Celtic languages

Introduction.—The Celtic languages form one group of the Indo-European family of languages. As might be expected from their geographical distribution, they hold a position between the Italic and Teutonic groups. They are distinguished from these and other branches of the family by certain well-marked characteristics, the most notable of which are the loss of initial and inter-vocalic p, cf. Ir. athair with Lat. pater; Ir. lan, “full,” Welsh llawn, Breton leun, with Lat. plenus; Gaulish are-, “beside,” Ir. ar. Welsh, Breton ar, with Gr. περ?, παρ?; and the change of I. E. e to i, cf. Ir.fir, “true,” Welsh gwir, Breton gwir, Lat. verus. We may further mention that the I. E. labialized velar gv is represented by b, e.g. Ir. bo, “cow,” Welshbuwch, Gr. βο?ς, Sanskr. gaus; Ir. ben, “woman,” Gr. γυν?, whilst the medial aspirates bh, dh, gh result in simple voiced stops. I. E. sonant r and lbecome ri, li. Other distinctive features of the modern dialects are not found in Gaulish, partly owing to the character of the monuments. Such are the -ss- preterite and the fusion of simple prepositions with pronominal elements, e.g. Ir. fri-umm, “against me,” Welsh wrth-yf, Breton ouz-inn. The initial mutations which are so characteristic of the living languages did not arise until after the Romans had left Britain. The Celtic languages betray a surprising affinity with the Italic dialects. Indeed, these two groups seem to stand in a much closer relationship to one another than any other pair. As features common to both Celtic and Italic we may mention: (i) the gen. sing, ending -i of masc. and neut. stems in o; (2) verbal nouns in -tion; (3) theb- future; (4) the passive formation in -r.

The various Celtic dialects may be divided as follows:—(1) Gaulish; (2) Goidelic, including Irish, Scottish Gaelic, and Manx; (3) Brythonic, including Welsh, Breton and Cornish. Gaulish and Brythonic, like Oscan and Umbrian among the Italic dialects, change the I. E. labialized velar guttural qv to p, whilst the Goidelic dialects retain the qv which later gives up the labial 613element and becomes k, e.g. Gaulish petor-, “four,” Ir. cethir, Welsh petguar, Breton pevar, Lat. quattuor; Ir. cia, “who,” Welsh pwy, Lat. quis; Gaulish epo-, “horse,” Welsh eb-ol, Breton eb-eul, Ir. ech, Lat.equus. Several attempts have been made to prove the existence of Celtic dialects with qv on the continent. Forms containing p occur in the Coligny calendar, discovered in 1897, by the side of others with qv, a state of affairs not yet satisfactorily accounted for. The Rom tablets, discovered in 1898, have not been interpreted as yet, but p forms are found on them exclusively. In an excursus we shall deal with the language of the Picts.

No comprehensive handbook of the Celtic languages on the lines of Gröber’s Grundriss der romanischen Philologie or Paul’s Grundriss der germanischen Philologie was available in 1909. The reader may refer to Windisch’s article “Keltische Sprachen” in Ersch und Gruber’s Allgemeine Encyklopädie der Wissenschaften und Künste, and V. Tourneur, Esquisse d’une histoire des études celtiques (Liége, 1905; vol. ii. with full bibliography). Also H. Zimmer, “Die kelt. Litteraturen” in Die Kultur d. Gegenwart, T. i. Abh. xi. I, Berlin and Leipzig, 1909. The materials for the study of the older forms of the languages are to be found in Zeuss’s Grammatica Celtica as revised by Ebel. A comparative grammar of the Celtic dialects has been prepared by H. Pedersen (Göttingen, 1908). See also Whitley Stokes and A. Bezzenberger, Wortschatz der keltischen Spracheinheit (Göttingen, 1894).

I. Gaulish.—Celtic place-names are found as far east as the Dniester and Dobrudja, and as far north as Westphalia. The language of the Galatians in Asia Minor must have stood in a very close relation to Gaulish. Indeed few traces of dialectical differences are to be observed in continental Celtic. Unfortunately no literary monuments written in the ancient speech of Gaul have come down to us, though Caesar makes mention of religious poems orally transmitted by the Druids, and we also hear of bardi and vates. But a large number of personal and place-names have been preserved. The classical writers have, moreover, recorded a certain number of Gaulish words which can generally be identified without difficulty by comparing them with words still living in the modern dialects, e.g. pempedula, “cinquefoil,” cf. Welsh pump, “five,” and deilen, “leaf”; ambactus, Welsh amaeth;petorritum, “four-wheeled chariot,” cf. Welsh pedwar, “four,” and Ir. roth, “wheel,” or rith, “course.” We have further between thirty and forty inscriptions (three in north Italy) which we may without hesitation ascribe to the Gauls. These inscriptions are written in either N. Etruscan or Greek or Latin characters. We are thus in a position to reconstruct much of the old system of declension, which resembles Latin very closely on the one hand, and on the other represents the forms which are postulated by the O. Ir. paradigms. Hence Gaulish is particularly valuable as preserving the final vowels which have disappeared in early Irish and Welsh. The few verb-forms which occur in the remains of Gaulish are quite obscure and have not hitherto admitted of a satisfactory explanation. The statements of ancient authors with regard to the Belgae are conflicting, but there cannot be much doubt that the language of the latter was substantially the same as Gaulish. Caesar observes that there was little difference between the speech of the Gauls and the Britons in his day, and we may regard Gaulish as closely akin to the ancestor of the Brythonic dialects. It is difficult to say when Gaulish finally became extinct. It disappeared very rapidly in the south of France, but lingered on, possibly till the 6th century, in the northern districts, and it seems unnecessary to discredit Jerome’s statement that the speech of the Galatians in Asia Minor bore a strong resemblance to the language he had heard spoken in the neighbourhood of Trier. There is no evidence that Breton has been influenced by continental Celtic. The number of Gaulish words which have come down in the Romance languages is remarkably small, and though at first sight the sound-changes of French and Welsh seem to bear a strong likeness to one another, any influence of Gaulish pronunciation on French is largely discounted when we find the same changes occurring in other dialects where there is little or no question of Celtic influence.

The proper names occurring in classical writers, on inscriptions and coins, have been collected by A. Holder in his monumental Altceltischer Sprachschatz (Leipzig, 1896-1908). The inscriptions have been most recently treated by J. Rhys in the Proceedings of the British Academy, vol. ii. See also a paper in this volume entitled “Celtae and Galli” by the same author for the text of the Coligny and Rom inscriptions. The value of Gaulish for grammatical purposes is set forth by Whitley Stokes in a paper on “Celtic Declension” in the Proceedings of the London Philological Society (1885-1886). For the extent over which Gaulish was spoken, its relation to Latin and its influence on Romance, see E. Windisch’s article on “Keltische Sprache” in the section “Die vorromanischen Volkssprachen” in Gröber’s Grundriss der romanischen Philologie², vol. i. pp. 373 ff. Cf. further the introduction to J. Loth’s Chrestomathie bretonne (Paris, 1890); G. Dottin, Manuel pour servir à I’étude des antiquités celtiques (Paris, 1906); R. Thurneysen, Keltoromanisches (Halle, 1884).

II. Goidelic and Brythonic.—When the monuments of the Celtic dialects of the British Islands begin to appear, we find a wide divergence between the two groups. We can only mention some of the more important cases here. The Brythonic dialects have gone very much farther in giving up inflectional endings than Goidelic. In Irish all final syllables in general disappear except long vowels followed by s or r and u < o preceded by i. But these reservations do not hold good for Brythonic. Thus, whilst O. Irish possesses five cases the Brythonic dialects have only one, and they have further lost the neuter gender and the dual number in substantives. In phonology there are also very striking differences, apart from the treatment of the labialized velar qv already mentioned. The sonant n appears in Brythonic as an, whereas in Goidelic the nasal disappears before k, t with compensatory lengthening of the vowel, e.g. I. E. *kmtom, Ir. cét, “hundred,” W. cant, Bret. kant; Prim. Celt. *jov?ko-, O. Ir. óac, Mod. Ir. óg, “young,” W. ieuanc, Bret, iaouank. t, k standing after a vowel and preceding l, n (and also r if k precede) disappear in Goidelic with compensatory lengthening of the vowel, e.g. Prim. Celt. *statla-, Ir. sál, “heel,” W. sawdl; Prim. Celt. *petno-, Ir. én, “bird,” O. W. etn, Mod. W. edn. Similarly b, d,g disappear in Goidelic when standing after a vowel and preceding l, r, n with compensatory lengthening of the vowel, but in Welsh they produce a vowel forming a diphthong with the preceding vowel, e.g. Prim. Celt. *neblo-, Ir. nél, “cloud,” W. niwl; Prim. Celt. *ogno-, cf. Lat. agnus, Ir. uan, “lamb,” from *on, W. oen; Prim. Celt. *vegno-, cf. Ger. Wagen, Ir. fén, “wagon,” O. W. guein, Mod. W. gwain. The Goidelic dialects have preserved the vowels of accented syllables on the whole better than Brythonic. Thus Brythonic has changed Prim. Celt, a (= I. E. a, o) to o (W. aw, Bret. eu); and Prim. Celt. u to i, e.g. Ir. bráthir, “brother,” W. brawd, Bret. breur; Gaulish dunum, Ir. dún, “fort,” W. din. Already in Gaulish the I. E. diphthongs show a tendency to become simple long vowels and the latter are treated differently by Goidelic and Brythonic. In early times I. E. eu, ou both became o and I. E. ei gave e. In Goidelic o, e, in accented syllables were diphthongized in the early part of the 8th century to ua, ia if the next syllable did not contain the vowels e or i, whereas in Brythonic o gave ü (written u) and e became in W. ui (wy), and in Bret. oe (oue), e.g. GaulishTeuto-, Toutius, Ir. tuath, “people,” W., Bret. tud; Brythonic Leto-cetum, Ir. tuath, “grey,” W. llwyd, Bret, loued. Similarly in loan-words, Ir. céir, fial, W. cwyr, O. Corn. guil, from Lat. cera, velum. Further I. E. ai, oi are preserved in Irish as ai (ae), oi (oe), Mod. Ir. ao, but in Welsh I. E. ai gave eitherai or oe, whilst oi changed to ü (written u), Ir. toeb, “side,” W., Bret. tu; I. E. *oinos, Ir. óen, “one,” W., Bret. un; Prim. Celt. *saitlo-, cf. Lat.saeculum, W. hoedl, “age,” Bret. hoal. In Goidelic accented e changes to i before i, u in the following syllable, cf. Ir. fid, “wood,” gen. sing, fedo, O. H. G. witu, and i changes to e before a or o under similar conditions. In like manner u becomes o before a or o, whilst o changes to u before i, u, cf. Ir. muir, “sea,” Prim. Celt. *mori, gen. sing. mora. Of Brythonic finals which disappear, a, i, (o), j alone influence preceding vowels, whilst an i (y) which received the stress in O. W. was also able to modify vowels which went before it. In Goidelic the combinations sqv, sv appear respectively assc, s (medially, f), but in Brythonic they both give chw; Prim. Celt. *sqvetlon, Ir. scél, “story,” W. chwedl; Prim. Celt. *svesor, Ir. siur, “sister,” but mo fiur, “my sister” (whence Scottish piuthar by false de-aspiration), W. chwaer, Bret. c’hoar. In 614Brythonic initial s becomes h in the 7th century, but this is unknown in Goidelic, e.g. Ir. salann, “salt,” W. halen, Cornish haloin, Bret, holenn; Lat. sé-men, Ir. sil, “seed,” W. hil. Initial v gives f in Goidelic in the course of the 7th century, whereas in Brythonic it appears as gu, gw, cf. Lat. vérus, Ir. fir, W., Bret. gwir. We may also mention that in Goidelic initial j and medial v disappear, e.g. Gaulish Jovincillus, W. ieuanc, “young,” Bret, iouank, Ir. óac, óc; W. bywyd, “food,” Ir. biad. Post-consonantic j in Brythonic sometimes gives -id (Mod. W. -ydd, Mod. Bret, -ez), e.g. Gaulish nevio-, novio-, O. Bret, nowid, W. newydd, Bret, nevez, Ir. núe. I.E. -kt and -pt both appear in Goidelic as -cht but in Brythonic as -ith, cf. Lat. septem, O. Ir. secht, W. seith, Bret. seiz.

