrine recognizes two major changes in the treat- ment Of OE /d/ and /9/ after vowels and resonants and before /r.~ r1~ vr/: that in Late Old English and Early Middle English, /9/ before /rV/ became /d/, but that in the fifteenth century (or beginning in the late fourteenth century) this /d/, along with all others similarly placed now before /rV/ or /Vr/, became /a/. One must be quite careful to distinguish between the environments of these two changes. The earlier change is indicated as being " . .-. vor unmittelbar folgendem nicht silbischem . . . 3;"13 that is in the environment /-rV/ to the exclusion of /-r/ or /-vr/. The second of these changes , the specific problem Of my investigation, is in the ene vironment /N;r/ or /v;vr/. 12Dobson, II ,p.956,sec . 584 . 13Jordan, p. 185, sec. 206; also Dobson, II, p. 954, sec. 581. 21 Many recent statements of the matter are Often oversimplifications. Fernand Mossé said only of the earlier of these changes: "2 (a) before g, g, _r, _1 becomes g: OE £2153 'fathom' > gaggle, OE bygaen >1ME birden."14 Eilert Ekwall, although not principally concerned with historical phonology, passes off the later change with, " . . . a change of g’to Eh is well known in words like father, mother (Old English.£gdg£, 15 modor)." Margaret Schlauch's dressed-up, but equally pat version reads: "The voiced stop [d] became the spirant [C] by way of [d8] before unaccented -er [er], as in mother < ggggg, gather < gaggg, weather < weder."16 More recently, Hans Kurath stated: "Medial [a] is in part derived from earlier [d] in the cluster 11+Fernand Mlosséz Handbook of Middle English, trans. James A. Walker BaItImore,-I952), p. . 15Eilert Ekwall The Concise Oxford Dictionary 2; Egglish.Place-Names (OEIOrE, I953), p. xxxiv. 16Mar aret Schlauch, Th2 English Langgage (Oxford, 1959 , p. 47. 22 [dr] as in father, ather, hither, mother, weather, whither. This change took place in late ME. As a result, weather became homOphonous with wether."17 The final interpretation in this historical survey is somewhat out of the purely phonological mainstream of most of the above explanation. It has influenced my thinking and suggests the strong pos- sibility of the necessity for a phone-graph rein- terpretation. It is Otto Jespersen's physiological explanation, which strongly echoes Sweet: The vacillation found between /8/ and /d/ especially in the neighbourhood of 3 must be explained through the interdental stop, an- alphabetically written pod: the tip of the tongue forms a stOp with the lower edge of the upper teeth, I heard this sound most distinctly in 1889 in the Yorkshire dialect of the Rev. C.F. MOrris, who writes it ggh‘before,(gz§; it is also found, I think, in the Irish pronunciation of loudShZer, broad(h)er. It may be popularly described as a [d] formed where [a] is usually formed or a [a] exaggerated to a stop. 17 Hans Kurath, A %onolo pand Hos Modern English.(Ann Arbor, 1965),17. 18Otto Jespersen, A Modern lish Grammar ' (Heidelberg, 1928), I, pp. .25832'59, sec. .2. See Efi. p. 387. sec. 14.02 for discussion of Jespersen' s an- ‘ alphabetic notation system: Greek letter . articulator, number a point of articulation, raised Roman a manner of articulation. 23 The change of postvocalic /d/ to /a/ before /Vr/ from Old English words to their MOdern English cognates has concerned not a few scholars. Historical phonologists agree that the change took place. And certainly it did happen if one accepts historical phonology as they present it, but maybe it did not happen at all. Certainly the analyses of the change have been serious, but they are not always neat. There are uncomfortable lapses and explanations that some teachers and students feel are not really ex- planations at all. It is reasonable to expect from the neogram- marian tradition an analysis plus perhaps a list of exceptions to whatever general pattern is established. All of the above analyses have in some form or other relied on analogical patterning, which in itself is fine, but by itself, I suspect, denies a regular phonetic change. Most present-day linguists sub- scribe to the notion that all sound changes can be described in purely phonetic terms and that any list Of exceptions can then be described in the same way until the exceptions are explained away, and the sound change can be described by a series of sound change patterns that will account for all the developmental 24 forms involved in the sound change. None of the above analyses does this. It is this lacuna that the next chapter will attempt to fill. III THE EVIDENCE AND TRADITIONAL ANALYSIS The dominant view of sound change is and has been based on the autonomy of phonology, . . . that in general, phonological change takes place under conditions and within limi- tations which are in.phonological terms . . . . The conditions under which.particular changes [take] place [involve] features such as voicing or voicelessness of neighboring sounds and the positions of accent with regard to the sound in question . . . .1 The application of such a theory to the change in question calls for a structural study, a part of which requires setting up charts which dialectally and chronologically illustrate relevant forms and should thereby reveal the sound change in process and in state. Turning to three principal sources: the Oxford English Dictionary'(9§2), the Middle English Dictionggy (Egg), and Th2 Engl'sh Dialect Dictionagy (EDD), I examined the entire corpus (including all 1Charles A. Ferguson, Review of The Sound Pattern of RuSsian by Morris Halle, nggggg§_XIHVIII: Wy-SEW. , 2), 289. 25 26 recorded spellings) of uninflected, non-derived, Modern English words which contain the graphic se- quences vdgdzvr or 23px;, which developed from cog- nate l’fd(d)(V)r forms in Old English.2 The relevant words are MOdE ggggg, bladder, ladder, madder, Edggg, fodder, mother, father, ather, hither, thither, to ether, whither, weather; their histories are clear and well attested. They are un- questionably active, native vocabulary. Several words: for example, MOdE crowder (crowther) < ME crouder (crowdere, crowther); MOdE dither < ME 7didderen; Mde dodder (dother) < ME ggggg; Mde heather < ME hather Shadder); MOdE slither < ME slitheran ('Islideran)3 have very uncertain or unknown etyma. Although important, they remain peripheral to this study; so do related words that do not enter into the mainstream of the sound change, such as 2See Appendix I for the words examined. Words in differentbut similar environments such as /ndvr/ or /VBm/, in which the same phenomenon may be supected, are noted. I have, however, restricted the study to the word types noted above. 3At first glance, slidrian > slideran > slitheren might appear to be a normaI EeveIopmenf for Ehis word, but there are two developments from two different words: OE sleaeran > ME slitheren > MOdE slither and OE slidrian >»ME sIeHeran > MEEE'HIEIT slidder. 27 farthigg and brothel, which likewise have variants withwg at the time in question. (One may hope that the conclusion of this study could offer some insights into the etymology of these words.) For this con- trolled study, we must, however, stay with CE frdgdzgvgr patterns having MOdern English reflexes, in respect of the tradition under study, for simplicity, and for clarity. A study of Appendix I shows that the none native vocabulary is principally from Scandinavian, Latin, and French. EXCept for sporadic anomalies, such borrowings come through Middle English into Early Modern English consonantally unchanged: for example, Danish buldre > ME boulder; L goderatus > ME moderate; OF giggg > ME giggg (either, cidr 2. Intervocalic /6/ (found only in nonpnative words) presumably derives from ME /t/ as in Mde author < ME auctor.4 “(SearJ is though: by some (e.g., Hel e KBkeritz, Shakespeare's Pronunciation (New Haven, 1955 , p. 520.) o ave eve Oped as a spelling pronunciation of the 1550 alternate aucthor for auctor. Since it is imr possible to say'wfien Ehe moHern pronunciation with [6] was established, it is equally valid to assume that the spelling arose out of the pronunciationp-the more usual direction when orthography has not been subject to the pressures of convention such as we presently 28 A number of borrowings are unattested, or they are relatively late forms; for example, pother dates from 1692 and is labeled origin unknown. Old English words having postvocalic (inter)dentals be- fore /(V)r/ retain those (inter)dentals in Early Medern English except for OE ppppp >> EModE rudder; OE spiara >> EMOdE spider. These two are apparent perversitiies until it is noted that the usual lexical form OE pppgp is related to and probably developed from the North Germanic cognate péppp, while all Low (West) Germanic cognates have medial gr-including, according to the _O_E_D, Old English. Thus OE _r6_g_u_r_ » MOdE rudder will be seen to be a regular change and ME rother, a Scandinavian borrowing. The early history of spider is extremely obscure, but the earliest written occurrence Of the word (in the West Saxon Leech-Book (9th century)) is spiden Swih 2. The Old English form I'spidra is a reconstruction as noted in some dictionaries, Webster's New World, for example. The usual developmental notation, experience--and even more so when one considers that the graph 2 had and.has long functioned as an aspiration diacritic, as in Greek‘pp a [p], th . 9A: Sanskrit bh s [b], dh - [a]; German EH - [t j; and was Edvocated‘TOr just that function by the orthoepist, John Hart, An Ortho a (1569). In WOrks, ed. Bror Danielsson (Stoc o m, 55), pp. 204-205. 29 ME spithre > §pyder, should read ME spithre ~.spyder > spyger, as given in the etymology of the word in Wébster's‘gpipg Np! International Dictionary. If the Old English cognates of rudder and spider do have medial st0p consonants rather than fricatives, then rudder will have undergone the same development as pgggp, etc., and spider can be explained as a retained or unchanged form. Its appearance with pp might be related to the tendency to change, or it might be a unique or unusual morphological develop- ment related to the adaption of the Germanic suffix eppg as seen in the form spinnan (v.) > I"spinara (n.). The great blur on the development of this word as a derivative from the verb spinnan permits little but the most speculative of analyses. The area identifications of the forms in Appendix I are based on the Middle English dialect 5 areas identified by Meore, Meech, and Whitehall. Dialect identification of OED and MED forms was 5Samuel Meore, Sanford B. Meech, and Harold Whitehall, "Middle English Dialect Characteristics and Dialect Boundaries," Essgys and Studies in En - lish and Co arative Literature,-UEiversity'3?‘MIEEF igan PESlications in Language and Literature, XIII (Ann Arbor, 1955), pp. 1-60. 5O established.principally, and insofar as possible, from.the identifications given in the MED "Plan and Bibliography,"6 or from those in the Cambridge Bib- 8 liography7 or from the wells' Manual, and occasionally from the introductions of scholarly editions of texts. On the whole, the forms were dialectally identifiable and sufficiently distributed to permit adequate analyb sis. Schematized versions of my working charts ap- pear in Appendix II. Diachronic formrsound charts revealed the following normalized developments of the relevant words in the corpus: 6Hans Kurath, "Plan and Bibliography," Middle En lish Diction , eds. Hans Kurath, Sherman K and 3ohn,Riedy, Ann Arbor, 1952-), pp. 11-12. 7T.W. Bateson, "General Introductign," Cams brid e Biblio re 2; En lish Literature Cambridge, mafia—1935‘? . pp. 36%. . 9 8John E. wells, A Manual 9;; the Writ' s in Middle En lish 1050—1305—353-SmppIemenEs I—IX 'ew Haven, [6- , paSSIm. 