A Welsh Grammar, Historical and Comparative/Accidence2

by John Morris Jones

Verbs

§ 171. i. (1) The Welsh verb has three moods, the indicative, the sub­junctive and the imper­ative.

(2) The indicative mood has four tenses, the present, the imperfect, the past (aorist or perfect), and the plu­perfect.

(3) The subjunctive mood has two tenses, the present and imperfect.

(4) The imperative mood has one tense, the present.

ii. (1) The pres. ind. is often future in meaning. In the spoken language the future is the usual meaning; the present sense is retained only in a few common verbs such as gwelaf ‘I see’, clywaf ‘I hear’, medraf ‘I can’, tybiaf ‘I think’. (Ordinari­ly the present meaning is expressed periphras­tically.)

(2) The impf. indic. is seldom a mere impf. in meaning; usually it expresses Eng. ‘would’ or ‘could’.

The impf. is derived from the Ar. optative, and preserves its original meaning. It is used now in spoken W. as it is used in Homer and the Rig-Veda. Taking Meillet’s examples (Intr.² 193): Vedic kāmáyeta rā́jā samrā́ḍ bháviṭum ‘a king would like to be a supreme ruler’ = W. carai brenin fod yn benadur, cf. Mi wn ple mynnwn fy mod D.G. 501 ‘I know where I should like to be’, Mynnwn, pe nef a’i mynnai do. 288 ‘I would, if heaven would, [that ...]'; χερμάδιον…ὃ οὐ δύο γ’ ἄνδρε φέροιεν, E 303 = W. maen … ni chodai deu-ddyn, cf. Ni thynnai saith einioes hwn T.A. A 14975/107 ‘seven (men) could not take his life’; θεός γ’ ἐθέλων … ἀμείνονας … ἵππους δωρήσαιτο, K 556 = W. rhoisai (plup.) duw ewy­llysgar well meirch; Vedic yát páceyuḥ kravyā́daṃ kuryuḥ = W. pes pobynt gwnaent [y tân] yn gnawd-ysol (carni­vorous), etc. It denotes a possible or hypo­thetical as opposed to an actual thing; cf. O na welwn Wms. 508 ‘Oh that I am unable to see’ i.e. would that I saw! The impf. use comes through forms like gwelai ‘he could see’ > ‘he saw’, as in ef a welei lannerch … ef a welei carw etc. w.m. 1. The form oeẟ ‘would be’ w.m. 17, l. 29, has passed over entirely to the impf. sense, and forms peri­phrastic impfs. in the spoken lang., which does not use the impf. of other verbs in that sense. In speaking, we do not say fe safai ’r dref or y bryn ‘the town stood on the hill’ as the expres­sion of a fact, but we do say fe safai Dafydd yn segur am oriau ‘D. would stand idle for hours’ express­ing a possi­bility; we say fe welai rywbeth ‘he saw something’ (could see), but not fe safai yno ‘he stood there’ (was standing).

(3) The past is in the vast majority of cases aorist in meaning, as it is pre­dominant­ly in deri­vation. It may however have a perfect meaning, as some verbs have perfect instead of aorist forms, as treuliais fy nghlod D.G. 138 ‘I have spent my repu­tation’.

(4) The plup. ind. is very rarely plup. ind. in meaning; it usually means ‘would have’, ‘could have’, etc.; see (2).

(5) The pres. subj. in a principal sentence expresses a wish. In a dependent sentence it expresses a general, as opposed to a par­ticular, contin­gency; thus doed a ddêl ‘come what may come’, as opposed to y byd a ddaw ‘the world which will come’.

(6) The impf. subj. is used in dependent clauses only; it either stands in the protasis before the impf. ind., or repre­sents the past of the pres. subj.

The uses of the tenses can only be dealt with fully in the Syntax.

iii. (1) Each tense is inflected for the three persons of the sg. and pl.

(2) Each tense has in addition an impersonal form, whose implied in­definite subject means ‘some one, some, they’, Fr. ‘on’, Germ. ‘man’; as dywedir ‘they say, there is a saying, on dit’.

The impersonal form is generally spoken of as a “passive”; but as it takes after it pronouns in the accu­sative case, it cannot be parsed as a passive. Thus fe’m cerir or cerir fi ‘on m’aime’ (not *cerir i ‘I am loved’). The older gram­marians pretended to inflect it for the different persons by adding accu­sative affixed pronouns § 160 iii (1); as cerir fi, cerir di, cerir ef, etc., though Dr. Davies confesses that “omnia verba passiua ad naturam imperso­nalium quam proxime accedunt” D. 101. It has been argued that a substan­tival object has a soft initial, as gwêl ẟyn ‘he sees a man’; but this is a late use; the soft is rarely found after the 3rd sg. in Early Mn. poets. It arose to distin­guish the subject from the obj., but in the case of the im­personal there is no ambiguity. Intran­sitive verbs including the verb ‘to be’ are frequent­ly used in the imper­sonal, and the forms are not felt to be in any way different from tran­sitive imper­sonals except that a trans. verb requires an object: cychwyn­nir am ddau ‘a start will be made at two’.

The impersonal with its object is generally most con­venient­ly trans­lated into English by a passive with its subject, thus cerir fi ‘I am loved’; but this should not blind us to the con­struction in Welsh.

iv. (1) Each verb has also a verbal noun and most have verbal adjec­tives.

(2) The verbal noun is not strictly an infinitive; it governs the genitive, not the accu­sative, case. It may be used, like an abstract noun, with the article or an adj., as the subject or obj. of a verb or the obj. of a prepo­sition; but it is suf­ficient­ly distinct from an ordinary abstract noun by reason of certain construc­tions in which it cannot be replaced by the latter. See e.g. § 204 ii.

(3) Verbal adjectives are used like ordinary adjectives, and have not developed the peculiar uses of parti­ciples.

The Regular Verb.

§ 172. i. The regular verb caraf ‘I love’ is conjugated as follows; Ml. forms are given in spaced type:

Indicative Mood.

Present Tense.

Ml. W.				Mn. W.
	sg.		pl.		sg.		pl.
1.	karaf	1.	karwn	1.	caraf	1.	carwn
2.	kery	2.	kerwch	2.	ceri	2.	cerwch
3.	kar	3.	karant	3.	câr	3.	carant
Impers. kerir				Impers. cerir

Imperfect Tense.

Ml. W.				Mn. W.
	sg.		pl.		sg.		pl.
1.	karwn	1.	karem	1.	carwn	1.	carem
2.	karut	2.	karewch	2.	carit	2.	carech
3.	karei	3.	kerynt	3.	carai	3.	cerynt, carent
Impers. kerit				Impers. cerid

Aorist Tense.

1.	kereis	1.	karassam	1.	cerais	1.	carasom
2.	kereist	2.	karassawch	2.	ceraist	2.	carasoch
3.	karawẟ	3.	karassant	3.	carodd	3.	carasant
Impers. karwyt				Impers. carwyd

Pluperfect Tense.

1.	karasswn	1.	karassem	1.	caraswn	1.	carasem
2.	karassut	2.	karassewch	2.	carasit	2.	carasech
3.	karassei	3.	karassynt	3.	carasai	3.	carasynt, ‑ent
Impers. karassit				Impers. caresid, -asid

Subjunctive Mood.

Present Tense.

1.	kar(h)wyf	1.	kar(h)om	1.	carwyf	1.	carom
2.	ker(h)ych	2.	kar(h)och	2.	cerych	2.	caroch
3.	kar(h)o	3.	kar(h)ont	3.	caro	3.	caront
Impers. kar(h)er				Impers. carer

Imperfect Tense.

1.	kar(h)wn	1.	kar(h)em	1.	carwn	1.	carem
2.	kar(h)ut	2.	kar(h)ewch	2.	carit	2.	carech
3.	kar(h)ei	3.	ker(h)ynt	3.	carai	3.	cerynt, carent
Impers. ker(h)it				Impers. cerid

Imperative Mood.

Present Tense,

1.		1.	karwn	1.		1.	carwn
2.	kar	2.	kerwch	2.	câr	2.	cerwch
3.	karet	3.	karent	3.	cared	3.	carent, -ant
Impers. karer				Impers. carer

Verbal Noun.

karu; caru

Verbal Adjectives.

karedic, karadwy; caredig, caradwy.

ii. Stems ending in ‑i̯- (Ml. W. ẏ) drop the i̯ before i, y, and u; as rhŏ́di̯af, rhŏ́dir, rhŏ́dynt, rhŏ́dut; sŏ́ni̯af, sŏ́nnir, etc.; and stems in ‑w̯- drop the w̯ before w as cắdw̯af, cắdwn; see § 36 i, ii.

Notes and Additional Forms.

§ 173. Pres. Ind.—i. In the 2nd sg. and pl. and the imps., a and aw in the stem are affected; thus Mn. W. teli ‘thou payest’, ceni ‘thou singest’, gw̯randewi ‘thou listenest’, gw̯randéwch for gw̯randéw̯wch ‘ye listen’, gofelir ‘care is taken’, amcenir ‘an attempt is made’.

ii. The 1st sg. ends in ‑af; in b.b. written ‑aw, as dywedaw, kyuodaw, credaw 82 (≡ dywedaf, cyfodaf, credaf); in O.W., ‑am (≡ aṽ), as ni choilam ox. 22b ‘I do not believe’.

Traces of an ending ‑if occur: gwneif B.A. 1 ‘I will do’, kuynhiw b.b. 100 ‘I complain’, kynn mudif lle r.p. 1037 ‘before I change [my] place’, cenif, dygif m.a. i 191 ‘I sing, I bring’; cf. Bret. fut. kaninn.

iii. (1) The Ml. ending ‑y of the 2nd sg. remains in D.G., see ceny 186, rhedy 132, and is sometimes met with later; but in Mn. W. generally it became ‑i, see § 77 iv.

(2) The Ml. ‑y itself seems to be for ‑yẟ = Bret. ‑ez by loss of ‑ẟ, § 110 iv (3); the form ‑yẟ occurs in Early Ml. verse: ymwaredit b.b. 19 (where ‑it ≡ ‑yẟ) ‘savest thyself’, digonit ib. ‘commit­test’, guneit do. 23 ‘makest’; roẟyẟ b.t. 57 ‘givest’, mal y kynnullyẟ yt wesceryẟ ib. ‘as thou gatherest thou scatter­est’. Before di the ‑ẟ was lost early: nerthiti ox. gl. hortabere, Ml. W. nerthy di ‘thou streng­thenest’ (cf. diwedyẟ for *diweẟ-dyẟ § 110 iv (2)).

(3) In some expres­sions in common use forms without an ending occur; thus beside wely dy yna w.m. 36 we have wel dy yna r.m. 23 ‘seest thou there?’ wel dy racco w.m. 59, etc. § 221 iv (2). So os mɥn di E.M. 93 ‘if thou wilt’; and dial, fɥ́n di ‘wilt thou?’ glɥ́w di ‘dost thou hear?’ Without di we have os mɥ́n D.G. 113 ‘if thou wilt’.

iv. (1) The standard form of the 3rd sg. has no ending. The vowel of the stem undergoes the ultimate i-affection § 83 ii; thus daliaf ‘I hold’, deil ‘holds’;—archaf ‘I bid’, eirch ‘bids’;—galw̯af ‘I call’, geilw̯;—safaf ‘I stand’, saif;—paraf ‘I cause’, pair, peir;—gannaf ‘I am contained’, v.n. genni ‘to be contained’ (< *ꬶhn̥d‑: Lat. pre‑hendo, Gk. χανδάνω), 3rd sg. gain, see example; in Ml. W. (g)ein r.p. 1055, see vi (3) below, also geing by § 106 i (2), whence ng spread to other forms;—agoraf ‘I open’, egyr;—collaf ‘I lose’, cyll;—torraf ‘I break’, tyrr, written tyr;—atebaf ‘I answer’, etyb;—gwelaf ‘I see’, gwŷl, in Late Mn. W. gwêl;—cynhaliaf ‘I hold’, cynnail, cynneil;—gwahar­ddaf ‘I prohibit’, gwéheirdd D.G. 20;—ataliaf ‘I withhold’, eteil;—gwasgaraf ‘I scatter’, gwesgyr and gwasgar;—tawaf ‘I am or become silent’, teu, tau;—(g)adawaf ‘I leave’, edeu, gedy;—tarawaf ‘I strike’, tereu b.b. 63, tery;—gw(a)randawaf ‘I listen’, gwerendeu, gwrendy;—gosodaf ‘I set’, gesyd;—cyrhae­ddaf ‘I reach’, cyrraidd;—sorraf ‘I sulk’, syrr, written syr;—somaf (si̯omaf) ‘I dis­appoint, cheat’, sym;—dïolchaf ‘I thank’, dḯylch;—parchaf ‘I respect’, peirch b.b. 50, b.t. 17, G.Gr. d.g. 254;—arbedaf ‘I spare’, erbyd Diar. vi 34, E.P. 269 (but arbed Es. lv 7);—rhangaf fodd ‘I please’, reingk boẟ s.g. 277.

Ni ain o fewn main y mur,

Ni bu ’n f’oes neb un fesur.—T.A., a 14967/91b.

‘There is not contained within the stones of the wall, there has not been in my time, any one of the same stature.’

Am na ain d’aur mewn un dwrn.—T.A. a 14975/16. ‘Because thy gold will not go into one hand’ (is more than a handful).

A fo doeth efo a dau;

Annoeth ni reol enau.—G.I.H., tr. 87.

‘[He] who is wise is silent; the unwise does not control [his] mouth.’ Pan vynner iẟi tewi hi a teu r.m. 122 ‘when one wishes it (a certain harp) to be silent, it is silent’. fel y tau dafad Es. liii 7 ‘as a sheep is dumb’.

O syr, lle gwesgyr gwasgwyn,

O’m dawr, Gwyn ap Nudd i’m dwyn.—D.G. 246.

‘If he sulks, where he scatters [his] gasconade, G. ap N. take me if I care.’

Nid yw anair ond ennyd;

Ni sym twyll mo bwyll y byd.—E.P. 271.

‘Calumny is but [for] a while; deceit will not cheat the good sense of the world’; ny’m sym r.p. 1198 ‘will not dis­appoint me’.

I Dduw Madog a ddḯylch

Gan i chwaer hael cael y cylch.—D.G. 292.

‘Madoc thanks God that he has had the ring from his generous sister.’ Cf. 167, L.G.C. 70.

(2) In many verbs which have a, the vowel is un­affected; thus câr ‘loves’, cân ‘sings’, tâl ‘pays’, gad ‘leaves’; also in some with e, as cymer ‘takes’, adfer ‘restores’, arfer ‘uses’.

Some verbs with a have both the affected and unaffected form; thus gallaf ‘I can’, geill Ỻ.A. 169, D.G. 29, or gall E.P. 259; dïalaf ‘I avenge’, dïail D.G. 162, G.G1., p 108/41 B., dḯal L.Môn § 186 ii; chwarddaf ‘I laugh’, chweirẟ r.p. 1240, chwardd D.G. 402, L.G.C. 379, Job xli 29, Ps. ii 4; barnn Ỻ.A. 64 ‘judges’, beirn r.p. 1321.

(3) The vowel of the stem, if mutable, is of course mutated when the ending is dropped in the 3rd sg.; thus cỿrchaf ‘I make for’, cɥrch; dỿgaf ‘I bring’, dwg ‘brings’; ceisiaf ‘I seek’, Mn. W. cais ‘seeks’, Ml. keis, § 81 iii (1). In many cases ‑o- is a mutation of ‑aw-, the latter appearing in the 3rd sg., thus toddaf ‘I melt’, tawdd ‘melts’; boddaf ‘I drown, or am drowned’, bawdd ‘drowns’; holaf ‘I ask’, hawl ‘asks’. But in di­syllabic stems, when the un­accented ‑aw- in the ult. became ‑o-, it was in some cases treated on the analogy of original ‑o- and affected to y; thus adroẟaf ‘I narrate’, adrawẟ b.a. 1 ‘narrates’, later edryẟ r.p. 1253; halogaf ‘I defile’ (denomi­native from halawc, Mn. W. halog), helyc Ỻ.A. 34 ‘defiles’. In the 16th cent. dichyn was used, § 196 ii, but was supplant­ed later by the original form dichon ‘can’, Ml. W. dichawn. The substi­tution of ‑ɥ as in tery c.m. 32 for the affected ‑eu of tereu, etc., see (1), is due to the same analogy; see § 83 ii.

A similar analogy gave rise to gwerchyd D.G. 175 ‘guards’ from the v.n. gwarchod for gwarchawd metath. for gwar-chadw̯; the old 3rd sg. was gwer-cheidw̯.

v. (1) Verbs with stems in ‑ha- had the vowel unaffected in the 3rd sg.; the affected forms bwyty ‘eats’, pery D.G. 441 ‘lasts’ are late; the original 3rd sg. of bwytá-af ‘I eat’ is bwyta w.m. 456 ‘eats’, ef a vwytta Ỻ.A. 170, and of parhá-af is pár(h)a, as parha b.t. 40 ‘lasts’, ny phara r.p. 1046, w.m. 86 ‘does not last’, ni phara T.A. g. 236. The accent falls regularly on the penult; and the ‑h- after it was lost, § 48 ii, as in pára, but not before changing a media to a tenuis as in bẃyta ‘eats’ (: bwyd ‘food’).

(2) The ‑(h)a of the 3rd sg. is thus the unaffected stem-forming suffix, but it came to be mistaken for a personal ending; and as ‑ha- forms the stems of denomi­natives, ‑(h)a seemed to be a 3rd sg. ending of denomi­natives, and was used to form the 3rd sg. of denomi­natives generally. This may have origi­nated in doublets like neshá-af, v.n. neshá-u and nés-af, v.n. nésu ‘to approach’ (: nes ‘nearer’); the 3rd sg. of the first is regularly nés-(h)a, which, being very naturally taken to be the 3rd sg. of the second, suggested a 3rd sg. ending ‑(h)a. For exactly the same reasons it became a 2nd sg. imper­ative ending, and is used as such in all verbs in which it appears in the 3rd sg. pres. ind. In older examples the form is ‑ha, the ‑h- hardening a media or remaining as an ‑h‑; such examples survive in Ml. W. side by side with others in which the ending has come to be regarded as ‑a simply. Thus we find gwata r.p. 1382 ‘denies’, oetta impv. r.p. 1254 ‘delay’, gwatta m.a. i 319a ‘denies’, ehetta do. 319b ‘flies’, tremycca Ỻ.A. 150 ‘despises’, poenha do. 28 ‘punishes’, dielwha do. 147 ‘ruins’ (makes worthless), gwyhwa do. 148 ‘withers’, cerẟha do. 168 ‘goes’, gweẟha do. 165 ‘beseems’, side by side with gwada r.p. 1256 ‘denies’, oeda impv. do. 1285 ‘stay’, llettya do. 1254 ‘lodges’, ogana ib. ‘satirizes’, a gylchyna m.a. i 319b ‘surrounds’, a boena Ỻ.A. 147, kerdda do. 165, kerẟa do. 167, gw̯eẟa r.p. 1272. In the last examples simple ‑a has become a 3rd sg. ending.

(3) It is added to nearly all denominative stems which represent a noun or adj. without a suffix; thus hw̯yẟa Ỻ.A. 148 (: hŵyẟ, chŵyẟ ‘a swelling’), a ge(i)thiwa ib. (: keithiw ‘captive’), argyweẟa do. 166 (: argyweẟ ‘harm’), saetha r.p. 1272 ‘shoots’ (: saeth ‘arrow’), amcana 1285 (: amcan ‘design’), gwarchaea ib. (: gwarchae ‘fortifi­cation’), dilyssa ib. dilyssa 1254 (: dilys ‘certain’), llaessa 1254 (: llaes ‘slack’), sura r.m. 123 (: sur ‘sour’), a gospa Ỻ.A. 30 (: cosp ‘punish­ment’), gwassan­naetha do. 28 (: gwasa­naeth ‘service’), kyfvɏr­golla do. 35 (: cýfr-goll § 156 i (9)), breinia m.a. i 318a (: braint ‘privilege’), yssiga ib. (: ysig ‘crushed’), diwedda do. 3186 (: diwedd ‘end’), cynnydda 319a (: cynnydd ‘increase’), mynycha 319b (: mynych ‘frequent’), lwydda ib. (: llwydd ‘pros­perity’), a gocha r.b.b. 146 (: coch ‘red’), kyflea r.p. 1286 (: cyf-le ‘situation’), metha 1253 (: meth ‘failure’).

(4) It is also added to some stems not obviously denominative; thus cerddaf ‘I walk, go’ has 3rd sg. cerẟa in Ml. W., see examples above, and in Mn. W., see Diar. iii 28, vi 3, but a gerẟ b.t. 15; so sathra Ỻ.A. 147 ‘tramples’, but sathɏr b.b.b. 144; damuna Ỻ.A. 148 ‘wishes’ (the noun is damunet ‘wish’), traetha b.b. 8 ‘relates’ (noun traethawd ‘treatise’ < Lat. tractāt-us).

(5) It is added to stems in ‑i̯- mostly denominatives; as tykẏa w.m. 14, Mn. W. tỿ́ci̯a ‘avails’ (: twg ‘success’ < *tu‑k‑, √teu̯ā- ‘increase’) used only in the 3rd pers., § 196 v, llywẏa r.p. 1285 ‘governs’, Mn. W. llywi̯a ‘steers’ (: llyw ‘rudder’), hwyli̯a m.a. i 318a, Mn. W. hwyli̯a ‘sails, governs’ (: hwyl ‘sail’, cf. Lat. gubernāre ‘steer, govern’), cili̯a do. 319b ‘recedes’ (: cil ‘back’), rhodi̯a Ps. i 1 (: rhawd ‘course’ < *rōt‑, L°-grade of √ret- ‘run’), Mn. W. preswyli̯a ‘resides’ Ml. W. presswyla Ỻ.A. 169 (: presswyl ‘residence’), distrywi̯a (: distryw ‘destruc­tion’). But some i̯-stems do not take it: dali̯af, deil (not dali̯a), ceisiaf ‘I seek’, cais (not ceisi̯a), peidi̯af ‘I cease’, paid (not peidi̯a), meiddiaf, beiddiaf ‘I dare’, maidd, baidd, ‘dares’.

(6) It is added to denom. stems in ‑ych‑; as gwledycha Ỻ.A. 169, m.a. i 318a ‘governs’, fflam­mycha do. 318b ‘flames’, except whennych r.m. 123, chwen(n)ych Ỻ.A. 73 ‘desires’ (: chwant ‘desire’).

(7) Lastly, it is added redundantly to ‑ha- itself, as mwynhaa m.a. i 317b, Mn. W. mwynhā́ ‘enjoys’, kyt-lawenhaa Ỻ.A. 72, Mn. W. llawenhā́ ‘rejoices', dynessaa r.b.b. 148, Mn. W. neshā́ ‘approach­es’, arwy­ẟockaa do. 144, Mn. W. arwyẟocā́ ‘signifies’, Mn. W. glanhā́ ‘cleans’, edifarhā́ ‘repents’, etc., etc.

(8) A few verbs have two forms, one with and one without ‑(h)a; as plycca impv. r.m. 97 ‘fold’, plyc b.t. 18 ‘bends’ (plygaf ‘I bend’, plyg ‘fold’); tybia D.I.D. tr. 98, tyb T.A. f. 16 ‘imagines’ (tybiaf ‘I imagine’, tyb ‘thought, fancy’); a dwylla Jer. ix 5, a dwyll Ỻ.A. 147 ‘deceives’ (twyllaf ‘I deceive’, twyll ‘deceit’); gweinyẟa r.p. 1254 ‘serves’, gweinyẟ do. 1238; barn iv 2, barna Ps. cxxxv 14.

vi. (1) Sg. 3. ‑id, used where there was no preverb, is found in Ml. W., and survived in proverbs, and rarely in verse; like the fut. ‑(h)awd it became ‑(h)id; thus O. W. prinit (without ‑h‑) ox. 22b ‘buys’; Ml. W. ottid b.b. 89 ‘falls’ (of snow), meccid do. 90 ‘nourishes'; Trenghit golut, ny threingk molut r.b. 1082 ‘wealth perishes, fame perishes not’; Tyfid maban, ny thyf i gadachan ‘an infant grows, its swaddling cloth does not grow’; Dirmycid merch … ŵr ni welo G.Gr. p 77/194 ‘a woman despises a man whom she does not see’.

‑yd occurs in ë-yt (rh. with byt ‘world’) r.p. 1055 ‘goes’. It seems to be confused with ‑id in megyt, meckyt r.p. 1029 ‘nourishes’, gwlychyt do. 1032 ‘wets’.

(2) An ending ‑(h)awd of the 3rd sg. occurs in Early Ml. W. with a future meaning: bithaud (≡ byẟhawd) b.b. 7 ‘will be’, r͑eddaud (dd ≡ tt for dh) do. 58 ‘will run’, dircha­vaud do. 61 ‘will arise’, parahaud do. 100, parahawt b.t. 23 ‘will continue’, gyrhawt b.t. 13 ‘will drive’. These forms were survivals, and appear sometimes to be misused as passives under the influence of the ‑t im­personals: cluttaud b.b. 10 ‘will be brought’, briuhaud do. 58 ‘will be broken’.

(3) There are traces of a 3rd sg. in ‑yẟ, as ny wneyẟ gwir ny ein ymro r.p. 1055 ‘he who does not do justice will not be suffered [lit. contained] in a country’; kyn noc y daw rwng ẏ ẟwylaw y gwesgeryẟ do. 1049 ‘[it is] before it comes between his hands that he scatters it’. This is quite distinct from ‑yd above, and comes, as seen, after relatives.

There is no sufficient ground for the assumption, Arch. Camb. 1873 150, of a 3rd sg. ‑haw; for chaffaw b.b. 8 ≡ chaffaf 1st sg., see ii above, gwnaw sk. 126 is an error for gwnaho b.t. 16 l. 2; a wnaỽ b.t. 30 l. 18 is prob. a sc. error for wnaaỽt; the other examples are from un­trust­worthy texts.

(4) In the dialects an ending ‑iff, in Gwynedd ‑ith, is in common use. D. 85 regards it as falsely deduced from caiff, “Vt Ceriff pro Car, Periff pro Pair … Quæ nunquam sine indig­natione audio.” (As ‑iff is not a syllable in caiff the suggested deduction is im­probable.)

vii. Beside the usual ‑wn of the 1st pl., we seem to have a 1st pl. pres. ‑en once in the O.W. cet iben juv. sk. ‘we drink together’.

viii. The affection of the stem vowel in the 2nd pl. cerwch shows that ‑wch must be for ‑yw̯ch § 26 vi (5). A trace of this form occurs in chedyw̯ch Ỻ.A. 157 ‘ye keep’ dissim. for *chedw̯yw̯ch; the usual form is cedwch for cedw̯wch: cadwaf ‘I keep’.

ix. (1) Cor­respond­ing to the 3rd sg. in ‑hawt, a 3rd pl. in ‑hawnt occurs rarely in the earlier periods: cuinhaunt juv. gl. defleb(unt), gwnahawnt b.t. 13 ‘they will make’.

(2) In O.W. a 3rd pl. pres. ‑int occurs, as limnint juv. gl. tondent, scam­nhegint juv. gl. levant, nertheint juv. gl. armant. Some examples occur in the early poetry: diwris­sint kedwyr … mi nyd aw b.b. 108 ‘warriors hasten … I go not’; vyẟ … pan ẟyorf(yẟ)yn b.t. 13 ‘will be when they conquer’, discynnyn ib. ‘they will descend’.

x. ☞ The final ‑t of the 3rd pl. of this and of every other tense is often dropped in poetry, even in Early Ml. W., § 106 iii (2): tirran (≡ tỿrran) b.b. 2 ‘they muster’, dygan ib. ‘they bring’, darparan do. 5 ‘they prepare’, vidan (≡ vỿẟan) ib. ‘they will be’; other tenses: deuthan do. 2 ‘they came’, wnaethan do. 4 ‘they did’, darvuan do. 6 ‘they perished’, cuitin (≡ cwyẟyn) do. 95 ‘they fell’. The ‑t is lost in the spoken language.

xi. (1) Beside the impers. in ‑ir, a form in ‑(h)awr, cor­respond­ing to the 3rd sg. in ‑hawt, occurs in Early Ml. W.; as talhaur b.b. 31 ‘there will be payment’, ffohawr b.t. 16 ‘there will be flight’, dialawr ib. ‘there will be vengeance’, dyrehawr do. 33 ‘will be mustered’, agorawr w.m. 456 ‘will be opened’; Dygɏn yw aẟaw a garawr r.b. 1062 ‘it is hard to promise what is loved’; Heul yn Ionawr ny mat welawr, M(a)wrth a Whefrawr ae dialawr r.b. 970 ‘Sun in January is not good to be seen, [in] March and February there will be retri­bution for it’.

(2) The ending ‑(h)er has a fut. ind. meaning in Early Ml. W., as moch guelher y niuer b.b. 2 ‘soon will the host be seen’; nyth atter ti ẏ mywn w.m. 457 ‘thou shalt not be admitted’.

(3) In the early poetry an impersonal in ‑itor, ‑etor, ‑ator, ‑otor occurs: kenhittor kirrn b.b. 52 ‘horns will be sounded’, canhator b.t. 75 ‘will be sung’, megittor b.b. 62 ‘will be brought about’, r͑ewinetor b.t. 68 ‘will be ruined’, trae­thattor, molhator do. 23, bri­thottor b.b. 33 ‘are varie­gated’. Forms in ‑etawr also occur: dygetawr b.t. 10 ‘will be brought’, galwet­tawr do. 41 ‘will be called’; in these the ending has come under the influence of ‑hawr.

174. Imperf. Ind.—i. The 2nd sg. ending in Ml. W. is ‑ud, as dianghut r.p. 1037 ‘thou wouldst escape’. In Early Mn. W. this remains, as wyddud, atebud rhyming with mud in D.G. 460; but ‑ud di became ‑it ti § 111 ii, § 77 ix; hence Late Mn. W. carit. The ‑i- not being original does not affect the ‑a‑; cerit is an arti­ficial form: “secunda sing. fit etiam sine mutatione vocalis, & fortasse rectiùs, Carit” D. 89. In the dialects the vowel of the 2nd pl. is intro­duced, as caret; and this debased form occurs in recent writings.

ii. In the early poetry a 3rd sg. ‑i is found, affecting ‑a- in the stem (as well as the usual ‑ei, not affecting); thus ef gelwi b.a. 22 ‘he called’, ef lleẟi ib. ‘he slew’ (beside pan elei ib. ‘when he went’), ny cheri do. 26 ‘he loved not’ (beside ef carei ib. ‘he loved’), eiẟuni do. 16 ‘he desired’, klywi ib. ‘he heard’, a weli b.b. 45 ‘whom he saw’.

For the 3rd sg. in ‑i̯ad see § 191 ii (3).

iii. (1) The vowel of the pl. endings is ‑e‑, which regularly becomes ‑y- before ‑nt, § 65 iii (1). The intro­duction of the ‑y- into the 1st and 2nd as in hoffym Gr.H. g. 98 (for hoffem) is rare, and doubtless arti­ficial. On the other hand the ‑y- of the 3rd has tended to be replaced by the ‑e- of the 1st and 2nd since the 15th cent., e.g. nis terfynen’ L.G.C. 244 ‘they would not end him’ (usually L.C.C. has ‑yn(t): a berynt 186, a’m ceryn’ 206). In Late Mn. W. the re-formed ‑ent became the usual ending, though ‑ynt remained in use in poetry, e.g. E.F. 36, 287, 316.

(2) In Ml. W. a re-formed 3rd pl. ‑eint, with the vowel of the 3rd sg., occurs; as y wawt a ganeint Ỻ.A. 95 ‘the song which they sang’, a ẟeueint … a syrthẏ­eint do. 97 ‘which came, [tears] which fell’, achubeint w.m. 466 ‘they seized’. It is sometimes found in Early Mn. verse: an­rhegaint D.G. 24 ‘they presented’, oedaint, fyddaint, rhedaint do. 25; occa­sional­ly later: rhedeint b.cw. 23.

175. Aor. and Plup. Ind.—i. The 3rd sg. aor. has a number of endings:

(1) Ml. W. ‑awẟ, Mn. W. ‑odd, is common in Ml. W., and almost supplant­ed all other endings in the Late Mn. period. Ml. W. examples: kerẟauẟ w.m. 9 ‘walked’ (beside a gerẟwys do. 8), cym­hellauẟ do. 17 ‘incited’, wharyawẟ do. 163 ‘played’, parawẟ ib. ‘caused’ see (4), gofynnawẟ do. 164 ‘asked’, diskyn­nawẟ do. 422 ‘descended’, r͑wymawẟ ib. ‘bound’, frwyn­glymawẟ ib. ‘fastened by the reins’, dechreu­awẟ r.b.b. 117 ‘began’, dewissawẟ do. 319 ‘chose’. ‑oẟ already appears in Late Ml. W.: parhaoẟ, arveroẟ, llaẟoẟ c.m. 92 ‘lasted, used, killed’, brathoẟ do. 93 ‘stabbed’.

(2) ‑as, in cavas b.b. 66, w.m. 10 ‘got’, gwelas b.b. 101, w.m. 13 ‘saw’. It survived as the regular ending in these two verbs in Ml. W.; in Early Ml. W. other verbs take it, bradas, twyllas b.b. 81 ‘betrayed, deceived’, erects G. m.a. i 196 ‘created’, gallas B.V. do. 372 ‘could’. In cafas it survived in Early Mn. W.:

Pwy mewn gaeaf a gafas

Fis Mai yn dwyn lifrai las?—D.G. 265; cf. 116.

‘Who in winter [ever] found a May-month wearing green livery?’

(3) ‑es is added to stems having ‑o- or ‑oe‑; as dicones juv. sk. ‘wrought’, r͑otes (t ≡ ẟ) b.b. 42, r͑oẟes w.m. 9 ‘gave’, torres w.m. 94 ‘broke’, arhoës do. 47 ‘waited’, ffoës r.m. 152 ‘fled’, ymhoeles r.b.b. 199 ‘returned’. It is common in Mn. W., more especial­ly in the earlier period: ffoës D.G. 61, si̯omes G.G1. c. i 196 ‘deceived’, colles I.T. f. 43 ‘lost’, codes do. 45 ‘rose’, rhoddes Phil, ii 9 ‘gave’, torres Gr.O. 41 ‘broke’. It survives in the spoken lang. in contract­ed forms rhoes, troes.—Contrary to analogy it replaced ‑as in gwelas in Late Ml. and Early Mn. W., as gweles r.b.b. 130, D.G. 279, T.A. g. 235.

(4) ‑is is added to stems having ‑a- (which it affects to ‑e‑), or ‑aw- ( > ‑ew‑): treg̃his b.b. 21 ‘perished’, cedwis do. 43 ‘kept’, erchis Ỻ.A. 2 ‘bade’, dienghis w.m. 56 ‘escaped’, peris do. 57 ‘caused’, ettellis (l‑l, vb. atali̯af) r.b.b. 174 ‘withheld’, cynhellis (l‑l, vb. cynhali̯af) do. 257 ‘held’, edewis r.m. 169 ‘left’, eẟewis r.b.b. 171 ‘promised’. Also dechreuis w.m. 27, r.m. 17 ‘began’ (beside dechreu­wys w.m. 413, r.m. 267). It is occa­sional­ly met with in Early Mn. W., as gadewis D.G. 61.

Ni wn a fûm yn iawn fis

Heb hiraeth,—hi a’i peris.—I.D. 20.

‘I do not know that I have been well for a month without longing, [it is] she that caused it.’

(5) ‑w͡ys is perhaps the commonest ending in Ml. W.: pechuis b.b. 41 ‘sinned’, guiscvis do. 43 ‘wore’, treulwys w.m. 9 ‘spent’, cyrchwys ib. ‘made for’, meẟylẏwys do. 10 ‘thought’, diffy­gẏwys do. 12 ‘failed’, trigẏwys r.m. 92 ‘resided’, gallwys do. 108 ‘could’, mynnwys r.b.b. 200 ‘desired’. It was simpli­fied early to ‑ws § 78 i (2), as bendigus b.b. 36 ‘blessed’, ffruin­cluymus (read ‑clymus) do. 93; cerẟws p14/6 r. (mid-13th cent.) ‘walked’, claẟws p 14/14 r. ‘buried’, kemerrws p 29/31 r. ‘took’. The form ‑w͡ys dis­appeared, but ‑ws is sometimes met with in Mn. lit. W., and became the usual ending in parts of S.W.

Hadlyd liw hudol o dlws,

Hudolion a’i hadeilws.—D.G. 447.

‘Perished colour enchantingly beautiful, it is enchanters that built it.’

(6) ‑t in t-aorists, see iii (1).

ii. (1) The above are strictly stem-forming suffixes, with no personal ending, added to the pres. stem. The 1st sg. has ‑as affected to ‑eis; the 2nd sg. has the same with added ‑t; the pl. has a similar suffix, which takes three forms, to which the personal endings ‑am, ‑awch, ‑ant are added. The forms of the suffix are Ml. W. ‑ass‑, ‑yss- and ‑ss‑, Mn. W. ‑as‑, ‑s‑.

(2) ‑ass- and ‑yss- are not sharply distinguished: thus dywedassam r.m. 44 = dywedys­sam w.m. 61 ‘we mentioned’, collassam r.m. 52 = collyssam w.m. 72 ‘we have lost’, cilẏassant, torras­sant r.m. 36 = cẏlyssant, torrys­sant w.m. 52 ‘they retreated, they broke’. Both forms occur through­out the Ml. period, ‑ass- encroach­ing in later mss. as the examples show. Later ‑yss- dis­appeared, and in Late Mn. W. ‑as- alone is used.

(3) ‑s(s)- is used after ‑l- and ‑r- and after the diphthongs ‑aw‑, ‑yw‑, ‑eu‑: gwelsom w.m. 50, r.m. 35 ‘we saw’, cymersant w.m. 169 (= cymeras­sant r.m. 235) ‘they took’, adcors­sant b.b. 46 ‘they returned’, ymadaws­sam h.m. ii 292, Ỻ.A. 148 ‘we left’, clywssont w.m. 33 ‘they heard’, dechreus­sant do. 41, 72 ‘they began’, beside dechreuys­sant 44. In Mn. W. it is regularly found in gwelsom, and always after ‑aw- as gwran­dawsom; sometimes in other cases, as talsom, cymersom. In the dialects the ‑s- form became general.

(4) Beside the usual ‑am, ‑awch, ‑ant in Ml. W., ‑om and ‑ont are often found, and are specially frequent in the w.m.; ‑och is very rare: doethoch w.m. 161 (= doethawch r.m. 228) ‘ye came’. In Mn. lit. W. ‑om, ‑och, ‑ant are the usual endings. In the spoken lang. mostly ‑on, ‑och, on’.

(5) In the old poetry there are traces of the 3rd sg. ending ‑id, as in the pres. § 173 vi (1), as delẏessit Ieuan … vab Duw … yn dwfɏr echwyẟ r.p. 1184 ‘John held the Son of God in the water of baptism’ (the context shows that it is not impers. plup.), prinessit (read prynessit) ib.; also ‑yd, as keressyt r.p. 1168, pregethys­sit (/kyt) b.t. 54.

iii. (1) A 3rd sg. ending ‑t added to the pres. stem is found in some verbs, as cant b.a. i, w.m. 120, r.m. 196 ‘sang’ (not cânt as wrongly assumed by some recent copyists), gwant r.m. 81, w.m. 111 ‘pierced’; *‑er‑t- regularly becomes ‑yrth § 65 iii (2), hence diffirth, kymirth b.b. 40 ‘saved, took’, with a-affection differth r.b.b. 213, kymerth w.m. 9, see § 181 vii (1).

(2) These 3rd sg. aor. forms had come to be regarded in Kelt, as aor. stems, and other persons were formed from them, § 181 vii (1); thus keint b.t. 33 ‘I sang’, keintum w. 18a ‘I sang’, ceuntost b.b. 21 ‘thou hast sung’; gweint m.a. i 194a ‘I charged’ (in battle).

iv. (1) The impersonal, like the 3rd sg., has various endings. Verbs which take ‑as, ‑es, ‑is, ‑wys in the 3rd sg. have ‑ad, ‑ed, ‑id, ‑w͡yd respec­tive­ly in the im­personal.

(2) ‑ad in caffat r.m. 141 ‘was had’, cahat w.m. 40, r.m. 27 ‘was had', contract­ed cat r.b.b. 396, Mn. W. cad D.G. 189, etc. ‘was had’, see § 188 i (6), and in gwelat w.m. 51, r.m. 36 ‘were (was) seen’. In the old poetry it is seen in other verbs, as artuad (t ≡ ẟ) b.b. 23 ‘was blackened’ se-suinad ib. ‘was conjured (?)’ (redupl. perf.?)

(3) ‑ed, after ‑o‑, ‑oe‑: anvonet w.m. 84 ‘were sent’, collet do. 472 ‘was lost’, dodet do. 32 ‘was put’, r͑oẟet do. 33 ‘was given’ Mn. W. rhodded f.n. 28, poened c.c. 11, hoeli̯ed p 49/54b, etc.

It is also found in ganet w.m. 28, Mn. W. ganed ‘was born’; and in Early Mn. W. gweled D.E. Ỻ 163/119 ‘was seen’ for Ml. gwelat, like gweles for gwelas i (3); Ml. W. r͑ived m.a. i 373, llaẟed do. 220.

(4) ‑id, after ‑a‑, ‑aw‑: y delit … ac y carcharwyt r.b.b. 338 ‘was caught and im­prisoned’, edewit w.m. 58 ‘were left’, eẟewit r.m. 162 ‘was promised’. Only the context, as seen in the first example, shows that this is not the imperf., which ends in ‑id in all verbs. In Mn. W. ‑id aor. gave place to ‑wyd to avoid the ambiguity: dali̯wyd ‘was caught’, gada(w)wyd ‘was left’.

(5) ‑wyd, as in magwyt w.m. 33 ‘was reared’, gollyngwyt, r͑yẟhawyt do. 25 ‘was released, was set free’, cyweirwyt do. 26 ‘was prepared’, treulwyt ib. ‘was spent’, gommeẟwyt ib. ‘was refused’, etc., etc. This is the usual ending in Mn. W., and has super­seded the others except in a few forms like ganed ‘was born’, rhoed ‘was put’, etc.—Reduced to ‑wt § 78 i (2), whence dial. cawd § 188 i (6).

(6) The *‑t‑ of this suffix came without an intervening vowel after some roots ending in ‑d‑, early enough to give "W". ‑s for the group ‑dt- § 87 ii. Thus llas w.m. 89 ‘was killed’, also in Early Mn. W. and later, beside llaẟwyt H.D. p 67/277 r.; klas D.E. j 17/478 r. ‘was buried’, usually claẟwyt w.m. 89; gwŷs D.G. 236 ‘is known’.

Ef a’m llas i a’m nasiwn

Yr awr y llas yr iarll hwn.—G.G1., c. i 193.

‘I was slain and my nation the hour that this earl was slain.’

(7) Some verbs take ‑pwyd, which is generally added to the perf. or aor. stem; thus aethpwyt w.m. 59 ‘there was a going’, deuthpwyt do. 141 or doethpwyt do. 96 ‘there was a coming’, gwnaeth­pwyt do. 32 ‘was done’. In these three verbs the form persisted and is the standard Mn. form, as used e.g. in the Bible; but in Recent W., dial, and quasi-dial. forms aed, deuwyd (dial. dowd), gwnaed are also found.

Other examples are ducpwyt w.m. 28 ‘were brought’ (perf. st. dug‑ § 194 iii), gorucpwyt w.m. 452 ‘was done’, clywspwyt r.b.b. 178 ‘was heard’, dechreus­pwyt s.g. 291, canpwyt § 182 iv (4).

It is added to the present stem in dalpwyt r.b.b. 388 ‘was caught’, kynni­cpwyt do. 398 ‘was offered’, gatpwyt do. 399 ‘was left’, dywetpwyt w.m. 52 beside dywespwyt do. 189 ‘was said’.

v. (1) The pluperfect is formed by adding the personal endings of the imperfect to the aorist stem.

The impers. ‑it and 3rd pl. ‑ynt affect ‑aw- in the penult, thus adewssynt r.b.b. 180 ‘they had left’, edewssit r.m. 288 ‘had been left’. But ‑ass- usually remains un­affected: buassynt w.m. 89 (beside buessynt Ỻ.A. 19) ‘they had been’, anvonas­sit r.b.b. 306 ‘had been sent’, mynnassit r.m. 13 = mynyssit w.m. 20 ‘had been desired’, collas­synt r.m. 42 = collys­synt w.m. 60. D.G. 279 has dygesynt (if weles before it is the correct reading; if welas, it would be dygasynt) for teby­gesynt; the plup. of this verb is often synco­pated, tygaswn etc. D. 134.

(2) Some verbs have a plup. formed by adding oeẟwn, oeẟut etc. to the aor. stem: caws­soeẟwn etc. § 188 i (7), r͑oessoeẟ § 186 iii, as well as athoed etc. § 193 vi (5).

(3) An impers. of the plup. formed by adding ‑adoeẟ, ‑ydoeẟ to the pres. stem occurs in some verbs: ganadoeẟ § 197 ‘had been born’, aẟawadoeẟ g.c. 122 ‘had been promised’, mana­gadoeẟ m.a. ii 103 ‘had been mentioned’, magadoet, defny­tadoet (t ≡ ẟ) do. i 254.

§ 176. Pres. and Impf. Subj.—i. (1) The subj. stem is formed by a suffix ‑h- which is added to the pres. ind. stem and hardens a media to a tenuis; thus nottwyf w.m. 479: nodaf ib. ‘I specify’. After vowels and sonants the ‑h- dis­appears because it follows the accent § 48 ii, but it is often written in Early Ml. W. as gwnaho b.t. 16, gunelhont b.b. 60.

In Early Mn. W. the tenuis generally remained, and survived later in a few expres­sions as gato in na ato Duw ‘God forbid’: gadaf ‘I permit’. But from the 16th cent, the ind. stem has mostly been used, and the media restored, as in Dyn a godo Duw’n geidwad S.T. g.b. [375] ‘A man whom God raises as a saviour’.

(2) Some verbs have special subj. stems, as el‑: af ‘I go’, etc. § 193 vii; b‑: wyf ‘I am’ § 189; Early Ml. W. duch, gwares § 183 iii (1).

ii. The ending of the 3rd sg. pres. is ‑o: talo w.m. 9 (: talaf ‘I pay’), adnappo do. 36 (adwaen § 191), dycco do. 465 (: dygaf ‘I bear’). This is a simpli­fication of ‑oe, which survives in creddoe (dd ≡ tt < dh) b.b. 53 (: credaf ‘I believe’), see § 78 i (1). The form ‑w͡y is a variant of ‑oe § 183 ii (1), and the former not un­common­ly occurs in Early Ml. W., as gwelhvy b.b. 74 ‘may see’, achupvy do. 75 (: achubaf ‘I seize’), nottvy do. 76 (: nodaf ‘I specify’), gule­dichuy do. 59 ‘may rule’, canhwi do. 48 ‘may sing’.

iii. (1) The 1st sg. ends in ‑w͡yf: cattwyf w.m. 125 for *catw̯w͡yf (: cadwaf ‘I keep’), ymgaffwyf a ib. ‘I may meet’, etc. This is the usual form in Ml. and Mn. W. The occur­rence of ‑of is com­parative­ly rare: gwiscof w.m. 97 (= gwisgwyf r.m. 71), cysgof h.m. ii 137, gofynnof do. 260. This is probably a re‑formation from the 3rd sg.

(2) The 3rd pl. ending is ‑ont; rarely in Ml. W. ‑w͡ynt, as in elwynt B.A. 2, 3 (: af ‘I go’), and ‑oent, as pan venoent a.l. i 22 ‘when they desire’. All are prob. formed from the 3rd sg.

(3) The 1st and 2nd pl. end in ‑om, ‑och: diwyccom, digonhom b.b. 30, crettoch r.m. 131.

(4) The impers. ends in ‑er; but there are examples of a form in ‑w͡yr: r͑othwyr b.t. 1 for the usual r͑oẟer ‘may be given’.

iv. The 2nd sg. ending is ‑ych: r͑oẟych w.m. 4 ‘thou givest’ (mayest give), gwypych do. 14 ‘thou knowest’, gellych do. 151 (: gallaf ‘I can’). In Late W. a dialectal form ‑ech sometimes occurs, § 16 iv (2) (β), as lletteuech Ruth i 16, gweddïech Matt. vi 6, poenech Marc v 7. In the present dialects the subj. is seldom used except in the 3rd sg. and pl.; and some recent writers have used ‑ot for the 2nd sg. Even ‑ost has been written; in Wms.’s verse Marchog, Iesu the last line Tyrd am hynny maes o law 849 appears in recent hymnbooks as Pan y byddost ti gerllaw.—gellyt ZE. 512 is a misreading of gellych r.m. 220.

v. The impf. subj. is formed by adding the personal endings of the impf. to the subj. stem; thus (subj.) bei dywettut ti … (ind.) minheu a ẟywedwn w.m. 118–9 ‘if thou wouldst say … I would say’. In Late W., owing to the levelling of the subj. with the ind. stem, the distinction between the moods is not preserved in the impf., except in af, gwnaf, dof, wyf, which have special subj. stems; see i (2) above.

§ 177. Pres. Impv.—i. (1) The 2nd sg. is the bare stem of the pres. ind. It differs from the 3rd sg. pres. ind. in never having its vowel affected; thus deil ‘he holds’, dal ‘hold!’ tau ‘is silent’, taw ‘be silent!’ pair Zech. x 1 ‘causes’, pâr Ps. xxv 4 ‘cause!’

(2) Verbs which have ‑a in the 3rd sg. pres. ind. take it also in the 2nd sg. impv.: kerẟa w.m. 83, r.m. 60, llunnya w.m. 25, r.m. 16, etc., see § 173 v.

ii. (1) The 3rd sg. ends in ‑ed: kymeret w.m. 30, r.m. 19 ‘let her take’, aet un w.m. 13, r.m. 9 ‘let one go’, gadawed, dychweled Es. lv 7.

(2) A 3rd sg. in ‑id added to the subj. stem is also met with: elhid b.b. 101 (: af ‘I go’), r͑othid do. 93 ‘may he give’, gwrthleẟit Duw Ỻ.A. 26 ‘may God ward off’, Trowyr (≡ trỿ-w̯yr), getid Duw’r ieuaf G.Gl. m 146/185 ‘three men, God spare the youngest’, gettid Mair D.N. g. 154, Telid Duw iddynt M.K. [viii] ‘let God repay them’.

iii. The 3rd pl. ending is ‑ent: diskynnent w.m. 22 ‘let them descend’, katwent a.l. i 138 ‘let them keep’, traethent Ỻ.A. 159 ‘let them speak’, deuent (rh. with stent) L.G.C. 66 ‘let them come’. This is obviously formed from the 3rd sg. in ‑ed (since orig. ‑ent would have become ‑ynt). There is also a 3rd pl. bint Ỻ.A. 81 ‘let them be’ formed from bid. In the Bible a 3rd pl. in ‑ant is used: gwybyddant Ps. lix 13 ‘let them know’, dychwelant do. 14 ‘let them return’. This is a late re-formation following the analogy of the 1st and 2nd pl. which are taken over from the pres. ind. In spite of the use of this form in the Bible, the older form persisted in the late period: Angylion doent … Rhoent eu coronau Wms. 320 ‘Let angels come, let them put their crowns.’

iv. The 1st and 2nd pl. have taken the forms of the pres. ind.; but an earlier 1st pl. in ‑em occurs rarely, § 184 ii (1).