We unfortunately know very little about the position of the stress in ancient Gaulish. According to Meyer-Lübke in place-names the penult was accented if the vowel was long, otherwise the stress lay on the preceding syllable, e.g. Augustodunum, O. Fr. Ostedun, now Autun; Cataláunos(Châlons), Trícasses (Fr. Troyes), Bitûriges (Fr. Bourges). In Goidelic the stress, which is strongly expiratory, is always placed on the first syllable except in certain cases in verbs compounded with prepositional prefixes. In Old Welsh and Old Breton, on the other hand, the final syllable, i.e. the primitive penult, received the stress, but in both languages the stress was shifted in the middle period to the penultimate. The Goidelic dialects, like the Slavonic, distinguish between palatalized and nonpalatalized consonants, according as the consonant was originally followed by a front (e, i) or back vowel (a, o, u), a phenomenon which is entirely unknown to Brythonic.

Finally, the two groups differ radically in the matter of initial mutation or, as it is often called, aspiration. These mutations are by no means confined to initial consonants, as precisely the same changes have taken place under similar conditions in the interior of words. The Goidelic changes included under this head probably took place for the most part between the 5th and 7th centuries, whilst in Brythonic the process seems to have begun and continued later. It is easier to fix the date of the changes in Brythonic than in Goidelic, as a number of British names are preserved in lives of saints, and it is possible to draw conclusions from the shape that British place-names assumed in the mouths of the Anglo-Saxons. In Goidelic, we find two mutations, the vocalic and the nasal. Initial mutation only takes place between words which belong together syntactically, and which form one single stress-group, thus between article, numeral, possessive pronoun or preposition, and a following substantive; between a verbal prefix and the verb itself.

1. When the word causing mutation ended in a vowel we get the vocalic mutation, called by Irish grammarians aspiration. The sounds affected are the tenues k (c), t, p; the mediae g, d, b; the liquids and nasals m, n, r, l, s, and Prim. Celt. v (Ir. f, W. gw). At the present day the results of this mutation in Irish and Welsh may be tabulated as follows. Where the sound is at variance with the traditional orthography, the latter is given in brackets. In the case of n, r, l in Goidelic we get a different variety of n, r, l sound. In Welsh in the case of r, l, the absolute initial is a voiceless r, lwritten rh, ll, which on mutation become voiced and are written r, l. In Irish s becomes h written sh and the mutation of f is written fh, which, however, is now silent. Examples:—Irish, cú, “hound,” do chú, “thy hound”; Welsh ci, dy gi (do, dy represent a Prim. Celt. *tovo); Irish máthair, “mother,” an mháthair, “the mother,” Welsh mam, y fam (the feminine of the article was originally *senta, senda).

Original sound	k	t	p	g	d	b	m
Irish	χ(ch)	h(th)	f(ph)	?(gh)	?(dh)	v,w(bh)	v,w(mh)
Welsh	g	d	b	nil	ð(dd)	v(f)	v(f)

2. When the word causing mutation originally ended in a nasal, we get the nasal mutation called by Irish grammarians eclipse. The sounds affected are k (c), t, p; g, d, b; Prim. Celt. v (Ir. f, W. gw). In mod. Irish and mod. Welsh the results are tabulated below. Irish f becomes w written bh, whilst W. gw gives ngw. Examples:—Irish bliadhna, “year,” seacht m-bliadhna, “seven years,” cf. Latin septem, Welsh blynedd, saith mlynedd; Irish tir, “country,” i d-tir, “in a country,” Welsh tref, “town,” yn nhref, “in a town,” cf. Latin in.

Original Sound	k	t	p	g	d	b
Irish	g	d	b	ng	n	m
Welsh	ngh	nh	mh	ng	n	m

3. In Welsh k (c), t, p undergo a further change when the word causing mutation originally ended in s. There is nothing corresponding to this consonantal mutation in Goidelic. In this case k (c), t, p become the spirants χ (ch), th, f (ph), e.g. lad, “father,” ei thad, “her father,” ei represents a primitive *esias. In the interior of words in Brythonic, cc, pp, tt give the same result as initial k, t, p by this mutation.

The relation in which the other Celtic dialects stand to this system will be mentioned below in dealing with the various languages. It will be noted from what has been said above that, with the exception of the different treatment of the labialized velar qv, and the nasal sonant ?, the features which differentiate the Brythonic from the Goidelic dialects first appear for the most part after the Romans had left Britain. At the beginning of the Christian era the difference between the two groups can only have been very slight. And Strachan has shown recently that Old Irish and Old Welsh agree in a very striking manner in the use of the verbal particle ro and in other syntactical peculiarities connected with the verb.

(i.) Goidelic.—The term Goidelic is used to embrace the Celtic dialects of Ireland, Scotland and the Isle of Man. In each case the national name for the speech is Gaelic (Ir. Gaedhlig, Scottish Gaidhlig, Manx Gailck), from Ir. Scottish Gaodhal, Gaedheal, Mid. Ir. Góedel, W. Gwyddel, “a Gael, inhabitant of Ireland or Scotland.” Old Irish may be regarded as the ancestor of Scottish and Manx Gaelic, as the forms of these dialects can be traced back to Old Irish, and there are practically no monuments of Scottish and Manx in the oldest period. Scottish and Irish may be regarded as standing to one another in much the same relation as broad Scottish and southern English. The divergences of Scottish and Manx from Irish will be mentioned below. The language of the Ogam inscriptions is the oldest form of Goidelic with which we are acquainted. Some 300 inscriptions have up to the present been discovered in this alphabet, the majority of them hailing from the south-west of Ireland (Kerry and Cork). In Scotland 22 are known, whilst in England and Wales about 30 have turned up. Most of the latter are in South Wales, but odd ones have been found in North Wales, Devon and Cornwall, and one has occurred as far east as Hampshire. The Isle of Man also possesses two. The letters in the oldest inscriptions are formed by strokes or notches scored on either side of the edge of an upright stone. Thus we obtain the following alphabet:—

This system, which was eked out with other signs, would seem to have been framed in the south-west of Ireland by a person or persons who were familiar with the Latin alphabet. Some of the inscriptions probably go back to the 5th century and may even be earlier. As illustrations of the simplest forms of Ogam inscriptions we may mention the following: Doveti maqqi Cattini, i.e. “(the stone) of Dovetos son of Cattinos”; Trenagusu Maqi Maqi-Treni is rendered in Latin Trenegussi Fili Macutreni hic jacit; Sagramni Maqi Cunatami, “(the stone) of Sagramnos son of Cunotamos”;Ovanos avi Ivacattos, “(the stone) of Ovanus descendant of Ivacattus.” It will be seen that in the oldest of these inscriptions q is still kept apart from k(c), and that the final syllables have not disappeared (cf. maqqi, O. Ir. maicc), but it appears certain that in Ogamic writing stereotyped forms were used long after they had disappeared in ordinary speech. Several stones contain bilingual inscriptions, but the key to the Ogam alphabet is supplied by a treatise on Ogamic writing contained in the Book of Ballymote, a manuscript of the late 14th century. It should be mentioned that the Welsh 615stones are early whilst the Scottish ones are almost without exception late, and several of the latter have so far defied interpretation. In addition to the Irish Ogams there are a number of Christian inscriptions in Latin character, but, with one exception, they are not older than the 8th century.

See R.R. Brash, The Ogam Inscribed Monuments of the Gaedhil (London, 1879); R.A. Stewart Macalister, Studies in Irish Epigraphy (London), vol. i. (1897), vol. ii. 1902, vol. iii. 1907. The Welsh inscriptions are contained in J. Rhys, Lectures on Welsh Philology² (London, 1879). The Scottish stones have also been treated by Rhys in the Proceedings of the Scottish Society of Antiquaries (Edinburgh, 1892). See also G.M. Atkinson for the tract in the Book of Ballymote, Kilkenny Journal of Archaeology (1874). The Irish Christian inscriptions were published by Margaret Stokes as the annual volumes of the Roy. Hist, and Archaeol. Association of Ireland (1870-1877), and have been republished by R.A. Stewart Macalister.

(a) Irish.—We are able to trace the history of the Irish language continuously for a period of 1200 years, and from the time that the literary documents begin we are better supplied with linguistic material for the study of the language than is the case with any other Celtic dialect. At the same time that form of Irish which is to be found in the oldest documents has preserved a number of features which have entirely, or almost entirely, disappeared from the Brythonic languages. For this reason scholars have largely occupied themselves with Irish, which for purposes of comparative philology may be regarded as the classic Celtic language.

The history of Irish is divided into three periods:—Old Irish (700-1100), the documents mainly representing the language of the 8th and 9th centuries; Middle Irish, extending roughly from 1100 to 1550; Modern Irish from 1550 to the present day. These periods merge into one another to such an extent that no firm division can be made. The language of some manuscripts of the 14th century contains forms which are really Old Irish, and Middle Irish orthography was partly employed by historians and antiquarians in the middle of the 17th century. Old Irish, as compared with Brythonic, preserves a wealth of inflectional forms in declension and conjugation, but many of these tend to disappear very early. In the modern dialects of Ireland and Scotland there is a rigid rule of orthography that a palatalized, or, as it is termed, slender consonant in medial or final position, must be preceded by a palatal vowel (i), and a non-palatalized consonant by a non-palatal or broad vowel (a, o, u). This is the famous rule of the grammarians known as caol le caol agus leathan le leathan (“slender to slender and broad to broad”), but it is not so strictly adhered to in the spoken language as is commonly stated. In the older language the quality of medial and final consonants is only denoted very imperfectly, thus non-palatalized final consonants are regularly not denoted as such, e.g. O. and Mid. Ir. fir, Mod. Ir. fior. In Old and Mid. Irish the initial mutations are only regularly denoted in the case of the vocalic mutation of c, p, t, s, f, and the nasal mutation of b, d, g. The vocalic mutation of c, p, t, s, f was denoted by writing ch, ph, th, sh, fh, the first three symbols of which were derived from the Latin alphabet. Another method of denoting the mutation was to write a dot over the letter, originally the punctum delens, which was justified in the case of mutated f as the latter early became silent. But no such devices were ready at hand in the case of the medial b, d, g, and the mutated forms of these consonants were consequently not represented at all in the orthography. The same remark holds good in the case of the nasal mutation (eclipse) of the tenues. But it is easy to demonstrate that the same condition of affairs as we find in the modern language must have obtained in Old Irish. This insufficiency of symbols renders the orthography of the early stages of the language very complicated. We find that b, d, g were used initially to denote the voiced stops, but medially and finally they represent spirants, the voiced stops in this case being denoted by c, p, t. It is not until much later times that the h in the mutated forms of the tenues, or the use of the dot, was extended to the mediae. Thus in Mid. Irish we find do bochtaib in choimded (Mod. Ir.dobhochtaibh), Mid. Ir. ro-gab = Mod. Ir. do ghabh. The nasal mutation of c, p, t was first denoted by writing these sounds double and finally in the 18th century by writing gc, bp, dt. The spirants arising out of Prim. Celt. g, d, b came in Old Irish to be confused with those which developed out of Prim. Celt, p, t, k, in other than initial positions. In final positions in polysyllables we commonly find d and b written but medially th and ph, e.g. didnad, “consolation,” gen. sing, dithnatha. For the ending -ad cp. Lat. -atu-. On the other hand we find g written medially and ch finally. These rules, however, are not yet applied in the oldest documents.

