Introduction to Greek Epigraphy. Part I. Archaic Inscriptions and the Greek Alphabet, by E. S. ROBERTS. Cambridge. 188.

MR. ROBERTS's first volume is one of many welcome signs of the growing interest which our Universities take in the development of independent investigation as one side of a high classical education. Epigraphy indeed is not a thing for babes; the necessity of a rigorous linguistic training before epigraphy is begun cannot be too much insisted on, and the present reviewer has often insisted on it. But on the other hand a dangerous fallacy is sometimes current, that epigraphy is a study for archaeologists, and that scholars have no relation to it and no interest in it. In truth it is impossible to divorce a literature from its surroundings and to study it without any reference to the circumstances in which it was produced; every commentary on every page attests the truth that even the meaning and still more the spirit of any great work of literature demands a reference to the life and manners amid which it is set. Existing commentaries, however, while they cannot wholly ignore this side of their task, are usually weak or inaccurate where they touch on it, and if we would give to the study of antiquity the living interest of search and discovery, it is in this direction that most remains to be done. In the case of some authors epigraphy throws only a side-light on incidental details: in the case of others it throws a strong light on their whole subject, and he who hereafter treats such authors without constant reference to epigraphic facts will be trifling with his subjects and his audience.

This volume embraces the inscriptions which bear on the history of the Greek alphabet in its many varieties. The subject is a large one, and it is particularly difficult to expound to students on account of the paucity of definite results. I may say at

NO. XVII. VOL. II.

But this

once that the materials are exhibited clearly and fully, that the specimens of inscriptions are given accurately and well, and that in the important branches of the subject the typical inscriptions are almost always to be found in this volume, though of course where selection has to be made opinions will sometimes differ, and each one will occasionally miss something that he would himself have preferred. self have preferred. Every page of the work before us bears witness to the careful labour that Mr. Roberts has spent on it; and we are all the more desirous to emphasize this fact when we think of the continuous and heavy official duties' (mentioned in the preface) which necessarily press so much upon him. Those who prefer a selection of materials to a complete apparatus could hardly hope for a better selection. leads me, after what is I hope a hearty and adequate acknowledgment of the great merits of the book, to a statement of the disagreement in which I find myself placed with its general plan. Mr. Roberts aims at setting forth the results and general principles which the science of epigraphy has reached. This point of view is one to which a teacher in an English University is almost necessarily inclined. The necessity of preparing generations of pupils for severe and testing examinations produces a habit of mind and a style of lecturing which enunciates the doubly distilled and triply refined essence of knowledge-food for the mind of the most highly concentrated nourishment in the most easily digested form, containing precisely what is useful without any refuse of unnecessary detail. But in the concrete world of actuality truth is not thus presented to the inquirer. Brilliant as such lectures often are, it is perhaps open to question whether a simpler and more natural style would not be healthier both for pupil and lecturer. The perfection of such a lecture is that nothing should be given which is not

to be remembered, and that everything in it should deserve to be retained by the hearer in ready memory, whereas a simpler style of lecturing would set forth a great deal that is not intended to be held in memory ready for instant service. Mr. Roberts carries some hing of this method into epigraphy; he would place before students its results in definite clear principles, and unluckily he has in this volume selected the branch of epigraphy in which results of the kind he seeks are as yet conspicuously wanting. The known data are as yet insufficient to form the foundation for a scientific history of the Greek alphabet; we must for some time yet be content with an empirical account, and continue to observe and to catalogue, until a wider array of details shall justify scientific generalizations. We watch the progress of discovery with the intense interest imparted by the belief that any day some new observation may give us the clue to guide us through the labyrinthine varieties of the Greek alphabet, and make clear their relations to each other. But Mr. Roberts's business is to set forth principles, and principles he must have. Not finding anything but empirical facts of partial agreement, he sets forth these empirical facts as general principles, and reasons from them.