Origins of the Welsh Verb.

The Aryan Verb.

§ 178. i. In order to trace the development of the Welsh verb, some account, though it be in the briefest outline, must be given of the Ar. verbal system. For a fuller, but still concise and most instruc­tive descrip­tion, see Meillet, Intr.² pp. 165–219.

ii. Stem form.—Two kinds of stem may be distinguished. Thematic forms were those ending in the thematic vowel ‑e‑: ‑o‑; it was ‑o- in the 1st sg. and 1st and 3rd pl., and ‑e- in the 2nd and 3rd sg. and 2nd pl. Athematic forms were those ending in a consonant or long vowel.

iii. Personal endings.—(1) The Ar. verb had personal endings for each of the three persons of the sg., dual and pl. These were either primary or secondary; and the primary endings differed to some extent for thematic and athematic stems. There were special endings for the perfect.

In the following list I omit the dual; and as the thematic vowel cannot be separated from the ending in some primary forms, I insert the vowel before the ending through­out, separat­ing it by a hyphen, where possible, from the personal ending proper: all the persons of thematic stems are thus put on the same level.

(2) Active voice.

Primary.—Thematic: sg. 1. ‑ō, 2. ‑ēis, 3. ‑ēit; pl. 1. ‑o‑mesi, ‑o‑mosi, ‑o‑mē̆s, ‑o‑mō̆s, 2. ‑e‑the, 3. ‑o‑nti.

Athematic: sg. 1. ‑mi, 2. ‑si, 3. ‑ti; pl. 1. ‑mesi, ‑mosi, ‑mē̆s, ‑mō̆s, 2. ‑the, 3. after a consonant ‑enti, ‑n̥ti, after a vowel ‑nti.

Secondary.—Thematic: sg. 1. ‑o‑m, 2. ‑e‑s, 3. ‑e‑t; pl. 1. ‑o‑mē̆, ‑o‑mō̆, 2. ‑e‑te, 3. ‑o‑nt.

Athematic: sg. 1. after a vowel ‑m, after a cons, ‑m̥, 2. ‑s, 3. ‑t; pl. 1. ‑mē̆, ‑mō̆, 2. ‑te, 3. after cons, ‑ent, ‑n̯t, after vow. ‑nt.

(3) Middle voice (medio-passive); 1st and 2nd pl. omitted. Primary.—Thematic: sg. 1. ‑o‑mai, ‑ōi, 2. ‑e‑sai, 3. ‑e‑tai, pl. 3. ‑o‑ntai.

Athematic: sg. 1. ‑mai, 2. ‑sai, 3. ‑tai, pl. 3. ‑ntai.

Secondary.—Thematic: sg. 2. ‑es‑o, 3. ‑e‑to, pl. 3. ‑o‑nto. Athematic: sg. 2. ‑so, ‑thēs, 3. ‑to, pl. 3. ‑nto.

(4) Perfect. The following endings only need be mentioned. Active: sg. 1. ‑a, 3. ‑e.

(5) The characteristic of the primary endings is final ‑i. The differ­ence in the sg. between primary thematic and athematic forms may have arisen by phonetic change in the parent language; thus we should expect themat. sg. 2. ‑esi, but (though Skr. has bhár-asi) the Ar. form seems to have been ‑ēis; possibly by metath. and compen­satory lengthen­ing, but this is quite uncertain.

iv. Mood and Tense Stems.—(1) The present stem was rarely the simple root. In most cases it was either the redupli­cated root, the root with thematic vowel, the root with stem‑forming suffix, or the root with the infix ‑n- or ‑ne‑.

The present stem with primary endings formed the pres. ind.; as *dí-dō-mi (Gk. δίδωμι) ‘I give’, √dō‑; *bhér-o‑nti (Gk. Dor. φέροντι) ‘they bear’, √bher‑.

The present stem with secondary endings, and with the augment before it, formed a past, as *é bher-o‑m (Gk. ἔφερον) ‘I bore’. This augmented past is called imperfect, because it is imperfect in meaning in Gk. In Skr. it is merely a past.

(2) The stem of the s‑aorist was formed with ‑s- (athematic); of the future with ‑e- or ‑se- (thematic); of the optative with ‑i̯ē- etc.; these forma­tions are noticed below.

(3) The simple root with or without the thematic vowel formed aorist stems as follows, all the endings being secondary: firstly, R‑grade of √ + them. vowel, as *é liq^u̯-o‑m (> Gk. ἔλιπον), √leiq^u̯‑; this may be called the thematic aorist;—secondly, F‑grade of √ (at least in sg.), athematic, as *é bheid‑m̥ (> Skr. ábhedam), √bheid- ‘split’; this is called the root-aorist.

v. The Augment was a separable accented preverb denoting past time. It was lost entirely except in Gk., Armenian and Indo‑Iranian.

The augment is always followed by forms with secondary endings. These forms were also used without the augment; they are then called injunc­tive; thus Skr. Ved. bhárat ‘bore’, Gk. Hom. φέρω ‘bore’ < Ar. *bhere‑t beside impf. ábharat, ἔφερε < *é bhere‑t. Injunc­tive forms are either past or pres. in meaning; the augment makes them definite­ly past.

The Welsh Verb.

§ 179. Pres. Ind.—i. In Ar. the verb was unaccented when it followed a preverb such as a negative particle, or a prepo­sition later compound­ed with it. This was un­doubted­ly the rule in Kelt. (despite devia­tions in Ir.), as it was in Italic. In the pres. ind. in Kelt. in the 3rd sg. the accented verb had the primary ending, that is, the regular present ending, but the un­accented verb had the secondary suffix, that is, the injunc­tive form. Thus the W. proverb Trenghit golut, ny threingk molut § 173 vi (1) repre­sents Kelt. *traŋkī́-ti u̯ò…, né ttraŋkī‑t mò… It has been suggested that this reflects the original use of the Ar. primary and secondary endings; and it certainly accords with the fact that the augment, an accented preverb, is always followed by forms with secondary endings.

ii. The Ar. athematic stems, excepting those of a few common verbs, ended mostly in the long vowels ‑ā‑, ‑ē‑, ‑ō‑. As medial ‑ō- became ‑ā‑, and ‑ē‑ became ‑ī‑ in Kelt., these charac­teristics were reduced to two, ‑ā- and ‑ī‑. The vowel had F-grade in the sg., R-grade in the pl., as in Gk. ἵστημι < *sí-sthā-mi, pl. 2. ἵστατε < *sí-sthə-the. The Kelt, forms of the 1st sg. pres. were therefore *‑ā-mi, *‑ī-mi. As the form was mostly un­accented, and un­accented ‑ā- > Brit, ‑a- § 74, the prevail­ing Brit, forms were *‑a-me, *‑ī-me. These give the W. ‑af, -if, the latter com­parative­ly rare, § 173 ii, and now obsolete. Examples: (1) Ar. *dí-dō-mi ‘I give’ > Kelt. *(p)ró (di‑)dā-mi > Brit. *ró-da-me > W. rhoẟaf ‘I give’;—(2) Ar. *dhí-dhē-mi ‘I put’ > Kelt. *(p)ró (di‑)dī-mi > Brit. *ró-dī-me, which would give W. *rhoẟif ‘I put’. But the latter ending was rare, and was supplant­ed by ‑af, the result being, in this case, that two verbs became one: rhoẟaf ‘I give, I put’. The re­duplicat­ing syllable was probably lost by haplology. Only the vowel of the syllable dropped in dodaf ‘I give, I put’ < *dó-tāme or *dó-tīme < *dó d(i)-dōmi or *dó dh(i)-dhēmi: Gk. δίδωμι or τίθημι. Usually dodaf is ‘I put’; for dod ‘give’ see Ps. lxxii 1, Gr.O. 87.

iii. (1) The accented forms of the 3rd sg. *‑ā́-ti, *‑ī-ti give the W. strong forms ‑awd, ‑id. These are used at the head of the sentence, like accented verbs in Skr. The intro­duction of ‑h- before the ending in Ml. W., where not etymo­logical as in trenghit (ngh < ŋk), is ana­logical, and partly arti­ficial. The second form tended to oust the first in this case, as seen in O.W. prinit ‘buys’ for *prinaut < Brit. *prinā́-ti: Ir. cren(a)id; see § 201 i (4). The ‑id form with the initial of the affixed pron. fo, thus *‑id‑f, gave *‑it‑ff and then ‑iff, the dial. ending, by loss of the t as in the 2nd pl., see § vii. The West Gwyn. ‑ith has recent th for ff.

Ml. W. ‑yd in ëyt, § 173 vi (1), is from *‑etī < *‑e‑tai the middle 3rd sg. ending: Gk. φέρεται; see § 193 x (1).

(2) But the usual form of the 3rd sg. in W. is the stem without or with vowel affection; this comes from the un­accented injunc­tive form; thus câr loves < Brit. *kara‑t; rhydd ‘puts’ < Brit. *ró-dī‑t. The latter, being more dis­tinctive, spread; thus rhydd ‘gives’ instead of *rhodd < *ró-da‑t.

iv. (1) The Ar. thematic endings *‑ō, *‑ēis, *‑ēit would become *‑ū, *‑īs, *‑īt in Kelt.; and these in W. would all drop after affecting the vowel. The 1st and 2nd sg. so formed were lost because they were not dis­tinctive; but prob. the 3rd sg. added to the number of affected stems forming the W. 3rd sg.

(2) The thematic injunctive ending ‑et of unaccented verbs dropped without affecting the vowel; thus Ar. inj. *bher‑e-t ‘bears’ gives Kelt. *kóm beret > W. cymer ‘takes’, and Kelt. *áti beret > W. adfer ‘restores’, etc. It is found not only in compound, but in simple verbs, as cêl ‘conceals’ < *kelet, rhed ‘runs’ < *retet, etc., because the un­accented was, as in the case of athematic stems, the commoner form; e.g. ni chēl grudd gystudd colon prov. ‘the cheek does not hide the sorrow of the heart’.

(3) There is no *‑ed, since the them. prim, ending was ‑ēit, not *‑eti § 179 iii (5). The strong form of the above verbs is taken over from the ‑ī‑ conju­gation; as rhedid car gan anwaered prov. ‘a car will run down hill’. (So Ir. berid for *beri, with anal, ‑d.)

v. The W. 3rd pl. ‑ant is from Kelt, ‑anti < Ar. *‑ə‑nti which was common to the ‑ā‑ and ‑ī‑ conju­gations; see ii above. There is no trace of the thematic *‑o‑nti, because ‑ont came to be associ­ated with other tenses. The O.W. ‑int, Ml. W. ‑ynt, may represent the athem. *‑enti or the middle *‑ontai, more probably the latter; ‑(h)awnt is certainly formed after ‑(h)awt.

vi. The 2nd sg. ‑yẟ (which is the oldest form of the ending ‑y, later ‑i) seems to come from accented forms of itera­tives in ‑éi̯e, or denomi­natives and deverba­tives in ´‑i̯e- the commonest stem-suffix in the Ar. languages. In Kelt, from *karo‑s ‘dear’ the i̯e-denom. would be *karé-i̯ū, *karé-i̯īs, *karé-i̯īt; all these would give W. keryẟ. But the 1st and 3rd sg. had more distinc­tive endings, and ‑yẟ survived in the 2nd only, though there are traces of it in the 3rd, see § 173 vi (3). The latter occur in relative sentences, where the verb was prob. accented, as in Skr. The accented 2nd sg. is frequent­ly used, and answered by accented na and the unacc. 1st sg.

vii. The 1st and 2nd pl. in W. are re-formations, and it is useless to attempt to derive them from Kelt, forms. The Kelt. 2nd pl. was, them. *‑e-te, athem, *‑a-te. The former would give W. *‑ed (Ml. Bret. ‑et); to this was added the initial of the affixed pron. chwi, thus *caret‑ch^w̯ > *carewch by loss of t, cf. iii (1); at this stage a 1st pl. *caren was formed on the analogy of the 2nd pl., with the initial of the aff. pron. ni ‘we’; this form is attested in O.W. iben, and survives to this day in West Gwyn. in caran beside carwn ‘we love’ (Gwyn. ‑an = ‑en). As the 2nd pl. clashed with the impf. it was re-formed with the vowels of the 2nd sg. thus *cerywch > cerwch ‘ye love’; sub­sequent­ly the vowel of this ending intruded into the 1st pl., giving carwn ‘we love’. A statement in the 2nd pers. is always answered in the 1st, hence the influence of the forms on one another in the less used pl.

viii. (1) In Pr. Ar. an ending *‑r- formed impersonals. It survived only in Indo-Iranian and Italo-Keltic. In Skr. it takes the form ‑uḥ (before a vowel ‑ur) in the active, and ‑re, ‑ire in the middle; ‑uḥ repre­sents *‑r̥ or *‑r̥s, Meillet Intr.² 203. These endings in Skr. form the 3rd pl.; this is natural enough when one considers that there is only a shade of distinc­tion in meaning between the impers. dywedir ‘on dit’ and the 3rd pl. dywedant ‘they say’.

(2) In Italo‑Kelt. it was used in two ways; first, it might be added to the tense-stem, as Umbrian subj. ferar ‘on portera’, pres. ind. ier ‘on va’, Oscan subj. sakrafír (with últiumam for object) ‘cysegrer’. Secondly it was added to the 3rd sg. or pl. middle, and then extended to other persons in deponent verbs in Ir., and deponent or passive in Lat., as Lat. itur, Osc. vincter ‘vincitur’, Umbr. emantur ‘emantur’. On the im­personal use of the Lat. passive see Ernout MSL. xv 273–333.

(3) In Kelt. the ending may be taken to have been *‑re (also *‑ro?). The Brit. shorter forms of the ‑ā‑, ‑ī‑, and thematic conju­gations in the pres. were *‑ā́-re, *‑ī́-re and *‑e-re re­spective­ly. These give the W. pres. impers. ‑awr, ‑ir and ‑er. The second survives to this day, see ix (2), and is in common collo­quial use. The first was used in Early Ml. W., and the third occurs also, but was obsoles­cent owing to its clashing with the subj. form. The ‑h- sometimes seen before ‑awr and ‑er is an intrusion from the subj.

(4) Longer forms, with *‑re added to the 3rd sg. middle secondary endings would be *‑ā̆‑to‑re, *‑ī̆‑to‑re and *‑e‑to‑re. These give the W. ‑ator or ‑otor, ‑itor and ‑etor. The dental should be ‑d‑, which occurs in dygedawr b.t. 75; the ‑t- is partly due to the intrusion of subj. ‑h‑, partly a mis­transcrip­tion of O.W. ‑t‑, as these forms were obsolete at the dates of our MSS. Since the above was written an O.W. example has come to light in cephitor cp., with one ‑t‑ as in retec ib., Ml. W. r͑edec.

ix. (1) The reason why the Welsh pres. has always had a fut. meaning is that it contains beside the pres. the Ar. ‑e- future, generally called subjun­ctive. This tense is formed by adding the thematic vowel e/o to the pres. stem. In the case of thematic stems the effect was to lengthen the thematic vowel through­out. In the sg. this would make no differ­ence (Gk. subj. φέρω. ind. φέρω; the subj. φέρῃς is a re‑formation; orig. *bhérēis would give *φέρεις in the ind.). In long-vowel stems the added thematic vowel simply converted them to thematic stems, as Gk. subj. διδῶ beside ind. δίδωμι; this intro­duces no new element. The 3rd pl. fut. *‑ōnti (Gk. Dor. φέρωντι) would have its vowel shortened § 74 iv, and so would not differ from the pres.

(2) In the impers. the fut. form for thematic stems would be *‑ē‑re > Kelt. *‑ī‑re, beside the pres. *‑e‑re. All thematic stems therefore would have a fut. in ‑ir beside the pres. in ‑er. This shows why ‑ir became the prevail­ing pres.-fut. form.

(3) In consonantal athematic verbs the distinction between pres. and fut. is much clearer; thus the pres. stem *es- ‘be’ has fut. stem *ese‑; the former gives the Ar. pres. *és-mi, *és-(s)i, *és-ti (> Skr. ásmi, ási, ásti); the latter gives the Ar. fut. 1. *és-ō (> Lat. ero), 2. *és-ēis, 3. *es-ēit, injunc­tive *es-et (> Skr. asat, Lat. erit).

The W. pres. is a mixture of pres. and fut. forms. The Kelt. fut. *ésū, *ésīs, *ésīt would give *oe for the three persons; of this a trace survives in oe‑f b.b. 50 ‘I am’. The pres. sg. 2. *ése (< Ar. *ési) and 3. inj. *eset would give *wy, whence sg. 1. wy‑f, 2. wy‑t, 3. *wy rnetath. to yw § 78 iv; in pi-eu ‘whose is?’ it is weakened to ‑eu, § 78 iii, § 192. The Ar. 3rd sg. pres. *ésti survives in W. ys, which has become imper­sonal. The W. 3rd pl. ynt (for *hynt) comes from Ar. 3rd pl. pres. *s‑enti (*s- is V-grade of √es‑). The W. 1st pl. ym (Ir. ammi) implies a Kelt. **ésmesi, a confusion of pres. *smesi and fut. *ésomesi. The W. 2nd pl. ych is, as usual, a new form made to match.

As byẟaf used for the fut., wyf has lost its fut. meaning except in certain idioms, as yr wyfyno yfory ' I shall be there to-morrow '.

§ 180. The Imperfect—i. As above intimated, § 171 ii (2), the W. impf. comes from the Ar. optative. This was formed by means of a suffix *‑ii̯ē‑, *‑i̯ē- with secondary endings.

ii. (1) In athematic verbs the suffix *‑i̯ē- was F-grade and accented in the sg.; the preceding vowel had R- or V-grade; thus 3rd sg. Gk, τιθείη < *dhi-dhe-i̯ēt (ei̯ R_1e of ēi̯), Skr. dadhyā́t < *dhe-dh‑i̯ḗt, the Skr. pre­serving the original accentu­ation.

(2) In Kelt. the ē became ī, so that the forms would be *‑a-i̯ī́t, *‑e‑i̯ī́t; these were levelled as *‑i‑i̯ī́t in Brit. and this gives ‑ai, § 75 iv, v (2); thus Kelt. *kara-i̯ī́t > W. carai ‘would love’. This form would also result from the 1st and 2nd sg. forms *‑a-i̯ēm, *‑a-i̯ēs; hence the endings for those persons were selected from thematic verbs.

(3) The consonant stem *es‑ ‘be’ gave Ar. *s‑(i)i̯ē‑t, which gives Skr. siyā́t or syā́t, O. Lat. siet; in Kelt. it would be *sii̯ī́t. Coming generally after a preverb, or after its comple­ment, it was un­accented; and *´siiīt gives regularly W. (h)oeẟ ‘would be, was’ § 75 iv (2); the h- is seen in yttoeẟ < *yd-hoeẟ < *íta sii̯īt ‘there would be’ § 219 ii. The whole tense oeẟwn etc. was built from the 3rd sg.

iii. (1) In thematic verbs the suffix ‑i̯ē‑ had its V-grade ‑i̯‑, which formed a diphthong with the thematic vowel, which was always ‑o‑; thus the optative of *bhérō ‘I bear’ was sg. 1. *bhéroi̯‑m̥ > Skr. bháreyam (for *bharayam). In Kelt. it would be *béroi̯‑m̥ > Brit. *béroi-an(n) > W. *cy-merw͡y‑n > cymerwn. The only possible expla­nation of ‑wn is that it is for *‑w͡yn, see § 78 i (2); on *oi > w͡y § 75 ii (2); on the retention of ‑n § 113 i (1).

(2) The W. 2nd sg. ‑ud comes regularly from the 2nd sg. middle *‑oi-thēs. The ending *‑thēs (: Skr. ‑thāḥ) is represent­ed in the ‑the‑r of Ir. deponents; and ‑ud spread from deponent to all verbs in W. because it was distinc­tive.

iv. (1) In athematic verbs, in the middle voice where the ending was syllabic, the suff. became R-grade *‑i̯ə‑; this coming before the accent remains as ‑i̯a‑; thus in the deponent verb gwnn ‘I know’ the 3rd sg. impf. is gwyẟi̯ad for *gwẟi̯ad regularly represent­ing the 3rd sg. opt. mid. *u̯id-i̯ə-tó.

(2) In long-vowel stems the reduced stem-ending and suffix would thus be *ə-i̯ə; by § 63 vii (5) this should give *ii̯ə > ‑ī‑, which is the usual form (though other reduc­tions are possible), as in Skr. da-dī-tá < *de-dī-to, √dō‑. Thus the 3rd sg. opt. mid. of Kelt. *karā-mi would be *kar-ī-tó, which gives regularly W. cerid, the impers. of the imperf. ind. This middle was un­doubted­ly a passive in Kelt., and was assimi­lated in its use to the impers. pres. in ‑r after the ‑r form for this tense, namely *‑ir, had gone out of use owing to its clashing with the pres.

(3) The 3rd sg. mid. of thematic stems ended in *‑oito. We should therefore expect ‑ud beside ‑id for the impers. in W. A trace of this actually occurs in ac y haruetud etc. b.b. 20, which should be *ac yth arweẟud etc. ‘and thou wert borne’, etc., where the scribe mistook the impers. for the 2nd sg., which makes no sense if it is active, and we can hardly assume the 2nd sg. to have retained a passive sense.

v. (1) In the 1st and 2nd pl. of athematic stems the Ar. form was *‑i̯ə‑: *‑i‑. We can probably assume for Kelt. *kár(a)-i̯ə-me; the m was doubled on the analogy of the aor.; and post-tonic *i̯a > i̯e > e in W., § 65 vi (1); hence W. carem. Similarly 2nd pl. *caret + chw̯- > karewch, carech.

(2) The 3rd pl. ending was *‑i̯ént (for *‑i̯ḗnt). The form *‑a‑i̯ént gives W. ‑i § 75 v (1); as tri ugeint canhwr a sevi b.t. 55 ‘6000 men stood’; hence the rare “3rd sg.” ‑i. The 3rd pl. ‑ynt seems to be a middle form < *‑ento < *‑i̯ə‑nto (or *‑into < *‑ī‑nto), which spread because it had the 3rd pl. sign ‑nt.

vi. (1) The impf. subj. is the optative of the s‑aorist, cf. Lat. vīderīmus < *u̯eid-is-ī‑m‑. Thus Kelt. *kara-sii̯īt > Ml. W. karhei.

(2) The plup. is an optative formed from the new Kelt. ss‑aorist. Thus Brit. *karassii̯ī́t > carassai.

The plup. is held to be a Brit. innovation. Strachan’s examples of the impf. subj. being replaced by the plup. in later texts, quoted in b.b. 157, prove nothing as to the antiquity of the plup.; its existence in Bret. shows that it goes back at least to Brit., so that the evidence of Ml. texts is ir­relevant. We also find the plup. in early texts where we should expect to find the impf. subj. as ri-uelssud b.b. 20. The fact is that the two aorists were not very sharply distin­guished.

§ 181. The Aorist.—i. The Welsh aorist comes from a Keltic re-formation of the ‑s- aorist. The orig. Ar. formation seems to have been (1) L‑grade of √ + ‑s‑ (in Kelt. R‑grade in the pl.), or (2) F‑grade of V + ‑is‑. The endings are secondary.

ii. (1) With long-vowel stems the suffix is ‑s‑; thus Skr. á-prā-sam < *é plē‑s‑m̥, √pelē‑ ‘fill’, Gk. ἐφίλη-σ‑α (inter­vocalic ‑σ- restored from cons. stems ‑ψα, etc.). Thus Kelt. *kárā‑s‑m̥ ‘I loved '. Bearing in mind that st > ss and that sm > mm the whole Kelt. tense may be restored thus: sg. 1. *kárāsm̥, 2. *kárāss, 3. *kárāss, pl. 1. *ká-rammo, 2. *kárasse, 3. *kárasn̥t.

(2) This tense was wholly reconstituted in Kelt., with stem sg. *kárāss‑, pl. *kárass‑. The 1st and 2nd sg. were made anew with thematic endings; thus 1. *kárāssū, 2. *kárāssīs (inj. ‑es). The 1st pl. became *káras­sammo instead of *kárammo; then followed 2. *kárassate instead of *kárasse. Un­accented ā was shortened in Brit, and Ir. and these forma­tions gave regularly Ir. sg. 1. ro-charus (2. ro-charais), pl. 1. ro-charsam, 2. ro-charsaid, and W. sg. 1. kereis, 2. kereis + t, pl. 1. karassam, 2. *karassat + chw̯- > karas­sawch. The ending of the 3rd pl. was made primary; thus *káras­santi > Ir. carsait, W. karassant. As a variant the thematic vowel was brought into the pl. also; thus Brit. *kárass-o-mmos, *kárass-o-nti giving W. karassom, karassont; from these followed carasoch.

iii. To the 3rd sg. two things happened, (1) It remained unchanged; thus *kárăss > W. *kar, which was extended to karawẟ to distin­guish it from the pres.; for ‑awẟ see § 182 iii. The Ir. ro-char implies *karass with short a from the pl.

(2) It was re-formed with the thematic vowel, following the 1st and 2nd sg.; thus *kárāsset; or with ‑a- from the pl. as *kárāssat. Either of these would give W. *karas (caf-as § 175 i (2)). The first gives Ir. carais.

iv. The treatment of ‑ī- stems was precisely similar. The stem-ending in the sg. was *‑īss‑; this survives in the W. 3rd sg. peris. In the 1st and 2nd sg. it was replaced by ‑eis of ‑ā- stems; but in Gwyn. dial. ‑is survives in these persons also. In the pl. the stem-ending was ‑ass‑, as for ‑ā- stems, the ‑a- represent­ing ə, the R-grade of the ‑ē- from which the ‑ī- is derived.

v. (1) Consonant stems formed the aorist with *‑is‑, cf. Lat. vid-is-tis, which developed similarly, and gives W. ‑ỿss- in eiste­ẟyssant, etc. In the 3rd sg. it appears in W. as ‑es from *‑iss‑at. In the 1st and 2nd sg. it was replaced by ‑eis.

(2) The *‑iss- suffix seems to have intruded into the thematic conju­gation; thus Brit. *kóm bere-iss-at > W. kymerwys, kymerws.

vi. The impersonal forms ‑ad, ‑id, ‑ed, ‑wyd seem to have been formed on the analogy of the impf. im­personal, with the vowels of the 3rd sg. aor.

vii. (1) The root-aorist, § 178 iv (3), was treated similarly in Kelt. Thus for the root *qan‑ ‘sing’ the orig. Kelt, root-aor. would be sg. 1. *kan-m̯, 2. *kan‑s, 3. *kan‑t. The 3rd sg. became the stem, and the new tense formed from it was sg. 1. *kantū, 2. *kantīs, 3. *kantet or *kantat. These forms gave W. sg. 1. keint, 2. *keint, 3. kant. To the 1st and 2nd sg. the perfect endings ‑um, ‑ost, § 182 iv (1), were added, § 175 iii (2).—gwant ‘wounded’ from gwanaf < *gwonaf: Ir. gonim, √ꬶ^u̯hen‑, is probably formed on the analogy of cant. The root *bher‑ has this aor., which survives only in the 3rd sg. in W.; thus W. kymyrth < *kóm bertet or kymerth < *kóm bertat, § 175 iii (1).

(2) Other examples that survived are from roots ending in gutturals: dyrreith b.t. 54 ‘returned’ < *do-(p)ro-rek-t-et, √reg̑‑: W. dyre ‘come!’ § 193 x (8);—maeth b.t. 74 l. 1 ‘nursed’ < *makt‑ < *māk‑t‑, √māk̑‑: magaf ‘I nourish’. The root *u̯ereꬶ‑ ‘work’ had sg. 1. *u̯rek-t-ū, 3. *u̯rek-t-et giving W. gwrith, gwreith; the former occurs in ef gwrith b.t. 26 (? 3rd sg.); the latter seems to occur in gwnaeth [read gw(r)eith] gwynnẏeith gwreith e law b.a. 2 lit. ‘work of vengeance wrought his hand’; but this verb (gwnaf) being in the pres. con­jugated like af, this tense was assimi­lated to the perf. of af, and became sg. 1. gwneuthum. 3. gwnaeth. The quotation shows that scribes changed old gwreith to gwnaeth, the wrong gwreith, viz. the noun, being changed here. In Bret. the old form survived: Ml. Bret. sg. 3. grez.

§ 182. The Perfect.—i. In Pr. Ar. the vowel-grade of the root was F° in the 1st sg., and L° in the 3rd sg., as Skr. cakára ‘I made’ < *qeqora, cakā́ra ‘he made’ < *qeqore.—Ml. W. kigleu ‘I have heard, he has heard’, Ir. ro-chūala, ro-chūalae. The W. form implies the 1st sg. k̑ū́-k̑lou̯-a: Skr. s̑u‑s̑ráva; for the long ū of the redupli­cator cf. Skr. tū-tava, √teu̯- ‘be strong’. See § 194 v (4).

ii. (1) The following old perfects are 3rd sg. only, and show L°-grade of the root: √u̯ereꬶ- ‘work’ gave *u̯e-u̯rōꬶe > Brit. *u̯o-u̯rā́ge > Ml. W. guoreu, goreu ‘did’ (u̯ lost by dissim., āg > eu § 71 iii);—√ret- ‘run’ gives gwa-red-af ‘I succour’; perf. sg. 3. *re-rōt-e > Brit. *u̯o-(re)rā́te > Ml. W. gwarawt ‘succoured’;—√u̯et/d- ‘say’ gives dy-wed-af ‘I say’; perf. sg. 3. Brit. *do-u̯ā́t-e or *dó-u̯at-e > Ml. W. dywawt or dywat, dywot ‘said’ (unacc. ā shortened § 74; wa: wo § 34 iv).

(2) √deuk- had R-grade *duk- in the Brit. pres., giving W. dyg-af (: Lat. dūco < O. Lat. douco, F°-grade); perf. sg. 1. *du-douk-a > W. *dy-ẟuc, 3. *du-dōuke > W. dy-ẟuc b.t. 4, 52. The tense was re-formed with the perf. endings ‑um, ‑ost § iv (1), § 194 iii (2).—The verb amygaf ‘I defend’ has similarly a 3rd sg. perf. amuc § 194 iv (2).

iii. In verbs like eisteẟaf ‘I sit’, gorweẟaf ‘I lie’, arweẟaf ‘I carry’, go(r)ẟiweẟaf ‘I overtake’, etc., the form of the above perf. is seen in goẟiwawẟ w.m. 42 ‘overtook’; this being re-formed as gorẟi­weẟawẟ r.m. 29 (so eisteẟawẟ w.m. 188, etc.), the ‑awẟ seemed to be a 3rd sg. past ending; and was added to suffix­less aorists like *kar § 181 iii (1) giving karawẟ, Mn. W. carodd ‘loved’.

iv. (1) Deponent verbs in Brit, had periphrastic perfects formed like those of Lat. deponents. Thus √ag̑‑: perf. sg. 1. *aktos esmi > *aktoimmi > aethum, euthum; 2. *aktos (e)si > *aktossi > *aethos + t = aethost; 3. *aktos ’st > *aktosst > aeth ‘went’. From these forms 1st and 2nd sg. endings ‑um, ‑ost were deduced, and added to other forma­tions, such as the root-aor. keint and the perf. duc. This perf. itself was completed in the pl. by the addition of the aor. endings ‑am, ‑awch, ‑ant.

(2) The Ml. plup. is sg. 3. athoeẟ for *aethoeẟ, which repre­sents *akto(s) sii̯īt. The diphthong ae was simpli­fied prob. by dissim. with the diphthong oe. The second perf. athwyf etc. seems to be a new creation formed on the analogy of the plup.

(3) The impers. llas ‘was slain’ is an example of this formation. It is not a root-aor. as it has R-grade of √qolād‑. It is probably a perf. passive; thus *slad-tos (e)st > *slass-osst > llas ‘was slain’. This passive has a pl. llesseint b.b. 63 ‘were slain’ which seems to be re-formed like impfs. in ‑ynt § 174 iii (2), for *llessynt < *slassī senti; lleẟes­synt b.a. 9 ‘were slain’ seems to be another re-formate.

(4) The impers. of the above perf. is formed by adding the impers. *bw͡yt of the verb ‘to be’ to the stem; thus aeth-pwyt, etc. This was extended to root-aorists, as *kant‑pwyt > kanpwyt, perfects, as duc-pwyt, and presents; § 175 iii (7). The form *bwyt does not occur elsewhere; prob. the whole formation is new.

§ 183. Pres. Subjunct.—i. The pres. subj. represents the Ar. fut. with suffix ‑se- (fut. in ‑e- of ‑s- aor.), which gives Italic subj. also: Lat. faxit. The W. forms are chiefly those of the ‑ā- conjugation. The accent in the sg. seems to have been on the ā.—In the b.b. it seems sometimes to be a mere fut., e.g. vvnahont 61 ll. 14–15.

ii. (1) Stem *karā-se- gives sg. 1. *karā́sō > *karā́sū > *karāi̯ī > W. *karwy; ‑f was added to distinguish it from the 3rd sg.; the 3rd sg. *karā́-sīt > karwy, and the unacc. injunct. *karaset, the usual form > *karoe > karo, § 75 i (2), (3), § 78 i (1); pl. 1. *kara-so-mos, with m doubled after the aorist pattern, gave kar-hom; pl. 3. *kara-sonti gave kar-hont.

(2) Impers. *kara-se-re > kar-her § 75 i (2). The form r͑othwyr § 176 iii (4) is most probably made from the 3rd sg. r͑othwy.

(3) According to the above the ‑h- belongs to the pl. and impers. only; in the sg., therefore, it is an intrusion. In Ml. Bret, it is not usual in the sg. but occurs regularly in the pl.

iii. (1) In consonant stems the ‑s- came immediately after the cons.; few examples survive because the conjugation had become vocalic in the indic.—√u̯ereg- ‘work’; pres. ind. *u̯rag-at > W. gw̯na ‘does’, subj. *u̯rek-se‑t > gunech l.l. 120 ‘may do’, ny ofyn ẏ neb a wnech b.t. 64 ‘he asks no one what he may do’;—√deuk‑: pres. ind. *duk-at > W. dwg ‘brings’, subj. *deuk-se‑t > duch b.b. 40, later duwch b.t. 28;—√ret‑: subj. *u̯o-ret-se‑t > gwares § 194 ii. The vowel of the root is seen to be F‑grade in this tense.

(2) Corresponding to the 3rd sg. gwnech the 2nd sg. *u̯rek-sīs would give *gwnych; this being re-formed as *gwne-ych and gwnel-ych, the latter form would naturally spread to el-ych and del-ych; and as these are three of the commonest verbs in the language, the ending ‑ych might spread from them to all verbs, as being the only distinctive form of the 2nd sg. pres. subj.

§ 184. The Imperative.—i. The 2nd sg. has always represented the bare pres. stem. Thus W. câr ‘love thou’ < Kelt. *karā; W. kymer ‘take’ < *kóm bere < Ar. *bhere: Gk. φέρε.

ii. (1) For the other persons the optative seems to have been once in use: ystyrẏem b.t. 33 ‘let us consider’. The 3rd sg. forms are difficult. In Ir. the endings are ‑at, ‑et; the lost vowel cannot be the ‑ō of Lat. ‑tō, or the ‑u of Skr. ‑tu (Thurneysen Gr. 351); it must be ‑o or ‑a. The forms are the same in Ir. for active and deponent verbs; this suggests that the ending was the middle secondary *‑to. In Ir. also the forms are the same as those of the impf.; the mid. forms of the 3rd sg. opt. *‑i̯ə‑to, *‑ī-to (W. gwyẟiad, cerid) would give ‑ed, ‑id if in the former the accent were shifted to the stem. The 3rd pl. may have been *‑ynt (Corn. ‑yns beside ‑ens), the form in the impf.; but it was re-formed with the vowel of ‑ed, rarely of ‑id as in bint § 189 ii (5).

(2) The 1st and 2nd pl. took the forms of the pres. ind. early; and in the late period the 3rd followed.

Contracted Forms.

§ 185. i. (1) Verbs whose stems end in ‑o- or ‑a- (mostly from Brit, ‑og- or ‑od- and ‑ag-) have many contracted forms, more especially in the Mn. language. The following tables show all the possible contractions; the accent is marked in each case, and the accented vowels which are long in the present pronunciation are so marked, all others being short. Forms that are never contracted are distinguished by a hyphen, as parhḗ-ais. Any other form may occur uncontracted; thus trṓ-af as well as trōf occurs in Mn. W.

Exx. trṓf for trṓ-af ‘I turn’ (paratṓf for paratṓ-af ‘I prepare’); parhā́f for parhā́‑af ‘I continue’ (glanhā́f for glanha-af ‘I clean’).

Indicative Mood.
Present Tense.
sg.	pl.	sg.	pl.
1. trṓf	1. trówn	1. parhā́f	1. parháwn
2. trói	2. trówch	2. parhéi	2. parhéwch
3. trȳ́	3. trṓnt	3. pára, péry	3. parhā́nt
(3. paratṓ-a)		(3. glanhā́)
Impers. tróir		Impers. parhéir
Imperfect Tense.
1. trówn	1. trṓem	1. parháwn	1. parhā́em
2. trṓut	2. troech	2. parhā́ut	2. parhā́ech
3. trṓi	3. trṓent	3. parhā́i	3. pārhā́ent
Impers. tróid		Impers. parhéid
Aorist Tense.
1. tróis	1. troesom	1. parhḗ‑ais	1. parhásom
2. tróist	2. tróesoch	2. parhḗ‑aist	2. parhásoch
3. trṓes, trṓdd	3. tróesant, ‑ont	3. parhā́‑odd	3. parhásant, ‑ont
Impers. trṓ‑wyd, trṓed		Impers. parhā́‑wyd
Pluperfect Tense.
1. tróeswn, etc.		1. parháswn, etc.

Subjunctive Mood.

Present Tense.

	sg.		pl.		sg.		pl.
1.	trṓ-wyf	1.	trṓm	1.	parhā́-wyf	1.	parhā́-om
2.	trṓech	2.	trṓch	2.	parhḗ-ych	2.	parhā́-och
3.	trṓ	3.	trṓnt	3.	parhā́-o	3.	parhā́-ont
Impers. trṓ-er				Impers. parhā́-er

Imperative Mood.

Present Tense.

1.		1.	trówn	1.	parháwn
2.	trṓ	2.	trówch	2.	pára	2.	parhéwch
(2.	paratṓ-a)			(2	glanhā́)
3.	trṓed	3.	trṓent	3.	parhā́ed	3.	parhā́ent
Impers. trṓ-er				Impers. parhā́-er

Verbal Nouns.

trói (paratói), pára, parháu (glanháu)

Verbal Adjectives.

tro-édig, tro-ádwy, trṓ

(2) It is doubtful whether ‑er occurs contracted; the 3rd sg. impv. in ‑ed is mostly uncontracted.

(3) The contracted forms ‑ṓi, ‑ā́i of the 3rd sg. impf. are often pronounced and written ‑oe, ‑ae, see § 52 iii (3).

(4) The aor. stem ‑oes- is generally misspelt ‑ois- in Recent W.; thus troesom Es. liii 6, paratoesant 1 Bren. xviii 26 (so in 1620) appear as troisom and parottoisant! in recent bibles.

(5) On the 3rd sg. pres. para, pery see § 173 v (1).

ii. Stems ending in w̯-diphthongs have contracted forms when the endings ‑wn or ‑wch follow; thus táwn for táw-wn ‘let us be silent’, téwch for téw-wch ‘be ye silent’; gw̯randéwch for gw̯randéw-wch ‘listen ye’; clýwn for clýw-wn ‘we hear’, clýwch (re-formed clɥ́wch) for clýw-wch ‘ye hear’ or ‘hear ye’.

iii. Other vowels and diphthongs are not contracted; e.g. gweddī́-ir ‘there will be prayer’, cáe-ent (cáy-ent) ‘let them shut’, béi-id ‘fault was found’, dilḗ-er ‘may be deleted’, cynorthẃy-ynt ‘they assisted’. But for ‑ā́-odd in the 3rd sg. aor. we sometimes find -ā́dd in the Late Mn. period, e.g. cashâdd E.P. 222 for cashā́odd, gwellhā́dd C.C. 338; cf. cā́dd § 188 i (6).

§ 186. i. The full form rhoddaf ‘I give, put’, v.n. rhoddi, survives throughout as a literary form in Mn. W.; but in the living language the -ẟ- had already disappeared in the Ml. period, and the verb is also conjugated like trṓ-af, trṓf, in Ml. and Mn. W.; as roet (1 syll.) R.P. 1217 ‘was given’, roy A.L. i 6 ‘to give’; see § 110 iv (2) and § 33 iii (1). In the 3rd sg. pres. ind. rhydd the -ẟ survives in the spoken lang. (and is sometimes wrongly transferred to try); but rhy is used commonly in lit. W. as Duw a ry gwymp i'r drwg ŵr H.A. F. 10 ‘God will give the evil man a fall’. For rhy however, the compound dỿ́-rɥ is often found ; and dỿ́-ro for the and sg. pres. impv. rho ; by assim. of ỿ, dỿro became doro W.M. 53, 478, which is the form used in Gwynedd. The bards use forms with and without -ẟ- indifferently:

Rhoddi gwin yn rhwydd a gaid,

Rhannu a rhoi i weiniaid. — D.I.D., G. 179.

‘There was a generous giving away of wine, a distributing and giving to the weak.’

ii. In the subj. mood, we have r̔o-ho A.L. i 6, contracted to r̔o W.M. 23; and *r̔oẟ-ho giving rhoddo (r̔oto, -t- ≡ -ẟ-, B.B. 29), or rhotho by the comparatively rare change of ẟh to th (≡þ) § 111 iii (2).

A ro gam i wraig o Iâl,

Fo ry Duw rai a'i dial. — L. Mon, A 31059/78.

‘Whoever deals injustice to a woman of Yale, God will provide those who will avenge her.’

Maer Rhuthun im a'i rhotho. T.A., A 14976/169.

‘May the Mayor of Ruthin give it [the bow] to me.’

iii. Beside the aor. 1st and 2nd sg. r̔oẟeis, r̔oẟeist (r̔oteist, -t- ≡ -ẟ-, B.B. 30), a perfect was formed for these persons by adding -um, -ost to the aor. stem r̔oes-; see 182 iv (i); thus r̔oessum W.M. 63, Ỻ.A. 124 ‘I have given’. There is also a plup. 3rd sg. r̔oessoeẟ, 3rd pl. r̔oessoeẟynt; this survived in Early Mn. W. but seems to be used as a perf.:

Llaw Rys nid llai a roesoeẟ.—H.D., p 99/482.

‘The hand of Rhys has given no less.’ Other Late Ml. forms are r͑oẟassoeẟ, r͑ossoeẟ, Cymmrodor ix 77. Gwent rysoeddyd h.g. 30.

iv. On the origin of rhoddaf, see § 179 ii.

§ 187. i. arhoaf w.m. 17 ‘I wait’, contr. arhṓf, is conjugated like tro-af, trṓf, except that the v.n. is arhos w.m. 17, Mn. W. aros; thus Mn. W. ind. pres. sg. 1. arhṓf, 2. arhói, 3. éry; pl. 1. arhówn, 2. arhówch, 3. arhṓnt; impv. sg. 2. áro ‘stay!’ Ml. W. arho w.m. 17, aro do. 125; etc.

 gwaew hir gwae a’i héry.—I.H.S. 26.

‘Woe to him who awaits him with a long spear.’

Neidia goruwch hen adwy

I’r maes, ac nac aro mwy.—D.G. 30.

‘Jump over an old gap into the field, and stay no longer.’

Nid arhṓn’ hwy draean hyn.—I.F., m. 148/721.

‘They will not remain one third of this [time].’

ii. The above conjugation persisted well into the Late Mn. period, e.g. arhoent b.cw. 23, arhowch do. 102; but in the late 15th cent, a new formation sprang up in which the v.n. aros is substituted for the stem aro‑, giving arhosaf, etc. The earliest examples I have noted are in I.F.

Od ymddengys Rhys arhosaf.—I.F., m 148/301.

‘If Rhys appears, I will stay.’ So Arhoswch farn, rhoesoch fedd I.F. p 83/33, pan arhoser do. P 100/79.

iii. The only possible original of the ‑s of aros is either ‑d‑t- or ‑s‑t- (the v.n. suffix being *‑tu‑). The latter would imply ‑os- for the orig. stem; but where ‑s- came between vowels in Brit., the vowel before it was either lost, or contracted with the following vowel in Brit. itself, so that we could not have arho-af. We must therefore assume that ‑ẟ- has disappeared in this word as in rho-af (the ẟ of rhoẟaf being more or less artificial); hence arho-af for *ar-hoẟ-af < *ari-sod‑, √sed- ‘sit’; and aros < *ari-soss- < *pₑri-sod-tu- ‘sit before’; § 63 ii, § 110 iv (2).

§ 188. i. (1) caffaf ‘I shall get’ has stem kaff‑, kah- or ka- in Ml. W., and ca- in Mn. W. with ‑ff- in 3rd sg. pres. ind. and in subj.; and is conjugated regularly, except in the aor. The forms that occur are as follows.

(2) Indic. pres.: Ml. W. kaffaf w.m. 459, cahaf h.m. ii 126, caf w.m. 3; keffy w.m. 3, 23, 80, etc. (spelt kyffy 3, 460), kehy r.m. 120, key do. 293, 57, 118; ceif w.m. 25, 43 ≡ ceiff r.m. 16, 30; caffwn w.m. 34, cawn do. 84, r.m. 61; ceffwch r.m. 19, cewch w.m. 29; caffant w.m. 183; keffir w.m. 83, r.m. 60, keir w.m. 85, keffitor a 14869/56, O.W. cephitor CP.

Mn. W. cāf; céi, cái; céiff, cáiff; cáwn; cewch; cā́nt; céir, cáir, § 81 iii (1).

(3) The impf. in Ml. W. has kaff- or ka- in the indic.: cawn w.m. 394, r.m. 251, caffut w.m. 396, r.m. 253; subj.: pei caffwn w.m. 18, r.m. 12. In Mn. W. cáwn, cā́ut, cā́i, etc., and sometimes caffwn etc. in the subj.

(4) The pres. subj. seems to have kaff- chiefly: caffwyf w.m. 454 (twice); keffych do. 480 (4 times); kafont, kafoent (f ≡ ff) b.ch. 4, etc.; but caho Ỻ.A. 150, caont do. 48. Mn. W. has caff- only.

(5) Impv.—The vb. implies an absolutely passive ‘getting’ or ‘catching’ (as ‘catching’ a cold), and so has never been used in the impv. except in the 3rd pers. (or impers.), in which case the command is not addressed to the subject, and its carrying out is independent of his will. The forms are Mn. W. 3rd sg. caffed, caed, 3rd pl. caffent, caent; impers. caffer.

(6) Aorist.—The Ml. W. forms (all of very frequent occurrence except the 2nd pl.) are, sg. 1. keveis, 2. keveist, 3. kavas; pl. 1. kawssom, ‑am, (2. kawssawch), 3. kawssant, ‑ont; impers. kaffat, kahat. (The apparent contraction a geis r.m. 253 is almost certainly a scribal error for a ge(ve)is, cf. w.m. 395.) The Mn. W. forms are sg. 1. cefais, 2. cefaist, 3. cafas § 175 i (2), later cafodd; pl. 1. cawsom, 2. cawsoch, 3. cawsant. In the 14th cen., the following contracted forms are found, sg. 1. ces D.G. 124, G.Gr. d.g. 254; sg. 3. cas D.G. 294; impers. a gat r.p. 1299, cad D.G. 189, 409, 429, 430. Later are found cḗs; cḗst; cā́s and cā́dd D. 130, cadd M.K. [61]; impers. cafad B.Br. f. 6, cā́d; cā́ed (prob. orig. a false spelling of cā́d); cafwyd (cáwd c.c. 271, a dial, form used in late verse § 175 iv (5)).

(7) Pluperf.—The forms are Ml. kawsswn, etc., Mn. cawswn, etc., conjugated regularly. In Ml. W. is also found a plup. formed with ‑oeẟ: sg. 1. kawssoeẟwn s.g. 278; sg. 2. cawssoeẟut do. 247; sg. 3. kawssoeẟ do. 303, cawssoeẟei h.m. ii 170, cawssoeẟ-ẏat s.g. 30, ‑at h.m. ii 224; pl. 3. kawssoeẟynt s.g. 11. It is seen that the forms are found in Late Ml. mss. They are also used occasionally by Early Mn. bards, e.g. cawsoedd L.G.C. 18.

(8) Verbal Noun.—Ml. W. caffael w.m. 12, kaffel r.m. 8, 141, cael w.m. 13, r.m. 8 (once, caffu b.b. 53). Mn. W. caffael, caffel, cā́el.

There is no *cavael; the form cafael w.m. 60 ≡ kaffael r.m. 43. Nettlau’s cauael does not exist; the word is gauael (≡ gavael) r.m. 7, see below.

ii. (1) gafaelaf ‘I take hold’ is conjugated regularly in Ml. and Mn. W. with the v.n. gavael as stem.

(2) The Ml. W. inflected forms are mostly those of the compound ym-afaelaf; e.g. 3rd sg. pres. ind. ymeveil w.m. 70, 71; 3rd sg. aor. ymavaelawẟ r.m. 50.

(3) The verbal noun is gavael w.m. 11, r.m. 7, ymavael r.m. 142, ymavel ib.; Mn. W gafael, gafel, ymafael, ymafel.

(4) Other forms of the verb occur in Late Mn. W.: ymaflaf, 3rd sg. pres. ind. ymeifl, v.n. ymaflyd; and ymaelaf, v.n. ymaelyd; and re-formations from the form gafel of the v.n. occur dialectally, as gafelaf etc.

iii. dyrchafaf ‘I raise, lift up’ is conjugated regularly. It is also written drychafaf. The form derchafaf occurs in mss. which use e for ỿ as m.a. ii 316. The v.n. is dyrchavael w.m. 39 or dyrchavel r.m. 271; in Late Mn. W. this is superseded by dyrchafu; v. adj. dyrchafedig ‘exalted’.

The 3rd sg. pres. ind. is Ml. W. dyrcheif h.m. ii 274 or drycheif r.b.b. 144, Mn. W. drychaif g. 138, there printed dyrchaif the usual form. There is also in Ml. W. dyrchevid b.b. 82 ‘raises’. The 2nd sg. impv. is dyrchaf s.g. 23, L.G.C. 144, becoming dyrcha Ps. iv 6 by the loss of ‑f § 110 iii (5). From this a 3rd sg. pres. ind. dyrcha came into use in Late Mn. W., e.g. Ps. xxvii 6, Gr.O. 88; which some recent writers have improved to dyrch, with v.n. dyrchu!

