Ancient Britain and the Invasions of Julius Caesar

by T. Rice Holmes

VII. THE ‘PICTISH QUESTION’

A view which has become fashionable of late years, owing to the influence of Professor Rhys and Professor Zimmer, is that the [dolichocephalic] neolithic people of this country were identical with the Picts,[1775] whose name first occurs in the panegyric addressed about A.D. 296 to Constantius Caesar.[1776] To clear the ground, I should say, first, that it is universally admitted that descendants of the neolithic race survived not only in the part of Scotland which was inhabited by the Picts but in most parts of Britain. The question is whether the Picts represented that race in a special sense, and still spoke the neolithic non-Aryan language. As we shall see, Professor Rhys himself, who maintains that they did, emphatically affirms that among the medley of tribes who were known as Picts some were Celtic and spoke a Celtic tongue. Secondly, it may be well to state certain elementary facts of Celtic phonology (although I dare say that to most of those who may read these pages they are already familiar), without a knowledge of which parts of the following discussion and of the later section on the Celts would be unintelligible. The ancient Gauls, for the most part,[1777] and the Brythons, from whose dialect modern Welsh is descended, are commonly called the P Celts; while the Goidels, whose dialect was the ancestor of Gaelic, Irish, and Manx, are known as the Q Celts. The reason of this distinction is that the Gauls and Brythons changed the original sound qu into p, while the Goidels retained it, and in the sixth century of our era modified it into c.[1778] It has been affirmed, however, on the evidence of the formularies of Marcellus of Bordeaux, that some of the Western Gauls in the fourth century spoke a dialect which was akin to Goidelic;[1779] and Professor Rhys and Mr. Nicholson[1780] regard the words Sequani and Sequana (the Gallic name of the Seine) as proving that this dialect was not confined to the west: but M. d’Arbois de Jubainville refuses to admit that these names are Celtic,[1781] and contemptuously denies that the formularies are to be taken seriously.[1782] Professor Rhys[1783] and Mr. Nicholson[1784] also infer from certain inscriptions found in the departments of the Ain and Deux-Sèvres, which probably belong respectively to the first and the fourth century of our era, that a dialect akin to Goidelic was spoken in those localities: but here again M. d’Arbois dissents;[1785] and he remarks that an inscription found at Géligneux in the department of the Ain contains a word, petru-decametos,[1786] which belongs to the language of the P Celts. Professor Rhys urges that ‘the presence of monuments in the language occupying the subordinate position may be taken as evidence presumptive of its being the vernacular in the immediate neighbourhood’:[1787] but, as we shall see hereafter,[1788] a pillar, bearing a Goidelic inscription, has been found at Silchester, where the vernacular was undoubtedly Brythonic; and the obvious explanation is that the inscription was the work of a stranger. M. d’Arbois,[1789] moreover, unlike Professor Rhys, maintains that when the Celts first invaded Britain, the Celtic language everywhere was one and the same: according to him, none of the Celts had then changed q into p, but that change was made at a later date by the Celts who conquered Gaul, and some of whose descendants afterwards conquered Britain. Until near the end of the nineteenth century Celtic scholars unanimously believed that all the Celtic dialects had rejected ‘Indo-European p’, except, as Mr. Nicholson says,[1790] ‘in borrowed words or in certain combinations of consonants’; in other words, that wherever the Indo-European or Aryan tongue from which Celtic was descended had the sound of p the Celtic dialects had all lost it: but Professor Rhys holds that Mr. Nicholson has proved from the above-mentioned inscriptions, found in the departments of the Ain and Deux Sèvres, that it was retained by the Sequani and the Pictones.[1791] M. d’Arbois de Jubainville of course rejects this conclusion; and he reminds his opponents that p is absent from all Ogam inscriptions.[1792]

