of all its parts, and discerns that immediate expression of reality which emanates from life, he has a right to demand that others, who merely throw their looks by the way on the region where he lives and has taken up his home, should not deny the correctness of his views because they perceive nothing of the kind. The learned naturalist, who has never left his native town, will not recognise the animal's track by which the hunter is guided; and if any one, on going into Benvenuto's prison, when his eyes had for months been accustomed to see the objects around him, had asserted that Benvenuto, like himself, could not distinguish anything in the darkness, he would surely have been somewhat presumptuous."

External evidence is usually the clearer and more precise in its intimations, as well as the more obtrusive or the more readily come by; it is in these respects like other superficial or outside things; but internal evidence, when its interpretation is free from doubt, is the more trustworthy and conclusive. The mind, however, is not satisfied without a concurrence of the two kinds of evidence whenever the case seems to admit of it. And, when there is such a concurrence, the one kind of evidence will often enable us to carry our deductions and conclusions farther than the other.

It is very rarely, if ever, that internal evidence is absolutely wanting; external evidence frequently is. A familiar instance of evidence which is purely internal is that with which Paley sets out in his work on Natural Theology, of a watch in motion found by a person who had never seen or heard of such a contrivance, but who at once and without any doubt infers it to be the work

of an intelligent and designing mind. His inference to that extent could hardly have been strengthened by the addition of any amount of external evidence.

In such questions, however, as the authorship of an anonymous book, or the parentage of a picture the painter of which is unknown, the internal evidence, which we always have, and without which in such a case no external evidence would be sufficient to produce perfect conviction, at least to a mind of any critical sagacity, is generally rendered much more forcible by the support of external evidence.

It is the same with questions relating to the affiliation and connexion of languages. Here, too, the internal evidence, or that presented by the languages themselves, is indispensable; but such external evidence as is to be had is not to be disregarded. It demands, at least, always to be explained, and to be shown to be consistent with the internal evidence; and it sometimes serves as a useful index to the direction in which the internal evidence is to be looked for or pursued.

For the latter reason it is convenient in questions of this nature to begin with the consideration of the external or historical evidence.

II. The First of the facts constituting the External or Historical Evidence that we have in regard to the sources of the English language is, that the country in which it is spoken and has grown up appears to have been occupied at an early date, in whole or in part, by a Celtic population.

THE earliest express statement that has come down to us in regard to the language spoken in the country now called England is that of Tacitus, who, writing in the first century of our era, says (Agricola, 11) of those of the Britons of his day who were nearest to Gaul, that they were probably of Gallic extraction, and that their speech was not very different from that of the Gauls (sermo haud multum diversus).

It may be questioned, however, whether this statement, rightly understood, would warrant us in coming to the conclusion which it has been commonly held to establish; namely, that the portion of the British population which had been derived from Gaul was of Celtic race and spoke a Celtic tongue.

For Cæsar (Com. v. 12), writing a century before Tacitus, although he says nothing about the language of the Britons, in asserting as a fact what Tacitus advances as a probability, that the Britons dwelling along the coast opposite to Gaul had originally come from that country, particularises Belgium as the part of

Gaul whence they had emigrated; and elsewhere (i. 1, and ii. 4) he tells us that the Belga were for the greater part of Germanic descent, and that both they and the Aquitani differed in language, as well as in institutions and laws, from the proper Gauls, or Celts as they were called in their own tongue.

From all this it has been contended that the language of at least a part of the population of Britain, when the country first became known to the Romans, must have been not a Celtic but a Germanic language. This view was first proposed by the Scottish antiquary, Sir John Clerk of Pennicuick, in a “Dissertation on the Ancient Language of Britain," written in 1742, but first published in 1782, in the first volume of the Bibliotheca Topographica Britannica; and it was afterwards maintained, with much ingenuity and learning, by John Pinkerton, in his "Enquiry into the History of Scotland preceding the Reign of Malcolm III." (2 vols. 8vo. Edinburgh, 1789; republished in 1794 and in 1814). Pinkerton states (i. 363) that he had not happened to see Clerk's Dissertation till after the materials for his own work were collected.

Nevertheless, it is by no means universally admitted that the people of Belgic Gaul, whatever may have been their origin, spoke a Germanic language. It is contended that Cæsar's statement can only mean that they spoke a different dialect from the people of Celtic Gaul; and that, if they were Germans by descent, they had, after their settlement in Gaul, exchanged their ancestral speech for the common language of that country. In this undecided state of the question respecting the language of the Belgæ, recourse has been had, for evidence in regard to the earliest language spoken in

Britain, to the ancient topographical nomenclature of the country, that is, the oldest names of places and natural objects in it. These, which are always originally significant, are the surest evidence we can have in regard to the language spoken in any country at the time when they were imposed. The ancient topographical nomenclature of Britain is elaborately investigated by George Chalmers in the first volume of his Caledonia (3 vols. 4to., 1810-24); and the subject has also been more recently discussed by the late Rev. Richard Garpett in a paper printed in the Proceedings of the Philological Society (i. 119). Whatever differences of opinion may still exist upon subordinate points, there is now no dissent from the general conclusion arrived at by both of these writers, that the oldest topographical nomenclature everywhere in Britain is Celtic. This is the case in the parts of the country which Cæsar states to have been colonised from Belgic Gaul, as well as elsewhere. Kent, for instance, is a Celtic name, and Thames is a Celtic name. Mr. Garnett further holds the topographical nomenclature of France and that of ancient South Britain to belong to the same form of the Celtic, namely, the Cambrian, or Welsh; and he conceives that to be the earliest and least corrupted form now subsisting.

It should be observed, however, that the fact of the most ancient topographical nomenclature of the Belgic parts of Gaul and Britain being Celtic does not prove that the Belgic colonists in either case spoke a Celtic language; for the names may have been imposed by preceding occupants of Celtic race. It only proves that the parts occupied by the Belgic colonists must, as well as the rest of the country, have been at one time in the