¶ A list of the forms of the above three verbs occurring in r.m. and part of h.m. ii is given by Max Nettlau in Cymmrodor ix 111 ff., but is inaccurate in some details, e.g. i (8) above.

iv. The facts in i show that the stem of caffaf is caff- or cah‑. The form cav- occurs in the aor. sg. only, and must have been deduced from the pl. at the stage between cawsant and *caffsant from the orig. caff‑. In Bret. kaf- (≡ kaff‑) remains in forms ordinarily unvoiced, and kav- is extended to others; but forms like kef (= W. ceiff), beside kav, survive to bear witness to the original stem kaf- in Bret. also.

caffael and gavael seem to contain the doublet *qap‑: *ꬶhabh- § 101 iii (2). The v.n. gavael has its exact equivalent in Ir. (ath‑)gabāil from *gab-ag-li- formed with suff. ‑li- from a compound of *ꬶhabh- and *ag̑- § 203 i (4). The vb. in Ir. is gabim, and the W. gafaelaf prob. replaces an old *gaf-af equivalent to the Ir. (Dialectal gafaf is no doubt new.)

The W. stem caff- or cah- represents *qap‑s‑, § 96 iv (3); hence caffaf from the fut. *qapsō, with the usual reconstruction which gives e.g. ad-feraf from *bherō. The pres. caffaf, caf is always fut. in meaning; and recent writers have used a fictitious 3rd sg. ca ‘gets’ because caiff means ‘will get’. (The pres. sense can only be expressed periphrastically: yr wyf yn cael ‘I am getting’.) The v.n. caffael, cael is perhaps formed on the analogy of gafael.

It may be objected that dyrchafaf ‘I raise, lift up’ shows stem *cav‑. But there is no reason whatever for the supposition that this verb has anything to do with the others. The prefix dyr- must represent *do-(p)ro- § 156 i (13), which cannot give ‑ch- from k- or g‑. The root seems to be *sqabh- ‘fix, hang’: Skr. skabhnā́ti ‘fixes, supports’, O.Bulg. skoba ‘fibula, clasp’, Lith. kabù ‘I hang’. *sqabh- gives ‑chaf- regularly, § 96 iii (4). The v.n. dyrchafael may be a similar formation to gavael, or, as is more likely, formed like gadael and gallael on its analogy, § 203 i (2).

Irregular Verbs.

The Verb ‘To Be’.

§ 189. i. The following table shows the Ml. W. forms of the verb ‘to be’. Nearly all are used in Mn. W., so that it is unnecessary to repeat them for that period. Forms that became obsolete in Mn. W. are marked †; where the Mn. form or spelling differs it is given in (   ).

Indicative Mood.

Present Tense.

	sg.		pl.
1.	wyf, ydwyf, yttwyf	1.	ym,ydym, yttym
2.	wyt, ydwyt, yttwyt	2.	ywch,ydywch (ych,ydych)
3.	yw, ydiw, yttiw (late ydyw), y mae, mae, oes	3.	ynt, ydynt, yttynt, y maent, maent
	†yssit ‘there is’, †ossit ‘if there is’, ‑s in os ‘if it is’		†yssydynt
	Relatival form: yssyẟ (y sydd), syẟ, yssy (y sy), sy.
	Impersonal: ys, ydys, yttys.
	Conjunctive: y mae or mae (late mai), †panyw, (dial. taw).

Consuetudinal Present and Future.

1.	byẟaf, †byẟif	1.	byẟwn
2.	byẟy (byddi)	2.	byẟwch
3.	byẟ	3.	byẟant
	Cons. bit (bid)		†byẟhawnt, †bint
	Fut. †bi, †byẟhawt, †biaw
Impers. (byddys, byddir)

Imperfect.

	sg.		pl.
1.	oeẟwn, †yttoeẟwn	1.	oeẟem
2.	oeẟut (‑ud, ‑it)	2.	oeẟewch (oeddech)
3.	oeẟ, yttoeẟ (ydoedd)	3.	oeẟynt, †yttoeẟynt
Impers. oeẟit (oeddid)

Consuetudinal Imperfect.

Sg. 1. byẟwn, etc. regular.

Perfect.

1.	buum, bum (bûm)	1.	buam, ‑om
2.	buost	2.	buawch (buoch)
3.	bu	3.	buant, buont
Impers. buwyt (‑wyd)

Pluperfect.

Sg. 1. buasswn (buaswn, baswn), etc. regular; pl. 3. buyssynt, beside buassynt, ‑essynt § 175 iv (1). Also sg. 3. †buei, etc.

Subjunctive Mood.

Present Tense.

	sg.		pl.
1.	bwyf, byẟwyf	1.	bom (bôm), byẟom
2.	bych, byẟych	2.	boch, byẟoch
3.	bo, byẟo, bytho	3.	bont (bônt), byẟont, bythont, boent, †bwynt

Imperfect Tense.

1.	bewn (bawn), byẟwn	1.	beym (baem), byẟem
2.	beut (baud, ‑it), byẟut (‑ud, ‑it)	2.	(baech, byddech)
3.	bei (bai), byẟei (‑ai), pei (pe)	3.	beynt (baent), byẟent
Impers. byẟit (‑id), bythit (‑id)

Imperative Mood.

Present.

	sg.		pl.
1.		1.	byẟwn
2.	byẟ	2.	byẟwch
3.	bit (bid), boet (‑d) poet (‑d)  (bydded)	3.	byẟent, bint

.

Verbal Noun.

bot (bod)

¶ For a list of Ml. forms, with references, by Dr. J. G. Evans, see bb. 109 ff.

ii. Notes.—(1) Pres. ind.—Forms with ytt- (yt‑) appear in poetry in Mn. W. but are comparatively rare.

Trist fu’r glêr tros dy fawr glwyf,

Trist éto trosot ýtwyf.—G.G1., m 146/161.

‘Sad have been the minstrels for thy sore sickness, sad still am I.’

The 3rd sg. ydiw was so written up to the 16th cent.; and rhymes with words in ‑iw, as friw/ýdiw D.G. 35, cf. 119, 144, 193, etc., and g. 186, 193, 203, 206, 235, 247, also with yw (≡ iw); see § 77 v. The Late Mn. ydyw is an etymological spelling, and is read ỿ́diw, except by a few affected persons. The N.W. dial. form is ỿ́di (and, in answering questions only, ń̥di, a curious attempt to sound ỿ with the tongue in the d position). S.W. dial., in questions and answers, ódi.

ydys is sounded ỿ́dɥs; on ys see § 82 ii (1). In Mn. W. ỿd- and the rare ỿt- come only before monosyllabic forms, and always take the accent, †yssydynt w.m. 457 is formed from yssit § 162 vi (1).

The Late Mn. spelling mai of the conj, form seems to come from mai ḗ § 219 vi (1); elsewhere the pronunciation is mae ≡ maɥ or ma’; the form mai owes its adoption to the popular notion that a conjunction ‘that’ must differ from a verb ‘is’. The word means, not ‘that’, but ‘that it is’; as gwn mae Dafydd a’i gwnaeth ‘I know that it is D. who made it’.

(2) The consuetudinal pres. is in use in that sense in the spoken lang. (in N.W.), but the fut. is a commoner use. The form bit (bid) is mostly impv., see (5); but it is sometimes indic. even in Mn. W., owing doubtless to the survival of proverbs such as bid anwadal ehud ‘the fool is changeable’; thus

Bid gwaeth gwybodau a gair

Beirdd gwedi bardd y gadair.—Gu.O., M 146/450 (m. D.E.)

‘The sciences and renown of bards are worse after the [death of] the bard of the chair.’ Cf. bid sicr ‘it is certain, to be sure, of course’.

The forms †bi b.t. 12, †byẟhawt w.m. 456, etc. are fut. only.

(3) In the impf. the consuet. ind. byẟwn is distinguished from the subj. bewn (bawn); the latter is never ind., but the former is used in the subj., as pei byẟut, etc. Ỻ.A. 67; also bythit w.m. 104, cf. (4).

The form pei for *pei y, before a vowel pei yt, ‘were it that’ is used in the sense of ‘if’ with the impf. subj. or plup. With the 3rd pers. infixed pron. ’s, it is bei ys w.m. 424, later pei ass w.m. 17. In Mn. W., the forms are pe, ped, pe’s; also with b‑: be g. 128, 238, etc.

As the subj. stem seems to have been b- or p- the orig. form of sg. 1. 2. should be bwn, *but like 3. bei; so in the pl. The phrase pei yt vwn, occurring as bei et-vwn w.m. 71, was contracted early to pettwn ‘if I were’, 2. pettut, 3. pettei; pl. 1. pettem, etc. Thus bettut kynn decket ac Absalon Ỻ.A. 67 ‘if thou wert as fair as A.’; pettei do. 68; Mn. W. pettwn b.cw. 10 ‘if I were’, petynt ‘if they were’. But pei byẟei Ỻ.A. 67–8, be bai H.D. p 99/494, etc., are also used.

Traces occur of an old plup. with stem bu‑: sg. 3. buei r.p. 1045, bwyat (read bu-ẏat) do. 1038, pl. 3. bỽyn (read bu-yn) ib., buyint b.b. 96.

(4) Beside the pres. subj. proper bwyf, the form byẟwyf with ind. stem is used; also bytho T.A. c. i 342, bythont w.m. 47, with byẟ + h‑, a new subj. stem.—The impers. boer m.a. i 20 is doubtful; the context suggests sg. 3. bo. But E.P. ps. xciv 13 uses boer.—3rd pl. bwynt b.t. 5; boent a.l. i 106, L.G.C. 240.

(5) As stated above (2), bit (bid) is usually impv.: Bit ẏ waet ef arnam ni s.g. 25, Ỻ.A. 83 ‘His blood be upon us’; na vit ofɏn arnawch r.m. 147 ‘let there be no fear on you’ i.e. fear not; bit w.m. 22, r.m. 14 ‘let there be’. The form bint Ỻ.A. 81 ‘let them be’ is formed from bid; it is rare in Mn. W., L.G.C. 240.

iii. (1) For the origin of w͡yf, w͡yt, yw, ym, ych, ynt, see § 179 ix (3). yd- is the affirmative particle § 219 ii; yttynt < *yd hynt; from this ytt- spread to other persons.

(2) y mae, mae occurs at the beginning of a positive statement, or positive rel. clause; it seems to have meant originally ‘there is’ or rel. ‘where is’, since mae at the beginning of a question means ‘where is?’ Thus mae ymma Matholwch w.m. 39 ‘there is here M.’, y lle ymae Abel Ỻ.A. 118 ‘[in] the place where Abel is’, mae y mab? w.m. 29 ‘where is the boy?’ The m- of mae is never mutated; this points to *mm (Corn. ‑mm‑) < *sm. The y m- is prob. ym- (often so written in Ml. W.) representing the locative in ‑smi of the *e- demonstrative (nom. sg. *es § 159 iv (1)), as in Umbr. loc. esme ‘in hoc’ < *esmi, Av. ahmi. Thus *esmi est, ‘here is, there is’ pronounced *ésmii̯est > *ymoeẟ § 75 iv (2), whence by loss of ‑ẟ and the change of oe to ae after a labial § 78 i (1) and ii (2) we have ymae. The rel. form similarly from *i̯osmi est. The interrogative form mae ‘where is?’ appears to be a new development in W., with the y- dropped because it seemed to be affirmative; it prob. comes from indirect questions in which mae is rel., as manac imi mae Arthur w.m. 123 ‘tell me where Arthur is’. Corn. has pyma? as if from *q^u̯osmi est? The pl. y maent (≡ ỿmā́ɥnt) must be a new formation from y mae.— The Bret. form is ema, ma, Corn. yma, ymma, ma, pl. ymons; the last form confirms the assumption of oe by preserving the o.

(3) oes occurs after nyt (nid), nat (nad), the interr. part. a, and od ‘if’, in each case when the subject is indefinite. nyt oes represents *nitaisti < *n’ ita esti ‘there is not’. The positive *esti ita ‘there is’ > *estīta > yssit. Similarly ossit ‘if there is’ < *ā ’stīta § 222 v (1). In Ml. W. yssit is only a survival, having been generally replaced by y mae. As nid oes means literally ‘there is not’, it is natural that its subject should be indefinite. But early examples of a definite subject occur: cinnit hoys ir loc guac hinnuith in pag. reg. cp. ‘though there is not that empty place in the regular page’; nat oes hi w.m. 470 ‘that there is not [such a one as] she’; in r.m. 113 this becomes nat ydiw y vorwyn ‘that the maid is not’.

(4) yssyẟ, syẟ, etc. < *estíi̯o < *estí i̯o § 162 vi (1). ys < *esti § 179 ix (3)—panyw ‘that it is’ § 222 x (2).

(5) oeẟ see § 75 iv (2), § 180 ii (3), yttoeẟ § 180 ii (3), q.v.

iv. (1) From √bheu̯ā- ‘be’ there was an iterative derivative *bh(u̯)ii̯ō which gives Ir. bīuu ‘I am wont to be’, Lat. fīo. The three persons of the sg. *bhu̯íi̯ō, *bhu̯íi̯ēis, *bhu̯íi̯ēit would all give W. byẟ, which was afterwards inflected byẟaf, byẟy, byẟ by analogy. In Kelt., Ital., Germ., there are also athematic forms of this verb; thus there were sg. 2. *bhu̯ī-si > Lat. fīs, 3. *bhu̯ī-ti > Lat. fīt, W. bid. [Lat. fīo takes its long ī from these.] The Early Ml. W. fut. bi is a future of this form, representing *bhu̯ī-sēit (or *bhu̯ī-ēit?). The forms byẟhawt, biawt are of course formed by adding ‑(h)awt to byẟ, bi.

(2) The opt. of *bh(u̯)ii̯ō, sg. 1. *bh(u̯)íi̯oi‑m̥ might give byẟwn, but prob. the whole tense is a later formation from byẟ.

(3) The perf. bu-um, etc. is obviously formed from the 3rd sg. by the addition of the perf. endings ‑um, etc. § 182 iv (1). The 3rd sg. bu, Ir. bōi, bāi represent Kelt. *(be‑)bāu̯e < Ar. *bhe-bhōu̯e: Av. bavāva; § 76 iii (5).

(4) The pres. subj. bwy(f) represents the ‑se- fut. of √bheu̯ā‑; thus *bh(u̯)ā-sō > bwy etc. § 183 ii.

The impf. subj. sg. 3. bei < *bii̯ī́t < *bai̯ī́t < *bh(u̯)ə-si̯ḗt. From bei was deduced bwn as in bei et-vwn ii (3); but later bewn, as if bei were *be-ei; in Mn. W. when bei had become bai, the 1st sg. became bawn; and in the late period bai itself came on the analogy of this to be treated as bâi and sometimes written bae, see § 185 i (3).

The initial p- is for *b-h- with ‑h- from pl. forms; see § 183 ii (3).

(5) The impv. sg. 2. byẟ is from *bh(u̯)íi̯e the crude stem of *bh(u̯)íi̯ō. The 3rd sg. bid is from *bh(u̯)ītó the 3rd sg. opt. mid. of stem *bhu̯ā‑; see § 184 ii (1) and § 180 iv (2). The 3rd sg. boed or poed is a re-formation from the subj. stem. The pl. forms are obvious re-formations.

(6) The v.n. bod implies Brit. *butā, which (as there is both in Ir. also) may be a Kelt. formation beside *bhu-t‑is which gives Ir. buith: Gk. φύσις. Like other v.n.’s bod has been made mas.; but in compounds it remains f., as ha-fod, eistedd-fod, preswyl-fod.

Compounds of the Verb ‘To Be’.

§ 190. i. (1) The verbs of the v.n.’s cánfod ‘to perceive’, dárfod ‘to waste away; to happen’; górfod ‘to overcome’; hánfod ‘to be from; to come’, are conjugated with the b-forms of the verb ‘to be’; as canffýddaf, etc. In Ml. W. canfod appears generally with the pref. ar‑.

Pres. (fut.) ind.: sg. 1. gorvyẟaf c.m. 61, 70;—2. henbyẟy w.m. 97;—3. dervyẟ c.m. 43, gorvit (≡ gorvyẟ) b.b. 52; dy-ẟerbi r.p. 578, dy-worpi do. 585;—impers. gorvyẟir w.m. 82, r.b.b. 152, c.m. 13.

Impf. ind. sg. 1. gorvyẟwn w.m. 131;—3. hanbyẟei w.m. 141.

Perf.: sg. 1. Mn. canfū́m § 191 ii (5);—3. kanvu r.p. 1143, arganvu c.m. 50, s.g. 7, darvu c.m. 59, gorvu w.m. 89;—pl. 1. darfuam b.b. 105;—3. darvuan b.b. 6;—impers. arganvuwyt w.m. 49, darvuwyt r.p. 1296.—Plup.: pl. 3. gorvuassynt c.m. 68.

Pres. subj.: sg. 1. hanbwyf m.a. i 301b;—2. hanpych gwell r.m. 87, w.m. 185, s.g. 1, hanbych well p 16/44, Mn. W. henffych well ‘may you come well!’ i.e. welcome! (gwell not orig. cpv. § 148 i (4));—3. darffo s.g. 17, c.m. 42, 59, gorpo b.b. 17, hanffo Ỻ.A. 131, c.m. 33;—pl. 3. gorffont r.b.b. 222;—impers. gorvyẟer c.m. 13, gorffer do. 22.

Impf. subj.: sg. 3. darffei c.m. 68, 29, gorffei r.m. 163, hanphei c.m. 55, hampei do. 58.

Irnpv.: sg. 3. derffit r.p. 1044, r.m. 155; dervhid b.b. 91.

V.n. arganvot w.m. 54, darvot c.m. 32, gorvot w.m. 56, hanvot do. 460;—v. adj. darvodedic Ỻ.A. 86, Mn. W. darfodedig ‘perishable’.

(2) In darfod two verbs have prob. merged: (a) darfod ‘to waste away, to perish’ < dar‑: Gk. φθείρω § 98 i (4);—(b) darfod ‘to happen’ < *do-ári- § 156 i (13). The latter is used in the 3rd sg. only, see § 196, as Beth a ẟarvu uẟunt wy? Ỻ.A. 7 ‘What happened to them?’; often as a so-called “auxiliary”; as pei na ẟarffei ẏ’r dwst gyvodi c.m. 68 ‘if the dust had not risen’; deryw in Ml. W. is generally thus used. In Mn. W. it is replaced by darfu; but the pres. had a past force from the sense of ‘afore(time)’ in the prefix. The v.n. darfod introduces noun-clauses corresponding to direct statements with deryw, as Ml. W. wrth ry-ẟarvot iẟaw ẏ r͑oẟi s.g. 32 ‘since he had given it’.

Examples: (a) derfydd f. 27 ‘will perish’, darfu D.G. (§ 160 i (1)) ‘is spent’, darfyẟant Job iv 9 ‘they perish’, ni ddarfu 1 Bren. xvii 16 ‘wasted not’, darvuan b.b. 6 ‘they perished’;—(b) darffo 1 Cor. xv 54, darfu ’m (for darfu ym) Gr.O. 98 ‘it happened to me’, i.e. I did; y darffai d. 112; a vynno Duw derffit r.m. 155 lit. ‘what God will let it come to pass’.

(3) canfod, gorfod and hanfod contain respectively the prefixes cannh- § 156 i (7), gor- do. i (17), and han‑ do. ii (3).

gorfod is chiefly used in the 3rd sg. in Mn. W., as gorfu i or gorfu ar ‘was obliged’ § 196 vii. For the verb the v.n. is often used, as gorfod iddo for gorfu iddo, cf. a heẟiw yn gorvot arnam … ymwahanu c.m. 50 ‘and to-day we must part’.

ii. (1) In addition to the above forms Ml. W. has a pres. and impf. formed with ‑wyf and ‑oeẟwn. These survived in Early Mn. W. Before ‑yw, ‑ym, ‑ywch, ‑ynt, ‑a- is affected to ‑e‑; the ‑e- often intrudes into forms with ‑wyf, ‑wyt, and vice versa ‑a- often occurs before ‑yw, etc. Thus:

Indic. pres.: sg. 1. hanwyf w.m. 3, henwyf r.m. 2, cannwyf D.G. 200;—2. hanwyt w.m. 3, 191, henwyt r.m. 2;—3. cennyw r.p. 1433, D.G. 205, deryw, derw w.m. 99, henyw s.g. 13, hanyw L.G.C. 9;—pl. 1. henym Ỻ.A. 164;—3. henynt Ỻ.A. 169.

Mil ar benn bryn a’i cennyw.—G.Gl., p 75/159.

‘A thousand behold it [the mansion] on the top of the hill.’

Na sonier am a dderyw.—I.G. 289.

‘Let there be no mention of what has happened.'

Mawrserch Ifor a’m goryw;

Mwy na serch ar ordderch yw.—D.G. 3.

‘The great love of Ivor overcomes me; it passeth the love of woman.’

Impf.: sg. 3. canhoeẟ w.m. 64, r.m. 46 ‘could see’, daroeẟ s.g. 25 ‘happened’, hanoeẟ do. 41;—pl. 3. hanhoeẟynt s.g. 15, r.p. 1047.

Ymddiried im a ddaroedd.—G.Gl., m 146/168.

‘He trusted in me.’ (Elliptical, for a ẟaroeẟ iẟo ‘happened to him’, i.e. he did.)

O’r hen arglwyddi’r hanoedd.—L.G.C. 2.

‘She was descended from the lords of old.’

O'r hen wŷdd yr hanoeddych.—I.H.S., Ỻ 133/2 1 2. ‘You are descended from the old stock.’

(2) Beside hanwyf etc., Ml. W. has handwyf, handwyd, handid, handym, handoetud, handoet (t ≡ ẟ) all in m.a. i 358, handid b.b. 33, 107, handoeẟ r.p. 1432, handoet w. 1a.

These seem to be formed from an extension of the prefix, such as *sani-ti, cf. hefyd § 220 ii (8), giving before a vowel hand- § 113 i (2); by analogy *hand-fid > hand-id, cf. § 110 iii (3); handyvyt m.a. i 358 makes the line too long.—ny handei w.m. 183, r.m. 85 makes no sense; a better reading seems to be ny hanẟenei p 16/43 (w.m. p. 92) ‘he could not rest’ (hanẟen, by dissim. > Mn. W. hamẟen ‘leisure’, hamẟenol ‘leisurely, slowly’; han- ‘without’ + den, √dhen‑: Skr. dadhan-ti ‘causes to run’).

iii. The verb cyfarfyddaf â ‘I meet’ is conjugated like the above verbs (v.n. kyvarvot w.m. 58, 125, perf. sg. 3. kyvarvu do. 170, plup. sg. 3. cyfarvuassei ib.), except that the old forms were obsolete in Late Ml. W. But D.B. has kyveryw a mi r.p. 1385 ‘has met me, happened to me’; and ry-gyveryw a occurs in w.m. 42, changed to ry-gynneryw a in r.m. 29, as if it were a compound of deryw, the form cyveryw being apparently unknown, and the u (≡ v) mistaken for n.

iv. In the dialects darfyddaf and cyfarfyddaf, the most commonly used of these verbs, are mostly conjugated as if they were regular verbs; and such barbarisms as darfyddodd, cyfarfyddais, canfyddais occur in recent writings. The impf. hanoedd seems to have survived the other obsolete forms; this was mistaken for an aor. hanodd, from which was inferred an imaginary v.n. hanu, common in recent biographies.

§ 191. i. (1) The verbs gwnn (gwn) ‘I know’, v.n. gwybot (gwybod), and adwaen ‘I am acquainted with’, v.n. adnabot (adnabod), are conjugated as follows in Ml. (and Mn.) W.

Indicative Mood. Present Tense.
sg.		pl.
1.	gwnn (gwn)	1.	gwẟam, ‑om (gwyddom)
2.	gwẟost (gwyddost)	2.	gwẟawch, -och (gwyddoch)
3.	gwyr (gŵyr)	3.	gwẟant (gwyddant)
Impers. gwys (gwŷs, gwyẟys)
1.	adwaen, adwen, atwen (adwaen, adwen)	1.	adwaenam, adwaenwn (adwaenom, adwaenwn)
2.	atwaenost (adwaenost, adweini)	2.	adwaenawch, atweynwch (adwaenoch, adwaenwch)
3.	adwaen, adwen, atwen (adwaen, edw̯yn)	3.	atwaenant (adwaenant)
Impers. (adwaenir, adweinir)
Future Tense.
sg.		pl.
1.	gwybyẟaf (gwybyddaf)	1.	gwybyẟwn (gwybyddwn)
2.	gwybyẟy (gwybyddi)	2.	gwybyẟbwch (gwybyddwch)
3.	gwybyẟ (gwybydd)	3.	gwybyẟant (gwybyddant)
Impers. gwybyẟir (gwybyddir)
1.	adnabyẟaf (adnabyddaf)	1.	adnabyẟwn (adnabyddwn)
2.	adnabyẟy (adnabyddi)	2.	(adnabyddwch)
3.	adnebyẟ, ednebyẟ (adnebydd)	3.	adnabyẟant (adnabyddant)
Impers. adnabyẟir (adnabyddir)
Imperfect Tense.
1.	gwyẟwn, gwyẟẏwn (gwyddwn)	1.	gwyẟem, gwyẟẏem (gwyddem)
2.	gwyẟut, gwyẟẏut (gwyddud, gwyddit)	2.	gwyẟewch (gwyddech)
3.	gwyẟẏat, gwyẟat, gwyẟẏei (gẃyddi̯ad, gwyddai)	3.	gwyẟynt (gwyddynt, ‑ent)
Impers. gwyẟit (gwyddid)
1.	atwaenwn (adwaenwn)	1.	adwaenem (adwaenem)
2.	atwaenut (adwaenud, ‑it)	2.	(adwaenech)
3.	atwaenat (adwáeni̯ad, adwaenai)	3.	atwaenynt (adwaenynt, ‑ent)
Impers. etweinit (adwaenid, adweinid)

Perfect Tense.
sg. 1.	gwybuum (gwybū́m)	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \ \end{matrix}}\right\}\,}$	etc. like canfū́m
	adnabuum (adnabū́m)
	Impers. gwybuwyt, adnabuwyt (gwybū́wyd, adnabū́wyd)

Pluperfect Tense.
sg. 1.	gwybuasswn (gwybuaswn)	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \ \end{matrix}}\right\}\,}$	etc.
	adnabuasswn (adnabuaswn)

Subjunctive Mood. Present Tense.
sg. 1.	gwypwyf (gẃypwyf, gwybýddwyf)	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \ \end{matrix}}\right\}\,}$	etc.
	adnapwyf (adnápwyf, adnabyddwyf)
sg. 3.	gwypwy, gwypo, adnapo (gẃypo, gwybýddo, adnápo, adnabýddo)

Imperfect Tense.
sg. 1.	gwypwn, gwybyẟwn (gẃypwn, gwybýddwn)	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \ \end{matrix}}\right\}\,}$	etc.
	adnapwn, adnabyẟwn (adnápwn, adnabýddwn)

Imperative Mood. Present Tense.
	sg.		pl.
		1.	gwybyẟwn (gwybyddwn)
2.	gwybyẟ (gwybyẟ)	2.	gwybyẟwch (gwybyddwch)
3.	gwypet, gwybyẟet (gwyped, gwybydded)	3.	gwypent, gwybyddent, (gwypent, gwybyddent, ‑ant)
Impers. gwyper, gwybyẟer (gwyper, gwybydder)
		1.	adnabyẟwn (adnabyddwn)
2.	ednebyẟ, adnebyẟ (adnebydd)	2.	adnebyẟwch (adnabyddwch)
3.	(adnabydded)	3.	(adnabyẟent, ‑ant)

(2) The verb cỿdnabỿddaf ‘I recognize’, v.n. cydnabod, has pres. ind. cỿdnabỿddaf, impf. ind. cỿdnabỿddwn, and the rest of the verb like adwaen.

ii. (1) In the pres. indic. the endings of the 2nd sg. and the pl. are seen to be those of the perf. and aor. In the dialects the 3rd pl. has ‑on beside ‑an.

But adwaen has also the pres. endings; thus beside adwaenam Ỻ.A. 164 ‘we know’ we find adwaenwn w.m. 25 ‘we know’; so atweynwch c.m. 12; Mn. W. adweini Es. lv 5 ‘thou knowest’.

(2) Both the 1st and 3rd sg. pres. ind. were adwaen or adwen; the Mn. W. 3rd sg. edw̯yn is a new formation from adwen (on the analogy of etyb ‘answers’ §173 iv (i)). Examples: 1st sg. adwaen b.b. 102, atwaen s.g. 72, atwen w.m. 390; 3rd sg. attwen h.m. ii 235, Atwen mab ae llocha, ac nyt atwen ae kar r.b. 964 ‘a child knows who fondles him, but does not know who loves him’; pawb adwaen pwy I.G. g. 79 ‘everybody knows who’.

Yr ydwyf, hyd yr adw̯en,

Yn dwyn haint ni’m gad yn hên.—D.G. 443.

‘I am, as far as I know, suffering from a disease that will not spare me to old age.’

A’r un sud, er nas edw̯yn,

Y inesur Duw amser dŷn.—B.Br., f. 15.

‘And in the same manner, though he knows it not, does God measure man’s life.’

(3) The 3rd sg. impf. ind. has the ending ‑ẏat, ‑at, Mn. W. ‑i̯ad. The ending ‑ei is rare in Ml. W.: gwyẟyei b.a. 6, r.p. 1264. The ‑ẏ- (≡ i̯) in gwyẟẏwn etc. doubtless spread from ‑ẏat; it did not come into general use. In Mn. W. ‑i̯ad survived in poetry, but gwyddai, adwaenai became the usual forms. See gwyẟẏat w.m. 183, r.m. 85, s.g. 11, atwaenat s.g. 72, w.m. 150.

Yr oedd i rai a wyddiad

Obaith dyn o fab i’th dad.—T.A., a 14694/117.

‘There was, to those who knew, hope of a man in a son of thy father.’ See adwaeniad D.G. 430, T.A. g. 234.

(4) On the ‑t- for ‑d- before w̯, see § 111 v (2).

(5) Note the accentuation of gwybū́m, adnabū́m, in which the last syllable has a late contraction, § 41 iii. Uncontracted gwybū́-um occurs as late as the 15th cent.; see § 33 iv. The 3rd sg. gẃybu, adnábu has no contraction, and is accented regularly.

O’r tad Hywel ap Cadell,

Nid adnabū́m dad neb well.—T.A., c 84/849.

‘[Sprung] from his father, H. ap C.,—I have not known a better father to anyone.’ Cf. adnabûm, so accented, b.cw. 105; so canfûm do. 16, 91. Ml. W. gwybuum w.m. 389, adnabuum ib.

iii. (1) gwnn probably comes from *u̯indṓ § 66 iii (1), or middle *u̯indṓi: Skr. vindá-ti ‘finds’, Ir. ro-finnadar ‘is wont to know’, √u̯eid- with ‑n- infix. The 3rd sg. gŵyr seems to be a deponent form made by adding the impers. *‑re directly to the root § 179 viii (2); thus *u̯eid-re > *u̯eig-re (§ 104 iv (3)) > gŵyr.—The 2nd sg. gwẟost represents a periphrastic form *u̯idós’si, verbal adj. + verb ‘to be’, the remnant of a tense like euthum, re-formed in the pl. with aor. endings § 182 iv (1). In Mn. W., and occasionally in Late Ml. W. gwẟ- becomes gwyẟ- on the analogy of the other tenses. The impers. gwŷs prob. represents a passive *u̯id-tos (’st).

The tense replaces the old perf. with pres. meaning, *u̯oida: Gk. οἶδα.

(2) The impf. 3rd sg. gwyẟiad may be for *gwẟiad § 180 iv (1). The 2nd sg. gwyẟut may represent a thematic *u̯eidoithēs, in which case its wy is original; and the 3rd sg. may have taken wy from this. The wy is the falling diphthong: Pob meistrolrw͡ydd a w͡yddud D.G. 460.

(3) The rest of the verb comes from periphrastic tenses formed of a present participle of some such form as *u̯eidans and the verb ‘to be’.

iv. (1) adwaen corresponds to Ir. ad-gēn, which comes from *ati-gegna, re-formed in Kelt. for *g̑eg̑nōu: Skr. jajn̑ā́u, Lat. nōv‑i, √g̑enē‑; but W. adwaen, which is for *adwoen § 78 ii (1) (2), contains ‑u̯o- as pointed out by Rhys, RC. vi 22; it seems also to have the vowel of the reduplicator elided; thus adwaen < *ati-u̯o-kn‑a < *ati-u̯o‑g’gn‑a. It may however represent *ad-wo-ein < *ati-u̯o-gegn‑a. The 3rd sg. had *‑e for *‑a and gives the same result in W. The rest of the tense is formed from adwaen‑ as a stem on the analogy of gwẟost etc., or with pres. endings.

(2) The impf. ind. is a new formation from the same stem, except the 3rd sg., which may be old. The form atwaenat may however be for atweinat s.g. 36 which would represent regularly *ati-u̯o-gn‑i̯a-to < *‑g̑n‑i̯ə-tó 3rd sg. opt. mid.

(3) The rest of the verb comes from periphrastic tenses formed with the prefix *ati- only, and a verbal adj. *gnau̯os < *g̑nə-uo‑s (: cf. Lat. gnāvus < *g̑n̥̄-u̯o‑s), with the verb ‘to be’. This implies that ‑nab- is for ‑nawb- (cf. clybot § 194 v (4)); the ‑aw‑ is attested in O.W. amgnaubot ox., which must be the same formation with a different prefix. (This ‑au- cannot be from ‑ā‑, which would give ‑o- in the penult.)

§ 192. i. (1) pieu (Mn. W. pḯ-au) ‘whose is?’ contains the dative of the interrogative stem *q^u̯i- and ‑eu ‘is’, a weak form of *wy, which elsewhere became yw ‘is’ § 179 ix (3). The forms of the verb that occur in Ml. W. are as follows; most of them are re-formations from pieu, the ‑eu- generally unrounded to ‑ei- before v or ff:

Pres. ind.: sg. 2. piwyt see ii (1) below;—3. pieu;—pl. 3. piewynt (for *pieu-ynt) w.m. 83.

Impf. ind.: sg. 2. pieuoetud (t ≡ ẟ) see ii (3) below;—3rd sg. pioeẟ w.m. 117, pieuoeẟ r.m. 196, piewoeẟ w.m. 121, pieweẟ do. 129, pioweẟ do. 178, pieoeẟ do. 135;—3rd pl. pioeẟynt s.g. 426.

Fut.: sg. 3. pẏeuvyẟ (ẏ ≡ i) a.l. i 179 ms.b., pieivyẟ ib. ms.d., h.m. ii 81;—pl. 1 pieifyẟwn c.m. 42.

Perf.: 3rd sg. pieivu w.m. 394, r.m. 252, pievu w.m. 394.

Pres. subj.: 3rd sg. pẏeẏfo (ẏ ≡ i, f ≡ ff) a.l. i 196.

Impf. subj.: 3rd sg. pieiffei s.g. 299, pieivyẟei do. 324.

(2) In Mn. W., only the 3rd sg. is used. The forms are—

Indic. pres. pḯau;—imperf. pḯoedd L.G.C. 168, I.Ỻaf. c.c. 352, accented pïṓedd by T.A., c 84/849;—fut. pïéuvydd L.G.C. 291;—the other tenses rarely occur.

In the dialects the pres. pḯau only is used, and other tenses are formed periphrastically by using tenses of the verb ‘to be’ with relatival piau; thus oedd pia(u) ‘was who owns’ for pioedd ‘who owned’.

ii. (1) The verb ‘to be’ in pieu generally means ‘is’ in the sense of ‘belongs’; but sometimes it has a complement, in which case the literal meaning of the compound is seen clearly; thus—

Hi a ovynnawẟ iẟaw pioeẟ mab s.g. 12 ‘she asked him to whom he was son’ (whose son he was). Piwyt gwr di do. 222 ‘to whom art man thou?’ (whose man art thou?).

(2) The interrogative meaning of the compound survived in Ml. W. and Early Mn. verse; but the usual meaning is relative. Interrog. pieu in a question is often followed by rel. pieu in the answer; and this may represent the transition stage, as in the case of pan ‘whence?’ § 163 i (6).

Pieu yniver y llongeu hynn? … Arglwyẟ, heb wynt, mae ymma Matholwch … ac ef bieu y llongeu w.m. 39 ‘To whom belongs this fleet of ships? Lord, said they, M. is here, and [it is] he to whom the ships belong’.

Pïau rhent Gruffudd ap Rhys?

Hywel pḯau ’n nhâl Pówys.—T.A., j 17/217.

‘To whom belongs the rent of G. ap R.? [It is] Howel to whom it belongs on the border of Powys.’

When the relative became the prevalent construction, pwy ‘who?’ was used before the verb to ask a question, thus pwy bïau ‘who [is it] to whom belongs?’ This occurs in Ml. W.; as Pwy biewynt wy w.m. 83 ‘who [is it] to whom they belong?’ Cf. § 163 v.

Pwy bïau gwaed pibau gwin?—T.A., a 14998/29.

‘Who has the blood of pipes of wine?’

(3) Relatival pieu sometimes introduces a dependent relative clause, as Dodi olew ar y gwrda bieu y gaer r.m. 174 ‘administering extreme unction to the goodman who owns the castle’. But it is chiefly used to form the subject-clause after an emphatic predicative noun, § 162 vii (2), as in ef bieu y llongeu (2) above ‘[it is] he who owns the ships’; Meuryc bevɏr bieuoetud M.A. i 225b ‘[it was] bright Meuryc to whom thou [sword] didst belong’; a minneu bieu y ẟwy iarllaeth R.M. 239 ‘and [it is] I to whom the two earldoms belong’.

(4) As pi- is itself relative it is not preceded by the relative a, ZfCP. iv 118; see examples above. Cf. also mi bieivu r.m. 252, mi bḯau … a thithau bḯau I.G. 318, Dafydd bieuvydd L.G.C. 291, etc. The initial of pi- is generally softened, as in most of the above examples, but it frequently remains unchanged, as E koc a’r dẏsteẏn pẏeu a.l. i 20 ‘[it is] the cook and the steward to whom belong…’; e gur (≡ y gŵr) pyeu do. 82; Hywel piau (2) above; Mi piau cyngor … mi piau nerth Diar. viii 14 (1620). In the spoken lang. both p- and b- are heard; the former prevails in N.W.

(5) As pieu seemed to be a verb meaning ‘owns’ though without a subjective rel., it is sometimes found so used with an accusative rel., as castell Kaer Vyrẟin yr hwn a bie(u) y brenhin r.b.b. 297 ‘the castle of Carmarthen which the king owns’; y castell fry a pieu Belial b.cw. 10; more rarely with subjective rel., ni ae pieifyẟwn c.m. 42. Still rarer are re-formations like ti biy c.m. 14.

iii. pi- cannot come from *q^u̯ū(i) < *q^u̯ōi the dat. of *q^u̯o‑, since q^u̯ became k in Kelt, before u; it is probable therefore that pi- comes from *q^u̯ī < *q^u̯iī < *q^u̯ii̯ei: Oscan piei dative of the stem-form *q^u̯i- § 163 vi.

Af, Gwnaf, Deuaf.

§ 193. i. af ‘I go’ and gw̯naf ‘I make, do’ are conjugated alike in Mn. W. except in the impv.; deuaf ‘I come’ is analogous, but has different and varying vowels in its stems. In the earlier periods each of the verbs has forms peculiar to itself. In the following tables Mn. W. forms are given in brackets, marked as in § 185.

ii. af ‘I go’.

Indicative Mood. Present.
	sg.		pl.
1.	af (ā́f)	1.	awn (áwn)
2.	ey (éi, ái)	2.	ewch (éwch)
3.	a, e-yt (ā́)	3.	ant (ā́nt)
Impers. eir (éir, áir)

Imperfect.
1.	awn (áwn)	1.	aem (ā́em)
2.	aut (ā́ut)	2.	(ā́ech)
3.	aei, aey, ai (ā́i, ae)	3.	eynt (ā́ent)
Impers. eit (éid, áid)

Perfect.
1.	euthum (éuthum)	1.	aetham (áethom, ‑am)
2.	aethost (áethost)	2.	aethawch (áethoch)
3.	aeth (áeth)	3.	aethant, ‑ont (áethant, ‑ont)
Impers. aethpwyt (áethpwyd)

Second Perfect.
1.	athwyf, aẟwyf, ethwyf, eẟwyf (éthwyf)	1.	ethym
2.	athwyt, aẟwyt (éddwyd)	2.
3.	ethyw, eẟyw (éthyw, éddyw)	3.	ethynt, eẟynt

Pluperfect.
1.	athoeẟwn (áethwn)	1.	(áethem)
2.	(áethud, ‑it)	2.	(áethech)
3.	athoeẟ, aẟoeẟ (áethai)	3.	athoeẟynt (áethynt, ‑ent)

Subjunctive Mood. Present.
1.	el(h)wyf (élwyf)	1.	el(h)om (élom)
2.	el(h)ych (élych)	2.	el(h)och (éloch)
3.	el (ḗl, élo) aho	3.	el(h)ont, el(h)wynt (élont) ahont
Impers. el(h)er (éler)

Imperfect.
1.	el(h)wn (élwn)	1.	(élem)
2.	el(h)ut (élud, ‑it)	2.	(élech)
3.	el(h)ei (élai)	3.	el(h)ynt (élynt, ‑ent)
Impers. (élid)

Imperative Mood. Present.
		1.	awn (áwn)
2.	dos (dṓs)	2.	ewch (éwch)
3.	aet, elhid (ā́ed, eled)	3.	aent (ā́ent, ā́nt)

Verbal Noun.
mynet (mỿ́ned, mɥnd) ‘to go’

iii. gwnaf ‘I make, do’. Indicative Mood. Present. sg. 1. gwnaf (gw̯nā́f), etc. like af (ā́f); exc. strong 3rd sg. gwneyẟ. Imperfect. sg. 1. gwnawn (gw̯náwn), etc. like awn (áwn); pl. 2. gwnaewch (gw̯nā́ech). Perfect.
A.	sg. 1. gwneuthum (gw̯néuthum), etc. like euthum (éuthum).
B.	sg.		pl.
1.	gorugum	1.	gorugam
2.	gorugost	2.	gorugawch
3.	goruc, goreu	3.	gorugant
Impers. gorucpwyt
Second Perfect. sg. 1. (gw̯néddwyf),   2. (gw̯néddwyt),   3. gwneẟyw (gw̯néddyw) Pluperfect.
	sg.		pl.
1.	gwnathoeẟwn (gw̯náethwn)	1.	(gw̯náethem)
2.	gwnathoeẟut (gw̯náethud, ‑it)	2.	(gw̯náethech)
3.	gwnaethoeẟ, gwynathoẟ, gwnaẟoeẟ (gw̯náethai)	3.	gwnathoeẟynt (gw̯náethynt, ‑ent)
Impers. gwnathoeẟit (gw̯náethid)
Subjunctive Mood. Present. sg. 1. gwnel(h)wyf (gw̯nélwyf) etc. like el(h)wyf (élwyf) throughout; also sg. 3. gunech, gwnech. Imperfect. sg. 1. gwnel(h)wn (gw̯nélwn), etc. like el(h)wn (élwn).

Imperative Mood. Present.
	sg.		pl.
		1 .	gwnawn (gw̯náwn)
2.	gwna (gw̯nā́)	2.	gwnewch (gw̯néwch)
3.	gwnaet (gw̯nā́ed)	3.	gwnaent (gw̯nā́ent, ‑ā́nt)
Impers. gwnel(h)er (gw̯néler)
Verbal Noun. gwneithur, gwneuthur (gwneuthur)
Verbal Adjectives. gwneithurẏedic (gw̯neuthurédig, gw̯neuthurádwy)

iv. deuaf ‘I come’.

Indicative Mood. Present or First Future.
	sg.		pl.
1.	deuaf, doaf (déuaf, dṓf)	1.	deuwn, down (déuwn, dówn)
2.	deuy, dewy, doy (déui, dói)	2.	deuwch, dowch (déuwch, dówch)
3.	daw (daw § 52 iii (1)), dyẟaw, do, dyẟo	3.	deuant, doant (déuant, dṓnt), dyẟeuant
Impers. dyẟeuhawr (déuir, dóir)
Second Future. sg. 1. dybyẟaf; 3. dyvyẟ, dybyẟ, dybyẟhawt, dyvi, dybi, dypi, deubyẟ, deubi, deupi; pl. 3. dybyẟant.
Imperfect.
	sg.		pl.
1.	deuwn, down (déuwn, dówn)	1.	(déuem, dṓem)
2.	deuut, dout (déuut, dṓut, ‑it)	2.	(déuech, dṓech)
3.	deuei, doei, doey, doi (déuai, dṓi)	3.	deuynt, doynt (déuynt, dṓent, déuent)
Impers. deuit (déuid, dóid)

Perfect.
A.	sg.		pl.
1.	deuthum, doethum (déuthum)	1.	doetham (déuthom)
2.	deuthost, doethost (déuthost)	2.	doethawch, ‑och (deuthoch)
3.	deuth, doeth (dā́eth, dṓeth)	3.	deuthant, doethant, doethont (déuthant, ‑ont)
Impers. deuthpwyt, doethpwyt (déuthpwyd)
B.
2.	dyvuost	2.
3.	dyvu, dybu, deubu	3.	dyvuant, dybuant
Second Perfect.
1.	dothwyf, doẟwyf	1.	dofym
2.	dothwyt, doẟwyt	2.	doẟywch, doethywch
3.	doethyw, dothyw, doẟyw, deẟyw (doddyw, deddyw)	3.	doẟynt
Pluperfect.
1.	dathoeẟwn (déuthwn)	1.	(déuthem)
2.	(déuthud, ‑it)	2.	(déuthech)
3.	doethoeẟ, dothoeẟ, dathoeẟ (déuthai)	3.	doethoeẟynt, dothoeẟynt (deuthynt, ‑ent)

Subjunctive Mood.

Present.

sg. 1. del(h)wyf (délwyf), etc. like the el- forms of el(h)wyf (élwyf) throughout; also sg. 1. dybwyf; 3. dyvo, dyffo, dyppo, deupo, dyẟeuho; pl. 3. dyffont, deuhont.

Imperfect.

sg. 1. del(h)wn (délwn), etc. like el(h)wn (elwn); also sg. 3. dybei, dyfei dyffei.

Imperative Mood.

Present.

sg. 2. dyret, dabre (dỿ́fydd, dýred, dɥ́rd, tỿ́red, tɥ́rd, dábre, dỿ́re, dial. dére); 3. deuet, doet (déued, dṓed, déled); pl. 1. down (déuwn, dówn); 2. dowch, dewch (déuwch, dówch, déwch); 3. deuent, doent (déuent, dṓent).

Verbal Noun.
dyvot (dyfod, dywod, dywad).

v. Pres. and Impf. Ind.—(1) The contracted forms ei, eir, eid, gwnei, gwneir, gwneid are now written and pron. with ei (≡ əi); but formerly ai was used as in uncontracted forms; § 81 iii (1); as Bwrdeisiaid a w̯naid yn waeth G.Gl. p 100/174. The 3rd sg. impf. aei, gwnaei are already contracted in Ml. W., as ai w.m. 117, 252, 451, gwnai 54, 250, 389, gunai b.b. 56; similarly doei had become doi w.m. 7. See § 52 iii (3).

(2) For a Dr. M. used dial. aiff (now eiff) § 179 iii (1); this is condemned by D. 86. Some late writers have used gw̯naiff also; but the lit. gwna prevails. The old strong form of af is ë‑yt § 173 vi (1); of gwnaf is gwnë-yẟ do. (3).

(3) The stems deu‑, do- are both used throughout the pres. and impf. except in the 3rd sg. pres.; thus doaf, doy r.m. 76, w.m. 55, deuaf, deuy s.g. 15. In Mn. W. doaf is contracted to dṓf D.G. 355, L.G.C. 206, 468; this is the usual spoken form, though doa(f) persists in Dyf. dial. The 3rd sg. is daw; also do b.t. 38, dyẟau b.b. 32 (‑u ≡ w), dyẟaw r.p. 1055, l. 16, dyẟo ib. l. 23.—O.W. gurthdo gl. obstitit.

(4) The second future of deuaf is a survival, chiefly used in poetry: sg. 1. dybyẟaf b.t. 19; sg. 3. diwit (≡ dyvyẟ) b.b. 51, dybit (≡ dybyẟ) do. 55, dyfyẟ b.t. 10, dybyẟ r.p. 1190, dyẟybyẟ b.t. 42, dybyẟhawt r.p. 1437, dyvi b.t. 72, dybi b.b. 60, dypi w.m. 478, deubyẟ b.t. 17, deubi b.t. 3, deupi b.b. 61; pl. 3. dybyẟant b.t. 26.

vi. Perf. and Plup.—(1) In late Mn. W. euthum, gwneuthum, deuthum, are often misspelt aethum, gwnaethum, daethum. In the dialects the 1st and 2nd sg. perf. are mostly replaced by new aorists ēs, gwnēs, dóis on the analogy of cēs and rhois, also eis and gwneis (“balbutientium puerorum mera sunt barbaries” D. 117).

(2) In Ml. W. the perf. stem of deuaf is deuth- or doeth‑; and the 3rd sg. is deuth or doeth. Ml. W. daeth is doubtful; y | daeth b.b. 3 is prob. yd aeth, cf. 97 marg. In the Early Mn. bards the form attested by the rhyme is doeth D.G. 259 (misprinted daeth), 287, as there is no rhyme to dauth the regular Mn. equivalent of Ml. deuth. Late Mn. W. daeth may be dauth h.g. 21 misspelt, as daethant is a misspelling of deuthant. The N.W. dial, form is dṓth, 3rd pl. deuthon’ or doethon’. In S.W. dā́th is also heard.—Impers. § 175 iv (7).

Dan i ddant erioed ni ddoeth

Ar i enau air annoeth.—D.N., m 136/123.

‘Under his tooth there never came on his lips an unwise word.’

(3) The second perf. of af and deuaf is of frequent occurrence in Ml. W. poetry, as athwyf, ethyw H.O.G. m.a. i 275, athwyd, ethynt P.M. do. 289, aẟwyf C. do. 216, etiw (t ≡ ẟ) do. do. 220; dothuif b.b. 79, dotyw (t ≡ ẟ) M. w. 1a, dotynt (≡ doẟynt) do. do. 3a, ethint b.b. 33. It is also met with fairly often in Ml. prose: eẟyw w.m. 456, ethyw r.m. 104, dothwyf w.m. 459, doẟwyf do. 20, doẟyw do. 457, doẟywch, doẟym do. 475, ethynt r.b.b. 205, but tends in later mss. to be replaced by the first perf.; thus dothwyf w.m. 459 appears as deuthum in r.m. 105; doẟyw w.m. 473 as doeth in r.m. 105. D.G. and his contemporaries continued its use in poetry; afterwards it became obsolete: deddyw D.G. 4, ethyw (misspelt eithiw, euthyw) I.G. 312;

Lliw dydd a ddaw^[1] lle doddyw;

Llewych haul ar y lluwch yw. D.G. 321.

‘Daylight comes where she has come; she is sunshine on the snowdrift.’ It was at this period, when the form was already an artificial survival, that it first appears for gwnaf: gwneddwyf D.G. 115, gwneddwyd do. 102, gwneddyw do. 429, gwneẟyw I.C. r.p. 1286. These imitations were shortlived.

