The most remarkable characteristic of Greek is one conditioned by the geographical aspect of the land. Few countries are so broken up with mountains as Greece. Not only do mountain ranges as elsewhere on the European continent run east and west, but other ranges cross them from north to south, thus dividing the portions of Greece at some distance from the sea into hollows without outlet, every valley being separated for a considerable part of the year from contact with every other, and inter-communication at all seasons being rendered difficult. Thus till external coercion from Macedon came into play it was never possible to establish a great central government controlling the Greek mainland. The geographical situation of the islands in the Aegean equally led to the isolation of one little territory from another. To these geographical considerations may be added the inveterate desire of the Greeks to make the πόλις, the city state, everywhere and at all times an independent unit, a desire which, originating in the geographical conditions, even accentuated the isolating effect of the natural features of the country. Thus at one time in the little island of Amorgos there were no less than three separate and independent political units. The inevitable result of geographical and political division was the maintenance of a great number of local characteristics in language, differentiating in this respect also each political community from its nearest neighbours. It was only natural that the inhabitants of a country so little adapted to maintain a numerous population should have early sent off swarms to other lands. The earliest stage of colonization lies in the borderland between myth and history. The Greeks themselves knew that a population had preceded them in the islands of the Cyclades which they identified with the Carians of Asia Minor (Herodotus i. 171; Thucydides i. 4. 8). The same population indeed appears to have preceded them on the mainland of Greece, for there are similar place-names in Caria and in Greece which have no etymology in Greek. Thus the endings of words like Parnassus and Halicarnassus seem identical, and the common ending of place-names in -ινθος, Κόρινθος, Προβάλινθος, &c., seems to be the same in origin with the common ending of Asiatic names in -nda, Alinda, Karyanda, &c. Probably the earliest portion of Asia Minor to be colonized by the Greeks was the north-west, to which came settlers from Thessaly, when the early inhabitants were driven out by the Thesprotians, who later controlled Thessaly. The name Aeolis, which after times gave to the N.W. of Asia Minor, was the old name for Thessaly (Hdt. vii. 176). These Thesprotians were of the same stock as the Dorians, to whose invasion of the Peloponnese the later migration, which carried the Ionians to Asia and the Cypriot Greeks to Cyprus, in all probability was due. From the north Aegean probably the Dorians reached Crete, where alone their existence is recorded by Homer (Odyssey, xix. 175 ff.; Diodorus Siculus v. 80. 2); cp. Fick, Vorgriechische Ortsnamen (1906).

Among the Greeks of the pre-Dorian period Herodotus distinguishes various stocks. Though the name is not Homeric, both Herodotus and Thucydides recognize an Aeolian stock which must have spread over Thessaly and far to the west till it was suppressed and absorbed by the Dorian stock which came in from the north-west. The name of Aeolis still attached in Thucydides’ time to the western area of Calydon between the mountains and the N. side of the entrance to the Corinthian gulf (iii. 102). In Boeotia the same stock survived (Thuc. vii. 57. 5), overlaid by an influx of Dorians, and it came down to the isthmus; for the Corinthians, though speaking in historical times a Doric dialect, were originally Aeolians (Thuc. iv. 42). In the Peloponnese Herodotus recognizes (viii. 73) three original stocks, the Arcadians, the Ionians of Cynuria, and the Achaeans. In Arcadia there is little doubt that the pre-Dorian population maintained itself and its language, just as in the mountains of Wales, the Scottish Highlands and Connemara the Celtic language has maintained itself against the Saxon invaders. By Herodotus’ time the Cynurians had been doricized, while the Ionians, along the south side of the Corinthian gulf, were expelled by the Achaeans (vii. 94, viii. 73), apparently themselves driven from their own homes by the Dorian invasion (Strabo viii. p. 333 fin.). However this may be, the Achaeans of historical times spoke a dialect akin to that of northern Elis and of the Greeks on the north side of the Corinthian gulf. How close the relation may have been between the language of the Achaeans of the Peloponnese in the Homeric age and their contemporaries in Thessaly we have no means of ascertaining definitely, the documentary evidence for the history of the dialects being all very much later than Homeric times. Even in the Homeric catalogue Agamemnon has to lend the Arcadians ships to take them to Troy (Iliad, ii. 612). But a population speaking the same or a very similar dialect was probably seated on the eastern coast, and migrated at the beginning of the Doric invasion to Cyprus. As this population wrote not in the Greek alphabet but in a peculiar syllabary and held little communication with the rest of the Greek world, it succeeded in preserving in Cyprus a very archaic dialect very closely akin to that of Arcadia, and also containing a considerable number of words found in the Homeric vocabulary but lost or modified in later Greek elsewhere.

