|
|
|
|
(delwedd F6355) (tudalen 000_1)
|
STUDIES
IN
WELSH PHONOLOGY
SAMUEL J. EVANS, M.A. (Lond.)
1870-1938
Author of The Elements of Welsh Grammar:
The Latin Element in Welsh;
Welsh
and English Exercises:
Welsh Parsing and Analysis:
Questions
and Notes on Welsh Grammar
Editor of "Drych y Prif Oesoedd " with Introduction and Notes
(Guild of Graduates' Series.)
Joint Editor Qf Chaucer's Prologue and Knight's Tale.
LONDON: DAVID NUTT, LONG ACRE, W.C.
NEWPORT, MON.: JOHN E. SOUTHALL, Dock STREET.
1909.
|
|
|
|
|
(delwedd F6356) (tudalen 000_2)
|
PREFACE.
THE right of Welsh to an honourable place in our system of education has been
amply recognised of late years in the schools and colleges of the
Principality. The movement for the introduction of this subject into the
curriculum is not an isolated fact, or the feverish dream of a few
irresponsible enthusiasts.
It
is a part of the Educational renascence which has stirred and possessed the
nation for the last forty years. It has a counterpart in the recent upheaval
in Western Europe and in America for the due recognition of the mother tongue
in education, both as a mental discipline and a material gain, and in the
main its methods must be those sanctioned for modern languages by the leading
educationists of all countries.
It is
universally recognised that a training in phonetics must constitute an
essential and organic part of the "Reform movement." But while
English, French, and German sounds have been carefully analysed, classified,
and compared, little has been done for this department of Welsh study outside
the lecture rooms of our National Colleges. Sir John Rhys's Lectures on Welsh
Philology - the only work yet issued that gives much attention to the subject
- is out of print, and its Phonology is mainly historical,
|
|
|
|
|
(delwedd F6357) (tudalen 000_3)
|
The
author is well aware that these Studies can lay no claim to any finality, but
he believes that as they embody careful observation extended over a number of
years, and a systematic attempt at equating and differentiating the sounds of
Welsh, English, and French — the three languages with which Welshmen are most
intimately concerned —they will not fail to advance the cause of Welsh
education, and enable the teacher to adopt more completely for Welsh and
English, the method already adopted in most secondary schools for French and
German. He has allowed himself the pleasure of a frequent digression into the
field of Etymology, wherever, by so doing, he could throw light on any point
of Phonetics.
The
author has derived useful hints from the writings of Sir John Rhys, Dr.
Silvan Evans, Professors Anwyl and Morris Jones, Zeuss and Loth, Brugmann,
Giles and Peile, and is in a special degree indebted for many suggestions to
the works on Phonetics by Viëtor and Passy, Sweet and Rippmann. He will be
grateful for any suggestions that may contribute to greater lucidity or
accuracy.
LLANGEFNI, December, 1908.
|
|
|
|
|
(delwedd F6358) (tudalen 000_4)
|
TABLE OF CONTENTS. CHAPTER CHAPTER
I.—VOWEL SOUNDS II. CONSONANTAL SOUNDS MUTATION OF INITIAL CONSONANTS CHAPTER
Ill. CHAPTER IV.—OrHER CHANGES AND THE ACCENT SOFTENING OF CONSONANTAL SOUNDS
PROVECTION AND BACK-FORMATION. • . EXCRESCENT SOUNDS ... LOST SOUNDS (AND
LETTERS) METATHESIS ACCENT PAGE 1 31 76 94 94 96 100 106 107 109
|
|
|
|
|
(delwedd F6359) (tudalen 000_5)
|
REFERENCES
& ABBREVIATIONS.
B.C. or Bardd Cwsc.: Professor Morris Jones's Edition, (Jarvis and Foster
Bangor).
Brutieu: The Oxford Edition, by Sir John Rhys and Dr. Gwenogvryn Evans.
D. ab G.: Barddoniaeth Dafydd ab Gwilym, o Grynhoad Owen Jones a William Owen
- Llundain 1789.
Dem FJydd:: .:: Gorchestion Beirdd
Cymru, Edited by Cynddelw, (Humphreys, Caernarfon). Cywyddau Goronwy Owen:
Mr. W. J. Gruffydd's Edition (Southall, Newport): Gwirionedd y Grefydd
Gristionogol, by Edward Samuel, Edited by Dr. Silvan Evans, (Spurrell,
Carmarthen). • Bishop Morgan's Translation (1588),. Edited by Dr. Gwenogvryn
Evans, (Henry Frowde).: The Myvyrian Archaiology of Wales— (Gee & Son,
Denbigh). Mabinogion--—rrhe Oxford Edition, by Sir John Rhys & Dr.
Gwenogvryn Evans. • Deffynniad Ffydd Eglwys Loegr. Guild of Graduates'
Series. /!/orqqan
Llwyd: Llyfr y Tri Aderyn. Drych y Prif. Oesoedd. do. do. )
's/orra: Ystorya de Carolo Magno, edited by Professor Powel (Cymmrodori01m
Society's Publications.) is sprung from, is derived from, passes into, is the
parent of.
|
|
|
|
|
(delwedd F6360) (tudalen 001)
|
WELSH
PHONOLOGY.
CHAPTER 1.
VOWELS.
No
study of the sounds of a language can be considered satisfactory which does
not investigate to some extent their physiological basis.
The
nature and the timbre or quality of every sound may be determined by
reference to the vocal organs concerned in its production. Thus p, t, c (as
in Welsh gardd, Eng. get) are called Mutes or Checks, because the passage of
the breath is momentarily stopped before the sounds are uttered. If the vocal
chords are not brought into action at all, the sounds are p, t, c, while for
b, d, and g some vibration is necessary.
The
Vowels on the other hand are primarily the result of the vibration of these
chords, but their distinctive timbre is determined by a number of secondary
tones produced by a peculiar configuration of the mouth. The shape of the
cavity is largely dependent upon the position of the stopper or tongue and
the lips. For one
|
|
|
|
|
(delwedd F6361) (tudalen 002)
|
WELSH PHONOLOGY.
position of the stopper the mouth may be closed or at any intermediate stage
up to wide open. Again, the lips may be drawn back, yielding ' unrounded '
vowel sounds, while " rounded " vowels are those produced when the
lips are brought over the mouth so as to diminish the orifice. French u as in
vu, lune, and Welsh w in cwm, swn, are extreme instances of rounded vowels.
TABLE OF VOWEL SOUNDS.
All attelul)t is here Ill'ade to
indicate diagranlmatically the position of the tongue, and the part ef it
specially adjusted in the articulation of various vowel sounds heard in Welsh
and English:
Front of Tongue. Tongue highest.
Close i open i close e close u open u Back of Tongue. close w open w close o open o (o) open e (G) neutral (e) very open e (æ) a a Tongue lowest.
In producing vowels no friction or
stoppage must occur. The voice has free play, but with a given opening of the
mouth, as measured by the angle of the jaws, the blade* of the tongue (e.g.)
may be brought near *
Tip, blade, front, middle, back are
the terms generally used for consecutive parts of the tongue from tip
backwards.
|
|
|
|
|
(delwedd F6362) (tudalen 003)
|
WELSH PHONOLOGY.
3 tille front palate, yielding the '
close i ' of the table: or if the mouth be somewhat more open and the
interval between the tongue and palate be in the slightest degree increased
the timbre is that of ' open i.' This shade of difference in the aperture,
together with an accompanying small change of the part raised of the tongue,
constitutes the distinction indicated in the table by the epithets ' close '
and ' open. '
The symbols in brackets are from the
Alphabet of the Association Phonetique Internationale, and will be used for
phonetic transcription in this book. We will now discuss the vowel sounds in
greater detail.
I.—MEDIO-PALATAL: clear a, neutral
a.
a: This sound is produced by very
slightly raising the fronto-medial part of the tongue. It is heard in
Welsh—calon, tad. French—ma, rage. It is unknown in Standard English, but in
the North the ' a ' in such words as ' pat ' and ' man ' has this value. 011
the other hand, English has two 'a'-sounds unknown in Welsh outside certain
dialects, viz. (i.) a, as in father. It is produced by slightly lifting the
middle of the tongue, that is, the part immediately behind that raised in
articulating the ' a ' in Welsh tad. It IS the sound of ' a ' in French åme,
pas. 'IAhis is the timbre often given to ' a ' in Cloriannydd, and the first
' a ' in Bala by English speakers.
|
|
|
|
|
(delwedd F6363) (tudalen 004)
|
4 WELSH PHONOLOGY.
The failure of many Welshmen to
produce without much practice this English a sound is one of the chief
elements in the so-called Welsh accent imported by them into English speech.
(ii.) æ: This perhaps is the commonest ' a '- sound in English. Heard in
glad, man, sad, &c. It is not as open as the Welsh ' a,' the angle of the
jaws being somewhat smaller and the tongue higher. An Englishman speaking
Welsh provokes a smile on the part of a native more on account of the
peculiar æ timbre he gives to the Welsh ' a ' than because of his blunders
over ' 11 and ' Ch.' IL—FRONT VOWELS. Open e (e): as in eto, erddi, echnos.
This sound occurs also in English, as ' let,' ' met.' ai ' in French paix,
and the Close e: as in rhed • first part of the diphthong in English pale,
make. Welsh close long e is a simple sound, while in English it is invariably
diphthongal, thus ' make ' and ' pale ' are pronounced meik, peil. Failure to
notice this difference accounts for the occasional articulation of these
words as mék, pél, on Welsh lips. Teachers in Elementary and County Schools
in the Principality will readily appreciate this point. Open i: as in erlid •
Close i: as in hir: in si. English, pill. English, ee in peel: French, i
|
|
|
|
|
(delwedd F6364) (tudalen 005)
|
WELSH PHONOLOGY.
Ill.—BACK VOWELS. 5 It should be
noticed that for these the lips are generally rounded, especially when the
tongue is much raised, just as for the Front Vowels already considered, the
lips are more and more unrounded in the passage along the series from ' a '
to ' i.' Open o (o): long and short in English and French, e.g. English
all—pronounced o:l, Paul ( crop ( = krop). In Welsh this sound is always
short, as in hon ( hon) , trom ( = trom). Hence a Welshman's first attempt at
pronouncing long open ' o ' is generally a failure. He will substitute for it
the long close ' o,' with which he is quite familiar in Welsh (see below), so
that on his lips the first in each of the following pairs of English words is
undistinguishable from the second: law and low, saw and so, Paul and pole,
hall and hole, naught and note, Saul and soul or sole. Close o: always long
and always a simple vowel sound in Welsh: as tro; 110: Paul and Saul as
popularly pronounced in Welsh ( =p61, s61): =eau in French beau, and o or ow
in Scotch no, know. In Stand ard English long close ' o ' is always
diphthongal. Hence the Welshman who pronounces English ' note ' as ' not,'
and the Englishman who
|
|
|
|
|
(delwedd F6365) (tudalen 006)
|
WELSH PHONOLOGY.
pronounces Welsh tro as trow, are
equally wide of the mark. Open w: always short, as trwm: so in English, full.
oo in English pool, Close w: examples llw, mwg and ou in French sou. lv.
MIXED VOWELS. 'A' and ' a ' have already been referred to (see i. above). As
the neutral will have to be discussed in greater detail, we shall begin this
series with Close u: as u in Welsh un, pur, and y in h9n. Open u: as ' y ' in
hyn, byr, and ' u ' in alltud. These two sounds differ from French ' u ' in
that the latter is (1) a blade and not a front or medial sound, and (2) more
rounded. Neutral: This sound is usually represented by phoneticians as an
inverted ' e ' thus a. It is produced by closing the mouth rather more than
is done in articulating the a of English father, and slightly elevating the
middle of the tongue. It is variously represented in English By ' a ' in
(e.g.) India, America, attend, villa, ar er or o beggar, altar. father, weaker. sudden. purple. sailor. son, harmony.
|
|
|
|
|
(delwedd F6366) (tudalen 007)
|
WELSH PHONOLOGY.
7 In works on English it is often
called the ' obscure natural vowel a natural ' because the sound is or
produced with the minimum of effort. It is also known o schwa ' or ' neutral
vowel,' and its commonest ill Latin, Teutonic, and Celtic is ' a;' thus the
Aryan or Indo-Germanic hypothetical form po-ter- is in Latin pater, Gothic
fadar, and Old Irish athir.* In Welsh it is usually referred to as the '
obscure y ' sound because it is oftenest represented by that symbol. It is
the regular value of ' y ' in non-final syllables, as in chwychwi, tydi,
hynod, and also in proclitics as y, yr, yn, fy, dy. ' V' however is not the
only symbol used in Welsh to represent this sound, though the fiction whereby
it is commonly asserted that Welsh words are phonetically spelt, has tended
to the assimilation of the sounds of a word to its spelling. We need no
further proof of this effort at harmonising sound and spelling than that contained
in the Possessive ei, his. In Me(liaeval Welsh this was written ' y,' but
since William Salesbury gave currency to the new and ullreal form ' ei,' all
effort llas been made in formal speech to pronounce t he word as a diphthong.
Now, vvhen in speaking t he 111011th is slightly more open than is the case
when uttering neutral a, we get a sound still obscure, but hardly
distinguishable from ' a,' and it is an interesting instructive phenomenon in
Welsh orthography t lint in originally unaccented syllables—and the origin *
Indo-Germanic p is lost in Celtic, except before another connonalit, when it
becomes a spirant.
|
|
|
|
|
(delwedd F6367) (tudalen 008)
|
WELSH PHONOLOGY.
of this sound is to be looked for in
the absence of accent a ' and ' y ' are very freely interchanged. We shall
give a fairly long list, as the question will come up again in various forms.
INTERCHANGE OF A AND V. Mediaeval yn and an English our. Mediaeval ych and
awch* = English your. The Relative Pronoun y for a is sometimes met with in
Mediaeval and early Modern Welsh V wneuthur yn gyuurd gwr ac y bei arglwyd ar
y sawl vrenhined hynny. '—Mab., 82. Nid ei gair nhwy y saif. '—Llyfr y Tri
Aderyn, 192. Y digred y welaist gynneu.' Bardd Cwsc, 32. P wy a haeddei
uffern well na chwi, y fyddei'n hel ac Yll dyfeisio chwedleu. '—Ibid. 95. O flaen pob Llith, y Gweinidog y
ddywed.' Rubric the Te Deum in the Prayer Book. For the same
reason a is occasionally used where we should now use y: Y mae ynys parth
hwnt y ffreine yn gadwedic or mor o bop tu idi, ac a uu gewri gynt yn y chyuanhedu.' Bruts 52. Od oes ddim ynom a ellir craffu arno.—-Defryniad Ffydd, 65 (cf
pp. 92, 150). * For the ' w ' in awch see below.
|
|
|
|
|
(delwedd F6368) (tudalen 009)
|
Ymha herwydd (y dolwg) ? Defyniad Ffvdd,
191. amgeledd and ymgeledd Do, ebr ynte, rai unic a dihelp a Phell oddiwrth
ymgeledd. '—Bardd Cwsc, 57. amddifad and ymddifad. anlddiffyn and ymddiffyn.
ychwaneg and achwanec (Mediaeval) NIab. 209. ath iarllaeth titheu heuyt yn
achwanec. y vreham for Abraham. Ystorya de Carolo, 20. anial and ynial: '
Vnteu peredur a gyehwynnwys racdaw. Sef y deuth y goet mawr ynyal. '—Mab.
200. Angharad and Vgharat (Mediaeval): Nachaf ygharat law eurawc yn kyuaruot
ac ef.' Mao. 215. {Illi(lano and ymdano . 'A pheredur a gyuodes, ac a wisgawd
y arueu ym(lanaw ac
ymdan y uarch.'—Mab. 217. y for y and afory: a thost yw gennyf
welet ar was kyn uonhedicket a thi y (lihenyd a vyd arnat avory.' Mab. 216.
|
|
|
|
|
(delwedd F6369) (tudalen 010)
|
10 WELSH PHONOLOGY.
canhebrwng and cynhebrwng: dyma
Ganhebrwng yn mynd heibio. '—Bardd Cwsc, 28. cyn and the somewhat unliterary
can, with the Equal Degree: A phe bai un gan ffoled a gwneuthur hynny. '—G.
0., Llythyrau. canfas and cynfas. dynwared and danwared: danwaret y
kyweirdabei a welsei . Vlab. 195. Mediaeval ys and as,—Extended forms of the
Infixed Pronoun of the 3rd person: Pei as gattei lit idaw. '—M ab. 274. Pei
as gwypwn mi ae dywedwn.' Mab. 130. Pei ys gwypwn ny down yma. '—M ab. 29.
ambell and ymbell: Rhaid yw cyd-ddwyn ag ymbell fai. '—Deffyniad Ffydd, p.
xiii. Ilysenw, and the colloquial North Wales llasenw. This word is sometimes
' blasenw ' in Anglesey. afagddu and y fagddu: 'A chroesaw fyd tiffernol ! a
thydi Afagddu, ddyfnaf lyngclyn, derbyn fi.' I. D. Ffraid: coli Gwyn/a, i.
303-4. Cydfydd y Fall a'i gallawr, Cår lechu'n y fagddu fawre '—G. O.: Y Farn Fawr. It will be noticed that the variation is commonest in the initial
syllable, but it occurs in other positions, as—
|
|
|
|
|
(delwedd F6370) (tudalen 011)
|
WELSH PHONOLOGY.
boly and bola, from the Brythonic
bolg. Beth yssyd yn y boly hwnn heb ef. '—M ab. 38, hely and hela, from the
Celtic root selg. 11 eiry and eira, from the root seen in Latin argentum: [V]
petheu draw sy'n perthyn i drafaelwyr mynyddoed(l eiryog. '—Bardd Cwsc, 57.
Silliilarly, while gwely is the usual form of the word for bcd, the plural
gwelåu is from the variant singular gwela* ( > gwela-au > gwelåu):
Gwelem rai ar welåu sidanblu. '-—Bardd Cwsc, 23. It Illust not be supposed
that, when these doublets arose, the -y and -a of hely and hela, &c., had
their values in these words. They are rather the two sytnbols which most
nearly represented the neutral 9 Nound which resulted from the vocalisation
of the -g. 'Phe same attempt at representing the obscure vowel Nound is seen
in /»v• a common variant of ' pa ' in Mediaeval litera'A gofyn a oruc idi py
dir oed hwnnw a phy lee' Mab. 184. vowels in unaccented syllables are not
clearly in any language, and this reacts on the for if ' a,' e,' and ' y '
acquire an approxitillitely coinmon value in that position, it follows that
the sound in any given word may be variously reprenetlted by different writers
or even by the same writer (lilTerent tillies sometimes by ' a' and at other
tittles by 'y ' or ' e.' Of course, in a language like re note in Prof.
Morris Jones 's of Bardd Cwsc.
|
|
|
|
|
(delwedd F6371) (tudalen 012)
|
WELSH PHONOLOGY.
Welsh, where popular etymology has
played such a prominent part, orthography will shew abundant evidence of it.
Further, adjoining sounds frequently determine the character of the vowel:
thus, ' n,' being an unrounded dental, tends to clarify or palatalize an
obscure or guttural sound, with the result that the sequence ' an ' or ' yn '
sometimes appears as ' en.' Compare English ' Harry ' and ' Hal ' with '
Henry.' In an unaccented syllable the reverse process is not uncommon (see
below). A AND E. Med. -ei > Mod. -ai, as, dysgei > dysgai. Med. -eu
> Mod. -au, as plural SUffX. agor and egor The latter is the regular form
in Lly/r y Tri Aderyn. angraifft and engraifft Kymer agreift o lawer o
betheu. Ystorya de Carolo, 22. anrhydedd and enrhydedd: Mi ath vyrywys yr
enryded a gwassanaeth idaw ef. ' Mab. 200. -deb: the familiar substantival
suffix is -dab in 'Ai cudab yv•l rhoi codwm ? '—G. O.: Cywydd y Calan. Duwdab, dyndab a swydd lestl Grist. '—Contents of 1st Ch. of Gospel acc.
to St. John. fal ( < hafal), now fel. wynebwarth and wynabwarth: 'Ac yn
wynabwerth idi hitheu dy vwrw o honaf i.' Mab. 210.
|
|
|
|
|
(delwedd F6372) (tudalen 013)
|
WELSH PHONOLOGY.