(4) Both the first perf. in ‑th‑um and the second perf. in ‑wyf are probably original for af only. The older perfects of the other verbs are:

gwnaf: sg. 1. gorugum w.m. 226–9; sg. 2. gorugost r.m. 192; pl. 1. gorugam, 3. gorugant w.m. 227, 226; sg. 3. goruc of extremely frequent occurrence, goreu surviving in poetry, b.b. 43, M. w. 2a, E.S. m.a. i 349a, guoreu b.a. 35, 38; impers. gorucpwyt w.m. 452 (= gwnaethpwyt r.m. 100), w.m. 454, r.m. 101.

deuaf: sg. 2. dyvuost w.m. 458 (= doethost r.m. 104); sg. 3. dyvu w.m. 457 (= doeth r.m. 104), dybu M. w. 1b, 2a; pl. 3. dybuant b.t. 6, r.p. 1405, G.B. do. 1192.

(5) In Ml. W. the plup. of all three verbs was formed by means of ‑oeẟwn; as doethoeẟ Ỻ.A. 17 ‘had come’, athoeẟ w.m. 13, aẟoeẟ do. 15 ‘had gone’, gwnaethoeẟ do. 30, gvnathoeẟ do. 440, gwnathoeẟwn s.g. 198, gwnathoeẟut do. 274; dothoeẟ r.m. 200, dathoeẟ do. 197. These forms are rare in Mn. W.: rhy-wnaethoeẟ D.G. 509. The Mn. plup. is a new formation made, as in regular verbs, by adding impf. endings to the perf. stem: gwnaethwn Ezec. xxxi 9, daethwn Matt. xxv 27, aethai Luc viii 2, etc. D. also gives elswn etc.; this formation is used for gwnaf in the Bible: gwnelswn 1 Chron. xxiii 5, gwnelsei 2 Chron. xxi 6.

vii. Subjunct.—(1) The subjunct. stems are el‑, gw̯nel- and del‑; as elwyf w.m. 457, delwyf r.m. 131, elych, delych do. 237, gwnelych w.m. 456, delhich b.b. 84, gwneloch w.m. 475, elont r.m. 34, elwynt b.a. 2; elhut b.b. 56, delhei do. 96; elher do. 33.

The peculiarity of the pres. subj. with these stems is that the 3rd sg. lacks the usual ending ‑o (or ‑wy); thus a phan el ef … yny el ef w.m. 22 ‘and when he goes … until he goes’, val nat el neb do. 49 ‘so that no one may go’, Y kyn a el, hwnnw a orẟir r.b. 1063 ‘the chisel that will go, that [is the one] that is hammered’, Guledic … a’n gunel in r̔it (i ≡ y, t ≡ ẟ) b.b. 40 ‘may the Lord make us free’, y dit y del paup do. 41 ‘the day when each will come’. So in Mn. W.; thus, expressing a wish: Dêl i’th fryd dalu i’th frawd D.G. 34 ‘may it come to thy mind to repay thy brother’, cf. 341; I henaint yr êl honno L.G.C. 10 ‘may she go [live] to old age’, cf. 476; Dêl amorth yn dâl imi Gr.O. 59 ‘may misfortune come as retribution to me’; in a dependent clause:

Pan ddêl y Pasg a’r glasgoed,

Bun a ddaw beunydd i oed.—D.G. 199.

‘When Easter comes, and the green trees, [my] lady will come daily to the tryst.’ Sometimes in Late W. the ending is added; as gwnelo § 162 i, doed a ddelo beside doed a ddêl ‘come what may come’.

(2) Other forms of the subjunctive occur as follows in Ml. W.:

af: pres. sg. 3. aho r.m. 140; pl. 3. ahont b.t. 17.

gwnaf: pres. sg. 3. gunaho b.b. 70, gwnaho b.t. 10, ll. 13, 27, gunech, gwnech § 183 iii (1); pl. 3. gvvnahont b.b. 61, gwnahon b.t. 34.

deuaf: pres. sg. 1. dybwyf r.p. 1183; sg. 3. dybo ib., dyvo do. 584, dyffo b.t. 10, dyppo b.b. 90, deupo b.a. 6, dyẟeuho, deẟeuho b.t. 29; pl. 3. dyffont m.a. i 136, diffont b.b. 59, 60, deuhont b.t. 3; imperfect sg. 3. dyfei b.t. 3, dyffei do. 13, b.a. 2, dybei b.t. 6.

viii. Impv.—(1) dos ‘go!’ e.g. dos ẏ’r llys w.m. 14 ‘go to the court’. This is the usual meaning; but the original meaning was doubtless, like that of the Corn. and Bret. forms, ‘come’. This is preserved in some parts of Powys to this day; and is sometimes met with in Ml. W.; e.g. dos yma r.m. 176, s.g. 221 ‘come here’.

(2) Ml. W. dyret w.m. 21, r.m. 173, Ỻ.A. 99, etc.; dabre b.b. 102, w.m. 17, r.b.b. 125, etc.—Mn. W. dyfydd D.G. 41, dyred do. 107, dabre (misprinted debre) D.G. 31, 134, 515, tyred, dyre I.G. 215, Gwna ddydd a dyrd, Gwenddydd dec W.Ỻ. 83 ‘make an appointment and come, fair Gwenddydd’, Tyrd i’r bwlch, taro di’r bêl I.T. Ỻ 133/213 ‘come to the breach, strike thou the ball’, § 44 vi, Dere â’r cafodydd hyfryd Wms. 273 ‘come with [i.e. bring] the gladsome showers’.

(3) Sg. 3.: aet w.m. 13, 35, elhid b.b. 101, gwnaet r.m. 261, gvnaed w.m. 406, deuet w.m. 186, deuhet r.m. 88, doet w.m. 122.

(4) Pl. 2.: dowch w.b. vi r., w.m. 407, 447, r.m. 261, 292, dewch Ỻ.A. 126.

ix. Verbal noun.—(1) On myned, mynd, see § 44 vi.

(2) The Ml. and Mn. v.n. of gwnaf is gwneuthur. D. 121 also gives gwneuthud, but this is rarely met with. It is printed in D.G. 107, but is not attested by the cynghanedd. In the dialects a new form gwneud arose; this is in common use in the late period; the earliest example I have noted is in rh.b.s. 1. (In D.G. 409 gwneud makes a short line, and should be gwneuthur; for it wr wneyd marnad arall c. i 200 read vwrw’n y dŵr farwnad arall p 77/158; so wherever gwneud is attributed to an old author.) V.a. gwneithurẏedic g.c. 114.

(3) The only v.n. of deuaf is dyfod; but the f became w § 26 v, and wo interchanges with wa § 34 iv, hence dywot Ỻ.A. 80, dywod T.A. a 14976/101, dywad D.G. 306, spelt dowad c.c. 369 (see § 33 iii), beside the original dyfod. The form dỿwad became dw̄́ad in the dialects, and this is the spoken form both in N., and S.W. But in part of Dyfed a form dṓd developed (apparently from *dowod < dỿwod); this was used by Wms., and has since been in common use, chiefly in verse in free metres.

The noun dovot W.M. 33 ‘a find’ is a different word, being for do-ovot a.l. i. 94 (also dohovet [read ‑ot] ib.) < *dỿ-w̯o-vot.

(4) All the forms given in dictionaries, containing the tense stems of these verbs, such as äu, athu, elu, eddu ‘to go’, dawed, dawad, delyd, doddi ‘to come’, gwnelyd ‘to do’, are spurious. Silvan Evans misquotes D.G. 306 dywad as an example of dawad, s.v.; but admits that the others do “not occur in the infinitive”! see s.v. delyd.

x. Origin of the forms. (1) af < *aᵹaf: Ir. agaim ‘I drive’ √ag̑‑: Lat. ago, Gk. ἄγω, Skr. ájati ‘drives’. The verb had middle flexion in Brit., cf. ë-yt ‘goes’ < *ag̑-e-tai (' drives himself, goes’) § 179 iii (1). Hence the perf. euthum < *aktos esmi § 182 iv (1), and the plup. athoeẟ ib. (2). For the voicing of th to ẟ in eẟyw, aẟoeẟ see § 108 iv (2). Stokes’s reference of eẟwyd ‘ivisti’ to √ped- Fick⁴ ii 28 (still quoted, e.g. by Walde² s.v. pēs) is made in ignorance of the facts.—On dos see (7); on mynet § 100 iv.

(2) The subj. stem el- comes from the synonymous root *elā‑: Gk. ἐλάω ‘I drive’; in the pres. ind. the stem was *ell‑, prob. for *el‑n‑, Thurneysen Gr. 314, as in Ir. ad-ella ‘transit’, di-ella ‘deviat’; in W. *ell-af was driven out by af, but the subj. elwyf remained. W. delwyf is probably, like gwnelwyf an analogical formation. The reason why the 3rd sg. has no ‑o may be that these forms superseded an old 3rd sg. middle *elhyt and 3rd sg. gwnech which had no ‑o. The view that gwnêl is a re-formation is borne out by the actual survival of gwnēch.

(3) The stem of gw̯naf is *u̯rag‑, √u̯ereꬶ- ‘work’ § 100 i (2). In the pres. and impf. ind., therefore, the flexion was exactly the same as for af, stem *ag‑; this led to its being assimilated to af in other tenses. The old root-aor. sg. 1. gwrith, 3. gwreith became gwneuthum, gwnaeth like the perf. of af, § 181 vii (2).—The old perf. of √u̯ereꬶ- is preserved in the 3rd sg. in Ml. W. guoreu, goreu § 182 ii (1), Ml. Bret, guereu, gueure, guerue.—It does not seem possible to derive goruc from the same root; this occurs as sg. 1. 3. in Corn. gwrûk (grûg etc.); it probably represents a synonymous form associated with goreu on account of accidental similarity; possibly < *u̯er-oik‑, √peik̑-: Skr. pįs̑áti ‘carves, adorns, forms, prepares ', pés̑aḥ ‘form’ (: Lat. pingo, with ‑k̑/g̑- altern.); cf. Duu an goruc b.b. 39 ‘God made us’.

(4) The v.n. gwneuthur is for gwneithur g.c. 112, 128, w.m. pp. 93, 94 (p 16), b.ch. 62 (cf. anghyfreith wneuthur r.p. 1296, i.e. wneithur) § 77 viii. The original v.n. was *gwreith < *u̯rek-tu‑; by the loss of ‑r- after the initial this became gweith, gwaith ‘work’. The form *gwreith occurs, written guereit, in enuir ith elwir od guur guereit b.a. 37, which appears elsewhere as enwir yt elwir oth gywir weithret do. 34, l. 4, though the rhyming word is kyvẏeith; but weithret is also a genuine variant rh. with kiwet ib. l. 9. Possibly the ‑r- was first lost in the compound *gwreithret by dissim. The ‑ur added to *gwreith ‘work’ may have come from the synonymous llafur < Lat. labōrem. The form *gwreithur might easily have become gwneithur by dissim. § 102 iii (2), as it was dissimilated to gwruthyl in Corn. The ‑n- might spread from this to the verb; but as gw̯n- is slightly easier than gw̯r- the change may have taken place in the vb. itself owing to its frequent occurrence. The old v.n. gweith with lost ‑r- came to be dissociated from the vb., and gwneuthur remained the only v.n. Ultimately from gwaith ‘work’ a new denom. gweithiaf ‘I work’ was formed, with gweithio ‘to work’ as v.n.—gweith ‘battle’ < *u̯iktā (: Ir. fichim ‘I fight’, Lat. vinco) is a different word.

(5) deuaf is a compound of the verb ‘to be’, as seen in the v.n. dy-fod. The prefix is *do- which appears regularly as dy- before a cons.—The pres. is future in meaning, and comes from the fut. *esō; thus *dó esō > *deu, which was made into deu-af § 75 ii (2), so the 2nd sg.; the 3rd sg. *do eset gave daw or do see ib. The pres. deuaf would be in O.W. *doüam; under the influence of 3rd sg. do this became *do-am > Ml. W. doaf; thus deu- and do- became the stems of the pres. and impf.; and deu- was even substituted for dỿ- in some other tenses as deu-bi for dy-bi, v (4). [Later the 3rd sg. daw was made a stem in S.W. dialects, and dawaf, dawai, etc. occur in late mss.]

(6) Other tenses contain the b- forms of the vb. ‘to be’; the fut. dyvyẟ, dyvi, pres. subj. dyvo, dyffo are regular; the perf. might be either dyvu < *do-(be-)baue or dybu < *do‑b’baue; from the latter the ‑b- spread to other tenses. The perf. dyvu or dybu was supplanted, see vi (4), by a new perf. formed in imitation of aeth but with the vowels of the pres. stems deu‑, do‑; thus deuth, doeth; and by a new second perf. similarly modelled on eẟyw, which like eẟyw itself became obsolete in Ml. W.

(7) The impv. of deuaf was dos, which was transferred to af, see viii (1). The Corn. forms are dus, dues, des, the Bret. is deuz. It is clearly impossible to equate these forms either with one another or with dos. What has taken place is that the vowel of other forms, especially the 2nd pl., has been substituted for the original vowel; thus W. dos after do-wch, Corn. dues after duech, des after de-uch, Bret. deuz after deu-it ‘come ye’; a late example is W. dial. (to a child) dows yma ‘come here’ after dowch. This leaves Corn. dus as the unaltered form; dus < *doistǖd < *do estōd: Lat. estōd, estō, Gk. ἔστω.

(8) The loss of dos to deuaf was supplied by the impv. of verbs meaning ‘come’ from √reg̑‑: Ir. do-rega ‘he will come’; thus dabre < *dabbirigā < *do-ambi-reg-ā; dy-re < *do-rigā < *do-reg-ā. The forms with ‑d are generally referred to √ret- ‘run’; but it would be more satisfactory if they could be connected with the above. Ir. tair ‘come’ < *to-reg shows *reg- athematic; to athematic stems a 2nd sg. impv. *‑dhi might be added (: Gk ‑θι); thus *do-reg-di > *do-red-di > dyred. It is true that ‑dhi was added to R-grade of root; but there are exceptions, as in the case of ‑tōd (Lat. estō for *s-tōd).

dyre also occurs as 3rd sg. pres. ind., r.p. 1036, l. 28.

Verbs with old Perfects.

§ 194. i. (1) dywedaf ‘I say’ has 3rd sg. pres. ind. Ml. W. dyweit Ỻ.A. 21, Early Mn. W. dywaid. In Late Mn. W. this form is replaced by dywed, which is not so much a re-formate from the other persons as a dial. pron. of dywaid, § 6 iii. (In Gwyn. the dial, form is dyfyd re-formed with the regular affection as in gweryd: gwaredaf.)

The 3rd sg. dyweit seems to contain the affected form of the R-grade *u̯at- (*u̯ₑt‑) of the root § 201 i (3); cf. beirv b.b. 101: berwaf.

(2) The aor. is dywedeis w.m. 10, dywedeist do. 63, dywedassam, etc., which is regular, except that for the 3rd sg. the perf. is used: Ml. W. dywawt r.m. 5, 6, dywat do. 23, dywot w.m. 6, 7; Early Mn. W. dywawd R.G.E. d. 141, dywad, dywod, dyfod. For these in Late Mn. W. a new formation dywedodd is used; but in Gwyn. dial. dỿwad, dw̄́ad may still be heard (Rhys, RC. vi 17).

Ni ddyfod ond yn ddifalch;

Ni bu na gorwag na balch.—D.N., m 136/123.

‘He spoke only modestly: he was neither vain nor proud.’

The impers. is the perf. dywespwyt r.m. 90, r.b.b. 10, dywetpwyt s.g. 17, Mn. W. dywetpwyd Matt, i 22 (1620). But the aor. dywedwyd is more usual in Mn. W., and also occurs in Ml. W.: dywedwyt Ỻ.A. 115.

(3) The 2nd sg. impv. is, of course, dywet w.m. 121; Mn. W. dywed. But in Early Mn. verse we sometimes find dywaid, D.G. 355, G.Gr. do. 247, owing to the influence of the irregular 3rd sg. pres. ind.

(4) The v.n. is Early Ml. W. dywedwyd > Ml. W. dywedut § 78 iv (2), written in Mn. W. dywedyd.

In the dialects S.W. gwĕ́ud (the vb. also gwedaf), N.W. (dwĕ́ud), dĕ́ud, dw̆́yd, (ĕ ≡ ə).

(5) dywedaf: √u̯et/d- ‘say’: Skr. vádati ‘speaks’, W. gwawd ‘song’, Ir. fāith ‘poet’, Gaul. (-Gk.) pl. οὐᾱ́τεις (whence Lat. vātēs, Walde, s.v.) < Kelt. *u̯āt- < L° *u̯ōt‑. Perf. dywawt, etc., § 182 ii (1).—V.n. dywedwyd § 203 iii (4), (8).

(6) The verb, with the root-form *u̯at‑, see (1), was used without the prefix dy- before na ‘that not’, thus gwadaf na ‘I say that not, I deny that’. Hence gwadaf came to mean ‘I deny’, v.n. gwadu, though an objective clause after it is still introduced by na. With neg. di- in Ml. W. diwat w.m.l. 92 ‘denies’.

Oes a wad o sywedydd,

Lle dêl, nad hyfryd lliw dydd?—Gr.O. 38.

‘Is there an astronomer who will deny that the light of day, where it comes, is pleasant?’

ii. (1) gwaredaf ‘I succour, relieve’: 3rd sg. pres. ind. O. W. guorit juv. sk., Ml. W. gweryt r.p. 1171, l. 5, Mn. W. gweryd Diar. xiv 25, Gr.O. 113;—v.n. Ml. W. guaret w.m. 3, Mn. W. gwared, gwaredu. The verb is quite regular. But in O. and Early Ml. W. the 3rd sg. past is the perf. guoraut juv. sk., guaraud b.b. 39, gwarawt r.p. 1159.

There is also a 3rd sg. pres. subj. gwares seen in gwares Duw dy anghen r.p. 577 ‘may God relieve thy want’, § 183 iii (1).

(2) gwared < *u̯o-ret- < *upo- ‘under’ + *ret- ‘run’: cf. Lat. suc-curro < sub ‘under’ + curro ‘I run’;—gwarawt § 182 ii (1).

iii. (1) dygaf ‘I bring’: 3rd sg. pres. ind. Ml. W. dwc w.m. 398, Mn. W. dwg (≡ dw̄g);—v.n. Ml. and Mn. W. dwyn (≡ dŵɥn). Old 3rd sg. pres. subj. duch § 183 iii (1).

(2) Perf. sg. i. dugum w.m. 42; 2. dugost s.g. 246; 3. duc w.m. 42; pl. 3. dugant c.m. 107, s.g. 246, re-formed as ducsant c.m. 59, dugassant s.g. 16. In Mn. W. the 3rd sg. dug (‑ū‑) remained the standard form, though a new dygodd has tended to replace it in the recent period. But the other persons were re-formed as aorists in the 16th cent., though the older forms continued in use:

Dy wg yn hir y dugum;

O dygais, di-fantais fûm.—W.Ỻ.

‘Thy resentment have I long borne; if I have borne it, I have been no gainer.’

(3) The compound ymddygaf is similarly inflected: v.n., Mn. W. ýmddwyn ‘to behave’, ymddw̄́yn ‘to bear’ § 41 i; perf. sg. 3. ymddug Can. iii 4, in late bibles ymddûg (and so pronounced).

(4) dygaf, dug § 182 ii (2); dw͡yn § 203 iv (3).

iv. (1) Ml. W. amygaf ‘I defend’: 3rd sg. pres. ind. amwc b.t. 29; v.n. amwyn.

am-w͡yn seems to mean literally ‘fight for’, since it is followed by â ‘with’; as amwyn y gorflwch hwn a mi w.m. 122 ‘to fight for this goblet with me’; amvin ae elin terwin guinet b.b. 57 ‘to fight with his enemy for the border of Gwynedd’.

(2) Perf. sg. 3. amuc b.b. 39, b.a. 12, neu‑s amuc ae wayw b.a. 11 ‘defended him with his spear’. There is also a form amwyth used intransitively, and therefore prob. a middle form like aeth; as pan amwyth ae alon yn Llech Wen b.t. 57 ‘when he contended with his foes at LI. W.’—Plup. sg. 3. amucsei r.p. 1044.

(3) am-wg < *m̥bi-(p)uk‑, √peuk̑‑: Lat. pugna, pugil, Gk. πύκτης, πυγμάχος, O.E. feohtan, E. fight.—The perf. amuc with ‑uc < *‑pōuke, like duc § 181 ii (2). The form amwyth prob. represents *amb(i)uktos ’st; as it has the R-grade of the root, it cannot be a root-aorist. The v.n. has ‑no- suffix § 203 iv (3).—See also § 54 i (1).

The perf. has not been preserved in gorchfygaf ‘I conquer’, Ml. W. gorchyfygaf § 44 ii < *uper-kom-puk̑‑.

v. (1) clywaf ‘I hear’: 3rd sg. pres. ind. clyw w.m. 54; v.n. Ml. W. clybot w.m. 474, clywet G.Y.C. (anno 1282) r.p. 1417, Mn. W. clywed.

(2) Perf. sg. 1. cigleu w.m. 36, 83 = r.m. 23, 60, r.m. 129, b.t. 33; ciglef r.m. 130, 168, w.m. 408, 423 = r.m. 262, 274, c.m. 46, 48; sg. 3. cigleu w.m. 144 = r.m. 214, c.m. 50, s.g. 10, 11, etc. The rest of the tense is made up of aor. forms: sg. 2. clyweist w.m. 230, r.m. 168; pl. 3. clywssont w.m. 33, r.m. 22; impers. clywysbwyt Ỻ.A. 117, clywspwyt s.g. 246.

In Early Mn. W. the 1st sg. ciglef survived in poetry, see ex., and I.G. 338. But the ordinary Mn. form is clywais D.G. 81. Similarly the 3rd sg. cigleu is replaced by clywodd Luc xiv 15; thus the tense became a regular aor. There is also a Late Ml. and Mn. 3rd sg. clỿbu s.g. 362, Ex. ii 15, and impers. clybū́w͡yd Matt. ii 18 beside clyw͡yd Ps. Ixxvii 18.

Doe ym mherigl y ciglef

Ynglyn aur angel o nef.—D.G. 124.

‘Yesterday in danger I heard the golden englyn of an angel from heaven.’

(3) In Early Mn. W. a 2nd sg. impv. degle is found, e.g. G.Gl. i. mss. 315; both form and meaning seem to have been influenced by dyre (dial. dere) ‘come!’

Degle’n nes, dwg i liw nyf

Ddeg annerch oddi gennyf.—D.G. 218.

‘Lend nearer ear! bring to [her of] the colour of snow ten greetings from me.’

(4) clywaf, see § 76 v (2).—cigleu § 182 i; the form ciglef is the result of adding 1st sg. ‑f to cigleu (euf > ef); it tends in late mss. to replace the latter; thus cigleu w.m. 144 = ciglef r.m. 214. The cynghanedd in the example shows that the vowel of the reduplicator is i (as it is generally written), and not y; hence we must assume original *k̑ū‑. The 1st sg. was most used, and prob. gives the form cigleu.—clybot is probably for *clyw-bot, cf. adnabot § 191 iv (3).

vi. goẟiweẟaf' ‘I overtake’: v.n. goẟiwes § 203 iii (7) so in Mn. W., sometimes re-formed in Late W. as goddiweddyd.—Perf. sg. 3. goẟiwawẟ, see § 182 iii.

Verbs with t-Aorists.

§ 195. i. (1) canaf ‘I sing’: 3rd sg. pres. ind. can b.b. 13 ≡ Mn. W. cân; v.n, canu.—Aor. sg. 1. keint, keintum, 2. ceuntost, 3. cant § 175 iii, § 181 vii (1), impers. canpwyt § 182 iv (4); there are no corresponding forms in the pl. The t‑aor. was already superseded in Late Ml. W.; thus sg. 3. canawẟ Ỻ.A. 117, Mn. W. canodd; but cant survived in the phrase X. a’i cant ‘[it was] X. who sang it’, ascribing a poem to its author, and is often miswritten cânt by late copyists § 175 iii (1).

(2) gwanaf ‘I wound’ is similar. Aor. sg. 1. gweint, 3. gwant § 175 iii; Mn. W. gwenais, gwanodd.

ii. (1) cymeraf ‘I take’, differaf ‘I protect’: 3rd sg. pres. ind. cymer, differ; v.n. kymryt w.m. 8, 9, diffryt r.m. 132, 141.—Aor. sg. 3. kymerth, differth, kemirth (≡ kỿmɥrth) a.l. i 126, diffyrth r.m. 139, § 175 iii (1). Beside these, forms in ‑w͡ys, ‑ws occur in Ml. W., as kemerrws § 175 i (5), differwys G.B. r.p. 1191. But cymerth survives in biblical W., e. g. Act. xvi 33, beside the usual Late Mn. W. cymerodd c.c. 318, Matt, xiii 31.

(2) The v.n. cymryt, Mn. W. cymryd c.c. 335, cam-gymryd M.K. [137], has been re-formed as cymeryd; but the prevailing form in the spoken lang. is cỿ́mrɥd Ceiriog o.h. 110 (or cỿ́m’ɥd). The translators of the bible adopted cymmeryd, evidently thinking that it was more correct than the traditional form.—On the other hand, the verb is sometimes found re-formed after the v.n.; thus kymreist r.g. 1128, cymrodd D.G. 356, cymrais E.P. ps. cxix 111.

(3) cymeraf < *kom-bher- § 90;—differaf < *dē-ek̑s-per‑, √per- ‘bring’: Skr. pí-par-ti ‘brings across, delivers, protects’;—cymryt < *kom-bhr̥-tu- § 203 iii (8).—cymerth, cymyrth § 181 vii (1).

iii. Early Ml. W. dyrreith ‘came, returned’; maeth ‘nursed’; gwreith ‘did’; § 181 vii (2).

Defective Verbs.

§ 196. The following verbs are used in the 3rd sg. only.

i. (1) Ml. W. dawr, tawr ‘matters’, impf. dorei, torei, fut. dorbi; also with di‑: diẟawr, diẟorei, v.n. diẟarfot. (The ‑ẟ- is inferred from Early Mn. cynghanedd, as deuddyn / diddawr D.G. 37.) The verb is chiefly used with a negative particle and dative infixed pron.; thus ny’m dawr r.p. 1240 ‘I do not care’, literally ‘it matters not to me’. It is generally stated to be impersonal; but this is an error, for the subject—that which ‘matters’—is often expressed, and when not expressed is understood, like the implied subject of any other verb. Thus, Ny’m tawr i vynet w.m. 437 ‘I do not mind going’; i is the affixed pron. supplementing ’m, and the subject of tawr is vynet, thus ‘going matters not to me’; so, Ny’m dorei syrthyaw … nef r.p. 1208, lit. ‘the falling of the sky would not matter to me’; odit a’m diẟawr r.p. 1029 ‘[there is] scarcely anything that interests me’.

Pathawr (for pa ’th ẟawr) w.m. 430 ‘what does [that] matter to thee?’ Ny’m torei kyny byẟwn w.m. 172 ‘I should not mind if I were not’. Nyt mawr y’m dawr b.t. 65 ‘it is not much that it matters to me’; ni’m dorbi B.B. 60, 62 ‘it will not matter to me’. Without the dat. infixed pron.: ny ẟiẟawr, ny ẟawr cwt vo r.p. 1055 ‘it matters not, it matters not where he may be’.

(2) In Late Ml. W. the subject and remoter object came to be confused in the 3rd sg.; thus nys dawr ‘it matters not to him’ came to be regarded as, literally, ‘he does not mind it’, ‑s ‘to him’ being taken for ‘it’. Thus the verb seemed to mean ‘to mind, to care’; as am y korff nys diẟorei ef s.g. 64 ‘about the body he did not care’; heb ẟiẟarbot py beth a ẟamweinei iẟaw r.b.b. 225 ‘without caring what happened to him’.

In Late Ml. and Early Mn. W. this new verb ‘to care’ came to be inflected for all the persons; as ny ẟiẟorynt r.b.b. 216 ‘they cared not’, ni ddoraf D.G. 529 ‘I do not care’, ni ddorwn i do. 296, ni ddawr hi, ni ddorwn do. 174. In spite of this perversion the phrase ni’m dawr persisted, e.g. D.G. 138, G.Gr. d.g. 248, Gr.O. 57; also o’m dawr ‘if I care’, D.G. 246, G.Gr. ib.

(3) The interchange of t- and d- suggests the prefix *to‑: *do‑; the fut. dorbi and the v.n. show that the verb is a compound of the verb ‘to be’, the first element originally ending in a consonant, as in adnabod, gwybod. Hence we may infer dawr < *dāros’st < *dó-(p)aros est; *paros: Gk. πάρος, Skr. puráḥ, all from Ar. *pₑros ‘before’; for the development of the meaning cf. Skr. purás kar- ‘place in front, make the chief thing, regard, prefer’; with the verb ‘to be’ instead of ‘to make’ we should have ‘to be in front, to be important, to matter’. The impf. dorei must therefore have been made from the pres. dawr.

The reason for dar- in the v.n. is a different accentuation: *do-áros- > dar- § 156 i (13). The form darbod survives as a v.n. without a verb, meaning ‘to provide’, whence darbodus ‘provident’. This may have been a separate word from the outset, with *pₑros meaning ‘before’ in point of time; ‘*to be before-hand’ > ‘to provide for the future’. The verb darparaf ‘I prepare’ seems to have the same prefix compounded with *par-: peri ‘to cause’ < *q^u̯ₑr‑, √q^u̯er- ‘make’ influenced by Lat. paro (parātus > W. parod ‘ready’).

From diẟawr were formed the abstract noun diẟordep M.A. ii 346 and the adj. diddorol only occurring in Late Mn. W. and generally misspelt dyddorol ‘interesting’.

ii. (1) Ml. W. dichawn, digawn ‘can’, Mn.W. dichon, is rarely used except in this form, which is 3rd sg. pres. ind.

ny ẟichawn efeu gwnneuthur Ỻ.A. 33 ‘which He cannot do’, cf. 34, 35; llawer damwein a ẟigawn bot w.m. 28, r.m. 18 ‘many an accident may happen’.—Chwi yn falch a ddichon fod T.A. a 9817/184 ‘you who may be proud’. Ni ddichon neb wasanaethu dau arglwydd Matt, vi 24. Llawer a ddichon taer-weddi y cyfiawn Iago v 16.

A subjunct. 3rd sg. occurs in kyn ny ẟigonho y gerẟ hon w.m. 488 ‘though he does not know this craft’. In g.c. 138 we find nas dichonaf vi ac nas dichonwn pei ‘that I cannot [do] it, and could not if...’

The form dichyn M.K. [ix.] is an artificial re-formation which was in fashion for a time, and then disappeared.

(2) dichon, dichawn < *diᵹ’ᵹawn < Brit. *dī-gegāne; digawn < Brit. *dī-g'gāne; < Ar. perf. sg. 3. *g̑eg̑ōne: Gk. γέγωνα ‘I make known’; for meaning cf. Eng. can: √g̑enē- ‘know’.—W. gogoni̯ant ‘glory’ orig. ‘*fame’ < *u̯o-g’gān‑.

(3) A stem of the same form (usually with -g-) is inflected

throughout in O. and Ml. W. in the sense of 'cause to be, do, make ', v.n. digoni M.A. i 359.

Ind. pres. sg. 2. digonit B.B. 19 ( = digonyS) ; aor. sg. i. digoneis M.A. i 27ia, sg. 2., 3. dicones JTJV. SK., 3. digones B.T. 40, dichones M.A. i 273, impers. digonet W.M. 477 ; plup. sg. 3. digonsei B.T. 24 ; subj. pres. sg. i. dichonwyfw.A. i 2710;.

(4) This seems to come from Vgene- 'cause to be, give birth to', of which the pf. was sg. i. *gegona, 3. *gegone : Skr. i.jajdna, ^.jajana, Gk. i. yeyova. Whether the two roots are originally the same has not been decided. If the original meaning was something like ' to be efficient', it might have become i. ' to produce, give birth to', 2. 'be master of, understand '.

(5) Ml. W. digawn, Mn. "W. digon ' enough ' may have originated in phrases such as digawn hynny ' that will do ' understood as ' that [is] enough ' ; cf. digawn a Sodet yman K.M. 14. From digon ' enough ' a new verb was made in Mn. W., digonaf, v.n. digoni ' to suffice '.

iii. Ml. W. deryw, Mn. W. darfu 190 i (a).

iv. Ml. W. gweSa R.P. 1286 ' beseems ' 173 v (3), impf.gwebei W.M. 178 ; Mn. W. gwedda, F. 30, impf. gweddai Eph. v 3, v.n. gweddu I Tim. ii 10. Followed by i.

Other persons are found : gweS-af, -wyfiL.A.. 122, gweddynt Gr.O. 63. gwedda is a denom. from gwedd ' appearance ' < *uid-d 63 iv.

v. Ml. W. tykya W.M. 14 ' avails ', impf. tygyei ib., v.n. tygyaw do. 1 6 ; Mn. W. tycia Diar. x 2, impf. tyciai, v.n. tycio Matt, xxvii 24. Followed by i.

Ny thykya y neb ymlit yr unbennes W.M. 1 4 ' it avails no one to pursue the lady ' ; the subj. is ymlit ; thus ' pursuing avails not '.

tycia is a denom. from twg : Vteua x -, see 111 v (2); but the -c- in the pres. is caused by the -h- of -ha.

vi. Ml. W. deiryt R.P. 1197 'pertains, is related' foil, by y 'to' ; impf. deirydei s.G. 105. Mn. W. deiryd L.G.C. 272, Gr.O. 47. A 'r lludw gorff, lie daw y gyt, YV lludw arall lie deiryt. G.V., B.P. 1299.

' And [I commend] the body of dust, where it will all come, to the other dust where it belongs.'

The last syll. -yt may be the 3rd sg. mid. ending 179 iii (i) ; this would explain the limitation of the vb. to the 3rd sg. In that case deirydei is a re-formation, and the prefix and stem are devr- < *do-gr- ; the root may be *gher- 'hold' (:Lat. co-Aors); thus deiryt from

• do-ghretai ' holds himself to '. vii. metha gan ' fails ', synna ar ' is astonished ' :

Pan fethodd genni' ddyfiisio B.C. 15 ' when I failed to guess', lit. ' when guessing failed with me ' ; metha gan y buan ddianc Amos ii 14 ; si/nnawdd arnaf D.G. 386 'I was astonished at', synnodd arnynt Matt, xiii 54.

These verbs began to take the person for the subject in the Late Mn. period ; as synnodd pawb Marc ii 1 2. The transition stage is seen in synnodd arno wrth weled Act. viii 1 3, where weled is no longer, as it should be, the subject ; the next step is synnodd ef; then synnais, etc., in all persons.

Other verbs are used in a similar way in the 3rd sg., but not exclusively ; hiraethodd arno ' he longed ' ; llawenhaodd arno ' he was rejoiced ' ; lleshaodd iddo ' profited him ' ; gorfu, arno or iddo ' he was obliged ' ; perthyn iddo or arno ' belongs to him ' ; digwyddodd iddo 1 it happened to him ', etc. The subject is usually a v.n. : digwyddodd iddo syrthio l he happened to fall ' ; gorfii arno fyned ' he was obliged to go '.

§ 197. i. The verb genir ' is born ' is used in the impersonal only ; ind. pres. (and fut.) genir , impf. genid, aor. ganed, also Late Mn. W. ganwyd, plup. Ml. ganadoeb, ganydoeb, ganyssit, Mn. ganasid ; subj. pres. ganer ; v.n. geni.

Although the forms, except in the pres., are, as in other verbs, passive in origin, they take the impers. construction, being accom- panied by objective pronouns. The v.n. takes the obj. gen. : cyn fy ngeni ' before my birth ', lit. ' before the bearing of me '.

genir, ganer, ganet IL.A. 37, genit, geni do. ii,ganadoe8 H.M. ii 263, ganydoeS R.B.B. in, ganyssit do. 286.

A 3rd. sg. aor. genis 'begat' occurs in C.M. 19, in a translation, and is prob. artificial.

ii. genir < Brit. *ganl-re < *g e ne-, V gene- : Lat. gigno, Gk. yiyvo- fjai, etc. The ganad- in the plup. is the perf. pass. part. *ganatos < *g e n9-to-s ; prob. -yd- is due to the anal, of ydoedd.

§ 198. i. Ml. W. heb yr, heb y, or heb ' says, said ' is used for all persons and numbers ; the yr or y is not the definite article, as it occurs not only before proper names, but before pronouns. The Mn. W. forms corresponding to the above are ebr, ebe, eb. In Recent W. the form ebe (with -e for Ml. y 16 iv (2) ) is some- times wrongly written ebai, the -e being mistaken for a dialectal reduction of the imp ending -ai 6 iii.

Oes, arglwyo, heb yr ynteu W.M. 386 ' Yes, lord, said he ' ; heb yr ef ib. ' said he ' ; heb yr wynt do. 185 ' said they ' ; heb yr Arthur do. 386 ' said A.' ; heb y mi do. 46 ' said I ' ; heb y jxiwb do. 36 ' said everybody ' ; heb y Pwyll do. 4 ' said P.' ; heb 0/do. 2 ' said he ' ; heb ynteu do. 3 'said he'; heb hi do. 10 'said she'; heb wynt do. 27 'said they ' ; etc. Its use without an expressed subject is rare, and occurs chiefly where it repeats a statement containing the subject : Ac yna y dywat Beuno, mi a welaf, heb H..A. 126 'And then Beuno said, " I see," said [he] ' ; A gofyn a oruc i8aw, arglwyb, heb H.M. 1 79 ' and he asked him, "lord," said [he] '; heb ef. . . heb R.M. 96.

Mn. W. (N.W.) eb ni Ps. cxxxvii 4 (1588), eb ef B.CW. 8 ' said he', eb yr angel ib. ' said the angel', ebr ef do. 10, ebr ynteu do. 15, eb ej M.K. [i i], hebr efdo. [20] ; (S.W.) ebe Myrddin D.P.O. 4, eb un do. 97, ebe I.MSS. 154 ff. The N.W. dial, form ebr, e.g. ebr fi B.CW. 10, etc. is now re-formed as ebra.

Yn 61 Sion ni welais haul,

Eb Seren Bowys araul. T.A., A 14975/107.

' Since [I have lost] Sion I have not seen the sun, said the bright Star of Powys.'

ii. C. used hebaf and heb^t, see ex. ; P.M. imitating him (the two poems are addressed to father and son) wrote ny hebwn hebod M.A. i 394 ' I would not speak without thee'.

Ti hebofnyt hebu oe8 ten;

Mi hebot ny hebaf inneu. C., E.P. 1440.

' Thou without me it was not thy [wont] to speak ; I without thee I will not speak either.'

The compound atebaf ( < *ad-keb-af) ' I answer ' is inflected regularly throughout : 3rd sg. pres. ind. etyb, v.n. ateb. The rarer compounds gwrthebaf* I reply \gohebaf l \ say' (now 'I correspond') seem also to be regular : gohebych B.F. K.P. 1154 (Mn. W. 3rd sg. pres. ind. goheba, v.n. gohebu).

iii. In O.W. only hepp M.C. ( = heb 18 i) occurs, before a consonant in each case. In Ml. W. heb yr and heb occur before vowels, and heb y before consonants. Assuming that the original form in W. was *Jiebr, this would become either *hebr or heb before a consonant; the former would naturally become hebyr, later heby ; this seems to be the sound meant by heb y, the y being written separately because sounded y as in the article. Before a vowel *hebr would remain, and is prob. represented by heb yr (the normal Ml. spelling would be hebyr = hebyr). In S.W. heb and heby survived, becoming eb, ebe ; in N.W. heb and hebr, becoming eb and ebr.

If the above is correct, the original *hebr must be from a deponent form with suffix *-re added directly to the root ; thus *seq*-re, Vseq*- 1 say ' ; cf. gwyr 191 iii (i). In the face of the compound ateb = Ir. aithesc, both from Kelt. *ati-seq*-, Strachan's statement, Intr. 97, that heb 'says' is of adverbial origin seems perverse. A sufficient explanation of its being uninflected is its deponent form. In com- pounds it was regularized, and C.'s hebaf is deduced from these.

iv. The verb amkawS ' answered ' is a survival which occurs frequently in the W.M. Kulhwch, and nowhere else ; the 3rd pi. is amkeubant W.M. 486, -8, which the scribe at first wrote amkeua6t do. 473, -7, -8, -9, mistaking n for u and writing it 6.

amk-awS, 96 iii (4); if the explanation there given is correct, amkeuSant is a re-formation, possibly at first *amkeuSynt with affec- tion of aw as in beunydd 220 iv (2).

§ 199. i (1) meddaf ' I say ' is inflected fully in the pres. and impf. ind. only : 3rd sg. pres. medd, impers. meddir 'it is said '. There is no v.n.

Exx. i. Me8 seint Awstin IL.A. 42 'St. Augustine says'; 2. me8 yr ystoria do. 129 ' says the account ' ; 3. Dioer, heb y kennadeu, Teg, me8 Pryderi oe8 y'r gwr . . . W.M. 88 ' " By Heaven," said the messengers, " Pryderi says it would be fair for the man ...'"; 4. Edyrn vab Nu8 yw, meS ef; nyt atwen inheu e/B.M. 259 ' He says he is Edyrn son of Nudd ; but I don't know him '55. Blawt, meBei y GwySel W.M. 54 ' " Flour," said the Irishman ' ; 6. Broch, rneSynt wynteu do. 24 ' " A badger," said they'.

Mn. W. : meddaf I.F. I.MSS. 319, Col. i 20; meddi loan viii 52 ; medd M.K. [20]; meddant 2 Cor. x 10.

(2) In the recent period medd has tended to take the place of eb, and has almost ousted it in the dialects. But in Ml. W. the two are distinct : heb is used in reporting a conversation, and is therefore of extreme frequency in tales ; me8 is used in citing authors, as in exx. i., 2., or in quoting an expression of opinion as in ex. 3., or an answer not necessarily true, as in exx. 4., 5., 6. Hence we may infer that me8 originally meant 'judges, thinks ', and is the original verb corresponding to me&wl ' thought ' : Ir. midiur ' I judge, think ', Lat. meditor, Vmed-, allied to Vine- ' measure '. To express ' think ' a new verb meSylyaf, a denom. from mebwl, was formed, 201 iii (6).

(3) The verb meddaf ' I possess ' is however conjugated regu- larly throughout : 3rd sg. pres. ind. medd, 3rd sg. aor. meddodd W.1L. C.IL. 105, v.n. meddu.

This verb is unconnected with the above, and probably comes from /med- ' enjoy ' : Skr. mddati ' rejoices ' (from the sense of ' refreshing ' comes ' healing ' in Lat. medeor, medicus). W. meddaf is often intrans., followed byar; meddu ar 'to rejoice in, be possessed of. A common saying is Mae hwn yn well i feddu arno ( this is better to give satisfaction ', lit. ' to have satisfaction on it '.

ii. (1) The verb dlyaf (2, syll.), dylyaf (3 syll.) ' I am entitled to, obliged to ' is conjugated fully in Ml. W. : 3rd sg. pres. ind. dyly, 3rd sg. aor. dylyawb IL.A. 15, v.n. dlyu, dleu, dylyu. But in Mn. W. the inflexion is restricted to the impf. and plup. ind. with the meaning ' I ought ', more rarely ' I deserve ', and the v.n. is not used.

D.G. has dyly 28 ; elsewhere the impf. dylyivn, dylyai (misprinted dyleuaf, dylai) 35 'I deserve, she deserves'; Ni ddylyut ddile-u (mis- printed Ond ni ddylit) 427 ' thou oughtest not to destroy'. The 3rd sg. dylyai became dylai 82 ii (3), also without the intrusive y, dldi. Hence sg. i. dyldwn, 2. dylaut. These forms may still be heard from old speakers; but in the Late Mn. period a re-formed tense dylion, etc. has come into use ; and the written form is dylwn 2 Cor. ii 3, dylit Es. xlviii 1 7) dylei loan xix 7, dylem, i loan iv n, etc. The plup. in any case would be dylaswn 2 Cor. xii n, etc. In the early i7th cent, an artificial sg. 3. dyl was sometimes used.

Gwirion a ddlae a drugaredd ; a MS. ddylae,

Gwae'rferch a'i gyrro i'wfedd. D.E., c 49/33.

' The virtuous deserves mercy ; woe to the woman who sends him to his grave.' On -ae for -di see 52 iii (3).

(2) The first y in dylyaf is intrusive, and comes from dyly < *dly% § 40 iii (3). Related forms are Ml. W. dylyet, dlyet 'merit; debt', Mn. W. dyledD.W. 80, died T.A. A 14967/29 'debt', 82 ii(3); the latter is the Gwyn. dial, form ; late Mn. dyled ; Bret, die ' debt ', dleout ' devoir ', Ir. dligim ' I deserve ', dliged ' law, right ' ; all these may represent either *dleg- or *d^g- in Kelt. : Goth, dulgs ' debt ' < *dhlyh-, O.Bulg. dliigu ' debt ' ; the underlying meaning is ' to be due, or lawful ' either ' to ' (' merit') or ' from ' (' debt ') ; hence *dhlegh- ( law '. There is nothing to prevent our referring to such a root O.E. lagu, E. law, and Latin lex (ilex, Sommer 293), if for the latter we assume -gh/g- § 101 iii (1).

200. i. hwde, hwdy ' here ! take this ' and moes ' give me ' are used in the imperative only ; in Mn. W. hwde has pi. hwdyyoch ; moes has Ml. pi. moesswch R.M. 182, Mn. moeswch Gr.O. 58.

Hwde vodrwy W.M. 168, R.M. 234 'take a ring'; hwde di y votrwy honn R.M. 173 ' take thou this ring ' ; hwdy ditheu ef CM. 31 ' do thou take it ' ; hwdiwch M.K. [78], B.cw. 38.

Moes 154 iii (2) ex.; moes vy march W.M. 17 'give me my horse ' ; moes imi y gorvlwch W.M. 164 ' give me the goblet ' ; Melys ; moes mwy prov. ' [It is] sweet ; give me more ' ; moes i mi dy galon Diar. xxiii 26; moes, moes do. xxx 15; moesswch rhyngoch air Barn, xx 7.

ii. hwde is not used for ' take ' generally, but is an exclamation accompanying an offer, cf. Gwell un hwde no deu a8aw B.B. 968 ' better one "take this" than two promises' ; hence possibly hw for *hwy 78 ii < *s(u)oi 'for (thy) self the reflexive *sue- being used orig. for all persons. In that case -dy or -de is the ordinary affixed pron. (= B.B. -de, 160 iv (3), used because hw was taken for a verb), or is perhaps voc. ; hwdy di then is *hw dydt. The S.W. hwre is late, M.IL. ii 108 (not by him, see do. 319).

moes < *moi esto(d) 75 ii (2) ' be it to me ', i.e. ' let me have it ' ; cf. est mihi ' I have '. If so, i mi ' to me ' after it is redundant ; but its frequent omission makes this probable.

Verbal Stems.

§ 201. i. The pres. stem of the \V. verb, from which in regular verbs the aor. and subj. stems can be regularly deduced, may be called the stem of the verb. It is found by dropping the -of of the ist sg. pres. ind. The ending -of, as we have seen, comes from ~Brit.*-ame for unaccented *-dmi, which is sometimes original, and represents Ar. *-d~mi or *-o-mi ; but -of was often substituted for -if < Brit. *-l-me < Ar. *-e-w,and for the affection caused by Brit. *-u < Ar. *-o, the ending in thematic verbs. The W. verbal stem represents—

(1) F-grade of V, as in cymer-af 'I take', ad-fer-af 'I restore', Vbher- : Lat. fero, Gk. <cpo>. So rhed-af 'I run', gwared-af 'I succour ', eh-ed-af ' I fly ', etc.

(2) F-grade of V , as in gwan-af l I wound' < *gwon-, IT. gonim, Vg*hen- : Gk. <}>ovd<a. So pob-af I bake ', a-gor-af ' I open ' 99 vi, etc.

(3) R-grade of V, as in dyg-af 'I bring' < *duk- 182 ii (2); also V-grade, as in co-sp-af ' I punish ', Ir. co-sc-aim < *con-sq*- (' talk with '), Vseq*- ' say '. (Though in rho-dd-af ' I give ' the dd appears to be V-grade of Vdo-, in reality -ddaf represents Ar. *-do-mi with F-grade, as in Gk. Si'&o/u.)

(4) R-grade of V with w-infix, as in gann-af ' I am contained ' <*ghnd- 173iv(i), V yhed- : E. get ; andin^wnTi'Iknow' <*uind-, Vueid- 191 iii (i). W. prynaf ' I buy ' < *q*rina~mi ) V<j*reia.- 179 iii (i). The infix comes before the last cons, of the root, and is syllabic (-ne~) before a sonant ; the last cons, in *q*reia- is (a=c#), and before 9 the syllable is -na~ 63 v (2), hence *q*rina- ; cf. Gk. Dor. dámhami, Vdema-.

(5) R-grade of V + j, as in seini-af ' I sound '< *st e n-f i -, Vsten- ; sain ' a sound ' is an old v.n., cf. darstain ' to resound ' 156 i (13).

(6) V-grade of V + *ii > W. -y8-, as in b~yS-af 189 iv (i); and gweiny-af K.P. 1244 'I serve', 3rd sg. gweinyB do. 1238, gioeinySa 1254 < *uo-gn-n-, Vgerie-, 196 ii (4); the v.n. is gweini <*uo- gnlm- 203 vii (4), These represent Ar. iteratives and causatives in -eie- ( : -i- : -I-}.

(7) K-grade of V + *-isq->W. -ych-, as llewych-af (late corruption llewyrchaf) < *lug-isk-, Vleuq/g- : Gk. -I-O-KW; F-grade of V + *-sq-

> W. -ch-, in Ml. W. pu-ch-af ' I wish ' < *quoi-sq-, Vquoi- : Lith. Jcveczu ' I invite ', O. Pruss. quoi ' he will ', Lat. ms, 0. Lat. vois ' thou wishest ', Lat. invltus, (qu > Lat. v), Gk. KOITCU ywaiKuiv cTrtdu/u'at Hes. Ar. suff. *-sqe-.

(8) Other Ar. stem-forms, mostly deverbatives and denominatives, such as -d- or -dh- stems, as rhathaf, rhathu 91 ii; -t- stems, as gadaf ' I leave ' < *ghd-t- ii (2) ; -u- stems, as (gwr)andawaf ' I listen ' 76 iii (i) ; stems with -m-, as tyfaf ' I grow ' < * tu-m- : Lat. tumeo, Vteud x - ' increase ' ; etc.

ii. (i) Many verbs are denominatives formed from the v.n. as stem. Old examples are gafaelaf ' I take hold ' from v.n. gafael 188 iv ; gwasanaethaf ( I serve ' from v.n. gwasanaeth ' to serve ' ; as the latter was also an abs. noun meaning 1 ' service ', a new v.n. gwasanaethu was made from the verb, 203 i (i) ; ymdcUr(i)edaf ' I trust ' from \.n.ymddir(i}ed ; andawaffrom andaw i (8) ; cadwaf etc. 202 v. For later examples see (3).

(2) (a) The verb gadaf ' I leave, let, permit ', v.n. gadu, gadael, gadel has a doublet adawaf'I leave, leave behind ', v.n. adaw (in Late Ml. and Mn. W. gadawaf, v.n. gadaw, gaclo\ The two verbs are conjugated regularly throughout ; thus—

i. gadaf : 3rd sg. pres. ind. gad, 2nd sg. impv. gad, 2nd pi. do. gedwch, 3rd sg. pres. subj. goto = galto B.P. 1271; na at K.P. 1299

> nat do. 1216, Mn. "W. nad ' let not ', na ato > nato ' forbid ' ; from these we have nadafL forbid', v.n. nadu c.c. 187, Card, nadel.

