application of the striving for distinctness which I have already mentioned as acting counter to ordinary phonetic change; I say, irregular, because it is impossible to predict in what way it may act. A great number of instances may be found in Cumberland where a Norse colony settled, probably in the latter half of the tenth century, and introduced words which in after time had a strange sound, and were identified with whatever English word they resembled. Thus "foss," a waterfall (as in Norway at the present day), was confused with "force" (fortis), and so we get Scale Force, &c.: the proper names Koli and Mioll are disguised in Coal Gill and Mill How, and the compound name Toli-Wagen has given us Dolly Waggon Pike upon Helvellyn'.

The results of this principle of change are very numerous in composite languages like the English. I do not however imagine that it operated much on the Greek and Latin languages. Greek in the stage at which we know it, could have little admixture which is not manifest at the first glance: and the Latin was not much more affected. In pure languages, I conclude (in spite of a few real and some apparent exceptions), phonetic change has a downward tendency; it causes in general weakening of the language, even though that weakening may be usefully employed. What then was the original, of which the Greek and Latin are copies, weakened each in its own peculiar way? This will be the subject of the next Chapter.

1 See Ferguson, Northmen in Cumberland, for these and many other etymologies.

CH. I.

P. E.

2

CH. I.

NOTE TO CHAPTER I.

ON THE DERIVATION OF LATIN WORDS FROM GREEK.

The facts are so very simple, yet there is so much misconception about them, that it seems worth while to say a word on the supposed derivation of Latin words from Greek. This theory is probably to be attributed to Niebuhr's hypothesis of a Greek and non-Greek element in the Latin language, which made its way into English works without much examination through the influence of Niebuhr's extraordinary genius; but which has been completely overthrown by Comparative Philology. The apparently Greek element in the Latin language is (generally speaking) that part of the common inheritance of the Greeks and Italians, which each nation retained and developed after the separation of the two branches of the original stock1. The apparently non-Greek element is that portion of the common inheritance which was neglected by the Greeks or, if retained by provincial and obscure dialects, was disused by those which possessed a literature; which therefore in process of time seemed to be to some extent actually waspeculiar to the Italians.

What then are we to say of words like lyra, &c.? Are not these derived from the Greek? Certainly not derived. No Latin word is derived from the Greek in the proper sense of the term. The Latin borrowed words fully formed from the Greek, which it spelt on different principles according to the different times at which they became nationalised. At the earliest period at which such borrowed words occur, we find them spelt with such Latin characters as most nearly represented those Greek sounds which had either been developed by the Greek after the parting of the two peoples, or which had been lost by the Latins out of the original common stock. Thus the

1 See this more developed in Max Müller, Chips from a German Workshop, 11. 41, &c.

Greek aspirates-peculiar developments of the Greek—appeared in Latin as unaspirated mutes; e.g. Aciles (Axievs), Burrus (Iuppós); this last word and Bruges (øpúyes) shew that the full Latin u was taken as the nearest Latin exponent of the Greek upsilon (a modified u), and in Plautus ss appears as the best representative of the strong Greek (which differed from the old weak Italian 2) in badisso, tarpessita, &c. In the Augustan age, on the contrary, Greek characters are borrowed as well as the sounds, the Y in lyra, the Z in zona, &c.: while a combination of letters represented the complex sound of the Greek aspirates-chorda, philosophia, &c. Now it is obvious that these words were not derived from the Greek; they were not formed from a Greek root by adding to it a Latin suffix; they were derived in Greece from Greek roots by Greek suffixes and transplanted when fully grown into Latin. They are as foreign to the Latin language and its development, as the men and things they represent were foreign to Rome. But from these borrowed Greek words it was inferred by a false analogy that numbers of genuine Latin words were borrowed from the Greek. Because lyra was the Greek λúpa, it was supposed that lacruma was the Greek Sáκpuua; and consequently it was written lacryma, or even by some curious fatality lachryma. But in truth the words have nothing in common except their base dakr (whence the A.-S. teagor, our "tear"); each was formed from that base, but by its own suffix in its own land: the emotional Italian was not likely to lack a word for a tear, till he had borrowed it from the Greek! In other cases-e. g. the Latin silua, no doubt the noun SVLVĀ existed in Graeco-Italian days, and was then modified by the two peoples in different ways according to their different phonetic laws. But it is an entire mistake to write silua with a y, that is, to imply that the word was borrowed from the Greek vλŋ. Indeed the Latin has kept the old form more nearly than the Greek; it has changed u to i, and ā to a, both regular Latin changes, and both weakenings; but An exhibits no less than four weakenings; s has passed into the rough breathing; u has (as always in Greek) been weakened to upsilon; v has passed out altogether, and a has been thinned