51 Group I EMde (16C) Late ME (15C) OE adder /§ds r/ adder /dddar/9 naddre bladder /blédar/ bladder /bléddar bliddre ladder /lador/ ladder /1§ddsr/ hl'iddre madder /médar/ madder /mdddor/ madere udder /fidsr/ udder /uddsr/ uder fodder /f5dsr/ fodder /f6(5)ddar/ f6dor Group II mother /m$asr/ moder [mcdsr/ mador father /féaar/ fader /fédsr/ fader gather /géaar/ gader /gédar/ gadrian together /tog;aer/ togider /togidsr/ tag-dere thither leiaar/ thider /eidor/ bider weather /w€aer/ weder /w§dsr/ weder whither /wiaar/ whider [hwidar/ hwider 9The positing of the double stop is, of course, arbitrary. The forms are idealized. The phoneme might have had phonic length /d3/. There are presently two schools of thought on this matter (relating particularly to Old English rather than Middle English, however). Quirk and Wrenn, for exp ample,interpret a double graph as representing a phonically long consonant, while Campbell considers the double graphs as two phonically short consonants. The whole contoversy is excellently summarized in RObert A. Peters, "Phonic and Phonemic Long Consonants in Old English," Studies ip Linguistics, XIX:1, 1-4. 52 In Group I, the number of gg spellings from original‘g increases after the year 1000. However, madder shows no gg form until c1550 and not constantly until c1550. Scottish Eggpp and.pgggp alternate‘g and.gg through the fifteenth century. Elsewhere these two show relatively stable‘gg forms. The remaining words, bladder, ladder, 23993;, and fodder, Show a slight tendency to a graphic shape with.:ppgp in two notable cases: (1) sixteenth century East Midland and (2) seventeenth century Northern; but the,:ggpp pattern occurs ninety percent of the time in all the dialects in all periods. Between 1500 and 1600 and continuing through the seventeenth century, one can find instances of pp forms of the entire first group in all the dialects. The history of the forms in Group II can be viewed from the following schema, which indicates the date and dialect of the first occurrence of the word with‘pp; date when the pp form took over as the dominant form (based on simple frequency of occurrence}; then some comments on what I thought to be most notable about these forms. In general the words in Group II are spelled with.:gpp throughout their early history, with‘:ppgp after 1525. .63 3...me opoodoao soon as possess sauce to .AoflB as 83 soosusoz 83o posses: .03 swoops» .M 3 scanned assoc w .63 smsosfi .z 5 on 3133mm o8 .8 fl odoaoaona assoc so 83 condos: .m 83o access: .8 one fl. assoc mm .2 03 as a on concedes wooden coma flospaoz Rio sossomop I .om oi as canon so ososogz .3an as 83 flosssoz 83o scopes» .om. pmcoxo peedwwdlos 5.” mpmwmncm pun .mnmcn lawman c mono pomp aw canwxnmswm .8 or.” 5 mason v.6 coma oncnpnoz oomao 923.3 .63 Amazes». .23 ..z ..om a.“ scanned manom w OOMH dd“ 3H a hHDHHflSUOHH AD. 3.. U oomH .0ande cm §HO Hmnpflm X/ Z/ .63 fl .om oaooxo opooamao So as Savanna wdauooaofiv s oosfiaoo .62.. success. .m can as esp 2s 3 reasons canoe so 1 o 83 seasons .m mamas Rosana .63 masons» .u 3 scanned assoc w .63 on» canons» .2 cos .8 as pdoasaosa one canon so 83 condos: .m ommao sospos mpneaaoo 028995000 In»: *om 0239308 In»! pmhah 34 The history of the forms in Group II words is confused and confusing. In the absence of a graph.reinterpretation, it is difficult to make any really discerning comments about this group. However, in general terms, the data on page thirtybthree support the general view that the sound change took place in the fifteenth century and was well on its way to completion.by the Early Mbdern English.period (1500). A.phcnological analysis of the Early Medern English.words from the list on page thirtybone reveals that in Group I the phonic environment is /Vdor/ where /E/ - /a/, /U/, /o/; in Group II it is /Vtsr/ where /V/ = /A/, /a/, lfiéf /2/. Or schematically: 'I I I I \s I I A k 9- This distribution suggests the possibility of a back-central versus a front-central vowel conditioning factor with an additional conditioning factor on the central /a/. Limitations can be placed on /a/ which in Group I occurs (1) initially, (2) after the cone sonant blend /bl/, (5) after the lateral [l/, (4) after the nasal /m/. In the second group, /a/ occurs 35 (1) after the voiceless labiO-dental fricative /f/, (2) after the voiced dorso-velar stop /g/. Thus considering only segmental features, the formulation for this change is:, E postvocalic /d/ > EMOdE /3/ before /ar/ except when preceded by (l) a back vowel or (2) a low vowel when free or when combined with a preceding lateral or nasal. If the allophonic values of /a/ were known and agreed upon, the solution might be a simple one. OE gpggp and gedrian contain OE,§, a front vowel, which gave rise to ME é, namely front [a’].10 On the other hand, ME.§ in pgggp, bladder, ladder, madder is usually held to have been, in the Early Middle English period, the back vowel [61.11 If these assumptions were true, there would be a clear separation of change according as /V/ is back on the one hand or front on the other. However, one need only turn to the discussion of the free development of stressed vowels in Kbkeritzl2 and Dobsonl3 in loDobson, II, p. 594, sec. 98. ll Dobson, II, p. 545, sec. 59. 12K6keritz, p. 167. 13Dobson, II, Chp. 6, passim. 56 order to be confronted with a lively (and yet to be resolved) controversy concerning the articulatory production of the allophones of /a/, the chronology of the raising of ME [a‘] to [m'], and its subsequent retraction and lowering to [o‘]. I personally have much faith in Dobson's theories, but it must be re- membered that his theories are Often based on his interpretation of orthoepic evidence, and that his theories assume a complete regularity of change. Assuming this back-front vowel conditioning and returning to the excluded words on the second page of this chapter, the normative developments crowder < crouder; dither < didderen; dodder < gpgpp are thus verified by the known data. Presumably, then, the alternate spellings or the divergent forms would indicate either a dialectal retention Of an older form or a dialectal fronting or backing of a radical vowel. Recalling the "quantity" theories of Jordan and Luick brings a slightly different result. There is an established tradition in Modern English that the vowel preceding a doubled graphic consonant is phonically short. Therefore, one can reasonably 37 assume that in the absence of the graphic geminate, the preceding vowel, though perhaps not long, is not markedly short. Some use the expression "half long." In other words, the phonic value of‘! in a,zgpp sequence is relatively longer than it might be in a Egggp sequence. Thus, on the basis of vowel quan- tity, reflected in the orthographic practice of the modern period, one can say that ME postvocalic /d/ > EMOdE /a/ before /ar/ except after a stressed short vowel, usually signaled in the orthography by*a geminate (phonically lengthened?) consonant. This analysis finds support in the theory of Hans Kurath,14 who in considering consonant quantity suggests that intervocalic /a/ < /d‘/ or that OE peppy a /fad'er/ >> ModE father = /faaor/; OE Eggpp = /fo’dor/ >> MOdE fodder = /fadar/. Such a theory based on single versus double consonants or quantitative phonetic features works well enough backwards, which is in one sense our aim -to posit the developmental forms of a word. But one must remember that a commonly used, authenticated 14Hans Kurath, "The Loss of Long Consonants and the Rise of Voiced Fricatives in Middle English," L a e, XXXII (1956), 455-445. 58 Middle English form might not be that which gave rise to the Early Modern English form. Reconstruction aids etymological studies, but it does not necessa- rily determine either a usual or a normalized pro- nunciation or spelling. "The aim of . . . structural analysis is, after all, to discover the extent of systemization in a language at a given time, not to impose a system on it."15 Either the preceding traditional analysis. based on discrete phonemic features and the environr ments in which they are found, or that based on vowel length could serve as an answer to those who are looking for purely phonetic phenomena as an explana- tion of the sound change under study here. Neither statement is, however, without exceptions and con- ditions because the sound change /dor/ > /aer/ is an isolated change, or perhaps merely a phoneme sub- stitution. The statements I have made differ from the well known sound laws of Grimm and Verner, which are net couched in terms of a single phoneme, but rather in terms of a class of sounds. On their 15Kurath, "The Loss of Long Consonants," 442. 59 example, what happened to /d/ should have happened to /b/ and /g/ in similar environments since tradi- tional theories hold that sound changes are regular. This is only true, however, when it is true. Lin- guists, like other scientists, aim to make their generalization as inclusive as possible. Grimm's statement, for example, is much broader than Verner's, but no less scientific; and so too, I should hope, are those offered here. Interestingly and unfora tunately there are no attested native words cognate in Old English and Modern English having the graphic configuration ngbzvr or figggzvr or with their re- lated fricative congeners [b] and [g], which have long been lost to Modern Englishe-and even Old Eng— lish can hardly be shown to have [b]. Without such sequences we cannot test distinctive feature formu- lation with finality. In addition, and so far as I can determine, no other MOdern English voiced fricative appears to have developed from a voiced stop in the environment /Y’—~V’r/.16 16An obvious example of a fricative developing from.a stop is ModE have < OE habban. The process is apparently: habe- > Bab - > Ha55-. However, in the second and EHird persons siEEEIEr and in the past tense, Old English retained hab- in the alternated havh ~vhaf-; thus OE habban,_EEIab (/f/ s [v], its 40 The difficulties of interpreting the sound change arise from the many forms over several centuries and into our own time. Meet of us would agree, I think, with Iakov Malkiel, who wrote: The weakness of.a phonological change may have a direct bearing on the infiltration of sporadic changes, lexical blends, and other modifications; it may also serve as an index of dialect mixture.17 What is perhaps the soundest phonological conclusion is to consider that both the segmental and suprasegmental factors are active conditioning forces; that one factor sometimes cancels out the other, and at times they work in harmonyb-just as Grimm's Law is modified by Verner's. Even together, they leave some problems unanswered, but with no large group of exceptions. intervocalic allophone). ME habb- forms were re- duced by leveling to hav—. TEese kinds of changes are described by the Egg-editors as weakenings. The QEQ records intervocalic /f/ as in hafu as early as Beowulf;‘y in this position is first recorded in 1225. 17Yakov Malkiel, "The Inflectional Paradigms as an Occasional Determiner of Sound Change," Di- rections for Historical Linguistics, eds. WinfEEE LeEfiann and YEEov MaIEieI (AuStin, 1968), p. 29. IV ON THE PHONETIC MANNER OF 2 The previous chapters of this study indicate an uncertainty concerning the theory and method of the change ME /d/ > EMOdE /5/ before /Yr/. This chapter, based on works concerned principally with the production of those sounds, is enlightening in that previous theories of the sound change very Of- ten are not based on any evidence of articulatory phenomena. It is difficult to write cogently about this point. One has to sense it, to feel it, rather than to know it. It has been so easy for historical line guists to be deceived (as it is for us all) through their own familiarity with spelling and the entire writing system. Nevertheless, I suggest that the many forms of my corpus appear to be the result of a long history of phonetic-graphemic ambiguity con- cerning g, which is clearly not always the voiced dental or alveolar stop which our handbooks would lead us to believe. Few handbooks of Old or Middle English suggest very forcefully that there is any alternative pronunciation for g other than [d]. Some do suggest the very general possibility in very 41 42 old texts. Quirk and Wrenn, for example, state that " . . . in the later eighth century the letter‘g was . . . often used for ([d] and [8])."1 Since all phonetic work in Old and Middle English must be based on written records, orthographic tradition plays an important part in any analysis. A succinct overview of orthographic practice comes from Kurath, who writes: The spelling of modern English is in a large measure a heritage of the late Middle Ages. With the emergence of English as a literary language in the fourteenth century, a system of spelling was devised to render the phonemes of the spoken language. At that time, ppg correlatigp‘between,gpp: pheme gag Ehoneme Egg systematic, although not all of the vowel phonemes were distinr guished in spelling. The introduction of printing a century later (1475) served to standardize this spelling, though it was already out of step with the phonemic system, which had undergone remarkable changes be- tween 1400 and 1500. During the last four centuries the two have drifted farther and farther apart. The spelling system of Middle English 1Randolph Quirk and C.L. Wrenn, _Ap Old English Grammar (New York, 1957), p. 8. 