Och arglwy8, heb y Gwalchmei, gat y mi vynet . . . Ae adu a wnaeth Arthur K.M. 181 '"Alas lord," said G., " let me go." And A. let him.' Ny adei ef hun vyth ar legat dyn W.M. 465 ' he never left sleep on eye of man.' Ym-ad a P.G.G. 22 'forego ' impv.

Gwedd ewyn, cyd gweddiwyf,

Gadu ar Dduw rannu 'r wyf. D.G. 17.

' [Maid of] the colour of foam, though I pray, I leave it to God to dispose.'

Ac ato'dd awn bei'm getid. G.G1. P 83/59.

' And to him would I go, if I were allowed.'

Nad i ferch newidio foes. D.G. 295. ' Let not a woman change my life ' (? read niweidio ' mar '). Nato Duw 159 ii (2), E.P. 274 ' God forbid' ; nadodd D.G. 105 ' prevented '. Gredwch i blant bychain ddyfod attafi Marc x 14.

2. adawaf : 3rd sg. pres. ind. edeu, Mn. W. gedy, 2nd sg. impv. adaw, 2nd pi. edewch, Mn. W. gadewch, 3rd sg. pres. sub), adawo, etc.

Adaw tiy lie hwnn IL.A. 105 ' leave thou this place '. Ac yn y llestyr yS ymolcho y8 edeu y modrwyeu W.M. 475 ' and in the vessel in which she washes she leaves her rings '. hyt nat edewis efwr byw do. 54 ' till he left no man alive '. A el y chware adawet y groen B.B. 965 ' whoso goes to play let him leave his skin behind '.

gadaf is itself prob. an old denom., i (8), from *gto-t-, Vghe- : Skr. jdhdti 'leaves', Lat. he-res, Gk. x^P 0s - adawaf is a denom. from adaw, which may be an ad-compound of the same root with u- verbal noun suffix 202 v(i); thus *ati-gh9-u- >Brit. *ate-gau- >ad-aw. Initial^- begins to appear in adaw in the I4th cent. : gedewis IL.A. 106.

The verb gadaf is in common use in the spoken lang., but recent writers seem to think that it is a corruption of gadawaf, and in late edns. of the Bible gedwch I.e. has been changed by vandals to gadewch.

(6) cyfodaf'I rise, raise', v.n. cyfodi, is generally reduced in

Mn. W. to codaf, codi (cflfod- > cywod- > co-wod- > cod-). But in lit. W. the 3rd sg. pres. ind. cyfycl Matt, xvii 23, and 2nd sg. impv. cyfod Gen. xxxi 13, remained. In the recent period, how- ever, a dial, form cwyd ( < * cw\i/d < cgwqd) is sometimes used for the former, and even as impv., e.g. Ceiriog C.G. 94.

In Gwyn. the dial, forms are cyfyd ' rises ', cw\ad ' rise ! ' the latter now being replaced by a new cod from the vb. stem.

cyf-od-af < *kom-(p)ot-, Vpet- ' fly ' : Gk. Trorrj, irero/xat, O. Pers. ud-a-patata' rises '\cyf-od- orig. ' rise ' (of birds, bees, etc.). The V also means 'to fall' Walde 2 573, hence ."W. od-i 'to fall' (of snow), as Ottid eiry B.B. 89 ' snow falls ' j hence od ' snow '.

(3) In Mn. W., especially in the late period, some verbs have been re-formed with the v.n. as stem ; thus arhoaf became arhosaf 187 ii; adeilaf became adeiladaf 203 iii(i); olrheaf'Itr&ce', v.n. olrhain 203 iv (i), became olrheiniaf-, and darlieqf'Ire&A' became darllennaf, or darllenaf, formed from the dial. v.n. darllen, for the standard form darllein, darllain.

As there is no early evidence of darllen it cannot be assumed to be from Hen < lleen < Lat. legend-, darllennaf instead of *darlleiniaf may be due to the influence of ysgrifennaf. But in S.W. it is sounded darllenaf with single -n-, as if influenced by lien. In the 1620 Bible the vb. is darllennaf Dan. v 1 7, but impv. darllain Es. xxix 1 1 , darllein Jer. xxxvi 6, v.n. darllein Act. viii 30. iii. The stems of denominatives are formed in W. either without a suffix, or with the suffixes -ycJi-, -yg-, -ha-, -ho- or -i- ; thus—

(1) Without a suffix: bwyd-af'I feed', v.n. bwyd-o, from bwyd 1 food ' ; meddiann-af ' I take possession ', v.n. -u, from meddiant * possession ' ; pur-af ' I purify ', v.n. -o, from pur ' pure ' ; arfog-af ' I arm ', v.n. -i, from arfog ' armed '.

(2) Suff. -yen- as in brad-ych-af I betray ', v.n. bradychu, from brad ' treason ' ; chwenychaf ' I desire ', v.n. chwenychu IL.A. 13, whenychu E.B.B. 89, cJiwennych D.G. 91, from chwant 'desire'; tewgch-af ' I fatten ', v.n. -, from tew ' fat ' ; on the suff. see i (7).

The relation between this and the abstr. noun ending -wch 143 iii (23) is seen in pas 'cough' < *q%dst- ( : O.E. hwosta), pesychaf'T. cough ', pesychu ' to cough ', peswch ' coughing ' ; the last is a suffixless v.n., and is still used as a v.n. in S.W. dialects. Ar. *-isq- > *-tf*x > -wch 96 iii (4), 26 vi (5).

diolwch 'to thank' W.M. n, 'thanks' do. 34, became diolch 'to thank' B.B.B. 134, 'thanks' do. 10, and *diolychaf ( I thank' became diolchafwM. 104 even earlier; diolwch < *de-ial-isq- : W. iolaf 'I praise ', v.n., ioli, eiriolaf ' I entreat ', v.n. eiriawl < *ar-idl- ; Kelt. *idl- ' speak fervently ' < Ar. *jdl- ' fervent ' : Gk. ^Xos, Dor. SAos 'zeal'.

(3) Suff. -yg-, as in gwaethyg-af I become worse ', v.n. -n, from gwaeth ' worse ' ; mawryg-af ' I extol ', v.n. -u, from mawr c great' .

The suff. is prob. a variant o-ych- after th, II, cf. -wg 143 iii (23). The stem-form of Ml. W. gwellygyaw from gwall ' defect ' has been influenced by the synonymous diffygyaw < Lat. de-fai-.

(4) Suff. -ha- ; the -k- unvoiced -6, -d, -g> and often -/, -8 111 iii. It has various uses :

(a) ' to seek ', added to nouns, forming v.n.'s without a v.n. ending : card6ta ' to beg ' (cardod ' charity ') ; biota ' to beg meal ' (blawd ' meal ') ; cica ' to beg meat ' (cig ' meat ') ; fta ' to beg corn ' (yd ' corn ') ; py%g6ta ' to fish ' (pysgod ( fish ') ; cneua ' to gather nuts ' (cnau ' nuts ') ; addra ' to go bird-catching ' (adar ' birds ') ; cynuta ' to gather fuel ' (cynnud ' fuel ') ; llygota ' to catch mice ' (llygod 'mice'); gwreica 'to seek a wife' (gwraig 'wife'); lloffa ' to glean ' (ttaw(f) ' hand '), etc. None of these has a corresponding verb, 204 i ; but many have a nomen agentis in -hai, as bldtai, cynutai 143 iv (2).

These forms are proper compounds of noun stems with *sag-<*80g-, V sag- : Ir. saigim ' I seek ', Goth, sokjan, E. seek ; thus *mlato-8ag- >*blod-ha-> biota. The noun suff. -hai < *sagifj ' seeker ' 104 ii (2).

(b) ' to go as, act as ', in Ml. W. marchockaaf ' I ride ', wa;-- chockaawb s.G. 34, marchocawn do. 35 ; v.n. marchogaeth do. I, 35. A variant of the verb is formed without a suff. : marchogaf, 3rd sg. pres. subj. marchoco A.L. i 24, imps, marckocer do. 264, also with v.n. marchogayth ib.

Brit. *markakos agdme ' I go as rider ', treated as one word, gave

• marchogha^af > marchocdaf. But the v.n. was a proper compound
• markako-aktd > *markdkdktd > marchogaeth ' to ride '. In Dyfed a

new v.n. was formed from the vb. stem : marchocdu, now corrupted to brochgdu. (-aaf implies active flexion, but the vb. wasorig. middle.)

(c) ' to become, be ' with adjectives ; as gwanhaf ' I become weak', \.n.gwankdu ; cryfhdf' I become strong ', v.n. cryfhdu, dial, cryffdu ; trugarhaf 'I am merciful, have mercy', v.n. trugarhdu, from trugar ; etc.

Brit. *udnnos agame ' I go weak ' > *uanno8-agame > W. gwanhdaf, Where a vowel drops before 8, the latter remains as h, cf. 183 ii (2). It is a common usage to stereotype the nom. sg. mas. in such phrases ; cf. Lat. 2>otis sumus, not *potes sumus, and Skr. pi. i . datasmas instead of datarah smas following the sg. ddtasmi ' I shall give ' < data asmi ' I am a giver.' W. parhaf I continue ' (v.n. parUdu, pdrd)< Brit.

• paros-agame '-I go on the same'< Lat. par.

(d) ' to make ' with adjectives ; glanhaf ' I clean ', \.n.glanMn ; ffioastataf f l flatten, straighten', v.n. gwastatdu', cadarnhaf 'I strengthen', v.n. cadarnhdu.

Brit. * glanosagame > W. glanhdaf. The nom. sg. mas. was used because it had been stereotyped in this form of phrase in group (c).

To this group should probably be referred difetha 'to mar, spoil' < *di-fe8-ha ( to make unusable, unenjoyable ', Vmed- ' enjoy ', 199 i (3). The verb was difetMaf, see difetha-awS R.B.B. 394, diffethe-eiat W.M. 29, diffetha-er W.M.L. 137 (old ff for / 19 ii (2)) ; it is now re-formed as difethaf, though the v.n. remains unchanged.

(e) ' to use ', etc., with nouns ; as dyfrhaf' I water ', v.n. dyfrhdu ; coffaf'I remember', v.n. coffdu or coffa; bwytaf ' I eat', v.n. Iwyta.

Gwyn. dial. byta<O.W. bit juv., ML "W. byd B.B. 84, variant ofbwyd 101 iii (2). This group follows the analogy of (b) as (d) does that of (c).

U Similar formations abound in Ir., Thurneysen Gr. 314; but Ir. does not help us to decide the orig. forms, as intervocalic -s- simply drops in Ir. The combination goes back beyond Prim. Kelt. ; in Lat. it is a proper compound : mltigdre, remigdre, ndvigdre, etc.

(5) Suff. -ho-; in paratoaf ' I prepare' 185 i ; crynhoaf ‘I compress, summarize ', v.n. crynhoi.

W. paratoaf < *parad-ho8-af < *pardto-sod- ' set ready ' < Lat. pardtus + *sod-, V sed- 63 ii ; cf. arhoaf 187 iii ; see also 74 i (1).

(6) Suff. -j- ; this is added to nouns, and is largely used : taniaf 'I fire ' (tan ( fire ') ; glaniaf ' I land ' (glan ' shore ') ; soniaf ' I mention ' (son 'rumour ') ; meddylwfL think ' (meddwl ' thought ') ; rhodiaf ' I go about ' (rkawd ' course ') ; etc.

This is the Ar. denominative suff. *-ie-, as seen in Skr. apas-yd-ti ' is active ' from dpas- ' act ' ; Gk. reXeio) ( < *reXr-<>) from reAeo-- ; &7\oo> (< * 817X0- <>) <&7\o-s ; etc., Meillet, Intr. 2 183.

In old formations the -i- of course affects the preceding vowel in W. ; thus niweidiaf'I injure ' : niwed, Ml. "W. er-nywed w. 480, 76 iv (4); imdiaf, \.n.peidio 'to cease, be quiet ' < Brit. *pat- < *g?()9-t-, Vq*eie- ; L&t.'quies.

In W. the suff. is not added to adjectives. But -at-, Ml. -ei- in the ult. may be caused by the of the lost adjectival ending *-ios ; and the i is kept in the vb. ; thus disglair ' bright ' < *de-eks-kl(t)ari6s ( : claer < *kliiaro8 75 vi (i)) ; hence W. disgleiriaf I shine , v.n. disgleirip. From these forms it was extended to other adjectives with -ei- as perffeithio, perffeibyaw from perffeith<^t&i. perfectus ; and with -i-, as gwirio ' to verify ' from gwir 35 iii (but cywiro from cywir).

The suffix is generally used in verbs borrowed from Eng. ; thus pasiaf ' I pass ', passiodd Can. ii 1 1 ; peintyaw ' to paint ' ( 1 6 v (2)) ; yatopyawS S.G. 72 'stopped'. In some cases two forms are used; thus ffadyaw s.G. 285 'to fail ', beside ffaelu do. 348, the latter being the treatment of native words with -ae-; both forms are still in use. So helpio and helpu.

VERBAL NOUNS.

202. i. (i) The v.n. often consists of the stem of the verb with no ending: aieb 'to answer', vb. atebaf\ edrych 'to look', vb. edrychaf\ dangos 'to show', vb. dangosaf\ adrodd 'to recite', Ml. W. adrawb, vb. adroddaf\ anfon ' to send ', vb. anfonaf\ bwyfa ( to eat ', vb. bwytd-af ^etc. ; Ml. and Early Mn. W./o 223 i (2), Late ffoi ' to flee '. (2) This form implies a lost monosyllabic ending, most probably neut. *-os as in Gk. γέν-ος, Lat. gen-us; thus ateb < *ati-seg*-os. The loc. *-es-i of this gives the Lat. inf. -ere; thus O. Lat. genere ‘gignere’ = genere abl. of genus, Brugmann² II i 525. The W. v.n. may be acc., in which case it often stands, as gallaf ateb ’I can answer’. But it may also be nom. as ateb a wnaf ‘[it is] answer that I will do’. The word is the same as the abstr. noun ateb ‘an answer’; and perhaps need not be assumed to come from an oblique case.

(3) The verb trawaf HM. ii 252 has 3rd sg. pres. ind. tereu B.B. 63, and v.n. taraw H.M. ii 253. The vowel in the first syll. of these two forms is probably intrusive 40 iii (3) ; it does not occur in the other forms of the verb in Ml. W. : trawei W.M. 24, B.M. 15 ; trewis W.M. 80, 90, H.M. 58, 66, C.M. 1 8 ; trawawS s.G. 18, trawssant do. 31, trawssei do. 6 1 ; trawher W.M.L. 3, trawhet do. 29. The Late Mn. tarawaf, tarawiad are artificial ; the natural forms are still trawaf, trawiad. If the etymology trawaf < *trug-ami (ru<ur] 97 v (3) is correct, it does not admit of a vowel between the t and r.

(4) Many verbs which seem to have sufiixless v.n.'s are denomina- tives formed from the v.n. 201 ii (i), (3), and v (i) below.

ii. The ordinary endings used to form verbal nouns are -u ; -o, Ml. -aw ; -i.

-u and -aw represent forms of v.n.'s of verbs of the a conjugation. -u < *-au-, prob. < \oc.*-a-uen (or nom.-acc. *-a-un) : Skr. dat. da-vdn-e, Gk. Cypr. 80- f ev-ai, Att. Sowcu ; (*-a-uos is also possible, with the suff.

of byw 204 ii (5) ; but this is a rarer form). aw for *-aw(f]< loc.

• -a-men (or nom.-acc. *-a-mn) ; see 203 ii (4); but Ir. has also -mu-

203 vii (4), and -ma-, beside -m e n- flexion.

-i belongs to the I conjugation ; the O.W. form was -im ( = -iv) 110 iii (5) ; hence from *-i-men (or *-l-mn), as assumed above for -aw ; thus rhoddi ' to put' < Brit. *ro-di-men<*pro-dhe-men.

In Ml. and Mn. W. the use of the above endings is determined by the form of the stem, as follows :

iii. -u is added to stems in which the vowel of the last syllable is a, ae, e or y ; as canu, pallu, diddanu, tarfu ; taeru, arfaethu, saethu, gwaedu ; credit, trefnu, sennu, ttedu ; nyddu, crynu, prydyddu, melysu. Exceptions : a few stems having a, v (3) ; gwaeddaf ' I cry ' has v.n. gweibi R.M. 174, IL.A. 154 ; tnedafbas medi B.B. 45.

Ml. W. cae-u W.M. 24 c to shut' is contracted in Mn. W. to caw 38 iv, 52 iii (3).

iv. -aw, Mn. -o is added to (i) j-stems ; thus mebylyaw W.M. 10, tygyaw do. 16, rhodio, diffygjo, teitfiip, gweitJiip, seilio, hoelio, etc. In Ml. W. the i is often omitted, as treufaw W.M. 6, Mn. \V. treulio; keissaw do. 487, Mn. W. ceisio, 35 ii (i). Some -stems have other endings, see 203 iii (a), vi (i), (a), vii (i).

A few i-stems have suffixless v.n.'s ; thus kynnigyaf has Icynnic W.M. 30 'to offer'; disirywyaf has distryw E.B.B. 159, distriw do. 89, now distrywio. In Late Mn. W. meddwl, son have superseded meddylio, sonio as v.n.'s. In daliaf the. -i- represents original -g-, and daly, ddla, late dal represent an original suffixless *dal%, see 1 10 ii (2). Similarly hely, hela, hel, vh. Jieliaf, heliodd Gen. xxvii 33 ' hunted ' ; but N.W. has beside hel ' to gather ', hel-a ' to hunt ' where -a may be the stem suff. -ha of a lost vb. *helhaaf, seen in O.W. in helcha gl. in venando, helghati ' hunt thou '. The -i- of bwriaf is from -g- which appears as w in the v.n. bwrw, see ib.

As ai is ei in the penult 81 i, and stems with -ei- take -*- 201 iii (6) it is seen that denominatives from nouns and adjs. with -ai- must have v.n.'s in -io ; thus areithio, disgleirio, diffeithio,gwenhieithio from araith, disglair, dijfaith, gweniaith. (If these had been araeth, disglaer, etc., as now often misspelt, the v.n.'s would be, by iii above,

• araethu, *disglaeru, etc., which are never spoken or written.) There

is only one exception ; cyfieiihu (a late word) has -u because the vb. cyfieithiaf became cyfieithaf by dissim. of i's ; the regular cyfieithio also occurs, P 2 1 8/1 79 K.

(a) stems having i, u, en, wy ; as blino, gwrido, llifo, rHfo ; euro, dymunO) grymM&o ; euro, heulo, ceulo ; bwydo, rhivyfo, arswydo, twyllo.

Ml. W. dinustyr M.M. 32 ' destroys ' has v.n. dinustraw K.P. 1246; in Mn. W. dinustr became dinistr by 77 ix, and the verb is re-formed with -i- suff., v.n. dinistrio Deut. xii 2. (The late dinystrio is a mis- spelling ; the sound in the penult is not y but i. ) dinustr < *de-nou(i)- stro- : niwed 76 iv (4), suff. as in Lat. monstrum.

Some stems ending in -eu have suffixless v.n.'s, as dechreu ' to begin ', madden ' to part with, to forgive ', ameu ' to doubt ', vb. amheuaf. Also in Mn. W. tramwy Job i 7, arlwy D.G. 104.

On account of the early change of wy to wy after a vowel, we have -u for example in tywyllu ; in these cases, therefore, the suffix is no guide to the orig. form, (tyunfll < tywyll 1 1 1 i (2).)

v. -i is added to (i) stems ending in w ; thus berwi, cAwerwi, enwi, sylwi, gwelwi. Some of these have suffixless v.n.'s, as cadw, galw, marw.

Two distinct formations are represented here. i. In verbs which take -i the -w- either forms part of the root, as in berwi, Vbhereu- 63 iv, or belongs to the stem of the noun or adj. of which the verb is a denom., as gwelwi, from gwelw ' pale '. 2. In verbs which do not take -t the w is itself the v.n. suffix, from *-uen (or *-uos), see ii above, and the vb. is a denom. formed from the v.n. ; thus cadw ' to keep ' <*kat-uen, Vqat- 'hide, cover, keep' : O.H.G. huoten'c&re for, keep ', E. heed, Lat. cassis ; galw ' to call ' < *gal-uen, Vgal- : Lat. gallus, E. call. For two of these v.n.'s, by-w and mar-w, no verbs were formed 204 ii.

Though the classes remained distinct, a v.n. of one class was liable to pass over to the other; thus merwi P 12/1 24 R. 'to die'. From Brit. *lanos (<*p|ru)-s 63 vii(2)) 'full', a v.n. *lan-uen would give *llawnw, from which may come Uanw E.M. 94, llenwi W.M. 23, R.M. 15, llewni R.M. 175. arddelw (now arddel) for arddelwi.

Stems ending in -aw are similarly divided : suffixless adaw 201 ii (2), gwrandaw do. i (8), taraw i (2) above ; with -i, tewi, distewi only (taraw having gone over to the other class). D.G. uses distawu 165, andMn.W. croesawafhas croesawu.

(2) stems having oe or o, whether the latter be original o or a mutation of aw ; thus oedi, troelli, oeri, poethi ; ttonni from llonn ' merry ', cronni from crawn { hoard ' ; torri, cyflogi, arfogi.

Some stems having o take no suffix, as dangos, anfon, adrodd i (i). Ml. W. ccgori W.M. 59, 60, R.M. 42, A.L. i 498, D.G. 134, Can. v 5 is later agaryd M.K. [30], Dat. iii 20, or agor M.K. [32], B.cw. 56 (agori R.M. 174 with punctum delens, 1 late).

(3) some stems having a, which is affected to e ; as ercki, vb. arcAaf; peri, vb. paraf\ senffi, vb. sangaf\ perchi, vb. parckaf; (fefni, vb. dafnaf.

mynegi, Ml. W. menegi, and trengi are stems with a ; in Ml. and Early Mn. W. the verbs are managafsuid trangaf; but in Late Mn. W. the e of the v.n. has intruded into the vb., and mynegaf, trengaf are the usual forms. In B.B. 8 delli (II ≡ l-T) occurs for the usual daly, dala iv (i). defni is also an abstr. noun ' dropping ' Diar. xix 13, xxvii 15 ; and is often taken for a pi. ofdqfii e.g. I.MSS. 232 ; v.n. defni IL.A. 23.

203. Verbal nouns are also formed by means of other suffixes, as follows :

i. (1) -aeth in marchogaeth 201 iii (4) (6), and Early Ml. W. gwasanaeth 'to attend, serve', later gwasanaethu.

Ef a 8ety guassanayth ar e vre[nhines] . . . Ny 8ely ef eyste, . . . namyn guasanayth oy sevyll B.CH. 22 ' He is to attend on the queen. He is not to sit but to serve standing*. In both these examples Aneurin Owen prints guas(s)anaythu A.L. i 54, 56. In 60 he gives guasanaeth correctly, with -u as the reading of MSS. C.D.E. In Late Ml. W. gwassanaethu R.M. 174 is the form used. See 201 ii (i).

-aeth became an abstr. noun suffix 143 iii (2). Thus a lost verb *hir-hd-af ' I long ' had a v.n. hiraeth ' *to long ' which came to mean ' longing ' ; from this was formed the denom. hiraethaf ' I long ' with v.n. hiraethu ' to long *.

(2) -ael or -el, in gafael, gafel 188 ii (3), 201 ii (i) ; ca/ael, caff el, cael 188 i (8) ; dyrchafael 188 iii; gadael beside gadu 201 ii (a) ; gallael beside yallu ' to be able '. Probably -ael is original only in gafael 188 iv.

(3) -ach, in cyfeddach ' to carouse ', prystellach, ymdesach, din- darddach, caentach p 5/x R. These have no verbs. The ending is sometimes substituted for another : chwiliach ' to pry ' for chivilip ( to seek *.

(4) The above are v.n.'s from Vag- seen in -ha- stems ; thus -aeth <*-ak-ta (<-dk-<-o-ak- ; the a would be shortened before let even though accented) 201 iii (4) (6), with *-ta suffix as in bod 'to be ' < *bhurta. -- ael : Ir. -ail < *-ag-li-s, with fern. *-li- suff. ; the suff. *-li- forms fern, abstract nouns in Slavonic also, and Armenian has an infinitive suff. -I. -- ach<*-aks-<*-ag-s-, perhaps *-ag-sen ; cf. Gk. φέρειν < *φέρε-σεν.

ii. (1) -ofain, Ml. W. -ovein, -ovant occurs beside -aw in wylofain, Ml. W. wylovein IL.A. 117, 154, cwynovein s.cf. 343, cvinowant B.B. 46. A similar formation is digofain G. 132 ' wrath ', later usually digofaint, abstr. noun.

(2) -fan(n), -fa, in cwynfan. Ml. W. kwynvan IL.A. 154, beside cwynaw, cwyno ; gribvan IL.A. 154, Mn. W. griddfan whence vb. griddfannaf, gribva R.M. 132; ehedfan Ml. W. ehetvan, beside ehedeg, vb. ehedaf * I fly '.

(3) -ain, Ml. -ein, in ttefain, ttevein R.M. 132, vb. lief of * I cry ' ; llemain, late llamu, Ml. llemein, vb. llamaf ' I leap ' ; germain ' to cry', Ml. germein, no verb; ochain, vb. oc/iaf; ubain, diasbedain with no verbs. With -t in diobeifeint IL.A. 129 'suffering', Mn. W. dioddefaint, only occurring as an abstract noun.

(4) cwynaw < Brit. loc. *kein-a-men, denom. from *qeino-, 101 ii (2) ; ct0ywo/imi from the dat. of the same stem, *&em-a-w e < *-m e nai : Gk. 8o-^u,vai, Skr. da-mane ' to give '. -- ovant represents another case, prob. loc. *-a-m e ni ; -ant < -ann < *- e n-, 62 i (2). -fan(n), -fa represents the same case as the last, but with a different accent ; thus griddfan ' to groan ' < *grido-m e ni, V ghrewF- : O.E. granian, E. groan ; cwynfan is similar, or formed by analogy. lief ' cry ' < TSritSlemen < *lep-men, < *lep- : Skr. Idpati ' chatters, murmurs, laments ' ; lief ain is the dative *lemanl< *lep-m e nai ; llefaf is a denom. from lief, Similarly garm 95 ii (3), germain < *gar-sm e nai ; no verb was formed for this ; llam : llemain, vb. llamaf denom. From these -ain was deduced and added to the intjs. ub and och and to the noun diaspad (diasbad) 'cry '; vb. ochy IL.A. 154 ' groanest', ocha G. 196.

iii. (i) -ad, Ml. -at, in adeilad 'to build', Ml. adeilat R.M. 93, R.B.B. 56, 58, 59, IL.A. 123, verb adeilaf} Ml. W. gicylat W.M. 74, R.M. 53 ' to watch ', gwylyat s.Q. 2, vb. gioylaf W.M. 74, K.M. 53, also with y; chwibanat c.M. 48 'to whistle', vb. chwibanaf; dyfiead B.CW. 124 'to gasp ' beside dyheu, vb. dyheaf.

In Late Mn. W. adeilaf has been replaced by a denom. of the v.n. : adeiladafi from which coraes a new v.n. adeiladu. The orig. meaning was ' to form a wattle ' ; the absence of i after I points to eil coming from *egl- 35 ii (3) ; hence adeilaf < *ati-egldmi < *-peglo- by dissim. for *pleg-lo-, Vplek/g- : Lat. pledto, Gk. TrAeVw, TrAcy/ia.

(2) -aid, Ml. -eit, added to e-stems: ystyryeit C.M. 61 'to consider ', synyeit W.M. 33, R.M. 22 ' to take thought ', tybyeit s.G. 75 ' to imagine ', mefylyeit M.A. i 251 ' to think ', ervynnyeit IL.A. 125 'to implore', y*fflyfyeit C.M. 5 'to snatch'; Mn. W. meddyljaitl D.G. 22, ysiyriaid, syniaid, tybiaid ; -o is also used with these stems ; and erfyn is now suffixless. -eit is added to one w-stem : ysgytweit R.B.B. 58 beside ysgytwaw IL.A. 1 66, Mn. W. y%gwyd y vb. ysgydwaf ' I shake '.

On account of the dial, reduction of at to e, 6 iii, this suffix is confused in Late Mn. W. with -ed ; thus, ystyried, synied, tybied ; these three are in common use. G.J. wrote ystyriaid correctly, Hyff. Gynnwys 28.

The form ysgwyd is v.n. and 3rd sg. pres. incl. ; it is for ysgydw by metath. of w, cf. echwyb ' evening, west ' < Lat. occiduus ; ysgydw 'brandishes' occurs M.A. i 285; stem ysgydw- < *8quiu- for *squti- < *8qu e ti- : Lat. quatio for *(s)qitatio, O.Sax. skuddjan ; W. sgytio f to shake violently ', ysgwd ' a push '.

(3) -ed, Ml. -et, in kerbet W.M. 486 ' to walk ', Mn. W. cerdded, vb. cerddaf; clywed 194 v (i) ; guelet W.M. 17, Mn. W. gweled, gweld ' to see ', vb. gwelaf; yfet W.M. 182 ' to drink ', Mn. ^N.yfed. vb. yfof\ myned ' to go ', vb. af 193 ii. The -ed became part of the stem in dan-ivared ' to mimic ', V uerei- 63 vii (3) ; ymddiried, ymtiret ' to trust ' : dir ' true '.

cerBed < *kerd-, Corn, cerdltes ' to go, walk ', Bret, kerzet id., Ir. ceird ' walk ', Vsqerdd- ' turn about ' ; Gk. Kop&a, Lat. cardo. gwelaf ' I see', Bret, gwelout, Corn, gweles ' to see ' <*ghud- 93 iv : Lith. zveilgiu ' I look towards ', zvilgeti ' to see ', Gk. ^e'Xyw ' I fascinate ' (" fascination is ever by the eye " Bacon), V ghud-, extd. *ghul-g-. There is also a gwelaf from Vud- ' wish ' : tra welho Duw W.M. 72 4 while God will ', Mn. W. os gwelwch yn dda ' if you please '.

(4) -nd, Ml. -ut, earlier -wyt, in dywedut 194 i (4) ; kyscwyt B.T. 27, usually cysgu 'to sleep'; cadvid ( = cadwyd) B.B. 62, cadwydvf. loa, usually cadw 'to keep'; ymchwelut W.M. 10, 14, s.G. 23 Ho turn', ymchoelut C.M. 5. There is some confusion, even in Ml. W., of this suff. and -yt ; thus the last word is written ymchoelyt in R.M. 7. In Mn. W. dywediit is spelt dywedyd ; but (T.J. wrote dywedud Hyff. Gynnwys, p. iv.

(5) -yd> Ml. -yt, in kymryt, diffryt 195 ii ; etvry't C.M. 24 ' to restore ', edryt R.B.B. 6 (by 110 iii (3)), Mn. W. edfryd D.P.O. 132, mostly replaced by adfer in the late period, vb. adferaf; ymoglyt W.M. 104 'to beware ', re-formed as ymogelyd in Mn. W. ; yochlyt IL.A. 26, beside gochel ib., R.B.B. 106 ' to avoid ', Mn. W. gochel, vb. gochelaf; diengyt IL.A, 72 (Gwyn. dial, dewnid) beside diang ib., Mn. W. dianc ' to escape '.

(6) -d, Ml. -t, in Ml. W. dilit W.M. 41, R.M. 28 ' to stick to, follow' (also written dilyt W.M. 41, 77 iii, and later assumed to have -y-^ but this is an error, the older rhymes having -i-, as Hid M.A. i 408), verb dilynaf ' I follow ', whence in Mn. W. the v.n. dilyn, also dilin D.G. 343 ; Ml. W. erlit w.M. 16, Mn. W. erlid ' to chase, persecute', vb. ertynaf; in Mn. W. a new vb. erlidiqf is formed from the v.n., and a new v.n. erlyn from the vb. ; ymlit W.M. 14 ' to chase ' from which a denom. was already formed in Ml. W., e.g. R.M. 64 ; bod ' to be ' 189 iv (6).

(7) -s, for 8 in the stem, in go(r]biwes t vb. go(r)Uwebaf 194 vi, Mn. W. v.n. goddiwes, late goddiweddyd ; aros, verb arAoqffor *arhobaf 187.

A megys nat ymoBiweS un creadur a Duw, ac ef yn y moSiwes a phob peth . . . IL.A. 10 ' And as no creature apprehends God, and He apprehending everything . . .'

(8) Verbal nouns were formed in Ar. by means of suffixes *-tu- (: Lat. supine -tu-), and *-ti-\ Kelt, had also f. *-ta, 189 iv (6). The preceding vowel is generally, but not always, R-grade. adeilad 1 to build ' = adeilad f. ' a building ' <*ati-eglatd < *-9-td. eit may be from a dsA,.*-ati < *-s-tiai. et < *-i-ta ; as yfed ' to drink ' < *pibi- ta, Vpoi-', also from *-e-to-, *-e-ta, cf. 143 iii (12). The y of -yt comes from the i of*re, *li representing Ar. *r, *l; thus corresponding to cymeraf ' I take ' < *kom-b}ier- with F-gi ade *bher-. the v.n. had R-grade *bhr-, as *kom-bhr-tu- > W. cymryd; so goglyt<*uo-ty-tu-, Vkel- 'hide . The -y- tended to spread from these. The v.n. dilit is a similar formation, < *de-ll-tu-, where *-ll- is R 2 -grade of Vleia x -' to stick'; the vb. dilynaf < *de-lind-mi, with w-infix 201 i (4). The verb *lynaf ( : Ir. lenim) disappeared in W., and its compounds, as erlyn (for *erUyn), show the influence of the synonymous glynaf. s (for -8) represents -d-t-, 187 iii. ud <-wyd is a different formation from the others ; the most probable explanation of the wy seems to be that it comes from new ai 75 i (3) ; thus dywedut< dywedwyt<*do-uet-aito< *do-uet-at%o : Lat. abstr. suff. -atio, see vii (3).

iv. (1) -ain, Ml. -ein, for -<?- in the stem, in olrein, darllein 201 ii (3), dwyrein ' to rise ', the latter surviving only as a noun meaning 'east'. Examples of the verbs: olrewys W.M. 469, (larllewyt do. 49, dwyreawb M.A. i 300. The v.n. arwein has vb. ancebaf in Ml. W., later arweiniaf; so kywein \cywebei W.M. 119.

Gioell kadw noc olrein B.B. 968 ' better keep than seek.'

(2) -wyn, for -yg-, -wg in the stem : dwyn, verb dygaf 194 iii ; ymddwyn, vb. ymddygaf'ih. ; amwyn verb amygaf 194 iv ; adolwyn, beside adolwg, also atolwg in Mn. W., verb adolygaf'1. pray '.

rac adolwyn y un vynetn.u. 197 'lest any should be asked to go', Adolwg a wnaf G.Y.C. M.A. 1517 'I pray '.

(3) The suffix is *-no- ; cf. O.E. -an< *-o~no-. *-egno- > -ein 104 ii (i). Medial -eg- before a vowel >-ig- > y or e. *-uk-no- >-wyn 104 iii (i). ar-, cy-wain seem to come from Vuegh- 65 ii (3), but the verbs imply Vuedh- 149 i; as dn did not become gn (e.g. llyn-e?>, not *blin- 104 iv (i)), we cannot assume Vuedh- for the v.n.'s.

v. -eg, Ml. -ec, in rhedeg ' to run ', vb. rhedaf; ehedeg ' to fly ', vb. chedaf.

-eg< *-ika abst. noun (orig. adj.) suffix, 143 iii (14).

vi. (1) -an, added to ?-stems, borrowed from O. or Ml. E. ; as hongian ' to hang ' (O.E. hangian), yttwyrian ' to stir ' (O.E. atyrian) ; hence added to others as trottyan B.P. 1272, mwml'tan 1 to mumble '. Added to W. stems -ian forms a sort of pejorative v.n., as gorweddian 'to lie about lazily', ymlwybran ' to plod one's way', sefyllian 'to loaf. It is not much used in the lit. lang. Without i it appears as an abstr. suff. : cusan, Ml. W. cussan ' kiss ' < O.E. cyssan ' to kiss '.

Eng. strong verbs generally become t-stems in W. with v.n. -o as yildip 'to yield '< O.E. gildan; cf. 201 iii (6).

(2) -al seems to be a variant of -an arising from dissimilation in nasal stems; thus Uncial beside tincian 'to tinkle', mevnal beside mewian 'to mew'; cyfnewidial D.G. 145 for cyfnewidio ; naddial for naddu, techial for techu ; sisial whence vb. sisialqf ' I whisper ' ; myngial ' to mumble ', no vb.

vii. Each of the following v.n.'s has a form peculiar to itself:

(1) lluddias G.G1. F. 14, Ml. W. lluddyas IL.A. 19 'to hinder', vb. lluddiaf, 3rd sg. pres. ind. lludd D.G. 105, aor. sg. 3. llubywyx W.M. 103, lluddiodd D.G. 105.

The suffix is prob. the same as the abstr. noun suff. -as 143 iii (6).

(2) arediff, Ml. W. eredic B.B. 44 'to plough', vb. arddaf 100 iii (2). There is a v.n. erti ( = erbfy in B.B. 55, and a recent artificial arddw ; but the v.n. in common use is arediff.

Pwy bynnac a dorro tervyn oy eredyc, y brenhyn a 8yty yr ychen ay harSo A.L. i 196 ' Whoever shall destroy a boundary by ploughing it, the king shall be entitled to the oxen that plough it.'

The ending is similar in formation to that of the v.adj. -ediy 206 vii.

(3) chwerthin 'to laugh', vb. chwarddaf, 3rd sg. pres. ind. chweirb or chwardd 173 iv (2).

chwerthin B.M. 185, 237, wherthin W.M. 171. chwerthin is also an abstr. noun meaning 'laughter/

Gweniaith brydferth a chwerthin Erioed a fu ar dy fin. D.G. 108. ' Pretty flattery and laughter have always been on thy lip.'

chwarSaf<*s-uar-d-ami, d-stem, Vuere(t)- : Lat. ndeo 63 vii (5) ; chwerthin, Bret, c'hoarzin, < *s-uar-tln-l < *s-u e r'-tln-ai, dat. oi *s-u e r'-tiio, abs. noun in *-tiio : Lat. -tio ; the oblique cases have *-tin- ( < *-tii e n-) in Kelt, as in Osc.-Umbr., not -tion- as in Lat., Brugmann* II i 319. The dat. of this stem occurs as infinitive in Ir. also : do saigthin ' to seek '. chwerthin seems the only survival in W. The use of -tio as an abs. noun suffix is a feature of Italo-Kelt.

(4) gweini ' to serve ', vb. gweinyddaf.

The -i of gweini may represent the I which stands in ablaut with -eie- ; *uo-gnx-mu- > gweini : O.W. gnim, Ml. W. gnif, Ir. gnlm, u- flexion. On the verb see 201 i (6).

(5) * e fyM ' ^ stand ', vb. safaf.

safaf is a denom. from a noun *sth9-mo-s, Vstha- 'stand*, like tyfaf I grow* from *tumo-8, 201 i (8); 8eJyll<*8th9m-i-U-s, with the iterative and causative -i- ( : -eie), and the suffix *-li-, as in gafael i (4) ; I between i's gives W. II 111 i (2).

(6) gwneuthur ' to make, do', vb. gwnaf 193 x (4).

(7) There are one or two other anomalous forms such as chwUiath D.G. 319 (beside chwilota) annos ' to incite ' (beside annog)\ gwastrod-edd Gr.O. 178, 300 from gwaxtraiod 'groom', suff. 143 iii (13).

§ 204. i. Many verbal nouns have no verbs, but are used exactly like other v.n.'s in construction. Most of them have been named : cardota, biota, etc. 201 iii (4) (a), cyfeddack, etc. 203 i (3), germain, etc. 203 ii (3) ; godro ' to milk'; ym-ladd 1 to tire one's self ' <*mbi-l9d-, Vied- : Gk. XrjSeiv ' to be tired', Lat. lassiis 156 i (2) ; but$m-la6 ' to fight', V qotid- 101 ii (3), is conjugated throughout ; 41 i.

ii. The most important v.n.'s without verbs are byw ' to live ' and marw ' to die '. They are also abstract nouns, and adjectives.

(1) They are v.n.'s after wedi, or yn with the radical, in periphrastic conjugation or forming participle equivalents :

Os marw bun, oes mwy o'r byd ?

Mae'r haf wedy marw hefyd. T.A., c. ii 79.

' If the maiden is dead does the world any longer exist 1 Summer is dead too.'

I fardd ydwyf, ar ddidol,

Yn brudd yn byw ar i 61. T.A., A 24980/166.

' His bard am I, in seclusion, living sadly after him.'

Also when qualified by an adverbial expression consisting of yn and an adj., as byivn gymwys W.IL. F. 32 ' to live justly '.

Gwell bedd a gorwedd gwirion

Na byw'n bir yn y been hon. D.G. 108.

' Better the grave and innocent rest than to live long in this pain.'

(2) They are abstract nouns when qualified directly by adjectives, as marw mawr ' great mortality ', byw da ' good living ', or when they follow yii, with the nasal mutation :

A m ych dwyn ym my w 'ch dynion

Yr oerai'r sir, eryr Sion. T.A., G. 229.

' Because you were taken in the lifetime of your men the shire became cold, eagle[-son] of Sion.'

Also generally with prefixed pronouns : odihenghy a'th vyw gennyt w.M. 476 ' if thou escapest with thy life ' : Mn. W. yn ft/ myw ' in ray life ', meaning ' for the life of me '.

(3) They are adjectives when they qualify nouns expressed or implied :

Y gwr marw, e gdr morwyn

Ddaear dy fedd er dy fwyn. T.A., G. 229.

' Dead man, a maiden loves the dust of thy grave for thy sake.'

Ar 61 y marw yr wyli,

Ar 61 y fyw'r wylaf fi. D.E., p 112/840.

' Thou weepest for him who is dead ; I weep for her who is alive ' ; lit. ' the dead ' mas. sg., ' the living ' fern. sg.

When following- yn with the soft mutation :

Ni bu 'n fy w, cyd bo 'n i fedd,

Ni bu 'n farw neb un fawredd. H.K.

' There has not been alive, though he is [now] in his grave, there has not been dead any one so great.'

Also when they are complements, without yn, of the verb ' to be', as byddfyw lit. ' be alive' i.e. live ! hwnmo afydd marw Ezec. xviii 4.

Fy Nuw, pei cawn fy newts,

Ni byddai fyw o'm bodd fis. D.Gr. 174.

' My God, if I had my choice, he would not be alive a month with my consent.'

Silvan Evans s.v. byw treats the word as a v.n. here ; but no v.n can stand in this position. We cannot say bydd rhodio for ' walk ! ' but we say bydd da or bydd dda ' be good ! '

(4) As adjs. they have pi. forms bywion E.P. PS. xxvii 13, Ivi 13, late and rarely used, and meirw, meirwon. In periphrastic conj. the pi. meirw is used for the v.n. when the subj. is pi., as y maenf wedi meirw, by a confusion of the v.n. and adj. ; cf. gwedy myfav M.A. i 228 for ffwedy mebwi pi. Compared : marwed'L.G.C. 218 'as dead', yn gynvywyet s.G. 77 ' as alive '; manoach, S.Ph. CY. ix 34.

(5) byw < *g*i-uo8 63 vii (3), which may be an adj. like Lat. vlvus, or a noun like Gk. /Si'os. marw < *m e ruos ( e r, not r, before u r 63 iii) similarly formed from 4/mer-; in Lat. mortuus<*mrtuos the t is intrusive according to Brugmann 2 II i 448.

§ 205. The v.n. is always mas. in construction. But many of the forms were originally fern., and some remain fern, when used as abstract nouns. Thus bod< *bhu-td is f. in hafod 189 iv (6); abstract nouns in -aeth are f. 139 ii ; gqfael noun is f., gafael gadarn W.M. n, Mn. W. yr afael 'the hold'; and cerbed in y gerbet honno R.M. 183; so adeilad f. 'building', cyfeddach f. as y gyfeddach. Doubtless rhedeg, ehedeg, etc. were originally f. ; O.W. trited retec CP. c third (f.) course '.

But since v.n.'s were made mas., abstract nouns of the same form have tended to follow them. Thus gwasanaeth ' service ' is mas. , because the same form continued in use as a v.n., 203 i (i). Similarly hiraeth rnas. ' longing ', see ib.

VERBAL ADJECTIVES.

§ 206. i. Verbal adjectives are formed from the stem of the verb either without a suffix or with the suffixes -edig, -adwy and in Ml. W. -awt, -ediw. The last two suffixes are rare, even in Ml. W. ; only a few verbs have suffixless v.a.'s ; but all regular transitive verbs may have v.a.'s in -edig and -adwy. The former has usually the sense of a past pass, participle, the latter of a fut. pass. part, or gerundive.

ii. Suffixless : plan E.P. PS. cxxxvii 2 ' planted ', verb planuaf ' I plant ' ; pryn l bought ' vb. prynaf ; cwsg ' sleeping ' in bardd cwsg, vb. cysgaf\ llosg in marwor llosg 'burning embers', vb. llosgaf\ tawb ' melting ' R.M. 169, Ps. Iviii 8, ' molten ' Lev. xix 4.

iii. Suff. -edig, Ml. W. -edic : darparedic W.M. 23 ' prepared ', betidicetic L.L. 121 'blessed', emelldicetic ib. 'accursed', lladdedig ' killed ', gweledig ( seen ', caredic W.M. 37 ' loved', Mn. W. caredig ' kind ', crwydredig ' wandering '.

iv. Suff. -awt : guit gvyrhaud (=ywyb ywyrhawff) B.B. 89 ' trees [are] bent ' ; At(wyn) lloer Uewychawt B.T. 9 ' sweet [is the] moon shining '.

v. SufF. -adwy : dywedadwy 169 iv (i) ' to be said' ; credadwy C.M. a i 'to be believed, credible' ; moladwy B.T. 71 'to be praised, praiseworthy ' ; ofnadwy ' terrible ', vb. ofnqf'I fear' ; cyraeddadwy 1 attainable ', vb. cyrhaebaf ' I reach ' ; safadwy ' stable ', vb. safqf

• I stand '. -ediw is a variant of -adwy : molediw 169 v (4).

vi. Suff. -ad in cnoydrad ' wandering '. -aid in honnaid, as bit honneit 222 x (2), vb. honnaf ' I publish.'

vii. The suffixless v.a. is prob. originally the pres. part. act. in

• -a-nts, *-e-nts : Lat. -ans, -ens ; thus llosg ' burning '< Brit. *losk-ans. The suff. -awt is the past part. pass, in -to- ofa-stems ; thus *-a-to-s

> -awt. Cf. barvawt B.T. 2 1 ' bearded ' < Lat. barbatus ; edig is an extension in *-l-ko- of this ; thus, *-a-tiko-s > -edig. aid 153 (2).

-adwy is the fut. pass. part, in *-teuio- ; the -a- before it may be from unacc. a or from ; thus ofnadwy < *obna-touws 76 v (4) : Skr. kar-tavyct-h ' faciendus ' ; ediw is a variant of this due to inetath. 78 iv (i). The rhyme (glyw) B.P. 1041 shows that the -iw is for -yw after the dental 77 v. The element *-teu- seems to have been used also with suff. -no- to form verbal adjs., which mostly became nouns in W. : credadun ' believing ' D.G. 54, ' credible ' M.A. i 5636, usually 'believer ' < Brit. *kreta-tou-no- ; amheuthun ' unac- customed fare, treat', adj. 'unwonted', for *am-(w)heithun 77viii <*inbi-suek-tou-no- : chweith ' taste ' 108 iv (a), pref. 'different' 156 i (4) (d), hence ' change of diet' ; yspardun s.G. 2 ' spur' : yspar 96 iv (i). Participles in -to- formed from cons, stems sur- vive as nouns: talctith 'frontlet, crown', Ml. W. talheith B.B. 106 <*tdlo-8ekta ' forehead-attachment ', V seg*- : Skr. sdjati ' attaches ', Lith. segu ' I fasten, attach ', saktls ' buckle '.

COMPOUND VERBS.

207. i. The prefixes compounded with verbs are mostly the same as those compounded with nouns and adjectives ; see 156.

ii. A verb may also be compounded with a noun or adj.; as efe a Iwyr-lanhd ei lawr dyrnu Matt, iii 1 2 ; see 45 iii. The verb forms the second element, and has its initial softened, like the second element of a noun-compound. The initial of the first element becomes that of the compound verb, and is softened after the rel. a, remains rad. after the rel. y, etc., like the initial of a verb.

PREPOSITIONS

208. i. Personal pronouns forming objects of prepositions in Brit, and Goidelic came to be agglutinated to the prepositions, and ultimately developed into mere inflexions. The " conjuga- tion " so formed was very similar in W. to that of the verb, and was influenced in its later development by verbal forms.

ii. (i) Inflected prepositions have two forms, m. and f. of the 3rd sg., and one for each of the other persons sg. and pi. Many have in addition an adverbial form ; and all preserve their uninfleeted forms, which are used when the object is other than a personal pronoun.

(2) The simple form of every conjugable prep, causes the soft mutation of the initial of a following noun ; except er, rhag and rhwng which cause the radical ; yn which causes" the nasal, rad. or soft, according to its function ; and uwck, is which cause the rad. except in mochldw, isldw, uwchben.

iii. (i) The 2nd sg. ending is Ml. W. -t = -d, and often appears as -d in Early Mn. W., e.g. arnad D.G. 2, gennyd/byd do. 3, atad do. 42, Jwbod do. 513; but like the -d of the 2nd sg. impf. ind. it became -t in Late Mn. W., see 174 i.

(2) The 3rd pi. ending is in Early Ml. W. rarely -u, -itS, later and usually -unt formed after the 3rd pi. of verbs. In Late Mn. MSS. and books this is misspelt -ynt, with rare exceptions, e. g. G.J. Hyff'. Gynnwys (1749). The final -t is frequently dropped in poetry, as in verbs; and in the spoken lang. is always dropped; see 106 iii (a).

iv. There are three conjugations of prepositions, distinguished by the vowel of the ist and and pers. endings ; thus for the 1st sg., i. -af, ii. -of, iii. -yf.

. 209. First Conjugation. i. To this belong ar 'upon' ; at ' to ' ; o dan ' under ' ; o ' from ', with the stem ohan- (ohon-}, which does not occur uninflected ; and am ' about ' with the stem amdan-,

ii. (i) ar ' upon ' may be taken as an example of the conjuga- tion :

Ml. W. Mn. W.

sg. pi.

sg. pi.

i. arnaf i. arn-am,-ann

i. arnaf i. drnom

3. arnat 2. arnawch

2. drnad, -t 2. drnoch

3. m. arnaw

(arnabu)

3. m. drno (drn-unt

Iarnei

arnabiib

f (arnai \ -ynt

erni 3- -

arnabunt

\ drni

arnu

.arnunt

adv. drnodd

The 3rd sg. f. -ai occurs in Early Mn. verse, though rarely : arnai D.G. 85, attai do. 195, danai do. 210, ohonai I.G. 390. The 3rd pi. in -addunt survived in poetry in onadditnt, see vi ; in other cases it is rare in Mn. "W. : arnaddynt Neh. ix I. The adv. form occurs in oddi danawdd D.G. 306, oddi arnodd and odd I tanodd Job xviii 16 ; danodd ' underneath ' is common in Late Mn. W. and the dialects. No other prep, of the 1st conj. has an adv. form. Ml. 1st pi. in -ann : amdanan(n), attann R.P. 1176, arnan(n) do. 1177, atann w. 1216, o honan ni C.M. 13.