1. In 296, when the panegyric addressed to Constantius was written, the Picts to whom the writer referred were confined to the part of Scotland which extends northward from the firths of Forth and Clyde; but Professor Rhys and Professor Zimmer maintain that the habitat of the Pictish people was once much more extensive. ‘Irish literature,’ says Professor Rhys,[1793] ‘alludes to Picts here and there in Ireland ... in such a way as to favour the belief that they were survivals of a race holding possession at one time of the whole country.’ That the Picts once inhabited the whole of Britain is proved, in the opinion not only of the two professors but also of M. d’Arbois de Jubainville, who differs from them on the question of Pictish ethnology, by the following linguistic facts.[1794] The Irish name of the Picts was Cruthni.[1795] Britain has, since the Middle Ages, been called in Welsh ynys Prydein: Prydein is the Welsh equivalent of Cruthni; and ynys Prydein means ‘the island of the Picts’. Now, as Professor Rhys remarks,[1796] Prydein, with its cognate forms, Prydain, Prydyn, and Pryden, represents an old Welsh word Priten; and accordingly, the Brythonic or the Gaulish name of the Picts, when it reached the ears of the Greeks, would have been written by them Πρετανοί. It must of course be borne in mind that Cruthni, Prydain, and Priten did not appear in literature until long after Caesar’s time; but the etymology which connects Πρετανοί and Πρεταν(ν)ικαὶ (νῆσοι)—the name by which Ptolemy and other Greek writers call the British Isles[1797]—with Priten is accepted by Celtic scholars who, on the question of the ethnology of the Picts, differ widely among themselves. M. d’Arbois de Jubainville[1798] concludes that in the time of Pytheas the masters of Britain were the Picts; while Professor Rhys holds that when, shortly before that epoch, the Brythons first landed in Britain,[1799] not the Picts but the Goidelic Celts were the dominant race. In other words, he believes that the Goidelic Celts called the island which they conquered by ‘some such a Goidelic name as Inis Chruithni, “Island of the Picts”’.[1800] M. d’Arbois identifies the Picts of the time of Pytheas with the ancestors of the Goidelic Celts: like Professor Rhys he regards the word Pretani as simply the Brythonic, or Gaulish form of a Goidelic word Qrtanoi, of which Cruthni was the later Irish equivalent;[1801] but he holds that no Brythons had set foot in Britain until after the time of Pytheas, and that the word Pretani was learned by Pytheas not in Britain but in Gaul.

Both the views that have just been stated seem to involve difficulties. If Professor Rhys is right in believing that the pre-Roman Goidelic invaders of Britain (whose very existence, as we shall afterwards see, is not universally admitted) called the people whom they found in possession by some such name as Chruithni or Cruthni, the name which, transformed by Brythons into Pretanoi, was applied by Pytheas to the inhabitants of Britain generally, it would appear either that the Goidelic invaders had no name of their own or that it was suppressed.[1802] Moreover, Professor Rhys does not explain how it happened that Pytheas never learned the name by which, as he tells us, the Brythons called themselves, namely, Brittones. On the other hand, M. d’Arbois’s view would compel us either to assign the first Brythonic invasion to a date a century later than that which is now generally accepted,[1803] or to assume that Pytheas, although he visited Britain, learned nothing there of the name of its inhabitants. I confess that I cannot suggest any satisfactory solution.

It remains to be inquired whether the Picts of history did really, in a special sense, represent the neolithic population, and whether they spoke a non-Aryan language.