When we turn to the inflections we find that most of the old terminations have disappeared, but that their influence on preceding consonants is still felt and serves to distinguish one form from another; thus in the declension of fer, “man,” nom. sing. fer, gen. sing. fir, dat. sing, fiur, acc. sing,fer n-, nom. pl. fir, gen. pl. fer n-, corresponding to Prim. Celt. (Gaulish) viros, viri, viro, viron, viri, viron, the influence of the following sound still differentiates the cases from one another. In the later language the initial mutations come more and more to be used for this purpose. In Middle Irish the declensions and conjugations are much simplified and the neuter gender is given up in substantives. In the verb the athematic conjugation has disappeared and the distinction of primary and secondary endings is not observed. On the other hand Irish has developed a peculiar system of absolute and conjoint inflection with different sets of endings. The conjoint endings are always used in the case of compound verbs, and in simple verbs they are employed after certain proclitics, e.g. the negative particles. Thus berid, “he bears,” is an absolute form; do-beir, “he gives,” ni beir, “he does not bear,” are conjoint forms. Further, the verb system is partly dominated by the various devices employed to express relatival function. There are three main types of conjugation in Old Irish corresponding to the Latin first, third and fourth conjugations, the Latin types moneo andaudio being difficult to distinguish in Irish. In the modern language there is in reality but one conjugation. The old Irish verb system comprises present and imperfect indicative, imperative, pres. subjunctive in -a- or -s- with corresponding past subjunctive, future in -f- or -s- or -e- or with reduplication along with corresponding secondary future, -s- preterite, -t- preterite, reduplicated preterite, a preterite containing a long stem-vowel, together with deponential and passive forms in -rd. This system is eked out with the verbal prefix ro, which among other functions changes a preterite into a perfect or a present into a perfect. Such a cumbrous system was bound to fall to pieces. A number of isolated forms have come down, but the only tenses which have survived into the modern period are the present and imperfect indicative, the imperative, the present subjunctive, the -s- preterite, the -b- and -e- future with corresponding secondary forms, and some of the passive forms in -r. At the same time in the modern language there is an increasing tendency to use analytical forms. Two noteworthy features of the Irish verb remain to be mentioned. The one is the use of pronouns as objects infixed between particle and verb, or in a verb compounded with a preposition between preposition and verb. There are two sets of forms according as to whether the verb occurs in a relative clause or not. Thus -m- is the ordinary infixed pronoun of the 1st pers. sing., whilst -dom- is the corresponding relative form. In the 3rd pers. sing. aspiration may be employed, e.g. ní ceil, “he does not hide,” ní cheil, “he does not hide it.” This has been given up in the modern language. Secondly in verbs compounded with prepositions the accent of the verb varies according as to whether the verb is used enclitically or not—thus after the negative ní or in the infinitive and imperative. Hence we have do-béir, “he gives,” by the side of ní tábair, “he does not give,” infin. tabairt; do-gníu, “I do,” ní d?nim, “I do not do,” infin. d?num. The changes caused by this alternation in addition to others due to the working of the Irish accent and to the initial and internal mutations have played havoc with the verb system and render it exceedingly difficult to reconstruct the paradigms. In the later periods of the language analogy naturally plays a great part, and many of the complicated forms are done away with, but even in the modern dialects the alternation between enclitic and orthotonic forms still survives in the commonest verbs, e.g. Irish bheir sé “he gives,” ní thabhair sé, “he does not give,” infin. tabhairt; Scottish bheir e, cha toir, toirt; Manx ver eh, cha der, coyrt; Irish ní sé, “he does,” ní dheanann sé, “he does not do,” infin. deanamh; Scottish nì e, “he does,” cha dean e, “he will not do,” infin. deanamh; Manx nee eh, cha jean eh, jannoo.

In the early period Irish borrowed a number of words from Latin. These are mainly connected with the church or with articles of civilization which would be imported from Roman Britain. Some of these show traces of British pronunciation, e.g. O. Ir. trindóit, from Latin trinitatem with ofor a. In others again Lat. p is represented in Ir. by c, which may be due to the substitution of q as being the nearest Irish sound to the foreign p. Thus we find Ir. corcur, “purple,” casc, “Easter”; cenciges, “Whitsuntide”; cruimther, “presbyter.” In addition to these several loans were received from Norse. In the Mid. Irish period many French words came in, and during the middle and modern periods the number of English words introduced is legion. Pedersen has tried to show in his Vergl. Gramm. that a considerable number of words were borrowed from Brythonic (Welsh) at an early date.

[For the Latin loan-words, see J. Vendryès, De hibernicis vocabulis quae a latina lingua originem duxerunt (Paris, 1902); Kuno Meyer has collected a number of loan-words from Norse, Anglo-Saxon, Early English, Latin and Early French in Revue celtique, xii. 460 and xiii. 505. See also Whitley Stokes, Bezzenberger’s Beiträge, xviii. 56 ff. For Celtic names in Norse see W. Stokes, Revue celtique, iii. 186 ff., and W.A. Craigie,Zeitschr. f. celt. Phil. i. 439 ff.]

With regard to the dialects of Irish, there is a well-known rhyme which states the peculiarities of the speech of the four provinces, and dialectical differences must have existed at an early period, though they do not make their appearance in the literary language until the 18th century. At the present day the Irish of Leinster has vanished entirely, and we have unfortunately no records of it. But in the other three provinces the vernacular still lives, and we find the Irish of Munster, Connaught and Ulster marked off from one another by well-defined peculiarities. In general it may be stated that the south of Ireland is more conservative than the north. In Munster there is a tendency to shift the word-stress from the initial syllable to a heavy derivative syllable, e.g. -an. This does 616not take place in Connaught, whilst in Ulster the tendency is to shorten the vowel. Again in monosyllables ending in ll, nn, m, and under certain other conditions a short vowel becomes a diphthong in the south, in Connaught it is merely lengthened, but in Ulster the original length is retained, e.g. Ulster ball, “member, limb,” Connaught ball, Munster baull. Final dh, gh in Munster are sounded as g. In certain cases the north prefers the vocalic mutation where the west and south have the nasal, thus notably in the dative singular after preposition and article, e.g. Munster-Connaught do’n bhfear, “to the man,” Ulster do’n fhear. In the south synthetic verb-forms are employed to a much larger extent than in the north.

In the early part of the 19th century Irish was still the speech of more than half the inhabitants of Ireland. A German traveller reckoned that out of a total population of seven millions in 1835 four millions spoke Irish as their mother-tongue. The famine of 1846-1847 was felt most in those districts that were purely Irish, and these were the parts that were and still are chiefly affected by the tide of emigration. Add to this the fact that the influence of O’Connell and his satellites, and above all that of the Roman Catholic clergy, was against the language. In spite of the efforts of the Gaelic League (founded 1893), which have met with considerable success, the language is rapidly dying of internal decay. The speakers of Irish are chiefly confined to the following counties, where over 20% of the population speak Gaelic:—Waterford, Cork, Kerry, Clare, Galway, Mayo, Sligo, Donegal. The following figures will illustrate the decay of the language since the famine:—

Year.	Monoglots.	Bilinguists.
1851	319,602	1,204,684
1861	163,275	942,261
1871	103,562	714,313
1881	64,167	885,765
1891	38,192	642,053
1901	20,953	620,189

According to the 1901 census report the speakers of Irish were distributed as follows:—Leinster, 26,436; Munster, 276,268; Connaught, 245,580; Ulster, 92,858. The Gaelic movement, which has thriven largely on account of its anti-English character, would have a much better chance of galvanizing the ancient language of Ireland if it were not for the supreme difficulties of Irish spelling and phonetics. Of the hundreds of thousands of persons who attend the classes of the League not more than one or two per cent. at the outside arrive at any state of proficiency. Presbyterian Gaels in Scotland are taught to read the Bible but Irish Catholics are not encouraged to do so. The result of this is seen in the fact that, whilst many, if not all, of the local Nationalist newspapers under the pressure of the League publish badly-printed and little-read columns in Irish, there are only two regularly appearing periodicals which contain any large amount of Irish. Half the contents—and those the most important—of the weekly organ of the league, An Claidheamh Soluis (“the flaming sword”), are in English. The latter was started in 1898 under the title of Fáinne an Lae (“the ring of day,” i.e. the dawn). The other periodical is the monthly Gaelic Journal (Irisleabhar na Gaedhilge), a would-be literary magazine of very inferior quality which has led a precarious existence since 1882. In 1898 it was decided to hold a festival called the Oireachtas (“hosting, gathering”) on the lines of the Welsh Eisteddfod. The venture was a great success and similar meetings have been held every year since, whilst each province and many of the counties have their annual local Gaelic feis (festival). The literary output of the movement has been prodigious, consisting in the main of a number of short stories and dramas (mostly propagandist), but nothing of any particular merit has as yet been forthcoming. The best-known writers are Dr Douglas Hyde (collector of folk-stories—Beside the Fire, 1890, An Sgeulaidhe Gaedhealach, 1895 (reprinted from vol. x. of theAnnales de Bretagne), Love Songs of Connaught, 1893, Religious Songs of Connaught, 1905); P. O’Leary (author of two lengthy stories, Seadna, 1904, Niamh, 1907); P. Dinneen (author of an historical tale, Cormac Ua Connaill, 1901); P. O’Shea, better known as “Conan Maol,” author of a collection of short stories entitled An Buaiceas, 1903.

Authorities on Irish Language.—For the study of Old Irish—Zeuss, Grammatica Celtica2 (Berlin, 1871); B. Güterbock and R. Thurneysen, Indices to the Irish words treated in Zeuss (Leipzig, 1881); E. Windisch published the first grammar of Old Irish in 1879 (trans. by N. Moore, Pitt Press, 1882), but Windisch’s treatment of the verb was rendered obsolete by the discovery of the laws of the Irish accent by H. Zimmer, Keltische Studien (Berlin, 1884), and R. Thurneysen, Revue celtique, vi. 309, J. Vendrèys, Grammaire du Vieil-Irlandais (Paris, 1908); R. Thurneysen,Handbuch des Alt-Irischen (Heidelberg, 1909). Mention should also be made of J. Strachan, Selections from the Old Irish Glosses (Dublin, 1904); and the same writer’s Old Irish Paradigms (Dublin, 1905), Stories from the Táin (Dublin, 1908). See also various papers on the Irish verb in theTransactions of the London Philological Society by Strachan (1895-1902); H. Pedersen, Aspirationen i Irsk (Copenhagen, 1898); C. Sarauw, Irske Studier (Copenhagen, 1901); G.J. Ascoli, Archivio glottologico italiano, vols. v. and vi. For the study of Middle Irish—E. Windisch, Irische Texte mit Wörterbuch (Leipzig, 1880). (Other volumes in conjunction with W. Stokes.)

Editions of texts by W. Stokes, Kuno Meyer and others in the Revue celtique, Zeitschrift für celtische Philologie, Ériu. K. Meyer has issued an exhaustive Mid. Irish glossary (A-D) as a supplement to the Archiv für celtische Lexikographie. The remainder is being published under the auspices of the Royal Irish Academy. The first grammar of Modern Irish was published by Francis Molloy in 1677 at Rome under the title ofGrammatica Latino-Hibernica. Molloy was followed by Jeremiah Curtin in 1728 with a book called Elements of the Irish Language. Numerous other grammars were published towards the end of the 18th and at the beginning of the 19th century, but few of them have any value. The more important of them are enumerated in the introduction to O’Donovan’s Grammar and to Windisch’s Kurzgefasste irische Grammatik, and in Pedersen’s Aspirationen i Irsk, pp. 29-47. We may mention W. Neilson’s Grammar (1808) as it is important for the Irish of E. Ulster. But the greatest native grammarian was John O’Donovan, who traversed Ireland in connexion with the Ordnance Survey, and published in 1854 a comprehensive grammar noting the differences between the various dialects. A little grammar published by Molloy in 1867 is instructive on account of the author’s peculiar point of view. The most useful books for the study of the living language are the series of booklets (five) published by Father O’Growney, one of the chief promoters of the present movement. Mention should also be made of J.P. Henry’s Handbook of Modern Irish, pts. i.-iv., and of the grammars by P.W. Joyce (Dublin, 1896) and the Christian Brothers (Dublin, 1901). For the northern form of Irish J.P. Craig’s Grammar of Modern Irish is useful (² Dublin, 1904). The phonetics of a Munster dialect have been investigated by R. Henebry, A Contribution to the Phonology of Desi Irish (Greifswald, 1901). The dialect of the Aran Islands off the coast of Galway has been described by F.N. Finck, Die Araner Mundart, i. Lautlehre und Grammatik, ii. Wörterbuch (Marburg, 1899). G. Dottin has given an account of a dialect of North Connaught (Mayo) in the Revue celtique, xiv. pp. 97-137. A study of the speech of the north was published by E.C. Quiggin under the title of A Dialect of Donegal, Phonology and Texts (Cambridge, 1906). For an account of the decay of Irish see H. Zimmer, “Die keltische Bewegung in Irland,” Preussische Jahrbücher for 1898, vol. 93, p. 59 ff., and the last chapter of Douglas Hyde’s Literary History of Ireland (London, 1901).