The intensive prefix ' en- ' assumes
the form in a few words on the analogy of the negative ' ananaele en + aele: anghysbell <
en + cysbell • 13 anannog < en + root seen in hogi, awch,
eg-ni,* diog,: English ' to egg: ' Latin acer, &c. annwn < en + dwfn,
hence the extremely deep or bottomless pit: anrheg < en + rheg, a gift ,
ansawdd < en + sawdd. Y AND E. brodyr and older broder.t 'Afiachus fu
faich oes fer, Echdoe fryd eich dau froder.' Tudur Aled: G.B.C. 228. deall
and dyall. 'Ac yna y dyallawd Peredur. '—M ab. 216. dyred (now tyred) and
Demetian dere • egni and yni ennill and ynnill: Ny allwn i vyth ennill vy
arglwyd i o dyn arall.' Mab. 176. Ni fynnem bei allem dy ynnill di. '—Ll. y
T. Ad. 188, ysgar and esgar , ysgymun and esgymun. * Dr. Silvan Evans
incorrectly analyses ' egni ' into e + gni. Compare the Author's ' Latin
Element in Welsh,' p. 8.
|
|
|
|
|
(delwedd F6373) (tudalen 014)
|
14 WELSH PHONOLOGY, Mercher—older
Merchur—is sometimes Merchyr: Dyw merchyr mae'n bybyr bwyll, Poed yn ddyw
sadwrn didwy11.'—-D. ab G. cxxix. Dydd Merchur y Lludw. '—Prayer Book. Er,'
for, because of, notwithstanding, is regularly ' yr in Mediaeval literature:
Vr y lawn werth. '—Mabe 179, and similarly yrof ( = erof) and yrdaw erddo) as
in Mab. 105. On the other hand the Definite Article ' y ' was sometimes
written ' e,' as E brenhin yna a disgynnawd. '—St. Great, S 85. The Mediaeval
Possessive Adjective ' y ' (his, her, its their) was regularly written ' e '
when (1) Suffixed to a or o . N yt eynt hwy oe bod.' Mab. 32. V wreic ae gwr
ae phlant.' Mabe 32. (2) Used with hun, self: 'A Chyn penn y pedwyryd mis wynt ehun yn peri eu hatgassau.'—Mab. 32. In the same way the Postvocalic or Infixed Personal
Pronoun was ' e ' after ' a,' as Mi ae dywedaf itt, '—Mab. 200. Mi ae Iledeis. '—--NIab. 200, The diminutive suffixes -yn (m.) , -en (f.) , and -an
(common) In Gwynedd the suffix -en is generally pronounced -an, as merlan for
merlen and hogan for hogen,
|
|
|
|
|
(delwedd F6374) (tudalen 015)
|
WELSH PHONOLOGY.
15 while doublets like archan and archen
have literary sanction. Further, the interchange of -yn and -an is not
infrequent in literature, as, dynyn and dynan. -an ' has gained currency as a
Hence the form doublet of both ' -yn ' and ' -en.' Examples of the
interchange of ' y,' ' a,' and ' e ' need not be further multiplied. As far
as its origin is accentual—i.e. where analogy or adjoining sounds do not
account for it—it is an illustration of the Law of Ablaut or Vowel Gradation
which is so marked a feature of Strong Verbs in English. The ultimate result
of this weakening due to the absence of accent is the total elimination of
the sound. Thus, while u in cynnull is distinctly articulated, it is
generally elided in cynulleidfa. Maurice Kymn and Elis Wyn leave the ' u '
out in writing. An ' y ' sound follows the ' g ' in tragywydd, but it is
seldom heard and not often written in trag(y)wyddol and trag(y)wyddoldeb •
Dirgelion tragwyddoldeb nis gwn i, A darllain gair o'u gwers nis gelli die'
Caniadau Pro/. J. Morris Jones, 166 For the same reason ' io ' is dropped
from ' Cristnogion ' in ' Defyniad Ffydd,' M. Kymn, as Ni a wyddom fod yn
amser yr Apostolion ddynion Cristnogion. '—65. Plas (compare English Palace):
baswn and basai for buaswn, buasai: rheiny for rhai hynny, and the Venedotian
Cleta and dled for caletaf and dyled, are all instances of the operation of
the Law of Ablaut.
|
|
|
|
|
(delwedd F6375) (tudalen 016)
|
16 WELSH PHONOLOGY.
O AND A. The interchange of o and a is
not uncommon in Welsh. In some instances a is merely a weakened o. Examples
'Meredic a ( = o)wyr' (you are) strange men. —Mab. 126. O'r a ( = Lat ex eis
qui), as Dugant bop peth or a oed reit herwyd eu deuawt wrth aberthu gantunt.
'—Brutieu, p. 52. This phrase in Modern Welsh is generally ' a'r a,' and even
' ar a,' as— Yr wyf yn gobeithio fod yr Hanes 011 mor gywir ac mor llawn hefyd ar a ellir ei
ddysgwyl. '—D. y P. O. 7a. Achos from Latin occasio. Achub from
Latin occupo. Sawdwr or sawdiwr from Mid. Eng. soudiour. Beth yw Sawdwr
Iledlwm addycco dy ddillad wrth ei gleddyf, wrth y Cyfreithwyr ? '—Bardd
Cwsc. (Y m) achlud from Lat. occludo. Yrwan is from yr awron ( = yr awr hon)
and not from a diminutive form of awr with the Definite Article ( = yr awran)
as has been sometimes asserted. -am and -om:—am common in Mediaeval
literature and even as late as the 17th century in the 1st plural of the
Aorist and of Pronominal Prepositions, e.g. Ni a doetham y erchi olwen yr gwas hwnn. '—-Mabe 117
(cf. pp. 129, 130, et Passim). Ie heb ynteu bwyll ni awn yr un niver y buam
doe. Mao. 10,
|
|
|
|
|
(delwedd F6376) (tudalen 017)
|
WELSH PHONOLOGY.
17 Diolwch y duw kaffel 0-honat y
gedymdeithas honno, ar arglwyddiaeth a gawssam ninheu. '—Mab. 8. With this
regular Mediaeval termination we may compare the -am of the Old Irish
preterite 1st person plu. As m is a rounded consonant, the tendency is to
convert the unrounded and palatal ' a ' into the rounded and more guttural o.
Hence -om is the regular ending in present day Welsh of the 1st plural Aorist,
and 1st plural of Pronominal Prepositions of the ' ataf ' class. As daethom,
buom, cawsom: arnom (Mediaeval arnam). Po and pa with Superlatives Pa uchaf
yr ymgodant isaf y cwympant.—Lly/r y Tri Aderyn, p. 212. 'A' changes into ' o
' especially when adjoining a guttural like ' w ' or a rounded sound like ' f
' or ' m.' In writers of the 16th century and later ' aw ' tends to pass into
' ow ' in words of more than one syllable a change practically unknown in
Mediæval literature. Maurice Kyffm in Deffynniad Ffydd Eglwys Loegr (1595)
shews an especial fondness for ' ow: ' thus he writes— nowfed, anhowsder,
howsach, yr owron, mowrion, cowri, iniownder, odidowgrwydd, &c. Bardd
Cwsc has Ilownion for llawnion. So in Demetian and Gwentian, forms like mawr,
mawredd, are regularly mowr, mowredd. Mediaeval cawad (a shower) is now more
usually cawod or cafod, though the ' a ' is still retained colloquially in
Demetian. But the student must here be warned against the somewhat natural
tendency to regard a 2
|
|
|
|
|
(delwedd F6377) (tudalen 018)
|
18 WELSH PHONOLOGY.
modern literary form as a lineal
descendant of that found in our Mediaeval literature. The best prose of the
pre-Reformation era was in South Welsh. The Mabinogion, the Bruts, and the
Laws of I-Iywel Dda were all written in that dialect. On the other hand, in
the 16th and 17th centuries and early 18th, a galaxy of able writers and
translators in North Wales eclipsed their contemporaries in the South, and
their productions came to be regarded as the standard of Welsh composition.
It. follows that many Venedotian peculiarities gained literary currency for
the first time, supplanting rather than growing out of the Demetian
peculiarities of older literature. Thus it is more correct to regard cawod
and cawad as coexisting in sister-dialects than as having developed one out
of the other. The same explanation holds good in the case of Demetian ef a
with): both are etymoor ef å) and Venedotian e/o ( logically the same and
mean literally ' he and ' or ' he with.' It is natural that in the form ' efo
' the etymology of the phrase should be obscured, and then ' efo acquired the
general meaning ' with,' and came to be used with other than the 3rd sing. masculine,
e.g. Dos di efo
foe Gweddiwn efo'n cyd-Gristianogion. '—G. Mechain. It may
be noted as further evidence of the origin of efo that before vowels the
fuller form ' efog ' is sometimes used ef åg), annogaeth i bob dyn i ystyried
eu ' Efog ( ffyrdd a'u crefydd. '—Agoriad (1703).
|
|
|
|
|
(delwedd F6378) (tudalen 019)
|
WELSH PHONOLOGY.
19 The simple preposition o (for å) is
often used in the dialect of Glamorgan to denote the instrument, as, ' Codi
glo o'r rhaw (å'r rhaw) = throwing up coal with a shovel. * It is interesting
to note that the history of the English language shews a similar break of
continuity. Modern standard English is a direct descendant not of the
vigorous i West Saxon dialect of King Alfred—the dialect in which all extant
Old English literature is written— but of the Midland or Mercian which gained
literary currency and precedence in the 14th century due to a complexity of
causes, but mainly because that consummate literary artist—Chaucer—wrote his
inimitable Canterbury Tales in that dialect. Thus it is that we write ' did '
and not ' dud,' which would be the Modern form of the Old English (i.e. West
Saxon) dyde, and, to take only one other instance, the word all is sprung
from the old Midland all, and could not be explained by reference to the West
Saxon eall. But to return to the interchange of ' a ' and ' o,' a few other
instances may be mentioned Vstondard ' (from Eng. standard). ——1/1 ab. 155.
Bonllef ' ( = banllef) .—Bardd Cwsc, p. 107. Na, nag, and Mediaeval no, nog,
after a comparative. Orgrephid for argrephid.—Llyfr Job (Dr. Morgan), 43.
Coron and coran. Cawgiau and cowgiau, &c. In Anglesey and Carnarvonshire
dafad is regularly pronounced dafod. * Quoted from Mr. John Griffiths's V
Wenhwyseg, ' p. 19. Published by J. E. Southall, Newport.
|
|
|
|
|
(delwedd F6379) (tudalen 020)
|
20 WELSH PHONOLOGY.
ow < yw: w,' being a high back
rounded vowel sound, tends to assimilate ' y,' as well as ' a,' into the intermediate
back rounded vowel ' o,' especially in colloquial speech: ' bowyd,' '
tragowydd,' for ' bywyd,' ' tragywydd.' Instances are found in literature:
cowir ' for ' cywir Codi gwal cauedig wedd Caer enwog, cowir rinwedd.' Edw.
Morus, Cywydd y Paen. cowydd ' for cywydd 'Am Englyn neu Owdl neu Gowydd e
wyr pawb nad cynefin ond i un dyn ar unwaith ganu'r un o'r rheini, '—Def.
Ffydd, xv. clowem ' for ' clywem . Er mwyn na chlowem ar eyn calonneu adel
Ilugru gwir wasanaeth Dduw drwy'r fath ofregedd•' Def. Ffydd, 46. Howel (Eng.
Howell, Howells) < Hywel. MUTATION AND SIMPLIFICATION OF VOWEL SOUNDS. *
Besides those already discussed, the leading vowel changes in Welsh may be
thus classified: * Compare Prof. Anwyl 's Grammar and Prof. Morris Jones's
Welsh Orthography.
|
|
|
|
|
(delwedd F6380) (tudalen 021)
|
21 A. WELSH PHONOLOGY.
Simplification in final unaccented
syllables e as byddai
bydde y o Final ' -aw eleirch elyrch gadael > gadel adwaen
> adwen ysgawl > ysgol brawd (judgment) > difrod (' -ew ') sometimes
passes into ' -au -eu,' as tawaf > 3rd sing. tau, and goleu side by side
with Ilewyrch. This ' au ' or ' eu ' is further simplified in a few
instances, as gwrendy for gwrandau or gwrandeu from gwrandawaf: ymedy for
ymadau or ymadeu from ymadawaf • tery < tarau or tareu from tarawaf.
Bo—Change in a final syllable due to the suffixing of another syllable all
> eu as ffau—ffeuau U llai—lleiaf crwm—cryman pawl—polion brawd—brodyr
uwch—uchder Iluwch Iluchio y (clear) > y (obscure) bryn—bryniau.
C.—Affection or mutation due to— I. lost final a: y > e as gwyn—gwen w
> o „ hwn—hon
|
|
|
|
|
(delwedd F6381) (tudalen 022)
|
22 WELSH PHONOLOGY.
vowel i, whether lost or II.
consonantal i or long retained: meirch a > ai or ei as march gweiniaid
drain —main dywaid ae > ai e > ei, ai oe > wy gwan draen maen dywed
nifer-- nifeiri cyfer—cyfair and cyfeiriad hén—hjn gwél—gwjl agoraf—egyr
porth—pyrth adroddaf—edrydd oen—--wyn Ill. The vowel i of the next syllable:
a > e as can—ceni saer—seiri ae > ei , maen—meini aw > ew as
tawaf—tewi u (y) > i „ dysgit from older dysgud ti, dysgyd tie As the last
change throws much light upon the history of an important inflectional form,
illustrative examples are appended— The first stage is the Mediæval -ut, as,
Pa yssyd arnat ti, pryt na delut y edrych y gofut a uu arnaf i. '—Mab. 176.
Later, with the unrounding of u towards the close of the Mediaeval period, u
and (clear) y were confused and
|
|
|
|
|
(delwedd F6382) (tudalen 023)
|
WELSH PHONOLOGY.
23 freely interchanged as may be
readily seen by even a casual perusal of Defynniad Ffydd Eglwys Loegr (1595),
or of Lly/r y Tri Aderyn. Hence the existence of the doublet forms gwneuthym,
deuthym, and euthym by the side of the better attested and older gwneuthum,
deuthum, and euthum, and hence the frequent appearance of ' y ' in the 2nd sing.
Past Imperfect, as, Goddau f'armerth, o'm nerthyd, V w Dydd Barn a diwedd
byd.' Goronwy Owen, Cywydd y Farn Fawr. Then y passed into i, due to the thin
i of the following ti.' Hence the present day forms ' delit,' nerthit,'
dysgit,' &c. The objection may readily be raised—why then has not the i
of the final syllable affected the y in the Pronominal Prepositions gennyf,
gennyt, wrthyf, wrthyt, ? The answer is not far to seek. The present-day
spelling of these is a case of an artificial reversion to the Mediaeval forms
after an intervening period when gennifb or gennif, genniti or gennit,
&c., were the accepted spelling. The different stages are illustrated in
the following . (1) ' Da yw genny/ heb y lunet nat achaws itt y hynny. '—Mab.
177. Chwedleu o lys arthur gennyt heb yr iarlles.' Mab. 178. Pette gennyf y
seibiant a'r hamdden sydd gan lawer.' Def. Ffydd (1595). 26 WELSH PHONOLOGY.
The change of ' uw ' into ' u,' as
uwch——uchder, is rather a case of reversion to a more radical form. The w '
in uwch, awch (Mediaeval, = your), and several other monosyllables, is
intrusive or inorganic: ch is articulated by bringing that part of the
tongue, which is raised for the sound ' w,' a little nearer to the soft
palate so as to cause an audible friction when breath is exhaled. Hence in a
leisurely and careless pronunciation of a vowel—other than ' w '—before ch*,
the tongue in taking up the consonantal position glides through the ' w '
position, and in so doing articulates more or less audibly the ' w ' which is
written in the above words. When another syllable is added, a shorter,
brisker, and consequently more precise articulation of the first syllable is
imperative, with the consequent elimination of the ' w. The ' w ' is of
somewhat later growth in uwch than in buwch, Iluwch, In the Mabinogion the
comparative is regularly spelt uch, as uch penn y pwll.' 216 (cf. 153, 175,
&c.). This is due to the influence of the positive uchel where of course
' w ' does not occur. Still, üwch is met with as in Ac yn dyvot yn ogyfuwch
ar orsed. '—Mab. 8: and in the poetry of Dafydd ab Gwilym (d. 1400) it is
quite comlnot) Nid gwen gwelwdon anghyfuwch, Nid gwyn ewyn Ilyn, na Iluwch.
'—xxix. * ch generally lengthens the preceding vowel in monosyllables.
|
|
|
|
|
(delwedd F6383) (tudalen 024)
|
24 WELSH PHONOLOGY.
(2) ' Pe profem ni (ddywedyd) gair
wrthit ti a fydde blin gennit ti ? ' Bishop Morgan—Lly/r Job, p. 9 (1588). Da
genni/ ddwyn y ddeilien lås a newydd da i'r rhai a achubir.'—Lly/r y Tri
Aderyn, p. 158 (1653). Mi fynnwn wybod genniti pa beth yw dirgelwch Arch
Noah, '—Ibid. p. 158. Drych yspio anmgen nac oedd genni ar y mynydd. ' Bardd
Cwsc, p. 9 (1704). (3) ' Nid oes genny/' ddilll i ddywedyd chwaneg. Drych y Pri/ Oesoedd, p. 6a (1740).
Thus it will be seen that the mutation or affection of u
(or y) into i in Pronominal Prepositions began to come into vogue in the
latter part of the sixteenth century, and held its ground until the early part
of the eighteenth* I V. The vowel u or y of the next syllable. a e as
nant—nentydd tant—tennyn gwnaeth—gwneuthtllil maes—meusydd All the changes
detailed under C are cases of assimilation—an active principle in all
languages. Thus in the English word many, it is evident that the change from
a to i or y is considerable, and before we have left the former we try to
adjust our organs of speech for the pronunciation of the coming i. This
reduces a to e or a diphthong intermediate between the original a and the i
of the next syllable. The symbol a continues to be written in many, but its
sound is that of the Welsh
|
|
|
|
|
(delwedd F6384) (tudalen 025)
|
WELSH PHONOLOGY.
25 open ' e,' and generally the
spelling is changed to represent the sound, as (1) English men fronm man,
through the intermediate stage mani, mannie (2) Welsh ceni from canaf.
Similarly while Brythonic vindos has becolne Welsh gwyn, the feminine vinda,
through the partial assimilation of the i to the a, has become Welsh gwen.
All the changes given in the above classification are not without exception.
Adwaen and gadael are as common as adwen and gadel, if not more so, while the
change into y in gwrendy, &c., is exceptional. Again the mutation of aw
into o, as in pawl—polion, does not take place in mawrion ( < mawr),
llawnion ( < llawn). Byddai is rarely written bydde to-day, Lost final ' a
' changes ' y ' into ' e,' but it leaves the diphthong ' wy ' unaffected.