On this historical foundation alone is it possible to understand clearly the relation of the dialects in historical times. The prehistoric movements of the Greek tribes can to some extent be realized in their dialects, as recorded in their inscriptions, though all existing inscriptions belong to a much later period. Thus from the ancient Aeolis of northern Greece sprang the historical dialects of Thessaly and Lesbos with the neighbouring coast of Asia Minor. At an early period the Dorians had invaded and to some extent affected the character of the southern Thessalian and to a much greater extent that of the Boeotian dialect. The dialects of Locris, Phocis and Aetolia were a somewhat uncouth and unliterary form of Doric. According to accepted tradition, Elis had been colonized by Oxylus the Aetolian, and the dialect of the more northerly part of Elis, as already pointed out, is, along with the Achaean of the south side of the Corinthian gulf, closely akin to those dialects north of the Isthmus. The most southerly part of Elis—Triphylia—has a dialect akin to Arcadian. Apart from Arcadian the other dialects of the Peloponnese in historical times are all Doric, though in small details they differ among themselves. Though we are unable to check the statements of the historians as to the area occupied by Ionic in prehistoric times, it is clear from the legends of the close connexion between Athens and Troezen that the same dialect, had been spoken on both sides of the Saronic gulf, and may well have extended, as Herodotus says, along the eastern coast of the Peloponnese and the south side of the Corinthian gulf. According to legend, the Ionians expelled from the Peloponnese collected at Athens before they started on their migrations to the coast of Asia Minor. Be that as it may, legend and language alike connected the Athenians with the Ionians, though by the 5th century B.C. the Athenians no longer cared to be known by the name (Hdt. i. 143). Lemnos, Imbros and Scyros, which had long belonged to Athens, were Athenian also in language. The great island of Euboea and all the islands of the central Aegean between Greece and Asia were Ionic. Chios, the most 497 northerly Ionic island on the Asiatic coast, seems to have been originally Aeolic, and its Ionic retained some Aeolic characteristics. The most southerly of the mainland towns which were originally Aeolic was Smyrna, but this at an early date became Ionic (Hdt. i. 149). The last important Ionic town to the south was Miletus, but at an early period Ionic widened its area towards the south also and took in Halicarnassus from the Dorians. According to Herodotus, there were four kinds of Ionic (χαρακτῆρες γλώσσης τέσσερες, i. 142). Herodotus tells us the areas in which these dialects were spoken, but nothing of the differences between them. They were (1) Samos, (2) Chios and Erythrae, (3) the towns in Lydia, (4) the towns in Caria. The language of the inscriptions unfortunately is a κοινή, a conventional literary language which reveals no differences of importance. Only recently has the characteristic so well known in Herodotus of κ appearing in certain words where other dialects have π (ὅκως for ὅπως, κοῦ for ποῦ, &c.) been found in any inscription. It is, however, clear that this was a popular characteristic not considered to be sufficiently dignified for official documents. We may conjecture that the native languages spoken on the Lydian and Carian coasts had affected the character of the language spoken by the Greek immigrants, more especially as the settlers from Athens married Carian women, while the settlers in the other towns were a mixture of Greek tribes, many of them not Ionic at all (Hdt. i. 146).

The more southerly islands of the Aegean and the most southerly peninsula of Asia Minor were Doric. In the Homeric age Dorians were only one of many peoples in Crete, but in historical times, though the dialects of the eastern and the western ends of the island differ from one another and from the middle whence our most valuable documents come, all are Doric. By Melos and Thera Dorians carried their language to Cos, Calymnus, Cnidus and Rhodes.