(2) Forms of ar : arnaf W.M. 2, arnat ib., arnaw ib., arnei do. 9, erni B.B. 43, arnatn W.M. 29^ arnawch ib., amaSut (~t = -8) M.A. i 258, amaSunt W.M. 470, S.G. 89, arnu M.A. i 403, arnun do. 223, arnunt

W.M. II, 39.

iii. at ' to ' (i. e. motion ' to ') is similar (but without 3rd sg. f. -* in Ml. W.) :

attafwM. 10, attat R.P. 1357, attaw W.M. 2, attei do. 6, attain do. 441, attunt do. 39, attaBunt p 21/29 ; 2nd pi. -och in attach S.G. 52.

iv. Mn. W. o dan, dan, tan ' under ', Early Mn. dan 51 vi, cf. D.G. 373 ; Ml. W. a dan,y dan, dan, O. W. guotan,gutan ox.

adan, ydan, dan W.M. 91, E.M. 66, dan W.M. i, 463 ; adanaw do. 94, ydanaw B.M. 68, ydanam do. 165, adanunt W.M. 67.

v. am ' around, about, concerning ' ; stem amdan-, ymdan-, Mn. amdan-. After gwiscaw the stem is generally used in Ml. W. instead of the simple form ; thus, instead of gwiscaw ymdanaw ac am y varch W.M. 165, the usual phrase is gwiscaw ymdanaw ac ymdan^ varch do. 162 'to accoutre himself and his horse ' ; so W.M. 147, K.M. 217, 229, 231 ; amdan W.M. 99. This form is still in use in the spoken lang., pronounced amdan.

amdanafvfM. 21, amdanat ib., amdanaw do. 2, amdanei do. 5, 13, 15, ymdeni R.M. 120, im-deni B.B. 43-4, ymdana8u8 M.A. i 197, amdana8unt IL.A. 40, amdanunt S.G. 43, ymdanunt do. 84 ; O.W. amtanndi B.S.CH. 2.

The compound yam has two meanings, i. 'besides, in addition to* W.M. 469. 2. = Mn. W. oddiam 'from about, off' : yamdanaw W.M. 5, 24.

vi. o 'from, of,' Ml. W. o (oc in oc eu, see vii (5)) ; stem ohon-, ohan-, Mn. W. oJi6n-. The forms of the 1st and and pers. fluctuate between this and the and conjugation, and the 2nd sg. has the ending -awt not added to any other prep. In Mn. lit. W. -of, -ot are used exclusively (but dial. -(/), -at). The 3rd sg. and pi. have 1st conj. forms only ; 3rd pi. onabunt. Aona/w.M. 35, K.M. 23, S.G. 22, ohonof W.M. 104; ohonat R.M, 7, 8, s.o. 85, o fawo< W.M. 12, 169, ahonot do. 10, o honaud B.B. 86, ohonaurt B.T. 53, W.M. 159, C.M. 53 ; o honaw W.M. i, 12 ; o fawez K.M. 2, s.G. i, 89, o Aewez s.G. 12, o fora W.M. 3 ; o Jwnawch W.M. 7, ohonawch do. 13; ona&unt B.M. 145, 151, ohonunt W.M. 22, K.M. 151, K.B.B. 48, ohonun B.P. 1280, Mn. W. otwnun L.G.C. 318, onaddunt do. 462, onaddun Gr.O. 94.

In the 1 6th cent, o^ow- was often contracted to on- or hon-, as cyu adnabod dim honi G.R. [xiv] ' before knowing anything of it ' ; cf. E.P., PS. cv 1 6 ; onynt M.K. [59], ono-fo do. [60] ; later Os ymddifad onot it "Wms. 438 (printed ohonot, but the metre allows only 2 syll.) ' if destitute of thee '. Analogy has restored the full form, and the contraction survives only in mono/, monot, mono etc. for ddim ohonof, etc. ; thus ni welais mono for ni welais ddim ohono ' I have not seen anything of him', i. e. I have not seen him ; 170 v (2).

vii. (1) W. ar is for *war, O.W. guar CP., Bret, war, Corn, war < *uor for Kelt. *uer< Ar. *uper 65 v (3). The personal forms are made from an adverb *uor-na; for the suffix, cf. Lat. superne 220 i (3), and for the ending, supra. The pronoun stood in a case not affecting a in the sg. or pi., hence prob. ace. ; thus sg. i . amaf< *u6rname<*u6rna me; pi. i. arnam or arnann<*arnanni<* i u6rnansme<*uomd nsme (: Av. ahma, Gk. Lesb. a/t/ic 'us'); sg. 2. arnat<*uornd te; pi. 2. arnawch formed on the analogy of the verb ; sg. 3. m. arnaw is prob. a re-formation after the 3rd sg. -8aw 210 x (i); sg. 3. fern. ei-ni<*uornaslm<*uorna sim; arnei<*uomatnm 75 i (2) ; *&im is the ace. of *sl ' she '. The most probable explanation of the -8- in the 3rd pi., which also occurs in the 3rd sg. of other conjugations, is that it is the prep. *do ; this took the dative, orig. instr. ; the instr. pi. of the pron. *es was *eibhis (: Skr. instr. pi. ebhih) as in Ir. doib 'to them '<*<&> eibhis ; this would give *duv in W. ; v after u disappeared early, but if altered to 8 (8 . . . v > 8 . . .8) would remain longer ; hence W. arna8u(8)< *uornadoibi8< *uornd do eibhis or some such form ; arnaSunt has the -nt of the verb added; arnu, arnunt are probably later formations. The modern equivalent i of the prep. *do performs the same function as that assumed above for *do ; it is added to an adverb to make it a prep. ; thus tu yma i 'this side of 216 ii (4), heibio i 'past' 210 iii.

(2) at is the stem of the personal forms substituted for *ad, which may be from *ato<*ad-do, a compound of *ad and *do both denoting ' to '. The personal forms seem to be derived from an adverb *ato-ta ; thus ataf<*ad-daf<*dto-ta-mc; etc. as in (i).

(3) o din (adan, O.W. guotari) is formed from *uo- ' under ' 156 i (16) (o-/a-<*uo-/ua- 65 v (i)) and *-tana < *-t e na as in Lat. pro-tinus<*pro-t e nos : Lat. temis, V ten- 'stretch'; *uo-tand~me >o danaf, etc., as arnaf above; adv. o dano8<*uo-tana-dt (suff.

• -dhi or *-dhc 162 vi (2)). On the accent of oddn see 47 i ; oddn

>ddn; see also 51 vi. (4) am<*inbhi 156 i (4); amddn- formed like oddn- above.

(5) o ' from, of, Corn, a, Bret, a < Brit. *a < Ar. *apo : Gk. aTTo, Lat. ab, dbs, Skr. dpa. The Bret, and Corn, a imply Brit, unacc. *a; the form a survived in Ml. W. in phrases of the form truan a beth lit. ' a wretched of a thing ', in which the ordinary o is substituted for it in Mn. W. ; see 71 i (2). Ml. W. oc, occurring only before eu, is due to the analogy of ac 160 iv (2), 213 iii (i) ; in Bret. Van. the analogy is carried a little further, Loth Voc. 28.—Bret, ac'h 'from' (ac'h Alre ounn ginidik 'je suis natif d'Auray', Troude), W. ach 'off, by' as in ach y law W.M. 472 'beside him'< *aks < *aps : Lat. dbs, Gk. jty. W. dhan- for ahan-, Corn, ahdn-, Bret, dhan-, ac'han-, from *ap-sand 'away from'; *sand<*s 6 nd : Lat. sine, Skr. sanitur 156 ii (3). The 2nd sg. ohonawt implies an accent on the -a : *ap-sana-te. The 3rd pi. ona8u (Bret, aneze, Corn. anedJie) seems to imply a simpler form, not a contraction of *ohona8- (of which there is no trace in W., Corn, or Bret.), possibly *po-na : cf. O.H.G. fona<*pu-na. The a- was generalized in Bret., Corn., and the o- in "W., where it intruded into the second syll., and even the third, causing a confusion of conjugations.

210. Second Conjugation. i. To this belong rJiag ' before ' ; heb ' (past) by, without ' ; yu ' in ' ; trwy ' through ' ; tros ' over ' ; er, Ml. W. yr ' for ' ; rJiwng ' between ' ; uwcTi ' above ' ; is ' below '.

ii. rhag ' before ', Ml. W. rac, is conjugated as follows :

Ml. W. Mn. W.

sg. pi.

1. raff of i. ragom

2. raff of, 2 ragoch

q. m. racbaw f racbu Q j

f. racbi ( racbunt

adv. racko.raccw

sg. pi.

1. rMffof I. rhagom,

2. rhdgod, -t 2. rhdgocli

3. m. rhdgddo irhdgddunt f. rhdgddi \ -ddynt

adv. rhdco, dew ' yonder '

ragof W.M. 4, ragot ib., racSaw do. 9, rac&i S.G. 63, recSi w.M. 423, rogSaw do. 444, roc8i A.L. i 452, 516, 522, ragom B.B. 29, ragoch R.M. 1 29, rac8u W.M. 53, K.M. 37, C.M. 37, IL.A. in, racSunt W.M. 86 ; racko W.M. 251, racco E.M. 8, raccw A.L. i 112 (MS.C. i3th cent.), Mn. W. rakw P 54/269 B., rhaco L.G.C. 32, usually acw do. 83 ; forms with -o- : Gwentian rhog (accented) H.G. 3, 70, rog8o I.F. P 83/66 ; S.AV. dial. oco. O.W. sg. 3. m. racdam JUV. gl. sibi.

iii. heb ' (past) by ; without ' : sg. i hebof, 3. m. hebbaw, Mn. hebddo, f. hebbi ; pi. i. hebom, $. hebbunt ', adv. heibyaw, heibaw, Mn. heibio c past '.

A cher&et heb gorr s.G. 257 'and walked past a dwarf, hebof, hebot B.P. 1440, see 198 ii, hebSaw ef W.M. 17 ' past him ', hepSaw ef do. 417 'without him', hebbi B.G. 1117; heibyaw B.M. 10, heibaw W.M. 15 ; Mu. W. hebod D.G. 513, hebom i Cor iv 8 (1620), heboch Ithuf. xv 28 (1620, changed in late editions to heibioch !). ' Past ' as a prep, is in the late period generally expressed by heibio i Job ix ii.

iv. yn ' in ' ym, yng 107, 'n, 'm, 'ng 44 vii (i) : sg. i yuqf, 3. m.yndaw, i.yndi, Mn. W. yndo, yndi, re-formed later as ynddo, ynddi.

ynof IL.A. 80, ynot W.M. 29, E.M. 19, ynoch s.G. 94 ; Mu. W. ynof G.K. [127], D. 70, ynot (3 times) Dat. xviii 22. The form indi B.B. 45, as opposed to hebti ( = JwbSi) do. 44, suggests d (yndi) rather than 8. So in Early Mn. W., as unAyn/ynd&w H.D. p 99/474, randir/yndi L.G.C. r. 34; but L.G.C. 231 has ynddo/ Wend&ydd (1 read Wendydd ; the dial, forms are S.W. yndo, N.W. ynnofo).

Llundain, ni chair lie yndi ; Liu Owain Mn ai lleinw hi. G.T., IL 134/167 K.

• London there is no room in it ; the host of old Owain fills it.'

v. trwy ' through ' : sg. 3 trwybaw, trivybi, pi. 3 irwybunt. The stem of the ist and and pers. is trw-, or trwy'b- taken from the 3rd. Adv. trwob (trwab). In Mn. lit. W. the forms are trwof) trwot, trwyddo, trwyddi, trworn, etc. ; adv. trwodd (dial. trwab).

drwyftqfi S.G. 9, 12, drwybot IL.A. 49, drwot do. 99, drwySaw, drwy&i W.M. in, trwyftaw B.P. 1418, drwySunt IL.A. 171, drwoB W.M. 51-2, B.M. 36, S.G. 68, drwa8 B.M. 36. Mn. W. trwof-i 2 Tim. iv 17, trwot Philem. 7, trwodd Mic. v 8.

vi. tros ' over ' : sg. i. trossof, Mn. "W. trosof, 3. trostaw, trosti, Mn. W. trosto, trosti; adv. Mn. W. drosodd.

drossof W.M. 88, drossot do. 25, drostaw ib., drosti A.L. i 536, drossom IL.A. 155, trostut (-t = 8) M.A. i 258, drostunt IL.A. 49. Mn. W. drosof-i Matt, xvii 27, trosom Eph. v 2, trosodd Matt, ix i ; etc.

vii. er ' for ', Ml. W. yr : Mn. W. erof, erot, erbo, etc. ; Ml. sg. I. yrof, erof, 3. yrbaw, yrbi ; etc. No adv.

2/ro/B.p. 1264, yrof, yrot W.M. 9 'for me ', 'for thee ', yr&aw do. 37 'for him', yrom B.P. 1294 (/dreith), yrSunt B.M. 49. Mu. W. erof-i Ps. cix 21, erom Rhuf. xvi 6, erddo Col. i 16, etc.

viii. (i) rhwng ' between ', Ml. "W. hong, yrwng : Ml. W. sg. I. |

3. f. yrydi, (y}ryngthi, (y)ryngti; pi. i. yfom, 2. yroch, $. yfybimt, (y)ryngthunt t (y)ryngtunt\ Mn. W. sg. I. rkof, rhyngof, 2. rhot, rhyngot, 3. rkyngtho, -i; pi. i. rhom, rhyngom, 2,. rhoch, rhyngoch, 3. rhyngthunt. In Late Mn. W. rhof, rhot, etc. are no longer used ; and -dd- is substituted for -th- in the 3rd pers., an artificial change, for in the spoken lang. the dental is still -th- or -t- as in Ml. W.

(2) yrof i a duw W.M. 2, 9, 10, etc. 'between me and God ', erofa duw do. 88, rofi a duw do. 18, yrynghot W.M. 109, y rom ni do. 10 'between us', y roch C.M. 41 'between you', yrySunt W.M. 64, c.M. 30, 41, ryngthaw W.M. 22, ryngtaw do. 19, yrygthi do. 176, y ryng- thunt, do. 6, 35, /f ungthunt do. 6. The forms yroto" IL.A. 119, yrwng do. 75, 1 20, yrygtunt 75 with the form r as in yrobes ( = yrhoes) do. 1 20, yrei (E^ rto') do. 75, show that the r is rh, as if initial (medial r after y is written 2).

(3) Strachan, Intr. 39, refers yrof to a simple ro, which is imaginary. Mn. W. rho in rAo Duw D.G. 227 is a contraction of rho a < rhof a. Zeuss confused yrof ' for me ' with yrof ' between me', ZE. 670; but the accentuation is different : Mn. W. erof 'for me ', Ml. W. yrof, yrom vii (accentuation attested by cynghanedd), but Mn. W. rhof ' between me ' Ml. W. rof see above, Mn. W. y rhom D.G. 201, rhom ' between us', as

Amodau, rhwymau oedd rhom,

Eithr dngau a aeth rhyngom. T.A., c. ii 79.

• Between us were covenants [and] bonds, but death went between

us.'

(4) The compound cyfrwng is similarly used : kywrug brodorioti B.B. 55 'between brothers'; kyfryngoch M.A. i 222, kyfryngthut (-< = -8) do. 233. cyfrwng is also a noun meaning 'interval' B.B.B. ii, In Recent written W. a neologism cydrhwng (ciid-rhwng) is sometimes used.

(5) Without initial yr- we find 1st and 2nd sg. forms used as adverbs : yngo D.G. 52, yngod do. 88, 280, G. 142 'hard by ', Ml. "W. yghot W.M. 1 1 8, yngot S.G. 304 ; cf. iso, isod.

ix. uwch ' above ', is ' below ', Ml. W. uch t is : Ml. W. sg. 3. m. uchtaw, istaw W.M. 455 ' above him ', ' beneath him', pi. J. uchom B.B. 29 'above us'. The ist and 2nd sg. are used as adverbs: uchof A.L. i 50, p 14/38 K. 'above', uchot IL.A. 115 ' above', Mn. W. ucho G. 234, uchod 'above', iso, isod 'below'.

In Late Mn. "W. uchod 'above', isod 'below' are used, but no other inflected forms. For uchof, isof periphrastic forms are used, such as uwch fy mhen, is fy nJiraed, or is fy Haw. x. (1) The -o- of the ist and 2nd pers. endings of this conjugation prob. represents the ending -o of the prep, in Brit. Although the thematic vowel -o was not a case ending in Ar. it was a common ending of adverbs and preps., e. g. *apo, *upo, *pro, and may have spread in Kelt. Hence perhaps *proko me > Brit. *rokome > W. (*rhogof), rhagof. For the 3rd pi. -Bunt see 209 vii(i); 3rd sg. -8i< *'-do-8i 75 ii (2), where ** is the instr. sg. of *sl 'she'. The 3rd sg. m. -Saw is difficult ; Ml. W. -8aw, and Bret, -zan, O.W. -dam ( = -Bav) seem to be two different reductions of *-8auv, in which au (aw) must be from *-ou- not from -a- (since dm > Bret, eun) ; botli *-8aw/and Corn, -iho may be from *-do-emi 75 ii (2); *e-mi, instr. in *-mi of *es ' he '.

(2) rhag < *prokos : Lat. reci-procus, procul < *procolos, dim. of *prokos; unacc. ok > ak in Brit. 65 ii (i) ; dialectal rhog, rhogBo, etc. < *prok-. The form before a noun seems to have ended in -ft causing the rad. initial : Brit. *rokos unacc. ; personal forms as above (i). The adv. raccw, racco ' yonder ' has a suffix *-hw or *-ho, prob. with loss of -nn (as yma ' here ' etc. 110 v (2)), for *-hwnn or *honn<*som-de or *som-da 'there' suff. *-d?ie or *-dha 162 vi (2), cf. hwnt 220 ii (5) ; thus Brit. *roko sonde ' in front there '>raccw. Initial r- was prob. first lost after consonants : y gwr racw > y gwr acw, y bryn racw > y bryn acw, *dracw > dacw, etc.

(3) heb ' (past) by, without ', Ir. sech id., Bret., Corn, heb ' without ' : Av. Jtaca ' away from ', O.Pers. hacd id. : Lat. secus, Skr. sdca ' at, in the presence of. These are believed to be all from Vseq*- 'follow ' ; e.g. Brugmann 2 II ii 894 ff. The development of the meaning in Kelt, and Iran, is not quite clear. In "W. heb with the vb. wyf means ' not having attained ' : yr wyf heb fy nghinio ' I have not had my dinner', perhaps <'*! am in pursuit of. This may explain the sense of ' lacking '. ' Past ' and ' away from ' may be from ' proceed- ing '. The adv. heibiaw (Bret, ebiou) seems to be a cpv. of the adj. *seq*os ; it might represent a loc. *seq*iosi 75 ii (2).

(4) yn l in'<*en, *eni and *en-do : Lat. in, O.Lat. en, Gk. ev, evi', etc. Although the last ends in -o, ynof, ynot, etc., cannot come directly from it, as they have only one -n- in lit. W. Ml. and Mn. They are prob. re-formations from yn on the analogy of rhagof, etc. The -d- in yndaw, etc. is due to provection of 8 after n, 11 1 vii (2).

(5) trwy ' through ', Ir. tri, ire, Bret, tre, dre. It causes lenition in W., Bret, and Ir., except in Ir. before the article. For the form in the last case Brugmann 2 II ii 900 gives *tres, comparing *pres in Gk. 7rpeo-/3i;s ; but as *pri, *prei existed beside *pres, so there were prob. *tri, *trei; these would account for the leniting forms. W. trwy < *trei ; trwy8u(nt) < *trei do eibhis. The ist and 2nd pers. forms and the adv. are analogical formations.

(6) tros 'over, across' is a weak form of traws 71 i (2), as in ar draws ' across '. It comes from a participial form *trans = Lat. trans < *trants. The 3rd pers.' trostaw, trosti, trostunK* trans do- ; the other persons and the adv. are analogical formations.

(7) er, Ml. W. yr, er. The meanings are ' for the sake of ; in spite of; in exchange for; since (a particular date)'; er ys, er's 214 vii. The prep, represents more than one derivative of *per, prob. *per, *peri, *pero-s : Lat. per, Gk. -rrepi, Skr. pdri, Skr. pardh. The prep, takes the rad. ; this would be the initial after *peros. The personal forms have the meaning of 'for the sake of, and may come from *pero-.

(8) rhwng ' between ' has replaced O.W. ithr ' between ' ; Ml. W. yrwng<*per-ongo-, 65 iii (i) : cyf-wng 'interval', cyf-yng 'con- fined ', e-ang ' wide ' < *eks-ang- : Gk. ayx<-, dy^ou ' near ', Lat. angustus, Germ, eng, Vagh-/ogh- 'narrow, strait'; the o- is seen in Gaul. Octo-durus " arx in angustia sita ". *(p]er-ongo-me by the usual loss of the second syll. would give *ymof] the seems to have been dropped, as before r, giving yrof, which kept its O.W. accentua- tion 47 i, like yrwng ; if so, the o in yr6m was originally short, and yron(n) occurs for it in B.B. 101 1. 2; the n seems to have been metathesized in O.W. igridu B.S.CH. 2 ' between them ' for *yrwyu < *(p)er-ongo-doibis. The forms yrynghof, yryngthaw etc. are probably new formations from yrhwng, perhaps originally ryiohof for

• yrnof. The curious 2nd sg. gryghod w. 36 seems to be a scribal

error for rynghod (M.A. i 192).

(9) uwch, is, see 148 i (14), (10). The 3rd pers. forms may be old, the adj. being used adverbially before do ; the other forms are prob. analogical.

211. Third Conjugation. i. To this belong gan 'with, by ' and wrth ' over against ' :

ii. (i) gan is conjugated as follows :

Ml. W.

sg. pi.

1. gennyf i. gennym

2. gennyt 2. gennwch

3. m. ganthaw,

-taw

f. genthi, -ti

.3-i

-unt -unt

Mn. W.

sg. pi.

1. gennyj- i. gennym

2. gennyt 2. gennych

3. m. gantho, (ganthunt,

-to 3 ' 1 -z!wwi5 f. genthi, -ti

Also in Ml. W. genhyf, etc. In Late Mn. W. sg. 3. m. ganddo, f. ganddi, pi. 3. ganddynt ; the dd is artificial.

(2) O.W. cant ox. ' with ', Early Ml. W. kan, as kan canyat e penteulu A.L. i 14 'with the permission of the chief of the house- hold '. The rad. is sometimes retained in Early Mn. verse : cennyd D.G. 329, cennym T.A. G. 252 (misspelt cenyd, cenyni). Ml. W. y gan ( from with ', as ugeynt ykan pop gur A.L. i 14 '20 from every man ', became gan already in Ml. W., as attep ny chavas ef genfhi hi W.M. 10 'he got no reply from her' ; Mn. W. gan ' from '.

(3) genhiw ( = genhyf) B.B. 101 ' with me', genhyf W.M. 55, gennyf do. 1 8, genhid B.B. 10, gennyt W.M. n, ganthaw do. 9, genthi do. 10, 15, genti do. 28, y gennym do. 12 'from us', genhwch do. 57 'with you', gennwch ib., B.M. 40, S.G. 92, ganthut (-t = -S) M.A. i 258, gantu W.M. 57, ganthunt B.T. 65, W.M. 16, s.G. i, gantunt HJ.A. 69.

There is also in Early Ml. verse what appears to be a 3rd pi. genhyn B.T. 13 (twice), 15, 1 6 (twice), 17, 77, nyd ant y kenhin B.B. 49 ' they will not go back ' ; also a form y genhyS B.T. 75, in an obscure passage.

(4) gan with the verb ' to be ' expresses ' have ' : y mae gennyf ' there is with me ', i.e. I have, y mae gennyt ' thou hast', etc. ; nid oes gennyf f I have not ', etc.

(5) On gennif, gennit, wrthif, ivrthit, see 77 iv.

iii. (i) wrth: sg i. wrthyf W.M. 10; 2. wrthyt ib. ; 3. m. icrthaw do. 2 ; f. wrthi do. 10 ; pi. i. w-rihym IL.A. 155 ; 2. wrthywch W.M. 39; 3. wrthu IL.A. 113, wrtkunt do. 119.

(2) O.W. gurt paup ox. ' against everybody ', gl. consistes. gurthdo Juv., gl. obstitit, seems to be a verb, 193 v (3).

Ml. W. y wrtk ' from beside ', as yicrthywch ac yrvrth ych tei IL.A. 157 ' from you and from your houses', Mn. W. ocldi wrth, oddi wrthyf y etc. ; Ml. W. y wrth also means ' compared with ', W.M. n, Mn. W. wrth B.CW. 5 ' compared with'.

iv. (i) gan, O.W. cant (Corn, gans, Bret, gant, Ir. prefix oet-, ceta-) has the meanings of *kom, of which it is a derivative. Thus cann < cant < Brit. *kanta < Ar. *km-ta = Gk. Kara < *km-ta. The pronoun suffixed affected the a ; it may have been abl. *ml(d) = Lat. me(d), or possibly a loc. *moi which as a mere suffix would become *-im. So for the 2nd sg. The first and 2nd pi. are prob. analogical. The affection of a before a labial became ^, 69 ii (4) ; hence gennyf, gennym, gennywch, which caused the sg. 2. to follow; thus the distinction kept in Corn, between genef and worthy/ (similarly in Bret.) is lost in W. Between vowels -nt- > -nnh- > -nn- regularly. In the 3rd pers. *do is used, as after other prepositions, taking of course the same case ; hence *cant-8aw > ganthaw or gantaw, so for f. and pi. The 3rd pi. genhyn, with verbal -n (add.d to the apparent stem genhy- ?).

(2) wrth, O.W. gurt(h), Corn, worth, arth, Bret, ouz, oz, implies some such form as *uerto 66 iii (i); Ir. frith 'against' < *urt- : Lat. red-, re- < Ital. *ured- < Ar, *uret- ; cf. Lat. versus ' against ', part, of verto : Vuer- 'turn'. wrthyf< *uerto-mi, etc. The 3rd sg. wrthaw < *wrth-$aw, the -8- merged in the -th-} so for the f. and pi.

212. i. The prep, i 'to', Ml. W. y, O.W. di is inflected anomalously ; the 1st sg. is qm, or with the affixed pron. qmi or ym?/ ; for y-, i- is common in Ml. W., and became the usual form in Mn. W. The inflexion is as follows :

pi. I. iinn, qnni, ^nnif, in, inni 1. qm, ymi, ?/mq, im, imi 2. ytt, iftti ytty, itt, itti 3. m. ibaw, Mn. iddo f. ibi, Mn. idcli 2. qwch, qchwi, ychwif, iwch, ichwi 3. ubu, nbub, ubunt, Late Mn. iddynt

Examples: ym IL.A. 98, 1. 4, ymi W.M. 20, 22, im do. 46 ; ytt IL.A. 95, ytti B.M. 5, yti W.M. 4, itt do. 3, 8, 9, it do. 20, itti ib. ; yni do. 29, inni do. 139, ychwi B.M. 7, tw?c^ c^toi W.M. n, ywch do. 50, w<w< ( = w8w8) B.B. 49-50, w8w B.T. 74; u8unt 77 viii; t'Saw, z'8i passim. The Late Mn. W. spelling iddynt is artificial; see 77 viii.

ii. Forms with y survive in Early Mn. verse, in which the rhymes show that the sound of the y is if.

Dafydd ap Gwili/m, nmti

Y bu fraw am na bai fn(. G.Gr. (m. D.G.) F.N. i.

' Dafydd ap Gwilym to me there was dismay because he was no longer [alive] there.'

Arglwydd gwtfnn, nid oes qnni

Un tad oil onid tydi. M.R., p 93/56.

' Holy Lord, there is to us no father at all but Thee.'

Ni all angel penfelyn

Na llu o saint ddim lies qn. G.I.1L., F. 8, M 130/4708.

' No golden-haired angel or host of saints can [do] us any good.'

iii. The affixed pron. is often accented ; in that case it is usually written separately, i mi, i (i, etc., Ml. W. y mi W.M. 8. As inni has undoubtedly a double n the form yni W.M. ao must mean yni (the double consonant being simplified before the accent 27 ii).

Gwell i ml golli 'mflwyd

Na chan boen nychu 'n y byd. T.A., A 14866/201.

' It is better for me to lose my life than in pain to pine in the world.' Earely in poetry i myfi D.G. 53, i nyni H.S. 22, etc.; thus: Mwya ofn yw i myfi

Ofn y paid ar f'anap i. I.D., A 14997/28.

' The greatest fear to me is the fear that my trouble will end.'

iv. When dydd da 'good day ' and nos da ' good night ' are followed by yd ' to thee ' and ywc/t ' to you ', da yd is contracted to daqd written daed ( 29 ii (2) ) and da ywch to AtfKwi, daewch, now reduced to dawch 30.

"Nos daewch" i'r ferch nis dichon-,

" Nos daed ti " nis dywaid hon. D.E., A 14967/18 (o. 1 18).

' [To say] " good night to you " to the maid avails not ; she will not say " good night to thee ".'

Breiniawl wyt o'r banvnwaed;

Barwn Ystepltwn, nos daed. L.G.C. 141; see 127, 480.

' Noble art thou of the blood of barons ; Baron of Stepleton, good night to thee.'

O'r cyff hwn a'r Cyffinwaed

Y caivn was dewr. Can' nos daed ! T.A., A 14975/102.

' Of this stock and the blood of Kyffin we have a brave youth. A hundred good nights to thee.'

Dydd daed D.E. p 83/103, dyddiau daed G.G1., M 146/203, D.G. 381.

Nos dawch is still in common use ; but daed is not now generally known. Silvan Evans quotes L.G.C. 141 (see above) and D.G. 381 under daed eqtv. of da,

v. oe 'to his, to her, to their' 160 ii (T); yw, i'w ' to his, to her, to their ' ib., 160 iv (2).

vi. y, i 'to ', O.W. di < Brit. *do is equivalent to the prefix dy- 65 iv (2), 156 i (13). It is strange that this prep, whose ist sg. is the only one in Ir. which has certainly a single -in (= W. *-/) is the only one in W. with -m ( = -mm). The -m is due, like the usual -mm in Ir., to the Kelt, doubling of the initial of an unacc. word following an accented monosyll., 217 iv(i); thus ym < *do~mml < *do moi. The corresponding form of the 2nd sg. would be *yth, cf. yth ' to thy ' ; but the form that survived was yd (id W.M. p. 279), as in daed ; by late analogical doubling this gave ytt (d-d > tt 111 ii (i)). It may be conjectured that the 3rd pers. forms were orig.

• daw, *di, *du ', as these were mere suffixes in the conjugation of

other preps, it is probable that *&y was prefixed here to represent the prep. ; y would be assimilated to a following i or u, and perhaps Haw takes its i from i8i, O.W. didi L.L. 120. But the prep. *dl 156 i (ii) may have been prefixed, with an intensive force, as befoie *do- in di-Saior 195 i. § 213. i. The prepositions a [spirant], ag, Ml. W. a, ac e with ' and wedi [rad.], Early Mn. and Ml. W. (g)wedym&j be followed by independent pronouns ; thus d mi ' with me ', d thi ' with thee ' ; d myfi, d m'ji, d tkydi, d tKdi 159 ii (3) ; ag ef, ag efo ' with him ' ; etc. ; gueti ef L.L. 1 20 ' after him ', gwydi ny B.B. 44 ' after us '.

Ac ni bydd oherwydd hyn

Gwedy ef gwiw dy ofyn. I.G. 312 (m. D.G.).

' And therefore it will be of no avail after him to ask for thee ' (i. e. for a cywydfi).

Y Deheu feirdd wedy fo

Sydd wannach eu swydd yno. Gut.O., M 146/398.

' The bards of the South after him are weaker in their performance there' (m. G.G1.).

Da oedd cyffion Huw Conwy,

A da yw Huw wedy hwy. L.G.C. 463.

' Good were the ancestors of H.C., arid good is H. after them.'

In Late Mn. W. the use of a pron. as above after wedi is rare.

ii. (1) a is now generally circumflexed to distinguish it from a ' and '. When it is accented it is of course long, but when un- accented it is short. The same is true of a ' and '.

(2) wedi has late -i owing to the frequency of its use, cf. 16 ii (3). In Early Ml. W. where it rhymes it has -i[ ; see e. g. L.G.C. 15, 66. In Ml. W. it has -y in MSS. in which i and y are distinguished.

iii. (1) a, ag < *agg6s ; it has two distinct meanings, and may therefore have a double origin : (a) ' together with ' < *ad-g- : Ir. ac, oc, occ < *ad-go-s : Lat. ad, Goth, at, E. at, Brugmann 2 II ii 793 ; this is the prep, used after cyf- and ym-, as cyf-arfod d 'to meet with ' ; and is the same as d, ag ' as ' after the equative ; (b) ' by means of, as O.W. ha crip ox. ' with a comb', Mn. W. d phyg Gen. vi 14 'with pitch' < *ab-g- : o 209 vii (5). For ag, a [spir.] in this sense o [soft] is used in Gwent : taro ci o asgiorn ' to strike a dog with a bone ' Seren Gomer, Mai 7 1814; cf. kymynynt o 8ur B.P. 1042 for k. a dur B.B. 72 ' they hewed down with steel ' ; con- versely, after a spv. ag is used for o before a relative, as yn oreu ac y gellynt C.M. 54, gyntaf&C y gallaw8 s.o. 408,

Y glanaf ag a luniwyd,

A'r goreu oil o'r gwyr wyd. T.A., A 14971/53.

' The handsomest of [all] that have been created, and the very best of men art thou.' In Ml. W. this is o before the demonst. 'r, as goreu . . . o 'r a vu K.M. 82 ' best of those that were ' ; rarely a, as o bop . . . aV a vet IL.A. 141 ; Mn. W. a'r a. The common origin and overlapping use of a 'by means of and o 'of, from' prob. con- duced to the formation of the analogical oc 209 vii (5). The last element in Brit. *ag-gos is the same as that of the orig. form of a 'and' 222 i (3).

(2) gwedy (: Bret, goitde) 'after' is also an adv. 'afterwards' 220 ii (9). As a prep, it is largely used before verbal nouns, and in periphrastic conjugation it forms the equivalent of a perfect. For its origin see I.e. ; it has the same final element as a, ag.

iv. The above are the only prepositions which may govern personal pronouns, except mal, megys 215 iv.

214. The following prepositions are of more or less re- stricted use :

i. ach is used only in ach law ' near at hand ' ; ach fy Haw ' near me ', etc. 209 vii (5).

A trace of a wider use is seen in ym ach mur Kaer Loyw R.M. 131 ' beside the wall of Gloucester ' ; ym = am < *mbhi 156 i (4).

ii. ger [rad.] ' near ', Ml. kir, ker, gyr, ger, geir, geyr, gar, is used chiefly in gerlldw,ger Haw ' at hand ', gerbron, ger Iron l before ' (ger fy Haw ' near me ', ger dy lam di IL.A. 125 f near thee ', gerfy mron 1 before me '), but may occur before any noun denoting a place.

The radical initial is k-, as kir Haw B.B. 10, ker Uaw E.P. 1246, M.A. i 230, cer bron do. 206, ker tir Tyssilyaw do. 341, Mn. W. a cJier bron Dat. iii 5. The origin of the word is uncertain ; it seems to form the prefix in cyr-haeddaf ' I reach ', Mn. v.n. cyrr-aedd : haeSaf ' I reach ' ; possibly allied to cwrr ' edge ', V(s}qer- ' cut ' ; both -ei- and -y- may be affections of -a- or -o- before -rr-, and -e- may be a variant of y 16 iv (2) ; gar W.M. p. 281 may have unaffected a.

iii. tra is used only in drachefn ' backwards, again ' (cefn ' back ') ; with infixed pronouns kilya drathgefpn C.M. 41 ' with- draw!' and sg. impv., draechefyn K.M. 177 'behind her'. In Late Ml. and Mn. W. by a wrong division of drachefn we have drach dy gevyn S.G. 275 'behind thee', dracli 'y nghefn D.G. 274 ' behind me ', drach ei chefn Gen. xix 26.

Tra mar tra Brython B.T. 76 ' beyond the sea, beyond [the borders of] the Britons' and tra run B.B. 49 'beyond Khun' preserve the remains of a wider use.

trachefn for *tarcMfn < *taros kebn- < Ar. *t s ros : Ir. tar : Skr. tirdh < *t e rds ; allied to trwy 210 x (5) ; see 156 i (22).

iv. pw (py) is used only in the phrase pwy gilydd 'to its fellow ', as o ben bwy gilydd ' from end to its fellow ' i. e. from end to end. For examples see 166 ii (3). A trace of a wider use survives in awr py awr R.B.B. 107 ' [from] hour to hour'.

pw : Ir. co ' to '. Initial gemination after the latter is secondary, according to Thurneysen, Gr. 456, who compares O.Bulg. M 'to' (< *qom : Skr. kdm after the dative). The Kelt, form would be *q*o; this may be the pron.-stem *q*o-, seen in e-grade in *q v e 'and ' (Lat. -que, Gk. re, etc.) : Lat. us-que < *ud-s ' out ' + *q*e ' to *.

v. eithr [rad.] ' without, except ', 99 v (4), is used before verbal nouns, as eitJifr bot yn well kyweirdeb y bwyt W.M. 227 ' except that the preparation of the food was better ' ; hence it came chiefly to be used as a conjunction. But it occurs also before nouns and pronominalia : eitJifr mob C.M. 2 'beyond measure'; eithfr y rei a oefynt W.M. 227 'except the ones who

were '.

Eithr Morfudd ni'm dihudd dyn. D.G. 51. ' Except Morfudd no one will appease me.'

vi. O.W. ithr M.C. ' between ' seems to occur only once ; it was obsolete in Ml. W.

ithr, Corn, inter, yntre, Bret, entre, Ir. etar, eter : Lat. inter, Skr. antdr.

vii. ys, es [rad.], Ml. W. ys c for . . . past ' is used before a noun denoting a period of time, er ys with a past verb : yr ys pell o amser R.M. 130 ' [I came] a long time ago ', cf. IL.A. 106, 107; er ys mis W.IL. G. 293 ' for a month past ' ; contracted er’s.

Ys guers y8 wyfyn keissaw a olchei vyg cleyfv?M. 487 'for some time I have been seeking one who would burnish my sword.'

Ofnus fyth fu'r fynwes fau Es deufis hyd nos Difiau. G.G1. p 103/193. ' My heart was constantly afraid for two months till Thursday night.'

ys ' for the space of, perhaps < *en-s : Gk. ts, 215 iii (i). If oed W.M. 123 1. 2 (omitted in B.M. 197) is oe8 'was 'for yr ys P 14/185 it shows ys taken for ' is ', cf. Bret, zo, Fr. il y a ; but yr ys is old, and implies ys prep.

viii. Ml. W. annat [rad.] ' before, in preference to ' is used before neb, dim, and other expressions in which 'any' is expressed or implied. In Ml. W. yn began to be used before it ; and in Mn. W. it became yn anad,, the nn being simplified owing to the word being unaccented, cf. canys 222 iv (i) : yn anad neb. Or clywy Siaspat dos torthi, a diaspat gwreic annat diaspat o'r byt B.M. 195 'if thou hearest a cry go towards it, and a woman's cry before [any] cry in the world'; W.M. 120 has anat, but other nn's are simplified in the same col. ; heb ymgyfarvot ac ef yn annat neb s.G. 34 * without meeting him of all men ', lit. ' rather than anybody ' ; cf. s.G. 142 ; yn anad neb D.G. 35, 107 ; Mor llygredig oedd ei wedd yn anad neb, a'i bryd yn anad meibion dynion Es. lii 14. It is also used adverbially : ac yn annat llawen oe8 Arthur S.G. 10 'and A. was especially glad '.

annat < *anta-tos an adv. formed from *anta ' before ' : Goth. anda-, Gk. avra, a doublet of *anti : Gk. dvri, Lat. ante.

ix. myn [rad.] ' by ' (in oaths) ; in N. W. sounded m$nn ; Ml. W. mynn, myn.

myn Haw vyghyveillt W.M. 458 ' by the hand of my friend ' ; so B.M. 105, mynn II. etc. do. 170 ; myn vy fy8 C.M. 57 ' by my faith '; myn Duw B.M. 115, myn Ztyw W.M. 473 'by God'; myn fenaid D.E. c 49/15 B. ' by my soul ' ; myn einioes Pharaoh Gen. xlii 15.

Nid oes ym, myn Duw, & o swydd * Printed dyn.

Ond olrhain anwadalrhwydd. D.G. 33. ' I have, by God, no task but studying fickleness.' Cf. W.M.L. 41.

mynn : Gael, mionn ' oath ', Ir. mind ' oath ' : Vmendh- extension of Vmen- 'thought'; cf. W. adduned 'vow' < *ad-moni- 100 v. Macbain connects Ir. mind ' oath ' with Ir. mind ' holy relic ' and this, with less probability, with Ir. mind ' diadem ', O.W. minn gl. sertum (: Lat. monile, see Walde s.v.).

x. ym [rad.] 'by' (in oaths).

Gwell ym ym Padricf B.P. 1277 ' It is better for me, by Patrick ! ' Ym Sant Grigorf L.G.C. 183 'by Saint Gregory!'; ym Beunol O.G1. M 146/188 ; ym lesu! T.A., G. 229.

The origin of the word is obscure (? ym ' to my ').

215. Nominal Prepositions. i. Some of the above pre- positions are of substantival or adjectival origin. Others are from adjectives :

(i) cyn (cyn) [rad.] ' before ', in time : kin lleith B.B. 22 ' before death ', kin myned do. 30 ' before going ', kin brand do. 41 ' before the judgement '. In Ml. W. it is followed by no ' than ' before pronouns pers. and demonst., and thus remains an ,adv. : kyn noc ef W.M. 178 'before him', kynn no hynny do. II

• before that '. In Mn. W. it is no longer used before pers.

pronouns, and has become a prep, before demonstratives : cyn Jiynny ' before that '. It is in common use before nouns. cyn is the adv. cynt ' sooner ', an obi. case of the cpv. adj. cynt 148i( 3 ).

(2) nes [rad.] ' until ', used before verbal nouns ; as, ny chysgaf hun lonyb nes gwybot W,M. 167 'I shall not sleep comfortably until I know ' ; nes ei orfedd T.A. G. 237 ' until his lying- (low) ', i.e. * until he lay (low) ' ; rarely before abstract nouns : nes Jienaint L.G.C. 445 'till old age'.

Galw am ddyfod diodydd, Gwyliaw tan nes gweled dydd. L.Gr.C. 430. ' Calling for drinks to be brought, watching the fire till day is seen.'

The construction survives in Late Mn. W. : nes i mi ddyfod Es. xxxvi 17 'until my coming', cf. B.CW. 83, 115; but a new construction, nes before a noun-clause beginning with y, arose, e. g. nes y del y dydd c.c. 211 (end of i7th cent.) ' until the day comes '. In the dialects the y is omitted and nes becomes a con- junction ; but nes with v.n. is still in common use.

nes ( = nes, though now sounded nes 51 vi) = nes ' nearer ', 148 i (i) ; ' nearer than ' > ' this side of ' > ' until ' ; cf. nys caffaf-i efo yn nes dioSef llawer s.G. 291 'I shall not get him before suffering much '.

ii. Nominal preps, from nouns are used not only simply, as cylch 'about', but with a preceding prep., as o gylch 'about'; the latter forms may be called composite nominal prepositions. When a pers. pron. is required to be the object, it takes the form of an infixed pron. in the composite prep., as o'th gylch ' about thee' ; o'th achos D.G. 101 'on thy account, because of thee '. The simple nom. preps, are the following, all taking the [rad.] except hyd :

(i) achos 'because of; compos, o achos id. : noun acJios 'cause'.

o achaws W.M. 12 'on account of; o'th achaws di K.M. 233 ; pa achaws E.B.B. 112' why 1 ' achos gwenfun I.MSS. 239.

(a) cylch, amgylch 'about'; compos, ynghylch, o gylch, o am gylch, o amgylch 6gylch ' round about ' : cylch ' circle '.

Yn bwhwman gan annwyd

Cylch drws dy dy, Lleucu Llwyd. 1L.G., BR. ii 171.

' Shivering with cold about the door of thy house, LI. LI. ', i. e. around thy grave; (v.l. YngliylcJi dy dy F.N. 29); cylch dolydd Dwylais L.G.C. 202, gyloh y Ddol G. 91 ; ygkylch y ty W.M. 47 'about the house '.

cylch is believed to be derived from Lat. circulus ; but the latter gives O.W. circhl CP. ' cycle ' regularly ; and cylch, Bret, kelc'h, may well be Kelt. < *q*u-q*lio- (by met.) : Gk. Kv/cAos, KvxXtos : E. wheel, etc.

(3) eisiau ' wanting, without ' ; compos, o eisiau ' for want of, o'th eisiau, etc. : eisiau ' want ', prob. orig. an adj. < Lat. exiguus (noun eissyivet < exiguitas).

Mis Jiaf oedd i ferch Ddafydd,

Ac eisiau hwn gaea' sydd, T.A., o. 245.

' It was a summer month to the daughter of Dafydd, ami without him [her dead husband] it is winter.'

(4) erbyn * by ' (a certain time or event), ' in readiness for ' ; compos, yn erbyn ( against ' (a person or thing) ; yn fy erbyn Matt, xii 30 ' against me ' ; also i'm herbyn Matt, xviii 2 1 ' against me '.

Hid im pen un brin erbiu en barnu B.B. 42 ' to the summit of one hill to be judged ', lit. ' for the judging of them '.

erbyn is itself originally a composite prep. < Kelt. *ari quennoi, made up of the prep. *ari 156 i (6), and the dat. of *quennos ' head ' : Corn, erbyn, Ir. ar chiund (in Ir. there is ar chend also, with chend ace.). The orig. construction with a pron. was Corn, er dhe byn ' against thee ', Ir. ar do chiund ' in front of thee '. The improper compound erbyn was mistaken for a proper in W., whence yn erbyn etc. ; but it did not become an ordinary noun though treated as such in this construction.

(5) herwyS ' according to, in the manner of ', and ' by ' (as in lead ' by ' the hand) ; gervyb in Late Ml. W. in the last sense, Mn. W. g erf y del ; compos, o herwydd ' on account of, o'm herivydd 'on my account', o'r herwydd { on that account', yn herwydd 1 according to ', yn ol yr herwydd ' on the average ', pa herwydd ' why ? '

herwyS y dyyll [read dyall] ... a roSes Duw yr neb ae trots IL.A. 1 60 'according to the understanding that God has given to him who translated it ' ; herwit guir in gueini B.B. 44 ' in the manner of men in service ' ; herwyS y afwyneu W.M. 142 [lead the horse] ' by his reins ' ; gervyS y avwyneu C.M. 47 ' by his reins ' ; erwyS y tract W.M. 55 [grasped the boy] 'by his feet'; oherwydd hyn 213 i; o'i herwyS D.G. 498 'on her account'; yn herwyS yueledigaeth W.M. 34 ' as regards appearance ' ; (y)r iaith Gymraec yn ei herwydd M.K. [vii] ' the Welsh language in general ' ; Ba herwydd na bat hiraeth T.A. 'why should there not be longing 1 ?' O.W. heruid duiutit Juv. ' according to divinity ' ; hihi erguid ox. ' in general ' (?).

hemvyS : Corn, herwydh ; Bret, hervez ; the h- though appearing in all (as opposed to W. and Corn, erbyri) can only be explained as accentual, cf. henw, Bret, hano, Corn, hanovj ; the rest seems to be <*ari-uid- ' *appearance, manner', Vueid- 'see'; erwy8 also occurs without h- in O. and Ml. W. ; and gervyS is a variant due to another treatment of e-, see 112 ii (2).

(6) hyd [soft] 'the length of, 148 i (8), in two senses, (a) 'as far as', (b) 'along'; compos, ar hyd [rad.] 'along', ar dy hyd ' along thee ', also ' at thy full length ' (on the ground).

o lost irinis hit bronn ir alt L.L. 73 ' from Llost yr Ynys as far as the breast of the Allt ' ; o hynny hyt trannoeth W.M. 6 ' from that [time] till the morrow ' ; hyt yr amser do. 1 9 ' till the time ' ; often followed by yn 216 ii (i); ar hit taf L.L. 258 'along the Taff' ; ar-i-hit do. 159, ar-y-hit do. 143 'along it', n-i-hit do. 43, 78, etc., yn-y-hit do. 146 'along it'.

In the dialects hyd developed an inflected 3rd sg. hqd8o, hydSi (the y, instead of y, shows it to be late) ; this is sometimes met with in Late Mn. W. : ar hyd-ddi Gen. xxviii 1 2.

O.W. bikit CP V ticket ib., bekeit ox., bekit L.L. 73, behet do. 73, 122, bet Juv., L.L. freq., e.g. 146 (7 times), 155 (n times) 'as far as' ; cekit L.L. 73 'along', cikitan do. 122 bis 'along', cihitun ox. 'along' ; Ml. W. vet (misprinted ver) K.M. 144 (see W.M. 201) ' as far as ' ; Gwentian ved H.G. 23, 52 ' till '.

bi- < Ar. *bhi (: *obhi) ' on (to) ' : Goth, bi, Skr. abhi (Lat. ob may be from *obhi or *opi) ; -het may represent ace. *-sitm ; the unique form -heit may be due to Imihatn which follows it; bet is generally regarded as a contraction of behet, but such a contraction is doubtful so early ; cf. also Bret, bet, bete, beteg ; can it be an adv. direct from bi- ? cehit = eqtv. cyhyd 148 i (8) ; cihitan an adv. like guotan etc., from *ko-si-tan- ; -'un error for -an?

(7) llwrw ' in the track or direction of, after, with, as re- gards ' ; compos, yn llwrw id., ar llwrw id., adv. ' forward ' ; S.W. dial. Iwrw i ben ' head foremost ' ; also Ml. W. llwry.

llwrw essiwet ket B.P. 1351 'after dearth of largess', llwrw alaeth . . . digrawn . . . deigyr do. 1 206 ' with grief the tear flows ' ; yn llwrw llwyth elvyt dovyt a'n dyd yn llawr P.M. M.A. i 306 ' following earth's tribe the Lord will place us in the ground'; dos heb argysswrw ar llwrw yr lie do. do. 292 ' go without fear forward to the place '.

llwrw : Corn, lerch ' track ', war lerch ' after ', Bret, lerc'h ' track ', Gael, lorg ' track ', Ir. Ivrg ; all < Kelt. *lorg- < *plorg- dissim. for *pro-rg-,Vreg- 193 x (8) : Lat. pergo, perrexi < *per-reg- ; etc.

(8) parth, parthed 'towards, as regards'; compos, o bartfi, o barthret G.C. 108 'as regards' ; imparthred B.B. 26 'in the region of ; parth is oftenest followed by a 216 ii (2).

parth espyt B.P. 1226 'as regards strangers'.

parth ' part ' < Lat. part- ; parthed is by dissim. for parthrfd with -red as in gweithred 143 iii (22).