Thus while the feminine of gwyn is gwen, the diphthong in mwyn and tywyll is
the same in both genders. Tywell, sometimes met with, is due to the
lilistaken notion that the ' w ' is consonantal, e.g. V nos dywell yn distewi,—caddug V n
cuddio Eryri, Yr haul yng ngwely'r heli, A'r Iloer yn ariannu'r
Ili'. '—Gwallter Mechaill. Silllilarly while ' y ' changes ' a ' into ' e '
as in nentydd, ' y ' in the diphthong ' wy ' is eclipsed by ' w,' and loses
its assimilating power: hence anwyl, arwydd, not enwyl, erwydd.
|
|
|
|
|
(delwedd F6385) (tudalen 026)
|
26 WELSH PHONOLOGY.
The change of ' uw ' into ' u,' as
uwch——uchder, is rather a case of reversion to a more radical form. The w '
in uwch, awch (Mediaeval, = your), and several other monosyllables, is
intrusive or inorganic: ch is articulated by bringing that part of the
tongue, which is raised for the sound ' w,' a little nearer to the soft
palate so as to cause an audible friction when breath is exhaled. Hence in a
leisurely and careless pronunciation of a vowel—other than ' w '—before ch*,
the tongue in taking up the consonantal position glides through the ' w '
position, and in so doing articulates more or less audibly the ' w ' which is
written in the above words. When another syllable is added, a shorter,
brisker, and consequently more precise articulation of the first syllable is
imperative, with the consequent elimination of the ' w. The ' w ' is of
somewhat later growth in uwch than in buwch, Iluwch, In the Mabinogion the
comparative is regularly spelt uch, as uch penn y pwll.' 216 (cf. 153, 175,
&c.). This is due to the influence of the positive uchel where of course
' w ' does not occur. Still, üwch is met with as in Ac yn dyvot yn ogyfuwch
ar orsed. '—Mab. 8: and in the poetry of Dafydd ab Gwilym (d. 1400) it is
quite comlnot) Nid gwen gwelwdon anghyfuwch, Nid gwyn ewyn Ilyn, na Iluwch.
'—xxix. * ch generally lengthens the preceding vowel in monosyllables.
|
|
|
|
|
(delwedd F6386) (tudalen 027)
|
WELSH PHONOLOGY.
AE AND A1. 27 Although in this essay
we are only indirectly concerned with the representation of sounds in
writing, a short digression may be made to refer to the notations ai and ae.
In Mediaeval Welsh our modern diphthong ai was written ei (ey), and the
sequence ai (or ay—for in Mediaeval writings y is often used with the value
of modern i) was then comparatively rare. A diphthong beginning in a followed
by a palatal vowel was regularly written ae: thus our modern a'i (and his)
was ae, and even a + consonantal i had the same form, as in daeoni. The use
of e for the palatal vowel after a, where we now write ai, long continued a
common feature of Welsh orthography. Hence the ai of English words passed
into ae in most Welsh derivatives, as < fray. ffrae maeden < maiden.
paent < paint. < plain. plaen trafaelio < Middle English or Anglo
French travail, whence Modern English travel. Vspaen < Spain. On the other
hand the verb-noun termination of these derivatives is -io, -o, the very
ending we should expect if the diphthong were ' ai ' as in the English
original, as, gwaith disglair * gweithio. disgleirio. * Cyfiaith—-cyfieithu
is an exception.
|
|
|
|
|
(delwedd F6387) (tudalen 028)
|
28 WELSH PHONOLOGY.
Words containing ' ae ' in their final
syllable regularly form their verb-nouns in -u, as arfaeth arfaethu. But for
these borrowed words we have ffraeo, paentio, trafaelio: 'A gymnmerech i
Fardd i'ch plith sy' n chwennych trafaelio ? ' Bardd Cwsc, 6. Still,
trafaelio seems dialectal, for in Dafvdd ab Gwilym the form trafaelu also
occurs . Tra fu'n trafaelu trwy fodd, Trwy foliant y trafaeliodd.' Cywydd
iv., and so regularly in Demetian to-day. THE VALUES OF ' V. Something too
should be said on this question: while the power of every other vocalic
symbol has been already sufficiently indicated, ' y is different. Besides the
neutral a sound, which is always short, and inevitably so, it has the clear
sound of ' u ' in (a) Monosyllables: e.g., dyn, Ilym, ty. Exceptions . it has
the primary or neutral a value in proclitics, that is, words that have no
accent of their own, but are for this purpose read with the following word.
The most common proclitics are ' fy • dy: Y," yr (Definite Article,
Relative Pronoun. Adverb, and Conjunction) .
|
|
|
|
|
(delwedd F6388) (tudalen 029)
|
WELSH PHONOLOGY.
29 (b) In the last syllable of words
of more than one syllable, as gelyn, plentyn. (c) In any syllable when
followed by a vowel, e.g. hyawdledd, gwelyau. (d) In the diphthong ' WY,'
e.g. hwyl, mwynhad. (e) Frequently when preceded by consonantal ' w,' e.g.
gwystlon, wynebau. (f) Generally in the prefix cyd-, and in the first
elements of compounds, if monosyllabic, e.g., cydweithio, Rhydychen,
byrfyfyr, Tyhén, brysneges. N.B.—' y ' has sometimes the value of Welsh ' i,'
as in megys, tebyg, heddyw. VOWELS IN DIALECTS. The simplification of
diphthongs has proceeded much further in dialects than is recognised in the
Welsh literature of the present (lay. It is hardly necessary to add that the
dialects differ considerably among themselves in that respect. This chapter
may be appropriately closed with a few examples of dialectal changes and
vowel values. E ' in final unaccented syllables is pronounced like a ' in the
North West portion of Gwynedd—i.e. in Anglesey, Carnarvonshire, and part of Denbighshire—
and often in Gwentian, as bach(g)an for bachgen. -Au ' and ' -eu ' final have
the value of a ' ill North West Gwynedd and in Gwentian, as— petha, gola.
|
|
|
|
|
(delwedd F6389) (tudalen 030)
|
30 WELSH PHONOLOGY.
e ' in the rest of Wales, as— pethe,
gole. 'Ae ' and ' ai ' in final unaccented syllables are pronounced as a ' in
North West Gwynedd and in Gwentian, (i.) as--gadal, perffath, (ii.) ' e ' in
the rest of Wales, as gadel, perffeth. 'Ae ' and ' oe ' in monosyllables are
simplified into å ' and ' 6 ' respectively in Demetian and Gwentian, as— man,
c6s, for maen, coes. 'Ai ' in sollle monosyllables—generally those ending is
pronounced ' ae ' in Demetian, as in 1'— gwaer, taer, for gwair, tair, also
Caen for Cain. 'Au ' in monosyllables is generally pronounced in Demetian and
Gwentian, as— doi, hoil, coi, oir, for dau, haul, cau, aur. oi ' U ' and the
clear ' y ' are generally pronounced ' i in all parts of Wales outside North
West Gwynedd, as— cani for canu, mid for mud, hin for hyn, hin for 11911 or
hun.
|
|
|
|
|
(delwedd F6390) (tudalen 031)
|
WELSH PHONOLOGY.
31 CHAPTER 11. CONSONANTS, B AND M.
The relation of these two sounds is intimate. The passage is closed at the
lips. If then the lips are separated to allow the breath to escape, and the
vocal chords are at the same time vibrated, b is articulated. But the lips
may be kept together, and the breath passed out through the nostrils, giving
rise to the sound m. Hence the frequent interchange of m and b, as W. blith
and Eng. milk. mieri and Eng. briar. But in Welsh this is not the only
reason. At least two other contributory causes of considerable importance
exist (1) The soft mutation of both m and b is f, and as the softened initial
is a very common feature in Welsh construction, a word may be more familiar
in that state than in its radical form. Hence a not uncommon re-
|
|
|
|
|
(delwedd F6391) (tudalen 032)
|
32 WELSH PHONOLOGY.
version to the wrong original. Thus
maban (diminutive of ' mab ') commonly occurs under the form ' faban,' as '
ei faban,' ' dy faban,' ' dyma faban tlws.' In Modern Welsh faban has been
referred back to a radical baban ' so often that this doublet has gained a
firm footing in the language. In this particular instance no doubt the
English word ' baby ' materially helped the growth of the doublet. But no
such outside influence can account for ' modrwy,' which is from ' bodrwy
(< bawd + a termination meaning band, cf. aerwy) through the intermediate
' fodrwy.' Fronl early times up to the seventeenth century it was quite
customary to wear rings on the thumbs: but as the custom changed and as rings
were worn on other fingers as well, the origin of ' fodrwy ' was obscured.
Hence its being accidentally referred back to a coined ' modrwy,' which has
now supplanted the more correct form. For a similar change of custom we need
only refer to the English thimble, which is no longer worn on the thumb
except occasionally by sailors in repairing the tough fabric of their sails.
It is clear that the etymological meaning of modrwy could not be known to the
translator of Hosea ii. 13.: Mi a ymwelaf å hi am ddyddiau Baalim, yn y rhai
y gwisgodd ei chlustfodrwyatl a'i thlysau.' A still more curious instance is
' bodo,' the familiar name for aunt in North Cardiganshire. In the light of
this interchange of b- and m-, its origin is at once made it is a shortened
form of modryb with the dimiclear: nutive suffix -o, just as -ie is added in
English to make
|
|
|
|
|
(delwedd F6392) (tudalen 033)
|
WELSH PHONOLOGY.
33
auntie. The shortening of words in this way is a fact
of daily occurrence. Hence Ned for Edward, with the
-n of the 1st sing. possessive adjective prefixed, so Nel
from -n Ellen. -O is a familiar SUffx used to denote
smallness or endearment: Gweno, Deio, Bilo ( Wil-
liam), 1010
Iorwerth), and others will occur to the
reader.
(2) The tendency to refer borrowed words in /- to
radical forms in m- or b- has given endless scope for
diversity of treatment, for while one speaker will refer
volet ' (a gauze veil worn by ladies in the middle ages)
to a radical boled, another may with equal justice con-
sider moled as the correct radical form.
This same keenness for provection accounts for m- in
mal < fal (whence fel) < y fal < hafal, cognate with
Lat. similis. This word is often ' bal ' in Gwentian.
In addition to those already mentioned, the following
are some of the most interesting instances of the changes
here described
bargod, as compared with Latin margo, English
margin
bainc and mainc < A.S. benc, whence English
bench
banon and manon
balaen, balain, balen ( < Milan), a steel blade for
the manufacture of which Milan was celebrated
in the Middle Ages '
bodrydaf and modrydaf '
borddwyd and morddwyd •
bore and English morrow, morning ,
3
|
|
|
|
|
(delwedd F6393) (tudalen 034)
|
34
WELSH PHONOLOGY.
bwysel and mwysel < English bushel
byswynog and myswynog ,
baeddu and maeddu •
beiddio and meiddio:
ben and men (a wagon)
benyw and menyw
bilain and milain ( < Eng. villain)
bignen and mignen, and Eng. bog
bigwrn and migwrn
bwyaid and mwyaid
bwytal and bitail, from Mid. Eng. or Anglo French
vitaille (whenee Mod Eng. victuals), food
Dyvot a oruc gwyr iwerdon hyt att Arthur a rodi
bwyttal idaw. '---Mab. 136.
bydwraig is a half adapted and half translated
form of English mid-wife: ' mid ' > ' bid ' (and
then by popular etymology ' byd,' though the
term makes nonsense) + ' gwraig,' translation
of ' wife.' The English word ' bracelet ' was sub-
mitted to the same piecemeal treatment .
breich ' ( braich) is a correct translation of
brace, which is no other than the French ' bras,'
an arm, and -led is the English -let with the regu-
lar softening of final t after a vowel. The popular
leather) in
etymologist saw the word ' Iledr ' (
the StifflX, and Lewys Glyn Cothi has the rather
amusing couplet*
Gwisgaw breichledr, os medraf,
O arian neu aur a wnaf.'
* See Silvan Evans's Dictionary under Breichledr. '
|
|
|
|
|
(delwedd F6394) (tudalen 035)
|
WELSH PHONOLOGY.
35
Dr. Silvan Evans (v. breichled) is disposed to follow
this fanciful etymology.
burgyn and Eng. morkin
melfed < Eng. velvet:
V tal dan y melfed du
A gae wirion dy garu. '—Bedo Rrwynllys.
miswrn < Eng. visor:
P wy nid yw'n canfod Rhufain .
yr hon a
beintiessid gynt å Iliwieu hyfryd, eithr yr owr'on gan
dynny ei miswrn, y mae'n haws yr olwg, ag yn llai'r
bris arni ' Def. Ffydd, 188.
mwydyn < bwydyn < y bwydyn < abwydyn
mach, meichiai, is cognate with Latin vas, a surety.
c.
c is always hard as in English cat.
Initial c + consonantal w does not occur in native
Welsh words. Cweryl, cwestiwn, cwarel, are all bor-
rowed, and are as yet unnaturalised, and cwinc, a finch,
is dialectal.
In the following couplet from Edmwnd Prys—
I rannu hon ar onest,
Ni cheir cydwybod na chwest,'
the last word in its radical form may be either chwest
or cwest.
It is clear that the genius of the language rejects this
sequence. In fact Aryan qu, that is q velar after which
|
|
|
|
|
(delwedd F6395) (tudalen 036)
|
36
WELSH PHONOLOGY.
a slight consonantal w was regularly developed, passed
into p in Brythonic. Compare Welsh pedwar, pump,
with Latin quattuor, quinque (see Sir John Rhys's Welsh
Philology). Words in qu borrowed into Welsh as they
become naturalised change q into ch, as
chwil and chwart, from English quill and quart.
C and the corresponding voiced sound g have two
values according to the part of the mouth concerned in
the articulation.
If the stoppage is between the ridge of the tongue
and
(i.) The back or soft palate, the sound is back or
velar, as in cwys, cwm: gwas, gordd.
(ii.) The front or hard palate, the sound is front or
palatal, as in cil, cig: gilydd, gerwin.
In Montgomeryshire and along the Merioneth coast
an extreme palatal value is given to these mutes before
a, with a consequent growth of a parasite i after the
thus in the above-named districts caws,
consonant:
cath, and gardd are pronounced ciaws, ciath, and giardd.
The same peculiarity is found in Gwentian, where for
example cant is pronounced ciant.
C AND T.
Now, if the stoppage is made by the tip instead of the
adjoining blade of the tongue, the sound produced is t
and not palatal c. The physiological difference is so
slight that one can easily understand why palatal c
and t pass so readily into each other. Indeed there is
|
|
|
|
|
(delwedd F6396) (tudalen 037)
|
WELSH PHONOLOGY.
37
abundant evidence that the transition from velar ' c
(' g ' or ' ch ') to the labio-dental ' t' (' d ' or ' th ') is
easy and of common occurrence. Thus in Latin we find
the doublets
juvencus and juventus:
novitius and novicius ,
while in Late Latin verbs the endings -itare and -icare
were frequently interchanged.
So, older English apricock is now apricot, though
probably here the change of ' c—c ' into ' c—t ' is due
in part to dissimilation.
In Welsh poetry -od and -og are sanctioned as correct
rhyme, e.g.,
Lle cyrch iyrchod, rywiog ryw,
Lle can edn, Ile cain ydyw. '—D. ab G. xix.
Interchange of—T (or D) and C:
D (or DD)
1. T (or D) AND C.
dyllhuan, tyllhuan and cylluan, cyrlluan ,
dyrchafael and cyrchafael:
Ascension Thursday is ' Duw leu Kyrchauel '
Brut y Tywysogion.'
ieuanc, ieuant, and iewaint:
'A'r cloyn a går yr ieuaint. '—Llywarch Hen.
in
(v. Pughe and Pryse's Dictionary, s.v. ieuant).
ysgol ( < Latin scala) and ystol.
|
|
|
|
|
(delwedd F6397) (tudalen 038)
|
38
WELSH PHONOLOGY.
chwilcath (vb. and noun) and chwiltath (verb).
Gwar gwrcath, gwydn chwilcath chwai.'
D. ab G. cexxx.
Pa beth sydd yna'n chwiltath ?
Tyngais i'm cyffais mai cath. ab G. clviii.
tywarch and cywarch.
For ' Glynn Cywarch ' see Bardd Cwsc, 54.
Tyddyn Cywarch ' is a farm near Llanfechell in
Anglesey.
The late Dr. Silvan Evans in his monumental dic-
tionary (see under Cywarch) enquires—
Can it be that cywarch, cywarchan, cywarchen
have originated in misreading c and t (tywarch, tywarch-
an, tywarchen), which are very like in old Mss. ?
While probably this suggested explanation has some
foundation in fact, it should be borne in mind that
the interchange of t and c is not confined to Welsh or
modern languages. Moreover, no such misreading
would account for the interchange of d and g (see below).
Ilosgwrn and Ilostwrn:
compare Ilostlydan, a beaver, and Cornish lost a tail
tlws and clws:
tlawd and clawd .
Clawd is the general pronunciation in parts of Gwy-
nedd (as in 141911) and also in Gwentian. Tl in the same
syllable involves a peculiar change of direction in the
emission of breath, and great mobility or muscular
activity in the tip of the tongue. The breath forces its
|
|
|
|
|
(delwedd F6398) (tudalen 039)
|
WELSH PHONOLOGY.
way out over the tip (t), which is then immediately
replaced, the air being passed laterally over the sides of
the tongue (l). The closely related sequence cl is much
easier, for after the stoppage between the back or
middle of the tongue and the roof of the mouth is forced
open (c), it is not difficult to bring the tip into contact
with the gums of the upper teeth so as to divert the
escaping breath over the sides
The change of tl (or dl) into cl is not peculiar to Welsh.
Latin very early turned medial -tl- into -cl- as in
Periclum, and probably English clever is from Middle
English deliver (
quick, active).
TH AND CH.
brith and brvch •
brithyll and (Gwentian) bryehyll •
dethe (Demetian) and decha (Gwentian) deheuig;
blith as compared with Eng. milk, Old Irish melg,
(For interchange of b and m see above).
3. D (OR T) AND G (OR C).
bwrdais and bwrgais, from Eng. burgess ,
eyfod alld cyfog:
Ae yna Ylngyuoe o bawp ar hyt y ty.'
'And then all in the house arose. '—Mab. 39.
Cyfog has now a specialised meaning.
cywyll (culture, tillage) and diwyll, diwyllio:
dillwng (to liberate, set free) and gollwng:
deor and gori
dweyd and colloquial gweyd •
pioden and piogen:
|
|
|
|
|
(delwedd F6399) (tudalen 040)
|
40
WELSH PHONOLOGY.
Piogen occurs in Llyfr y Tri Aderyn, p. 229. There
is a Tyddyn Piogen in the parish of Holyhead, Piod is
common.
sudd and sug: see Lly/r y Tri Aderyn, p. 263.
Welsh softens final -c after a vowel into -g, except in
a few borrowed words which have not been completely
naturalised. Thus
Mediaeval rhac is now rhag, and English catholic is
catholig.
The preservation of ' ac,' and, and ' nac,' no, neither,
nor, side by side with ' ag,' ' nag,' is due to a conscious
effort to differentiate forms on account of difference of
function. This very desirable distinction is scrupu-
lously observed in present-day writings, but in works
of the 16th, 17th, and 18th centuries ' ag ' and ' nag
are frequently used for ac and nac. Thus in De,i/ynniad
Ffydd Eglwys Loegr by M. (1595),
ag ' is the
regular form used for and before vowels. Again Edward
Morus in Cywydd v Llwon Ofer writes
Buddiol, gweddol ag addas,
O bur gred i beri gras.
CH is a guttural spirant bearing the same relation to
c as Welsh f/ does to and th to t. It is somewhat
harder than the guttural in Scotch loch, Irish lough,
German nacht, and Old English liht. It may be
|
|
|
|
|
(delwedd F6400) (tudalen 041)
|
WELSH PHONOLOGY.
41
(1) an immutable radical as chwi, chwant, or
(2) a spirant form of c as tri chant.
As a radical initial it is invariably followed by w,
which is in some words nothing but a parasitic sound
heard as the tongue moves away from the ch position.
Further, this w is always consonantal except in the
word chwydd* (a swelling) and its derivatives. Com-
pare the consonantal value of u after q in English.
Indeed the Welsh initial chw- differs in phonetic value
from English qu only in that q is a guttural stop, while
ch is a guttural spirant. And as cw- is foreign to Welsh
(v. under C above) it is not strange that English qu-
words pass into Welsh under the form chw-.
For the same reason provected gw (consonantal)
passes into chw, as—
chware from gware
chwaen from gwaen ,
chwedi from gwedi •
chwysigen from gwysigen.