43 that stands back of the spelling of Modern English was devised by scholars who wrote and spoke Latin and French, the language of scholarship and high society. In the main, they adapted the spelling of Latin to rendering the English phonemes, but French spellings were retained in words taken from that language and in a few instances carried over into native words.2 Neither Middle French nor vulgar Latin had an interdental fricative phone, and no graph to rep- resent one. The digraph pp, as in Classical Latin spatha, is a strongly aspirated dental stop, as in the emphatic pronunciation of MOdE terrible.3 In Old French, pp is a development from.Gallo-Romanic intervocalic [dJ,as in‘ypgp, which became [a], as indicated by the Old French spelling‘ypppg; it then weakened to silence in later Old French and Middle 2Kurath, Phonology and Prosody, p. 55. Italics are mine. Although true in a general way, the statement implies a one to one relationship, a traditional tenet, which the following evidences do not, I believe, generally support. 3W. Sidney Allen, Vox Latina (Cambridge, Eng- land, 1965), p. 112. 44 French E94 Or more specifically, in late Old French, the earlier postconsonantal aspirate p became a mute (i.e., a breath stop, no longer a fricative), weakening [a] to inaudibility. As a result, the pp digraph disappeared in Old French only to be occasionally revived as a scribal variant in fifteenth and six- 5 teenth century France where it has been presumed to have been a non-functional mute (i.e., not pro- nounced).6 A usual linguistic practice is to fill an unknown quantity with the nearest known quantity. Long ago, Skeat made the following Observations on the orthographic practices of a Norman scribe ren- dering the Octavian Imperator: . . . It is worth while to see what he makes of the English.pp, a sound which it cost the Norman a good deal of trouble to achieve, though he learnt it at last. His “John H. Fox and Robin Wood, A Concise Histopy of the French Language (Oxford. 1968). p. 50. or Frederick B. Luquiens, pp Introduction to 9gp French Phonology ppp MO holo (New Haven, I959), p. 45, sec. ; p. 6 , sec. 5; p. 67, sec. 278. 5Urban T. Holmes jr. and Alexander T. Schutz, A Histopy of the French Langpage (New York, 1958), p. , sec. 557— 6OEID’ 80v. £0 45 method was simple, viz., to confuse it with p; and when once we know this it is easy to tell what he means. When he writes pp, he means ppp (105, 128, 206); but he only writes de occasionally, in moments of re- lapse. Skeat follows his comments with numerous ex- amples, which graphically illustrate a remarkable fluctuation: e o e Ede]? for tEZder, i.e., .thither' (257) . . . brod , for brotpyp, Oder, for other . . . . It is still more odd to find the scribe substituting‘pp for the English 9; as in onther, for 6ndgr, i.e.,‘under' (515), correctly spelt onder (with the NOrman'g for ‘3) in l. 550; thonryght (1114), a variant of donryght (1560), i.e. 'downright'; with other traps for the unwary.7 Since Latin and French were the languages of the literati, we may note references to mispro- nunciation in Latin and French before we look at some commentary on English. The earliest records that relate to the 7walter w. Skeat, "The Use orgp for g in Middle English " Notes and Queries, series I (October 25, l 2)’32I, 422. ' 46 problem are those of the sixteenth century orthoepists, spelling reformers, grammarians, and the like. They were very conscious of the pronunciation [d] for the grapheme g. Their reactions to this pronunciation are couched in terms like "abused," "corrupt," "bar- barous." EXcept for such value judgments, however, these early linguists discuss the phenomenon little and make only cursory, if any, analysis of the situa- tion. But the fact of the change, the alternation, the substitution, or whatever, and its social aspects are obvious. One commentator, William Salesbury, writes of " . . . som barbarous lyspers . . . who deprave the true Latins pronunciation . . . [and] do read . . . 9333p legith in stede of gpig,lpgip."8 And William Lily, also on Latin pronunciation, de- plores " . . . the disfigurement created by our wandering when we pronounce p and g as if they were "9 aspirates. French‘p and'g were similarly mispronounced. 8 o o o 0 William Salesbury Ip.Pl e and Familiar lptroduction (London, 1567), pp. 53,467- 9williem Lily. A Shorts Introduction oi , II,Brevissima Institutio (London, 1567), sig. A4 . "Foede quoque erratur a nostris, ubi p,& g tanquam aspiratas pronuntiant." 47 In John Palsgrave's early writing on French.g, with the usual oblique Latin reference, he states: ,2 in all maner thynges confermeth 4 him to the general rules above rehersed, so that I se no particular thyng wherof to warne the lernar, save that they sound not Q of _a_d. in these wordes adultére, pg: 0 tiOn, adoulcér, like pp as we of our tonge do in wordes of latine ath athjuuandum for pg adjuvandum corruptly, for in all wordes where g hath his distinct sounde, he shalbe sounded 1yke as the latine tong soundethp.10 In works on English, we find statements like Simon Daines' comment on the " . . . promiscuous use of,p and pp, descended hereditarily to us from the Saxons. [The result: is that] many pronounce pp like g."11 In the late sixteenth century, the early phonetician John Hart wrote: "The Q we abuse in the sound of pp the figure of which element we ought to write even as we speak: and pronunce the ,g writen in his proper sound which is as in ladder, and not as we abuse yt yn derivations: as in adoption 10John Palsgrave Leclarcissement de pp langpe Francoyse, ed. F. Genin (Paris,‘1852), p. 28. 11Simon Daines, Orthoepia An licana (1640), eds. M. Roesler and R. Brotanek (HaIIe, I958). PP. 53‘54- 48 12 The matter was not when we pronunce yt athoption." just a problem of cultivated versus vulgar pronuncia- tion; it was an academic concern, for Hart adds that his insistence that Latin g be pronounced [d] was " . . . moch against the good wills of . . . [his] masters."13 Hart also notes that Latin Q be— fore another p should always be a stop.14 If this is an accurate observation, it might well explain or in- dicate an influencing factor in single /d/ becoming a fricative and double /dd/ remaining a stop. These remarks suggest that not a few speakers of Early Mbdern English pronounced postvocalic Q as [a] or something very like [a] in Latin loans--not in the 9g context, however. Whether single‘g in native words was treated in the same fashion, Hart does not explain. But again, a usual linguistic practice is to give foreign words English pronunciations, not the converse. Thus French or Latin g if pronounced [al'could only have come from some English phone-graph 12John Hart, The Qpen ing of the Unreasonable Writppg of our lisEIEEppg,e '-Bror Danielsson 0 Fa War 8 ’oc O m, 1955), Pt.1 . p. 13-4. 13Danielsson, Hart's Works, Pt. 1, p. 144.. ’ l4Danielsson, Hart's Works, Pt. 1, ad dico, p. 205; but advertized, p. 205. 49 correspondence. Another important early work, considered to be the first grammar of English as a foreign language, is that of J.B. Gen. Ca. (now known as James or Jacques Bellot), £3 Maistre d'Escole A_nglois (1580). In this text, Bellot attempts to describe the pro- nunciation of‘g and th for the Frenchman aspiring " . . . to attayne the true pronouncing of the Eng- lishe tongue."15 0f g, he says, it is " . . . pro- 16 that is, as [g]. nounced as in the French tongue;" Of‘th, {2h are pronounced in blowing with the tongue against the fore teeth before the sounding of them and taketh the voice of one 22153, both before and after,§,§,9."l7 There is little doubt that Bellot is here describing a voiced dental fricative [a]. In the seventeenth century, George Mason, considering the same problems, stated: Now when pronouncing these and similar 15James Bellot, Le Maistre d'Escole énglois, ed. Theo Spira (Halle, 19I27j‘57‘1. leSpira, Bellot‘s pg Maistre, p. 14. l7Spira, Bellot's £3 Maistre, p. 25. 50 words, the Frenchman should be careful lest instead of sayingll thanke 223, he says tanke 122 or danke 122, as the Flemish do. If I were speaking to learned people who know the Greek language, I would tell them to pro- nounce these words as if they were written with a theta; but when speaking to those who do not know this language, I would tell them that when they pronounce these words, they gently touch their front teeth with the tip of the tongue, imitating approximately in pronouncing th either serpents or goslings when they hiss. ' Somewhat fancifully, this statement indicates that /9/ is a voiceless dental or postdental fricative articulation, not, however, an interdental, nor [9] nor [1;]. Nevertheless, at least three times in his work, Mason phonetically renders 3h with the symbol g; namely $12.52. 9.9.1.2. 512 for Eggs, Egg; and _t_1_l_e_ 19 18George Mason, Grammaire m loise (1622), ed. Rudolf Brotanek (Halle,W .IfiZIB. "Or an prononqant ces mots & sembla'blesP que le Francois se donne garde que au lieu de, I thanke ou, il profere, tanks ou, ou danke you: comme ont les Flamans: male a essant ma parolle aux gens doctes qui entendent la langue Grecque, qu' ilz prononcent ces mots, comme s 'ilz estoyent esscrits avec vn, a, semblablement parlant aux ignorans d'icelle langue, quand ilz viendront a prononcer ces mots, qu 'ilz touchent doulcement les dents de devant du bout de la langue, imitans en partie en la prolation de th, ou les serpens, ou les oisons quand ilz sifflent." 19Brotanek, mason's Grammaire, p. 2. 51' meaning, I should think, that _d_ signals some kind of fricative articulation. After Mason, one finds no further really significant statements concerning the phonetic manner of g and th until Alexander Ellie's monumental work 92 133311 English Pronunciation, which proved to be a disappointing source for the matter under study here. Relating chiefly to Gower and Chaucer, Ellis states "20 that "D‘was (d) of course. His reference to 31;, however, suggests that the phone [d] which became [a] should have undergone the change by Chaucer's time, for he says: TH . . . had probably the same sounds as at present and distributed in the same man- ner. Occasionally we meet with g in places where we should have expected _th = (dh), as in fadur 100 = father, hider 674, thider, slider 1265, where the rhyme shews that the sound was really (d) and not (dh), but the (6.) seems to guarantee the pronunciation of 1:2 as (dh) when written in these words.21 20Alexander J. Ellis, 93 Earl; figlish Pro- nunciation (E.E.T.S.,'Extra Series , , , , 5, Eondon, 1569), p. 508, sec. 3. 21Ellis, p. 317, sec. 5. Numbers refer to lines in Thomas Wright's' edition of the Canterb Tales'(Har- 161m MS.). In F.N. Robinson's edition E§ew York, 1933), 1. 674 I 6720 52 Ignoring the cogency of the evidence, what Ellis is concerned with is not the value 0f.Qa which he stated earlier and unequivocally was [d], but that [d] > [a], not [a]. As I noted earlier, what Ellis has to say about EMOdEIQEh is interesting: "Some literary men write gth to indicate the sound ([8])."22 This appears to be a reasonable answer to the orthoe- pists' plea for the resolution of the ambiguouslth, a plea which is still heard today, as well as a reasonable explanation of the supposed transitional state noted in Chapter II. Following Ellis, a few twentiethrcentury scholars have something to say about the past phonetic manner d and graphemic—phonetic correspondence in Old and Middle English. Sporadic, isolated references in language texts suggest the possibility that ‘Q, a [a] in many instances, or that d represents an allo- phonic distinction. Karl Bfllbring, for example, writes that, " . . . voiced [d] is frequently represented bygg in the oldest texts; very rarely is voiceless [p1,u23 :aEllis , p. 593. 