2. Was the word Pict, in its original form, pre-Aryan or Celtic? The answers that have been given to this question only serve to amuse the ignorant scoffer, and to illustrate the truth that even if the labours of Zeuss placed the study of the ancient Celtic languages upon a scientific basis, Celtic scholars still know very little about them. When we inquire of Professor Rhys, we are perplexed by the quick changes of front to which his most devoted disciples have by this time become accustomed. In the second edition of his Celtic Britain[1804] he said that ‘neither the Picts nor the Scotti probably owned these names, the former of which is to be traced to Roman authors’; and he described the theory which ‘connected the Pict with the Gaulish Pictones’ as a ‘clumsy invention’.[1805] In his Rhind Lectures he assured us that ‘the principal non-Aryan name of the inhabitants of both islands [Great Britain and Ireland] was some prototype of the word Pict’,[1806] and gave reasons, which are now generally accepted, for believing that that name was not connected with the Latin pictus.[1807] At the same time he definitely committed himself to the view which he had previously derided as a ‘clumsy invention’, and affirmed that ‘the word Pict ... is hardly to be severed from the Pictones of ancient Gaul’. In The Welsh People, which first appeared in 1900, and in a later edition of the same work, dated 1902,[1808] he argued that ‘Ictis [the name of an island mentioned by Diodorus Siculus[1809]] and Icht [the old Irish name of the English Channel] represent possibly a Celtic pronunciation of the same Aboriginal word which the Romans made into Pictus ... we must’, he added, ‘suppose it an early name which the Aborigines adopted, while the Celts ... applied another name Qṷrtani, Pretani, Cruithni,’ &c. But in the same year in which the first edition of The Welsh People appeared he told the members of the British Association that ‘pictos was a Celtican word of the same etymology, and approximately, doubtless, of the same meaning as the Latin pictus; that the Celticans had applied it at an early date to the Picts on account of their ... tattooing themselves; and that the Picts had accepted it’.[1810] It is not absolutely clear whether by ‘the Celticans’ he means only those people of Gaul who spoke a language akin to Goidelic or the first Celtic invaders of Britain. As, however, we are told that the Picts accepted their name from ‘the Celticans’, it would seem that those ‘Celticans’ were, or at all events included, the British Goidels; and we ask ourselves in bewilderment why, if the ‘Celticans’ applied the name pictos to the Picts, they also applied the name Qṷrtani.[1811] But when we open the latest edition of Celtic Britain,[1812] we find that the professor’s views are still in process of development, or of flux. He now reverts to the theory that ‘the native name which suggested the Latin [Pictus] was not of Celtic origin either, though only found treated as Celtic’. He adds that ‘the term Pictones, as occurring in Gaul in Caesar’s time, makes it probable that it was also a name of long standing in Britain’; and finally he avows with characteristic candour that ‘we know not from what language it comes’. Turning to our other authorities, we learn from Zimmer that Picti is obviously a Latin translation of the name [the ancestor of Prydain] which the Romans learned from the Britons.[1813] In other words, the German savant holds that the word Pictos [if it ever existed except as a Latin accusative plural] was neither aboriginal in Britain, nor Celtican. It has been suggested[1814] that Picti is connected with the old Irish word cicht,[1815] a carver or engraver, and is the Cymric form of a Goidelic word Qicti;[1816] while Mr. Nicholson, who insists that Picti is not Cymric but Goidelic, claims to have ‘fully shown that this name is ... from the root peik- “tattoo”, with Ind.-Eur. p preserved’.[1817]

The one absolutely certain conclusion to which the student of ethnology can come is that the name of the Picts has not been proved to be of pre-Aryan origin.