Now ch will readily pass into h if the back of the
tongue be not held so near the back of the palate as to
produce the friction heard in the former sound. And
that is exactly what takes place regularly in the radical
initial group chw- in Demetian. There chwant is pro-
nounced hwant, and chwaer is hwar. Chwi and its
derivative chwithau are exceptions, and retain even
in that dialect the full value of the ch, though it may be
mentioned in passing that the w is usually dropped, the
words being pronounced chi and chithe. The com-
* Consonantal w in Demetian.
|
|
|
|
|
(delwedd F6401) (tudalen 042)
|
42
WELSH PHONOLOGY.
pound chwychwi is not heard in that part of Wales
except in the very corrupt form ' y chi.'
Instances of hw- (or rather wh-, see below) occur in
Mediaeval works written in Demetian .
Gossot v deulu ar whech milltir o wastattir.'
Ystorya, 11.
Duw a talo itt vyg whaer heb y peredur. '—Mab. 220.
The same characteristic marks Gwentian, except that
in that dialect a further change is often made whereby
the spirant disappears altogether, thus—' whar ' or
war ' (the ' a ' having the same value as in English glad,
man), ' whech ' or ' wech.'
A similar change occurs in late Cornish and in Breton.
But why the spelling wh and not hw in the above
instances ? Etymologically hw would be more correct,
while phonetically the digraph stands for a simple
sound, and the aspirate comes neither before nor after
the w. A simple symbol would more accurately repre-
sent the sound, or, failing that, a mark of aspiration
over the w as in Greek. The reason for the changed
order is two-fold
(1) The analogy of ph, rh, and the
(2) An idea prevalent in Mediaeval times that the
digraph represented two distinct sounds, and that the
aspirate followed the w. A similar mistake accounts
for the change of Old English ' hwa ' into Middle Eng-
lish ' who.'
Just as chw- in South Welsh passes into wh, the re-
verse process may be and is resorted to in order to give
a word a more distinctly Welsh sound and appearance.
|
|
|
|
|
(delwedd F6402) (tudalen 043)
|
WELSH
PHONOLOGY.
Thus English words in wh- are borrowed into Welsh
under the form chw-, as—
chwip from Eng. whip
ehwirligwgan from Eng. whirligig ,
ehwisgi from whiskey.
It is a natural corollary that there are no words in
Standard Welsh to-day beginning in hw or wh. In the
few words in hw-, tv is invariably a vowel, as—
hwy, hwyl.
No words outside the dialects begin in wh.
Any enquiry then into the origin of radical ch- and
chw- Illust take note of the Illany avenues along which
words containing these sounds have come to their pre-
sent form.
Thus chwysigen is borrowed from Latin vesica,
through the intermediate form gwysigen, where the g
itself is prosthetic, and therefore not an organic part of
the word. In Chwefror the chw is from the s of mis and
the / of Latin Fcbruarius, the s passing into ch through
the intermediate h. Similarly chwech goes back to a
hypothetical Aryan sueks, a doublet of seks which is the
parent of English six and Latin sex. Hence W. chwech,
Eng. six, and Latin sex are all cognate. As in the case
of h (see below), radical ch may often be referred to
Aryan s. (For secondary ch see chapter on Mutations.)
Ch in chwip is from the inorganic h in English whip.
It. was stated above that Old English had a guttural
spirant which was represented in writing by h, as—
liht ' (modern light), ' broht' (
brought) ,
neah
nigh), ' heah ' ( = high), ' hreoh ' ' ruh ' (
rough) ,
hleahhan
to laugh), ' hoh hough).
|
|
|
|
|
(delwedd F6403) (tudalen 044)
|
44
WELSH PHONOLOGY,
The Mediaeval scribe, who sought to spell the English
of his day phonetically, represented the guttural spirant
by the nearest digraph at his command, that is by gh,
just as successfully as the Welsh scribe who adopted ch
for the parallel but more surd Welsh sound. Of course,
the limited number of symbols in either alphabet made
a more accurate representation impossible, unless the
bold (and better) step were taken of coining or borrow-
ing new symbols for these individual sounds. It is true
that in neither language did the digraphs have the
separate values of their component parts: that is, gh
in ' brought ' at no time in its history had the value of
g + h ' in ' big horn,' nor was ch in Welsh chwi in-
tended to represent the distinct sounds heard in ' ac
hefyd.' But just as the Greek aspirated t, p, c—that is,
theta, phi, and chi
in the post classical period became
continuants, so had English ' p-h ' and t-h ' passed
into the spirants or continuants ph (
f) and th ( =th
in thin, this). Hence Mediaeval orthographists in adopt-
ing gh (in English) and ch (in Welsh) as symbols for the
guttural continuants in those languages were only
extending a principle already recognised in the case of
ph and th.
The reader will not fail to notice another instructive
parallel between English and Welsh in the development
of the parasite u in Inost instances before the guttural.
As in the case of the tv in Welsh awch (Y'our), uwch
(higher), &c., the guttural vowel sound is somewhat
inevitably produced as the tongue takes up its position
for the articulation of the guttural spirant. This is
especially the case after a back or low vowel like ' o
or ' a.
|
|
|
|
|
(delwedd F6404) (tudalen 045)
|
Further, just as in South Wales ch in ' chw ' has allnost
everywhere passed into h, so even in Old English h at
the beginning of a syllable was a simple spirant as in
Modern he, has. Hence while the second h in Old Eng.
heah passed into Middle English gh, the initial h remained.
In Modern English the guttural ' gh ' sound has dis-
appeared. The muscular effort involved in its articula-
tion was found too severe, and either the sound was
dropped with compensatory lengthening of the vowel as
in though, high, Vaughan (< W. Fychan), or it was
changed into the more easily produced labio-dental
spirant f as in laugh, rough, Gough ( < W. Goch) .
No such difficulty has been experienced in Welsh
except initially as mentioned above—and the inter-
change of ch and ff is not common. Dichlais and difflais,
safe, secure, may be an illustration, but even if they are,
? di,
the change is probably from the f/ of difflais (
not + flexus, bending) to the ch of dichlais.
D.
D is a dental sonant of which the corresponding spi-
rant is dd ( = th in English then), just as th is the corres-
ponding spirant to the dental surd t. The inter-relation
WELSH PHONOLOGY.
t.
45
therefore of these
sounds may be expressed by the
formula
or again
D-t >
dd
th
th.
st > s:
|
|
|
|
|
(delwedd F6405) (tudalen 046)
|
46
WELSH PHONOLOGY.
The change of ' d—t ' into ' s ' in ' llas
a form once
common for the aorist Impersonal ' lladdwyd,' ' lladded '
is a striking case of dissimilation. In early Welsh
d—t,' in lladt, a short doublet of lladded, lladd
wyd, came too near together to be distinctly and
separately articulated: the mobility of the tip of the
tongue was not equal to the task. Hence an attempt
was made to articulate the ' d ' by bringing the adjoin-
ing blade of the tongue into contact with the palate,
leaving the tip free for the immediately succeeding ' t.'
But that division of labour of necessity yielded ' s t,'
and eventually the s assimilated or eclipsed the t:
llad-t > llast > llas.
The same result is familiar to us in the English ' Inust
and ' wist,' and in Latin ' est,' he eats. ' Must ' is from
A.S. ' moste
(for mot-te) the preterite of ' mot,' and
wist ' ( < wit-te) is the preterite of ' wit,' still used in
to wit.' ' Est ' is for ' ed-t ' a doublet of edit from edo.
The same process of dissimilation will probably ex-
plain ' Wstrws,' the nan-me of a large house near Capel
Cynon, on the road between Llandyssul and New Quay,
in Cardiganshire. The origin of the natne has been the
subj ect of much discussion at different times.
It
seems likely to be nothing but a modified form of
Wyth-drws,' just as ' wythnos ' is pronounced ' Wsnos
in Anglesey and other parts of Wales. For such a name
as ' Wyth-drws ' compare ' Saith-aelwyd,' a farm by
Llangefni.
Di > Dzh
Di,' when followed by a vowel, tends to acquire the
|
|
|
|
|
(delwedd F6406) (tudalen 047)
|
WELSH PHONOLOGY.
47
dzh ' sound heard in English judge, gin. In formal
speech this change is not countenanced, due partly to
the fiction that Welsh words are spelt phonetically, and
that therefore per contra they should be pronounced as
they are spelt. But colloquially diogel and diofal are
approximately dzhogel and dzhofal, while only a minc-
ing pronunciation of ' diawl ' in colloquial speech would
differentiate the sound from that in ' dzhawl.' Com-
pare the pronunciation of t in English nature and similar
words.
Dd is the soft mutation of d, and is never met with as
a radical initial in Welsh (cf. chapter on Initial Muta-
tions).
For inorganic dd and lost dd see sections below de-
voted to excrescent and lost sounds.
Final dd after a vowel is apt to be hardened into d,
thus, while gormodd has literary sanction, e.g.,
Marw mab mam, mawr ymhob modd,
Mair a Gannon ! marw gormodd. '—Tudur Aled,
G.B.C., 230.
and is still used colloquially in Gwentian, gormod is
the recognised literary form to-day. Similarly machlud
(from ym + Latin occludo) is for an older form ymach-
ludd, while ansawdd has been kept by the side of
ansawd, but with differentiation of meaning.
N tends to harden dd into d, as bendith from benddith.
F and Ff.
Both are labio-dental spirants, f being voiced like
English v, or f in of, and the corresponding voiceless
sound like f in English for.
|
|
|
|
|
(delwedd F6407) (tudalen 048)
|
48
WELSH PHONOLOGY.
Initial F- does not occur in the radical or dictionary
form of a word, except
(1) by Aphaeresis, as ' fel ' < ' fal ' < hafal.
(2) by crystallization of a mutated form as the radical,
as TV, /y. In many instances this is due to the habitual
use of the definite article ' y ' with the word, as (y)
fagddu, (y)fory.
(3) In a few borrowed words which have resisted
provection into m- or b-, due to the influence of the
familiar original spelling, as finegr, fernlilion.
DD AND F.
The interchange of the voiced spirants / and dd is
common, due to the ease with which the vocal organs
can pass between the two positions concerned in their
articulation, If the tongue be dropped from the dd-
position, the lower lip is readily brought into contact
with the upper teeth yielding f, while again the
dropping of the lip is often accompanied by the raising
of the front and tip of the tongue. Examples .
Afanc and addanc:
Ac wynteu a dywedassant bot adanc mywn gogof. '
Mab. 224.
Balwyf and balwydd .
Balwydd is probably to a great extent due to a
fancied connection with gwjdd.
Cufigl and cuddigl.
|
|
|
|
|
(delwedd F6408) (tudalen 049)
|
WELSH PHONOLOGY.
Caerdyf and Caerdydd
49
Etymologically the word denotes the caer on the
river Taff (Welsh Tåf), though there is a local tradition
that it denotes the caer of Aulus Didius.
A'm bod er's talm, salm Selyf,
V n caru dyn uwch Caerdyf.'-—D. ab G. Il. (p. 3).
Eifionydd and Eiddionydd
eddain
edryd
efain
edryf •
ach ac edryd (or edryf),' stock and lineage.
godwrdd, dadwrdd, godwrf, tarfu, cynnwrf
gwyryf, gwyrydd, and adwerydd
gwyddon and gwyfon
hwyfell
hwyddell
lladd and cyflafan:
llawryf and llawrwydd.
For llawrwydd compare balwydd above.
nwyddau and nwyfau •
plwyf and plwydd.
We may also compare rhudd and English ruddy with
Rufus and rubric.
The corresponding voiceless th and are not nearly
as often interchanged, for the muscular tension required
for their articulation in the vocal organs is unfavourable
to the necessary fluidity. Still examples do occur:
The 3rd Sing. Pres. Indic. termination of verbs is
-ith in Demetian, -iff in Gwentian, as rhedith,
rhediff, runs ,
benthyg is a corruption of benffyg
4
|
|
|
|
|
(delwedd F6409) (tudalen 050)
|
50
WELSH PHONOLOGY.
and pencnath (N. Wales), penneth (S. W.) are common
pronunciations of English Penknife. Similarly English
children sometimes say ' nuffink ' for ' nothing,' and
some always pronounce ' three ' as ' free.
DD AND TH.
Dd in an energetic articulation may easily pass into
th, and on the other hand a languid pronunciation may
reduce th into dd:
arglwydd and arglwyth
colwydd and golwyth
cynysgaeth and cynysgedd
diwaethaf and diweddaf '
(gwen)ith and yd •
Old Ir. ith corn: compare English te'heat — the white
grain.
wmbredd and wlllbreth.
We may conmpare the voiceless ' th ' often heard in
the articulation of ' with ' in English, when special
emphasis is laid upon it: also the emphatic adverb
off ' and the weakened preposition ' of.'
The adjoining ' n ' accounts for the unvoicing or
hardening of the ' dd ' in
(1) ganthaw, gantaw: ganthi, genti (common in the
Mabinogion), from ganddo: ganddi.
(2) Mediaeval ganthunt and gantunt for ganddunt
(—Modern ganddynt, the u > y on the analogy of the
3rd plu. Past Impf. and Plupf. of verbs):
' A diheu oed ganthunt na welsynt eiryoet gwr a march
ac arueu hoffach gantunt eu meint noc wynt. '—Mab,
|
|
|
|
|
(delwedd F6410) (tudalen 051)
|
WELSH PHONOLOGY.
F ( = V), AND OR U.
51
The interchange of the voiced spirant ' f ' and ' w,
whether vocalic or consonantal, is one of the best at-
tested facts in Welsh phonology. In value the ' w ' is a
vocalised f, as in
awrddwl fronm afrddwl
evwoed fronn cvfoed.
Further, the softening of the labials ' 111 ' and ' b
may follow one of two lines:
(1) the lower lip Inay be
brought into gentle contact with the upper teeth, allow-
ing a mild escape of breath. The resulting sound is f
v). Or (2) with equal ease the tension may cease in
the lips, which will then be held loosely together, and,
due to the gentle pressure of the breath in the resonance
chamber behind, they will be pushed forward: the
back of the tongue will at the same time be raised as an
autonaatie accompaniment: the resulting sound is ' w.'
Thus the Welsh cognate of Old Irish ' cumachte,' and
of English ' lilight,' assunnes the doublet forms
cyfoeth and cvwaeth.
Cawn o ddawn a eiddunwyf,
Cyzcyae(hog ac enwog wyf. ab G. iii.
and the ' b ' seen in Old Irish Serb, delb, marb, tarb, &c.,
is ' w ' ill their Welsh cognates chwerw, delw, marw,
and tarw.
The change of ' w' or ' u ' into ' f
is sometimes
fanciful and due to false analogy, as,
yntef for yntau:
Ac yntef oedd yn eistedd yng-hanol y llwch. '—Llyfr
Job, ch. ii.
|
|
|
|
|
(delwedd F6411) (tudalen 052)
|
52
WELSH PHONOLOGY.
and, anghefol for angheuol .
(Fel) y nessao ei enaid i'r clawdd: a'i fywyd i _-loes-
ion] anghefol. '—Llyfr Job, xxxiii.
Other examples of the interchange of ' f,' and w,
u ' are—
archfa for arehwa .
archfa'r bwydydd.' Bardd Cwsc, 23.
berfa, from Old. Eng. berewe:
colloquial bregwast or brecwast, fronl Mod. Eng.
breakfast.
criafol and criawol
cenafon and cenawon
cleddyf and cleddau •
-aw and -ew in final unaccented syllables are some-
times weakened into -au, -eu. Compare cenawon with
ceneu (also cenaw), Ilewych and goleu.
cafod and cawad
Llyma gawat o nywl yn dyuot. '—Mab. 46.
Cafodydd didorr o saethau Inarwol.' Bardd c qté'sc, 47
Diofryd and diowryd
Rho' ddiowryd rhyw ddirwy,
Adeiliad serch er merch mwy. '—D. ab G. clxxix.
diawl and diafol •
Diafol is a modern learned borrowing. The word
occurs repeatedly in Deffynniad Ffydd Eglwys Loegr
(1595), and its form is invariably ' diawl
Yr ydych o'ch tad Diawl, ai ewyllys ef a fynnwch 'i
gyflowni. '
191.
|
|
|
|
|
(delwedd F6412) (tudalen 053)
|
WELSH PHONOLOGY.
dwyfol and dwywol •
dwyfes and dwywes ,
defnydd and deunydd •
53
Os dilyni onestrwydd, nid wyti ond ffwl di-ddeu-
nydd. '—Bardd Cwsc, 38.
goreu and goref, goraf
lau and English Jove
pythefnos and colloquial pythewnos
ysgafn and ysgawn
ysgaw and Esceifiog
tywod and tyfod •
Canys yn awr trymmach fydde na thyfod y m6r. '
Llyfr Job, 14.
Also the dialectal
brifo (in Gwynedd) for briwo, briwio
Ilifo
dyfyd
dwad (Demetian)
towlad
Iliwo, Iliwio .
dywed (he) says
dyfod •
taflod,
G.
Alxvays hard as ill English get, never as in gilt
Initial g- is often prosthetic, as
gallt for allt:
godidog < odid
oonest < English honest ,
•win, gwineu < Latin vinum
gwynt < Latin ventus.
|
|
|
|
|
(delwedd F6413) (tudalen 054)
|
54
WELSH PHONOLOGY.
Why ? The reason is two-fold
(1) Initial g- in soft mutation, after passing through
the intervening stage of a very soft spirant, something
like the ' r ' grasseye of French or the lisped r sometimes
heard from English and Welsh speakers was eventu-
ally dropped, e.g.
gardd—ei ardd •
gwaith—ei waith.
Hence the radical form of ardd and waitli will be
found not under a- and zv-, but under g-. Analogy and
uncertainty have greatly extended the prosthesis of g-,
and quite a host of words which originally began in a
vowel or w have now to be looked for under the gut-
tural. Others have gained currency in dialects, as gaddo
for addo in Gwynedd, and giar for iar in Demetian.
(2) Analogy and uncertainty will not adequately
explain the remarkable fact that hardly any Welsh word
in the radical form begin in consonantal ' w.' Another
principle has been at work. The back of the tongue in
being raised to the ' w ' position is carried a little too
far and touches the back of the palate, with the inevit-
able result that g is articulated. This may be a case of
indistinct articulation, but it seems more likely to be
an illustration of the law whereby initial I and r- in
Welsh are almost invariably proveeted into ll- and rh-
w ' as an initial lacks precision and definiteness, and
the effort at giving it a Inore distinctive character leads
to the muscular tension in the back of the tongue,
which brings it into contact with the back of the palate,
yielding the guttural ' g.'
|
|
|
|
|
(delwedd F6414) (tudalen 055)
|
WELSH PHONOLOGY.
55
The Celts of France introduced the same peculiarity
into French. Hence English war is guerre in French,
and the forms guard, guise, guarantee, &c., are nothing
manner), and war-
but the English words ward, wise (
rant, borrowed by the French who recast them in their
linguistic mould, and returned them to the English in
a new garb. French has gone further, and has lost the
lip element altogether in pronunciation, though the ' u
generally remains in writing. Hence phonetically in
that language initial w sound has passed through gw
into g.
A similar change is found in the prefixes go- and gor-
for older gwo- and gwor- in Welsh. And colloquially
initial gwn- gwl-, gwr- are continually pronounced in
most parts of Wales without the w, as
g'na for gwna, g'lyb for gwlyb, and in Gwentian
gwraig ' is ' graig.'
W ' in initial ' gw-' + vowel, is generally consonantal,
as in
gwyn, gwas, gwaith, gwych
but in a few words it is a vowel if ' v ' follows, as
g€ydcl, g€yl, g€yr.
H.
H: as in hardd, hwn. Never silent as in English
In Mediaeval writings it seems to have been
honest.
in some words either silent or very nearly quiescent.
Thus in ' hun,' sleep, h was a strong spirant as it is
to-day, but in hun, self, it had probably no more value
than in English ' historian.' The grammatical rule
|
|
|
|
|
(delwedd F6415) (tudalen 056)
|
56
WELSH PHONOLOGY.
which requires ' a ' before history, but ' all ' before
historian points to the weakening of the ' h ' in the
latter word. Similarly in Mediaeval Welsh, the phrase
for ' his sleep ' was ' y hun,' but for ' himself ' the pos-
sessive ' y ' was always written ' e,' thus in Maxen
WIedic we find on the same page (85) .