To begin with, some classification of the Greek alphabets must be adopted. Mr. Roberts follows Kirchhoff, and in so doing he is certainly right. Kirchhoff's classification is the most widely known, and the most commonly followed; a writer who at present sets forth epigraphic results must work on the lines traced by him. But Mr. Roberts goes further than this; he believes that the classification which he has adopted is the natural one, corresponding to the historical development of the early Greek alphabet. In section 3 he says, 'it will be sought in the course of the investigation to establish the following facts:' then follows a dogmatic summary of Kirchhoff's classification. He goes further still: through a great part of the book he speaks as if in this preliminary statement he had set forth principles, and he reasons from them as if they were general truths. Kirchhoff himself does not make this confusion; he speaks of his classification as empirical observation, insufficient as yet to justify any general statements with regard to the early history, origin, and diffusion of the alphabet in Greece. He appears to incline towards the theory that the Greek alphabet either originated independently, or at least was

developed independently in two different centres. This may hereafter be found true, but at present it is mere hypothesis; and Kirchhoff claims to have stated his study of the alphabet without committing himself or his readers to any hypothesis whatever-in short, to make a mere arrangement according to superficial appearance, without assuming that this arrangement corresponds to the facts of historical development. It is indeed true that he is not in fact so impartial as he claims to have been. He strains and even outrages appearances in order to force all alphabets into one or other of his two classes. He has taken a principle of division which suits the great majority of known alphabets, but which, in my humble opinion, fails in the remainder; the outlying alphabets belong to neither of his classes, and can only be placed in one or other by violence. But while Kirchhoff himself at least attempts to be impartial and does not claim to state principles, Mr. Roberts goes far beyond him in advocating his views. Independent criticism, clear statement of merits and acknowledgment of inadequacies in the views which he states, appear to me to be absent from Mr. Roberts's book, and this renders him often a dangerous guide to a young student.

The disadvantage of adopting Kirchhoff's classification of the Greek alphabets becomes apparent when Mr. Roberts begins to discuss the half-bellenic alphabets of Asia Minor. Having bound himself to fit all alphabets into one or other of the two classes, eastern or western, Mr. Roberts puts only the Procrustean question about each of the Anatolian alphabets, 'whether it was an alphabet of the eastern or of the western type which furnished the pattern,' and the crucial symbols in Asia Minor must be interpreted in accordance with one type or the other: no third interpretation is admissible. How utterly inadequate this method of treating the alphabets of unintelligible languages is, becomes plain in the case of Lycian to any one who looks over Deecke's series of articles in Bezzenberger's Beitr., 1886-7. As to the Phrygian alphabet I hope soon to have something more to say. As to Pamphylian, I may here add a few words, though I had thought that the question was finally set at rest, until Mr. Roberts's book showed me that in some Cambridge circles Bergk's ideas are still accepted as gospel, and that no English proof to the contrary can be even weighed against them. The first work which I ever ventured to publish was a proof that the Pamphylian

My interpretation gives actual or possible Greek words in a Greek inscription, Bergk's interpretation gives impossibilities.

But why does Mr. Roberts omit to mention, merely as a matter of bibliography, that the view which he rejects was proposed separately and independently by Deecke, and has been unhesitatingly adopted by Roehl, in the Inscriptt. Antiquiss. praet. Att. in Attica repertas, and by Bezzenberger in the Sammlung griech. Dialekt-Inschriften? Indeed, so far as I know, not a single German philologist, who has treated the Pamphylian dialect, in recent years, has even expressed doubt on the subject. Mr. Roberts himself seems to be a little inconsistent. On p. 177 he adopts part of my theory, that at Halicarnassus and Mesembria was the palatal sibilant, and says that 'the Pamphylian alphabet may be cited as additional evidence of a special sign for a sibilant in the S.W. district of Asia Minor (§ 123)': but turning to pp. 272, 317, we find that the sibilant he means is not 4 or Y, which on my theory are merely modifications of T, but W, which according to Bergk is a badly written sigma.

Apart from this want of independent tone and suggestive treatment, I can only praise the book for its general knowledge and accuracy. A few of the imperfections in detail, which I have noted, may be here mentioned. We find on p. 59, no. 24, a copy of a well-known fragment of a fifth