(9) plith ' in the midst of ; compos, ymhlilh ' among ', yn en plith 'among them', o With 'out of the midst of, d'ch plith 'from your midst', iWith 'into tbe midst of, iw plith 'into their midst ', plith draphlith 47 iii.

plith from Lat. ; perhaps < Brit. *plikt- for Lat. plicit- : W. plygu ' to fold ' < Lat. plic-o.

iii. Many composite nominal prepositions have no corresponding simple form (i e. the noun alone is not used as a prep.). All are followed by the [rad.]. The most important are

(1) mewn, Ml. W. y mywn, myuon * in ' (though apparently a simple form, myicn is a mere phonetic reduction of ymywn} ; o fewn ' within ' ; with inf. pron. i'w mewn hi Num. v 24 ; och meion Luc xvii 21 ; also in Mn. W. i fewn y Uys Marc xv 16 ; adv. i meicn^ oddimeicn.

(y)mywn ' in the middle of ' has come to be used for ' in ' before indefinite, yn being restricted to definite, objects ; thus ymywn ty W.M. 53 ' in a house \ yn y ty do. 54 ' in the house ' ; in Ml. and Early Mn. W. mywn, mewn is sometimes used before the latter.

i < *ens : Gk. ets < evs < *en ( in ' + -s as in *eks. i mewn, ymywn = Ir. inmedon, immedon ; Ir. medon ' middle '. The W. form has lost 8 110 iv (2), and was therefore orig. disyllabic *my\wn < *mywn, which most probably represents *my8-wyn 78 i (2). Both this and Ir. medon would be regular from Kelt. *mediokno : Lat. mediocris, spv. medioxtmus. If this equation is right, mediocris can hardly be ' *middle-hill ' (: ocris, Sommer 488, Walde s.v.) but may be an adj. in -ri- (cf. dcri-, sacri-) from *medioque formed from medio- like proj)e (for *proqiie) from pro, as the spv. medioximus beside proximu* suggests. The Kelt, would be a noun in -no- from the same (It.-Kelt.) extd. stem. Orig. stem *medh(i)io- : Skr. mddhya-h, Gk. /i(TO-S.

(2) er mwyn ' for the sake of, on account of, in order to ' ;

er fy mwyn ' for my sake ', etc,

Er dy fwyn yr ydwyffi

Mewn eira yma 'n oeri, D.G. 107.

' It is on thy account that I am shivering here in snow.'

As a noun mwyn meant ' value, enjoyment ', but except in the above phrase was generally replaced by mwynyant in Ml. W. ; thus in K.B. 963, Ni wybyBir mwyn (v.l. mwynyant io > ]6)fi/nnawn yny el yn yspi8 (8ispy8 1076) 'the value of a well will not be known until it goes dry '. As an adj. mwyn means ' gentle, kind, dear ', and is still in use ; cf. E. dear ' costly ' and ' loved '.

mwyn ' value ' < *mei-no-, Vmei- ' exchange, barter ' : Lat. munus, munia.

(3) ymysg (ymijsg} 'in the midst of; yn eu mysg 1 in their midst' ; on mysg ' out of our midst'; i'ck mysg ' into your midst' ; emysc hynny W.M. 33 'in the midst of that' i.e. those hap- penings ; ofyscr, ifysgr.

mysg : W. mysgu 96 iii (5). The idea is 'mixed up with ' ; and there seems no need for Henry's attempt, s.v. em/esk, to connect the word with *medhio-.

(4) yn wysg ' in the track of, after ', Ml. W. yn eu hwysc see below, yn wysc y benn W.M. 55 ' after his head ', i. e. head fore- most. Mn. W. yn wysg fy mhen y yn wysg dy drwyn, yn wysg i gefn, etc.

Mae yr aniveileit y8 aethawch yn eu hwysc P W.M. 86 ' Where are the animals which you went after ? '

wysg ' *track ' implies *ei..sk-, and seems like a case of metath. of t 100 v (? *ped-skio- : Gk. TrcSa '/xera', Lat. pes 'foot', etc.).

(5) yn ethryb 'because of, o ethryb id. J.D.R. [xiv].

Pellynnic vyg khof yg kyntevln

Yn ethrip caru Kaerwys vebin. G., w. 76.

' My mind is far away this Spring, ou account of loving the maid of Caerwys.'

ethryb ' causa, occasio ' D.D. s.v. seems to contain *-9q*- affected 69 ii (4) ; perhaps as a noun-sufF. added to *nter- (* enter 214 vi) ; ' circumstance ' (?).

(6) yn 61 after ', yn dy 61 ' after thee ' ; ar 61 ' after ', ar eu hoi or ar eu holau ' after them ' ; o'm hoi ' behind me ', i'th 61 ' after thee'.

All in common use. 61 149 i. (7) yng^yS 'in the presence of, yn fy ngwydd 'in my presence ', i'tk wydd ' into thy presence ', o'i gwydd ( from her presence ', etc.

gwy8 63 iv.

(8) o blegid 'on account of, o'th llegid 'on thy account'; ym plegyd M.A. i 306 ' on account of.

plegid (t for y after g, 77 ii) < Lat. placitum.

(9) ar gyfair (now misspelt ar gyfer) 'opposite', arfy ngJiyfair 1 opposite me ' ; ynghyfair ' opposite, against, instead of ; Ml. W. ar gyveir, yngkyveir, etc. ; y gyveir W.M. 449 ' the direction '.

O.W. ar cyueyr L.L. 141, ar ciueir do. 196 ; Ml. W. ar gyfeir W.M. 250; yghyveir do. 449, ygkyveir R.M. 293 'opposite', yn y gyveir R.M. 141 'in front of him, straight ahead', Mn. W. ar gyfj,ir D.G. 189 (rh. with leddf-air).

The reason for the misspelling is partly the dialectal pronunc., 6 iii ; and partly perhaps the form cyfer- in cyf-erbyn etc. All the derivatives, cyfeiriad ' direction ', cyfeirio ' to direct ', etc., are from cyfair.

cyfair < *kom-ario-, a compound of *kom- and *ario- < *p e ri-o-, a noun formed from the prep. *p e ri : Lat. prae, etc. 156 i (6).

(10) o flaen ' in front of, ymlaen id., o'm blaen ' in front of me ', dos yn dy flaen ' go in front of thee ', i.e. go on, ymlaen llaw ' beforehand '.

dyvot ymlaen llu Ynys y Kedyrn W.M. 54 ' to come in front of the host of the Isle of the Mighty'; yn ^ vlaen ac yn y ol B.M. 149 'hefore him and after him'; kernel oe blaen do. 49, W.M. 68 'to walk before them '.

blaen, O.W. blain : Corn, blyn ' tip ', Bret, blein, blin ' bout, extr6- mit^'. The meanings of the noun in W. are i. ' source' (of a river) frequent in L.L., and common later, 2. 'point' (of a needle, blade, epear, twig, etc.), 3. pi. blaenau in place-names ' outlying parts where valleys are hemmed in by mountains '. The orig. meaning seems to be therefore ' discharge, project ' ; hence prob. V g*ele- : Gk. /3oAAa> ' I throw ', /3Xr)fjLa, fioXr), ySo'Xos ' a throw ', /3oAi9 ' arrow ', O.H.G. qudlan ' to well, to gush ', O. Norse kelda ' source ', Gk. yScXov?/ ' needle ', Lith. gelti ' to prick ', gelonls ' needle ', etc. The formation is not quite clear ; the Corn, and Bret, forms seem to imply Brit. *blani- ( < *g*h-ni- : cf. Lith. gelonis) ; and the W. may represent the same with met. of t, 100 v; *ai > *oi > ae after the labial, blaenaf 149 i.

(11) heb amlaw R.M. 179 'besides, in addition to', hebldw or heb law Matt, xv 38, rarely amlaw GRE. 327 id.

Haw ' hand ' in the sense of ' side ' ; heb law ' out-side ', am law 1 be-side ' ; heb i Haw D.G. 148 ' beside her '.

(12) o ran 'on account of, e.g. W.IL. 173; o'm rhan i 'for my part', etc. ; o waith ' because of.

rhan ' share, part ', 63 vii (2). gwaith ' deed ' 193 x (4).

(13) ynghyfyl S.G. 35 * near ', ar gyfyl id. ; yn i chyfyl BE. iv 427 ' near her '. is oil ' behind ' ; is y gil R.M. 151 ' behind him '.

cyf-yl : ym-yl ' edge ' 101 iv (2). cil 59 vi.

(14) ach law 214 i ; gerllaw, ger Haw do. ii ; gerbron, ger bron ib. ; drachefn do. iii, trachefpn y bor W.M.L. 32 ' behind the door ' ; ar draws 210 x (6) ; ymrdn c.c. 34 ' on the point of, nearly ', in Late Mn. W. bron.

iv. (1) Ml. W. mal, ml, Mn. W. mal, fal, fel 'like', and Ml. W. megys, Mn. W. megys, megis ' like ', are followed by a noun, a verbal noun, or a noun-clause introduced by y. They generally stand in an oblique case, and are therefore prepositional. But sometimes they qualify nouns, as

Pan el y gwallt hir-felyn

A'i frig fal y caprig gwyn. D.G. 441.

Lit. 'When the long yellow hair goes with its tips like white cambric '.

Y ddyn fegis Gwen o'r Dd&l,

Kywiog araf ragordl. D.G. 379.

' The woman like Gwen of the Dale, gentle, patient, peerless.'

(2) fel and megis may be followed by independent pers. pronouns, as mal ef B.P. 1403 ' like him ', fel myfi, etc., or by demonstratives as fel hyn. (e}fel hyn (Corn, evel henn) though still surviving by re- formation, became (e)fell hyn, whence efelly y felly, felly 'so', 110 v (2). In Gweiit fell hyn became llyn, and subsequently yn llyn with adverbial yn, BAB. i 376, 378.

Ni fwriadwn fawr rodiaw

A gwr fell hyn gar fy Haw. T.A., c. i 338.

' I did not intend much to roam with a man like this near me.' ac evelly A.L. i 6 'and similarly'; Ay yvelly y gwnaethant wy ? W.M. 41 ' is it so that they did 1 '

(3) val, O.W. amal (: Ir. amal) is a weak form of hafal < *8 e m e l- 94 i ; Ml. W. mal may represent an early elision of the first syllable, thus *a'm e l- ; the e in Mn. W. may come from forms like felly where the a is affected by the y ; but Bret, has evel also. Ir. amal governs the ace. case.

tnegys seems to be a spv. (eqtv. 1) of the same word corresponding to comparatives in -ach ( < *-ak-aori) ; thus *8'm-dk-t8t6- > megys ;

• sam-dk-i8td- > Ml. W. yvegys G.c. fac. i. The use of megys as a

noun, as yn y megys liwnn C.M. 39 'in this manner', does not prove it an orig. noun ; cf. Mn. W. yr un fel ' in the same way '.

216. Compound Prepositions. This term may be used to denote expressions in which the last element is a preposition, to distinguish them from composite prepositions, in which the last element is a noun. They fall into two classes : i. prep. + prep. ; ii. noun, adv. or pron. + prep.

i. (i) Ml. W. y am ' from about ; besides ', 209 v ; y air ' from on ' ; y gan ( from with ' ; y wrth ' from by ' ; Mn. W. oddi am ' from about ' ; oddi ar c from on ' ; gan ' from ' 211 ii (2), more rarely oddi gan, see oddi gennyf 194 v (3); oddi wrth, now mostly oddnbrth ; Ml. W. 08 is R.M. 151, 172. Inflected: oddi amdano Gu.O. G, 193, oby uchtaw R.M. 141, ete. Ml. W. Si-eithyr S.G. 8 'except', o-bi-eithyr IL.A. 143 'outside'; Mn. oddieithr ' except ', dial, corruption oddigerth.

oddi is itself a compound of o 4- Si ; in Ml. W. it is comparatively rare before vowels : o8yarnati IL.A. 159, but occurs before adverbial expressions as o8yyma S.G. 7, 40, in which, however, it is generally 08- before a vowel, as o8ymma S.G. 4, o8yna W.M. 19 ; this is also the usual spoken form. The -i (mostly t before a vowel) is taken in Late Ml. and Early Mn. W. from forms in which a consonant follows, as oddi draw, Ml. W. obydraw C.M. 46. (oddieithr is for o ddieithr.)

In the Gweutiau dial. 08 was taken from these connexions, and used for o before a vowel, and 18 for i was made on its analogy. These forms occur in late Gwentiau writings ; and Pughe made a determined but unsuccessful attempt to substitute in the written language the new Gwentiau 08 y 'from the', etc., for the ancient o'r, etc., in order to avoid the apostrophe !

(2) Mn. W. er ys, ers 214 vii ; er cyn, as in er cyn cof 1 from before memory ' i. e. from time immemorial.

(3) gor-uwch, gor-is 45 iv (2) ; cyf-rwng 210 viii (4).

(4) The forms odan, amdan, ohonof, 209, are compound preposi- tions, and are often written o dan, etc. ; 209 vii.

(5) The combinations a chan ' having ', Jieb gan ' without having ' are not compound prepositions, because each prep, has its own object ; thus in heb ganddynt fagail Matt, ix 36 the obj. of heb is bugail, and the obj. of gait, is the suffix, so that the phrase may also take the form heb fugail ganddynt, lit. ' without a shepherd with- them '. Similarly cyn i, wedi i, er i, etc., before verbal nouns; the first prep, governs the v.n., as in cyn i mi ddyfod, which may also be expressed by cyn, dyfod ohonof lit. ' before the coming of me ', i. e. before I come.

ii. (1) hyd yn, hyd ar, hyd at ' as far as, up to, till, to '.

hyt ym penn y vlwyfyn W.M. 4 ' till the end of the year ' ; hid attad B.B. 3 'to thee '; diaspad . . . hid ar duu y dodir do. 106 ' the cry to God is it raised/

hyd yn oed ' as far as, even '.

In Ml. W. it has two meanings : (a) ' up to but not including* i. e. all except : a cafael cubel hyt enoet un keynyauc A.L. i 100 'and all is had except one penny ' ; (b) ' up to and including ' : hyt ynn oet eu pechawt IL.A. 34 ' even their sin '. The latter is the meaning in Mn. W. : hyd yn oed Marc ii 2 ' even '. The phrase is in common coll. use.

The origin of oed or ynoed here is quite uncertain ; as no pref. or inf. pron. is used with it, it would seem to be an adv. ' even ' (1 noet < *nai-t-, variant of neut ' indeed' 219 i (i)).

(2) tu a(g), tua(g] l towards ', tiiag at id., parth a(g) id., parth ag at id. ; Ml. W. y gyt a(c), gyt a(c), Mn. W. gyd a(g), gyda(g}> ynghyd a(ff) ' together with ', gyferbyn a(g) ' opposite ', gyfarwyneb a(g) \&^ynglyn a(g) ' in connexion with ', etc.

tu ha L.L. 272 'towards'; tu &th wlat IL.A. 125 'towards thy country ' ; y tu ac attaw C.M. 47 ' towards him ' ; tu ac at IL.A. 158 ; parth a'r berth W.M. 69 ' towards the bush ' ; parth ac attunt do. 38 ' towards them ' ; aros . . . hyt parth a diweS y dyS do. 70 ' to wait till towards the end of the day ' ; ygyt ac ef W.M. 7 ' together with him ' ; y gyt ac wynt do. 5 ' with them ' ; gyverbyn a hi B.M. 293 'opposite her', gyvarwyneb ac wynt W.M. 185 'opposite them ' ; tu-ag-at am M.K. [xi] ' with regard to '.

tu 'on the side', like parth, is definite without the article an old construction which survived in a few idioms; the tendency to use y before tu, as y tu ac above, is shown by the early tu ha to be a Ml. W. neologism, which did not become general.

(3) Ml. and Early Mn. W. vi a, ti a, ef a, efo a, hi a, before vowels vi ag, etc. ' with, together with', literally ' I with ', ' thou with ', etc. The pronoun had lost its pronominal force, and its antecedent was frequently u pronoun of the same person coming immediately before it. Thus :

A minJieu vi a'r morynyon a wiscaf ymdanaf inheu W.M. 99 'and I with the maids will dress myself, kysgeist di ti a Lawndot s.o. 302 ' thou didst sleep with Lancelot.' bwyt a llynn . . . y'th neithator di ti a'm merch i R.M. 120 'food and drink for thy nuptials with my daughter '. Sef a wnaeth ynteu ef ae lu y noa honno R.B.B. 76 ' this is what he did with his host that night '.

A rhif gwlith o fendithion A fo i ffuto ef a Aon. L.G.C. 463 ; cf. 4, 308. c And blessings numerous as the dew be to Huw with her.'

Yr otdd Esyllt urddaswawr

Draw hi a't mob Rhodri Mawr. L.G.C., M 146/140. ' Esyllt the noble was there with her son Rhodri Mawr.'

T nef i hwn efo a At. T.A., A 14975/107. ' Heaven [be] to him with her.'

efo a(g) was contracted to efu(g], as the metre requires in the last example; see efo honn, efo hi S.V. c.c. 361. In Gwynedd efo(g) came to be used for ' with ' irrespective of the person of the ante- cedent ; this is noted by Simwnt Vychan as a grammatical fault, P.IL. xcvi. His example is Mi efo Sion ' I with Sion ', literally ' I, he- with Si6n ', which should obviously be Mi vi a Sion ' I, I-with Si6n ', and may have been so written by the author of the line, as it yields equally good cynghanedd. [Ab Ithel, knowing efo only as a dial, word mean- ing ' with ', entirely misses the point in his translation, and italicizes Mi and Sion, as if ' / with John ' could be ungrammatical in any language !]

(4) tu . . . i forms a numerous class of prepositional expressions, as tu yma i 'this side of, tu draw i 'beyond', tu Jiwnt i id., tu cefn i ' behind ', tu uchaf i ' above ', etc.

tu ' side ', Corn., Bret, tu, Ir. toib, Gael, taobh < Kelt. *toibo- ; origin uncertain ; Macbain 2 359 gives Vsteibh/p- 'stiff, erect', which seems far-fetched from the point of view of meaning.

ADVERBS

217. Negative Particles. i. The forms of negative particles are as follows :

(i) Before verbs : in a direct sentence, Ml. W. ny, nyt, Mn. W. ni, nid ; in an indirect sentence, Ml. W. na, nat, Mn. W. na, nad ; in a relative sentence usually the first form, sometimes the second, see 162 v (i) ; in commands, na. nac ( = nag) ; in answering 1 a question, na, nac ( = nag). The forms nid, nad, nac are used before vowels only ; the forms ni, na before consonants, and a mutated g, as ny wnn . . . ny allaf W.M. 21 ' I do not know ... I cannot ', na at > ndd 201 ii (2) ' let not ' (nid allaf is not in accordance with traditional usage). With infixed pronouns : mm, na'm, ni's, nyw etc.

The initial mutation after ny na in Ml. W. is as follows : p-, t-, c- spir. ; 6- rad. ; m-, II- rad. or soft ; d-, g- soft ; f- not shown (r- may be r- or r-). But na (nac) takes the soft of b-, m-, II-. In Mn. W. II- is always softened ; b-, m- are generally softened, though the rad. remained also in the Early Mn. period, and persists in ni bu etc. beside nifu; p-, t-, c- spir. In Early Ml. W. ny relative softens the tenues.

Examples : Ml. W. : p-, ny forthint B.B. 34 ' they cherished not ' ; t-, ny thykya w.M. 14 'avails not'; c-, ny chan B.B. 31 'he sings not ' ; b-, ny by8 W.M. 4 ' he will not be ', ac na bo B.B. 54 ' and that there may not be ' ; m-, ny mynneis W.M. 18 ' I would not ', ny mynn IL.A. 148, na me8 do. 147, but nyvynhei W.M. 58 ' would not ' ; 11-, ny lluit reuuet (-t = -8) B.B. 8 ' wealth avails not ', ny llesseint do. 63 ' were not slain ', but ni laSaf i di W.M. 8 ' I will not kill thee ; g-, ny wnn, ny allaf above ; d-, ny tiuuc ( = ny 8iw(y)g) B.B. 8 ' makes no amends '. Relative : corph ni glivit (-t = -8) B.B. 20 ' body that hear- est not'; ny bara B.P. 1175. Mn. W. : ni mynnafl.F. P 97/179 'I will not', ni feddodd W.IL. C.IL. 105 'he possessed not'; ni bu T.A. G. 251, nifu T.A. 37 iii (i).

(2) Before a noun, adj., pron., adv. or prep. : Ml. W. nyt, Mn. W. nid [rad.] 'it is not', used before vowels and conso- nants ; indirect nat, nad [rad.].

Nyt gwaratwyS gwelldu B.B. 962 ' it is no disgrace to reform ' ; Nid cur llavur urth din [read dim] da B.B. 7 ' it is not pain to labour at anything good '.

ii. (i) The negative adverb na ' no ' may answer any question introduced by a or ai ; it may be used alone, but is generally followed by a neg. part., as na, nid hynny ' no, not that '.

(2) A question introduced by a is answered in the negative by na, nac ( = naff) with the verb ; as A ddaw ef? Na ddaiv ' Will he come ? No ' ; but if the verb is in the aor. (or perf.) the answer is na ddo, sometimes written naddo, but wrongly, for the a is long, not medium as in a penult ; thus A aeth ef? Na ddo ' Did he go ? No '. Na bo W.M. 425.

(3) A question introduced by ai is answered in the negative by Ml. W. nac ef, Mn. W. nag e (often written nage) 'not so', as Ai tydi a*i gwnaeth ? Nag e ' Is it thou that didst it ? No.' More rarely thus : Ae guell . . . ? Na well W.M. 85.

iii. A negative part, is frequently supplemented by bim ' at all ' ; see 170 v (3).

iv. (1) Ml. W. ny < Kelt. *ne < Ar. *ne. Ar. *ne was ordinarily accented, and the verb unacc. 179 i. In Kelt, the initial of the unacc. word seems to have been doubled after the accented vowel ; in Jr. gemination occurs after ni ' not ' and the preverbs ro, no. Thus Brit. *ne kkarame gives Ml. W. ny charaf. Hence the spir. of tenues after m. So *bb->b-, *mm->m-, *ll->ll-. The soft 8- may be due to late simplification of double d 93 iii (i) ; lenition of g- may have spread from gw-<*u-, which even if doubled would prob. give w- after a vowel. From these and the relatival form, lenition spread to b-, m-, tt-, rh-. The neg. rel. lenited because it was orig. unacc., and the verh accented, so that the regular softening took place after the vowel, 162 vi (3).

(2) Ml. W. nyt was orig. 'there . . . not'<*w(e) ita 189 iii (3) : and was used before consonants as well as before vowels, as 0. W. cen nit boi (prob. b- = v-) CP. ' though there be not '. The difference in meaning between ny ' not ' and nyt ' there . * . not ' was lost, and both are used in the two senses, ny before consonants and nyt before vowels.

(3) Ml. W. nyt ' it is not ' before a noun, etc., may come from *ne tod 'it [is] not', where *tod 'it' is the neut. sg. nom. of *so, *sa, *tod>Gk. 6, rj, TO, 159 iv (i). It is improbable that nyt contains the verb ' to be ' as Strachan assumes, Intr. 98.

(4) Though the vb. was unacc. after *ne in direct sentences in Ar., it was accented in dependent clauses ; this may have led to a reduced unacc. *n e giving Kelt. *na, W. na. If so, the mutation after na and the form nat followed the analogy of ny, nyt ; but this is probable in any case.

(5) W. na, nac before the impv. may be referred to Kelt, unacc. na + a particle beginning with k-, possibly cognate with Lith. -ki, a particle suffixed to imperatives.

(6) W. na, nac in answering questions. In na ddo (: Ir. na-tlw) we have simple na ; in nac ef ' it [is] not so ' the -c may represent some form of the *ke- pronoun.

§ 218. Interrogative Particles. i. The interrogative par- ticles are : (i) before verbs, a [soft] ; before nouns, etc., Ml. W. ae, Mn. ai [rad.] 'is it? ' (2) before verbs, O.VV. anit, Ml. AV. pony(t), pany(t), Mn. poni(d), pani(d), pond, panel, oni(d), ond ' nonne ? ' ; before nouns etc., Ml. ponyt [rad.], Ma. ponid. pond, pand, onid, ond ' is it not ? ' The initial mutation after pony etc. is the same as after ny ; so the use of -t before verbs. (3) Mn. W. ai S ' is it so ? ', onid 6, onite ' is it not so ? ' dial. N.W. ai e ? yrit e ? S. W. ai ef e ? t ef e ? ont ef e?

Examples : (i) Ml. W. A w&ost ti B.T. 27 ' Dost thou know? ' ae ti a eirch vy merch W.M. 479 ' is it thou that seekest my daughter 1 ' (2) O. W. anit arber bit JTJV. gl. nura vescitur 1 Ml. W. Pony welwch chwi B.P. 1418 'do you not see?' Pany chredwch chwi ib. 'do you not believe ? ' Ponyt ydym ni yn kredu IL.A. 83 ' do we not believe ? ' Ponyt llygoden a welaf i yth law di W.M. 78 'is it not a mouse that I see in thy hand ? '

Pand hir na welir ond nos ?

Pe byr, hir yw pob aros. I.F., M 148/59.

' Is it not long that only night is seen ? Though short, all waiting is long.'

Ond hir yr wyd yn tario ? W.1L., G. 293.

' Is it not long that thou art tarrying ? ' Onid oes dinistr i'r anwir ? Job xxxi 3. Ond rhaid i trdd fyw ? B.C. 119 ' must not trade live ? '

Preverbal a may be followed by an infixed pron. in Ml. W. : a'm dywedyS IL.A. 134 ' wilt thou tell me ? ' ae gwBost di S.G. 4 ' dost thou know it? '

In Late Mn. W. the p- forms are obsolete ; the forms used are oni, onid, more rarely ond. Wm.S. has ani, anid, which may have been dial, forms in the i6th cent.

ii. These particles originated in indirect questions : Ac amovyn a Pheredur a welsei y kyfryw varchawc W.M. 138 ' and inquiring of Pere- dur whether he had seen such a knight ' ; ny wnn a glyweist ywrthaw do. 166 'I know not whether thou hast heard about it' ; a gofyn a oruc Owein ae dyn bydawl K.M. 187 'and Owein asked whether it was a living man '. The point of transition is represented by Dywet . . . a weleisti W.M. 1 1 8, which may be rendered ' say whether thou hast seen ' or ' say, hast thou seen ? '

ae . . . ae ' whether . . . or ' : A w8osti peth wyt . . . ae corff ae cneit B.T. 27 'dost thou know what thou art, whether body or soul ? ' y rofti dewis uSunt ae gwrhau i8aw ae ymwan ac ef, see 222 ii (2).

iii. a [soft] ' whether ' may represent unacc. Brit. *d ' if ' instr. sg. f. of the pron. *o- : cf. Gk. ^ ' if ' which however is from *e, variant of *o instr. sg. m. ; for the instr. f. as adv. cf. Lat. ea, qua. See 222 v (i).

ae [rad.J is a contraction of a and a vocable *y, which orig. ended in a cons., and may be from *id ' it ', BO that ae may be lit. ' whether it [is]'; cf. nyt 217 iv (3).

l>o-ny, pa-ny<~Brii. *q^a ne ' whether not' ; *q v d instr. sg. f. as *d above ; if unacc. in Brit, it would give pa- ; if uuacc. later, po- ; see 71 i (2). ai e < ai ef is it so ?' yn't e for *arid fief ' is it not so 1 ' : (h)ef <

• 8emo-s, -d ' that, it, so' 159 iv (i). The S. W. second e repeats

the pron. of ae. Mn. W. ai comes from ai e, which is easier than ae (aq) e.

219. Affirmative Particles. i. (i) Ml. W. neu, neut before verbs, the former before consonants and with the same mutations as ny, the latter before vowels ; with infixed pron. neu'm, neus etc. ; with the perfective particle neur. Before nouns, adjs. etc. neut. [rad.] ' it is ' ; with neg. part, neut na(f).

neu cheint B.T. 19 ' I have sung ' ; neut atwen nat yr vy lies K.P. 1039 ' I know that it is not for my good ' ; neu'm due i Eljfin B.B. 67 ' E. brought me ' ; neu's ro8es W.M. 20 ' he has given it ' ; neur vum B.B. 7 'I have been' (also in full neu ry do. 74, W.M. 80) ; neut kyn- tevin, neut ru8 rycTt, neut crych egin K.P. 1036 ' it is spring, the furrow is red, the sprouts are curly' ; neut na'm dawr do. 1227 'I care not ' ; neut nat ry8 ib. In Early Mn. W. neu is a rare survival :

E fu amser neu dderyw

Ochfi! ban oeddwn iach fyw. 0.0.425.

' There was a time it is past ah me ! when I was alive and well.'

(2) neu for *nwy, 78 iii, < Brit. *nei loc. sg. m. of the pron. *no- : Gk. vat, Lat. nae 'indeed' (ei/ai 63 v (2)), Gk. vy, Lat. rie ' indeed ', instr. sg. m. of the same. The mutations after neu and the two uses of neut are to be explained like those of the parallel ny, nyt 217 iv.

ii. (1) Ml. W. y, c, y8, et ; yd, ed, yt ; yd-, yt(t)- ; Mn. W. y, y&> yr? yd-, yt-- In Mn. W. these are used almost exclusively before the pres. and impf. of the verb ' to be '. yd- was agglutinated to these tenses early, and ytt- spread from yttynt and yttoeb 189 iii (i), 180 ii (3). The compounds yd-wyf etc. were used like the simple forms, and might take other pre verbs before them, as neut yttiw dros amser W.M. 182 'it is past the time ', nit yttoy^wn i do. 8 ' I was not ', a yttiw Lawnslot yma S.G. I ' is Lancelot here ? ' Even yr yd- is common ; yr ydwyf 191 ii (2). In answers and denials the yd- forms only are used in the pres., except in the 2nd sg., as ydwyf ' I am! ' ydych 'you are ! ' but wyt ' thou art ! '

Ml. W. Yd wele(i)s-e Guendoku B.B. 53 ' I have seen Gwendolen '. Y roSet y march yr mob, ac y deuth hi . . . W.M. 33 ' The horse was given to the boy, and she came . . .' Ac y dyvu Glewlwyt y'r neuaS do. 457 'And G. came to the hall '. Pan 8oeth yti y peir ? E doeth im . . . do 45 (cf. 46) ' Whence came the caldron to thee ? It came to me . . .' Na wir, y8 ym wyrda R.M. 105, "W.M. 458 'No, indeed, we are goodmen '. Yt oet ( = y8 oe8) in y diffrid , . . Ysprid Glan B.B. 45 ' The Holy Ghost was protecting her/ Mn. W. : Ac y dyweit Iwl Kesar Y.L.H. [8] ' And Julius Caesar says ' ; yr wyf, yr wyt, yr oedd, yr ydym, yr ydoedd, etc.

(2) These particles are adverbial forms similar to the forms of the oblique relative 162 vi (2); but the base of these was probably the pron. stem *i- or *e-. If the suffixes survived in Kelt., there is no reason to suppose that they were added to only one base.

iii. (i) Early Ml. W. ef. This is found not only (a) before the 3rd sg., but also (b) before the impersonal, and (c) before the ist sg. The initial following is usually rad., sometimes soft (ef labhei B.A. 37, ef ' enir below) ; d- is ambiguous.

(a) Ac ew dybit ( = ag ef dybyS) B.B. 6 1 ' and it will come '. Ef diodes gormes, ef dodes fin B.A. 10 'He repelled invasion, he set a boundary'. Ef dyfu dreic llu P.M. R.P. 1419 'The dragon of the host came'. (6) Ef molir pawkwrth y weith R.P. 1056 ' Everybody is praised according to his work'. Ef gwenit B.A. 22 'There was an attack '. (c) Ew kuynhiw iny wuiw ( = Ef cwynif yni{ fwyf) B.B. 100 ' I shall complain while I am '. Ef gwneifbeirS byt yn llawen B.T. 63 ' I will make the bards of the world merry '.

It might be preceded by the negative nyt or another preverb :

(a) Nyt ef eiste8ei en tal lleithic B.A. 10 ' He would not sit at the end of a bench '. (6) Nid ew rotir new i'r neb nuy keis B.B. 86 ' Heaven will not be given to him who does not seek it '. Nyt ef enir pawb yn 8oeth B.P. 1056 'Everybody is not born wise '. (c) Nyt ef caraf amryssonyat B.T. 8 ' I love not strife ' ; kyt ef mynasswn do. 65.

It is probably an accident that it is not found before other persons.

(2) The pronouns mi, ti, hi etc. might come before the verb, agreeing in person with the subject. They might be preceded by nyt or another particle.

0. W. Ti dicone(f)s a a di(ar) a mar JTJV. SK. ' Thou madest both land and sea*. Early Ml. W. A mi 8ysgoganaf-e B.B. 48, 49 'And I predict'. Pan esgynnei baub, ti Bisgynmit B.A. 31 'When every- body ascended, thou descendedst '. Nyt mi wyf kerS wit B.T. 31-2 'I am not mute of song'. Neu vi erthycheis do. 62 'I groaned'. Pel mi ganwn B.A. 26 ' If I sang '.

(3) In Ml. W. the rel. a was inserted after ef&nd. mi etc. in the above constructions ; examples occur as early as the last

diconeis for what would be later iJigoneist ; -e- for -ei- occurs several times in the fragment. pages of the B.B., bat are not found in the B.A. It may have arisen partly as a support to an infixed pron., as Mi a,e dywedaf yt IL.A. 4 ' I will tell it thee ' ; hi &y gwelei ef W.M. 251 ' she saw him ' ; Hi ay provwn do. 66 ' We will try it', cf. iv below ; and partly mi a wn may be a confusion of mi wn ' I know ' with mi a wyr K.P. 1227 = Bret, me a oar ' [It is] I that know'. The a is often written where the metre shows that the author did not use it, as in hi a vu several times in R.P. 1365 for hi vu.

(4) In Mn. W. ef a, mi a etc. remain in use, as Mi a euraf 38 ix, Ef a borthes yr lesu D.N. F.N. 94 ' Jesus fed [the multi- tude] '. In the Bible ef a becomes efe a, except where it is clearly a particle, when it is written fe or fe a, asfe allei Gen. xvi 2, fe a allei I Bren. xviii 27, or fo as fom lleddir Diar. xxii 13.

But the natural Mn. forms seem to be ef, e,fo,f' y fe\ mi, ti etc. ; as Ef aelh D.G. 374, 527, E fti amser i (i), E gaeodd Mai 129 ii (i), Fo ddaw D.G. 175, fo'm cafodd do. 177 ; Mi vu do. 501, Mi welwn T.A. G. 238.

Tra fo gwlith mewn tref a gwlad Fo sdn dynion am danad. W.1L. 18.

' While there is dew in town and country men will talk of thee.'

F' aeth anwir ar faeth ennyd ;

F' aeth y gwir ar feth i gyd. I.F. F. 42.

' Untruth has prospered for a season ; truth has wholly failed.' Fe wna hon a fynno hi. D.G. 516.

' She will do as she pleases.' Note fe with fern, subject. The form was prob.fo, a.sfe is late; it occurs in the i6th cent. : ve golhid yr hen lyfreu Y.L.H. [8] ' the old books would be lost '.

In the spoken lang., in S. W. t (forjtf, mil) and/g are heard; but in some parts the pron. of the same person as the subj. is used, as chi welwch ' you see ', nw dn' ' they will go '. In N. W. mi alone is used for all persons, having ousted fo, which survives only in parts of Powys. In Sweet's specimens of N. W. dialect TPS. 18824, 477 many assertions begin with the verb, with rad. initial, which is utterly impossible in pure dialect. Every such verb is introduced by an affirmative particle, except in answers and denials consisting of single words, as Clywof ' Yes, I hear '.

(5) Ml. W. efaa in (i) above is the same as the ef in noc ef ' not so ; no', at e 'is it so?' and i-ef 'it is so'. The construction mi ganaf may be originally ' as for me, I will sing ', which explains the oblique mi instead of the nom. i. Undoubtedly later the pronoun was identified with the subject, though ef largely retained its character of a particle.

iv. (1) The rel. a is used in Early Ml. W. to support an infixed pron. before a verb ; thus

A.'th kivarchaw B.B. 98 = A?th gyfvarchaf R.P. 578 ' I greet thee '. A'/A vendiguis-te Awraham B.B. 35 'Abraham blessed thee.' Ac 8,'wch bi wynnyeith B.T. 12 'And there will be vengeance upon you'. A!s attebwys DofyS do. 24 ' The Lord answered him'. A's kynnull gwenyn do. 40 ' Bees gather it '.

It is used not only in affirmative sentences, but also before the subjunctive to express a wish ; as

A!m bo forth B.B. 34 ' May there be a way for me '. A'n eirolve ne ( = eirolwy ny) Mihangel do. 32 ' May Michael intercede for us '.

(2) This form prob. arose where the subject was expressed, as in A's attebwys Dofydd, the a anticipating Dofydd ; and is perhaps a survival for a particular purpose of the habit of putting the rel. clause first, which prevails in Skr. (Whitney 512 a), and may have been primitive.

v. (1) Ml. W. ry, the perfective particle, with the past makes it perf. in sense, ws, pawl} ry gavas y gyvarws W.M. 470 ' everybody has had his gift ' ; with the pres. subj., makes it perf. subj., as kanys ry gaffo o arall do. 453 ' though he may not have had him from another ' ; with the impf. subj., makes it plup., as kyn nys ry welhei eiroet do. 454 ' though he had never seen her ' ; with the plup., causes no modification of meaning, y ryn ( = yr hyn] ry abawsei do. 453 ' that which he had promised '. See Strachan, Intr. 57-60. It is sometimes reduced to r after neu i (i) ; ny, as nyr bar/o W.M. 230 ; , as ar boetfioeb do. 123. In Early Mn. verse ry is a rare survival : Annoethwas a'i rhy- wnaetlioedd D.G. 509 ' A booby had made it '.

It is prefixed to a verbal noun giving it a perfect sense ; and is mostly found redundantly after gwedy, as yb oeb kawat o eira gwedy ry-odi . . . a gwalck wyllt gwedy ry-lab hwyat W.M. 140 ' a shower of snow had fallen, and a wild hawk had killed a duck'; this is reduced to (g)wedyr S.G. 53, which survives in Early Mn. W. verse, as gwedy r* odi D.G. 27 quoted from the above ; wedy r* euraw L.G.C. 363 ' having been ennobled '.

It is seen from the first example above that the rel. a was not used with ry, which may contain the rel. without alteration of form. But in the Late Ml. period a began to be inserted before it, as ac a ry- umaethoeS W.M. 30 ( 151 ii (2)). The mutation after it was orig. the same as after ny ; thus in direct statements ry chedwis detyf B.B. 14 'he kept the law'; relatival, pawb ry gavas above. The lenition of the relatival form was generalized.

(2) Ml. W. ry - Ir. ro < *pro- : Lat. pro, etc., 156 i (21). The relatival use may be due to the analogy of ny, though it is not im- possible that rel. ry may have been formed like ny itself, by contraction, thus ry < *r(t)o < *pr(o) to.

vi. (i) Positive answers : to questions introduced by a, the answer is the verb repeated, or its equivalent, as gwnaf ' I will do [so] ', except when it is aor. or perf., in which case the answer is do ' yes '. To questions introduced by ai the answer is Ml. W. ief, ieu, Mn. W. t-e ; indirect, Ml. W. mae ef B.M. 29 ' that it is', Mn. W. mai e.

In Ml. W. the verb may be repeated in the aor. also : A ovynneist tl a oe8ger8 ganihwrdf Govynneis W.M. 487 ' Didst thou ask whether they had a craft 1 I did.'

Whether ef W.M. 42 corresponding to mae ef B.M. 29 is a scribal error, or a shorter form of reply, is not clear.

(2) do : Ir. to ' yes'. Thurneysen, Gr. 492, derives the latter from Ar. *tod ' that' ; but W. d- is inconsistent with this. Rhys, LWPh.' 242, assumes that it is the preverb *do, the verb being omitted so that do became a generalized past verb meaning ' he (I, we, etc.) did ' ; *do- survives in Welsh only as the prefix d>j- : Ir. to-, do- Vendryes Gr. 239 ; there are survivals in Ir. of do used as a perfective particle : ndigid 1 milks ', perf. sg. i. do-ommalg, tongid ' swears ', perf. du-cui-tig, Thurneysen Gr. 322. The alternation t- : d- occurs in this, cf. 196 i (3) ; and the answer expected is a verb.

i-ef < *l semo-s ' that [is] so '. *l : Gk. ovrocr-f, Umbr. -t : Goth. ja, O.H.G. ia, E. yea. mai e ' that it is so' ; mai 222 x (2), e as in ai e, see 218 iii.

220. Adverbs of Time, Place, Manner and Measure. i. (i) In Ar., adverbs or words which were later used as adverbs had the following- forms : (a) Bare stems, as *ne 217 iv (i), *pro > Gk. irpo 210 x (i). (b) Cases of noun, adj. and pron. stems, including the nom. sg., as Lat. versus 211 iv (2). (c) Stems with special adverbial suffixes ; see (3) below.

(a) (a) A demonstrative or similar adj. forming with a noun in an oblique case the equivalent of an adverb was often compounded with it as Lat. ho-die. (6) A preposition with its object generally forms an adverb equivalent, and many such expressions became improper compounds, as Gk. fK-TroScov.

(3) The special adverbial suffixes were (a) forms with a dental, see 162 vi (2); (6) forms with gh- as Gk. Si-xa", see 222 i (3); (c) forms with r, as Lat. cur, W. pyr ' why 1 ' {d) forms with a nasal, as Lat. superne, see 209 vii ; (e) the suffix -s, as in Gk. Si's, Lat. bis. See Brugmann 8 II ii 728-738.

ii. The following- W. adverbs represent old adverbial forms :

(1) Early Ml. W. nu 'now', as Nu nym car-i Guendit B.B. 50 ' Now Gwenddydd loves me not '. The sound was doubtless nw (: Ir. mi), and the Late Ml. nu e.g. W.M. 413, instead of *nw, is a mechanical transcript of the earlier spelling, the word having become obsolete.

nu < Ar. *nu bare stem, beside *nu : Gk. vv, O.H.G., O.E. nu,

Skr. nu, nu.

(2) Early Ml. W. moch ' soon, early, quickly ' e. g. B.B. a.

moch, Ir. mos ' soon ' < *moks = Lat. max, prob. nom. of a cons, stem like vix (: vinco) Brugmann 2 II ii 679 : Skr. maksu 'quickly, soon '.

(3) doe ' yesterday '.

doe = Lat. Am' both from *gMiesei : Gk. x#c's 75 vii (2), 98 i (3).

(4) yrhawg, rhawg 'in future, for a long time to come', Ml. W. yrawc R.P. 1034.

yrhdwg < *j)era-ko-(s) foimed from *pera like *prokos (> Lat. -procus, W. rhag) from *pro : Gk. Trtpa, *irpai<o~ in Ion. Trp^o-o-w (Brugmann 2 II i 481).

(5) hwnt ' hence, yonder ', as Ef hwnt, ef yma B.T. 37 ' It (the wind) [is] there, it [is] here '. Saf hwnt Gen. xix 9 ' stand back '. Dos hwnt M.E. i 125 'go away '.

hwnt, Bret, hont < *som-tos consisting of the demoust. stem *som- 'this', 164 vi, and the suffix *-tos 'from' as in Lat. in-tus 162 vi (2).

(6) yno ' there, thither, then ', yna ' then, there (near you) ', Early Ml. W. ynoeth B.B. 66 'thither', inaet/i do. 58 'then', ot-ynoeb B.T. 19 'then, thereafter', ob-ynaeth R.P. 581 id.

yn ' there, thither ' before the rel. y, yb, yd ' where ', as yn-y tereu tonneu tir B.B. 63 ' there where waves beat the shore ' ; en e-bo dadeleu A.L. i 62 ' where there is a suit ' ; A'r vorwyn a boei/t yn yb oeb Peredur W.M. 1 48 f and the maid came to where P. was '. Also, similarly used, myn, men, as myn-yd vo truin yd nit trev ( = vyb trew) B.B. 83 ' where there is a nose there will be a sneeze'; cf. 26; a bode* vy ren men y maent ryb R.P. 1367 1 which my Lord has put where they are free ' ; cf. 1 244.

The older forms of yno, yna are ynoeth, ynaeth ; the B.T. ynoeb represents the intermediate stage between ynoeth and yno 78 i (i). ynaeth>yna has followed the analogy of ynoeth; Powys dial, ene shows the change of ae to e 31. ynoeth and ynaeth imply Brit. *enokt-, *enakt-, the latter doubtless for *enakt- 74 iv. These are prob. derivatives of the pron. stem. *eno- ; but the formation is not quite clear. We may assume forms *eno-ko-8, *end-ko-s formed like *pro-ko-s, *pera-ko-8, and adverbs with a ^-suffix formed from these, on the analogy of *ek-tos ( : Ir. acht, Gk. CKTOS) ; thus *eno-k-te ' thither ' > ynoeth. For the base cf. Skr. ana ' then ; ever ', Gk. Ivr; ' the third [day] ' (' that [day] '), Umbr. inum-k, inum-ek, enom ' turn '. Ml. W. yn ' there ; thither ' may represent the loc. and ace. *eni and *enom of the pron. Ml. W. myn, men seems to be the same with initial (y)m- < *esmi, see(n). The rhyme ren/ 'men shows that the -n is single, and that the vowel was long ; hence the word cannot be an oblique case of mann ' place ', though so treated later, and written man.

(7) eto 'again, yet', Ml. W. etwo, etwa, earlier edwaeth C. R.P. 1173, etwaeth B.T. 29, M. w. 30, eddwaeth (dd = d-d, not 5*) B.B. 88. Also etton B.P. 1264, 1309, etonn do. 1321, etwan K..A. 37, W.M. 6 1.

The t is for d by provection before w 111 v (2), so that the older form was edwaeth, *edwoeth (wa : wo interchange), which implies Brit. *et..uokl-. This seems to be a formation like yno, see (6), from a base *eti-uo ; *eti : Gk. In, Skr. dti (which may represent *ati or *eti) ' over, beyond ' ; uo < *upo : Skr. upa, as adv. ' moreover, further ', see (9). The form eton, etwan < *edwon < Brit. *eti-uo-nd, an adverb formed with an n-suffix, see i (3). For loss of w before o see 36 iii. The existence of *eti as well as *ati in Kelt, is shown by Gaul, eti-c ' and '. It does not seem possible to explain the e- of eto except as original *-.

(8) hefyd c also, besides ', Ml. W. hevyt. In Late Mn. W. it is used in positive statements only; but in Ml. and Early Mn. W. its use is not so restricted ; see e. g. W.M. 8.

^[2]

Ni threithir y gwir i gyd

Yn llyfr nac unlle hefyd. G. Gl., P 114/458.

' The whole truth is not stated in a book or anywhere else.'

hefyd < Brit. *sami-ti ; suff. of manner *-ti 162 vi (2) added to *s e m-i-, with z-flexion following *s e m e li- (: Lat. simili-s) : Tr. samlith, same meaning, < *samali-ti < the fuller *s e m e li- : cf. Lat. simitv, apparently formed with suff. -tud from loc. *semei, "Walde 2 s.v.

(9) wedi ' afterwards ' e. g-. Matt, xxvi 73, Act. iii 24, B.CW. 21 1. 10, gwedi 1. 22 ; Early Mn. W. and Ml. W. (g}wedy, O. W. guotig ox., guetig B.S.CH. 2 ' afterwards ' ; na cTiynt no, gwedy R.M. 1 68 ' neither before nor after ', cynt na chwedy L.G.C. 66.

The final -i is late 213 ii (2). In the recent period wedi adv. has given place to wedyn, a dial, contraction of wedy hyn ( after this '.

gwedy, O. W. guotig, Bret, goude < Brit. *uotig(os) which may be for *uo-te-gos (eg > ig 65 ii (3)) ; *uo < *upo which as an adverb of time meant ' after ', cf. Skr. -tipa adv. ' moreover, further ', and Lat. s-ub- in sub-sequof, succedo; *-te suffix of time 162 vi (2) ; to *uo-te seems to have been added the suff. *-ghos as in ac ' and ' 222 i (3). Its consonantal ending is proved by the rad. initial which follows it as a prep.

(10) draw 'yonder' ; yma a thraw 'here and there'.

draw is probably for *trawf 110 iii (i) < *tram-, perhaps loc. *tramei of stem *tramo- : cf. *j)ramo- in Lat. prandium. " From Vter- there are old nominal m-formations, which have become adverbial and prepositional " Brugmann 2 II ii 901. See 156 i (22).

(11) ^ma 'here', poet. yman\ Ml. W. yma W.M. 22, ymma do. 32, 39, yman IL.A. 30 ; hyt yman W.M. 186 { hither ' ; draw ac yman R.P. 1369.

A chats un o'i chusttnau * Misprinted yma.

Yman a i'w ddwyn ym, neu ddau. D.G. 186, cf. 264.

' And ask for one of her kisses to bring here to me or two.'

Chuilio yman (misprinted ym man) . . Chwilio hwnt Gr.O. 32 ' Searching here, searching there '.

W. yma, yman, Corn, yma, omma (o ≡ y Williams Lex. s.v.), -ma man, Bret, ama, aman, -ma, -man, Van. ama, amann, amenn. Oil the loss of final -nn see 110 v (2). The word is perhaps to be divided *ym-ann < *esmi loc. sg. of the pron. *e- 189 iii (2) + *anda prob. < *an-dha ; *an- variant of *cn- of the *eno- pron. (cf. Goth. anfiar 'alius ' Brugmann 2 II ii 336) with suff. -dha 162 vi (2) as in Skr. i-hd ' here ', Gk. lv-0a ; *anda survives in Bret, ann ' here ', Ir. and ' there, in it '.

(12) allan 'out, in the open', Ml. W. attann K.P. 1044, IL.A. 106, 167, usually written allan bat rhyming with -ann in Early Ml. verse, thus cann / lloerganin) /allan (ri)/lan(n) B.T. 27.

The adj. allanol ' external ', so written and pronounced, is not older than the xyth cent., and so was formed long after the distinction between '-an and '-ann had been lost, 56 iii. There was no deriva- tive of allan, and therefore nothing to show whether it had -n or -nn.

allann < Brit. *alland(a), which represents *p e l-iam-dha or a similar formation from Vj>eld- ' stretch out ' : Lat. palam ' openly ' : O. Bulg. polje ' field ', O.lE.feld, E. field; cf. imaes '*in field' vi (2), which has ousted allan in S. W. dialects. Cf. also Mn. Ir. o soin ale ' from that time forward ' O'Don. Gr. 263 : o hynny allan W.M. 12 (soin Mn. W.) ' thenceforth '.

(13) Ml. W. rwy 'too much', as rwy yt werihey Arthur W.M. 470 ' overmuch dost thou asperse Arthur ' ; see viii (i).

(14) y, y, yd adverbial rel. 162 ; pyr ' why ? ' pan ' whence? ' cw, cwb, cwd 'where?' 163; arnodd etc. 209; heibio, acw t trwob, drosodd, yngo, yngod, ucho, uchod, iso t isod 210.

iii. The following adverbs are oblique cases of nouns and adjectives :

(1) fry 'up', obi. case, prob. loc., of Ire 'hill' 103 ii (i).