(y) uorwyn e hun, the maiden hersel/
y wreic vwyhaf a garei a welai trwy y hun,
he saw in his sleep the lady he loved most.
As the h in hun, self (hun < un, one) was of little
value, it was necessary for distinctness to differentiate
the sound of the possessive ' y ' more from that of ' u '
in hun.
In origin, h is frequently
(1) from s, as in hun, sleep, cf. Latin somnus, or
(2) accentual, as in cenhedloedd, plu. of cenedl.
In ' hun,' self, after the feminine possessive e, it may
be regarded as the Welsh modified value of the s
originally ending the Feminine Sing. Possessive. But
the normal h in this word should rather be considered
as a somewhat quiescent intervocalic h* , arising from
the fact that the word is regularly used with the Pos-
sessives ' vy,' ' dy,' ' y,' &c.
The fact that ninnau was written both as nillneu and
ninheu in Mediaeval literature points to the same in-
audible h.
* For instances of this h see Sir John Rhys's Welsh Philology,
P, 232.
|
|
|
|
|
(delwedd F6416) (tudalen 057)
|
WELSH PHONOLOGY.
57
THE HARDENING of b, d, g into p, t, c in the comparison
of Adjectives, and of b, d, g, f into p, t, c, ff in the Present
Subjunctive of Verbs and in Verb-nouns is, in most
cases, due to an accentual h: thus the Noun bwyd
yields the Verb-noun bwydha', which, with the shifting
of the accent to the first syllable, became ' bwyta,' so—
ceiniog—ceiniog-ha'—ceinioc'a ,
cardod—cardod-ha'—cardot'a
Cwpla is the familiar pronunciation in Gwentian. In
the same dialect one meets cryffa, ' to grow strong,'
from cryf-ha' from cryf.
llaw > Ilof-ha' > Iloff'a,
So in the Verbs
gwybod > gwyb-hwyf/ > gwyp'wyf
gadaw, gado > gad-ho' > gat'o:
dwyn > dyg-ho' > dyc'o:
and, finally in Adjectives
teg—teg-hed'—tec'-ed
teg-haf' tec'-af
gwlyb gwlyb-hed'—-gwlyp'ed
gwlyb-haf'—gwlyp'af.
The hardening in the Comparative Proper (tecach,
gwlypach, &c.) is due to the analogy of the other two
degrees—the Equal and the Superlative.
Intervocalic h is sometimes hardened into ch,* es-
pecially in the Gwentian dialect, as cychyd for cyhyd,
decheu for deheu, echon for ehon, eon, echedydd for
ehedydd. The change is euphonic.
* Cf. Silvan Evans's Welsh Dictionary s.v. cychyd.
|
|
|
|
|
(delwedd F6417) (tudalen 058)
|
58
WELSH PHONOLOGY.
THE ADDITION OF INITIAL H.
This occurs only in words beginning in a vowel.
is added
(1) To Nouns, after the Possessive Adjectives
m, my; ei, 'i, 'w, her
ein, '11, our
eu, 'u,
their, as
ei hwyneb: ein hawydd.
It
W'
(ii.) To ugain, when governed by ' ar,' on, in com-
posite numbers, as un-ar-hugain.
(iii.) To Verbs after the Postvoealic or Infixed) Per-
sonal Pronouns
'i, him, her
• 'u, them.
n, us ,
m, me
Efe a'u hanfonodd hwynt.' Luke ix. 2.
1.
I: may be a vowel as in
hir, ewin.
or a consonant as in—
iawn, iechyd, cuddio.
Compare y in English heavy, ready (vow el), and yes
young (consonant).
Long i like long a, long e, and long o are sinnple sounds
in Welsh, never diphthongal as in English.
The change from vocalic to consonantal i is very
simple. For the vowel the middle of the tongue is
brought near the palate: let there be a momentary
contact, and the ' i ' is consonantal.
|
|
|
|
|
(delwedd F6418) (tudalen 059)
|
WELSH PHONOLOGY.
Sl ( + VOWEL) =SH.
59
I after s, if another vowel follows, has no separate
value. The si together are sounded like s in English
sure.' The explanation is simple: the tongue seeks
to anticipate the ' i ' position during the articulation of
the sibilant, and neither the s nor the i is clearly pro-
duced. We get instead the mixed sh heard in English
shun, sure, that is, an s articulated with the tongue in the
i position. Examples
siom, slwr.
The reader will immediately recognise in these two
examples the English words ' sham ' and ' sure.' Fur-
ther investigation will lead to the suspicion that the
sound under discussion is confined to borrowed words.
This is not quite borne out by all the facts, at any rate
in dialects, for a few words like ' brysio ' are pronounced
by many people with an ' sh ' rather than an ' s-i.'
Still it would be generally conceded that to sound ' sh
in ' ceisio,' ' blysio,' is to use a corrupt pronunciation,
and it lilay be fairly argued that ' brysio ' should not
be an exception. Many people in Gwynedd find ' sh
difficult to articulate, and pronounce ' ship ' like ' sip.'
It will be instructive, therefore, to enquire where
we have borrowed this sound from. It cannot fail to
throw some light upon the limitations of Welsh speech-
sounds.
(1) First then, it stands for English ' sh '—a sound
usually represented in works on Phonetics by the old
symbol f:
|
|
|
|
|
(delwedd F6419) (tudalen 060)
|
60
siom
siop
siwr
sir
WELSH PHONOLOGY.
Eng. sham
Eng. shop
Eng. sure ,
< Eng. shire ,
Eng. shuffle
siffrwd <
Eraill yn siffrwd y Disieu a'r Cardieu. '
Bardd cwsc, 23.
siri < Eng. sheri//
Dyna, ebr ef, Rowndiad sy'n mynd yn Siri.'
Bardd Cwsc, 36.
(2) It is our way of representing Eng. ch (tsh) as in
chance, by ignoring as Modern French also does—the
initial t heard in that complex sound •
Sieb < Eng. Cheapside, as ' Siopeu Sieb ' in
Dafydd ab Gwilym.
sir ( > siriol) < English cheer:
Dy sir dda ar y cyrff meirwon.
Llyfr y Tri Aderyn, 165.
Gwnaethtlln iddynt gymmaint o sir ag a allwll.'
William Morris's Letters, 248.
sias (a turn, galile, set) from Eng. chase
Siarlymaen
siawns
siars
simnai
Rhisiart
Charlemagne ,
chance
charge
chimney
Richard.
(3) English i (and g in gaol --- dzh) is the soft sound
corresponding to the surd ch mentioned in (2) above.
Welsh has no such sonant sound corresponding to the surd
|
|
|
|
|
(delwedd F6420) (tudalen 061)
|
WELSH PHONOLOGY.
61
si. Hence English i is provected in borrowed words and
is represented by si, the initial d being again ignored like
the t of ch ( t J). French did not follow as simple a
i in
course: when the Classical Latin mid-front i (
Welsh iawn) was shifted forward and made a blade-
point sound ( = Eng. i) in Popular Latin, French
adopted the change, a course which English followed
under the influence of French. But while English has
adhered to the prosthetic d sound, French has long re-
verted to the pure spirant. Welsh Examples
Sior and Siors from Eng. George ,
siaced
Siams
Sian
Sion
Siasper
siwglaeth
jacket ,
James ,
Jane
John ,
Jasper ,
jugglery •
Hyd y Stryd allan gwelit chwareuon Interlud, siwg-
laeth a phob castieu hug. '—Bardd Cwsc, 23.
siel from Eng. jail, gaol:
Pa beth, ebr y Cyffeswr, tan edrych ar ryw siel ddu
oedd yno gerllaw. '—Bardd Cwsc, 34.
siwrnai fronl Eng. journey, or French journee.
These equations are subject to some limitations:
(a) They hold true only at the beginning of a word.
At the end, when the speaker's energy is already some-
what spent, his ability to imitate these foreign sounds
is more impaired, and they are all regularly reduced to
the pure sibilant s:
|
|
|
|
|
(delwedd F6421) (tudalen 062)
|
62
WELSH PHONOLOGY.
awrlais from Mid. Eng. orlage < French, horloge.
The ' ai ' is due to popular etymology.
colas < Eng. college.
Coleg is a learned borrowing from Lat. collegium.
ceisbwl < Eng. catchpole
hais < Eng. hedge
mantais < Eng. vantage
mursen < Eng. virgin .
For v > m, see under B: also the section on Provec-
tion.
< Eng. ostrich
estrys
scwrsio < Eng. scourge:
Tan scwrsio tri o ddynion a gwiail o scorpionau tan-
llyd.' Bardd cwsc, 3.
piser
PO tes
bernais
< Eng. Pitcher
< Eng. potage
< Eng. varnish
wits, witsiaid < Eng. witch
siars
< Eng. charge.
(b) The fact that (i) in si + a consonant the s is
generally a pure sibilant as in ' sicr,' and (2) the semi-
dialectal difficulty of articulating ' sh,' have led to
some uncertainty in the pronunciation of words like
sir ' and ' simnai. ' That is, the s is sometimes given
swn,' and not the more
the value of ' s ' in ' saer,
usual ' sh.' To these two causes, but more particularly
the latter, must be ascribed instances, like those given
below, of initial j (=dzh), and ch ( =tsh) becoming
( = sh). See reference above
Welsh ' s ' and not ' si
to the pronunciation of ' ship ' as ' sip.' Compare also
English chase and lance borrowed from the Central
|
|
|
|
|
(delwedd F6422) (tudalen 063)
|
WELSH PHONOLOGY.
French dialect, with their doublets
launch ' derived from Norman French:
siriol
' catch
som,' ' somi,' doublets of siom, siomi, e.g.,
Llyma ymodwrd yn iwerdon am y
63
and
som a
wnathoedit (i vatholwch) am y veirch.—Mab. 34.
Dros eich car disomgara
David Davis, dewis da.'
Edw. Morus, Cywydd y Paen.
Sermania ( = Germany)
Sarphes gadwynog falch anrhugarog,
A'i hesgyll yn arfog o Sermania.'
(quoted in) Drych y Pri/ Oesoedd, 24.
And Suddas < Judas
Gwnaeth Suddas a chusan
Diau gwnaid gloyw enaid glan
Dihenydd I)uw ei hunan. '—M.A. 342a: c 1300 A.D.
Milweis eiddig mal Suddas,
Heb son ana Dregaron gas. '—Gor. Owen, Cywydd i
leuan Brydydd Hir.
(c) The i (dzh) sound in English is a comparatively
modern development, and is unknown in classical Latin
and Greek. It follows that Welsh words borrowed from
Greek retain the consonantal i sound
lago, Iddew, lesu, loan, lorddonen,
(d) The modern innovations, Jacob, Jerusalem, Jere-
miah, &c., are unnaturalised forms.
|
|
|
|
|
(delwedd F6423) (tudalen 064)
|
64
WELSH PHONOLOGY.
Initial l- does not occur in the radical or dictionary
form of a word, except
(1) By aphaeresis: as, larwm < Eng. alarum.
(2) By crystallization of a mutated form as the radi-
loin) side by side with llwyn.
e.g. lwyn (
cal
(3) In a few borrowed words which have not been
quite naturalized, as
lifrai < Eng. livery
lili < Eng. lily.
In all other words original initial l- has been provected
into 11-, as
llaes < Latin, laxus.
Ll represents a peculiar sound. If the vocal organs
be in readiness to pronounce tl in English antler, the
sound Il can be arrived at by attending to two particu-
lars
(i.) Emit the breath more freely than for tl, and
(ii.) Let the emission of the breath be continuous,
without the sudden explosive sound of t at the beginning.
It is identical with the sound of ' dhl ' in Isan-
dhlwana as pronounced by a native.
The relation of these three sounds is intimate, and
their interchange might be illustrated from many lan-
guages. The diffculty of distinguishing r and I is felt
|
|
|
|
|
(delwedd F6424) (tudalen 065)
|
WELSH PHONOLOGY.
65
in our own time by the Chinese and Siamese. Christ
in Chinese is Kilisetu: a Siamese will pronounce ' the
flames rolled on ' as ' the frame loll on.' * When we
enquire into their physiological basis the ease of tran-
sition from one sound to another will be at once appre-
ciated. For all three the tip of the tongue is raised to
touch the sockets of the upper teeth. If then the
breath is passed through the nostrils, n is sounded. Let
it glide over the sides of the tongue, and I is produced.
Or finally, the tongue may be held loosely in the same
position, and the breath forced out over its centre and
tip, setting up a sharp vibration of the latter against
the front of the palate: Welsh r is the resulting sound.
The reasons for the interchange are various.
many cases it is the desire for dissimilation to avoid the
repetition of the same sound, as in
Latin caerulus from caelum ,
cancer from the root carcr.
French Pelerin and English Pilgrim from Latin pere-
grinus. Note that SvVe1s11 has not felt the need of a
change here, hence the form ' pererin.
Welsh galler often heard for gallel, gallael.
Again r is a difficult sound for many people. It has
disappeared almost entirely from English speech except
when it is followed by a vowel, thus reader is pronounced
ri:da. Some people in attempting the trill find the tip of
the tongue glued to the sockets of the upper teeth, and
their articulation of arall cannot be distinguished from
that of alall.
* Giles's Compara{ive Philology, p. 122,
|
|
|
|
|
(delwedd F6425) (tudalen 066)
|
66
WELSH PHONOLOGY.
The most active causes of the interchange of the
three sounds in Welsh words are carelessness of articula-
tion and a failure on the part of the hearer to catch the
precise sound uttered.
The following list is far from complete, but it will
help to shew the extreme fluidity of these liquids, and
incidentally suggest to the etymologist lines of enquiry
Intensive al-, an-, en-, er- .
alban, alaeth: annwn, enfawr: erfawr, erfyn.
addoli and Latin adoro
blagur alid brag (malt)
clul and cnul, also English knell ,
Cordovan and cordwal (
cordovan leather) •
Ac yna dechreu prynu y cordwal teckaf a gafas yn
y dref.' Mab. 48.
cornel < English corner ,
(ei) ddarllain yn ddifraw, ai fwrw i
Gochel di
gornel i rydu Yll dy erbyn. '—Morgan 141wyd, 252.
< L. consulor
cysuro
cythraul < L. contrarius
clwch, cnwc, and crug .
Clwch may be an Irish cognate. It occurs in ' Clwch
Dernog,' a farm in Anglesey, and perhaps in Maen
Clochog in Pembrokeshire. Cnwc. and crug, besides
being familiar as common nouns, enter into the place-
names Knucklas, in Radnorshire: Cruglas, in Anglesey;
Maesycrugiau, in Carmarthenshire (not Maesycrygiau) „
Chwefror and Chwefrol, Cornish Huerval.
toli, didol, toliant; ton, didwn.
torri, didor
|
|
|
|
|
(delwedd F6426) (tudalen 067)
|
WELSH PHONOLOGY,
67
The Def. Article 'yr;' 'yn ' in y naill (for yn aill), yn awr;
yl ' in y llall for yl all. In Gwentian and to a less extent
in Demetian ' y nall ' is the colloquial form of ' y lla1L'
Compare the Definite Article in Breton—ann, al, are
elydr and elydn (brass) .
'Ac ar y maen hwnnw y mae y delw honno or elydyn
teckaf wedy rydinetl ar lun dyn. '—Ystorya de Carolo, 4.
ernes and Gaelic earlas
fflangell and ffrewyll •
ffunen and Eng. fillet
gwawr, gwawl, and gwawn
Irish, gni: Cornish, gura
gwna, he does
groa, gra.
Ior(werth) and 1010 •
Ior and Ion
ulw, hulyf, and Lat. comb-uro ,
141eve1ys is called leuerys in Mab. p. 94
Mari, Mol, Mali, Malen
mesur and mesul
maenor, maenol, and maelor(?)
mwrn and mwrl (sultry)
mor(wydd) and Eng. mul(berry tree)
Sara, Sali, and Sal
shume and siffrwd
ser and Lat. stella.
N.
Breton
N. The normal value of n is the same as in English
man. Exceptions
|
|
|
|
|
(delwedd F6427) (tudalen 068)
|
68
WELSH PHONOLOGY.
(1) Before a labial it is assimilated into m not only in
the body of a word, as
an + parch < amharch,
but also in word groups, as
fyn brawd > fymbrawd > fy nmrawd
yn + Bangor > ymBangor > ym Mangor.
The predicative yn and the preposition yn + verb-
noun are notable exceptions in writing but not in pro-
nunciation, thus •
yn + parchus > yn barchus
yn + parchu > yn parchu.
The n in Llan- is not changed in writing before b-, p-,
as
Llanbedr, Llanberis,
but its phonetic value is distinctly that of nt. The
English for the fornmer Lannpeter -shews the correct
sound of the nasal.
(2) Before a guttural it has the value of ng in English
sing, as Bangor Bang-gor.
Final n is apt to pass into m, especially after a back
The labial no in its turn re-
rounded ' vowel like ' 0.
acts on the vowel assimilating it to the still more
rounded ' w: ' as
botwm < Eng. button
cotwm < Eng. cotton
latwm < Eng. latten
offrwm < offeren < L. o//erenda:
From the practice of making contributions (offrwtn)
to various causes at the celebration of the mass (offeren).
|
|
|
|
|
(delwedd F6428) (tudalen 069)
|
WELSH PHONOLOGY.
patrwm < Eng. battern
rheswm < Eng. reason, or French, raison:
saffrwlll < saffron.
69
So also in lladmer < Eng. Latimer for Latiner, the
original meaning of a Latin scholar becoming general-
ized in Welsh into an interpreter, translator: and
maentumio < Eng. maintain, or Fr. maintenir.
The same change of n to m is sometimes found in
English, as in Latimer above. Compare also lime-tree
for linden-tree: and pilgrim < Lat. Peregrgnus.
Final -in often passes into -ing, as
Llading < Latin •
colloquial pring < prin.
P. Aryan p is lost in Celtic, but velar q was advanced
into the labial position in Welsh, yielding a new p. It
will be interesting here to explain very briefly how the
change of velar q to p came about: velar q is a k sound
followed by w: that is, the q was sounded with rounded
lips. Once the lips were called into use, the whole duty
of articulating this ' velar q ' was thrown upon them in
Welsh. That is, from being a labio-guttural the sound
passed into a pure labial. This change took place in
Brythonic but not in Goidelic. Hence while ' four ' and
five ' are ' pedwar,' ' pump,' in Welsh and ' pevar,'
pemp ' in Breton, Old Irish shews the guttural, but
yet a guttural relieved of its excrescent ' w '—cethir,
coic.
|
|
|
|
|
(delwedd F6429) (tudalen 070)
|
70
WELSH PHONOLOGY.
R. Always a strong trill in Welsh. Initially it is
provected into rh.
When preceded by e, r has a tendency to develope
an i-glide after the vowel: thus in Mediaeval and early
Modern literature ger is regularly geyr, geir, and ger
bron is geirbron: so in eirmoet though here the i may
be on the analogy of eirioet where it can be accounted
for as an affection of the e due to the i after the r. Cyfeir,
cyfair, like its Irish counterpart comhair, shews the
glide. Dr. Silvan Evans (see Dict. under 'Cyfair') con-
nects the word with cyf and år, aru, to Plough, and adds
the note ' Cyfar seems the better form, but cyfair is the
older and the plural is more regularly formed from it.'
The fact is, it has nothing to do with år, aru. It is
simply the word ' cyfer,' opposite, against, and its
secondary meaning of acre is an interesting parallel to
English ' country,' which is borrowed from French
contree < Med. Latin contrata < L. contra against, oppo-
site, and therefore ' the land over against the town:
compare also German gegend < gegen.
s.
S. Never voiced as in English is, was.
Certain diffcult combinations of consonants lead to
the growth of a preceding euphonic vowel in some
languages. This is not the case in English, but it is
common in Greek, Popular Latin, French, Spanish, and
Welsh.
|
|
|
|
|
(delwedd F6430) (tudalen 071)
|
WELSH PHONOLOGY.