The work of the earlier compilers of glosses will be mentioned in the literature section below. The first dictionary of the modern language of any importance was that published by J. O’Brien in 1768. Next came E. O’Reilly with his Irish-English Dictionary (Dublin, 1817). This book contains a vast store of words gathered on no principle whatever from all manner of sources, and has therefore to be used with caution, but even at the present day it renders considerable service. A second edition with a supplement by O’Donovan was published after the latter’s death in 1864. The first trustworthy dictionary of the modern language was published under the auspices of the Irish Texts Society by P.J. Dinneen (London, 1904). English-Irish dictionaries have been compiled by D. Foley (Dublin, 1855); E.E. Fournier (Dublin, 1903); T. O’Neill Lane (Dublin, 1904).

(b) Scottish Gaelic.—Scottish Gaelic is the form of Goidelic speech which was introduced into Scotland by the Dalriadic Scots who came over from Ireland in the early centuries of our era. We possess practically no early monuments of the language. We have one or two inscriptions in Latin characters, such as that at St Vigeans and the Ogams mentioned above, which have not yet been solved. In the Book of Deir there is a colophon of a few lines probably written by an Irish scribe in the 9th century, and as the language of these lines differs in no wise from the Irish of the period, we do not know if they accurately represent the Gaelic of Scotland or if they may not be pure Irish. In the same MS. there are further Gaelic scraps belonging to the 11th and 12th centuries. The word-forms in these entries are identical with those current at the time in Ireland, but the historical orthography seems to show more signs of decay than is the case in Irish. The medieval Scottish MSS. in the Advocates’ Library at Edinburgh are only just being published, but they seem either to hail from Ireland or to be written in pure Irish. The end of the 15th century brought a change. The Lordship of the Isles, the great bond between Ireland and Scotland, was broken up. The Gaels of Scotland, thrown on their own resources, advanced their own dialect to the position of a literary language and tried to discard the Irish orthography. The Book of the Dean of Lismore, compiled about 1500, is written in a kind of phonetic orthography which has not as yet been sufficiently investigated. The language of those poems which are not directly ascribed to Irish poets, and which may therefore be regarded as representing the literary language of the Highlands at the time, seems to occupy a position midway between Irish and Scottish Gaelic. But until the beginning of the 18th century the Highlands were 617under the literary dominion of Ireland, so much so that Bedell’s Irish version of the Scriptures was circulated in Scotland with a glossary from 1690 to 1767, and Bishop Carsewell’s version of Knox’s Prayer-book (1567) is pure Irish. The language of the people is poorly represented in the 16th and 17th centuries, and the orthography is not fixed until we reach the 18th century.

Irish and Scottish Gaelic differ considerably in point of vocabulary, but there are also important divergences in phonetics and inflections. In the first place, Scottish Gaelic as written has entirely given up the nasal mutation (eclipse), e.g. Scottish ar bò, “our cow,” Irish ar m-bó; Scottish nan tìr, “of the countries,” Irish na d-tír. It should, however, be observed that in Skye and the Outer Isles the nasal mutation has been partly restored and in some places there are even parallels to the Welsh nasal mutation of c, p, t to ngh, mh, nh. Secondly, post-vocalic c, p, t are commonly preceded by a breathed sound not represented in writing, thus mac “son,” is pronounced mahk; slat, “rod,” as slaht. Again there is a tendency to insert a sibilant in the group rt, thus ceart, “right,” is sounded kearšt, and the distinction between palatalized and non-palatalized sounds is not so rigidly observed as in Irish. The group cht is in Scotland pronounced as if chk. We may also mention that Scottish Gaelic preserves an old e in a number of words where Irish now has a, thus, Old Ir. fer, Scottish G. fer, Irish far, but in both cases the spelling is fear (in this respect Scottish Gaelic goes hand in hand with Manx and the almost extinct Irish of Down). Similarly, we find that in Scottish Gaelic and Manx stressed vowels preceding a palatalized consonant have not undergone palatalization to the same extent as in Irish, e.g. in Ireland duine, “man,” < *dunjo-, is pronounced din’ð, but in Scotland dun’ð (in Manx written dooinney). A further peculiarity of Scottish Gaelic is that it substitutes lenes or voiceless mediae for the voiced stops, and even l, r, n sounds show a great tendency to give up the voice. Scottish Gaelic goes farther even than Irish in the confusion of vowel-sounds, e.g. Lat. coxa, Ir. cos, “foot,” Sc. cas; Ir. codal, Sc. cadal. When we turn to the inflections we find that analogy has here played a much greater part than in Irish. There is a tendency to make the plural of all substantives except masculine monosyllables end in -an. In the conjugation the synthetic forms have with one or two exceptions entirely disappeared and the present forms have become momentary in force. Hence in ordinary grammars it is stated that the present has become a future, thus ni mi means “I shall do.” The past participle chiefly ends in -te as against Irish -the, -te, or -tha, -ta, according to the quality of the preceding sound. The present (future) and past subjunctive (conditional, representing both the imperfect indic. and secondary future of Irish) supply the place of the Irish consuetudinal forms. In idiom also Scottish has diverged very considerably from Irish, e.g. in the use of tha (Ir. tá) for is.

It seems now to be agreed that the various dialects of Scottish Gaelic fall into two main divisions—northern and southern. Mackinnon states that the boundary between the two passes roughly up the Firth of Lorne to Loch Leven, then across country from Ballachulish to the Grampians. The country covered by the northern dialect was of old the country of the Northern Picts, whilst the portion of Argyllshire south of the boundary line, together with Bute and Arran, made up the kingdom of Dalriada. The Gaelic district south of the Grampians belonged to the Southern Picts. The southern dialect is commonly regarded as the literary language. It approaches more nearly to Irish and preserves the inflections much better than the speech of the north.

The following characteristics of the northern dialects may be mentioned:—(1) The diphthongization of open e to ia is carried much farther in the north than in the south. (2) The vowel ao in the north is more regularly the high-back-narrow-unrounded vowel-sound, whereas the south in many cases has a low-front-wide-round sound. (3) The north has str in initial position where the south prefers sr. Further, the northern dialects go very far in dropping unaccented final vowels. It may be remarked that in the reduction of derivative endings containing long vowels Scotland goes hand-in-hand with Ulster Irish, thus Connaught aran, “bread,” is in Ulster and Scotland aran. Again, Scottish agrees with North Irish in the loss of synthetic verb-forms and in using as negative cha, Mid. Ir. nico, nocha. But, on the other hand, Scotland, with the exception of South Argyll and some of the Isles, diphthongizes accented a, o, e, in monosyllables, before ll, nn, m, thus resembling the speech of Munster. In South Argyll the original short vowel is half lengthened.

As to the southern limits of Gaelic speech in Scotland, the boundary between Gaelic and English in medieval times was the so-called Highland line, and at the War of Independence it is probable that it extended to Stirling, Perth and the Ochil and Sidlaw Hills, the Inglis being limited to a very narrow strip along the coast. Dr J.A.H. Murray traced the linguistic frontier in 1869-1870 with the following results. The line started about 3 m. west of the town of Nairn on the Moray Firth and ran in a south-east direction to the Dee, 4 m. above Ballater. On the other side of the Dee it began 4 m. above Balmoral and followed the boundary of Perth and Forfar as far as Glen Shee, where it went off to the south-west as far as Dunkeld. After passing Birnam Hill it turned due west until the upper part of Glen Almond was reached, where it bent to the southward, passing through Comrie and along the braes of Doune to the Teith, 3 or 4 m. below Callander. Thence it ran along the north shore of Lake Monteith to Gartmore, and from there to Rowardennan on the east side of Loch Lomond. On the west side it passed through Glen Douglas down Loch Long and the Firth of Clyde, leaving Bute and Arran to the west. At the present day this boundary has probably receded to the extent of several miles, and even in 1870 there were districts such as Bute and the region round Dunoon where Gaelic was only spoken by the oldest natives and the immigrant population. The language is not found in the north-east of Caithness, the boundary running, according to Murray, roughly from a little north-east of Lybster to the mouth of the Forss. Celtic was driven out of Shetland and Orkney by Scandinavian some time during the middle ages. (See further J.A.H. Murray, The Dialect of the Southern Counties of Scotland, London, 1875; Revue celtique, vol. ii. pp. 180-187.)

Until the 18th century Gaelic was spoken in Galloway and on the uplands of Ayr and Lanark. The following figures from the census returns illustrate the decrease in the number of persons who speak Gaelic:—

	Monolinguists.	Bilinguists.
1881	No return	231,594 (this includes Gaelic monolinguists)
1891	43,738	210,677
1901	28,106	202,700

In the last-mentioned year it appears that nearly one-half of the speakers of Gaelic are reported from the counties of Inverness and Ross (23,893 monolinguists and 82,573 bilinguists). From about 1300 we find Scottish emigrants filtering into the glens of Antrim, where the Gaelic that is spoken is still unmistakably Scottish. There have long been local societies of Highlanders for the cultivation of their native tongue, the most important one being An Comunn Gàidhealach (founded 1891). This society holds an annual gathering called the Mòd (= Eng. “moot”) on the lines of the Welsh Eisteddfod, and recently the Scottish Education Department has countenanced the teaching of Gaelic in Highland schools. But the political element plays little or no part in the language movement in Scotland, and the latter is not likely to assume the proportions of the Gaelic League in Ireland. As a rule, however, Highlanders are better able to read their own language than Irish Gaels, for, the majority being Protestants, they are encouraged to read their Bibles. There are only two periodicals which devote half their space to Gaelic. The one is An Deo-Greine (“the sunbeam”), founded October 1905; and the other is the Catholic propagandist quarterly Guth na Bliadhna (“the voice of the year”), started in 1904. Up to 1905 a fortnightly newspaper printed wholly in Gaelic appeared in Prince Edward Island, under the title of An Mac-talla (“the echo”), and efforts have been made to revive it. A weekly newspaper wholly in Gaelic was started in 1908 by R. Stuart Erskine under the title of Alba.

Authorities on Scottish Gaelic.—The first grammar of Scottish Gaelic was compiled by W. Shaw (An Analysis of the Galic Language, 1778). The most useful one was that published by Alexander Stewart, Elements of Gaelic Grammar (Edinburgh, 1801). A revised edition of this work with many additions and corrections was published by H.C. Gillies, London, 1902. This book is rather spoilt by the author’s attitude, and requires to be supplemented and corrected. G. Henderson and C.W. Robertson have published important papers on the modern dialects in the Zeitschrift für celtische Philologie, the Celtic Review and the Transactions of the Gaelic Society of Inverness. The most useful work on Gaelic philology is Alexander Macbain’s Etymological Gaelic Dictionary (Inverness, 1896) (a later edition by W.J. Watson). The chief dictionaries are Dictionarium Scoto-Celticum, published by the Highland Society of Scotland (Edinburgh, 1828); R.A. Armstrong, Gaelic Dictionary in two parts (London, 1825); N. McAlpine, Pronouncing Gaelic Dictionary (Edinburgh, 1847) (this book gives the pronunciation of Islay); Macleod and Dewar, Gaelic and English Dictionary (latest edition, Edinburgh, 1901); Faclair Gàidhlig, published by E. Macdonald, Herne Bay, appearing in parts since 1902.