to n.

The rule then to follow in writing Latin is very simple:

CH. I.

CH. I.

we must use the letters Y, Z, and the compounds CH, TH, PH, in words borrowed from the Greek and in no others. Such words are not difficult to recognise. They are mostly words relating to the arts and sciences which the Romans borrowed from the Greeks. All other words are, in ninety-nine cases out of a hundred, genuine Latin, and should be written in the Latin character. The only exception which should be allowed is in cases where we have express testimony that Roman writers in the last century of the Republic employed Greek characters or the equivalent compounds in Latin-in words which are beyond doubt genuinely Latin, but which by a mistaken analogy were then supposed to be derived from the Greek. In such cases we may write, e.g. pulcher'-though we believe it to be etymologically wrong-on the same principle that we write, e. g. caussa, and querella; because they represent the spelling which, rightly or wrongly, was in use in Cicero's day among educated men; not because we believe it to be the Greek Toλúxpoos. Lucretius truly says, "Utilitas expressit nomina rerum;" and it is equally true that use must always be the standard of orthography, and must override etymological considerations. Only let our standard in Latin be the usage of Cicero's time, not of the period of the Renaissance.

1 See Cic. Orat. c. 48. § 160.

2 It is possible however the h in this and similar words, Cethegus, Carthago, &c., may have nothing to do with the Greek, but may be a vulgar use of the aspirate which was passing into the literary language in Cicero's day. His phrase "usum loquendi populo concessi" rather supports this view. See additional evidence in the section on "Aspiration" in the last chapter of this book, and Roscher, de Aspiratione apud Romanos in Curtius, Studien, II. 2. 143, &c.

CHAPTER II.

RELATIONSHIP OF THE INDO-EUROPEAN PEOPLES.

I TRANSLATE from Schleicher1 the very brief and clear account of the main divisions and subdivisions of the variously called Indo-European, Indo-Germanic, or Aryan language: to which can be traced nearly all the languages of Europe, and two at least of those of Asia, the Sanskrit and the Zend.

The name of Indo-Germanic has been given to a certain class of the languages of the Asiatic-European portion of the earth, which are so accordant with each other, and which differ so much from all other languages in their nature, that they clearly show themselves to have sprung from a common original language. Within this IndoGermanic family of languages, some, which are more closely geographically connected, shew themselves certainly to be the most nearly allied, so that the Indo-Germanic family divides into three groups or divisions. These are

CH II.

ClassificaIndotion of the

European peoples.

or Asiatic.

"I. The Aryan3 division, consisting of the Indian and (i.) Aryan Iranian, or more correctly Eranian, families of languages, which are very closely related to each other.

"The oldest representative and original language of the Indian family, and the oldest known language of the

1 Compendium der Vergleichenden Grammatik, pp. 6—8.

2 See note at the end of the chapter.

3 It will be seen that the term Aryan is here applied only to the two Asiatic peoples who can be certainly proved to have called themselves by that name.

Prof. Schleicher of course does not mean that the Sanskrit existed before the Greek and Latin, but that it is known to us in an older stage than any other. The error which arises from regarding every Sanskrit form as older than the corresponding forms in Greek and Latin will be noticed at p. 31.