These settlements, Aeolic, Ionic and Doric, grew and prospered, and like flourishing hives themselves sent out fresh swarms to other lands. Most prosperous and energetic of all was Miletus, which established its trading posts in the Black Sea to the north and in the delta of the Nile (Naucratis) to the south. The islands also sent off their colonies, carrying their dialects with them, Paros to Thasos, Euboea to the peninsulas of Chalcidice; the Dorians of Megara guarded the entrance to the Black Sea at Chalcedon and Byzantium. While Achaean influence spread out to the more southerly Ionian islands, Corinth carried her dialect with her colonies to the coast of Acarnania, Leucas and Corcyra. But the greatest of all Corinthian colonies was much farther to the west—at Syracuse in Sicily. Unfortunately the continuous occupation of the same or adjacent sites has led to the loss of almost all that is early from Corinth and from Syracuse. Corcyra has bequeathed to us some interesting grave inscriptions from the 6th century B.C. Southern Italy and Sicily were early colonized by Greeks. According to tradition Cumae was founded not long after the Trojan War; even if we bring the date nearer the founding of Syracuse in 735 B.C., we have apparently no record earlier than the first half of the 5th century B.C., though it is still the earliest of Chalcidian inscriptions. Tarentum was a Laconian foundation, but the longest and most important document from a Laconian colony in Italy comes from Heraclea about the end of the 4th century B.C.—the report of a commission upon and the lease of temple lands with description and conditions almost of modern precision. To Achaea belonged the south Italian towns of Croton, Metapontum and Sybaris. The ancestry of the Greek towns of Sicily has been explained by Thucydides (vi. 2-5). Selinus, a colony of Megara, betrays its origin in its dialect. Gela and Agrigentum no less clearly show their descent from Rhodes. According to tradition the great city of Cyrene in Africa was founded from Thera, itself an offshoot from Sparta.

Chief Characteristics of the Greek Dialects

1. Arcadian and Cyprian.—As Cyprian was written in a syllabary which could not represent a consonant by itself, did not distinguish between voiced, unvoiced and aspirated consonants, did not represent at all a nasal before another consonant, and did not distinguish between long and short vowels, the interpretation of the symbols is of the nature of a conundrum and the answer is not always certain. Thus the same combination of two symbols would have to stand for τότε, τόδε, δότε, δοθῆ, τόνδε, τῶδε, τὸ, δή. No inscription of more than a few words in length is found in either dialect earlier than the 5th century B.C. In both dialects the number of important inscriptions is steadily increasing. Both dialects change final ο to υ, ἀπό passing into ἀπύ. Arcadian changes the verb ending -αι into -οι. Arcadian uses δ or ζ for an original gw-sound, which appears in Attic Greek as β: ζέλλω, Attic βάλλω, “throw.” In inflexion both agree in changing -ᾶο of masculine -α stems into αυ (Arcadian carries this form also into the feminine -α stems), and in using locatives in -αι and -οι for the dative, such locatives being governed by the prepositions ἀπύ and ἐξ (before a consonant ἐς in Arcadian). Verbs in -αω, -εω and -οω are declined not as -ω, but as -μι verbs. The final ι of the ending of the 3rd plural present changes the preceding τ to σ: φέρονσι, cp. Laconian (Doric) φέροντι, Attic φέρουσι, Lesbian φέροισι. Instead of the Attic τίς, the interrogative pronoun appears as σίς, the initial σ in Arcadian being written with a special symbol ϟ. The pronunciation is not certain. The original sound was qw, as in Latin quis, whence Attic τίς and Thessalian κίς. In Arcadian καν the Aeolic particle κε and the Ionic αν seem to be combined.