3. Still, Professors Rhys and Zimmer will have it that the Picts must have been a non-Aryan people. Caesar,[1818] in a well-known passage, states that among the Britons groups of ten or twelve men had wives in common; in other words, that one of the British customs was polyandry. It has generally been assumed that he meant to say that the custom was prevalent among the Britons generally; but Zimmer, after reviewing the whole chapter in which the passage occurs, concludes that it refers only to interiores—the Britons of the interior[1819]—whom Caesar contrasts with maritimi,—the descendants of the Belgic invaders. The latter, he argues, according to Caesar’s express statement, differed but slightly in their customs from the Gauls:[1820] therefore the words in which Caesar describes the British custom of polyandry cannot refer to them, but must refer to interiores.[1821] The two professors agree in thinking that Caesar, owing to his ‘inability to realize a state of society exclusively based on birth’,[1822] misunderstood the institution which he tried to describe; in other words, that that institution was not polyandry but matriarchy,—the rule of succession by which rank and property are transmitted in the female line; a king, for example, being succeeded not by his own son but by the son of one of his sisters.[1823] Zimmer, referring to Schrader’s Prehistoric Antiquities of the Aryan Peoples,[1824] remarks that among all Aryan-speaking peoples and among the primitive Aryans the custom by which a father is succeeded by his own son (das Vaterrecht) was the foundation of social ordinance.[1825] Professor Rhys,[1826] indeed, thinks that this generalization cannot be proved, and refers to a well-known passage in the 20th chapter of the Germania of Tacitus,—‘Sisters’ sons are held in as much esteem by their uncles as by their fathers: indeed, some regard the relation as even more sacred and binding’,[1827] &c. (Sororum filiis idem apud avunculum qui apud patrem honor: quidam sanctiorem artioremque hunc nexum sanguinis arbitrantur); but he suggests that the tribe of which Tacitus speaks may have been mixed with some ‘aboriginal race practising the same institution as the aborigines of the British Isles’. And I suggest that the Picts were Celts mixed with aborigines who practised this same institution, and consequently that if it prevailed among the Picts, its prevalence does not prove that they were in any special sense representatives of the aborigines, or that they spoke a non-Aryan language.[1828]

Having corrected Caesar’s narrative to his own satisfaction, Professor Rhys sets himself to prove that matriarchy was a Pictish institution. He observes[1829] that ‘a Pictish king [during the later period of the Roman occupation and afterwards] could not be succeeded by a son of his own, but usually by a sister’s son. The succession,’ he continues, ‘was through the mother, and it points back to a state of society which, previous to the conversion of the Picts to Christianity, was probably based on matriarchy as distinguished from marriage and marital custom.’ To show that matriarchy had formerly prevailed in Britain outside the territory within which the Picts of history were confined, he adds[1830] that ‘the ancient literature of Ireland abounds in allusions to heroes who are usually described with the aid of the mother’s name’, and that ‘this kind of nomenclature implies the Pictish succession as its origin’. Again, he quotes an inscription found at Colchester, which ends with the words

DONVM. LOSSIO. VEDA. DE SVO POSVIT. NEPOS. VEPOGENI. CALEDO.

(‘This gift has been dedicated at his own expense by Lossio Veda, the son of the sister [?] of Vepogen, a Caledonian’), and remarks that when Lossio calls himself a Caledonian, that ‘is for our purpose much the same as if he had called himself a Pict’, and that, moreover, both Veda and Vepogeni ‘may be said to occur in the list of Pictish kings’, where the latter is ‘written Vipoig’. Vepogeni, indeed, is a Celtic word, borrowed, the professor assures us, in accordance with Pictish custom; but ‘the reduction of Vepogen to Vepog, which is what underlies Vipoig, is impossible on Celtic ground ... while Pictish offers a simple and natural explanation’.[1831]

Professor Morris Jones remarks, in support of Professor Rhys’s argument, that ‘the Pictish succession’ has ‘come down to our own times among the Berbers’[1832] (or rather Kabyles), who, he says, have been shown, on craniological grounds, to be akin to our neolithic race.

Apparently Professor Rhys does not regard the custom of reckoning descent ‘by birth alone’ as confined in these islands to the Picts, or to the pre-Aryan aborigines: if, as he is inclined, like Professor Zimmer, to believe, it was non-Aryan, ‘it must,’ he says, ‘have been accepted by the Goidelic Celts from the aborigines.’[1833]

Now, in regard to this last observation, the comment suggests itself that what Professor Rhys has not yet proved is that those aborigines were Picts. The Picts, as we shall presently see, were, according to some Celtic scholars, themselves Goidelic Celts (mixed of course with aborigines whom they had subdued and Celticized); according to others, their speech was akin to Brythonic.[1834] And if, as Professor Rhys insists, matriarchy may have been accepted by the Celts from the aborigines, it is perhaps not incredible that, as Mr. Sidney Hartland suggests, the Celts themselves, in prehistoric times, may have passed through the matriarchal stage,[1835] and that the survival of matriarchy among the Picts is not necessarily attributable to pre-Aryan ancestry.[1836] But, be that as it may, the survival of matriarchy among the Picts proves nothing more than that among the Picts, as among every other British people, the substratum of the population was pre-Aryan: it does not prove that the dominant element among them was pre-Aryan, or that they spoke a non-Aryan language.