5Karl Bulbring, Altenglisches Elementarbuch. (Heidelberg, 1902) Pt. I: PhonOIo , p. 137, n. 2. "Das stimmhafte [d wird in den §§¥esten.Texten.hfiufig ‘Q durch wiedergeben; sehr seltan das stimmlose [P]." 53 By this it is conceivable that‘d is the traditional grapheme for [a] in Old English, and that it is retained as such throughout Middle English. The graph shows phonetic rather than phonemic value. Similar parallels can be noted in the indistinction between palatal [c] and velar [k] or [g] in the runes of the Ruthwell Cross where the graph of the 24 front phone stands for both front and velar phones. Similarly in Common Celtic 3 = [c] or [k].25 Looking at scribal variations in fifteenth century correspondence, Norman Davis found the fol- lowing vagaries of‘g before 33 in the letters of three educated men, all in the employ of the Pastons: James Gloys. The usual spelling is'd whether the word is written full, as fadere, moder, gadered or abbreviated, as mng. But with 'other' there is some uncer— tainty, 9Q; and the toder occurring early in contrast [graphic, but not phonetic?) with prevailing othre, othere. Richard Calle. The use of‘g is familiar, 24Brunner, Siever's Grammatik, p. 180, sec. 205, n. 1; cf. Campbell, QE Grammar, p. IV}, sec. 427, n. 1. 25Kenneth Jackson, Lan a e and Histo in Early Britain (Cambridge, 33., 53), p. , sec. 56. 54 as modre, gadre, thider, but th is more fre— quent: fathers, gather-—(at least four times), hether, together. [The others] . . . show only‘g in all such words. John wykes has whethere . . . , and hethere in a letter written for Margaret IPaston; but otherwise edere, hedere, theder.26 Such examples support the view of modernists over that of the nineteenth century tradition, as we note in these comments by Lehmann: . . . recall how explicit Saussure [l9mc.] was about the ineffective role of the speaker in initiating, and even controlling language. . . . He speaks of a blind force operating against the sign system. Kurylowicz [204c.] in contrast views the speaker as deciding between alternate forms and through such decisions controlling the selection of two or more competing forms.27 Saussure implies a general linguistic naiveté, Kury- lowicz, on the other hand, a general linguistic s0phia- tication (or at least a linguistic awareness) which 26Norman Davis, "Scribal Yariation in Late Fifteenth Century English," in Hel es g2 L' isti ue et de Philologie: Fernand Mossé ii figmoriam er s, $9597.tp- 99. 27Winfred P. Lehmann, "Saussure's Dichotomy between Descriptive and Historical Linguistics," in Lehmann and Malkiel, Directions, p. 15. 55 I readily grant the medieval man, lettered or unr lettered. Mbst of my original corpus and related words show alternate (competing?) -der/-ther forms in Middle English, principally in.native words rather than foreign (not to the complete exclusion of the latter, however). The list in Appendix III offers evidence that Q often alternates with‘gh in the same word, in the same dialect, and in the same century. For example, fifteenth century brother has d4~ th forms from the south through Scotland-~even within the same text.28 This suggests either alternative pronunciations or that'g and‘th are free variant graphs of some allophone of either /8/ or /d/. It is reasonable to assume that spellings arise out of pronunciations, especially in the be- ginning of an orthographic tradition. However, lettered men.are known for their conservatism, and it is a commonplace that there is an orthographic lag in the rendering of a sound change. Too, the graphs‘g and 9 had spent many centuries being cons 28 The "Ge t toriale" of the Destruction of Tro , eds._33orge I. Eatton and‘David’Donaldson,‘ETE.T.S. (London, 1869), p. 512, 1. 9589; p. 451, 1. 15167. 56 fused because of their nearly identical graphic shapes and/or possibly their nearly identical pho- netic renderings. In late Old English, the graph '3 stood for both the voiced and voiceless (inter—) dental fricatives. It was ostensibly replaced with th, butIQ has had a long history of carrying the 29 double function of stop and fricative, and it has often been editorially changed to 5,30 Thus the writing of‘d or‘g might be nothing more than what Paul Postal calls " . . . the general tendency of human cultural products to undergo 'nonfunctional' stylistic change,"31 like schoolgirl circles over‘i's. Therefore, I propose that textbooks of Middle English would do well to modify statements like Mbssé's, " . . . the greater part of Middle English consonants were pronounced pretty much as in ModE."32 They should 290har1es L. Wrenn, "The Value of s elling as Evidence," Werd and Symbol (London, 1967, p. 152. 50The most convenient, illustrative example is the Exeter Book (eds. Krapp and Debbie). "Christ," pp. 5—49, contains fifteen insttnces of d.~ a or d clearly resulting from a by scribal amendation. 31F aul Postal, As ects of Phonological Theory (New York, 1968), p. 28+. " 32mosses, Handbook @, p. 14. 57 note that Middle English single d between a stressed and a. non-stressed vowel which develops into Modern English Eh was probably pronounced [as] by many cul- tivated speakers. I do not deny that such statements exist. They do, but minimally, or out of the mainstream of the discussion. With reference to wright, Mbssé, and Bfilbring, MQH. Scargill states, "The voiced spirant (a) is frequently represented bygg in Old English.*55 Meanwhile, A. Campbell leans in a slightly different direction.(whichhe later modifies) when he states: file assumption that 13.1.1. at first re- presented a voiceless spirant is based on Irish usage, and the parallel with,gh. No OE manuscript preserves this usage: some use Eh only initially, and have g medially for both the voiced and voiceless sound . . . .54 Early manuscripts . . . use th.ini- tially and‘d.media11y and finally, for a dental spirant. The distinction seems to be one of position in the word, not voiced MMMWWWWN Proto-Germanic and west Germanic into Old Egg ”Matthew H. Scargill, m _Qg 3g; Develog- 1 3 (Toronto, 1951), p. 32. 34Campbell,9§ Grammar, p. 23, sec, 55, n. 3. 58 and unvoiced sounds, for 9 often represents a voiceless spirant . . . .35 Campbell continues to suggest that later _d_ =- [a] or [63,36 but never that 1:3 ~ g s [d]. He concludes his discussion of Old English orthography and pronunciation with this note: "The use of the _d symbol for 2 [the most frequent graph for both the voiced and the voiceless fricative] in runes . . . is due to the influence of the Latin alphabet."37 OED editors reflect the middle-of—the-road position regarding the value of g. In discussing father, they note: The spelling in our quotations is uniformly with _d_ until the 16th century. Except that faker occurs sporadically in the Cotton and thtingen MSS. of the Cursor Mundi (a. 1300); but the pronunciation (a) may have been widely current in the 15th century or even earlier; in the 14th and 15th centuries the spelling with 39.33 is very common in words 3scannplmll, 0} Grammar, p. 24, sec. 57 (5). 36Campbell, p. 26, sec. 63; p. 29, sec. 69. ”Campbell, p. 29, sec. 70, n. l. 59 like brother, feather, leather, though this spelling cannot in all cases be supposed to indicate that the writers pronounced the words with (d).. The Mbdern English -ther (a) of OE.‘:g§§, :ggg . . . is really the result of a phonetic law common to the great majority of English dialects.38 The "law" to which the‘ggy editors refer is the fifteenth century change under discussion here. They obviously regard the change as regular. This reflects one weakness of past work in historical pho- nologyb-the passing off of substantive evidence in favor of supposed phonetic laws. One can only suggest further possibilities about phonology when it is based solely on the written record. we can sometimes be rather certain about matters, but we can never be dogmatic when talking about past pronunciations which are only partially recoverable. I am rather certain that the orthoepic and philological analyses Just presented can lead to only one conclusion-~thatlE/l7der/ > EModE /Vaar/ is a leveling of allophones. In other words, what has long been considered to be a phonemic shift, is really a phonemic split, brought about by Nerth T 38039, s.v. father. 60 Germanic influences. One often reads of Scandinavian lexical contributions but seldom of any phonological influence except in place-name studies. Eilert Ekwall remarks, " . . . a short g never occurred in Scandinavian between vowels, g (dh) generally corresponding to Old English g. In such positions g was replaced by Old Scandinavian'g. Hence MITHOP from original Midho , LOUTH from original giggg. «39 This is a very common phenomenon. Such a process might carry over as well into the general vocabulary and result in competing forms first inside the Danelaw, then through dialect mix» ’ture, throughout the country. Finally, through convention and the crystalizing influence of printing and learning, the resultant form as we now know it prevails. A few more illustrative place-names, their developments and earliest dates as garnered from Ekwall are: (EM) Atherstone (1221) < Aderston (710); (NM) Batherton (1260) < Baderton (1086); (K) Bethersden (140) < Bedersden (1100); (S) Cheddar (1100) < Chedar (880); (N) Adderston (n.d.) < Edreston (1233). These and similar examples indicate that the medial graphIg 39Ekwall, English.Place-Names, p. xxv. 61 is regularly changed to th in place-names in or by the mid-thirteenth century and very seldon.later than that. It is probably safe to assume as Henry Nyld did almost a half century ago: In the spelling of Middle English many dialectal varieties of pronunciation, and doubtless also of individual peculiarities, are expressed; but in a highly-cultivated literary language the spelling is usually crystallized and expresses merely a general average of the extant pronunciations, the same symbol being used by "correct" writers without regard to differences. Thus . . . symbols . . . which for practical purposes of philological statement and investigation, we consider as representing severally the same sound . . . with perfect consistency, may in reality have been conventionally used, in the same words, by writers whose pronunciation differed more or less consi- derably.4O I, of course, wholeheartedly subscribe to this state- ment which supports my view of variant phone-graph correspondences. Unfortunately, wyld withdraws to safer, traditional ground as he continues: However, until a spelling has become 40Henry C. wyld, The Historical Stugy'gf the Mother Tongue (London, 1920), p. . 62 absolutely fixed, like that of Classical Greek and Latin or Modern English, it is safe to assume that the use of a symbol is fairly consistent, and that it expresses at the worst, a group of closely-related varieties of sound. Earlier in the same work, wyld approached modern theory when he wrote of sound change in general: What really happened is that the underlying memory pictures of sound and movements undergo gradual modification, and are different in one age from what they were in a former . . . ; meaning thereby a change in the aggregate of mental pictures possessed by all the individuals of a com- munity, the result of which is that a series of substitutions takes place of one sound for another, until the sounds actually pro- nounced by a later generation in the same word differ widely from those pronounced by an earlier generation. Compare this with a modernist statement: Transition or transfer of features from one speaker to another appears to take place through the medium of bidialectal speakers, or more generally, speakers with 41wyld, Historical Stugz, p. 116. #Zwyld, Historical Study, p. 69. 