properly Lesbian, and is unknown to pure Doric ; here f is as a rule retained, but in the Laconian dialect when initial it becomes 3 (8a>yu.Js=a>yu<k), when medial 58 (/j.vari5$ta = /j.v6ifa, iror65Sei = 7rpo<r6fi). A double a is retained where this is the more ancient form, changed in ordinary Greek into a ; thus the Heraclean tables give o<Tffos, neffffos, effcrovrai, &c. The crff often found in Dorian inscrip tions (and sometimes La the earlier Attic also), where there is no historical explanation of its presence, seems to be an attempt to represent the sound of the earlier sibilant san, which was retained by the side of sigma. For the earlier guttural Tcoppa, the distinctive sign p is found in old inscriptions, almost, but not quite exclusively before o, e.g. , f>opivd66fi>, oppos. We may notice finally a free use of assimilation especially in Laconian (irov/j.fAa. irv y/j.-l), Kdppuv=Kpttffo (ey,i.C.,Kaprioiv, a.KK&p = a.<rtt6s) , and on the other hand the retention of vs by the Argives and Cretans (r6vs, yucWa, eVs, n6fvs, lipvvs, &c.). The characteristic Dorian inflexions are almost entirely such as are due to these phonetic laws, or to the tendency to metaphrastic or heteroclite formations, already noticed,
Ionic.—The Ionic dialect is commonly divided into three stages, the Old Ionic or Epic dialect, the New Ionic, represented most com pletely by Herodotus and Hippocrates, and the Attic. This division is not satisfactory ; for, in the first place, the Epic is a mixed dialect or more properly a style, and cannot be taken as a faithful represen tative of a spoken language ; and, in the second place, Attic is not a later stage of the New Ionic, but in many respects remains faithful to forms in which even the Old Ionic has departed from the earlier usage. The three sub-dialects, however, agree on the whole much more closely than any one of them does with either YEolic or Doric, and they may therefore be grouped together. We know from an express statement of Herodotus (i. 142) that there were many sub ordinate varieties of the ordinary Ionic ; he mentions four within a comparatively narrow extent ; but neitherjjthe extant inscriptions nor the statements of grammarians enable us to distinguish these with any precision. It is probable that the differences lay rather in slight shades of pronunciation than in any extensive variations, and that, on the whole, the varieties closely resembled each other. As the general character of Doric is due, at least in a measure, to the hardy mountain life of the Dorians, so the Ionic type was determined by the easier and more effeminate life of the lonians. All harshness is carefully avoided ; the spirants, especially the /, were dropt here earlier and more completely than in any other dialect ; the a is more extensively changed into e and o ; aspiration is frequently lost or transposed so as to be easier to pronounce ; r, especially before i, regularly passes into <r ; gutturals are replaced by dentals or labials. The vowel-system is especially rich and free ; sometimes an easy flow is given by the avoidance of contraction ; sometimes again a full colouring is produced by the variety of the diphthongs. The varied literary activity of the lonians in different directions gave a manifold development to their language, which makes it especially well adapted to poetry, and adds not a little of poetical charm even to their prose.
Epic Dialect.—The language of the Homeric poems is doubtless based upon the popular spoken dialect of the district in the midst of which they grew up. But as every scholar would now admit that they were constructed out of a large mass of previously existing material, however widely opinions may differ as to the person or school to which they owe their present form, and as much of this material must have dated from a great antiquity, it need not sur prise us to find in the midst of a dialect, which is of a much more recent type than ^Eolic or Doric, traces of archaisms, earlier in some respects than anything to be found elsewhere. It is one of the greatest services which comparative philology has done for the in terpretation of these poems, that it has enabled us to recognize as relics of an older language much which had been previously set down as poetic licence, or held to be inexplicable. One of the most interesting of these relics is the effect produced by the earlier existence of a spirant, no longer written, upon the quantity of a preceding syllable. As late as the time of I. Bekker all such cases were unhesitatingly ascribed to the digamma; and this accounts for many instances ; but in others the cognate languages point to cr or j : e.g., we find not only (pia fei/j.ara SiKTai, ovrca 877 /oi/co^Se, and hundreds of similar cases (La Eocho gives 84 Homeric words with the digamma), some of very common occurrence, but also Oebs (j)&s, /oi /caSe (j)i(j)en4vur t en yap ((r)tx ov eA/cea vypd, vUi ff<f (cr)e7rd ( urjj , els aAa (ff)a.ro, and many other instances. On the other hand, the occasional neglect of the digamma, even in words for which it is most certainly estab lished, points, not necessarily, as some have argued, to a later origin of those lines in which this occurs, but to a fluctuating usage, akin to though much more extensive than our own poetic use of forms like loveth and loves, formed and form d, my and mine. In the form in which the poems now appear, it is often of much importance to remember that they must have been tran scribed at a comparatively late date from the earlier into the later Ionic alphabet (see ALPHABET, vol. i. p. 610), and that doubtless many words were inaccurately represented. The limits of this sketch do not admit of a statement of the characteristic epic forms. They will be found given with very full references in the introduc tion to LP Eoche s school edition of the Iliad (Berlin, 1870), and with admirable clearness and scientific exactness in the sketch of Homeric grammar prefixed by Mr D. B. Monro to his edition of the First Book of the Iliad (Oxford, 1879).