(c) Manx.—Our sources of information with regard to the language of the Isle of Man are even more scanty in the early period than they are in the case of Scotland. There are a number of references to the island in Irish literature, but the earliest monument of the vernacular we possess is the version of the Book of Common Prayer made by Bishop Phillips in 1610. In this translation the traditional Irish orthography is not followed. The spelling resembles the orthography which was employed in Scotland by the compiler of the Book of the Dean of Lismore. How far this system was used is a question which it is difficult to decide. In Scotland the Irish orthography has prevailed in a slightly modified form, but Manx writers adhered to a mode of spelling which was as phonetic as any system based on English, or, probably more correctly Anglo-Scottish, orthography could be. This fact, combined with the rapid phonetic decay of the 618language, makes it extremely difficult to discover what sound-values are to be attached to the various symbols. At the beginning of the 18th century English was not understood by two-thirds of the natives, and in 1764 the S.P.C.K. issued a paper containing this statement: “The population of the Isle is 20,000, of whom the far greater number are ignorant of English.” But from this time English gradually crept in. The last edition of the Manx Bible was issued in 1819, and of the New Testament in 1840. The present writer’s great-grandmother refused to speak English, his grandfather (b. 1815) preached in Manx and English, and his father (b. 1844) only spoke English. The following figures illustrate the rapid decline of the language:—

	Monolinguists.	Bilinguists.
1875	190	12,340 (out of a population of 41,084 exclusive of Douglas)
1901	None	4,419

Manx stands in a much closer relation to Scottish Gaelic than Irish, and fishermen state that they could understand a good deal of what is said in South Argyll, though they are quite at a loss at Kinsale. Manx exhibits the same tendency as Scottish to use analytical and periphrastic forms in the verb, thus jannoo, “to do,” is used like Scottish deanamh with an infinitive to express the past and future. The present has acquired a momentary (future) signification, and the past participle ends in -it (Scottish -te). The negative is cha as in Scotland and Ulster. Manx goes as far as northern Scottish in dropping unstressed final vowels, e.g. chiarn, “lord,” Irish, tighearna; -yn is the favourite plural ending in substantives. The nasal mutation has been partly given up. Old Irish stressed e is frequently retained, e.g. fer, “man,” Irish far (spelt fear), and the vowels o and a are confused as in Scottish, e.g. Manx cass, “foot,” Scottish cas, Irish cos. Manx is divided in itself about the treatment of short accented vowels beforell, nn, m. According to Rhys the south side lengthens, whilst the north side diphthongizes; e.g. Irish crann, “tree,” clann, “offspring,” S. Manx kron,klon, N. Manx, kroun, kloun (written croan, cloan). In the matter of stress Manx is quite original, going farther even than the dialects of the south of Ireland. Not only does it shift the stress in the case of heavy derivative suffixes like -an and reduce the preceding vowel, e.g. Ir. fuaran, Sc. fuaran, Manx fran, “spring,” but even in cases like caghláa, “variety,” Sc. Ir. caochladh, O. Ir. coimmchloud; coráa, “voice,” Ir. comhradh. The Mid. English stress on the final is further retained in words from the French such as ashóon, “nation,” livréy, “deliver.”

As other features peculiar to Manx we may mention the following. An intervocalic s or sh shows a tendency to become lisped and voiced to d. In monosyllables post-vocalic final m, n, are often preceded by an intrusive b, d respectively, thus ben “woman,” may be heard as bedn. Ir. a becomes more palatal and is often æ. Ir. sc becomes st, sht, e.g. Ir. fescor, “evening,” Manx fastyr; Ir. uisce, “water,” Manx ushtey.

Authorities on Manx.—The place and personal names of the Isle of Man have been collected by A.W. Moore in Manx Names² (London, 1903) (33% of the proper names are Scandinavian). The chief source of information about the spoken language is J. Rhys, The Outlines of the Phonology of Manx Gaelic (London, 1895) (the book has unfortunately no index and no texts). The only serious attempt to represent spoken Manx graphically is the transcription of a song by J. Strachan in the Zeitschr. für celtische Philologie, vol. i. p. 54. The native grammarian is J. Kelly, who in 1803 published A Practical Grammar of the Ancient Gaelic or Language of the Isle of Man, usually called Manks. This book was republished by W. Gill for the Manx Society in 1859, and a facsimile reprint of this latter was made for Quaritch, London, 1870. A useful little book entitled, First Lessons in Manx was published by Edwin Goodwin (Dublin, 1901). There are two dictionaries, one by A. Cregeen, Douglas 1835, which is now being reprinted for An Cheshaght Gailckagh, a Douglas society which is endeavouring to encourage the use of Manx and to get it introduced into the schools. The other dictionary is by J. Kelly in two parts—(i) Manx and English, (2) English and Manx, published by the Manx Society in 1866. Kelly also prepared a Triglot of Manx, Irish and Gaelic, based upon English, which has never been published. A useful paper on the language appeared in the Transactions of the London Philological Society for 1875 by H. Jenner, “The Manx Language: Its Grammar, Literature and Present State.”

(E. C. Q.)

(ii). Brythonic. The term Brythonic is used to denote the Celtic dialects of Wales, Brittany and Cornwall. Unlike the Goidels the Brythonic peoples have no common name for their language. Forms of Brythonic speech were doubtless current throughout England and Wales and the Lowlands of Scotland at the time of the Saxon invasion. The S.E. of Britain may have been extensively Romanized, and it is not impossible that remnants of Goidelic speech may have lingered on in out-of-the-way corners. No literary documents dating from this period have been preserved, but some idea of the character of Brythonic may be gathered from the numerous inscriptions which have come to light. In the middle of the 6th century Brythonic was confined to the western half of Britain south of the Clyde and Forth. The colonization of Britannia minor or Armorican Brittany during the 5th and 6th centuries will be described later. In the latter part of the 6th century the W. Saxons pushed their conquests as far as the estuary of the Severn, and from that time the Brythons of S.W. Britain were cut off from their kinsmen in Wales. Early in the 7th century the Brythons of Strathclyde were similarly isolated by the battle of Chester (613). The kingdom of Strathclyde maintained a separate existence until the 10th century, and it is generally stated that Brythonic speech did not die out there until the 12th century. The question as to how far Brythonic names and words have survived in these districts has never been properly investigated. Certain it is that Brythonic numerals survived amongst shepherds in Cumberland, Westmorland and N.W. Yorkshire down to the second half of the 19th century, just as herrings are still counted in Manx by Manx fishermen otherwise quite innocent of the language. Accordingly, from the 7th century onwards Brythonic became gradually limited in Great Britain to three districts—Strathclyde, Wales, and Cornwall and Devon. During the 7th century the Brythons of Wales and Strathclyde often fought side by side against the Angles, and it is from this period that the name by which the Welsh call themselves is supposed to date, Cymro < *Combrox, pl. Cymry < *Combroges, i.e. “fellow-countrymen” as opposed to W. allfro, Gaul. Allobroges, “foreigners.” We have no means of determining when Celtic speech became extinct in the petty states of the north which retained their independence longest.

The chief features which distinguish the Brythonic from the Goidelic dialects have already been enumerated. In the course of the 6th and 7th centuries final short vowels disappeared. In compound names the final vowel remains in the first component until the 7th century. Short vowels in other than initial syllables when immediately preceding the stress (on the historical penultimate) disappear, whilst long ones are shortened, e.g. Welshcardawt from Lat. caritatem. Other vowels in unstressed position are apt to be reduced, thus o, u, give i in O.W. (Mid. W. y). A marked characteristic of Welsh as distinguished from Cornish and Breton is the treatment of a under the influence of a following i. In Welsh the result is ei, in Corn. and Bret. e, e.g. Welsh seint, “saints,” Bret. sent, sing. sant. The mutations seem to have started in the second half of the 6th century in the case of the tenues.

See J. Loth, Les Mots latins dans les langues Brittoniques (Paris, 1892); J. Loth, Chrestomathie bretonne (Paris, 1890).

(a) Welsh (Cymraeg).—It is usual to divide the history of the Welsh language into three periods—Old, Middle and Modern. To the oldest period belong the collections of glosses, the earliest of which go back to about 800. The middle period extends from 1100 to 1500.

As a rule the medial mutation of the tenues and mediae is not denoted in O. Welsh. Intervocalic g is sometimes retained but generally it has disappeared, whilst after r and l it is still written. In the course of the 9th century initial w (v) becomes gu (later gw). As the O. Welsh documents consist almost entirely of isolated words, we know scarcely anything about the morphology of the language during this period. To the middle period belong the ancient poems from the Black Book of Carmarthen, but the language of these compositions is evidently much older than the date of the manuscript (12th century), as it preserves a number of very archaic features. Other important sources of information for this period are the O. Welsh Laws contained in a MS. of the 12th century. To a somewhat later date belong the Mabinogion (14th century MS.), and the prose versions of French romances published by R. Williams (15th century). In Middle Welsh the consonant mutations are in general denoted in writing, though not consistently, and from this period dates the introduction of w and y (O.W. u, i) to denote vowel sounds. The symbol ll to denote a voiceless l was already employed in Mid. W. but rh (= voiceless r), dd (= Eng. th in “thou”) and f (= v) either do not appear or only become regular during the modern period In Mod. W. the orthography is regularized and does not differ 619materially from that of the late medieval documents. In O.W. the old stress on the final syllable (the historical penult) appears to have been preserved, but during the middle period the accent was shifted to the penult. In consequence of this change aw (< a) in final syllables is reduced to o in Mod. W., e.g. Mid. W. pechawt < Lat. peccatum, Mod. W. pechod.

The comparative wealth of inflection preserved by O. Ir. has almost entirely disappeared in Welsh. There are only the faintest traces of the case forms, the dual and the neuter gender. Compared with the Irish nominal declension according to -o- (-jo-), -a-, -i-, -u-, -s-, guttural, dental and nasal stems, Welsh only distinguishes the nom. sing. and plur., the latter sometimes retaining an old formation. Thus masc. -o- stems show palatal modification, e.g. corn, “horn,” plur. cyrn < *korni; the plural ending of -u- stems, O. Gaulish -oves, gives O.W. -ou, Mid. W. -eu, Mod. W. -au, e.g.penneu, “heads.” The termination -ones of the -n- stems appears as -on. The infixation of pronominal objects between a verbal particle and the verb itself continues in use down to the present day as in Breton. In the third person sing. of the pres. ind. there are instances in the oldest Welsh of the peculiar alternation between orthotonic and absolute forms which characterize the Irish paradigms, e.g. pereid, “it endures,” but ny phara. The several types of conjugation represented in Irish have become obscured, traces remaining only in the endings of the third sing. of the pres. ind., the pret. ind. (Mid. W. -as, -es, -is) and the pret. passive (Mid. W. -at, -et, -it). The verb system of Welsh comprises the following tenses: indic. present (also used as future), imperative, imperfect, preterite (in Mid. W. forms with s have become prevalent as in Irish, but forms corresponding to the Irish preterites in t or with reduplication or unreduplicated with long vowel are not infrequent in the early poetry), pluperfect (a new formation), pres. and pret. passive. In the subj. early W. distinguishes pres. and past, but the latter comes to be replaced by the pluperfect indicative. The sign of the subj. is -h- < s, which reminds one of the Irish s-subj., though the formation is somewhat different. There are also traces of a future formation containing h < s. (See also under Wales.)