2. Aeolic.—Though Boeotian is overlaid with a Doric element, it nevertheless agrees with Thessalian and Lesbian in some characteristics. Unlike Greek generally, they represent the original qw of the word for four by π before ε, where Attic and other dialects have τ: πέτταρες, Attic τέτταρες. The corresponding voiced and aspirated sounds are similarly treated: Βέλφαιος the adjective in Thessalian to Δελφοί, and φήρ for θήρ. They all tend to change ο to υ: ὄνυμα, “name”; ου for ω in Thessalian: Ἄπλουν, “Apollo”; and υ in Boeotian for οι: ϝυκία (οἰκία), “house.” They also make the dative plural of the third declension in -εσσι, and the perfect participle active is declined like a present participle in -ων. Instead of the Athenian method of giving the father’s name in the genitive when a citizen is described, these dialects (especially Thessalian) tend to make an adjective: thus instead of the Attic Δημοσθένης Δημοσθένους, Aeolic would rather have Δ. Δημοσθένειος. Thessalian stands midway between Lesbian and Boeotian, agreeing with Lesbian in the use of double consonants, where Attic has a single consonant, with or without lengthening of the previous syllable: ἐμμί, Attic εἰμί for an original *esmi; στάλλα, Attic στήλη; ξέννος for an earlier ξένϝος, Attic ξένος, Ionic ξεῖνος, Doric ξῆνος. Where Attic has -ᾶς from an earlier -ανς or -αντς, Lesbian has -αις: ταὶς ἄρχαις accusative in Lesbian for older τὰνς ἄρχανς. Lesbian has no oxyton words according to the grammarians, the accent being carried back to the penult or ante-penultimate syllable. It has also no “rough breathing,” but this characteristic it shared with the Ionic of Asia Minor, and in the course of time with other dialects. The characteristic particle of the dialects is κε, which is used like the Doric κα, the Arcadian καν, and the Attic and Ionic ἄν. Thessalian and Lesbian agree in making their long vowels close, η belonging ει (a close ē, not a diphthong), πατείρ, “father.” The υ sound did not become ü as in Attic and Ionic, and hence when the Ionic alphabet was introduced it was spelt ου, or when in contact with dentals ιου, as in ὀνίουμα = ὄνυμα, “name,” τιούχα = τύχη, “chance”; the pronunciation, therefore, must have been like the English sound in news, tune. Boeotian developed earlier than other dialects the changes in the vowels which characterize modern Greek: αι became ē, καὶ passing into κή: compare πατείρ and ϝυκία above: ει became ι in ἔχι, “has.” Thessalian shows some examples of the Homeric genitive in -οιο: πολέμοιο, &c.; its ordinary genitive of ο- stems is in -οι.

There are some points of connexion between this group and Arcadian-Cyprian: in both Thessalian and Cyprian the characteristic πτόλις (Attic, &c., πόλις) and δαυχνα- for δάφνη are found, and both groups form the “contracting verbs” not in -ω but in -μι. In the second group as in the first there is little that precedes the 5th century B.C. Future additions to our materials may be expected to lessen the gap between the two groups and Homer.

3. Ionic-Attic.—One of the earliest of Greek inscriptions—of the 7th century, at least—is the Attic inscription written in two lines from right to left upon a wine goblet (οἰνοχόη) given as a prize: hός νῦν ὀρχεστον πάντον | ἀταλότατα παίζει τοτο δεκᾶν μιν. The last words are uncertain. Till lately early inscriptions in Ionic were few, but recently an early inscription has been found at Ephesus and a later copy of a long early inscription at Miletus.