The New Ionic dialect is found first in the writings of the iambic New elegiac poets, Archilochus, Callinus, and Mimnermus (where the digamma has already entirely disappeared), and is known more completely from Herodotus and Hippocrates. We are told that the language of the former was varied (TTOIKI ATJ) as compared with the pure (/caflap?/) Ionic of preceding logographers ; this seems to refer to the occasional introduction of epic forms and expressions, which give a delightful poetic tinge to his language (cf. Quintil., ix. 4, 18, Turn ipsa SiaAe/cros habct cam iucunditatem,ut latentcs ctiam numcros complexa videatur) and not to any dialectic variations. Besides the general tendencies of Ionic mentioned above we may notice the retention of the earlier /c for TT in interrogative and relative words (KO!OS, (5/coVos, &c. ), the interchange of ej and ot; with the simple vowels (ftpofi.su, KetfSs, ^eivos, but /j.e^uf, 5eo>, ra^ea ; and /J.DVVOS, ovvofj.a, rb ovpos, vovffos), the contraction of or; into w (Pupat, (fiuOtf, fvv&ffus}, the use of TJI for ei (/Sao-iArj hj, fiavT^ iov), the Ionic crasis in wvrip, wAAoi, &c., the entire absence of the appended v, the gen. plur. in -tuv for Homer s -a.<av, Att. -lav, and the use of -a-rat, -aro for -VTO.I and -vro wherever these are added directly to the tense- stems (cf, fffKfvdoaTat, tin marai, ySe/SAeorai, riGearai, ayotaro, &c. ). The dialect of Herodotus has been most fully discussed by Bredow, Quccstionum criticarumde dialecto Herod, libri duo, 1846); there are some excellent remarks upon it by Mr Woods in an introduction to his edition of Book i. pp. 40-45. The text of Hippocrates is in too unsettled a state, and the genuineness of many of the treatises ascribed to him too doubtful, to make it possible for us to build much upon his authority. From inscriptions but little can be gained. See Erman in Cudius s Studien, vol. v. pp. 250 ff.
The Attic dialect may be regarded as on the whole a slightly modified representative of the Ionic spoken before the foundation of the Ionic colonies. It is not so much a daughter of Ionic as its mother, as Bergk justly calls it. In Ionic the tendency to soften the lan guage which had already commenced before the separation went on its way unrestrained in the luxurious life of the Asiatic cities 1 . In Attica, possibly owing to the free admission of non-Ionic citizens by Solon and Cleisthenes, this tendency was checked, and there arc even some signs of a reaction in the direction of the earlier and more vigorous speech. There is a celebrated inscription found at Sigeum in the Troad (C. I. G., 8), the antiquity of which, though attacked by Boeckh, has been established by Kirchhoff ; this is in two parts, the upper in Ionic dialect, the lower (which is probably a little later, but also belonging to the time of Pisistratus) in Attic, and we can already see the reaction at work. The Attic dia lect thus adapted itself admirably to the character of the Athenian people, which knew better than any other Hellenic community how to unite energy and dignity with grace and refinement, to preserve the ffefjLv&rt]s of the Dorian without sacrificing the x^P LS f the Ionian. The Attic of the inscriptions may be most conveniently divided according as these are written in the old alphabet of sixteen letters or in the so-called Ionic alphabet of twenty-four. The latter was introduced for public documents inthearchonshipof Euclides(403 B.C.); the inscriptions written before that date have been collected and edited by Kirchhoff in the first volume of the new Berlin Corjnw Inscriptionum Grcccarum, and their linguistic peculiarities well com mented upon by Cauer in Curtius s Studien, vol. viii. pp. 223-301 and 399-442. The Attic of literature is divided into the Old and the New, the point of division being earlier than the archonship of Euclides, and coinciding more nearly with the beginning of the Peloponnesian War (431 B.C.). The division is, however, not strictly a chronological one, for, while Thucydides and the tragedians adhere to the older forms, contemporary comic writers adopt the later ones; in Plato both are found side by side ; but in the orators the change to the new is fully established. The difference is not deeply marked, and lies for the most part in minute details. In some cases these seem to point to the adoption in literature of popular forms which had always been current, and which were really older than the forms that (probably owing to the influence of the Ionic poets and historians) had become fashionable with older writers. Thus the TT which in New Attic supplants <r<r cannot possibly have come from this weaker sound 2 ; they are both independent modifications of an earlier KJ or ij ; and the inscriptions show clearly (cf. Cauer, p. 284 sq. ) that era was never used except under Ionic influence. In other cases there are undoubtedly indications of the weakening
1 The tablets of Styra, engraved not later than 480 B.C. (Kirchhoff, Zur Geschichte der Griech. Alph., p. 139 sq.), give an interesting example of Ionic of a Jess complete development.
2 Ou this question, however, the arguments of Ascoli deserve careful consideration. They have considerably modified the judgment of Curtius in the fifth edition of the Urundzwje, cf. pp. 666 ff.