We have seen already that Wales began to exist as a separate entity roughly at the end of the 6th and beginning of the 7th centuries. In the second half of the 8th century the Welsh were confined in pretty much their present limits by History and extent.Offa, king of Mercia, who constructed the Dyke going by his name, which has approximately remained the political boundary between England and Wales ever since. From this time onwards the bitter feeling against England which we find expressed in the fervid compositions of Iolo Goch and other political bards served to prevent any serious inroads of English on Welsh-speaking territory. With the advent of the Tudors, however, there came a great change. Henry VII. owed his throne in large measure to the support he had received from Wales and he prided himself on his Welsh ancestry. A consequence of this was that throughout the 16th century Wales received exceptionally favourable treatment at the hands of the English sovereign and parliament. In 1562 a decree was issued ordering a translation of the Bible to be made into Welsh. All this could naturally not be without effect on the attitude of the leaders of the people towards England. The change is already apparent in the poems of Lewis Glyn Cothi and others. And the striking difference in the manner in which the Reformation was regarded in Ireland and Wales is worthy of remark. During the Stuart wars the Welsh nobles fought invariably on the Royalist side, and there is plenty of other evidence that the aristocracy of Wales was becoming thoroughly anglicized both in sentiment and language. At the same time the practice of the Tudors was reversed in many particulars. Thus it became the custom to appoint Englishmen ignorant of the national language to the Welsh bishoprics. In this manner it is not a matter for surprise that a feeling of estrangement should grow up between the bulk of the population, who only knew Welsh, and the clergy and nobles, their intellectual leaders. The neglect of the national language is evident from the large number of English words which have even crept into such classical works as Prichard’s Canwyll y Cymry and Ellis Wynn’s Gweledigaethau y Bardd Cwsg. It is stated that, of the 269 works published by Welshmen between 1546 and 1644, 44 were in Latin, 184 in English and only 41 in Welsh, and of these 37 consist of works of piety. Thus at the beginning of the 18th century there seemed a fair chance that Welsh would soon become extinct like Cornish.

An extraordinary change was brought about by the Methodist movement in Wales. The preachers, in order to get hold of the masses, addressed them in the vernacular, and their efforts were crowned with enormous success. At the same time a minister of the Established Church, Griffith Jones, went about Wales establishing lay schools to which young and old might come to learn to read the Welsh Bible. Between 1737 and 1761 3395 such schools sprang up, at which no fewer than 158,238 persons of all ages learned to read their native language. After Griffith Jones’s death this work was carried on by others, notably by Charles of Bala (1755-1814), who passed over to Calvinistic Methodism and whose schools were transformed after the model of the Sunday schools instituted in 1782 by Robert Raikes. Charles of Bala was largely instrumental in the founding of the British and Foreign Bible Society, and Wales was provided with 100,000 copies of the Bible and Testament at very moderate prices. Bishop’s Morgan’s version of the Scriptures made in 1588 (final revision 1620) represents the speech of North Wales which had remained more or less free from English influence, so that the language of the Welsh Bible is rightly regarded as the literary model. Three-fourths of the inhabitants of Wales belong to the various Nonconformist sects, and therefore pass almost without exception through the Sunday school, where they are drilled in its sole object of study, the Welsh Bible.

With the increasing employment of Welsh owing to the Nonconformist movement there was also awakened a new interest in the past history of the principality. A society calling itself the Cymdeithas y Cymmrodorion was founded in London in 1751, and during the succeeding half-century two periodicals exclusively in Welsh were started, the one, Trysorfa y Gwybodaeth, in 1770, the other, Cylchgrawn Cymraeg, in 1793. The year 1792 witnessed the creation of an important society, the Cymdeithas y Cymreigyddion, in London, in which the moving spirits were William Owen (Pughe), Owen Jones and Edward Williams. The results of their indefatigable search for ancient Welsh manuscripts were published in three volumes under the title Myvyrian Archaiology (London, 1801-1807). Owen further published an edition of the greatest medieval Welsh poet Dafydd ap Gwilym, and also the first copious dictionary. But this was not all. In Goronwy Owen (1722-1769) a poet had arisen whose works could stand comparison with the compositions of the medieval writers, and it was owing to the efforts of the three men above mentioned that the national Eisteddfod (= session, from eistedd, “to sit”) was revived. The origin of these literary festivals is shrouded in obscurity. It is recorded that a S. Welsh prince, Gruffydd ap Rhys, held a festival lasting forty days in 1135 to commemorate a victorious campaign at which poets and minstrels competed for gifts and other rewards. Gruffydd’s son Rhys ap Gruffydd is reported to have instituted a similar contest in 1176, at which the successful competitors received a chair whilst the others were given presents. It would seem that after the loss of Welsh independence a carefully graded order and a system of jealously guarded rules came into existence. Similar national festivals were held under royal patronage under Henry VIII. in 1523 and again under Elizabeth in 1568. From 1568 until 1819 no general eisteddfod for all Wales was held. Since 1819 the national festival has been held annually and every little town has its own local celebration. Hence the Nonconformist Sunday school, the pulpit and the eisteddfod may be regarded as the most potent factors in resisting the inroads of English. The whole question of the vitality of Welsh and what may be called the political and social history of the language is treated in great detail by H. Zimmer, “Der Pan-Keltismus in Gross-britannien und Irland,” i., in Preussische Jahrbücher, vol. xcii. (1898). In elementary schools in Wales the use of Welsh has been permitted since 1893.

With regard to the extent over which Welsh is spoken a detailed map is given in J.E. Southall’s Welsh Language Census of 1891 (Newport, 1895). A line drawn from the southern end of the estuary of the Dee about 2 m. W. of Connah’s Quay to Aberthaw in Glamorgan would practically include all those districts where Welsh is spoken by 60% of the population, and considerable deductions would have to be made for parts of Flint, Montgomery, most of Radnor and the N. part of Brecon. Little is spoken in the southern half of the Gower peninsula or in S. Pembrokeshire. Over much of Anglesey 97½% of the population spoke Welsh and in parts of Cardiganshire 98.3%. Of a total population in 1901 of 2,012,876, 929,824 were returned as speakers of Welsh, of whom 280,905 were monoglots. That Welsh is a very living language may be gathered from the following statistics. Between 1801 and 1898 no fewer than 8425 volumes were published in the vernacular, whilst in 1895 there were appearing regularly 2 quarterlies, 2 bi-monthlies, 28 religious and literary monthlies and 25 weekly papers. In 1909 the number was probably greater. The danger for Welsh lies rather in the direction of internal decay. The speech of the people is saturated with English words and idiom, and modern writers like Daniel Owen submit to the same influence instead of returning to the classical models of the 17th century.

Much remains to be done as regards the classification of the modern Welsh dialects. It is usual to divide them into four groups—(1) Powys (N.E.); (2) Gwynedd (N.W.); (3) Dyfed (S.W.); (4) Gwent (S.E.). One of the chief points on which N. and S. diverge is the pronunciation of the vowels i, u, y, which in the S. all tend to become i. The difference between N. and S. was noticeable as early as the time of Giraldus Cambrensis. See M. Nettlau, Beiträge zur cymrischen Grammatik (Leipzig, 1887), also Rev. celt. ix. pp. 64 ff., 113 ff.; T. Darlington, “Some Dialectal Boundaries in Mid-Wales,” Trans, of the Hon. Soc. of Cymmrodorion, 1900-1901. The only scientific description of a living dialect is “Spoken N. Welsh,” by H. Sweet, Trans, of the London Phil. Soc., 1882-1884.

Authorities on Welsh Language.—For the study of older Welsh:—J.C. Zeuss, Grammatica Celtica (Berlin², 1871)—an index to the O. Welsh glosses cited in this work was compiled by V. Tourneur, Archiv f. celt. Lexikographie, iii. 109-137; J. Strachan, An Introduction to Early Welsh, with a Reader (Manchester, 1909); J. Rhys, Lectures on Welsh Philology (London², 1879). Editions of texts—The Black Book of Carmarthen, facsimile edition by J. Gwenogvryn Evans (Pwllheli, 1906); J. Rhys and J. Gwenogvryn Evans, The Text of the Mabinogion (Oxford, 1887); The Myvyrian 620Archaiology of Wales (1801-1807; reprinted Denbigh, 1870); W.F. Skene, The Four Ancient Books of Wales (2 vols., Edinburgh, 1868); Aneurin Owen, Ancient Laws and Institutes of Wales (London, 1841); facsimile edition by A.W. Wade-Evans, Welsh Medieval Law (Oxford, 1909); K. Meyer, Peredur ap Efrawc with glossary (Leipzig, 1887); R. Williams, Selections from the Hengwrt Manuscripts (London, 1876-1892); J.E. Southall, Wales and Her Language (Newport, 1892). The earliest Welsh grammar was published as long ago as 1567 in Milan by Griffiths Roberts, reprinted in facsimile as supplement to the Revue celtique (Paris, 1883). An account of the language was prefixed to Owen Pughe’s Dictionary (1803). During the 19th century many manuals of indifferent value saw the light of day. The most authoritative works are:—T. Rowland, A Grammar of the Welsh Language (Wrexham, 18531, 18764), (still the most complete work), the same author also published a companion volume ofWelsh Exercises (Wrexham, n.d.); W. Spurrell, A Grammar of the Welsh Language (Carmarthen3, 1870); E. Anwyl, A Welsh Grammar for Schools, (i.) Accidence, (ii.) Syntax (London2, 1898). Other useful manuals for the beginner:—T. Jones, A Guide to Welsh, pts. i. ii. new ed. (Wrexham, n.d.); S.J. Evans, The Elements of Welsh Grammar (Newport3, 1903). Dictionaries:—The first Welsh dictionary was compiled by William Salesbury (London, 1547; facsimile reprint, London, 1877); W. Owen Pughe, A Dictionary of the Welsh Language (2 vols., London, 1803; reprinted Denbigh, 1870); W. Spurrell, Welsh-English and English-Welsh Dictionary (Carmarthen6, 1904); a smaller one by W. Richards in 2 vols. (Wrexham, n.d.), and many others. A dictionary on a large scale was planned by D. Silvan Evans and subsidized by the government. Only A-Dd has, however, appeared (Carmarthen, 1893-1906), cp. J. Loth in Archiv. f. celt. Lex. vol. i. for additions and corrections. A survey of Welsh periodical literature is contained in T.M. Jones’s Llenyddiaeth fy Ngwlad (Treffynnon, 1893). For Welsh folklore see J. Rhys, Celtic Folklore, Welsh and Manx (Oxford, 1901). H.H. Vaughan, Welsh Proverbs (London, 1889), also Rev. celt. iii. 419 ff. See also G. Dottin, Revue de synthèse historique, vi. 317 ff.; H. Zimmer and L.C. Stern in Kultur der Gegenwart, Teil 1, Abt. xi. 1.

(E. C. Q.)

(b) Breton.—Breton (Brezonek) is the name given to the language spoken by those Britons who fled from the south-west of England to Armorica (see Brittany) in the 5th and 6th centuries of our era to avoid being harassed by the Saxons. The first migration probably took place about 450. The Dumnonii and Cornovii founded small states in Brittany, or Britannia Minor, as it was termed, and were followed in the second half of the 6th and into the 7th century by a long stream of refugees (cf. J. Loth, L’Émigration bretonne, Paris, 1883; A. de la Borderie, Histoire de la Bretagne2, vol. i., 1905).