71
The consonantal group is usually initial s + conso-
nant, as sc, sp, st, sm. Thus in French
esprit < L. spiritus
école < Old Fr. escole < L. scola ,
< Old Fr. escu < Popular Latin, iscutum:
écu
< Latin, scutum.
In Welsh the euphonic vowel usually prefixed is the
obscure y-, as
ysgrif < L. scribo ,
yspryd < L. spiritus
ysmygu < Eng. smoke
< Old Eng. stane (a stone jug).
ysten
Sometimes the vowel is e, as
esmwyth: compare Eng. smooth
esgud, swift, nimble: compare Eng. scud.
Sometimes the glide, heard when the tongue takes
up the position for articulating the nasal n, is repre-
sented in writing. Thus neidrwydd, the temples, appears
in Lly/r Ancr (93) as eneidrwyd. But this is of a differ-
ent order, as the insertion of the e in eneidrwydd is due
to sonlewhat leizurely and slipshod pronunciation.
In the case of yspryd, ysgwydd, ysmygu, &c., on the
other hand, a real difficulty of pronunciation was ex-
perienced, and the problem was solved by dividing the
consonants between two syllables, the first of which
was improvised for the purpose. Even this expedient
failed to solve the difficulty in French, for after the
addition of the euphonic e- the s was nevertheless
dropped from the maj ority of the words.
|
|
|
|
|
(delwedd F6431) (tudalen 072)
|
72
WELSH PHONOLOGY.
In Welsh, where the y- (or e-) was not prefixed, either
the s was dropped, as in
mwg, mwyth (contrast ysmygu, esmwyth) ,
or the following consonant, as
sefyll for stefyll, from the same root as Latin sto
English, stand
sofl < Lat. stipula.
It must not be assunmed that the euphonic y- was not
heard because William Salesbury, Morgan Llwyd, Ellis
Wyn, Theophilus Evans, and other writers of their
time generally ignored it in writing. We might just as
reasonably say that b was sounded in the 16th century
in English det and dout, for b was then reintroduced into
these words by •writers of an etymological school, which
sprang up with the Revival of Learnincs in Western
Europe. The very same thing happened ill France
during the same period.
It is unfortunate that
Welsh is not yet free frolll the practice and doctrines of
those etymologists from Salesbury to Dr. William Owen
Pughe, who seriously tampered with our orthography,
and who propounded absurd and impossible theories
in Welsh etymology. It is true that Edward 141Mvryd
and Sir Willia1Y1 Jones were brilliant exceptions, and
did a great work on sound lines in Celtic and Aryan
Philology, but their influence upon the current of lin-
guistie opinion in Wales at (he tirne was comparatively
small.
Even the sound s was found difficult of articulation,
for in Old Welsh, where it was not flanked by a pro-
|
|
|
|
|
(delwedd F6432) (tudalen 073)
|
WELSH PHONOLOGY.
it almost*
tecting consonant, usually t,
passed into h, as
haul, compare Latin sol
73
invariably
hir
haf
serus
Eng. summer.
T.
Intervocalic t has become d in Modern Welsh, as
adar from older atar.
What then is the origin of our modern intervocalic
t? It is from
(1) dh as ateb < ad-heb, tecaf < teg-haf.
(2) d-d as cytuno < cyd-dy-uno.
dh and d-d alliterate with t in Welsh poetry, e.g.:
Wych mwy teg, a chenadhau. '—-Edw. Morus Cywydd
y Paen.
Oleu teg a elwid dydd.' G. O. , Cyvyydd y Calan.
(3) d-d < d-b as petawn < ped-dawn < ped-bawn.
(4) Analogy as rhatach for rhadach, because of the t
in rhated, rhataf.
(5) t in learned borrowings, as natur < L. natura.
Intervocalic -c- and -p- in Modern Welsh are to be
similarly accounted for.
Th: always voiceless as in English thin, never voiced
as in this.
* s In saith (cognate with Latin septem), sil (doublet of hil),
and one or two other words resisted the change into h.
|
|
|
|
|
(delwedd F6433) (tudalen 074)
|
74
WELSH PHONOLOGY.
As no words in Welsh begin with th, the digraph is
changed into t in borrowed words, as
trefa < Eng. thrave ,
tron, trwn < Eng. throne.
(See also under C, Ch, G, and Chapter I.)
It may be a vowel as in gwn, or a consonant as in
gwas.
The vowel 'w' in the diphthong 'WY' is apt to become
consonantal: thus w in awydd and awyr in older
writers was usually vocalic, but in present-day
Welsh they are generally pronounced aw-ydd, aw-yr.
Compare also cadwen from cad€yn, and the use of
tywell, sometinmes Illet with, as feminine of tywyll im-
plies a new consonantal value for the w.
The w in marwnad, meddwdod, delw, &c., is conso-
nantal in origin ( O. Ir. b), and it is seldom given a
syllabic value in poetry, delw being a Inonosyllable and
meddwdod a dissyllable for metrical purposes. For
the same reason the accent on Inarwnad, &c., is on the
first syllable ( mar'-wnad).
Thus in Goronwy Owen's Cywydd Marwnad Marged
Morys, ' garwfrwyn ' and ' weddwdod ' are dissyllabic:
Tristyd ac oerfyd garwfrwyll
Llwyr brudd a ehystudd a chwyn.
Ac o'i herwydd dwg hiraeth
Ormod, ni fu weddwdod waeth.'
|
|
|
|
|
(delwedd F6434) (tudalen 075)
|
WELSH PHONOLOGY.
75
The final w in hwnnw and acw are exceptions in that
they are not consonantal. Both come from o by back
formation: acw is from Mediaeval racco, which with
the loss of the r through confusion with the Definite
Article, gives modern aco (yco colloquial) > acw.
Hwnnw is from hwn + yno, just as honno is from hon
yno, and, to give it more the appearance of a masculine,
hwn-yno was changed into hwnnw (for hwnno), just as
Tom becomes Twm and bord > bwrdd. Probably the
w ' of ' acw ' is due in part to the ' w ' of hwnnw.
|
|
|
|
|
(delwedd F6435) (tudalen 076)
|
76
WELSH PHONOLOGY.
CHAPTER 111.
MUTATION OF INITIAL CONSONANTS.
The words of a sentence fall into groups according to
the closeness of their interdependence: thus in
Cefais y llyfrau yn rhåd,
we have two phrase units
cefais y llyfrau,' and ' yn
rhåd.' In reading observe an interval between the
two: but a stop cannot be made in the middle of the
unit, say between ' y ' and ' llyfrau,' without tending
to obscure the meaning. Again in-
Daeth llythyr i'm llaw.
O'i dlodi a'i hiraeth I cododd yr alltud I lef I oedd
yn adlais o feddwl Cymru
the units are indicated by the perpendicular lines.
Just as in French, liaison is regularly observed within
the phrase, so in Welsh the nine liltltable consonants
undergo changes according to the following table and
the appended rules:
|
|
|
|
|
(delwedd F6436) (tudalen 077)
|
Radical.
t
b
11
m
rh
WELSH
Soft.
b
d
dd
1
r
PHONOLOGY.
Nasal.
mh
nh
ligh
m
ng
77
Aspirate.
th
ch
* Formerly f from m was nasalized and this dis-
tinction between f ( < b) and f ( < m) is still kept up
in Breton.
Originally a very soft guttural spirant was audible
in this Inutation of g.
PRINCIPLES OF SOUND-CHANGE.
MUTA'DION IS ASSIMILATION, and its object is to econ-
omize effort in pronunciation. It takes place to some
extent in other languages: thus in + perfect becomes
imperfect, and dogs is pronounced dogz in English.
Note in these two examples that the first and last
consonants of the significant or root word assimilate
the adjoining consonants of prefix and Stiffx. That too
is the case in Welsh, where it is extended to word groups.
|
|
|
|
|
(delwedd F6437) (tudalen 078)
|
78
WELSH PHONOLOGY.
NASAL MUTATION: The first stage in the assimila-
tion of yn + pen is ymPen, a form familiar in Mediaeval
literature. But Welsh does not stop here: the -n, after
becoming -m by partial assimilation to the p-, in turn
eclipses the latter, that is, converts it into mh. Hence
the stages are •
yn pen > ympen > ym mhen or ymhen.
yn Caer > yng Caer > yng Nghaer or y' Nghaer.
Fyn Duw > fynDuw > fy Nuw.
The same changes occur in COMPOSITION, e.g. .
Cvmbro
Con + bro >
Cvmro.
con + paro > Lat. comparo > Welsh cymharu.
an + doeth >
annoeth.
SOFT MUTATION: Note that a vowel is a voiced or
sonant spirant, that is, a so/t sound produced by a con-
tinuous though very slight emission of the breath, caus-
ing a vibration of the vocal chords. On the other hand
p, t, c, are voiceless or surd mutes, that is, hard sounds
produced by a sudden explosive opening of the vocal
organs after a complete stoppage of the breath.
In these there is no ' voice,' for the vocal chords do
not vibrate. Hence a vowel-flanked p, t, or c, will be
partially assimilated from surd or voiceless to sonant or
voiced (b, d, g). Thus:
Old Welsh: aper > Mod. Welsh aber.
map >
mab.
and ei + pen > ei ben, his head.
Similarly the sonant mutes, b, d, g, are changed into
the sonant spirants, f, dd,—(g disappears), under the
same conditions.
|
|
|
|
|
(delwedd F6438) (tudalen 079)
|
WELSH PHONOLOGY.
79
The vowel preceding the mutated consonant has often
disappeared from Modern Welsh, but the mutation
in so far as it is phonetic in origin (see the disturbing
influences referred to below)—is as positive a proof of
its presence in the early stages of the language, as the
nasal mutation in ' nhad,' used by a child in calling or
addressing its father, implies a lost ' fy ' (older—fyn)
before the noun. So the Definite article ' yr '* softens
the initial Inutable consonant of a feminine singular
noun because the former word in this gender and num-
ber originally ended in a vowel. Similarly it softens the
d of dau just as dau itself turns ' dyn ' into ' ddyn,' e.g.
y ddau ddyn,' because the dual in Celtic as in Greek
ended regularly ill a vowel. Similarly, as -o- connects
cun and belinos in Cunobelinos (Mod. Cynfelyn) and
accounts for the softening of the b into f, and as we
write Anglo-Norman for English-Norman, so mar-
gaine and diod-lestr must have been connected by a
vowel, which accounts for gaine ( < cainc) and lestr
< Ilestr.
ASPIRATE MUTATION. Stages
ac + pan > aman > a Phan.
ac + lair > attair > a thair, &c.
ac + can > accan > a Chan.
The "P, tt, and cc involve such a strong emission of
breath that they resist ' voicing ' into b, d, g. The alter-
* The reader must not think that the mutation of (e.g.) m.
into f in ' y fam' ( < mam) is due to the vocalic ' y.' The
softening dates from a time anterior to the loss of the r of yr and
implies an early feminine form such as ' yre. '
|
|
|
|
|
(delwedd F6439) (tudalen 080)
|
80
WELSH PHONOLOGY.
native is assimilation to the spirant character of the
vowels flanking them. Similarly in composition
Lat. occasio > W. achos.
cippus > W. cyff.
sagitta > W. saeth.
Two FACTORS have considerably affected the range
and use of mutations:
(a) ANALOGY: e.g., wyth nasalizes the b of blwyddyn
because saith and naw do so. But saith and naw in
Brythonic ended in -n ( = m in Latin septem and novem).
Wyth never had a final -n.
(b) The growing use of mutations for purposes of
SYNTAX. Thus, on phonetic grounds, in the 18th cen-
tury and earlier, it was customary to soften the initial
consonant of the subject after verb-forms in -ai, that is,
the 3rd sing. Past Imperfect and Pluperfect, no doubt
due to the influence of the final vowel in the verb, e.g.,
O chai fachgen wrth eni,
Wyd awen deg, dy wen di. '—Gor. Owen.
This softening of the initial of the subject, in the case
of the substantive verb, was extended by analogy to
oedd and sometimes to other parts, both of the sub-
stantive verb and others, e.g.,
(Syn fyfyriais) wyched oedd gael arnynt lawn olwg. '
Bardd cwsc.
Dygymydd Duw ag emyn
O awen dda a wna ddyn. '—G. O.
|
|
|
|
|
(delwedd F6440) (tudalen 081)
|
WELSH PHONOLOGY.
81
Efe oedd gannwyll yn Ilosgi, ac yn goleuo. '—loan v.
35.
Ac o'i herwydd, dwg hiraeth
Ormod, ni fu weddwdod waeth.' G. O. vi.
Nid oes ddim cywir-Grefydd yn yr offeiriad. '—Def.
Ffydd, 66.
Still exceptions occur, as
Fel y galle Cymro ei ddarllen ef yn rhwydd. '—Def.
Ffydd, xviii.
Amhossibl fydde dwyn. '
Ibid. xiv.
But in present-day Welsh it has become the rule
that the initial consonant of the subject is to remain
radical after all parts of the verb (see below).
RULES OF MUTATION.
The commonest change is
THE SOFT MUTATION.
(a) After the Definite Article, in any noun or adjec-
tive—including ordinal numerals—of the Fem. Sing.,
y bob], the People.
a'r fwyn ferch, and the gentle maid.
y bumed salm, the fifth psalm.
N.B.—The following remain radical after the Def.
Art. .
(i.) Words in 11- and rh-, e.g.,
y llaw, the hand.
y rhaw, the shovel.
6
|
|
|
|
|
(delwedd F6441) (tudalen 082)
|
82
WELSH PHONOLOGY.
This is rather a case of reversion after a period of
mutation. The stages are
lawf — yr law —- yr llaw — y
Vre llawf yr (e)
llaw: for rl becomes rll, as perllan from perlan: con-
trast ydlan. Harlech for Mediaeval Harddlech (see Mab-
inogion) and erlyn for older erglyn are scarcely excep-
tions, for the intervening ' d ' and
' were present
during the period when rl changed into rll. So h in
rr before an accented syllable be-
rhaw is accentual,
comes rh.'*
(ii.) The initials of plurals, except ' pobloedd,' e.g.,
y gwragedd, the women.
Vr lesu a ddechreuodd ddywedyd wrth y
bobloedd,' Matt. xi. 7.
(iii.) Initials of Cardinal numerals of either gender,
y tair hudoles hyn, these three witches.'
Bardd cwsc.
y Deng Air Deddf, the ten commandments.
Dau (m.) and dwy (f.) , two, on the other hand, shew
softening:
y ddau ddyn, the two men.
y ddwy wraig, the two women.
(b) In nouns, as agnomens—that is, when following
proper nouns of either gender as titles or epithets, e.g.
loan Fedyddiwr.
Mair Forwyn.
Duw Dad.
lesu Fab Dafydd.
serigi Wyddel.
Hugh Flaidd.
* See Welsh Orthography, pp. 22—3.
|
|
|
|
|
(delwedd F6442) (tudalen 083)
|
WELSH PHONOLOGY.
Also in adjectives qualifying proper nouns
83
Hywel Dda.
Noah Gyfiawn.
Exceptions: Rhodri Mawr:
Deio ap leuan Du, where D (for Dd) is due to the
hardening effect of the n. Compare bendith for ben-
ddith, and see Chapter iv.——Provection.
(c) In addresses with or without an interjection, e.g.
Frodyr anwyl.
O Dduw !
Da, was da (from awas). Ha, was drwg.
(d) In an adjective after a noun feminine singular,
gwraig dda.
(e) Likewise, in the latter of two nouns, if used
attributively with a feminine noun singular:
llwy de.
llwy fwrdd.
pont bren.
canwyll gorff.
Pobl Wynedd—G. O.
' Gwlad Gymry. Kymn.
nodwydd ddur.
V Forwyn Fair.
canwyll wér (fro gwér). Gwlad Fon.
y dduwies Wener (fr. Gwener).
Contrast .
llawr coed.
llwybr troed.
y dyn Crist lesu.
where the initial remains radical, because the nouns
llawr, llwybr, and dyn are masculine.
De,' ' fwrdd,' and most of the second words in the
above groups, are syntactically Attributive Genitives.
Similarly any noun in the Genitive dependent on
another noun is more or less attributive in character,
|
|
|
|
|
(delwedd F6443) (tudalen 084)
|
84
WELSH PHONOLOGY.
and on that account tends to show initial soft mutation
if the noun, on which it depends, is feminine, e.g.
Teyrnas Dduw.
Din Dryfol.
Caer Wrangon.
Awr weddi (gweddi).
(f) Even after a masculine in a very few instances,
soft mutation occurs, e.g.
Ty Dduw.
Ty Ddewi.
This is rare, and is probably due to
(1) Analogy with feminine nouns, and
(2) The fact that these phrases are of comnaon oc-
currence, and therefore more or less subject to the rule
of mutation in compounds.
(g) If nouns of either gender, singular or plural, are
preceded by an Adjective in the Positive, or a noun
used attributively, Welsh tends to regard them as com-
pounds, and the initial consonant of the second word
is softened
hen dy:
y gwir Dduw
gau brophwyd
cam ddefnydd •
gwag ogoniant (fr. gogoniant)
y Duw Ddyn •
diod lestr:
mor gainc.
N.B.—Genera11y in Welsh such word groups coalesce
and form genuine compounds, as milgi, greyhound ,
gwirfodd, good will, consent: gwinllan, vineyard: mor-
fran, a cormorant: byrbryd, luncheon: byrgoes, short-
legged: byrbwyll, rash, hasty.
Still, though the native tendency of the language is
to reduce them all into compounds the process is gra-
dual, and it is sometimes difficult to decide whether the
two words have already coalesced: e.g. the forms
|
|
|
|
|
(delwedd F6444) (tudalen 085)
|
WELSH PHONOLOGY.
85
coel grefydd and coelgrefydd may be defended,
gwag ogoniant and gwagogoniant, &c.
As the mutation is the same in either case, the accent
is the only test whether a particular example is to be
written in two words or as a compound: the two words
are separate if they retain their respective accents, as
hen' dy',' an old house: but if one is dropped they form
a compound, e.g. ' hen'dy.
(h) After the INDEFINITE ADJECTIVES ambell, am-
ryw, cy/ryw, holt, rhyw, unrhyw, ychydig
Also after the INTERROGATIVE ADJECTIVE pa .
ambell waith (fr. gwaith)
yr holl /yd: unrhyw le •
Pa dduw sy'n maddeu fel Tydi ?
(i) In nouns after the Possessive Adjectives .
dy, 'th —thy: ei, 'i, 'w —his, e.g.
dy gleddyf •
i'th law
ei ben
i'w ferch.
(i) 111 Felninine nouns after Ordinal Numerals, e.g.
y bedwaredd bennod.
yr ail ferch
N.B.—In present day Welsh, even in Masculine nouns
after ail, e.g.
yr ail lyfr.
yr ail ddiwrnod
The Bible shews the transition, e.g.,
Ail Llyfr Samuel, but Ail Lyfr y Brenhinoedd.
(k) After the cardinal numerals
(i.) un (fem.
un wraig (fr. gwrai2), one woman.
|
|
|
|
|
(delwedd F6445) (tudalen 086)
|
86
WELSH PHONOLOGY.
but Il and rh remain radical, e.g.
un llaw:
un rhodfa.
Or rather, as after yr, the ll- and rh- are instances of
reversion, the I > Il due to the hardening effect of n,
and nr before an accented syllable > nrh. Contrast
anrhydedd with anwiredd ( < gwir).
(ii.) dau, dwy:
dau /rawd; dwy eneth.
So in the qualifying adjective, and for the same reason:
Fal y ddau ychen, hen hy
Fanog, ba beth y fyny. '—D. ab G. cxiii.
Gwae fi, na chair deuair deg,
Oni chawn un ychwaneg. ab G. clxx.
Deu wydel uonllwm. '—Mab.
Occasionally, c-, p-, t-, remain radical after dau, as—
dau Parth,' two Parts.—2 Kings ii. 9.
y dau cymaint,' double.—Rev. xviii. 6.