63 heterogeneous systems characterized by or- derly differentiation. Change takes place (1) as a speaker learns an alternate form, (2) during the time that the two forms are in contact within his competence, and (3) when one of the forms becomes obsolete. The transfer seems to take place between peer groups of slightly differing age levels; all the empirical evidence gathered to date indicates that children do not preserve the dialect characteristics of their parents, but rather those of the peer group which dominates their preadolescent years. Then more exactly to our point: At some point the social and linguistic issues are resolved together; when the op- position is no longer maintained, the receding variant disappears. This view fits the general observation that chggge ig‘gggg regglar ig‘ghg outcome thgg‘in process. Thus the Old English graph‘g might well have had two pronunciations, freely alternated, but pro- bably socially conditioned. OE'Q was either native Germanic [d] or its Romance counterpart [d], which 43U’riel Weinreich, William.Labov, Mhrin Herzog, "A Theory of Language Change," in Lehmann and Malkiel, Directions, p. 18#. ‘ l|weinreich, et a1.,"Theory of Change," p. 149, Italics are mine. 64 was introduced into Celtic Britain by Roman conquerors and Roman clergy during the periods of Roman occu- pation and sustained into Old English through the immense cultural impact of the Roman Christian mis- sionaries. The articulatory differences between [d] and [d] are minimal and could distinguish social dialects--with [d] predominating in the influential circles of governance and religion. Staying within this language contact frame of reference, any late Old English [d]‘~ [a] could then be called the re- sult of North Germanic influence from the Scandina- vian invasions and occupations; these would distin~ guish regional dialects. In Middle English proper, the Romance pronunciation, the first of these two older traditions, would have been reenforced in the same measure by Norman conquerors who exerted the same kind of linguistic pressures in the same kind of social circles--government, religion, and education. The increased mobility of the people might then have led to a confluence of these factors [d]l~ [QJav [a], resulting in a leveling by spirantization of the dental stops--another minimal phonetic step and a nonfunctional change. After considering the foregoing observations 65 and theories, I am convinced that we must posit at least one new allophone of Middle English./d/, pre- dicated on the supposition that the traditional pronunciation of Middle English is too often equi- valent to a reading in terms of Modern English spel- ling conventions. It would appear that, in recon- structing the manner or,g in Middle English, we have to consider the possibility of what is now a noannglish sound or even an English sound signaled by what is not now the usual symbol. Rather than simply the voiced alveolar stop [d] in ME /Ndar/, three homorganic possibilities suggest themselves: (1) an aspirated medial [dh], or, (2) Jespersen's earlier suggestion, dental [g], or (3) a hybrid of the two, [db], an aspirated dental stop. Of the three, I prefer the last over the other two because it is the more convincing relative of the interdental fricative we now have. In addition, there is a his- torical parallel in the tendency to aspiration of original /t/. John Hart's "breathed't" is at work in these pairs: author < auctor, lethargy < litargie. It appears that /t/ : /e/ :: /d/ : /a/. The change /t/ > /9/ is, however, much broader based since it involves loan words and is not restricted to inter- 66 vocalic /t/. There is, nevertheless, a similar principle at work in both pairs. If Latin and French had any phonological influence on English, and I see no reason to assume that they did not, since one was the language of learning and the other the language of culture, it is very likely that a common presti- gious pronunciation of‘d (at least intervocalically) was the pure dental [d]. The original sound is des- cribed.by Terentianus Marus in this manner: "The tip of the tongue, moderately curved at its highest point, strikes low at the front of the teeth; then is the,g sound completely articulated."45 Quite likely, the“crude and barbarous'speaker of English aspirated this stop to [dh]-as is yet his wont. If we accept the [dh] allophone (and like- wise the [th]) in the /€'cvr/ sequence, and these allophones are in turn "breathed" (spirantized) in Early Modern English, there is little else that they can become except [a] and [9]. Too, we must remember that an orthography often remains undisturbed by “Blllen, Vox Latina, p. 95. "At portio dentes quotiens suprema Iinguae? pulsaverit imos modique curva summos/ tunc‘g sonitum perficit explicatque vocem." 67 changes in pronunciation, and Middle English orthog- raphy like Mbdern English is not necessarily phonetic but largely symbolic and somewhat ideographic. Thus, I think that the evidences of this chapter suggest the real possibility of no phonemic shift at all, but the resolution by Modern English spelling con- vention of a graph-phone ambiguity. Another alter- native might be simply to accept [a] as the inter- vocalic allophone of ME /d/ before /vr/, which might be the simplest and truest answer in the long run. V SOME NONPHONETIC FACTORS Recent statements about phonological change derive in the main from the concepts behind the theory of transformational-generative grammar. The phono- logical aspect of this theory is embodied principally in a recent work by Noam Chomsky and Morris Halle. The theory utilizes an approach that embodies a COD? cept of the triune impact of the semantic, syntactic, and phonetic components of language. The Chomskyb Halle concept of language refers to both spoken and written interpersonal communication; their concept of grammar refers to " . . . a compendium of specific and accidental (that is, nonessential) properties"1 of a lpecific language. This compendium of properties is a system of rules specifying sound-meaning correspondences. In order to communicate on an inter-personal level , it is necessary for the speaker-listener to have a fundamental knowledge of the grammar of the language in use. That knowledge might well be intuitive and not 1Noam Chomsky and Morris Halle, The Sound Pattern.g§ English (New York, 1968), p. 55. 68 69 at all rationalized by the user, but it must exist. The idea of the transformational cycle stems from the notable consistency of speech and writing patterns among users of the English language. The transformational cycle asserts that the overt struc- ture of a complex expression is intuitively ordered through a fixed set of operations that considers the significance of each of its components. For Chomsky and Halle, the significant components for considera- tion should be classified in three general categories: semantic, syntactic, and phonetic. In their study, they deny the relevance of the phonemic level regar- ding the final language structure operating through the transformational cycle. They rely solely on the phonetic level of language description. The syntactic property of grammar refers to the symbol or form structure; the semantic to the symbol meaning; the phonetic to the symbol sound. Chomsky and Halle believe that certain demands are concurrently placed upon the formless and ununified elements of language (such as the letter or sound symbols) by these three components of grammar, so that a resultant, meaningful, overt expression is 7O generated according to certain intuitively prescribed procedures. The outlining of these regulating pre- cepts maps the transformational cycle. It is the principle of the transformational cycle that underlies a semantic interpretation of writing or utterance, the syntactic interpretation of writing, and the pho- netic interpretation of utterance. Chomsky and Halle state the correspondence in this manner: (l)a branching tree diagram of the deep structure, (2) a linear terminal string illustrating symbolically and actually the morphemic sequence, and finally, (3) a phonetic rendering of that morphemic string; for example: 71 +doa+ +opmoquaaoo+ +ow+ ':> N .oum .ma .oaasm ens ssmsosom .somamsosrmasos + seememoass + smaanmsnosmr Ame mnmfimmzmzm+dow+>h+opwoflndaaoo+>uzudm+ow+ zfizmamm+samsm+zmemH+oaop+zHaumzH>n+pnear>h+nmaaneemo+u>um>umzhzm+oz+zumzumg Amv rnmwnw+ +oH m swam W _ z _ ‘41—: 9+ +pmmm¥ +nmwanwpmo+ +034 L>_., a a 72 It can be seen from the above that the thrust of this generative phonology effort is directed to- ward an analysis of the surface structure of English grammar. Therefore, the syntactic component will have special significance, since it seems to cons- titute the genesis of the overt ordered system of written communication beyond the abstract conceptions of the mind. Too, the phonetic interpretation is important only in the surface structure; it is only loosely related to the deep structure. The semantic component refers specifically to the deep structure and assists in generating and interpreting the sur- face structure. Its relationship to the surface structure is not, however, much discussed by Chomsky and Halle. In the simplicity of everyday speech, it appears that deep structures are the mind's thoughts to be conveyed in written or spoken sentences. People often have similar thoughts or mental pictures of an idea, whether they speak the same language or not. Their separate surface structures according to their separate languages are simply their distinct means (as culturally defined) of expressing these me aningful thought 5 . 73 The surface structure of English, according to Chomsky and Halle, consists of a string of formatives to which certain lexical features are attached. For example, father is an element that begins with a voiceless fricative and ends with a retroflex glide; it is a "noun," "animate," "masculine,"... --a classification not unlike the taxonomy of the classical empiricists Locke, Berkley, and Hume, who considered such concepts as "substance," "primary goals," "secondary goals," (i.e., a thing is as you perceive it). These features of language receive specific names if their importance is realized in a specific language. Chomsky and Halle suggest that most of these feature categories are univerally utilized but culturally defined. In addition, they offer a category of physiological features: certain articulators, points of articulation, and manners of articulation that are perhaps meaning-conveying universals. This is not to say that every language uses all these different categories or all these different features to convey meaning. However, many categories are widely used in diverse languages. Chomsky and Halle go on to posit universal categories referring to the phraseology or bracketing 74 substrings of "feature" elements. Some items within this categorization are sentence, noun phrase, verb phrase, adjectival phrase... . Universal lexical categories include such items as noun, verb, adjective... . Taken together, the universal phrase categories, uni- versal lexical categories, and universal phonetic categories as well as universal lexical features form an infinite set of plausible surface structures. These features can then be defined along syntactic or semantic or phonological lines. Unfortunately, the utility of the Chomskyh Halle generative phonology theory in its fullest form has little apparent power to explain the change 0E /VdVr/>> EModE /\fdar/. Specifically, Chomsky and Halle call for a rule addition in the phonological component to account for sound change. The sound changes of diachronic English phonology, they say, canbe exemplified in terms of the rules of two large classes: readjustment rules and phonological rules. Readjustment rules express the properties of the lexical formative in a restricted syntactic context. Readjustment rules specify the change with formulas of the familiar type x —-> Y / z (x becomes Y in the environment Z). 