century Olympian inscription in Ionic dialect recording a long series of victories in Pythian, Isthmian, and Nemean games. Mr. Roberts follows the restoration of Treu and Roehl, who suppose that Theagenes of Thasos was the athlete whose victories are recorded; the difficulties of this interpretation, which are observed by Treu himself, are not mentioned by Mr. Roberts. Everything that we know about Thasos shows that the Thasians used their own peculiar alphabet until the general reception of the Ionic alphabet. In this Olympian inscription we find an early form of the Ionic alphabet, older than the form which that alphabet had assumed before its general reception over Greece. Other difficulties also suggest themselves to the scrutinizing critic. Just a month or two before Mr. Roberts's book appeared,' M. Foucart published in the Bulletin de Correspondance Hellénique, 1887, p. 289, a discussion of the same inscription, in which he stated conclusively the reasons against Treu's restoration, and showed that the victor in question is probably Dorieus, son of Diagoras, of Rhodes. The uncontested (ákovITEL) Pythian victory, mentioned by Pausanias as gained by Dorieus, and recorded on the stone, is strong evidence in favour of this restoration. What we might have hoped for from Mr. Roberts is, not indeed the restoration, for such lucky thoughts are not common and 1 Too late to be of use for Mr. Roberts's work yet in his preface he mentions still more recent papers without noticing this one.

must not be demanded, but an independent weighing and statement of the difficulties in the restoration proposed by others; this is of the first importance in a writer who would place epigraphic results fairly and dispassionately before students. This free and masterly point of view never appears to me to be occupied by Mr. Roberts.

On p. 192, no. 164, Mr. Roberts says of a small inscribed vase found in Rhodes, as both the alphabet and the dialect of this inscription are Ionic, it is probable that the vessel came to Rhodes in the way of commerce.' But there is every probability that the Ionic alphabet and dialect were used in the fifth century in Rhodes as in the Doric colonies on the mainland (see Bulletin de Correspondance Hellénique, 1887, p. 292), and the inference is therefore unjustifiable.

On p. 234 and p. 346 no reference is made to E. Curtius's interesting discussion of the historical aspect of the famous Naupactian inscription, Hermes X., p. 237.

On p. 126-7, under 'Sicyon,' no reference is made to the small class of inscribed Sicyonian vases, which (once assigned to a Corinthian fabric) are now referred to Sicyonian potteries on account of the symbol for epsilon, X.

On p. 5 Mr. Roberts lays down that the most ancient Greek inscriptions were written right to left; 'then followed a period of transition during which the Bovoтpoondóv βουστροφηδόν arrangement prevailed,' and this new method, in the sixth century, 'prevailed so far

as to be characteristic.' In the beginning of the fifth century 'as a natural consequence of' the former change, there ensued 'the complete adoption of the left to right direction.' This account is very superficial; it is almost ludicrous to say that the Greeks, after learning from other races to write towards the left, began for some unstated reason to write towards the right, and resolved to write every second line in the new way, keeping the alternate lines in the old fashion; then, after doing this for a century, until they were well accustomed to it and it had become 'characteristic,' they resolved to write every line towards the right. The truth seems to be that the Greeks learned from the Phoenicians to write towards the left, and that in Asia Minor they became acquainted with the boustrophedon method of writing, which was there used regularly from a remote period in the Syro-Cappadocian or Hittite' hieroglyphics. Boustrophedon arrangement is found among the earliest Greek inscriptions, and is no more characteristic of the sixth than of the seventh century, taking the proportion between the entire number of inscriptions known and those which are written Boustrophedon.

An

Boeotian school, by the Cimmerian invasion of Asia Minor, B.C. 690 (which Sayce, however, puts some twenty years later on the evidence of the Assyrian records). earlier date is precluded, he thinks, by the tradition (for which Philochorus is responsible) that Stesichorus was the son of Hesiod and born in his old age; and likewise by the mention, whether direct or indirect, of Aetna, Theogon. 860 (where with Schömann he reads 'Αΐτνης (ας) for ἀϊδνης), since the mountain can hardly have been known till after the foundation of Naxos. The poet had quitted Ascra after the unjust award and the loss of his patrimony, and settled near Naupactus: so Fick infers from Op. 635-9, Tyde meaning the place where his father had landed, not the place where he afterwards