(iii.) Saith, wyth, if the following consonant is p-,
saith dorth,' seven loaves.—Matt. xvi. 10.
saith dymor dyn, the seven ages of man.
wyth bennill, eight stanzas.
and often in the case of the other Intltable consonants,
as—
saith /asgedaid,' seven basketsful.—Matt. xv. 37.
dinas deg ar saith fryn, a fair city on seven hills.
|
|
|
|
|
(delwedd F6446) (tudalen 087)
|
WELSH PHONOLOGY.
(iv.) Deng, if the next consonant is g- .
as—
87
deng wr a thrugain,' seventy men. —Numbers xi. 24.
(l) In nouns, including the verb-noun, when following
a PERSONAL VERB that governs them in the Accusative
case .
Gwelais ddyn, I saw a man.
Clywaf ganu, I hear singing.
NOTE. (i.) The accusative following a verb in the
impersonal form shews no mutation, e.g.,
Gwelwyd dyn, a man was seen.
Clywid canu, singing was heard.
(ii.) A noun dependent on a verb-noun is in the
Genitive (not accusative) case, and therefore shews no
mutation as a verb-noun is always masculine (see (f)
above), e.g.,
Gweled dyn, seeing a man.
Clywed canu, hearing singing.
(iii.) The subject of a verb usually shews no initial
mutation in present-day Welsh, e.g.,
Daw diwrnod cyfrif ar 01 hyn.
A day o/ reckoning will come after this.
(iv.) In older Welsh the initial consonant of the
object often relnained radical, e.g.,
ef a welei carw: ny welsei cwn ,
Ef a glywei Ilef •
but also——ef a welei lannerch
all 011 p. 1 of the Mabinogion.
And ill the following line from Deio ap leuan Du:
V Ddraig Goch ddyry cychwyn.'—G. B. C. 172:
|
|
|
|
|
(delwedd F6447) (tudalen 088)
|
88
WELSH PHONOLOGY.
the cynghanedd required c- in cychwyn to alliterate
with the two g's in y Ddraig Goch.
(m) In an Adjective after a noun feminine sing., e.g.,
Gwraig dda, a good woman.
But final -s sometimes restores the radical, as
nos da: ewyllys da: Vnys Brydain.
(n) Sometimes in the Comparative after a masculine
noun, if the sentence is negative or implies a negative,
Diheu oed gantunt na welsynt Ilongeu gyweiriach y
hansawd no wynt. '—Mab. 27.
But contrast
Gwaew a maneg am einioes,
Gwalch na dyn glewach nid oes. '—Tudur Aled,
G.B.C. 229.
(o) In the Personal Pronouns mi, minnau—Nomina-
tive or Accusative—if the verb precedes, e.g.,
dysgaf fi or dysgaf i (mi > fi > i)
dysgais innau (minnau > finnatl > innau)
Gwelodd finnau.
Similarly in ti and tithau, unless the verb ends in t
Gweli di -dithau
Gwelais di -dithau.
(p) In the m- and t- of mi, minnatl, ti, tithau in the
Genitive Case, e.g.,
fy llaw i
dy awydd dithatl.
|
|
|
|
|
(delwedd F6448) (tudalen 089)
|
WELSH PHONOLOGY.
(q) In a verb following its subject •
Mi a welais (fr. gwelais)
Vr arglwydd a roddodd (rhoddodd).
89
(r) After the Adverbs ni, oni, na, if the verb begins
in b-, d-, g-, 11-, m-, rh, as
ni ddaw efe
oni welwn ni 'r mar ?
na lwfrhewch.
N.B. The b- in byddaf and other parts of the sub-
stantive verb is sometimes softened but generally re-
mains radical
Oni bai gywiro o'u harglwydd ei addewid.'
Edw. Samuel.
(s) In Adjectival and Adverbial expressions after the
predicative yn, as
yn dda, zvell: yn deg, fairly
(i.) ll- and rh- remain radical
but note
vn llawen, gladly: yn rhwydd, easily.
ii.) n is usually dropped at the beginning of a
sentence, and the consonant returns to its radical form,
Gwn yn dda, but da y gwn.
(iii.) If yn is dropped in the middle of a sentence, the
consonant may or may not be softened
A phump ohonynt oedd gall ' (fr. call). Matt. xxv. 2.
Na phecha mwyach, rhag digwydd i ti beth a fyddo
gwaeth ' (fr. gwaeth). John v. 14.
|
|
|
|
|
(delwedd F6449) (tudalen 090)
|
90
WELSH PHONOLOGY.
(iv.) Predicative yn is always dropped after the pre-
position yn, and the consonant remains radical, e.g.,
Pan ydynt yn manwl gyfrif. '—Edw. Samuel.
(t) After the prepositions
am, ar, at, gan, heb, i, o, tan, tros, trzvy, wrth
hyd when it means as far as
Heb Dduw heb ddim. '—Proverb.
Wrth law, at hand: O Gaer, from Chester.
also
N.B.—' i'
is followed by the radical in mi, minnau,
ti, tithau.
(u) After pan, when, and the interrogative particle a:
Pan ddaeth
A ddaeth efe ?
(v) In present-day Welsh the soft mutation has be-
come the rule in a word when separated by an inter-
vening word or phrase from the word with which it is
closely connected, e.g.,
Ac wedi iddo ddywedyd hyn.
And when he had said this.
Contrast with this
Wedi dywedyd hyn, having said this.
N.B.—This rule overrides all others, e.g.,
Daeth llawer yno, but
Daeth yno lawer:
notwithstanding the rule requiring the radical initial in
the subject of a verb.
|
|
|
|
|
(delwedd F6450) (tudalen 091)
|
WELSH PHONOLOGY.
91
Hence dyma, dyna, dacw, &c., are followed by the soft
mutation in accordance with rules (l) and (v), e.g., in—
Dyna ddyn da.
There is a good man.
ddyn is accusative case governed by gwel (understood),
and therefore d- becomes dd-. The soft mutation more-
over is required because ddyn is separated from the verb
gwel by the parenthetic dyna.
This beautiful rule is of comparatively recent growth.
Thus in Defynniad Ffydd (1595) the radical is continually
used after a parenthesis, e.g.,
Ceisio bwrw i lawr braint llywodraethwyr,' p. 7.
V n yr hyn (Swpper yr ArÅ1wydd) rydys yn rhoi fegis
geyrbron eyn llygeid marwolaeth ag adgyfodiad Crist,'
38.
TILE SPIRANT MUTATION: This takes place in words
after
(i.) The adjectives tri = three: chwe =six: ei, 'w = her
(possessive): all of which originally ended in -s, e.g.,
tri phen, three heads
chwe cheiniog, sixpence.
(ii.) The adverbs tra, ni, na, oni.
V dyn tra Phenderfynol,
The very determined man.
INa Phechweh, sin not.
(iii.) The prepositi011 å ( < åg—with),
pounds gyda, with: tua, towards.
Tua chyfeiriad y mar,
In the direction of the sea.
and its com-
|
|
|
|
|
(delwedd F6451) (tudalen 092)
|
92
WELSH PHONOLOGY.
(iv.) The conjunctions a, and: na, nor, than, that,
not: o, if •
' O cherwch fi, cedwch fy ngorchymynion,'
If you love me, keep my commandments.
The NASAL MUTATION takes place after—
(i.) The possessive adjective fy, my: e.g.,
fy mhen, my head.
(ii.) The preposition yn, in: e.g.,
yn niwedd y byd, in the end o/ the world.
NOTE.—(i.) Vn, however, does not nasalize the initial
of
(a) A verb-noun: e.g.,
yn gwerthu, selling.
(b) Cymraeg, Welsh, when used alone:
yn Gymraeg, in Welsh.
But note: yng Nghymraeg Rhydychen, in Oxford
Welsh: yng Nghymraeg y Beibl, in Bible Welsh.
(ii.) The predicative Yll when followed by pell, far,
and sometimes cynt, earlier: e.g., ym mhell; ynghynt.
This is no doubt on the analogy of the prep. yn.
(iii.) The cardinal numerals pum, saith, wyth, naw,
deng, deuddeng, Pymtheng, ugain and its compounds,
cau, when followed by
(1) blwydd, blynedd: e.g.,
Y mae efe yn ddeng mlwydd oed,
He is ten years old.
|
|
|
|
|
(delwedd F6452) (tudalen 093)
|
WELSH PHONOLOGY.
93
(2) diwrnod, though this not infrequently retains
its radical initial
deng niwr110d. '—Rev. ii. 10 ,
deg diwrnod . . ugain diwrnod. '—Num. vi. 19.
N.B.—(a) Deng always nasalizes the d- of diwrnod.
(b) The nasal mutation after un in—
un mlynedd ar ddeg.' Jeremiah i. 3
seems due to a syntactical confusion, the writer regard-
ing un-ar-ddeg as a phrase unit. If it were possible to
put blynedd after deg ( > deng) the nasal mutation
would of course be correct.
|
|
|
|
|
(delwedd F6453) (tudalen 094)
|
94
WELSH PHONOLOGY.
CHAPTER IV.
OTHER CHANGES AND THE ACCENT.
I—SOFTENING OF CONSONANTAL SOUNDS.
We shall refer only to the voicing of p, t, c into b,
d, g, and the aspiration of b, d, g, m into f, dd,
Something of what will be discussed here might have
been given in the chapter on the mutation of Initial
Consonants, but it seemed to conduce to simplicity
and clearness to adopt the present arrangement.
As the initials b, d, g may be either radical, as,
bardd, dafad, gardd:
or the mutated forms of p, t, c, as,
ei ben ( < pen), ei dafod ( < tafod), ei gynllun
( < cynllun),
it is natural that much uncertainty should often exist
as to whether a form in b, d, or g should be treated
as a radical, or referred back to a form in p-, t-, or c-.
Further, the frequent use of many words in the soft
mutation tends to stereotype those forms: they lose
their resiliency or elasticity, and fail to revert to their
|
|
|
|
|
(delwedd F6454) (tudalen 095)
|
WELSH PHONOLOGY.
95
correct radical form. Thus several prepositions are
regularly used in the mutated forms, as ' wedi ' and
gan,' from ' gwedi ' and ' can.'
So the first part of beunydd, beunoeth is the ac-
cusative of pob.
bellach < pellach:
bitw—-a word used in Cardiganshire for ' small,' ' very
small,'
is from English petty, or its parent French petit.
drachefn < trachefn •
elor is from O. W. gelawr;
euog < geu-og < gau •
grofft (see Mab. pp. 52, 53) < Eng. croft:
Vychan and English Vaughan < Bychan > bach;
Vchydig < bychydig < bach,
Intervocalic p, t, c are softened into b-, d-, g-, as
aber < aper •
gadael < gatael:
rhagor < rhacor, racor,
Present-day intervocalie p, t, c have sprung from other
sources: see Chapter Il. under t.
Final -p, -t, -c after a vowel are regularly voiced, as—
carped, cwpled, &c., < Eng. carpet, couplet:
apostolig, catholig, &c , < Eng. apostolic, catholic ,
sieb, mab, &c., < Eng. cheap(side) & Old W. map.
The few modern words ending in a vowel + -p, -t, -c
are not so much exceptions to this phonetic law as in-
stances of counteracting influences, which either resist
the change or cause a reversion from the softened form.
A few examples will serve as illustrations.
|
|
|
|
|
(delwedd F6455) (tudalen 096)
|
96
WELSH PHONOLOGY.
Chwip and not chwib from Eng. whip. The p is re-
tained because the i is short. -B tends to lengthen
the preceding vowel, but lengthening the i in this
onomatopoetic word wonld be to weaken the sense: so
rhic ', a cut or groove.
The conjunction ' ac ' retains its ' c ' in order to differ-
entiate it from ' ag,' ' åg.
to) is probably a back-formation from ata/ for
adaf, the d iu this word being provected. Compare the
practice of hardening intervocalic voiced consonants in
Gwentian, as retws for rhedodd. The need of distin-
guishing the pronominal preposition from (g) adaf, I
leave, has been another contributory cause, as also the
analogy of English at.
Brat, perhaps an Irish word, has a short a, and to
change t into d would be to lengthen the vowel, and
yield a word identical in sound and spelling with brad,
treachery.
The -t in the 2nd person singular of the Past Im-
perfect and Pluperfect of Verbs (as dysgit, dysgesit) and
of Pronominal Prepositions is provected from older d
under the influence of the t of ti habitually following
these words.
ll.—PROVECTION AND BACK-FORMATION.
(Cf. Chapter 11. under 'B and M
In the case of initial consonants this is the opposite
of soft mutation. The uncertainty over the correct
radical form, due to the prevalence of mutation, ac-
|
|
|
|
|
(delwedd F6456) (tudalen 097)
|
WELSH PHONOLOGY.
97
counts for nuany coined radical forms, as bicer, bicar,
micar, biceriaeth, from English vicar. So almost all
borrowed words in /-, 1-, r- have been provected into
rh-, as a perusal of a Welsh
forms in m- or b-,
dictionary will amply prove.
Sometimes, however, provection is due to an accen-
tual stress as in tyred ( dy + rhed), the Imperative
of ' deuaf.
(a) The following are instances of initial provection '
berfain < Eng. vervain;
bogail < Lat. vocalis;
bernais < Eng. varnish,
berf < Latin, verbum;
bernagl < Eng. vernicle;
bitail & bwytal < Mid Eng. or Anglo French,
vitaille;
becso (colloquial in Cardiganshire) < Eng. vex.
telaid, telediw < del, delw (Old Ir. delb);
tom < dom (cf. Eng. dung, from which it is
probably borrowed. For change of ng to m
compare Trallwng and Trallwm = WelshPool,
carlwng and carlwm)
tracht < dracht < Mid Eng. draught (in Mod. Eng.
the guttural has been advanced into a labio-
spirant)
trum (Irish, druim) < drum ,
7
|
|
|
|
|
(delwedd F6457) (tudalen 098)
|
98
WELSH PHONOLOGY.
tylluan < dyllhuan, dallhuan:
'Ac o achaws hynny y mae digassawc yr adar
yr tylluan. '——Mab. 80:
trem < drem •
tychan < dychan •
tyred < dyred
tywynnu < dy + gwyn.
pe < be < bei ( = bai, byddai) •
peibl (rare) < beibl,
Pawb a gasul ful foelrhawn,
Pibl iaith yn dwyn pobl i iawn.'
D. ab G. ccxvii.
potel < Eng. bottle:
Pabyloniaid < Babyloniaid, as—
Pobl anwar Pabyloniaid.'
C.O., fliraethgan am Fon.
post < bost < (Eng. boast):
Da gwyddost, ein post a'n parch
Troi dwylaw fal traed alarch. '—G. B.C., 175.
Prydain < Brydain (whence Latin, Britannia).
Much uncertainty has been experienced by writers in
dealing with this word owing to the common use of
ynys before it. As ' ynys ' is feminine the rule re-
quires the softening of the following initial; thus,
Vnys Frydain. But -s tends to resist this change
(compare ' nos da' for ' nos dda '): hence ' Vnys
Brydain ' is common. Compare ' Vnys Mon ' side by
|
|
|
|
|
(delwedd F6458) (tudalen 099)
|
WELSH PHONOLOGY.
99
side with ' Vnys Fon.' But this influence of s is not
as readily observed as the rule of soft mutation.
Hence B in ' Vnys Brydain ' was supposed to be
derived fronl radical P. Hence the new forln Prydain.
This uncertainty may be well illustrated from Drych
y Pri/ Oesoedd
Hen fatterion Brydain ' (radical), 3a.
V m Mhrydain ' ( < radical—Prydain), 31.
par ( < Welsh and Breton ' bar ' whence English
bar ') '
Mynet ir Ile yr oedwn i o honaw ef pan ym byr-
yawd ar par
a gadel y minheu y vwrw ef a
phar.' Mab. 80.
P < ph
Pysygwriaeth (and soft bysygwriaeth) frequently in
Llyfr y Tri Aderyn, pp. 228, 236,
cér < Eng. gear:
A 'ch holl ger eraill o'ch 01. '—Bardd Czosc, 42.
crand (colloquial) < Eng. grand
croesaw < oroesaw,
See also under Ch in Chapter Il.
ff- < f-, partly on the analogy of ll- < I:
fferyllt, fferyllydd < L. Vergilius ( = Vergil, as the
author of the Georgics: then —
herbalist,' and finally
chemist ') .
ffacbys < Eng. vetches, by popular etymology.
mol—if from English vial, viol.
|
|
|
|
|
(delwedd F6459) (tudalen 100)
|
100
WELSH PHONOLOGY.
m < v (compare also Chapter Il. under ' B and M 'h •
Illywyliau ( vigils) < Lat. vigilia. A mediaeval
borrowing, and therefore a doublet of ' gwyl,
borrowed into Old Welsh.
milain < Eng. villain .
1). ab G. clxxiv.
Llefain O'r milain i 'm 01.
The provection of medial ' b,' ' d,' ' g ' has already
been discussed.
N tends to provect d into t, and it is a noteworthy
fact that no Welsh words end in -Rd, unless we except
the contracted myn'd (for nlyned), Ilon'd (for Ilonaid)
and ond for onid. Thus:
garland is in Welsh garlant ,
diamond
pound
Rowland
diemwnt •
punt
Rolant •
and onide has developed the doublet ynte by dropping
the unaccented short i.
Similarly ndd > nd
Latin benedictio > ben-ddith > bendith '
and nl > nll:
Henllan, gwinllan are for older forms where the Il
of llan had been softened due to a connecting vowel
(now lost) between the n and Il: compare Franco-
British, Anglo-French.
111. EXCRESCENT -SOUNDS.
These are due to various causes, but in the great
majority of instances they may be explained in one of
two ways.
|
|
|
|
|
(delwedd F6460) (tudalen 101)
|
WELSH PHONOLOGY.
101
(1) Certain sounds are apt to be followed by other
sounds due to careless articulation: thus if the tip of
the tongue be held against the sockets of the upper
teeth with unusual tension when Il and n are
sounded, and then rather suddenly dropped, the liquid
will end in a d or t sound. The salne result follows if
a little of the breath required for Il or n is left unex-
pended when the tip of the tongue is removed. This
is such a common practice that illustrations I-nay be
found in many languages, as English cinder from Latin
cineris, and thunder from older thuner, thunor.
Hence the t in Welsh fferyllt from Latin Vergilius,
and in ' dallt,' tlle regular pronunciation of ' deall ' in
Anglesey. Hence also the t in lanto, a doublet of
Ifano, from Ian or Ifan, with the diminutive or endear-
ment suffix -o.
(2) The other leading factor in the development of
excrescent sounds in Welsh is false analogy and mis-
Thus initial g in soft mutation
taken etymologv
disappears, e.g.
ei ard(l.
gardd
Hence ' ardd ' must be referred back to ' garcld ' as
its radical forni. And by analogy Illany words which
never had a g are now written with this initial, thus
od —English odd) is novv
odidog ' ( < odid, strange,
godidoe,' and English honest has become ' gonest ' in
Welsh. Of course, this g- is not always sanctioned in
thus the Gwynedd ' gaddo ' (for addo,
literature .
addaw) has no place in our classics, though William
Morris in one of his letters (v. Morris's Letters, p. 204)
uses gadd awiad for addewid. And it is doubtful if
|
|
|
|
|
(delwedd F6461) (tudalen 102)
|
102
WELSH PHONOLOGY.
the persistent attempt made throughout the 19th
century to write wyneb with a guttural initial will re-
ceive much further support. In our earlier works the
word is regularly ' wyneb,' as
Nid wyneb-prvd sarrig, sur, cylilylog.
Lly/r y Tri Aderyn, 251.
The ' w
in ' gwlaw ' (for ' glaw ') on the analogy
of ' gwlyb ' is quite a modern innovation, and is con-
demned by some of the best of our present-day writers.