75 Phonological rules specifically identify a sound change as a replacive, a softening, a laxing, a spiranti- zation, a cluster simplification... . The assum- ption is, however, that changes in the phonetic ac- tualization of a sound are very slight; and further, that there is considerable lack of evidence for his- torical sound change. They quote from Hoenigswald: "It [sound change] has always been a speculative pic- ture whose best feature is a surface plausibility which it once possessed, but it does not possess anymore."3 I Unlike the traditional approach to sound change-~the observed correspondence between two stages of a language, purely at the lexical level (e.g., gygg > gtggg)--Chomsky—Halle call for a gram- mar addition rule that may function for many genera- tions without causing changes in the lexical repre- sentation. This addition rule may be formalized in abstract, formulaic notation, but it will not be observable in the written words themselves. Chomsky-Halle further suggest that if within a syn- 5Chomsky'and Halle, p. 250 citing H.M. Hoenigs— wald, Graduality, Sporadicity, and the Minor Sound Change Processes," Phonetica, p. 207. 76' chronic grammar, a rule cannot be traced to any sound change, it is the result of the dominance of one grammar over an alternative grammar. They state that a child is exposed to at least three grammars: G1 (his parents' grammar), 02 (his own grammar), G3 (his peers' grammar). G2 differs from 61 by grammar rule addition in order to approximate G3, the desired grammatical goal and therefore evalua- tively higher in the child's view of a grammatical hierarchy. In the process of striving for G3, the child loses G2 and acquires a new,§3, which is eva- luatively higher than G1, G2, or G3. It is his desired goal. ,9} is constructed on the basis of the actualization of GI and G3, from which the child has chosen those features most highly valued. In other words, sometimes the child will choose the grammar incorporating a historically attested change, and sometimes he will not, depending upon his eva- 1uative measure of the choice given. It is the re- sult of these unpredictable synchronic choices that establishes the diachronic pattern. The ChomskybHalle approach to the consonant system of English begins with this remark: "Although 77 it is not without its problems, the consonant system seems less interesting than the vowel system and we will not treat it in anything like the same detail."4 They hold to their word. Their discussion leads to sample readjustment rules (sound change formulas) of words derivationally inflected, then to specific change types: velar softening rules, stress rules, spirantization rules, palatalization rules, and the like, all utilizing Romance words, none of which shed light on the change under study here. At best what they do with historicalphonology is to provide a formulaic notation system based on the text's dis- tinctive feature notation, and they apply that no- tation to the linear rules already provided by the past. For example, "slightly modified forms"5 of Grimm's and Verner's Laws are formulaically noted: -VOC [+0 ont] / E2338] :::: (a) +con -s> __ +voc ~nasal [+voice]/ [wont] [-cons] (b) This is an abstract generalization of the statements that: 4Chomsky and Halle, p. 223. 5Chomsky and Halle, pp. 540-541. Actually, the modification is not so slight, since it eliminates initial and final stops and bases verner's Law on a following vowel, not a following accent--liberties the authors neither justify nor explain. 78 (l) the stops /p,t,k/ are actualized as their voice- less fricative congeners [q,9,x] if preceded by but not followed by a vowel (a); (2) the stops /P,t,k,/are actualized as their voiced fricative congeners [b,a,g] if preceded and followed by a vowel ((a) plus the top half of (b)); (3) the sibilant /s/ is actualized as its voiced congener [z] if followed by a vowel (bottom half of (b)). Although the principal thrust of the Chomsky- Halle theory is concerned with the synchronic des- cription of sound change in nonsnative words, although their discussion of historical change is primarily a notational innovation rather than a new explanation, it does suggest that perhaps we should look beyond the traditional approach of autonomous phonology. we can borrow from the larger concept behind the theory as expressed by Paul Postal: " . . . some regular phonetic changes take place in environments whose specification requires reference to nonphonetic morphcphonemic and/or superficial grammatical structure."6 §Postal, Aspects Phonological Theor , p. 240. 79 Some extra-phonetic processes that immediately suggest themselves are homonymic conflict, functional shift, and possibly stress readjustment. Group I words, those presently spelled with.gg, have or have had homonyms; they have propensities for word-level transformation different from Group II words; their surface structures, in other words, have subtle dif- ferences which might account for Group II entering and Group I not entering into the sound change /d/ > /a/. It is such possibilities that I should like to consider here. In the Early Modern English period, /éder/, /blader/, /1adsr/, /fider/, /i's’der/ formed minimal pairs with /aasr/ 'either,' /blaaer/, /laaor/, /fiaar/, /f589r/ 'load,' respectively; thus contrasting /d/ with /a/. No such clearly cut minimal pairs contrast /d/ with /8/ in Early Modern English words that have undergone the sound change. Therefore, any change from./d/ to /a/ in the former group of words would have resulted in a conflict of homonyms, which, according to the theory of Jules Gilliéron and Mario Roques,7 would have resulted in the ultimate suppression 7A5 synthesized by Robert Menner in "The Con- flict of Homonyms in English," Language, XII (1936), 229-244. 80 of one or the other formative. In general, the Gilliéron-Roques homophonic conflict theory calls for three steps: (1) homophonic development, (2) homophonic conflict, (3) homophonic suppression.8 The fact that variant spellings (and pre- sumably variant pronunciations) of gdggg, bladder, etc. with Eh (gghgg, blather ...), did develop in Middle English indicates a homophonic development. However, the second stage, that of conflict, would probably be infrequent in the words of my corpus because the homonyms would seldom fall into the same sphere of ideas or appear in similar contexts. The result, in these words at least, is no overt suppression of forms, and, therefore, some dialectal substantive homonyms having the same phonic and graphic shape. Some of these conflicting forms survive today. For example, in northern British dialects, fodder and fother both contrast and alternate their medial consonants and might on occasion create an ambiguity. A degree of suppression can be seen in typical dic- tionary entries like: fodder (dial. var. fother) 'food: 8For example, OE cwén 'queen, princess' > ME /kwe'n/ > EModE /kwi'n/; 0E cwene 'woman, harlot' > ME / kwc'n/ > EMCdE /kwi’n/ (the latter suppressed). 81 but fother or fodder 'load' is labeled an infrequent dialectal form whether it be identified as a morphemic 9 variant or a separate morpheme. What this suggests is that Group I words had an occasionally realized propensity for under- going the sound change /NdVr/ > /Vaer/, but in the process, the change created a phonetic as well as a graphic shape already signifying another morpheme. Since an alternative was available (that is, no change at all), that which was the least ambiguous form sur— vived into Modern English. Many dialect forms that were Middle and Early Modern English lexical homonyms showed or began showing contrasts in the radical vowel, thus resolving their ambiguities in another way. For exr ample, the radical vowel of fother (fodder) 'load' had length as opposed to fodder Sfother) 'food' which did not.10 This further suggests that the resistance to change might be corollary with something we could call homophonic aversion. This could be one explana- tion for Eggyg 'rudder' not becoming rother 'ox' or Egggg 'progeny' not becoming tuther 'one who toothes (saws)'. 9OED or Webster's Third New International Dic- tionary (Springfield, 1967?, s.v. fodder, Totfier. 10MED, s.v. fodder, fother. 82 From Group I, Egggg, bladder, madder, and Egggg are principally substantives in English; they seldom generate another form class except as adjuncts. On the other hand, the formatives of Group II, mother, father, ather, hither, weather, whither, to ether, thither are substantives, verbs, and adverbs. The three nouns are frequently subject to derivational processes and often appear as verbs ("to mother the baby," "to father a child," "to weather the storm"). Therefore, perhaps a nonphonetic, morphological fac- tor governs the sound change; namely, that MEjpost- vocalic /d/ becomes /8/ before /ar/ in verbs and adverbs, derived or regular, the verb form thus give ing rise by leveling to the resultant Early Modern English substantive. As tenuous as this might at first appear, there are related data that seem to bear on this. The first occurrence of weather with th is in l400-—coincident with its first occurrence as a verb. At the same time, fggig appears as a verb; the 01332 entry lists only two occurrences of it as such with a‘d, thereinafter it is always with‘th. Mbtherdoes not appear as a verb until 1548, but as a verb, it always appears in the 31; form. The ano- malous fodder does not behave like the others in 83 Group I, but its first occurrence as a verb is in 1300 with a‘g in medial position. From the E92 one finds that all the verb forms quoted for fodder are with th. The verb 39 bladder is an infrequent technical term and etymologically difficult to sift from the verb blather.{ One further interesting form is the ME smo(r)theren, an alternation of the OE verb smorian; which process of change suggests that the epenthetic 3h or the suffixal :ghgg, whichever might be the case, could have been a verb morpheme (or perhaps a phonestheme)11 in Middle English. Although presently all the words under consideration here are monomorphemic, :Ehgg does have a history as a suffix expressing comparison and/or alternation12 as in further, hither, thither, either. There is, then, the possibility that the speaker of Early Mod- ern English saw :Ehgg or heard [-aar] as a verb/adverb productive subclass, or conversely and more simply -der [dar] as an exclusive noun formative. Such considerations give rise to the further ll 12 As is gp’in s it, gpew, etc. Klein, Efiymological Dictionary, s.v. -ther. 84 possibility of the blending of morphological and suprasegmental phonological conditioning. When a word shifts its form class, it often shifts stress, a common enough feature of Romance borrowings in present-day English: perfect (adj.) >'perféct (v.), for example. Similar changes in native words are not known. One reason for this is that Old English and Middle English stress has only recently'begun to be studied within an adequate phonemic frame. The reconstruction of the suprasegmentals is only tentative at this time. However, possibilities can be offered. Using the terminology of Hans Kurath,13 father, a fully stressed noun under sentence accent, could possibly have become a half-stressed verb in the transitive verb phrase "to father a child." Thus when the Middle English noun féggg became the verb £2 fgggg, the weakened stress of the radical vowel in conjunction with the following unstressed /ar/ might well have given rise to a phonological downgrading of the voiced stop /d/ to the voiced fricative /a/. (Recall Sweet's "relaxation of articu- lative energy.") That is, an original /d/ in nouns Y- ‘7' 13Kiurath, Phonology and Prosody, p. 141. 85 under the influence of two weakly stressed vowels weakens from a forceful stop articulation to its fricative congener. This_fails, however, to account for the change in adverbs, which under both sentence accent and phrase stress usually receive primary stress. So too do the noun and verb father under sentence stress in Modern English. We can suggest, but we must question distributional statements about stress until more data are available. we must re- member too, however, that the use and function of stress in Old and Middle English was probably more significant and functional than it now is in Modern mg]. 