In the earliest stages it is difficult to distinguish Breton from Welsh. The history of the language may be divided into Old Breton from the 7th to the 11th centuries, Middle Breton from the 11th to the 17th centuries, and Modern Breton. In Old Breton the only material we possess consists of glosses and names occurring in lives of saints, Frankish authors, and charters. However, we find a few characteristics which serve to show that the old glosses are really Breton and not Welsh. Thus, an original a never becomes a diphthong (au, aw) in Old Breton, but remains o. In Bret, gnbecomes gr. Further, in O.W. pretonic u is weakened to an indeterminate sound written i and later y, a phenomenon which does not occur in Breton,e.g. Lat. culcita appears in O.W. as cilcet, but in O. Br. as colcet. A marked characteristic of Breton is the confusion of i and e, e.g. Ir. lis, “court,” W. llys, Br. les. In Old Breton as in Old Welsh neither the initial nor the medial mutations are expressed in writing, whilst in Middle Breton only the latter are regularly denoted. In this period the language diverges very rapidly from Welsh. As prominent features we may mention the following. Stressed o (= Prim. Celt. and Ir. a) becomes eu, in unstressed syllables e; thus the suffix -aco becomes -euc and later -ec, but in Welsh -auc and later-oc, -og. Postvocalic -tr, -tl become -dr, -dl as in Welsh, but in Middle Breton they pass into -zr, -zl, which in the modern language appear as -er, -el; e.g. Mid. Br. lazr, Mod. Br. laer, “robber,” W. lleidr, Lat. latro. Further, -lt becomes -ot, -ut, e.g. Br. aot, aout, “cliff,” W. allt; Br. autrou, “lord,” Ir. altram, W. alltraw, athraw, Corn. altrou; and, more important still, th, ? (W. dd) become s, z, e.g. Mid. Br. clezeff, “sword,” Mod. Br. kleze, W.cleddyf. The orthography only followed the pronunciation very slowly, and it is not until 1659 that we find any attempt made to reform the spelling. In this year a Jesuit priest, Julien Maunoir (Br. Maner), published a manual in which a new spelling is employed, and it is usual to date Modern Breton from the appearance of this book, although in reality it marks no new epoch in the history of the language. It is only now that the initial mutations are consistently denoted in writing (medially they are already written in the 11th century), and the differences between the dialects first come into view at this time. As in Welsh the accent is withdrawn during the middle period from the final to the penultimate (except in the Vannes dialect), which causes the modern unstressed vowel to be reduced in many cases. Again, in Old Welsh and Old Breton a short stressed vowel in words of one syllable was lengthened, e.g. tad, “father,” pl. tadau, but in Modern Breton the accent tends to lengthen all stressed vowels. Breton has gone its own way in the matter of initial mutation. The nasal mutation has been entirely given up in the initial position, whilst a new mutation, called medial provection, has arisen in the case of b, d, g, which become p, k, t after a few words which originally ended for the most part in z or ch. The vocalic mutation of initial g in Breton is c’h. We may also make mention of one or two other points on which Breton differs widely from Welsh. Breton has given up the combination ng, e.g. Mid. Br. moe, Mod. Br. moue, “mane,” W. mwng, Ir. mong. The language betrays a fondness for nasalized vowels, and in this connexion it may be noted that v representing an original m (W. f, Ir. mh), though generally written ff in Middle Breton, now frequently appears as nv; Mid. Br. claff, Mod. Br. klanv, “sick, ill,” W. claf, M. Ir. clam. Final g after r and l and sometimes in monosyllables after a vowel is represented in Breton by c’h, whilst in Welsh in the one case we find a vowel and in the other nil, e.g. Br. erc’h, “snow,” W. eiry, eira; Br. lec’h, “place,” W. lle. In Welsh mb, nd immediately preceding the stress appear in the modern language as mm, nn but in Breton we find mp, nl, e.g. Br. kantol, “candle,” W. cannwyll, Lat. candela; Br. kemper, “confluence” (in place names), W. cymmer, Ir. combor.

With regard to the extent of country over which Breton is spoken we shall do well to note the seats of the old Breton bishoprics. These were Quimper, St Pol de Léon, Tréguier, St Brieuc, St Malo, Dol and Vannes. Under Count Nominoe the Bretons succeeded in throwing off the Frankish yoke (841-845) and founded an independent state. At this time of greatest political expansion the language boundary was formed by a line which started roughly a little to the west of Mont St Michel at the mouth of the Couesnon, and stretched to the mouth of the Loire. During the next three centuries, however, in consequence of political events which cannot be enumerated here, we find French encroaching rapidly on Breton, and the old dioceses of Dol, St Malo, St Brieuc, and in part Vannes became Romance-speaking (cp. J. Loth, Revue celtique, xxviii. 374-403). So that since the 13th and 14th centuries the boundary between French and Breton begins in the north about Plouha (west of St Brieuc Bay), and stretches to the mouth of the Vilaine in the south. That is to say, the Breton speakers are confined to the department of Finistère and the west of the departments Côtes-du-Nord and Morbihan. Lower Brittany contains a population of 1,360,000, of whom roughly 1,250,000 speak Breton. The number of monoglot Bretons is stated to have been 768,000 in 1878, 679,000 in 1885, and over 500,000 in 1898. There is an infinity of dialects and subdialects in Brittany, but it is usual to divide them into four groups. These are the dialects of (1) Léon in Finistère; (2) Cornouailles in Finistère, the Côtes-du-Nord and a part of Morbihan; (3) Tréguier in the Côtes-du-Nord and Finistère; (4) Vannes in Morbihan and a portion of the Côtes-du-Nord. The first three resemble one another fairly closely, but the speech of Vannes has gone its own way entirely. The dialect of Léon is regarded as the literary dialect, thanks to Legonidec.

The modern language is unfortunately saturated with words borrowed from French which form at least a quarter of the whole vocabulary. The living speech is further characterized by innumerable cases of consonantal metathesis and by parasitic nasalization. Loth gives specimens of the most important varieties of Breton in his Chrestomathie bretonne, pp. 363-380, but here we must confine ourselves to pointing out the two most salient differences between the speech of Vannes and the rest of Brittany. In Vannes the stress has not been shifted from the final syllable. In Haute-Cornouailles and Goelo there is a tendency to withdraw the stress on to the antepenultimate, whilst in Tréguier certain enclitics attract the accent to the final. s, z of the other dialects representing Welsh th become h in Vannes, e.g. W. caeth, Br. keaz, kez, “poor, miserable,” Vannes keah, keh. This phenomenon occurs sporadically in other dialects. It may also be mentioned that Prim. Celt, non-initial d, W. dd, is retained as z in Léon but disappears when final or standing between vowels in the other dialects, e.g. O. Br. fid, W. ffydd, “faith,” Léon feiz, in Cornouailles, Tréguier and Vannes, fé. It is doubtful if the most serious differences between the dialects are older than the 16th century.

In the middle ages the language of the Breton aristocracy was French. Upper Brittany was politically more important than the western portion. The consequence was that no patronage was extended to the vernacular, and Breton sank to the level of a patois with no unity for literary purposes. But a new era dawned with the beginning of the 19th century. The national consciousness was awakened at the time of the Revolution, when the Bretons became aware of the difference between themselves and their French neighbours. It may be mentioned by the way that the Breton language was regarded with suspicion by the leaders of the First Republic and attempts were made to suppress it. A Breton named Legonidec had to flee to England for fighting against the Republic. He came under the influence of the movement in Wales, and on his return sought to create a Breton literary language. He published an excellent grammar (Grammaire celto-bretonne, Paris, 1807) and a dictionary (Dictionnaire breton-français, Paris, 1821), from which he omitted the numerous French words which had crept into the language and for which native terms already existed. Legonidec’s 621example fired a number of writers with zeal for their native tongue and the clergy became interested. Under their auspices manuals of Breton were published and the language was utilized in a number of schools. A society called the Association Bretonne was founded in the year 1844. But under the Second Empire, for reasons which are not easy to discover, this Breton awakening was declared to be contrary to the interests of the state, and all the means at the disposal of a highly centralized government like that of France were employed to throttle the movement. Down to the present day the use of Breton is strictly forbidden in all the state schools, and the influence of the Roman Catholic clergy has for the most part been hostile to the language. However, the attitude of the government aroused considerable dissatisfaction in the early ’nineties, and in 1896 theAssociation Bretonne (disbanded in 1859 and reconstructed in 1873) appointed a permanent committee with the object of preserving and propagating the national language. At the same time some of the clergy headed by Abbé Buléon began to move, and Breton was introduced into many of the schools not under state control. In 1898 was founded the Union Régionaliste Bretonne, the most important section of which endeavours to foster the native speech in conjunction with the Comité de préservation du breton (founded 1896). In 1899 the annual meeting of the U.R.B. was modelled on the lines of the Irish Oireachtas, the Welsh Eisteddfod and the Scottish Mod, and festivals of this kind have been held ever since. Many Breton newspapers publish columns in Breton, thus Ar Bobl (a weekly newspaper founded in 1904 and published at Carhaix) frequently devotes half its columns to the language. But there is also a weekly four-page newspaper which is wholly in Breton. This is Kroaz ar Vretoned, edited by F. Vallée and published at St Brieuc. In addition to this there are three monthly magazines wholly in Breton. The first is Ar Vro, edited by the poet Jaffrennou, and in 1908 in its fifth year. The second is Dihunamb, written in the dialect of Vannes and started in 1905. The third is Feiz ha Breiz, started 1899.

Authorities for Breton.—For the external history of Breton see H. Zimmer, “Die keltische Bewegung in der Bretagne,” Preussische Jahrbücher for 1899, xcix. 454-497. For Old and Middle Breton, J. Loth, Chrestomathie bretonne (Paris, 1890), and the same writer’s Vocabulaire vieux-breton (Paris, 1884). Loth and E. Ernault have been indefatigable in investigating the history of the language. Their numerous contributions are mainly to be found scattered through the Revue celtique, Zeitschrift für celtische Philologie and the Annales de Bretagne. Ernault has also published Glossaire moyen-breton in 2 vols. (Paris, 1895-1896); Dictionnaire étymologique du moyen-breton (Paris, 1888). Another etymological dictionary was published by V. Henry (Paris, 1900). Grammars, &c.:—Dialect of Léon: Legonidec, Grammaire celto-bretonne (Paris, 1807, 1838², also contained in H. de la Villemarqué’s edition of Legonidec’s Dictionary); F. Vallée, Leçons élémentaires de grammaire bretonne (St Brieuc, 1902); E. Ernault, Petite Grammaire bretonne (St Brieuc, 1897, the latter also takes account of the dialects of Tréguier and Cornouailles). Dialect of Tréguier: L. le Clerc, Grammaire bretonne (St Brieuc, 1908); J. Hingant, Éléments de la grammaire bretonne (Tréguier, 1868); P. le Roux, “Mutations et assimilations de consonnes dans le dialecte armoricain de Pleubian,” Annales de Bretagne, xii. 3-31. Dialect of Vannes: A. Guillevic and P. le Goff, Grammaire bretonne du dialecte de Vannes (Vannes, 1902); Exercises sur la grammaire bretonne (Vannes, 1903); H. d’Arbois de Jubainville, “Étude phonétique sur le dialecte breton de Vannes,” Revue celtique, i. 85 ff. 211 ff.; E. Ernault, “Le Dialecte vannetais de Sarzeau,”Rev. celt. iii. 47 ff., 232 ff.; J. Guillome, Grammaire française-bretonne (Vannes, 1836). As a curiosity we mention P. Treasure, An Introduction to Breton Grammar (Carmarthen, 1903). Dictionaries: Legonidec, Dictionnaire français-breton (St Brieuc, 1847), Breton-Français (St Brieuc, 1850), both republished by de la Villemarqué and representing the Léon dialect; A. Troude, Nouveau Dictionnaire pratique français et breton du dialecte de Léon avec les acceptations diverses dans les dialectes de Vannes, de Tréguier, et de Cornouailles (Brest, 1869), and Nouveau Dictionnaire pratique breton-français (Brest, 1876); E. Ernault, “Supplément aux dictionnaires bretons-français,” Revue celtique, iv. 145-170. The Breton words in Gallo, the French patois of Upper Brittany, were collected by E. Ernault, Revue celtique, v. 218 ff.

(c) Cornish.—The ancient language of Cornwall (Kernûak, Carnoack) stood in a much closer relation to Breton than to Welsh,1 though in some respects it sides with the latter against the former.