As for Professor Morris Jones’s argument, it may perhaps raise a probability that the ‘Pictish succession’ prevailed among the neolithic race, although, if the argument is worth anything, the professor ought to be able to show that the same institution belonged to the ‘Iberians’ of Spain, of Gaul, and of other countries who have also been shown ‘on craniological grounds’ to be akin to the Kabyles: but at all events it lends no support to the theory that the Picts were, in any special sense, descendants of the neolithic aborigines; for, assuming that they were Celts, they might have accepted the Pictish succession from them. There remains Professor Rhys’s statement that ‘the reduction of Vepogen to Vepog, which is what underlies Vipoig, is impossible on Celtic ground’. Is the professor quite sure? A few years ago he would certainly have said that the retention of ‘Indo-European p’ was ‘impossible on Celtic ground’; but in 1900 he announced that the ‘Celtican language’ which was spoken in the country of the Sequani ‘preserves intact the Aryan consonant p’.[1837] He has himself assured us that both the Celtic dialects spoken in the British Isles were greatly modified by a pre-Aryan language.[1838] Assuming, for the sake of argument, that the Pictish language was Celtic, is he prepared to deny that it could have been so far modified by a non-Aryan tongue that ‘the reduction of Vepogen to Vepog’ would still have been ‘impossible on Celtic ground’?[1839] Finally, when he tells us that Lossio’s description of himself as a Caledonian ‘is for our purpose much the same as if he had called himself a Pict’, we cannot help recalling his own statement[1840] that ‘the Caledonians were, as we understand their history, Goidels’; though, to be sure, in the latest edition of Celtic Britain[1841] he expunges this compromising sentence, and substitutes for it ‘the Caledonians were Picts’.

For my part I accept the professor’s emendation unreservedly. Picts the Caledonians certainly were; for does not the author of the panegyric addressed to Constantine speak of ‘the Caledonians and other Picts’?[1842] But for me the Picts were a mixed people, comprising descendants of the neolithic aborigines, of the Round Barrow race, and of the Celtic invaders,—a mixed people who spoke a Celtic dialect. And what puzzles me is that the professor should not have been struck by the anthropological facts that are fatal to the theory that the Caledonians were Picts in the sense which he attaches to the word,—that is, pure survivors of the neolithic aborigines, who spoke a non-Aryan language. For the neolithic aborigines, as we have seen, were, speaking generally, small dark men of the ‘Iberian’ type: the Caledonians were big fair or red-haired men. Doubtless there were, as I have said, ‘Iberian’ survivors among them; but who will deny that the powerful race whom Tacitus describes were predominant, or that their Aryan tongue had prevailed?[1843]

4. It is usually inferred from statements in Claudian[1844] and Herodian[1845] that the Picts tattooed themselves; and their testimony is supposed to be strengthened by the etymology of the names by which the Picts were known to the Irish and Welsh respectively,—Cruthni and Prydain. The former is said to be derived from cruth,[1846] the Gaelic word for ‘form’ or ‘shape’; and the latter from its Welsh equivalent, pryd.[1847] Thus Cruthni and Prydain would mean ‘the people whose bodies were decorated with figures’; and, as we have seen, Zimmer has no doubt that the Roman name for the Picts—Picti, or ‘painted men’—was simply a translation of Prydain or its older equivalent. Professor Rhys, who, in one of his many and diverse utterances on the subject, affirmed that pictos was a Celtican word,[1848] drew this conclusion from the fact, pointed out by Mr. Nicholson,[1849] that a coin of the Gallic tribe of the Pictones[1850] bears on the obverse a tattooed face; and he supposes that the reason why the Celticans applied this word to the Picts was that the latter tattooed themselves. ‘The Picts of Britain and Ireland,’ he remarks, ‘are found also called Pictones’; and ‘ancient Egyptian monuments represent the Libyans of North Africa with their bodies tattooed’.[1851]