(2) orig 'for a little while' dim. of awr; ennyd 'for a little while' (also am orig, am ennyd) ; ennyd awr D.G. 102 id.; oil 'wholly' 168 ii (2); lawer 'much' 169 ii (i); beth 'to some extent' 169 iv (i) ; ddim 'at all' 170 v (3) ; syrn 'a great deal' obi. case of swrn 'cluster, crowd' 129 ii (i) ex. 3 ( < *s-tur-no- : Lat. tur-ma, Vtuer-} ; gylcTi 6gylck, etc. 47 iii ; agos ' nearly ' ; nemawr. fawr in neg. clauses ' much ' ; achlan ' wholly '.

achldn is used like oil, generally following the word or phrase which it limits, as aV byt achldn ' and the whole world ' M.A. i 376, Prydein achlan K.P. 1402, y lluoeb achlan K.M. 136 ' all the hosts '. It is prob. an adj. which as an adv. retains its old accentuation like yrhdwg, erioed 47 i, ii. The most likely Brit, form is *aK>kladnos which may be for *n-ql9d-no- ' un-broken ', Vqoldd- 'strike, break' : Lat. incolumis ' un-harmed, whole ' ; cf. E. whole in two senses ; cf. also W. di-dum ' unbroken, whole ', di-goll ' whole ', coll<*qol'd-, Vqolad-.

(3") After an adj. : iawn ' very ', as da iawn, ' very good ' ; odiaeth ' very ', Gen. xii 14 ( : odid) ; aruthr ' amazingly, very ',' as merch landeg aruthr B.CW. 9 ; ofnadwy ' terribly ', etc. (4) Before an adj. with rad. initial: llawer before cpv., 169 ii (i) ; mwy, mwyaf 151 i ; similarly llai, lleiaf ; and in Mn. W. digon, as digon da 'good enough'; numeral with cpv. (with mutation peculiar to the numeral) 154 iii (a).

(5) gynt 'formerly'; cynt 'previously'; gynneu 'a shovt time (few hours) ago ' ; mwy, mwyach ' henceforth ' ; byth 'ever'; weithiau 'sometimes'; unwaith, etc. 154 iii (i); chwaith, ychwaith f either ', which replaces hefyd in neg. clauses in the late period, as na Herod chwaith Luc xxiii 15 'nor H. either'.

byth is the Ir. bith ' ever ' borrowed, the etymological equivalent of "W. byd ' world '. W. byth is generally sounded with short if, more rarely bifth which follows the W. analogy of monosyllables in -th. As the word is always accented the short ?/ can only be accounted for by the assumption of borrowing. The form a phyth R.P. 1028, L.G.C. 264 is due to the false analogy of a cJiynt in which the orig. rad. is c-.

chweith in Late Ml. W. occurs chiefly before a noun, and means ' any ', as na chlywei arnaw chweith dolur S.G. 55 ' that he did not feel jiny pain ', chweith antur do. 34, chweith pechawt do. 46 ; more rarely y chweith ' at all' do. 62. In Mn. W. it is found with an adj., as rhag na chaj)hom aros chwaith hir G.R. [95] ' lest we may not stay very long', Canys nid yw chwaith teg do. [124] 'for it is not very seemly ', chwaith hir B.cw. 40. These expressions seem to show that chwaith is orig. a noun ; perhaps gwaith ( occasion ' 100 i (2), as in unwaith above (with pref. *eks- 1) : Bret, choaz, Corn, whdth, wheth, 'yet, again ' (*-uokt- : *-uekt-).

(6) mwy (no) 'more (than)'; wellwell, waethwaeth 152 ii ; haeach in neg. clauses, meaning with the neg. ' not much, hardly at all ' ; oreu ' best ', gyntaf ' first ', etc.

Nyt arhoes ef haeach S.G. 38 ' he did not stay long '. The word is often used as a noun, as heb wneuthur hayach o brwc S.G. 39 ' without doing much wrong '; cf.iL.A. 122. hayachen R.M. 142, G. 234 has the sense of ' almost '. haeach seems to be a cpv. of an adj. *hae < *sag-io- or *sog-io-, Vsegh- : Gk. o^o. ' much ' adv., V segh-, Boisacq s.v. |^a>. haeachen is perhaps the full stem, and so the true obi. form, 147 iv (3).

(7) Noun or adj. in an obi. case followed by the obi. rel. y, yb, yr> neg. na, nad, (loc.) ni, nid: (a) in a dependent clause : modd y 'in the manner in which, so that', modd na 'so that . . . not ' ; pryd y ' at the time when, when ', pryd na ' when . . . not ' ; lie y, lie y8, lie yr, generally lie, lle'r ' in the place where, where ', Ml. W. lie ny, Mn. lie ni ' where . . . not '. (b) Predicatively at the head of a sentence, 162 vii (2) : odid y ' [it is] a rarity that, [it is] improbable that ', odid na ' [it is] improbable that . . . not ', i. e. it is probable that ; hawdd y ' [it is] with ease that ' ; da y ' [it is] well that ' ; print y ' [it is] scarcely [the case] that ', braidd y ' [it is] hardly [the case] that', as breib y diengi* R,B.B. 319 'he hardly escaped ', braidd na ' [it is] hardly that . . . not ' i. e. ' [it is (was)] almost [the case] that', as braidd na bunt bridd yn y bedd D.G. 296 ' I was almost dost in the grave '.

braidd may represent the instr. *bradu of an adj. cognate with Gk. /JpaSus ' tardy ', Lat. gurdus. Except in the above construction it generally has a governing prep, in Ml. "W., vii (i), but later it is used as an adv. in any position. It is not used as an adj.

An adj. preceding a vh. directly (without y), as mad Sevthoste B.B. 87 'well hast them come', forms a loose compound with it, 207 ii, and takes pre- verbal ny (not nyt\ as ny mad aeth B.B. 70, ny phell gwy8 B.A. 26 ' falls not far '.

iv. The following adverbs are formed of nouns in obi. cases with a demonstrative or similar adj., see i (2) (a),

(1) he^ddiw, Late Mn. W. Jieddyw 37 iii ; heno 78 i (i) ; e-16ni ' this year ' for *Ae-fleni, Bret, hevlene.

he&iw for *heSyw 77 v < *se-diues Skr. sa-divah ' at once ' beside sa-dydh ' on the same day ' prob. loc. sg. of an s- stem, and so not formed directly from *diieus ' day ', but an old formation going back to Pr. Ar. The others are prob. formed in Brit, on its analogy : he-no < *se-nokti loc. of *nokts; e-leni for *he-lyni (owing to prefer- ence for e..i sequence, cf. 65 iii (2)) < ? *blidnii loc. of *bleidonl which gives blwyddyn ' year '.

(2) beunydd ' every day ', beunoeth ' every night '.

The noun in these was ace. But Brit. *pdpon diien ( < *q*aq*om diiem) should give W. *pawb ny8 ; it seems to have been made into an improper compound early, and the aw treated like ordinary pen- ultimate aw (which normally comes from *ou) and affected to eu 76 iv (3), giving *peubnyS >peuny& ; then by analogy peunoeth (and S. W. dial, o beutu for lit. o boptu); Bret, bemdeiz, Treg. baonde.

(3) yn awr 'now' 114 iv ; yr awron, weithion, etc., 164 iii ; ymdnnos ' the other night' R.P. 1264, D.G. 82, 158, 200.

ymannos is probably to be placed here although the exact form of its Brit, original is doubtful. It stands for *ymannoeth which may represent loc. *esmi anda nokti lit. ' this here night ', see ii (i i). (4) pa le, pie ' where ? ' pa ddelw, pa fodd ' how ? ' pa bryd ' when ? ' etc. 163 ii.

(5) ry wbryd ' some time ', rywfodd ' somehow '.

v. Adverbs formed of a noun or adj. preceded by a conjunction or neg. part. :

(j) ond + noun or pron. : ond odid B.CW. 31 'perhaps' (lit. ' except a rarity ') 169 v (4) ; ond antur D.G. 266, G.Gr. D.G. 238 'almost', with neg. 'hardly' (lit. 'but by chance'); ond hynny ' any more ' IL.M. 94, 96, T. ii 1 76.

(2) nid + cpv. adj.: nid hwyrach I Cor. xvi 6 'perhaps'; nid gwaeth 'even' e.g. D.N. c. i 161, D.G. 410; nid amgen ' namely ' (lit. ' not otherwise ') Ml. W. nyt amgen.

It is curious that nid hwyrach is generally reduced to hwyrach in the recent period, though it survives as tw(yjrach in Gwyn. dial.

vi. Adverbs formed of nouns governed by prepositions :

(1) The prep, and noun compounded : ech-nos ' the night before last'; ech-doe 'the day before yesterday'; tran-noeth ' the following day ' ; tren-nydd ' the day after to-morrow ' ; tra-dwy ' the third day from to-day ' ; Ml. W. a-vory, W.M. 4, IL A. no, Mn. W. y-f6ry 'to-morrow'; yr-llynedd, er-llynedd 'last year ' ; 6-bry ' down ' ; &soes, eisioes ' already ', Ml. W. eissoes 'nevertheless'; g6r-moB, Late Mn. W. g6r-mod ' excessively'; adref IL.A. 109 ' homewards ', so in Mn. W.

ech-doe is an improper compound formed when *ech < *eks was a living prep. ; ech-nos is formed on its analogy, or is changed for an older *ech-noeth. On trannoelh, trennyB see 156 i (22); tra-dwy for *tar-dwy < *taros duuo ' beyond two [days] ' ; in such a phrase it is possible that the accent of *duuy might be on the -o, the original position ( : Skr. duva) ; and *duuo > *duui would give -dioy not *-deu 76 v (4) ; a-vory for *a8-vory < *ad marig-i (prob. loc. ; *ad takes loc. in Germ, also) ' to-morrow ' ; yr-llyneS < *per blidniian ace. of

• bleidorii ' year ' ; eisoes < 1 *es-i-oes ' ever ' (: oes ' age ') formed like

eiroet (4) ; cf. Fr. toujours ' nevertheless ' ; adref, an old compound, 99v( 4 ).

(2) The prep, and noun uncompounded, or forming improper compounds accented on the ultima : i fyny 'up', Ml. W.y iyny(b) 110 iv (3); i lawr 'down'; i waered 'down'; i mewn 'inside' 215 iii (i); i maes 'out', Ml. W. y mae* c.M. 58, R.M. 172, IL.A. 122, 1 66 ; o vywn IL.A. 166 ' inside ' ; o vaes ib. ' outside ' ; yn dl 'back', ar 61 'behind' 215 iii (6); ar hynt 'imme- diately' S.G. 274; oddi fyny 'from above', oddi lawr 'from below ', oddi mewn ' inside ' ; ymlaen ' in front ' 215 iii (10) ; ynghyd ' together', Ml. W. ygkyt W.M. 103, R.M. 75 (for which if gyt is oftenest found, see ib.), i gyd ' wholly ', Ml. W. y gyt 156 i (8); ar lied 'abroad', late ar led; ar frys 'hastily', rhag llaw 'henceforth', Ml. W. rac Haw R.P. 1418, dra-chefn 'backwards, over again' 214 iii; ymaith 'away', Ml. W. ymdeitk for earlier e ymdeith W.M. 2 ; i ffwrdd id.

i waered ; gwaered < *upo-ped-ret- ' under-foot-run ' ; i maes = Bret, emeaz, Corn, ernes < *ens magess- ' into field ' ; ar hynt : lynt ' way ' 63 iii (i) ; i ffwrdd \ffvrdd 140 ii.

(3) With the article : o'r blaen < formerly ' ; o'r neilltu f on one side', o'r herwydd ' on that account'.

(4) With an infixed pron. : o'i fron, f. o'i bron L.G.C. 122 ' throughout ', lit. ' from its breast ' ; in Late Mn. W T . with the art., o'r bron ' wholly ' (used in S.W., and mistaken by some recent N.W. writers for ymron, bron 'nearly' 215 iii (14) which is now used as an adv.) ; er-m-6ed ' during my time ', Ml. W. eirmoet R.P. 1259; er-i-6ed 'ever' 34 iii, Ml. W. eiryoet, eiroet; the form erioed with the 3rd sg. pron., 'during his time', was generalized, and of the forms with other persons only ermoed survived ; it is used in poetry down to the Early Mn. period, e.g. D.G. 22, L.G.C. 194. Ml. eir- is regular for eri- 70 ii ; in eirmoet it is due to the analogy of eiroet.

vii. Adverbs formed of adjectives governed by prepositions :

(1) ar fyrr B.cw. 18 'in short' ; ar hir D.G. 352 'for a long while'; ar iawn D.G. 5 'straight'; ar waeth R.G.D. 149 'in a worse state ' ; trwy deg ' fairly ', trwg deg neu hagr ' by fair [means] or foul ' ; trwy iawn ' by right ' ; wrth wir ' truly ' ; o fraidd 'scarcely', Ml. W. o vreib IL.A. 108, a-breib W.M. 131.

(2) * Any adj. following yn, as yn dda ' well ', yn well ' better ', yn ddrwg ' badly ', yn fawr ' greatly ', yn gam 'wrongly'. The adj. has the soft initial except when it is 11 or rh 111 i (i); but in many expressions forming improper compounds it has the nasal ; as ynghynt ' sooner ', ymhell ' far', ynghdm ' wrongly ', ynghudd ' secretly ' etc. 107 v (6). W. yn, Corn, yn, Ml. Bret, en, ent, Ir. in, ind < *en-do ; W. yn fawr = Ir. in mar. In Ir. the adj. was generally in the dat. ; and Zeuss ZE. 608-9 explained ind as the dat. of the definite article. This ex- planation has been widely received, and is repeated e.g. by Thurneysen Gr. 228. Against it may be urged: i. Other prepositions are similarly used in W., see above. -2. The prep. *en-do like *do governed the dat. 3. In Ir. co (Mn. Ir. go, W. pw 214 iv), which is syno- nymous with *endo. was often substituted for it, and has superseded it in Mn. Ir. 4. W. ymhell, etc., show that simple *en could be used as well as *en-do ; yn bell ' far ' and ymhell ' far ' are a doublet, both forms being in use ; ymhell is the same construction as ymlden where the yn, is a prep. 5. In W. leniting yn is also used to introduce the indefinite complement of verbs of being, becoming, making, etc., which makes it difficult for a speaker of the language to believe that leniting yn is the definite article. 6. The analogy not only of W. and Ir. but of other languages is all in favour of the prep., e.g. E. a-long, a-broad, etc.

(3) Special cases of comparatives after yn : yn hytrach ' rather', yn chwaethach W.M. 10 ' not to speak of, yghwaethach H.M. 85, ygkwaethach do. 150, agltwaethach do. 156, yg kyvoethach \V.M. p. 91 #, anoethach do. 182 ; also later chwaethach B.cw. 14.

hytrach is cpv. of hydr ' strong, prevailing' : O.Bret, hitr, Ir. sethar, of unknown origin. chwaethach (misspelt chweithach\>y Silvan Evans) is generally supposed to be from chwaith iii (5), e. g. D.D. s.v. ; if so it hns F-grade *-uok-t- ; -nchw- > -nhw- 26 vi (3) ; gk = nh 21 i ; an- < *n-do- : * 'en-do- ; yg kyv- seems to have pref. kyv- ; anoethach, with no pref., but with w lost before o 36 iii.

(4) Superlatives with the art. : o'r goreu ' very well ! ' o'r rhwyddaf Gr.O. 31 'most readily'; i'r eithaf ' extremely '; ar y cyntaf ' at first ' ; dial, ar y lleiaf ' rather too little ', ar y mwyaf ' rather too much '.

viii. (1) The prefixes rfiy-, go- and tra- by being accented separately before adjectives have come to be regarded as adverbs rhq, go, and tra ; thus rht[ (Ida ' too good ', go (Ida ' rather good ', tra da 'very good' 45 iv (2). See also 156 i (16), (ai), (22).

In the late period rhi{ is used as a noun ' excess ' for Ml. "W. rwy, as in Nyt gwell rwy no digawn E.B. 963 ' too much is not better than enough ' ; this is prob. the adv., ii (13), used as a noun ; rhwy adv.<

• prei (: *jrrai, Lat. prae) 210 x (5).

(2) lied and pur forming loose compounds with adjectives, 155 iv, are to the present linguistic consciousness adverbs; so prin in prin dclau Gr.O. 58 ' scarcely two ', etc. 221. Many adverbs are improper compounds formed of sentences fused into words. The following may be noted inW. :

i. (i) ysywaeth ' the more the pity ', Ml. W. ysywaeth IL.A. 157, s.G. 252, for ysy waeth ' which is worse '.

(2) gwaethiroeS duw C.M. 30 for gwaetk yr oeb duw(?) ' woe worth the day' ; Gwentian gwaitkiro dduw H.G. 106.

(3) yswaethe'roeS L.G.C. 38, seemingly a confusion of (i) and (2).

ii. (i) agatfydd Gr.O. 262, J.D.R. 134 'perhaps', Ml. W. agatvyb S.G. 224, ac atvyfc vr.M. 2, K.M. 2, for ag a atvyb ( with what will be ' i.e. per- ad venture ; cf. a advo B.B. 8 ' what may happen '.

(2) agattoeS H.M. ii 85 'it might be', ac attoeb K.M. 212, for ag a *ad-koeb ; for *tioeb see 180 ii (3).

(3) ysgatfydd ' perhaps ' i Cor. xv 37 for ys ag a atfydd.

iii. ysgwlr, 'sgwir L.G.C. 444 ' truly ', for yi gwir ' it is true ' ; malpei J.D.R. [xiv] ' as it were ; so to speak' for mat pel ' as it were ' ; sef c this is, that is, namely ', for y% ef.

iv. (i) llyma ' voici ', llyna' voila', for sytt yma 'see here', syll yna 'see there', cf. Bret, setu ' voici, voila' prob. for sellet Jiu ' see ye ' ; cf. syll dy racco E.M. 133.

(2) Mn. W. dyma ' voici ', more fully weldyma B.CW. 24, Late Ml. W. weldyma s.G. 221, for icel dy yma E.M. 58, wely dy yma W.M. 80 ' seest thou here ? ' So Mn. W. dyna ' voila ' for wel dy yna ? and Mn. W. dacw ' see yonder ' for wel dy raccw ? see 173 iii (3). Similarly ducho 'see up above', welducko for wel(y) dy ucho\ disc 'see below', weldiso D.G. 113, dial corr. dusw ; dyfry ' see up ', dobry ' see down ', dyngo ' see close by ' (yngo 210 viii (5)).

CONJUNCTIONS

222. The Welsh conjunctions are the following : i. Annexive : a, ac 'and', (i) The -c of ac is a survival of Ml. spelling 18 ii ; the word is sounded ag, and is treated as ag in cynghanedd, as seen by the correspondences marked below ; cf 111 v (4). In many Mn. MSS. it is written ay.

Ac yno ym medw Gwynedd

Imi ar bdr y mae'r bedd. D.G. 60.

' And there among the birch-trees of Gwynedd the grave is heing prepared for me.' Ag in the text here, but Ac in the previous couplet.

Ni thorrais un llythyren

O bin ao inc heb enw Grwen. D.N". M 136/147.

' I have not written one letter with pen and ink but Gwen's name.'

Am Fon yr ymofynnaf;

Mwnai ao aur Mon a g/. L.G.C. M 146/140.

' Mon will I seek ; I shall have the money and gold of Mon.'

(2) ac ( = ag) is used before vowels ; a [spir.] before consonants, including ^, and in Ml. and Early Mn. W. i ; as lara a chaws ; dwr a kalen.

Ni chwynaf od wyf afiach,

Os yfo sy fyw a iach. R.G.G. IL.B.M. 23.

' I shall not complain if I am ill, if he is alive and well.' The MS. has ag, which is usual in the late period before i ; but such combinations as ac haul sometimes seen in recent cynghanedd have no lit. or dial, justification, except perhaps in Gwentian where h is dropped. The same rules apply to a, ag ' with ' ; na, nac ' nor ' ; no, noc ' than '.

(3) ag : Ir. acus, accus, ocus; the Ir. -c- or -cc- represents -gg- as proved by Mn. Ir. -g- ; W. ag then represents *aggos ; the final -s and oxytone proved by the spirant initial which follows it ; the Ir. acus older occuis for *agguis < *aggos-ti. Brit. *aggos < *at-g}i6s formed of *at ( : *ei) 63 v (2) and a #A-suffix as in Gk. ' Si-xa, &-xou, 8i-xo-6ev, etc.

The base *at (: *et) is connected with *ati (: *eti) ' beyond ', whence ' and, but ' ; thus Lat. et, Umbr. et ' and ', Goth. ij> ' and, but ' < *et: Lat. at ' but ', Goth, ap-fian ' but ', Gk. dr-ap ' but ' < *at. The suffix -ghos is also seen in ag ' with ' 213 iii (i) ; and in agos ' near ', the base of which is probably *ad- ' to, near ' : Lat. ad, E. at ; thus *agos < Brit. *aggostos < *ad-ghos-to-s.

ii. Disjunctive : (i) neu [soft] ' or*.

neu < *n6ul < *ne-ue : Ir. no, no, nu < *ne-ue. The second ele- ment is Ar. ue ' or ' : Lat. -ve, Skr. va f or '. Thurneysen takes the first to be the neg. *ne- so that the orig. meaning was ' or not ' : Skr. nd-va ' or not '. But the development of the meaning is in that case not obvious. The *ne- may be the stem of the *eno-, *no- pronoun, as Gk. -ve in Thess. ro-ve ' rdSe ', Skr. na ' as ', Lat. ego-ne etc., of which the loc. is the affirmative part, neu 219 i (2); thus the original meaning would be ' or indeed, or rather '.

(2) Ml. W. ae . . . ae ' whether ... or; either ... or' ; Mn. W. ai . . , ai ; strengthened, naill ai. . . ai yntau.

ae [rad.] comes before a verbal noun, noun, adj., adv., or their equivalents, but not before a verb, cf. 218 i. A personal pron. after the second has the conjunctive form, minneu etc.

y ro8i dewis uBunt ae giorhau iSaw ae ymwan ac ef W.M. 160 'to give them [their] choice whether to do homage to him or to fight with him'; dewis ti ae o'th vo8 ae o'th anvo8 do. 124 'choose thou whether willingly or unwillingly ' ; ae tydi . . . ae titheudo. 162, 171, cf. 159 iii.

ae 218 iii, yntau 159 iii (2), iv (3).

(3) na, nac ' nor ' ; na(c) . . . na(c) ( neither . . . nor ' ; na [spir.] before a consonant, including- h and i ; nac before a vowel ; nac = naff ; exactly as for ac, see i above.

Er i gig ni rdi'r gegin

NaG er i groen garrai grin. G.G1. M I/DO. 43. ' The kitchen would not give for his flesh or for his skin a sear thong.' The MS. has actually nag, as is often the case; see i (i). nag < *naggos < *n(e) at-glws ' and not '.

iii. Adversative : (1) Mn. W. onid, ond [rad.] 44 vi ' but', Ml. W. onyt ; this is the form before a noun, etc., of ony ' if not ', v (i) below.

(2) eithr [rad.] \but', e.g. Act. iv 4, 15, 17, 19, 21 = prep. eilkr 214 v.

(3) namyn [rad.] ' but ', namn 44 vi, Ml. W. namyn, namen, namwyn, 78 ii (i) ; O.W. honit nammui ' but only'.

namyn os mivi a gdr yr amherawdyr, deuet lyt yman y'm hoi W.M. 1 86, cf. 185 'but if it is I that the emperor loves, let him come hither for me.'

Hael oedd, ac ni hawl iddi

Na'i main na'i haur, namyn hi. D.G. 293.

' He is chivalrous, and atks of her neither her jewels nor her gold, but only herself.'

namuyn, O. W. nammui, Ir. namda ' not more '. It is sometimes found without n-, by false division, as amyn B.CH. 16, amen A.L. i 288 1. 3. The example from D.G. shows how the meaning developed : ' not more [than] ' > ' ouly ' > ' but '.

(4) Ml. W. hagen ' however ', coming after the opening word or words of the sentence, and prob. an enclitic.

cam's rywdsti ef ; wynteu hagen ni wybuyssynt i eisseu ef W.M. 9 'for he had not seen them ; they, however, had not missed him ' ; mjt oeS nes hagen i&i no chynt do. 17 ' he was no nearer, however, to her than before '.

hagen, O. W. hacen M.c. gl. at ' hut ', Bret, hogen 'but ' (not enclitic). It has been suggested that the first part is identical with ac ' and ' (Loth. Voc. 150, Henry 165) ; as *at the base of ac also means ' but' 1(3) this is not improbable, but it is not easy to account for the form. O. W, lias ha, hac as well as a, ac, but the h- is not the aspirate, and is lost in Ml. W., 112 i. If, however, \ve suppose a cpv. in *-ison of

• aggos, its loc. *aggiseni would give *ag-hen, which by early metath.

of h ( 94 ii) might give hagen. For a similar cpv. cf. haeachen 220 iii (6) ; amgtn 148 ii (2).

iv. Causal : (i) canys [rad.] ' since ', cans 44 vi ; Ml. "W. can, kanyS) cans W.M. 487 ' since ' ; kan(n]y y han(n}yt, ' since . . . not ' ; kan(n}ys, canis iii (4) ' since . . . not . . . him (her, them) '.

ergliv wi ( = erglywji) can dothuif B.B. 75 'hear me since I have come'; kann colles HJ.A. 147 'since he has lost'; A chan derw yt fy/wednt y geir w.M. 21 ' and since thou hast said the word'. canys priflys oe8 do. 64 ' for it was the chief court' ; eisteS di yn y lie hwnn kanys tydi lieu S.G. 6 ' sit thou in this place for it is thou to whom it belongs'. Cany welas ef W.M. 16 ' since he did not see ' ; canyt oes vrenhin ar holl Annwvy-n namyn ti do. 8 ' for there is no king over all A. but thee'. canis, see iii (4) ; Kanys gwyBut K.M. 282 ' since thou didst not know it '. Later Kanys ny S.G. 17.

can is the same woid as the prep, gan 211 ii, iv (i) though possibly with a cons, ending, as it seems to take the rad. canys ' since ' = cann ys ' since it is ' and is often written kannys e.g. IL.A. 9, 10, 13, etc.; the -nn- is simplified because the word is generally unaccented; cf. anad for annat 214 viii. It rarely conies directly before a verb : cans oe8 W.M. 487 =kan oes R.M. 1 26. The neg. kany is for can ny ; it was pi ob. accented on the last s} 11., hence the simplifi- cation of the -nn-. The accent would suffice to distinguish kanys ' since . . not . . him ' from the positive kanys ' since '.

(2) achos ' because ', Ml. W. achaws.

Galw Gwrhyr Gwalltawt leithoeS, achaws yr holl ieitJioeS a wyoyat K.M. 114 'Gwrhyr Gwalstawt leithoedd was called, because he knew all languages'. The conj. is omitted in W.M. 471.

achos 65 ii (i), 215 ii (i). o achos is used before v.n.'s and noun-clauses, and so remains prepositional : Deut. i 36, iv 37, vii 12, Num. xxx 5.

(3) o ran 'for', 215 iii(ia).

Fob byw wrth i ryw yr aeth,

O ran taer yw'r naturiaeth. W.IL., C.IL. 73.

' Every living thing goes after its kind, for nature is insistent.' (4) Other composite nominal prepositions are used as con- junctions in the Late Ma. period: o blegid Act. i 5> " 345 o herwydd i Cor. xv 53 ; o waith, in S.W. dial, waif A.

v. Conditional: (i) o, od 'if, Ml. W. o, ot, or', os 'if it is' ; ossit 'if there is'; o'm 'if... me'; o'th 'if...thee'; os ' if . . . him (her, them) ' ; oni, onid ' if . . . not, unless ', Ml. W. ony, onyt ; oni 'm 'if ... not . . . me ', oni-s 'if ... not . . . him (her, them) ', Ml. W. onym, onyx, etc. As above indicated the -* of os is either y& 'is', or else the 3rd sg. or pi. infixed pron. ; but in Late Mn. W. os came to be used instead of o, od for ' if simply ; examples are common in the i6th cent. : os rhoed Haw W.IL. 60.—o is followed by the spirant, also in Early Mn. W. by the rad., of jo-, i- t <?-, and by the rad. of other mutables ; od is used before vowels.

Before verbs : o chlywy Siaspat . . . o gwely flws W.M. 1 19-1 20 ' if thou hearest a cry ... if thou seest a jewel ' ; o chat D.G. 30 ' if thou shalt get ' ; o ca/do. 20 ' if I get ' ; od ey W.M. 446 ' if thou goest ' ; ot agory do. 457 'if thou openest'; with infixed pronouns: O'TH lleSi D.G. 59 'if thou killest me ' ; o'th gaf do. 524 ' if I may have thee'; os canyhatta W.M. 412 'if she allows him [to go]'; with r(y) : or bu do. 172 'if there has been ' ; or kaffaf i-yyhyvarws do. 459 'if I get my boon' ; or mynny IL.A. 165 ' if thou wilt '. Before nouns, etc., followed by the relative pron., os ' if (it) is' : Ac os wynteu &e me8 hi W.M. 190 'and if it is they who hold it'; os oS (read o'th) vo8 y gwney ditheu do. 429 ' if it is of thy free will that thou dost ' ; or followed by a simple subject : os pechawt hynny IL.A. 38 ' if that is sin '. Ml. W. ossit before an indef. subject : ossit a Sigrifhao . . . C.M. 27 ' if there is [any one] who enjoys . . .' The neg. forms ony etc. follow the rules for ny; before verbs : ony by& W.M. 95 'if there be not ' ; with infixed pron. : onys kaffaf do. 459 ' if I do not get it '. Before nouns etc. onyt ' if it [is] not ' : onyt edivar IL.A. 47 ' if not repentant'. This form became onyt, later onid, oiid ' but ' ; ny Seuthum i yma onyt yr gwellau vy mwcfoS S.G. 184'! have not come here but to amend my life ' ; ny mynnaf-i neb onyt Duw do. 178'! desire no one but God '. Instead of OS ' if it is ' we find before a past tense or bu 'if it was' in W.M. 458 (modernized to os in K.M. 104) : or bu ar dy gam y dyvuost ' if it was at a walk that thou earnest '. For oni a new os na is used in Recent W.

o ' if '< Brit, *a ' if ' 218 iii ; on the form see 71 i (2). ot may represent *a-ti or *a-ta, see 162 vi (2), which survives only before vowels. But an old ot before a cons., in which the -t is an infixed pron., survives in the stereotyped phrase ot gwnn W.M. 12 'if I know it'; this may well be *a tod ' if it '. o* ' if it is ' < *d 'iti ; ossit ' if there is ' < *a 'stlta < *d 'sti ita. The mutation after accented *a was the same as after accented *ne, but made more regular owing to the word being of less frequent occurrence ; the rad. c- etc. seems to be due to further levelling.

(2) pel [rad.] 'if Late Mn. W. pe. The form pei is short for pei y ' were it that ' ; see 189 ii (3) ; the real conj. y, yt which follows pei is the citative eonj. ; see x (i). Before a noun there is, of course, no conj. after pei, which is then simply ' were it ' ; as pei mi rywascut velly W.M. 474 ' were it I that thou hadst squeezed so '.

pei ran S.G. 212 ' supposing that', cf. 256, 368, pei rhon D.G. 118, 271, 304, followed by a v.n. clause. The formation is not clear (*? pei rhoent ' if they granted ').

vi. Temporal: (i) pan(n) [soft] 'when', 162 iv (3), 163 vi ; sometimes dan, especially in poetry.

A phan 8oeth yno W.M. 8 ' and when he came there ' ; a phan welas do. 1 3 ' and when he saw ' ; pan gly whont do. 2 2 c when they hear '. Pa le V oeddit ti pan sylfaenais i y ddaear 1 Job xxxviii 4. Ban elom ni IL.A. 168 'when we go'.

Syrthiais, llewygais i'r llawr, Bann welais benn i elawr. T.A., G. 234. ' I fell, I fainted to the floor, when I saw the head of his bier.'

pan being relative a prep, may govern the antecedent, expressed as the r in o'r pan agoroch y drws W.M. 57 ' from the time when you open the door', but generally implied, as in erbyn pan do. 33 'by [the time] when', hyt pan do. 470 'until ', yr panda. 161, Mn. W, er pan 1 since '.

(2) tra ' whilst ' ; also hyd tra. It is usually followed by a soft initial ; tra parJiao W.M. 26 is a rare exception in Ml. W. In Late Mn. W. the rad. is common (sometimes by confusion with the prep, tra, the spir. e. g. Gr.O. 12).

ny ommeSwyt neb tra barhauft (read barhaa6S) W.M. 26 ' no one was refused while it [the feast] lasted ' ; tra gejfit do. 65-6, 68, 72 'while one could have ' ; tra vynho Duw do. 7 1 ' while God will ', tra welho Duw do. 7 2 id. ; tra gerSych W.DJ. 6 ' while thou walkest ' ; tra fyddai Matt, xiv 2 2, tr&fyddwyf Marc xiv 32 ; trajyddo haul Ps. Ixxii 1 7. hyt tra ym gatter yn vyw W.M. 479 ' whilst I am left alive ' ; hyt tra vei K.B.B. 79.

tra allied to the prep, tra, but coming from a Brit, form ending in a vowel, possibly *tare < *t e ri cf, *are- < *p f ri ; if so it is for *tar, see 214 iii.

(3) cyn [rad.] 'before' 215 i (i). It is used as a conj.

proper, coming immediately before a verb, see examples. In the recent period it is treated as the prep, by having y put after it.

kin bu tav y dan mein B.B. 68 ' before he was silent under stones ' ; kyn bum B.T. 25 ' before I was ' ; gwr a roteigad kyn dybu y dyt w. 2a ' a man who gave battle before his day came ' ; cyn elych s.G. 269.

Dduw I cyn el i ddaear,

A ddaw cof iddi a'i cdr ? B.Br., p. 112/264.

' God ! before he goes to earth will she remember [him] who loves her?'

(4) Ml. W. hyny, yny ' until ' ; Early Mn. W. yni ; Late Mn. W. only onid by confusion with oni v (i) ; and tauto- logically hyd oni.

A humiiw a 8yscawS Dewi hyny vu athro IL.A. 107 'And [it was] he who taught Dewi till he became a doctor ' ; A'r yny 8 a gerSassant hyny Soethant y Eryri W.M. 185 'And they traversed the island till they came to Eryri'; Ac yny agoroch y drws do. 57 'and until you open the door'; ynyveiyn llawn do. 56 ' until it was full'.

Ni ddof oddiwrth nai Ddafydd

Yni ddel y nos yn ddydd. L.G.C. 210.

' I will not come away from David's nephew till night becomes day.' onid oedd yr haul argyrraedd ei gaereuv.cw. 5 ' until the sun was reaching his battlements' i.e. setting; hyd oni Matt, ii 9.

hyny is for hyd ny, and appears in full in CP. : hit ni-ri-tarnher ir did hinnuith f until that day is completed '. hyd ny lit. ' while not ' ; the ' length ' (hyd) of time during which an event is ' not ' (ny) reached is the time ' until ' (hyny) it is reached.

(5) gwedy y, hyd y, etc., see xi.

vii. Concessive : (i) cyd [rad.] ' although ', Ml. W. kyf, ket, ki/Hy cen ; neg. kyn ny, kyny, keny.

kyt keffych hynny W.M. 480 ' though thou get that ' ; ket bei cann wr en vn ty B.A. 12 ' though there might be 100 men in one house ' ; Kyd carhuriv-e nwrva cassaav-e mor B.B. TOO ' though I love the strand I hate the sea '. Cyd byddai nifer meibion Israel fel tywod y mor Rhuf. ix 2 7 ; Cyd bai hirfaith taith or wlad hon yno Gr.O. 1 1 6 ' though a journey from this country thither would be long.' A chyn bei drut hynny B.M. 169 ' And though that was a brave [fight] ' ; A chyn bo W.M. 62. a chyn-nyt ymoialwyf a thiw.JA. 2 ' and though I may not avenge myself on thee ' ; kyn-ny bwyf arglwySes, mi a wnn beth yw hynny do. 5 1 ' though I am not a lady, I know what that is ' ; A chyny bei do. 62. 0. W. cen nit boi . . . Cinnit hois CP. ' though there be not . . . though there is not '.

cyd : Ir. ce, cla ' though ' ; cyny : Ir. cent, cini, cenl. The -d is to be compared with that of od ' if, see v (i) above ; as it is followed by the rad., cy-d may be for *ke tod ' if it ' a form which spread from kyt bo ' if it be ' etc. Before ny there was prob. no -d, and cyn ny is prob. a wrong deduction from cyny on the analogy of Jean ny iv (i) ; cyn before a positive verb spread from this. Traces of cy- without -d are found : ke-rei diffeith B.A. 7 ' though it were waste ' ; nyt arbedus ke-vei yr egluysseu G.c. 1 30 ' he spared not even the churches' ; Jcyffei B.B. 87. Kelt. *ke may be the stem of the *Jce- pronoun, as in Lat. ce-do ; loc. in Gk. e-/cei, Kel-@e.

(2) er na, see xi.

viii. Comparative : (i) cyn [soft] ' as ' before the equative ; see 147 iv (4).

(2) a [spir.], ag ' as ' after the equative, Ml. W. a, ac ; see i (2). This is the same word as a, ag ' with' ; see 213 iii (i). It is often found before cyn ' though \pei ' \$.\pan ( when'.

A chyn dristet oe8 bop dyn yno a chyn bei ayheu ym pop dyn onaSunt B.M. 188 ' And every man there was as sad as if death was in every man of them '.

(3) Ml. and Early Mn. W. no [spir.], noc f than ' after the cpv. ; Late Mn. W. na, naff ; see i (2). Also Ml. W. nogyt, noget, noc et ' than '. no chyn ' than if etc.

no chynt iii (4) ' than before ' ; ny wy&wn i varch gynt . . , no hwnnw W.M. 14 ' I knew no fleeter steed than that ' ; no hi do. 63 ' [he had not seen a more beautiful woman] than her'; no hwnnw do. 67 ' than that ' ; hyt na welsei oyn wenith tegach noc ef do. 7 3 ' so that no man had seen fairer wheat than it'. Tegach yw honno no neb D.G. 440 ' Fairer is she than any '. perach ac arafach nogyt y rei ereill IL.A. 101 ' sweeter and calmer than the others ' ; iawnach yw ioaw dy gynnhal nogyt ymi W.M. 37 ' it is juster for him to support thee than for me', cf. K.P. 1039, 11. 10, 30 ; Ny by8 hyn, ny byo ieu, noget y Becfvreu B.T. 36 ' it will not be older, it will not be younger, than at the beginning ', cf. 28.

The initial n- is the old ending of the cpv., see 147 iv(3); cf. Bret, eget, Corn, ages corresponding to W. nogyt. The remaining -o, -oc ( = -og) has the same formation as a, ac ' and ', i (3), and the spirant after o, as after a, implies the accent on the lost ult. Since unacc. d, and unacc. o before a guttural, both give a, we must refer our o to u- 66 v ; hence -oc < *uggos, which may be for *ud-gMs : Lith. uz- ' up ' < *ud-gh-, Ir. u- with gemination, Skr. ud- ' out, up ', Goth, ut, E. out', for meaning cf. E. out-shine. Ir. occ ace seems to be a mixture of *ud-g- and *ad-g- mostly with the meaning of the latter. The affixed particle -yt, -et is prob. *eti ' beyond ' i (3).

ix. Illative : yntau ' then, therefore ' in Late Mn. W. usually written y*tt; Ml. "W. ynttu ; 159 iii (2), iv (3). In this sense the word always comes after the opening word or words of the sentence.

Gimawn glot ynteu o'th draws gampev, B.P. 1219 ' Let us fashion praise, then, of thy feats of arms '.

x. Citative: (1) before verbs, y [rad.], yr 'that', Ml. W. y, (yd, yb). It is used to make a sentence into a noun equivalent not only after verbs of saying, believing, etc., as gicn y daw ef ' I know that he will come ', but generally where a noun-clause is needed, thus diau y date ef ' that he will come [is] certain '. The neg. form is na, nad, Ml. W. tta, nat.

ac a bywedassant y gucneynt yn yr un kyjfelyb s.G. 1 1 ' and they said that they would do likewise'; ac yn dywedut y'th UBir di do. 369 ' and saying that thou shalt be killed ' ; ac a wnn y car Duw ynteu IL.A.. 112* and I know that God loves him ' ; ef a wyddyat y collet ef do. 58 ' he knew that he would lose '.

Son fth gylch, oe hum a'lh gdi,

Ni thygasicn i'i/t gotcsai. T.A.A 14866/229.

' Saying about thee, if this man got thee, I should not have thought that he would have had thee.' On the spelling t see 82 ii (i).

The probable orig.meaning is 'how', so that yd may come from *io-ti,

• io- relative stem, *-ti suff. of manner 162 vi (2) : Gk. on. The

Skr. citative particle i-ti, coming generally after the quotation, is similarly formed from the demonstr. stem *i-. The mutation after it follows that of the oblique rel. in its other uses.

(2) Before nouns, etc. : Ml. W. panyw f that it is ', rarely before the impf. pan oeb ; and ymae, mae Mn. W. mae ' that it is', in the late period written mai 189ii (i) ; also dial. (S.W.) taw. Neg. Ml. nat, Mn. nad.

A bit honneit panyw bychydig a dal de&yf Duw y mywn Cristawn onis cwplaa C.M. 15 ' And be it known that it is little that the law of God avails in a Christian unless he performs it ' ; pann yw IL.A. 152, 160. Gwir yw ymae Duw a wnnaeth pob peth IL.A. 27 'It is tnie that it is God that made everything ' ; cf. do. 2 1 1. 1 3 ; Hyna ry attep i iii . . . ymae ti a Seicisettm W.M. 1 8 ' that is my answer to thee, that it is thou whom I would choose'; mae ti a iewisswn B.M. 12. ny wybyem pan oeS ti a grogem B.T. 1 2 ' we knew not that it was Thou whom we crucified '.

pan yw lit ' when it is ' ; to know ' when ' it is may as easily as to know how ' it is become to know ' that ' it is. ymae is doubtless relative = y mae ' where (it) is ', hence from *totmi est 189 iii (2). The loc. *jftsmi may mean ' how ' as well as ' where '.

xi. (i) A preposition governing the implied antecedent of an oblique rel. y (or neg. na) forms with the latter the equi- valent of a conjunction :

gwedy y5 lit. ' after [the time] when ', greedy yr, greedy jr, greedy na\ gwedy y is usually contracted to greedy \ MIL. W. reedy 'dd, reedy 'r, reedy.

gwedy yr efont o'r byt ftumn C.M. no 'after they go from this world ' ; gwedy y garffei car y alon B.B.B. 7 ' after he had conquered his enemies ' ; A guedy byryer ttcncer yndi W.M. 21 ' and after much has been thrown into it ' ; guedy na cheffit gcmthunt try do. 66 ' after it was not obtained from them '. WedyMd el y drydedd oe* L.G.C. 394 'After the third generation is gone'. With inf. prom gwedy as coUont IL.A- 167 ' after they have lost it '.

hyt yS, lyf y ' as far as, as long as'; hyt na ' as far as not ' > ' so that not ' ; Mn. W. kyd jr(r), tyd na.

hyt y sych gwynt, hyt y gvclych glow W.M. 459 ' as far as wind dries, and rain wets' ; cf. D.G. 2 ; hyt na W.M. 4, hyt nat do. 71.

gyt ac y * as soon as ' ; Mn. W. gyd ag y.

Ar hynny gyt ac y kyvodeg ef W.M. 52 "Thereupon as soon as he rose '. Ac val y gyt ac y do. 88, K.M. 64 ' And as soon as '.

am na ' because . . . not ' : er na ' though . . . not* ; eithyr na ' except that . . . not ' ; trwy y * so that ', lit. ' through [means] whereby ' ; Mn. W. am na y er na, and am y ' because f , ery * though'.

am na trybuum pan aeth W.M. 389 ' because I knew not when be went ' ; eithyr na tllynt SywedtU do. 56 ' except that they could not speak'; trwy y colletto IL.A. 143 ' so as to cause loss', trw yt W.M. 453.

mal y(8) { how, so that ', mal na(t) ' as if, so that . . . not ' ; megys y(S) ' as, so that ', megys na(t) ' as if, so that . . . not ' ; Mn. W. fal jr(r), fel y(r), . . . na(d) ; megy* jr(r), mfyi* J< r ).

val y gaUfi W.M. 13 ' as he could ', ral na teyptm do. 429 ' as if I knew not ', mal na trybuum do. 389 ' so that I knew not ' ; megys y dyweit yr ystori/a do. 165 'as the story says ' ; megys na E.B.B. 186 ' as if . . . not '.

(2) Similarly an adverb, or noun in an adverbial case, with the obi. rel. and forming its antecedent, as pryd y ' at the time when ', 220 iii (7) (a).

In the recent period, in imitation of these, y is sometimes written after conjunctions, as pan y delo or osy daw instead of pan ddelo or o(s) daw.

INTERJECTIONS

223. i. (i) The following interjections proper occur in Ml. W. : a passim ; ha R.M. 235 ; oy a W.M. 57, oi a do. 147, wy a w. 1200 ; oian a B.B. 52 if., hoian a do. 61-2 ; och B.B. 50, 91, W.M. 20; och a do. 170; ub do. 473; gwae R.P. 1150 1. 31, generally followed by the dat. ; haha W.M. 123 ; tprue ( = tprwy ?) K.P. 1277-8, Mn. W. trw (used in calling cattle).

(2) Many others occur in Mn. W. : o ; ust ' hush ' ; ffl ' fie ' (whence ff'iaidd ' loathsome '), later ffei, foil, by o, see ex. ; wflt ' fie ' ; hu, huw D.G. D. 148, used to lull a baby to sleep, later hwi (short proper diphth.), hwi\an ; dyt 'pooh', dyflyt D.N. j 9/230 (the /s in the MS., and the accent implied in the cynghanedd). D. 148 gives, in addition, hys, ho, he, hai, ochan, w, 'wb, wlan wfavb, waw, wew y ffw, whw, wi, haihow, haiwhw, hoho, Iw, oio, wichwach. Other forms are ow, pw t wchw, hai wchw, hwt, heng ; also twt ' pshaw ! ' ach, ych ' ugh ! ' and others.

Ffei o ieuenctid am ffo ;

Ni ffy henaint, flfei 'hono. S.T. p 313/2 12.

' Fie upon youth for fleeing ; old age will not flee, fie upon it.' [The MS. has o-ffei in line i and ohono in line 2.]

(3) gwae 78 ii (2). och 51 iii exc. (3) ; *-h, rounded after o- may have given the -ch, 26 vi. The diphthong oi does not appear elsewhere in Ml. W., and may be a survival of O. W. oi < *ai; the doublet wy < *di : Gk. a?. Interjections, like the forms of child- speech, are liable to continuous re-formation ; and a may be from original a (: Lat. a, etc.), which ought regularly to give *aw.

ii. Some interjections are followed by nouns or pronouns, expressed or implied, in the dat., as gwae vi K.M. 40 ' vae mihi ' ; Guae agaur a graun maur verthet B.B. 31 ' woe to the miser who hoards great riches' ; Gwae a gcfowy uw R.P. 1150 ' woe [to him] who offends God '. So, och ft D.G. 435 ; Och flnnau F.N. 90 ; also Och imi ib., Och ym D.G. 21 ; Ochan fi do. 38 ; dial, och a fl. Also, of course, by the vocative : Och Dduw G. 255, etc.

iii. An interjection proper is sometimes preceded by a numeral, as naw-och IL.G. R.P. 1306 ; wyth w&ejinnau G. 229 ; can' och ; naw wfft.

224. As in other languages, utterances of an interjectional character are made from other parts of speech, and from phrases and sentences, often mutilated.

i. Nouns, with or without adjuncts : (i) Duw e.g. W.HJ. 232 last line, Duw an(n)wyl Gr.O. 39 ; later by euphemism dyn and dyn annwyl.

(2) dydd da ' good day ', nos da ' good night ', etc. 212 iv.

(3) hawS amor R.P. 1310 ' good luck ! '; gwynfyd i . . Gr.O. 88 'joy to . . I'; gwyn fyd na . . D.W. 71 'would to heaven that . . !' (na on the anal, of na 171 ii (2)); diolch ' thanks ! '

haw8 amor /tor C. M.A. i 2056 shows that havodd-amawr I.G. 624 is a false archaism, amor < *ad-smor-, Vsmer- 'part' ( 156 i (13)), hence ' destiny, luck ' : Gk. /j.6po<;, pmpa ' lot, destiny ', Horn. Kara [Afjiolpav (/*ft- < *sm-), Kctcr/x.o/305 ' Svonyvos Hes. < *KaT-oyAopos.

(4) rhad arno ' a blessing upon him ! ' (usually sarcastic) ; yr achlod iddynt Gr.O. 200 ' fie upon them ! ' yr achlod iddo T. ii 1 94 ; druan ohono ' poor thing ! ' ; etc.

ii. Adjectives used adverbially, and other adverbial expres- sions: (i) da 'good!'; purion 'very well!'; truan 'alas!'; da di, da dithau, da chwi, da chwithau ' if you will be so good '.

(2) yn iach 'farewell 1 s e.g. 166 i; yn llawen W.M. 19 ' gladly ! with pleasure ! ' ; yn rhodd B.CW. 80, P.G.G. 17 ' pray ! '

(3) ymaith ' away ! ', adref D.G. 165 ' home ! ' hwnt 'avaunt! ' Ml. W. nachaf W.M. 73, 225 ' behold ! ', enachaf (e- =$-) M.A. ii 302, ynackaf do. 170; later written nycha D.G. 135.

ynachaf, perhaps ' *yonder ! ' a spv. of the stem from which yna is made, thus from *ena-Jc-s e mo- ; see 220 ii (6).

(4) er Mair D.G. 18; er Duw ib. ; ar f'enaid L.G.C. 223 ' by my soul ' ; etc. myn . . . / ym . . . / 214 ix, x. iii. Verbs : aro ' stop ! ', late aros ; adolwg ' pvay ! ', atolwg Ps. cxviii 25, for which the v.n. adolwyn 203 iv (2) is sometimes found.

Paid,- lor nefol, adolwyn,

fyd yn danllyd am dwyn. S.C. I.MSS. 291. ' Do not, heavenly Lord, I beseech thee, take me away in flames from the world '.

iv. Sentences : (i) henffych well 'hail' 190 i (i).

(2) Contracted into single words, and sometimes corrupt: dioer 34 iii ' by heaven ! ' for Duw a wyr ' God knows ' ; Late Mn. wele ' behold ! ' for a wely di ' dost thou see ? ' 16 iv (i), also wel 173 iii (3) ; llyma ' voici ' etc. 221 iv ; dyma ' voici ' for wely dy yma, etc., see ib. ; ysgwir ' truly ! ' do. iii ; ysywaeth etc. do. i.

Ysowaeth, nos o ayaf

Tm sy hwy no mis o haf. D.E. p 76/29, c 7/649. ' Alack ! a night of winter is longer to me than a month of summer.'

1. Misprinted y daw.
2. Cf. ryddnant 68 for ryd\nant ; the d doubled because the syll. is closed; see 54 i (3).