It agrees with Breton on the following points:—It has given up the nasal mutation of initials but provects the mediae. Prim. Celt. a is not diphthongized, but becomes e, e.g. Corn, ler, “floor,” Br. leur, W. llawr, Ir. lar. Ng is lost as in Breton, e.g. toy, “to swear,” Br. toui, W. tyngu, Ir.tongu; nd becomes nt before the stress and not nn as in Welsh, e.g. Corn. Br. hanter, “half,” W. hanner. Cornish like Breton does not prefix a vowel to words beginning with s + consonant, e.g. Corn. spirit, later spyrys, Br. spered, W. yspryd. On the other hand, O. Cornish does not confuse i and eto the same extent as Bret., e.g. W. helyg, “willow,” O. Cornish heligen, Br. halek. Further, Cornish does not change th, d to s, z as in Breton, e.g. beth, “grave,” Br. bez, W. bedd, and initial g disappears in the vocalic mutation as in Welsh. Peculiar to Cornish is the change of non-initial t, d to s, z. This occurs in the oldest Cornish after n, l, e.g. O. Corn, nans, “valley,” W. nant; Corn. tâs, “father,” W. tad. A feature of later Cornish is the introduction of a d before post-vocalic m, n, e.g. pedn, “head,” W. pen. In later Cornish the accent seems to have fallen on the penultimate as in Modern Welsh and Breton.

In 936 the “Welsh” were driven out of Exeter by Æthelstan, and from that time the Tamar appears to have formed a general boundary between English and Cornish, though there seems to be evidence that even as late as the reign of Elizabeth Cornish was spoken in a few places to the east of that river. The decay of Cornish has been largely attributed to the Reformation. Neither the Prayer-book nor the Scriptures were translated into the vernacular, and we find the same apathy on the part of the Church of England in Cornwall as in Wales and Ireland. Unfortunately the Methodist movement came at a time when it was too late to save the language. By 1600 Cornish had been driven into the western parts of the duchy and in 1662 we are informed by John Ray that few of the children could speak it. Lhuyd gives a list of the parishes in which Cornish was spoken, but goes on to state that every one speaks English. In 1735 there were only a few people along the coast between Penzance and Land’s End who understood Cornish, and Dolly Pentreath of Mousehole, who died in 1777, is commonly stated to have been the last person who spoke it, though Jenner seems to show that there were others who lived until well into the 19th century who were able to converse in the dialect. However, the modern English speech of West Cornwall is full of Celtic words, and nine-tenths of the places and people from the Tamar to Land’s End bear Cornish names. Celtic words still in use are to be found in Jago’s Dialect of Cornwall (Truro, 1882); thus the name for the dog-fish is morgy, “sea-dog.”

Authorities for Cornish.—A mass of details about Cornish is collected in H. Jenner’s Handbook of the Cornish Language (London, 1904). (Cf. J. Loth’s review in the Revue celtique, xxvii. 93.) Lhuyd’s Archaeologica Britannica (1707) contains a grammar of the language as spoken in his day, and a Sketch of Cornish Grammar is to be found as an appendix to Norris’s Ancient Cornish Drama. A dictionary was published by R. Williams entitled Lexicon Cornu-Britannicum (Landovery, 1865), to which W. Stokes published a supplement of about 2000 words in theTransactions of the London Philological Society for 1868-1869. We may also mention the English-Cornish Dictionary, by F.W.P. Jago (Plymouth, 1887), and a Glossary of Cornish Names, by J. Bannister (Truro, 1871). W. Stokes published a Glossary to Beunans Meriasek in the Archiv für celtische Lexikographie, i. 101, and important articles by J. Loth have appeared in the Revue celtique, vols. xviii. to xxiv. W.S. Lach-Szyrma, “Les Derniers Échos de la langue cornique,” Revue celtique, iii. 239 ff. H. Jenner, “Some Rough Notes on the Present Pronunciation of Cornish Names,”Rev. celt. xxiv. 300-305.

III. The Language of the Ancient Picts.—The evidence from which we can draw any conclusions as to the affinities of the language of the Picts is so extremely scanty that the question has been the subject of great controversy. The Picts are first mentioned by Eumenius (a.d. 297), who regarded them as having inhabited Britain in the time of Caesar. In the year 368 they are described by Ammianus Marcellinus as invading the Roman province of Britain in conjunction with the Irish Scots. In Columba’s time we find the whole of Scotland east of Drumalban and north of the Forth divided into two kingdoms—north and south Pictland—and it is reasonable to identify the Picts, at any rate in part, with the Caledonians of the classical authors. Galloway and Co. Down were also inhabited by Picts. Bede in enumerating the languages of Britain mentions those of the Britons, Picts, Scots and the English. The names by which the Picts are known in history have aroused considerable discussion. It seems natural to connect Lat. Picti with thePictones and Pictavi of Gaul, but in Irish they are known as Cruithne, which appears in Welsh as Prydyn, “Pict”; cp. Prydein, “Britain,” forms corresponding to the earliest Greek name for these islands, ν?σοι πρετανικα?.

Three conflicting theories have been held as to the character of the Pictish language. Rhys, relying on the strange character of the Scottish Ogam inscriptions, pronounces it to be non-Celtic and non-Indo-European. In this he has been followed by Zimmer, who bases his argument on the Pictish rule of succession. Skene maintained that the Picts spoke a language nearly allied to Goidelic, whilst Stokes, Loth, Macbain, D’Arbois and Meyer are of opinion that Pictish was more closely related 622to Brythonic. Of personal names mentioned by classical writers we have Calgacus and Argentocoxus, both of which are certainly Celtic. The names occurring in Ptolemy’s description of Scotland have a decidedly Celtic character, and they seem, moreover, to bear a greater resemblance to Brythonic than to Goidelic, witness such tribal designations as Epidii, Cornavii, Damnonii, Decantae, Novantae. In the case of all these names, however, it should be borne in mind that they probably reached the writers of antiquity through Brythonic channels. Bede mentions that the east end of the Antonine Wall terminated at a place called in Pictish Pean-fahel, and in Saxon Penneltun. Peanresembles Old Welsh penn, “head,” Old Irish cenn, and the second element may possibly be connected with Gaelic fàl, Welsh gwawl, “rampart.” The names of the kings in the Pictish chronicles are not an absolutely trustworthy guide, as owing to the Pictish rule of succession the bearers of the names may in many cases have been Brythons. The names of some of them occur in one source in a Goidelic, in another in a Brythonic form. It is of course possible that the southern part of Pictish territory was divided between Goidels and Brythons, the population being very much mixed. On the other hand there are a number of elements in place-names on Pictish ground which do not occur in Wales or Ireland. Such are pet, pit, “farm” (?), for,fother, fetter, foder, “lower” (?). Aber, “confluence,” on the contrary, is pure Brythonic (Gaelic inver). Though the majority of scholars are of opinion that Pictish was nearly akin to the Brythonic dialects, we are entirely in the dark as to the manner in which that language was ousted by the Goidelic speech of the Dalriadic Scots. In view of the comparatively unimportant part played for a considerable period in Scottish affairs by the colony from Ireland, it is well-nigh incredible that Pictish should have been supplanted by Gaelic.

Authorities.—J. Rhys, Celtic Britain (London², 1905), The Welsh People (London3, 1902), “The Language and Inscriptions of the Northern Picts,” in Proceedings of the Society of Antiquaries of Scotland (1892); H. Zimmer, “Das Mutterrecht der Pikten,” in Savignys Zeitschrift (1895); also trans. by G. Henderson in Leabhar nan Gleann (Inverness, 1898); W.F. Skene, Celtic Scotland (Edinburgh, 1876); A. Macbain in appendix to reprint of Skene’s Highlanders of Scotland (Stirling, 1902); A. Macbain, “Ptolemy’s Geography of Scotland,” in Transactions of the Gaelic Society of Inverness, xviii. 267-288; W. Stokes, Bezzenbergers Beiträge, xviii. 267 ff.; H. d’Arbois de Jubainville, Les Druides et les dieux celtiques à forme d’animaux (Paris, 1906). The various theories have been recently reviewed and criticized by T. Rice Holmes in an appendix to his Caesar’s Invasion of Britain (London, 1907).

IV. History of Celtic Philology.—For many centuries the affinities of the Celtic languages were the subject of great dispute. The languages were in turn regarded as descended from Hebrew, Teutonic and Scythian. The first attempt to treat the dialects comparatively was made by Edward Lhuyd in his Archaeologia Britannica (Oxford, 1707), but the work of this scholar seems to have remained unnoticed. A century later Adelung in Germany divided the dialects into true Celtic (= Goidelic) and Celtic influenced by Teutonic (= Brythonic). But it took scholars a long time to recognize that these languages belonged to the Indo-European family. Thus they were excluded by Bopp in his comparative grammar, though he did not fail to notice certain resemblances between Celtic and Sanskrit. James Pritchard was the first to demonstrate the true relationship of the group in his Eastern Origin of the Celtic Nations (London, 1831), but his conclusions were not accepted. As late as 1836 Pott denied the Indo-European connexion. A year later Pictet resumed Pritchard’s arguments, and Bopp himself in 1838 admitted the languages into the charmed circle, showing in an able paper entitled Über die keltischen Sprachen that the initial mutations were due to the influence of terminations now lost. But it was reserved to a Bavarian historian, J.C. Zeuss (1806-1856), to demonstrate conclusively the Indo-European origin of the Celtic dialects. Zeuss, who may worthily rank with Grimm and Diez among the greatest German philologists, rediscovered the Old Irish glosses on the continent, and on them he reared the magnificent structure which goes by his name. The Grammatica Celtica was first published in 1853. The material contained in this monumental work was greatly extended by a series of important publications by Whitley Stokes and Hermann Ebel, so much so that the latter was commissioned to prepare a second edition, which appeared in 1871. Stokes has rendered the greatest service to the cause of Celtic studies by the publication of countless texts in Irish, Cornish and Breton. In 1870 the Revue celtique (vol. xxviii. in 1908) was founded by Henri Gaidoz, whose mantle later fell upon H. d’Arbois de Jubainville. In 1879 E. Windisch facilitated the study of Irish by publishing a grammar of Old Irish, and a year later a volume of important Middle Irish texts with an exhaustive glossary, the first of its kind. Since then Windisch and Stokes have collaborated to bring out some of the greatest monuments of Irish literature in the series of Irische Texte. The text of the Würzburg glosses was published by Zimmer (1881) and by Stokes (1887), and that of the Milan glosses by Ascoli. An important step forward was the discovery of the laws of the Irish accent made simultaneously by Zimmer and Thurneysen. This discovery led to a thorough investigation of the difficult verb system of Old Irish—a task which has largely occupied the attention of Strachan in England, Thurneysen and Zimmer in Germany, and Pedersen and Sarauw in Denmark. In a sense the publication of theThesaurus Palaeohibernicus (Cambridge, 1901-1903) may be regarded as marking the close of this epoch. The older stages of Irish have hitherto so monopolized the energies of scholars that other departments of Celtic philology save Breton have been left in large measure unworked. J. Strachan had begun to tap the mine of the Old Welsh poems when his career was cut short by death. J. Loth and E. Ernault have concentrated their attention on Breton, and can claim that the development of the speech of Brittany has been more thoroughly investigated than that of any other Celtic language. The number of periodicals devoted entirely to Celtic studies has increased considerably of recent years. In 1896 K. Meyer and L. C. Stern founded the Zeitschrift für celtische Philologie (now in its 7th volume), and in 1897 the Archiv für celtische Lexikographie began to appear under the direction of K. Meyer and W. Stokes. As a supplement to the latter Meyer has been publishing his invaluable contributions to Middle Irish lexicography. In Ireland a new periodical styled Ériu was started by the Irish School of Learning in 1904. The Scottish Celtic Review, dealing more particularly with Scottish and Irish Gaelic, began to appear in 1903, and the Transactions of the Gaelic Society of Inverness are in the 26th volume. For Wales we haveY Cymmrodor since 1877, and the Transactions of the Hon. Society of Cymmrodorion since 1892, and for Brittany the Annales de Bretagne, published by the Faculty of Letters at Rennes (founded 1886).

See V. Tourneur, Esquisse d’une histoire des études celtiques (Liége, 1905).