1511 0 One final and rather simple nonphonetic so- lution might be based on the features "human" and ”nonhuman". All the formatives of Group I are non- human nouns. Group II presents the exceptional form weather plus human nouns and adverbs. It looks as if /-dvr/ became /-aar/ except in the case of non- human nouns. This gains strength when one notes the large number of Q; forms for weather. In other words, weather tends to the nonhuman pattern, especially in the Scottish and Northern dialects. An exception to the change? Yes. But as Paul Postal unequivocally 86 states in his anti-neogrammarian claim, "Exceptions exist.”].'4 Of course, all the neogrammarians ever claimed, and all I claim,is that exceptions exist by virtue of some actual if not always ascertainable fact or principle. And, in fact, I have assumed this all along in my attempt to discover why some words develop differently from others. If something is an article of faith, rather than a demonstrable fact, we must still try to describe it, or give up trying to explain anything. We must assume order, not chaos, if we are to discover the principles of order, obviously. What these nonphonetic factors suggest is that each word has its history; and with many forces at work on a word, some give way to one force while others give way to another, and few do so with parallel consistency. Sound changes, in other words, are linguistic tendencies rather than linguistic laws. As I stated in the introduction, this chap- ter was intended to do no more than note what might be possible explanations outside the traditional mode of autonomous phonology for the sound change 14Postal, Aspects Phonological Theory, pp. 276-277, n.5. 87 /Vdvr/ > /Vaar/. The generative phonologists strike at some basic tenets of the traditional theory of sound change when they deny analogyl5 and poohspooh purely phonetic environments as explanations for all sets of change, " . . . the claim that all rules 16 But I fail to are purely phonetic is empty." see how nonphonetic considerations, in an attempt to come up with the simplest answer, can do better than generate some interesting near—facts. Rather than providing the answer to sound change, I see nonphonetic factors embellishing and adding to the phonetic features that describe a sound change. Per— haps one day we will find that they work in harmony rather than in conflict. I think they are doing just this to some degree in this study. The weakness of both the traditional and modern theories lies not so much in their separate but equal aims, their sense of rigor and regularity, but in our lack of sufficient evidence and especially our inability to recover that which is apparently r v— 1SPostal,'p. 234, n. 3, and.passim. Too, Chomsky and Halle, p. 356, n. 11. 16Postal, p. 258, n. 11. 88 unrecoverable (in any wholly complete sense). Hcpe- fully the search for the truth will go on and new possibilities might somehow, someday, help uncover that one true answer (if it exists) to a not-so—simple problem. VI SUMMARY AND CONCLUSION In summary and conclusion, the supposed his- torical sound change, Middle English.postvocalic /d/ becomes Early Modern English /a/ before /vr/, has never been adequately explained. Some serious and important insights into the question were shown by the neogrammarians, but their answers were partial ones and sometimes contradictory. When one turns to the writers of modern handbooks, the usual student tool, he finds that they often have little time for deep analysis of this crux of English historical phonology; they are content with superficial restate- ments. The assumed change has been based principally on the fact that some Old and Middle English.words which were spelled with a medial g are now spelled with pp in the same position. Therefore, /d/ has become /a/ in certain instances-an apparent phonemic shift. If one assumes a genuine phonetic change, the traditional theory of autonomous phonology (that all phonetic change can be explained in.purely'pho- netic terms) offers some answers to explain the process. Such an explanation is not, however, 89 9O conclusive because we are uncertain of the real phonic quality of much of the corpus: that is, what the vowel equivalences were at various stages of the lan- guage and which form truly represents the dominant phonic feature of a particular dialect during a par- ticular period. Based on idealized states, auto- nomous phonology does suggest that a back versus front vowel conditioning factor appears to determine /Vdvr/ > /Naer/, with a front vowel determining change. In an attempt to expand the possibilities already formulated, I have suggested that an analysis of the sound change can be couched in terms of vowel length; that the change occurs only'in a word whose radical vowel had length. This quality must be extrapolated from the evidence of the Medern English word, which verges.upon imposing a system rather than revealing one. One would hope that forthcoming studies such as the Halle-Keyser theory of stressl(provided they are on the right track) would cast some light on this problem, but like pitch, there is a real question as to whether the metrics of discourse 1 Tentatively The Emolution 93 Stress 32 English. 91 are really recoverable. For the present and on the basis of ortho- graphic, orthoepic, and language contact evidence, this study, rather than explaining a sound change, supports a graphemic-phonetic reanalysis of Old English and Middle English g: that g is an ambiguous grapheme in the environment 2::Xp; in this position its manner of articulation often was closer to that of [5] than of [d]. I further conclude that this value generated from developments that had their origin in the Old and Middle English periods with the ins troduction of (l) the non-English [d] as an allophone of /d/ in Romance words, with a subsequent falling together with west Germanic /d/ [d] in native words; and (2) the North Germanic medial [a] in Scandinavian borrowings and cognates where [d] in the same position normally occurred in native words. The result was a falling together of variant pronunciations yielding a medial [db] or [a] in certain commonpcore native vocabulary. I recognize that the sounds and forma- tives of English are not the developments from a single dialect and that they develop through mixed and confusing states, ultimately leveling in a stanr dard pattern of language features which can'be exp 92 pressed in the terms of some idealized language state. Language features are the result of a cone fluence of factors. Every contribution to the comp plete picture enhances in some way our perceptions into the way of language, both its vagaries and its regularities. Therefore, I turned to a modern theory of language change, generative phonology, in the hope that it might answer the unanswered. Like traditional theory, generative phonology shed light, but it offered no final answer. It suggested to us to look to none phonetic factors, to posit change in morphological rather than in phonetic terms. Generative phonology presents some interesting possibilities, but it still lacks a rigorous methodology, fully worked out. The theory has little to offer diachronic historical linguistics. It does not measurably improve upon a structural study. It does, however, tell the struc- turalist to broaden his view, to consider both pho- netic and nonphonetic factors, and it suggests that the factors might work together in effecting a sound change. I am convinced that the best insights into language are those that are not constrained by a 93 single theory, new or old, but are enlightened by all theories and observations; evidence of any kind should be incorporated into an analysis. Apparently explanations of a sound change which are confined to one or the other theory (e.g., autonomous or generative phonology in this study), no matter how well con- structed, will fail to account for all the regularities and irregularities that can be observed in empirical studies of a language. Nevertheless, even incomplete or partial studies have value; as Chomsky and Halle note: "Any investigation of grammar is . . . a con? tribution to the study of performance, but it does not exhaust this study."2 Any formulation abounding with unknowns, as we have here, has many difficulties. Nevertheless, I hope that in some way my contribution has helped to broaden our understanding of one problem of a complex, language state transition. 2Chomsky and Halle, p. 110. APPENDICES APPENDIX I: Old English (or as noted) n5d(d)re alor ML apothecarius 0F auctor bl§(d)dre bladdre blaara < blaar ?Scand. MF bordure ? IR. bodhar ? Scand. MF broder br68er AN calender Middle English adder naddre alder apotecarie autour bladre bledder blather blether blunderen bordure border XXX bulder(stan) broideren brother kalender THE CORPUS 1 Modern English adder alder apéthecary author bladder bléther bléther blunder border béther boulder br6ider br6ther célendar 1The forms noted here derive in the main from Ernest Klein,‘§‘gppprehensive of the En lish Lan aid from CEarIes . Et olo ical Diction e (Amsterdam.an§ New York, 1966) one, The Oxford Dictionary 2; Epglish Etymology (Oxford, 1966) OF cedre cildru OF sidre sinder cliaan '51. CI'Wth Pfi‘considérer MF derision MF deriver XXX XXX XXX §ghwmaer mgaer eldra ellarn fader feaer ? Scand. 95 cedre chyldern children childeren cidre sidre sither cinder ? XXX crouder crowdere crowther consideren consitheren derision deriven ? didderen dodden doder doderen aither either elder eller eldre fader fether flenderis cédar children older cinder clider clither créwder crowther consider derision derive dither d6ddered dodder dother éither élder élder father feather flinders f6dor <'f5da foaer furara < furaor ganra < gandra gaderian gedrian MF gendre XXX ? *haaer < hep hider hl§d(d)er hladder léabor AF lavendre leaer MF litsrgie ML liturgia madere L moderatus modor m6raor myradrian foder fother fudder fother further gandre gaderen gendre hadder hather hider hither laddre lather lavendre lether liturgie mader moderate moder modre moder morthre murtheren mortheren f6dder féther further gander gather génder héather hither ladder lather lavender léather létharny madder méderate méther murder (n.) murder (v.) nauber naPer nawaer neolera < neosor niaera < niaer nahwider nahwiaer AF odour MF ordré HF ordure 0881' MP ponderer XXX ? L pandarus XXX MF poudre hraaor MF rendre réaer sculdor sculder XXX 0F escandele esclandrer esclandrir naither neither neyther nother nouther nethere nicer nithere nowider nowther odor ordre order ordure other ponderen pander Y H XXX poudre rather rendren rother shuldre shulder shodderen shoderen shuderen sclaundre slaundre 97 néither néther I nowither ddor order ordure 6ther pénder pander péther pudder péwder rather rénder rudder shéulder shudder slander slidrian XXX smorian Y splara MF soudure < souder sundrian MF surrendre OF tendre ? 0N tjdfir ON beir(r)a bar bér Per bader hider byder XXX 1311110]? tynder tindre teogathian 98 slideren slitheren smolderen < smoltheren smotheren < smortheren spyder < spithre soudour souldour sundren surrenderen tender tedir tethir their thair thar ther thider thither diderward thitherward thunder < thuner tinder tither < thithen+er slither smolder sméther spider sélder Sunder surrender ténder téther théir thére thither XXX thunder tinder tither t5gadere tdgadere tfigmdre uder under ‘wandrian AN‘wardere wardour weder we Her hwraer hweaer hwider ? weder wiser wundor wundrian togedere togidere udder under wandren wanderen warder weder wether whether hwider widderen wideren~wederen c> and H>uw> a glass n i I. new a a new , on 05H 00H om.” baa oma ONH mo 104 \uoewu\ mo vv \us.eGVmu\ masses v \uoemu\ meozm _"fodder" usees {Msees ussss usnss usees usaees seees il.m..se..ers i usees usnss . £2.38. 