FERRAR'S COMPARATIVE GRAMMAR.* Tins would be acceptable as the first

systematic work that has appeared in English on the subject, even were the handling of it less careful and thorough than it actually is. It is now some time since grammar has been raised by the labours of Bopp, Grimm, and others to its proper scientific rank. Formerly it meant little more than an artificial and cumbrous arrangement of barren facts ; it has now become a storehouse of fruitful knowledge, whence we may gain a daily enlarged insight into the early history of the races that now people the civilized world, the course of their migrations, the development of their thought, and the steps of their advance in the arts of social life. And as grammar is thus exalted to its fitting eminence and usefulness as a living force in the hands of comparative philologists, we hope that the empty shell of it may also in due time abandon the bad eminence it has too long enjoyed as an engine of scholastic torture. We are not unaware that there have appeared in these latter days, and are still in course of appearing, Public School Primers and other monsters of new and unspeakable hideousness. Rather we look on these with a certain satisfaction, as being, in truth, tokens of a death-struggle. But let us leave the idols and their sacrifices, wish- ing once for all a speedy deliverance to their victims the schoolboys, and turn to a more pleasant sight, for already a temple is rising in fair proportions in our land to do honour to the true divinity. Mr. Ferrar's contribution is substantial and timely, and deserves a welcome from all who have the work at heart.

The reader is supposed to know the cardinal facts which philo- logists have established by comparing one with another the family of languages called by some Aryan, by others Judo- Germanic, by others Indo-European. These facts have now become tolerably familiar to English readers by the works Of Max Miiller and others. Among the more recent expositions, we may mention Mr. F. W. Farrar's Families of Speech as well fitted to undeceive any one who has formed a prejudice against comparative philology, as being a fruitless or uninteresting study. It is presumed, then, that the Sanskrit, Zend, Greek, Latin, Teu- tonic, Keltic, and Slavonic types of language, and the living tongues now representing them, have all sprung from a single language, once spoken by an undivided ancestral people. Thus, the known Judo-European tongues stand to this original one in the same relation as the Romance languages do to the Latin. So that "as we could approximate to the roots and grammatical forms of the Latin language, even if we had no monuments of it, from a comparison of the roots and grammatical forms at present existing in the Romance languages, so analagously we may approximate to the roots and forms of the language of the Ludo- Europeans from a comparison of the languages spoken by their descendants. For example, if we take the case of the numerals, we see at once that the names for the first ten numbers in any Romance language are not derived from those in any other, but from the Latin." Similarly, the Sanskrit, Greek, Latin, &c., names for the same number "are all independent of each other, but they all presuppose:an Judo-European form."

And this process of approximation to the original forms has, in fact, been carried out so far that the vocabulary of the mother tongue as it existed before the divergence of the different tribes has been actually in great measure reconstructed ; and this makes

* A Comparatint Grammar of Sanarit, Greek, and Latin. By William Hugh Ferrer, M.A, Fellow and Tutor of Trinity College, Dublin. Volume I. London: Long- mane. 1869. it possible to draw conclusions of surprising extent and certainty concerning the habits and civilization of our parent race. The importance of the family in their eyes is shown by the very complete development of words expressing relationship ; the most important of these are the same that we use to this day. We know that they had made some advance in agriculture, and had houses and walled towns ; moreover, that they practised the art of weaving, knew something of the working of metals, built boats, and were accustomed to navigating the sea or large lakes. And it is possible that evidence of the same kind may in no long time settle definitely the question, as yet not at rest, as to the site of the common home whence the several Judo-European peoples were dispersed.