Now what does this community of custom prove about the ethnology of the Picts? The inhabitants of the Tonga and Society Islands and of New Guinea tattoo themselves: so do the Burmese, the Shans, the Maoris, and the people of British East Africa;[1852] so do very many Englishmen. All the available evidence tends to show that among the ancient inhabitants of the British Isles tattooing was not confined to the Picts. Herodian does not mention the Picts at all: he merely says that the Britons tattooed themselves. Professor Rhys admits, or rather strenuously maintains, that in the territory inhabited by the Picts in Scotland there were also numerous Celts;[1853] and he would hardly deny that they were included among the people whom Herodian describes. He himself remarks that ‘the Scotti (that is to say the Goidels)’[1854] practised tattooing.[1855] Mr. Nicholson, to whom he appeals, argues from the evidence of coins that tattooing was customary not only among the Pictones, but also among several other tribes of Gaul,—the Ambiani, the Baiocasses, the Caletes, the Coriosopites, the Osismi, the Sequani, and the Unelli. All these peoples were undoubtedly Celtic; that is to say, they were Celtic-speaking tribes among whom the Celtic element, ethnologically speaking, was, I do not say numerically, but politically predominant. Professor Rhys would certainly not argue that they were Picts: yet if he admits, as he does, that they were Celtic, the argument which he bases on the practice of tattooing collapses.

5. Some years ago Professor Rhys attempted to prove that the Pictish language was related to Basque;[1856] ‘but,’ he says, ‘whether it is related or not, my attempt to prove that it is has been pronounced, and doubtless justly pronounced, a failure.’[1857] At the same time, however, pointing to a famous ogam inscription, he wrote, ‘my challenge still remains, that if Pictish resembled Gaelic or Welsh, or in fact any Aryan language, those who think so should make good their opinion by giving us a translation of such an inscription, for instance, as the following from Lunasting, in Shetland:—Xttocuhetts : ahehhttmnnn : hccvvevv : nehhtonn.’[1858]

The lay reader will perhaps mentally endorse the comment of another Celtic scholar, Dr. Alexander Macbain, who disposes of the cacophonous puzzle by observing that ‘it is neither Welsh nor any other language’.[1859] For the present, at all events, it is safe to say that Dr. Macbain is as likely to be right as Mr. Nicholson, who, having boldly accepted Professor Rhys’s challenge, first judiciously reconstructed the text of the inscription, and then made an heroic attempt to translate his own version. It is Goidelic, so he assures us; and it means

‘Place of O’ Cuhetts his place within: CUAIBH of Nehton’.[1860]

On the other hand, the translation which Professor Rhys ‘provisionally’ offers of his text runs

‘“Kin—Ahehhtmnnn King Nechtan”.

That is to say, King Nechtan of the kin of Ahehhtmnnn’.[1861]

Perhaps it shows a slight lack of humour to attempt, even ‘provisionally’, to translate an inscription assumed to be written in a language the very existence of which is doubtful. Still it is conceivable that Professor Rhys’s text means what he says. But, supposing that it resembles neither Gaelic, nor Welsh, nor any Aryan language, what does it prove? Not that the Picts represented the neolithic aborigines, but simply that in the remotest of the British isles there still survived the non-Aryan language which, as every scholar admits, was once spoken in Britain.

