materially 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 (< ā) in final syllables is reduced to o in Mod. W., e.g. Mid. W. pechawt < Lat. peccātum, 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-), -ā-, -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 < *kornī; 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 Offa, king of Mercia, who constructed the Dyke going by History and extent. 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 971% 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. GwenogvrynEvans, The Text of the Mabinogion (Oxford, 1887); The Myvyrian