'settled' (váσoaro). Here, therefore, not in Boeotia, he composed the lays which were afterwards enlarged in the Hesiodic school, viz. the 'Reproof' addressed to Perses, the 'Works,' and the 'Song of the Five Ages.' These poems were composed in his native Aeolian dialect, which can be restored throughout all the genuine portions. The Theogony was composed, under Delphian auspices, in a dialect closely related to the Thessalian and other known dialects of Northern Greece, its special character being due to the contact of the amphictyonic tribes at Delphi. The poem originally consisted of three books. The first treated of the primeval powers, probably on the basis of an earlier Cyprian Theogony; hence it is that Aphrodite is established as Urania by the side of Uranus, but Chaos (vacant air) is substituted for Oceanus (the air-stream) as the original parent. The second treated of Cronos and his brother Titans; the long description of the birth of Zeus was introduced in honour of Crete, whence the Delphian priesthood was founded. The subject of the third book was Zeus, his victory over the Giants, his marriages and offspring. This Theogony was composed before the Works,' as the introductory description of Eris in the latter refers by way of correction (v. 11, ouk aρa K.T.λ.) to the former (Theogon. 225). Two proems were subsequently added (in the same dialect), the first intended for the 'Epwrídela at Thespiae, and accordingly glorifying Helicon and the local river Permessus (with the subsequent addition of Hippocrene and Holmeios), the other apparently for the 'Appiápaia at Oropus; 'Eλevonp, v. 54, denoting the city of Eleutheris in that neighbourhood.

·

These poems of Hesiod, like the original Homeric poems, were transposed into the quasi-Ionic or so called 'Epic' dialect, with large additions (including the 'Days') and a mass of 'rubbish' subsequently thrown in. The Ionian editor was, according to Fick's conjecture, Cercops of Miletus.

Diogenes Laert. ii. 46 says of this poet, épiλoveĺkel Hoιód (avri, which Fick, arguing from the context, explains as meaning that he professed to embody Hesiod, that is, engrafted his own productions on him. The two names are also coupled together in Athenaeus (xi. p. 503 D) in reference to a Hesiodic poem entitled Aiyiuos. Again, Αἰγίμιος. the statement that the Ναυπάκτια ἔπη (which must have originated in Locris, and therefore in the Hesiodic school) were ascribed by the majority to a Milesian (Pausan. x. 38, 11), notwithstanding that the original

composer was (according to the logographer Charon) Carcinus of Naupactus, may be explained on the supposition that Cercops had revised these as well as the other poems of the Locrian (Hesiodic) school. The fact that he was the author of an Orphic Theogony (known as the iepoì λoyo) tallies with several remarkable indications of mysticism in the Ionic additions, e.g. the praise of the 'sole-begotten' (uovvoyevs) Hecate, the curious ceremonial rules, which the Pythagoreans borrowed from the Egyptian priesthood (e.g. Op. 731-2, and 742-3), and the word-puzzles (Tévtošos, ávóσteos, &c.). That

the list of lucky and unlucky days was inspired by the same Egyptian influence is implied in the statement of Herodotus (ii. 82), even if he does not directly refer to 'Hesiod.' There are direct indications of Ionian origin in the strange reference to the month Lenaeon (Op. 504), and to the Biẞuvos olvos, not the wine of Biblis in Thrace, nor 'straw '-wine (as Hehn suggests, Kulturpflanzen, p. 492), but the wine of Biblos, a district of Naxos which still preserves its old name (comp. Mr. Bent's 'Cyclades,' p. 369). The mention of Amphidamas of Chalcis (Op. 655) belongs to another Ionic section and therefore, of course, gives no clue to the date of the real Hesiod.

The

It is recognised by all modern commentators that the Hesiodic text has been extensively worked over and corrupted. Fick's division, however, is mainly based on the number and metrical value of the Ionic forms, the genuine portions being those which he has found it possible to retranslate, with occasional alterations and excisions, for which he accounts in all cases. Ionic residuum includes, besides the 'Days' and many shorter passages, the list of rivers (Theogon. 338-45), which can hardly be older than Hecataeus, the hymn to Hecate (ib. 411-52), and the description of Hades (ib. 720-815), which implies the later idea of a division of Tuai after the dethronement of Cronos. Op. 564-618 is a short farmer's calendar, separately composed; 504-63 is manifestly Ionic in its elaborate description, and replete with Ionic forms.

The Ionisms consist of the numerous words in which the digamma is neglected, the Ionic use of ἐς, ἔσω, ἄν for κε, ἕως for avws, čoav, čŋv (which, in the Theogony, cannot ex hypothesi be altered to Aeolic čev), Ionic contractions (gen. έων for άων, accus.

for ea; contracted dative of the eo stem, e.g. ὄρει, σελᾷ: ̓Αχελῷον, Σιμοῦντα, γῆς for yaías, dolos, &c.). With these are found non-Epic or false forms, e.g. åpnpûiav for