But the truth is that the so-called Pictish inscriptions, even in the hands of the philologist, are so intractable that for ethnology they are practically useless. ‘I can hardly do more,’ says Professor Rhys,[1862] ‘than pick from previous attempts by others and by myself what seems to me the most probable reading.’ This is only one of numerous instances in his well-known article on the inscriptions which show how impossible it is to construct the text with any approach towards certainty.

Professor Rhys remarks, further,[1863] that ‘we have indications in Adamnan’s Life of Columba that [in the sixth century of our era] the language of the aborigines was still a living tongue’. The indications are that when Columba, who spoke Goidelic, visited the province of the Picts, he preached ‘to peasants or plebeians by interpreter’. To those who hold, with Dr. Whitley Stokes and Dr. Macbain, that the Pictish dialect was akin to Brythonic, the fact on which Professor Rhys lays stress presents of course no difficulty. M. d’Arbois de Jubainville, however, while he agrees with Dr. Macbain,[1864] makes a reply to Professor Rhys which might be used by those who hold, with Mr. Nicholson, that Pictish was akin to Goidelic. He tells a story of a Breton priest of the diocese of Quimper who assured him that he himself could not understand the Breton dialect of a woman who belonged to the diocese of Vannes.[1865]

Mr. Nicholson[1866] says that ‘we have abundant materials for deciding whether Pictish was or was not (1) Aryan, (2) Keltic, (3) Goidelic, in (a) the place-names recorded by ancient geographers and one or two mediaeval documents, (b) the person-names given by one or two ancient historians and in mediaeval chronicles, (c) the inscriptions’. From these materials Mr. Nicholson undertakes to demonstrate that Pictish was Goidelic, and that ‘it stands to Highland Gaelic in exactly the same relation in which Anglo-Saxon stands to modern English’;[1867] while Dr. Whitley Stokes[1868] and Dr. Macbain[1869] undertake with equal confidence to demonstrate that it was related to Brythonic. According to Bede,[1870] the place which marked the western termination of the wall of Severus was called in Pictish Peanfahel. Pean is commonly identified with the Welsh word penn, ‘a head’; and accordingly it has been inferred that Pictish was ‘a Kymric or semi-Kymric dialect’.[1871] Mr. Nicholson, on the other hand, claims to have shown that Pean is ‘a Goidelic borrowing from the Latin penna or pinna’. Professor Rhys[1872] formerly clung to the view that Peanfahel was a Brythonic name, but was not in the least disconcerted thereby; for, he explained, ‘the Picts must have learnt it ... from the Verturian Brythons.’ On the question of etymology he has now become a convert to Mr. Nicholson’s view:[1873] but on the question of ethnology he retains his own opinion; for, he explains, ‘The non-Celtic Picts, when we find them coming southwards, seem to have been fast adopting the idioms of their neighbours.’[1874] Mr. Nicholson[1875] analyses with laborious ingenuity a large number of names in Adamnan’s Life of Columba, of place-names in the Pictish Chronicle, of Pictish historical names, and of words which occur in the ‘Pictish inscriptions’, and insists that they are Goidelic: Dr. Whitley Stokes[1876] and Dr. Macbain[1877] produce words from the same sources, from Ptolemy’s Geography, and from Dion Cassius, and insist that they are Brythonic. Dr. Stokes’s authority is so great that his verdict is worth quoting:—‘The foregoing list of names and other words contains much that is still obscure; but on the whole it shows that Pictish, so far as regards its vocabulary, is an Indo-European and especially Celtic speech. Its phonetics, so far as we can ascertain them, resemble those of Welsh rather than of Irish.’[1878]

