So far so good; but if we go farther, we do not at first fare better. It would be grossly unjust to charge Mr Arnold with all the nonsense which has since been talked about Celtic Renascences; but I fear we cannot write all that nonsense off his account. In particular, he set an example, which has in this and other matters been far too widely followed, of speaking without sufficient knowledge of fact. It cannot be too peremptorily laid down that the literary equivalent of a "revoke"--the literary act after which, if he does it on purpose, you must not play with a man--is speaking of authors and books which he has not read and cannot read in the original, while he leaves you ignorant of his ignorance. _This_ Mr Arnold never committed, and could never have committed. But short of it, and while escaping its penalty, a man may err by speaking too freely even of what he confesses that he does not know; and of this minor and less discreditable sin, I own (acknowledging most frankly that I know even less of the _originals_ than he did), I think Mr Arnold was here guilty. Exactly how much Gaelic, Irish, or Welsh Mr Arnold knew at first-hand, I cannot say: he frankly enough confesses that his knowledge was very closely limited. But what is really surprising, is that he does not seem to have taken much trouble to extend it at second-hand. A very few Welsh triads and scraps of Irish are all that, even in translation, he seems to have consulted: he never, I think, names Dafydd ap Gwilym, usually put forward as the greatest of Celtic poets; and in the main his citations are derived either from _Ossian_ ("this do seem going far," as an American poetess observes), or else from the _Mabinogion_, where some of the articles are positively known to be late translations of French-English originals, and the others are very uncertain. You really cannot found any safe literary generalisations on so very small a basis of such very shaky matter. In fact, Mr Arnold's argument for the presence of "Celtic magic," &c., in Celtic poetry comes to something like this. "There is a quality of magic in Shakespeare, Keats, &c.; this magic must be Celtic: therefore it must be in Celtic poetry." Fill up the double enthymeme who list, I am not going to endeavour to do so. I shall only say that two sentences give the key-note of the book as argument. "Rhyme itself, all the weight of evidence tends to show, came into our poetry from the Celts." Now to some of us all the weight of evidence tends to show that it came from the Latins. "Our only first-rate body of contemporary poetry is the German." Now at the time (1867), for more than thirty years, Germany had not had a single poet of the first or the second class except Heine, who, as Mr Arnold himself very truly says, was not a German but a Jew.