It is scarcely likely that any reader will seek to prove
the antiquity of the ' w ' in ' gwlaw ' by reference to
the poems of Dafydd ab Gwilym, as published by
Owen Jones and Dr. William Owen Pughe, for the
cynghanedd proves conclusively that the ' w
is an
editorial emendation:
Cawrlais udgorn y curlaw. '—xliv.
Again final / is regularly dropped ill all the dialects'
and tref, eithaf, &c., are pronounced tre, eitha. But
when another syllable is added the / is retained, thus
trefydd, eithafoedd.
On the analogy of these and a host of others, words
that never ended in / are apt to have it inserted where
otherwise a hiatus would arise, thus llw* is in the plural
llyfon, llyfau, as well as llwon.
Ein Iliaws g•vv ar yn Ilesghau,
Alli oer lefain Inawr lyfau.' -Edward Morris, Llzc•on
O/er (v. Mr. M. dwards' Edition).
* It may be argued that as llw (Irish, luge), Ile (compare
Latin locus), and bro (for brog), all had a g originally, the f
is nothing but the guttural advanced into a labio-dental: com-
pare Aryan qu > Welsh p. Probably both theories are correct,
and one principle served to reinforce and direct the other.
|
|
|
|
|
(delwedd F6462) (tudalen 103)
|
WELSH PHONOLOGY.
so Ile makes Ilefydd as well as Ileoedd,
and bro > brofydd and broydd.
103
A few other instances of excrescent or inorganic
sounds (and corresponding symbols) are here added:—
dd in iddo ( < i + o), ganddo (?) , iddynt, onaddynt,
and in the Gwentian idd (
to). Idd is a bungling back
formation from ' iddo,'
iddi,' where the ' dd ' has
grown between the preposition and the pronoun. It
would be equally correct to say ' ddo ' for ' o ' or ' fo,'
and ' ddi ' for hi. In Gwentian, and to some extent in
Demetian, ' mae hi ' is pronounced ' maedd i.'
d in cefndraff < cefn + rhaff.
g: (i.) initially, see chapter ii.
(ii.)
Inedially between two vowels, as
oddigerth < oddieithr:
ar i g61 i '—Gwentian, and ' ar i giol i
Demetian for ' ar (e)i hol hi.'
(iii.) Finally in words ending ill -in, as
dwsing < Eng. dozen;
dirfing < durfin:
pring < prin
filoring < fflorin
141ading < Eng. Latin, as
Lladingiaith. '—Edward Samuel: Gwirionedd y
Gre/ydd Gristionogol, xvii.
n, especially after r, as arnaf ( < ar on + fi), canwyll-
arn, 'lialwrn and colloquial pinswrn, siswrn, &c. ( < Eng.
Pincers, scissors).
r in Ilewyrch for Ilewych, caprwn ( < Eng. capon).
|
|
|
|
|
(delwedd F6463) (tudalen 104)
|
104
WELSH
t in ysgyfaint from
parallel in idea to tfie
same oroans.
w in uwch, lluwch,
PHONOLOGY.
ysgafn, light,
English word
'The word is
lights ' for the
y-, e- ill words beginning with s + consonant (v.
above).
Epenthetic or parasite vowels frequently develope
themselves before the vowel-like consonants r, l, n, 11b
at the end of a monosyllable, if another consonant of
In this way the liquids
less schall or voice precedes.
acquire nuore syllabic value.
The excrescent vowel is generallv that of the pre-
ceding syllable repeated, as
brodor
cafall
eagal
congol
hagar
POb01
ofon (ofan) <
and ffwrwnm
brawdr (whenee bra wd)
cafll
cagl
congl
hagr
POb1
ofn
English form
Vn rhodd a welsoch Illewn un Ile,
Rhyw gongol gre rhag angeu ?
Bardd ctesc, 80.
Not infrequently the vowel is
Y,' and this is es-
pecially the case when the preceding vmvel sound is
diphthongal, ending in ' y:
awdyr < awdr < Latin, auctor.
|
|
|
|
|
(delwedd F6464) (tudalen 105)
|
WELSH
The ' w' in ' awdwr,'
&c., is of recent growth,
PHONOLOGY.
105
creawdwr, '
ymherawdwr, '
due to the analogy of (g)zvr as
a termination of the agent in such words as garddwr,
prynwr.
Maxen wledic oed amherawdyr yn ruuein. '
Mab. 82.
< brawd(r) (' brodyr ' is now used as
brodyr
the plural)
brwydyr < brwydr
llwybyr < llwybr '
neidyr, neidar < neidr.
Failing this excrescent vowel—and it is not much
countenanced ill present-day literature—the tendency
is—either (1) to drop one of the two consonants, as—
banal < banadl:
Teg fydd, pob dydd, dedwydd dål
Cnawd bun dan y cnwd banal.'*—Dafydd Nanmor.
brawd for brawdr: and colloquial
anal for anadl.
Compare Irish anal W. anadl, cenel IV. cenedl.
Or (2) to seek, by metathesis, an easier sequence, e.g.
oddigerth < oddieithr: and colloquial
ewyrth < ewythr.
Welsh has an even stronger objection to ending a
word in a mute + s, so much so that no word of native
If any such words
origin closes with that sequence.
are borrowed, a parasite vowel is regularly inserted
before the ' s,' as
Quoted by Dr. Silvan Evans, s.v. Balladl.
|
|
|
|
|
(delwedd F6465) (tudalen 106)
|
106
WELSH PHONOLOGY.
bocys < English box ,
bribis, bribys < English bribes, * which originally
meant fragments, from the root of Welsh ' briw '
cocos < Middle English cocks, of which cockles is
a diminutive:
mgys < English figs ,
hocys < English hocks, plural of hock, a species of
mallow:
micas, sop, bread and milk < English mix.
Bribis occurs in Bardd Cwsc, 81 .
Nid yw'r bythol betheu mawr,
I 'th dyb di nawr ond bribis.'*
IV. LOST SOUNDS (AND LETTERS)
This part of the subject would require several chap-
ters to deal with it exhaustively, but sonle of the chief
facts may be very briefly referred to.
(1) Short unaccented vowels are apt to be dropped:
compare
tragwyddol with tragywydd
cynlleidfa with eynnull
awdurdod from Latin auctoritateme
(2) t in the 3rd plural of Verbs and Pronominal Pre-
positions is everywhere dropped colloquially and often
even in literature, e.g.
Dy frawd O'r dwfr a oedir,
A'th nai da, aethon' i dir.
* Cf. note by Prof. Morris Jones to the word in Bardd Cwsc,
|
|
|
|
|
(delwedd F6466) (tudalen 107)
|
WELSH PHONOLOGY.
Gwreichionen a gaed o honyn'
1 07
Gwniwyd wyth bwyth ymhob un. '—G. B. C. 236.
Pan doethan dracheuyn. '—Mab. 86.
Compare -nt dropped in 3rd plural of French verbs.
So also in a fe.w other words, as
arian < ariant, dyffryn < dwfr + hynt.
(3) Final -f: see above under Excrescent Sounds.
(4) n before s is liable to disappear as in Latin (com-
pare Latin Ordinal Numerals), as, yswirio < insure (the
change of wr to wir is by popular etymology through a
desire to connect the word with ' gwir ' ).
cysuro < L. consolor ,
pwys < L. Pensum ,
and in the dialect of Gwynedd Jones is J6s.
i fyny for i fynydd: eiste for eistedd, &c.
A characteristic of Cardiganshire Welsh.
(6) 1110 < ddim o ,
dyma, clyna, < gwel di _vjna, gwel di yna,
V. METATHESIS.
The transposition of letters, sounds, or syllables in a
word is a fanliliar fact in all languages.
The liquids especially, 011 account of their fluidity,
or perhaps we should say porosity, are very liable to
interchange position with other sounds. But other
sounds also are transposed to a less degree. Metathesis
accounts for
brown by the side of burn in English
third and Riding (i.e. thirding) in Yorkshire.
|
|
|
|
|
(delwedd F6467) (tudalen 108)
|
108
WELSH PHONOLOGY.
The following are some of the most interesting in-
stances in Welsh
creddyf for crefydd, eounnonly used as late as the
end of the 18th century ,
dierth for dieithr, regularly in Lly/r y Tri Aderyn ,
see also Ffvdd, 117
dremygus for dirnmygus .
Hwy a ledasant ei safnau yn fy erbyn, tarawsant fy
-nghernau yn ddremygus.' Llyfr Job, 37.
dryntol from dwrn + dolen
Cyrraedd dol clryntol y clrws.
G. O. , Calendar y Carwr.
ewyrth for ewythr •
dogyn and digon in Nlabinogion
Ilewni and Ilenwi ( < llawn)
Ac ny welei nac Ylll_vl nac eithaf yr lluoed yn Ilewni
yr heolyd. '-—Mab. 175.
oddieithr and oddigerth
pylgain and ply-gain .
Hir bylgeint gwydyon a gyvodes.
sidan and English satin:
tangneddvf regular-ly in early Modern Welsh for
tangnefecld •
'A gwir Dcluw r danghneddyf ar goleuni fvdd gvdath
di ac ynot ti.' Lly/r v Tri A dervlt, 265,
tlfydd and llddyf
iwch < iwehi < ichwi •
Similarly the colloquial or clialeetal
blorio for brolio, in Carnarvonshire
|
|
|
|
|
(delwedd F6468) (tudalen 109)
|
WELSH PHONOLOGY.
clasgu for casglu, in Gwynedd and in (hventian
eyrnhoi for crynhoi, in Gwentian and Denletian
chwaler for chwarel, in Anglesey ,
drychynllyd for dychrynllyd, in Gwentian and
Denletian
pyrnu for prynu, in Gwentian and Dellletian
Llawner and Llafner for Llanfair, often heard iu
Carnarvonshire.
Examples need not be further multiplied. The reader
will no doubt be able to add many from his own ob-
servatlon«
VI.- ACCENT.
Definiti011: Accent, often called the syllabic accent,
is the special stress of voice laid upon a particular syl-
lable in a word, as English man' y, impos'sible, and
Welsh dag'rau, tears: gallu'oz, Powerful.
N.B. Accent Illust be distinguished frotll ernphasis,
sometimes called thought-accent, which is the special
voice stress laid upon a word in a sentence in order to
express the Ineaning of the latter, e.g. in
Myfi a'i gwnaeth.
It was I that did it.
The emphasis is on the whole word Illyfi, but the
accent of that word is on the ' fi,' thus lily-fi'
At the dawn of the Mediaeval period in Welsh, the
final unaccented syllable had disappeared, and the
accent had not been drawn back from retnaining ultima.
|
|
|
|
|
(delwedd F6469) (tudalen 110)
|
110
WELSH PHONOLOGY.
It was during that period that the y of the penult and
earlier syllables, being unaccented, acquired its obscure
y (neutral 9) sound.
Position of the accent in Modern Welsh: in words of
more than one syllable it is almost invariably on the
penult, or last syllable but one.
Examples: dag'rau, tears ,' gallu'-og, able: gall-
uog'-i, enabling.
Exceptions: The following are the most important
exceptions. Words accented on the last syllable
(a) Verb nouns in -oi ( < o-i) au and hau ( < ag-u)
and derivative nouns in had, for these are dissyllabic
in origin.
Examples: osgoi, avoiding: coffau, reminding: byw-
hau, enlivening: glanhad, a Purifying.
(b) The reduplicated pronouns—myfi, tydi, efe, nyni,
chwychwi, hwynt-hwy.
(c) Certain dissyllables beginning in ys-, ym-: e.g.
ysten', a can ystorm', a storm: ymlyn', adhere: ym-
gudd', he hides, drachefn': but note ysten'-au, cans ,
ystorm'-ydd, storms: ymlyn-u, adhering: ymgudd'-io,
hiding: according to rule.
N.B.—(i.) Borrowed words often retain their original
accent, as, ' philos'ophi.'
(ii.) The accent in diphthongs falls upon the leading
vowel:
e.g. gw'ydd, a goose.
(iii.) The following have no accent apart from the
word to which they are attached '
a, yr, y, ydd, ys, fy, dye
|
|
|
|
|
(delwedd F6470) (tudalen 111)
|
NEWPORT, MON,:
JOHN E. SOUTHAI,L, DOCK STREET.
|
|
|
|
|
(delwedd F6471) (tudalen 112_i)
|
SUBJECTS
INDEX
AND THE MORE
EXAMPLES.
PAGE
IMPORTANT
A, clear 2, 3: neutral 2, 3',
a-affection 21, 25',
a. and e
interchanged 12-13: a >
e 22', a and o interchanged
16-17; a and y inter-
-an
Analogy 54, 57, 80, et passim
annwn
Aorist
ar (Breton) the
a'r a, ar a < O'r
as, Pronoun...
Assimilation
athir
-au < Med.
dialects
au < aw
aw > ow
awch your
13
changed
a, Relative
Aberffraw, y Berffro
Ablaut
Abraham, y Vreham
ac
Accent
-ae- < -ai-
ae, in dialects
Adjectives
afagddu
afory
. 8-12
8
9
15
9
40, 96
109-110
27
30
58
10
6
-ai < Mediaeval -ei 12: How
pronounced
30
Alphabet: Association Pho-
netique Internationale 3
-am, Mediaeval Aorist 1st
plural termination .
17
Amwythig, y Mwythig
an our
an- < en-
9
8
13
B 31: B and m
Bala
balaen
basai, baswn
bodo
bodrwy
bore
breichled
broder
9
16
a
10
77
24, 25,
7
-eu 12
30
52
17
8, 26
3
33
15
33
34
13
|
|
|
|
|
(delwedd F6472) (tudalen 113_ii)
|
INDEX.
Brydain
Brythonic
bwytal, bitail
bydwraig
C 35: intervocalic 73, 95 ,
72, &c
PAGE
98-99
25, &c.
34, 97
34
-dab, -deb
dafod
PAGE
12
19
Demetian 17, 18, 29, 30, 39,
41, 109, et passim
deuthum, deuthym
23
I Dialects, dialectal 29, 30, 53'
palatal 36; velar
final 40: c and t
Caerdydd, Caerdyf
36:
di + vowel
diawl
et passim
46-7
47, 52
cafod
can as
cawad
Celtic
7'
Chinese
chords, vocal
ciath
clowem
cw-
cyn as
Ch 4, 40, 41-45: th
41, 42, 43: ch and
chw-
chwech
chwi
49
10
65
1
36
20 1
10
and h
41, 43
43
41
Dd 47;
excrescent 103
lost 107; dd and th inter-
changed
50
E, close 2, 4; neutral 2, 6;
open 2, 4', very open 2,
4: in dialects 29,' eu-
phonic 71: not always
written 72: e and y inter-
D 37, 38, 39, 45-7; dandg
39, 40, 45', d > t 96, 97;
dt > s 45-6: excrescent
103
changed
e, his, her, its, their
efo, efog
el
-ei and -ai
English vowel sounds
-en
est (Latin)
Etymological School
-eu and eau .
-eu < -ew
euthum, euthym
14
18
7
12
46
72
12
52
23
|
|
|
|
|
(delwedd F6473) (tudalen 114_iii)
|
INDEX.
PAGE
F, Ff, 47-53', F lost 102,
in Ilefydd 102-3:
107
f and dd
-fand -th
fadar
fal..,
fferyllydd
48-50
49
7
99
French 2. 54, 55, 60, 61, 62,
63, et passim
G, palatal 36; prosthetic
53-5: velar 36: excrescent
103: passing into a voiced
o- and d inter-
spirant 54: b
l, close 2, 4: open
lanto
-iff. -ith termination
sing, Present Indic.
lorddonen .
Isandlhwana
J > si 62, > s
Jerusalem
Jones, Sir William
iil.
PAGE
101
of 3rd
49
63
74
64
62-3
63
changed
glaw
godidog
gormod
Gothic
Gough
Gradation of vowels
gwelåu
39, 40
102
101
47
7
45
41
11
L, L], 4, 64: l- provected
64', analysis of the Il sound
64: l, n, r interchanged
64-7
18
Laws of Hywel Dda .
lladmer
Lam peter
Latin
liaison
Liquids
lluwch
...61,
Gwentian 17, 18, 36, 38, 42,
47, 58, 109, et passim
gwneuthum, gwneuthyn) 23
Gwynedd 18, 29, 30, 38 et
passim
accentual 56: prosthetic
Llwyd, Edward
M, see B.
M abinogion...
machlud
69
68
46
65 et seqq
76
64, r 07
26
38
72
18
57-58: hw and wh
Howel, Hywel
42
20
Mediaeval Orthography 44-5
|
|
|
|
|
(delwedd F6474) (tudalen 115_iv)
|
INDEX,
PAGE
Mercian
19
Merioneth coast dialect
36
Metathesis
107-109
Midland dialect
19
nuswrn
35
modrwy
32
modryb
moled
33
Montgomeryshire dialect 36
no, nog
-nt in French verbs
PAGE
[9
107
O, close 2, 5; open 2, 5
o and a interchanged
O, diminutive suffix
obscure a, obscure y
33
7
-om and -am in verbs 16, 17
-ow < yw
20
must
48
Mutation, extended by anal-
ogy 80: same as assimila-
tion 77: aspirate or spirant
77, 79, 91-2: in composi—
tion 78-9; of initial conso-
nants Chap. Ill.; nasal 77,
78, 92-3: soft 77, 78, 81-91:
used for syntactical pur-
poses 80: table of initial
mutations 77; of vowels
P: Indo„Germanic 7; inter-
vocalic 73, 95; palatal
36: velar
pa
palatal
Parasite vowels
po
Proclitics
36, 69
3
98
17
7
mywyliau
20-25
100
Provection
57-8, 96-100
Pronominal Prepos'itions
N, 65, 67, 68; > m 68;
l, n, r, interchanged 64-67:
excrescent 103: lost 107
Prydain
PY, pa
Q,
na, nag
natural a sound
Ned
Nel
neutral vowel sound
69
19
7
roo
33
33
35'
2,
R 70:
excrescent
grasseye
l, n interchanged
r,
Reversion
. 98-9
11
36, 69
103
54
26
|
|
|
|
|
(delwedd F6475) (tudalen 116_v)
|
|
|
|
|
|
(delwedd F6476) (tudalen 117_vi)
|
INDEX.
PAGE
Rounded vowels
rhediff, rhedith
49
S 70: > h 72-3: s + con
71
sonant
Salesbury, William
7
63
Sermania
sh (English) > si 60-1: > s
63
65
'Siamese
softening of consonants
94-5
63
som
Subjunctive, Provecti01i in 57
PAGE
-unt in Pronominal Preposi-
tions > -ynt
50
unrounded vowel sounds
2
-ut, -ud, > yt > it, in verbs
Suddas
63
uwch
Vaughan
Venedotian ,
Verbs
Vocal chords
Vocal organs
voice
22, 23
26
45
18
22-3, 49
1
1 et seqq
T 1, 73: excrescent 101, 104:
intervocalic 73, 95: in 3rd
plural of Verbs and Pro-
nominal Prepositions 106-7
timbre
tlawd
tlws
affection
Vowels, Chap. I,:
or mutation of
21-6:
changes in penultimate
table of 2:
syllables 21
simplification of 21
Tongue 1, 2, 3: back
front
blade of 2 •
middle of 2: top of...
of 2
of
U, close 2, 6; open
French u 6;
23: in dialects
2, 6; not
38
38
2
2
74
22,
30
dialects
W, close 2,
sonantal
intrusive
wh > chw
wist
wsnos
Wstrws
29-30
6, 20,: con-
74: open 2, 6;
26, 44, 105
43
46
46
46
|
|
|
|
|
|
VI.
INDEX,
PAGE
Y, clear 6: euphonic 71,
not always
106:
105,
written 72: obscure 7 ,
values of
y and a interchanged
8-12
y and e
y, Possessive
Y, Relative
ych, your
1 15
21-22
20
14
8
8
y fagddu
y fory
yn, our
yntau, yntef . , ,
yr, the
yr = er
yr awron
Y Vreham
ys, pronoun
y w > ow
PAGE
10
9
8
52
9
14
16
9
10
20
|
https://translate.google.com/
B5237: 
yn 
aith δ δ £