The growth of comparative philology is so recent that there must still be much room for conjecture ; the conjectures, however, are no longer chaotic, but guided by established certainties. The task -of the philologist is not to be accomplished by any mere guess work founded on random comparison of instances. We have learnt to resist the temptations which so sorely misled the older grammarians. We do not now attempt to derive one extant language from another of equal or less antiquity, instead of look- ing beyond both to their true common origin, though the figment of Latin coming from "Belie Greek may perhaps still linger in some school-books. Nor do we run astray into really unrelated languages in search of isolated resemblances. But we must not only know how to confine ourselves to the bounds within which there is really something to look for: we must have a definite rule to govern our search. To ensure a scientific procedure we must reduce to some certain uniformity the changes which the original sounds undergo in passing into the several -correlated languages. This we are enabled to do by "the remark- able law of the dislocation of the consonants discovered by Grimm." Mr. Ferrar's statement of the law, and the examples he gives in illustration of it, are exceedingly full and clear. It is impossible here to follow him in detail, and any attempt at con- densation would be impertinent for readers who are at all versed in these matters, and unintelligible for those who are not. But we may use the relations of modern German and modern English, in which the law is sufficiently apparent for this purpose, to exhibit the general -effect of Grimm's discovery in a rough and familiar way. We may collect from German and English an immense number of correspond- ing words, such as these (to take at random a few specimens) :— zehn and ten, Zahl and tale, Zehe and toe, Zahre and tear; again, we compare That with deed, Tag with day, Noth with need; and again, 'der with the, Ding with thing, diinn with thin. And in many words, as Zeit and tide, Tod and death, a double change of consonants may be seen. The interchanges of another class of letters appear in Pfeife and pipe, Pfahl and pale,hapfen and hop, Laub and leaf, Stab -and staff. The inspection of the corresponding words thus brought 'together discloses at once that the variations are not capricious. There is a uniformity in each group of instances, and there is lurther a certain order and definite sequence in the transition, so -to speak, from one set of changes to another. Now Grimm's law is the scientific expression of the uniform mode of change of which the above are a few particular cases, and which prevails through the whole extent of the Indo-European tongues.

The law being once established on the evidence of obvious -analogies, we are enabled by its application to trace real identities under an external disguise that would have defied any less rigorous -scrutiny. It is thus revealed, for instance, that not only Zahre and tear, but also the Greek hcizpu and Latin lacryma (daeruma in its older form), have descended from one original 'word; that our feather is formed on the same root as . the Latin penna ; and that words so unlike as pugnus and fist are in fact closely akin. Besides this kind of positive help, the law also gives us a safeguard

We have departed from the order of Mr. Ferrar's book, his first chapter being occupied with the "general alphabet" of all articu- late sounds produced by the human voice. He does not enter on

the question whether it is possible or worth while to construct a universal phonetic alphabet for actual use. If such a thing were

to be had, it would be curious and interesting, and some very novel results might be expected. A reporter, for example, skilled in this ideal alphabet should be able to take down accurately a speech delivered in an unknown language ; and another person, equally skilled in the character, but equally ignorant of the language, to read out the speech from this report with perfect correctness.

Such speculations, however, are not within the province of a comparative grammar ; indeed, the general analysis of speech seems to be here carried somewhat farther than is needed for the special purpose of the book. This introductory part is also less pleasant reading than the rest, insomuch as it shows less of the independent judgment which Mr. Ferrar elsewhere pre-

serves. Here he distinguishes between sounds nearly alike (e. g., English v and German w), not because he hears a difference, or

because any one who listens to correct speakers of the two lan- guages may hear it, but because the German writers he follows say there is a difference.

The latter part of the book is devoted to the discussion of the Sanskrit, Greek, and Latin alphabets, the changes of letters in the respective languages, and the exhibition of their systems of declension in comparison with one another and with the primitive Judo-European forms. The present volume, which is to be followed by a second in due course of time, goes only as far as the pronouns. Out of such a mass of condensed and minute exposi- tion it is almost impossible to select particular points for comment.

We can only sly that Mr. Ferra.r is generally a careful and trust- worthy guide. But we stumble sorely at the table of Sanskrit vowel-modifications on page 43. A statement there made about the vowel a is wholly in defiance of all the tradition of the elders. The slip, for such we presume it must be, though it affects a whole paragraph, is not less extraordinary than if a writer on French grammar were to assert that a is in that language capable of taking the acute accent.

The discussion which follows in its proper place as to the pro- nunciation of Latin has at this time a special interest, as a move- ment for the abolition of our present unutterably barbarous practice has now arisen in this country, and is gathering strength. It may be difficult, however, to fix and conform to an ideal standard of the manner in which Latin was actually spoken at some classical period ; certainly no other nation has attempted it. The pronunciation of Continental scholars is more tolerable than ours, only because with them the vowels of the Roman alphabet have for the most part retained their original force in the vernacular.

We may finally notice a rather subtle difference, incidentally mentioned in a special appendix to this volume, between our speech and that of the Germans (and we believe most other Continental nations) which is remarkable for the roundabout way in which it is established. The Sanskrit and other Indian alphabets have two

sets of t's and d's, which belong to different classes of letters, and

are completely distinguished from one another both in writing and in speaking. Now it is found that a native scholar taking down European words from the dictation of a German always uses

the t and d of one class ; while if the words be read out by an Englishman, he employs the corresponding letters of the other

class. Thus a distinction, which probably very few Germans or Englishmen would naturally and consciously discern by ear, and which no European alphabet recognizes, is brought to light by the testimony of Indian transcribers.