The most noticeable characteristic of Attic and Ionic is the change of α into η which is universal in Ionic but does not appear in Attic after another vowel or ρ. Thus both dialects used μήτηρ, τιμή from an earlier μᾱτηρ, τιμα, but Attic had σοφία, πρᾶγμα and χώρα, not σοφίη, πρῆγμα and χώρη as in Ionic. The apparent exception κόρη is explained by the fact that in this word a digamma ϝ has been lost after ρ, in Doric κόρϝα. That the change took place after the Ionians came into Asia is shown by the word Μῆδοι, which in Cyprian is Μᾶδοι; the Medes were certainly not known to the Greeks till long after the conquest of Ionia. While Aeolic and the greater part of Doric kept ϝ, this symbol and the sound w represented by it had disappeared from both Ionic and Attic before existing records begin—in other words, were certainly not in use after 800 B.C. The symbol was known and occurs in a few isolated instances. Both dialects agreed in changing u into ü, so that a u sound has to be represented by ου. The short o tended towards u, so that the contraction of ο + ο gave ου. In the same way short e tended towards i, so that the contraction of ε + ε gave ει, which was not a diphthong but a close ē-sound. In Attic Greek these contractions were represented by O and E respectively till the official adoption of the Ionic alphabet at Athens in 403 B.C. So also were the lengthened syllables which represent in their length the loss of an earlier consonant, as ἔμεινα and ἔνειμα, Aeolic ἔμεννα, ἔνεμμα, which stand for a prehistoric *ἔμενσα and *ἔνεμσα, containing the -σ- of the first aorist, and τοὺς, οἴκους, ἔχουσι representing an earlier τόνς, οἴκονς, ἔχοντι (3 pl. present) or *ἔχοντσι (dative pl. of present participle). Both dialects also agreed in changing τ before ι into σ (like Aeolic), as in ἔχουσι above, and in the 3rd person singular of -μι verbs, τίθησι, δίδωσι, &c., and in noun stems, as in δόσις for an earlier *δότις. Neither dialect used the particle κε or κα, but both have ἄν instead. One of the effects of the change of ᾱ into η was that the combination ᾱο changed in both dialects to ηο, which in all Attic records and in the later Ionic has become εω by a metathesis in the quantity of the vowels: νᾱός, earlier νᾱϝός, “temple,” is in Homeric Greek νηός, in later Ionic and Attic νεώς. In the dative (locative) plural of the -ᾱ stems, Ionic has generally -ηισι on the analogy of the singular; Attic had first the old locative form in -ησι, -ᾱσι, which survived 498 in forms which became adverbs like Ἀθήνησι and θύρᾱσι; but after 420 B.C. these were replaced by -αις, θύραις, &c. The Ionic of Asia Minor showed many changes earlier than that of the Cyclades and Euboea. It lost the aspirate very early: hence in the Ionic alphabet H is ē, not h; it changed αυ and ευ into αο and εο, and very early replaced to a large extent the -μι by the -ω verbs. This confusion can be seen in progress in the Attic literature of the 5th and 4th centuries B.C., δείκνυμι gradually giving way to δεικνύω, while the literature generally uses forms like ἐφίει for ἐφίη (impft.). In Attica also the aspiration which survived in the Ionic of Euboea and the Cyclades ceased by the end of the 5th century. The Ionic of Asia Minor has -ιος as the genitive of ι-stems; the other forms of Ionic have -ιδος.

4. Doric.—As already mentioned, the dialects of the North-West differ in several respects from Doric elsewhere. As general characteristics of Doric may be noted the contractions of α + ε into η, and of α + ο or ω into ᾱ, while the results in Attic and Ionic of these contractions are ᾱ and ω respectively: ἐνίκη from νικάω, Attic ἐνίκα; τιμᾶμες 1 pl. pres. from τιμάω, Attic τιμῶμεν; τιμᾶν gen. pl. of τιμᾱ “honour,” Attic τιμῶν. In inflection the most noticeable points are the pronominal adverbs in locative form: τουτεῖ, τηνεῖ (this from a stem limited to a few Doric dialects and the Bucolic Poets), τεῖδε, ὅπει, &c.; the nom. pl. of the article τοί, ταί, not οἱ, αἱ and so τοῦτοι in Selinus and Rhodes; the 1st pl. of the verb in -μες, not in -μεν, cp. the Latin -mus; the aorist and future in -ξ-, where other dialects have -σ-, or contraction from presents in-ζω; δικάζω, δικάσω, Doric δικάξω, &c.; the future passive with active endings, ἐπιμεληθησεῦντι (Rhodes), found as yet only in the Doric islands and in the Doric prose of Archimedes; the particles αἱ “if” and κα with a similar value to the Aeolic κε and the Attic-Ionic ἄν. Doric had an accentuation system different both from Aeolic and from Ionic-Attic, but the details of the system are very imperfectly known.