1 hummus as suees us me. “es al '3 suees a n fads sees I; 1'1 ll Flzuazuznn m «5%! AH; .z om one 09H oma o¢a om." UNA mo 105 "mother" \nucwa\ no vv \Hodswaua ”Egan v \Hmpua\ Mwozm users ussss users a users usfis ussps usss s . uses usfi es uses uses usnps ussps s .. a u . .Hbvb .H>.Ub. . H>U> > , . a e . u usues uses ses .55 "~de >n>v> sues - - ue , abuses uses uses ‘ {Ls .s . . . sues . sues 935 Hwy? all. . . . Has u__ sles wlses and ragga. u a Er r as , as v on." 00H RH 01; om.” ONH 106 " father" \ueemu mo vv \uoe.eu\ masses v \uoe.mu\ meosm ussss usfis users use» us . usees usees n ummmw» uses usAe ces a s . es mssus uses us. es_ __ _ ._ _u.s..Iless... uses Asvuses suses T sues uses uses uses uses uses Asflusas sues sues sues uses Asquses uses u uses uses sues As Vuses uses uses Lt uses uses sues nlur» 1 [1.11 I'll 4% em 23 s cm one omd one or.” 03 ONH m0 107 " gather" \ssuueem .. usuuseem\ so vv \ume.em\ assess v \us.em\ exeosu . s. s .. as users , usees usnss ussss uses usfies usnees uses 03 uses uses . . 552$ su fie ces one . . . a H>d> a 9gp 0m H sues sues - . ..> . [IdHuH £75 93S a sues mwmw» sues uses oi uses es sues . Malia-.3143 suses sues em a a sues siluses...._r|.s._u.lse_s suees 1 sues suses emu IIIHHHusT a a sues sues mo Illa!!! ; - zsofizs a ml as - as s cm 108 ."hither "a \euewu 1 use? so vv \uoeéuQ Basses v \uon.wu\ seofi usees usees Hbddb ussus usfis usfis usees usees . _ suns usees _ usees usfls suses : . ._ uses .Hbdb 55> .Hbg di “96> all a. 1.0 229,... Hbdb bum.» “SE on." 00." and” one.” on." ONH 109 "thither" \Hucwe mo vv $3ch I Han.wo\ gonna v \Hon.mc\ ”Goa L Ila - gas + usees usees usees uses [Ins-Lu u uses Asvuses sues uses uses-ill rlllllsusees s us. . w us e es uses suuses usAeves uses Juses llmdlls s s a .. Nu “flab as." cod oma 0an oma ONH mo buntewgi mo vv \uoe.wmos\ Basses v \uos.wmos\ Mesa b L H>AP> u e Asvuses sues 110 >H§> "together" h>m> Abdnb seems sssms H>w> Hag 05H Omd oma .O#H omH ONH m0 111 "weather" \uoems_mo vv \uoe.ws\_m=osse v \uoe.ws\.meosm ..ll Hbcdb u.. . es usuu.. usnees usees useeves usees .useeoes uses uses sueeves sues w nsvuses uses usee es uses usees _ sues uses Hbdb % . s o. . .0 Q. o . all . . . . .h.’ [ab . .. 1 sues .. . :2. J a... Er r one owe one use one owe mo 112 \uoew§\ mo vv \uoe...§s\ Samson v \uoe.ms\ Meozm "whither" .Hbfipb H>PP> .Hbeepb .Hbfipb H>n9> .thpb H>e3> H>GU> H>Ud> .Hbdcb H>Ud> 9,. E235 anpb H>dd> .Hbdb .Hbdb H>AP> .H> b M>QHH “Ha flbddb Hbdb l. Idlllsssus Us . rlua . oebvb . bunch. . s usenlldseel L J us moves L Wang r H J d1 flf L :3 , z - om esu owe one 0+2” on e owe Ho alder: 130 140 150 160 brother: 14C 15C 16C cider: 14C either: 140 (K) (EM) (N) (EM) (N) (EN) (80) (WM) (EM) (S) (SC) (EM) (EM) APPENDIX III: SOME MIDDLE ENGLISH DIALECTAL D an TH FORMS aldre, alder, alder, , aldir, broder, broder, bredur, breder, broder, brodyr, breder, broder, syderr, albre, alther althir althir alPer, alther brober brother brether brether, breoaer brother brother brother brother sither (N) ayder, aeiber, ether 113 elder: 150 (N) (EM) farther: 140 (S) 150 (S) 16C (EM) farthing: 16C (N) (EM) feather: 150 (EM) (WM) (S) fodder: 130 (N) 140 (EN) further: 16C (SC) (EM) (WM) 114 elder, elther elder , elther ferder, farther farder, farther ferder, farther farddyne. ferthyne ferdyns. ferthyns feder, feather feder, fether feder, fether fodder, fober fodder, fother furder, further furder, further furder, forther gather: 140 150 160 heather: 14C 150 hither: 140 15C leather: 150 16G mother: 15C murder: 140 150 (EM) (EM) (EM) (SC) (N) (EM) (8) (N) (EN) (EN) (EM) (so) (EM) (EM) (N) 115 gadyr, Syther’ geber gader, gather gad(d)re, gather haddyr, hather haddyr, hather heder, hether hyder, hither hider, hiPer hider, hither ledyr, lether ledder, leather modyre, mother moder, mother mordre , mrderour, morthere mordre, mourthered neither: 14C 15C nether: 160 other: 14C 16C rather: 13C rudder: 150 16C tether: 14C 150 tother: 140 (N) (WM) (SC) (N) (EM) (SC) (N) (WM) (EM) (SC) (N) (SC) (N) 116 neyder, neber neider, neiber neder, nether oder, obair oderr, oaerr uder, other udder, other rader, raber, raaer rodder, rother rudir, ruther tedyr, thether, tethire tedder, tether toder, tober under: 130 (K) 150 (N) whether: 140 (S) wither: 160 (EM) wonder: 140 (N) yonder: 140 (S) 11?. onder, onber ondire, onther, unber weder, whether wyder, wither wonder, wonber gonder, yonber BIBLIOGRAPHY 0F WORKS CONSULTED AND CITED BIBLIOGRAPHY 0F WORKS CONSULTED AND CITED Allen, William Sidney. Vox Latina. Cambridge, mglwd, 1965. Anderson, George K. The Literature of the gpglo- Saxons. Princeton, . Bellot, James (Jacques). Le Maistre d'Escole lois (1%0). Edited by’Theo Spifia. HEIIe, 1912. Bosworth, Joseph and T. Northcote Toller. An épglw Saxon Dictionary_ and Supplement (1921). rOnaon ’ O Braune, nghelm. Althochdeutches Lesebuch. Tfibingen, l 52. Brook, George L. Epglish Dialects. London, 1963. . ,A.Histo of the Epglish Langpgge. Fair- lawn, . ., 58:- Brunner, Karl. Altepglische Grammatik nach der el-sac sc en, rammatiE von Eduard Sievers. Egg edition, revised. HaIle, 1951. Bfilbring Karl. Altepglisches Elementarbuch. Part I, Phonology. el e erg, 1902. Cambridge yBibliograppypf gpglishfl Literature. Edited . . a eson. vo s. an supp ement. Cambridge, England, 1966. Campbell, Alistair. Old Epglish Grammar. Oxford, 1959. Chomsky, Noam._ "Current Issues in Linguistic Theory." The Structure of Lan a e: Readi s in the Philosopgy fiEdlte y Jerry A. or an JarrEIHsfigfii. Englewood Cliffs, N. J., 1964, pp. 50~118. and Morris Halls. The Sound Pattern.pf Epglish. New York, 1968. Classen, E. Outlines of the Histo '2; the English amas- Elfin. I9 118 119 Crystal, David and Randolph Quirk. ESgstems of Pro- sodic and Paral' istic Pea es ggngan. IKE—Hague, I965. _- Daines Simon. Orthoepia figglicana (1640). Edited ’ by M. Roes er an . ro anek. Halle, 1908. Danielsson, Bror. John Hart's Works on lish Orthoggaggy and Pronunciation—I , 1520. vols. tockfidlm, I955 {I}, 1 3 {I . Davis, Nbrman. "Scribal Variation in Fifteenth Century English." Helggges,gg,Lingfiistigue et gg,Philolo ie: Fernan sse gg emoriam. Pans. . pp. 9;Ib'3'." "'"'" "'"' ' ' DenholmfiYoung, N. Handwritin ,i§,§ggland and Wales. Cardiff, 19 . Dobson Eric J. lish.Pronunciation 1200-1200 ’ 2 vols. ‘gggdid: . , Ellis, Alexander J. On Earl lish.Pronunciation. Early English-Text Sociegy, Extra SEries, nos. 2, 7, 14, 23, 55. London, 1869. Ekwall, Eilert. The Concise Oxford Dictionggf of Egglish.PIace~Names. 55 edJ-Oifdrd, 960. Ferguson, Charles A. Review of the Sound Pattern of Russian by Mbrris HalIE. ngggggg7'XXXVIII? 5 (1962}, 284-297. ' Fox, John H. and Robin Wood. A Concise Histo ,2; the French ngggggg. ‘oxrasat'rgas. Gleason, Henry. ‘52 Introduction to Descri tive Linggistics. Revised Editdon. New York, 1961. Hall, John R.'5 Concise Anglo-Sax Dictionggy with Su lement By Her ert . r t. ed. CamErIdEe, England, 1960. ~ Halle, Mbrris and Samuel Keyser. ,"Chaucer and thII Study of Prosody." College En lish. XXVI : 3 (1966). 187-219. ’ Hockett, Charles F. A Course 33 Modern Ligggistics. New York, 19580 120 Hoenigswald, Henry M. a e e Lng Linguistic Reconstruction. L8E¥cago%.— Holmes, Urban T. and Alexander H. Schutz. A Histo 9; the French M. New York, I 58. Jackson Kenneth. e and Histo in Early ’ Cam Lfir $33,713 Britain. $38., 3. Jespersen. Otto. A InModern ngglishm Grammar on His- toricalh Roi es. el- Eerg, I9Pr Jones, Daniel. aPronounc Dictionggy. 4"! ed. New or1 , 9 . The Pronunciation of English. 4“ ed. Cambridge, H'a'ss” I953: Jordan, Richard. Handbuch der mittele lischen Grammatik. Part I. HeideIEerg, I925. Kennedy, Arthur D. Biblioggapgy of Writ s on the lish e rom e'Beeeg Ego TPr1nti_z_1g 880, o _ng En 9__ 22. 0'55er Klein, Ernest. _A ComErehensive figological Dictiogagz of the %g Ringgfi; vo s. s e an, Landon, ew or , K6keritz, Helge. _S__lhake eare' s Pronunciation. New Haven anc on on, 9 Krapp, George and Elliot V. K. Dobbie (eds.). Exeter Book. Morningside Heights, N.Y., 61956. Kurath, Hans. A Phonolo and Proson 23; Modern English. or, I . "The Loss of Long Consonants and the Rise of Voiced Fricatives in Middle English. " Lan a e, mm (1956), 435-445. Lehmann, Winfred. Historical nggistics: an Introduction. New York, 1 . A Re der in Nineteenth-Gen Historical Indo-fiopean Ifiggg' 1' stics. 1BI“oomIn_g'Edn—,"I967. 121 Lehmann, Winfred and Yakov Malkiel (eds.). Directions for Historical Linguistics. Austin and London, Lily, William. ,A 8horte Introduction of Grammar II. BrevissideIEEEItEEIBT'LEEHBE;‘I . Luick, Karl. Historische Grammatik der egglischen Sprache. I. Leipzig, I9IEZI9EOT Luquiens, Frederick B. Ag Introduction 32 Old French Phonology and MbgphoIogy. New Haven, 1959. Martinet, André. Elements of General Lin istics. Translated 5y EIisaFEtE PdImer. EEicago, 1960. Mason, George. Grammaire An loise nach den Drucken von 1622 un . Edited'by'fiuddIf Brotanek. ‘Hdile, I905: Menner, Robert J. "Conflict of Homonyms in English." ngggggg, XII (1956), 229-244. Middle En lish Diction . Edited by Hans Kurath, SEerman , an John.Reidy. Ann Arbor, 1952. In progress. Moore, Samuel. Historical Outlines of ggglish Sounds and Inflections. Revised By-Zl er cEfiardt. or, o and Thomas Knott. The Elements 2; Old lish. 10“ ed. Revised by James HuIFert. Ann or, 1960. , Sanford Meech, and Harold Whitehall. "Middle English Dialect Characteristics and Dialect Boundaries." E35 3 and Studies in lish and Co arative Literature. UnivdEsigy of 'Hddhigan PdEIications in Language and Liter- ature, XIII. Ann Arbor, 1935. Mbssé, Fernand. A Handbook of Middle En lish. Translated By flames I: WaIEer. timore, 1952. Nbulton, William D. "The Stops and Spirants of Early Germanic." Lan a e, XXX (1954), 1-42. 122 Nist, John. Phonemics and Distinctive Features of Beowulf." Studies _i_._n__ Linguistics, XIII: 1-2 (1958), 25'356 . £6 Structural History pf English. New York, 19 . Onions, Charles T. The Oxford Dictionary 2; English Etymology. Oxford, 1966. Oxford English Dictionary. 1} vols. Oxford, 1955. Palmer, Leonard R. The Latin Langpage. London, 1954. Palsgrave John. Leclarcissment _d_e 1a lan e Francoys . Edited by-Ft Genin. Paris, IESZT‘SE' Patton, George A. and David Donaldson, (eds.). The "Gest storiale" of the Destruction of T750 . my gglish Text-Eociety. London,_I8 . Penzl, Herbert. " A Phonemic Change in Early Old English." Langpgge, XX (1944), 84-87. Peters, Robert. A Légggistic Histopy 93 English. New York, 1 . . "Phonic and Phonemic Long Consonants in Old English." Studies in Linguistics, XIX: 1, 2, 5, 4 (196;), In. Policardi, Silvio. Historical Outlines 2f the English 1111311369. MlIanl, 132:8. , Postal, Paul. Aspects pf Phonological Theory. New York, . Pyles, Thomas. The Origins and Develo ment pf the English.IEngage. NEG‘YEEij‘ISEE. Quirk, Randolph and C.L. Wrenn. _A_.r_1 Old English Grammar. New York, 1957. Salesbury, William. _A Plgfle and Familiar Introduction. London, 1567. 123 Scargill, Matthew H. NOtes pp the Development of the Principal Sounds pffIEdo-European thfopgh Proto—Germanic and West Germanic into Efiglish. Toronto, I95I. Schlauch,9Mgrgaret. The Epglish.§§pgpggg. Oxford, 1 5 . Skeat, Walter W. "The Use of‘Q for TH in Middle lish." Notes and @ueries, 9“ Series, X Get. 25’ I g6:), 32 - O Stockwell, Robert and C. Westbrook Barritt. Some Old English Gra hemic-Phonemic Correspondences --ae, 23, an ‘3. ew or , 51. Stratmann, Francis H. ,A Middle-English Dictionary. Revised by Henry ra ey. or , . Sweet, Henry. King Alfred's West Saxon version 2; Gregory 3 astordI Care. Kppendix I. London, . ,éONew English Grammar. 2 Parts. Oxford, 19 . . The Oldest En lish Texts. (Early English Teif'Soc1eE . Eondon, I885. Waterman John T. ‘A Histor of the German Lapguage. Seattle and London, 1966. Webster's New World Dictionar . College edition. Cleveland and New Yorfi, 1966. Webster's Third New International Dictionary. Spring- fieId , 1%. wells John E. ,A Manua1,p£ Writ' s in Middle Epglish. ’ Supplements I-IX. New Hiven,’I9I6:I952. wrenn, Charles L. Word and gypbol: Studies $3 the English Langgage. London, 1967. wright, Joseph. The Epglish Dialect Dictiongpy. 6 vols. Ldfidon, -I905. (Reissued, New York, 1963). 124 Wright, Joseph. English Dialect Grammar. Oxford, 1905. Also incorporated into a50ve. . Grammar of the Gothic Langpag . 2nd ed. Oxford , I953. wyld Henry C. Collected Papers pp Henry Sweet. ’ Oxford, I913} . The Historical Study pf the Mother Topgpe. London, 1920. . ,A Short History pf English. New York, 1914. MTlfl'l‘fiflifllifljfllfllfitflffliujflfiljl'lijfljflflrflfimflms