But the arguments for Brythonic, on the one hand, and for Goidelic, on the other, leave Professor Rhys unmoved. Prove as many Pictish words as you please to have been Goidelic, as many as you please to have been Brythonic: he will regard them with serene indifference.[1879] For, he tells you,[1880] ‘the Pictish language would seem to have been rapidly becoming overloaded with loan-words from Goidelic or Brythonic when we first hear anything about it. So, failing to recognize this borrowing of words by the Picts, some have been led to regard Pictish as a kind of Gaelic, and some as a dialect akin to Welsh. The point to have been decided, however, was not whether Gaelic or Welsh explains certain words said to have been in use among the Picts, but whether there does not remain a residue to which neither Gaelic nor Welsh, nor, indeed, any Aryan tongue whatever can supply any sort of key.’ The professor is still thinking of that outlandish inscription which, according to Mr. Nicholson, is Goidelic, and the professor’s reading of which, according to Dr. Macbain, is no language at all. But, admitting provisionally the existence of ‘a residue’ to which no Aryan language ‘can supply any sort of key’, we should, I must repeat, only have to conclude that in certain remote parts of the extensive territory occupied by the Picts a non-Aryan language survived into the Christian era, just as in a remote part of France a non-Aryan language survives at this day: we should not have to conclude that that language was spoken by the Picts in general. ‘La question,’ says M. d’Arbois de Jubainville, in a notice of Professor Rhys’s article on the Pictish inscriptions,[1881] ‘la question est de savoir si cette population [the pre-Aryan population] est restée dominante. Les noms de peuples tels que Smertae ... des noms d’hommes tels que celui du Calédonien Argentocoxos ... me semblent décisifs.’ It is absolutely certain, and is insisted upon by Professor Rhys himself, that in Roman times many of the tribes which were included under the general designation of Picts bore Celtic names, and that many of the geographical names in the country which they inhabited were Celtic also. On the other hand, not a single Pictish name, tribal, or geographical, or personal, not a single Pictish word which has been preserved by Ptolemy or by our other authorities, has been proved to be non-Celtic; and if, as Professor Rhys maintains, Pictish was a non-Aryan language overlain by loan-words from the two Celtic dialects, it was so buried beneath them as to be no longer discernible. Argentocoxos,[1882] as the professor says, was a Pict, and one of the many Picts whose names were Celtic: if the Picts had spoken a non-Celtic language, however much overloaded with Celtic loan-words, would not their own names have been non-Aryan? As their names were Celtic, it is reasonable to infer that their language was Celtic also. The professor, it is true, points out that ‘in Wales many a man has the English name John Jones, though he cannot speak English’.[1883] Yes, but the Welsh are a conquered or, let us say, absorbed people, whereas the professor himself assures us[1884] that before the time of Ptolemy ‘the Goidels and the Picto-Brythons [of the North] had come under the power of the more purely non-Celtic tribes beyond them’.[1885] But this is of course a pure assertion. The professor fails to prove that any Celtic people in Britain came under the power of non-Celtic tribes. Many centuries before the time of Pytheas the neolithic population had for the most part been reduced to subjection; and, although remote clans may possibly have retained their individuality, in many parts of the island the descendants of the aborigines had become intermingled, first with the ‘Round Barrow’ invaders, the earlier of whom at all events, as I shall presently show,[1886] were not Celts, and secondly with the Celts themselves. Professor Rhys[1887] himself admits that the name of the Picts ‘was never, perhaps, distinctive of race, as Brythons and Goidels seem to have been sometimes included under it’; and, although he goes on to say that ‘the term probably applied most strictly at all times’ to ‘the non-Celtic natives’, it is not likely that the name of non-Celtic natives should have prevailed over that of the Celts.

For all these reasons it appears to me infinitely more probable that in Pictland as, according to Professor Rhys himself, in the rest of Britain,[1888] the non-Aryan language should have been absorbed by Celtic than that Celtic should have been absorbed by the non-Aryan language.

There is probably this grain of truth in Professor Rhys’s theory, that the non-Celtic natives continued to exist in greater purity in the country which was occupied by a group of tribes who, during the latter part of the Roman occupation and afterwards, were called Picts, than in any other part of Britain. But I doubt whether this eminent scholar could have spent his time less profitably than in striving to demonstrate, first, that the language of the Picts was related to Basque, and, when he was forced to abandon this attempt, in clinging to the theory that it was a non-Aryan tongue.