and the infinitives volere, vīere, vulere, velere, vellere, with perhaps a note stating that these infinitives are “seldom used” (see his Gr.² 66, 68), or alternatively a footnote to the effect that velle “is as often used” (do. 67). Examples are quoted of such forms as are genuine; and the impression is conveyed by the suggestio falsi of “seldom”, “as often”, and the like, that the others also occur. To the author truth meant conformity with his theory; facts, perverse enough to disagree, were glossed over to save their character.
In 1853 appeared the first edition of Rowland’s work, which was regarded for more than a generation as the standard grammar of Modern Welsh. It is for the most part a description of the written Welsh of the 19th century; but the paradigms contain many of Pughe’s spurious forms. The author had practically no knowledge of any Welsh older than that of the Bible translation; he records recent usages, but is unable to throw any light on them, or to decide between genuine and counterfeit forms. The use which he makes of Dr. Davies often shows that he was incapable of understanding him; e.g. in professing to give Davies's table of diphthongs, after including iw wy among the falling diphthongs he imagines that he has done with those combinations, and omits them from the rising class, without perceiving that the very object of the classification is to distinguish between falling iw̯ w͡y and rising i̯w w̯y. But his book contains a quantity of sound, if ill-digested, information about Late Welsh; and marks the return to common sense after the domination of Pughe.
The foundations of modern Keltic philology were laid by I. C. Zeuss in his great Grammatica Celtica, which was published in 1853. The sections devoted to Welsh grammar contain a wonderfully complete and accurate analysis of the language of the Red Book Mabinogion (ed. Lady Charlotte Guest, 1849), the Black Book of Chirk (in a.l., 1841), and the Welsh passages in Liber Landavensis (ed. Rees, 1840).