In older works Doric is often divided into a dialectus severior and a dialectus mitis. But the difference is one of time rather than of place, the peculiarities of Doric being gradually softened down till it was ultimately merged in the lingua franca, the κοινή, which in time engulfed all the local dialects except the descendant of Spartan, Tzakonian. Here it is possible to mention its varieties only in the briefest form. (a) The southern dialects are well illustrated in the inscriptions of Laconia recently much increased in number by the excavations of the British School at Athens. Apart from some brief dedications, the earliest inscription of importance is the list of names placed on a bronze column soon after 479 B.C. to commemorate the tribes which had repulsed the Persians. The column, originally at Delphi, is now at Constantinople. The most striking features of the dialect are the retention of ϝ at the beginning of words, as in the dedication from the 6th century ϝαναξίβιος (Annual of British School, xiv. 144). The dialect changed -σ- between vowels into -h-, μῶhα for μῶσα “muse.” Later it changed θ into a sound like the English th, which was represented by σ. Before o-sounds ε here and in some other Doric dialects changed to ι: θιός, σιός for θεός “god.” The result of contraction and “compensatory lengthening” was not ει and ου as in Attic and Ionic, but η and ω: ἦμεν infinitive = εἶναι from *esmen; gen. sing. of o-stems in ω: θεῶ, acc. pl. in -ως: θεώς; dy was represented by δδ, not ζ, as in Attic-Ionic; μύσιδδε = μύθιζε. The dialect has many strange words, especially in connexion with the state education and organization of the boys and young men. The Heraclean tables from a Laconian colony in S. Italy have curious forms in -ασσι for the dat. pl. of the participle πρασσόντασσι = Attic πράττουσι. Of the dialect of Messenia we know little, the long inscription about mysteries from Andania being only about 100 B.C. From Argolis there are a considerable number of early inscriptions, and in a later form of the dialect the cures recorded at the temple of Asklepios at Epidaurus present many points of interest. There is also an inscription of the 6th century B.C. from the temple of Aphaia in Aegina. ϝ survives in the old inscriptions: ϝεϝρεμένα (= εἰρημένα); νς, whether original or arising by sound change from -nty, persists till the 2nd century B.C.: hαντιτυχόνσα = ἡ ἀντιτυχοῦσα, τὸνς υἱόνς = τοὺς υἱούς. The dialect of the Inachus valley seems to resemble Laconian more closely than does that of the rest of the Argolic area. Corinth and her colonies in the earliest inscriptions preserve ϝ and ϙ (= Latin Q) before ο and υ sounds, and write ξ and ψ by χσ and φσ, the symbols which are used also for this purpose in old Attic. In the Corcyrean and Sicilian forms of the dialect, λ before a dental appears as ν: Φιντίας = Φιλτίας; and in Sicilian the perfect-active was treated as a present: δεδοίκω for δέδοικα, &c. From Megara has come lately an obscure inscription from the beginning of the 5th century; its colony Selinus has inscriptions from the middle of the same century; the inscriptions from Byzantium and its other Pontic colonies date only from Hellenistic times. In Crete, which shows a considerable variety of subdialects, the most important document is the great inscription from Gortyn containing twelve tables of family law, which was discovered in 1884. The local alphabet has no separate symbols for χ and φ, and these sounds are therefore written with κ and π. As in Argive the combination -νς was kept both medially and finally except before words beginning with a consonant; -ty- was represented by ζ, later by -ττ-, as in Thessalian and Boeotian: ὁπόττοι, Attic ὁπόσοι; and finally by -θθ-; λ combined with a preceding vowel into an au-diphthong: αὐκά, Attic ἀλκή, cp. the English pronunciation of talk, &c. In Gortyn and some other towns -σθ—was assimilated to—θθ, where θ must have been a spirant like the English th in thin; ζ of Attic Greek is represented initially by δ, medially by δδ, but in some towns by τ and ττ: δοός (= ζωός), δικάδδεν (= δικάζειν). Final consonants are generally assimilated to the beginning of the next word. In inflection there are many local peculiarities. In Melos and Thera some very old inscriptions have been found written in an alphabet without symbols for φ, χ, φ, ξ, which are therefore written as πh, κh or ϙh, πσ, κσ. The contractions of ε + ε and of ο + ο are represented by E and O respectively. The old rock inscriptions of Thera are among the most archaic yet discovered. The most characteristic feature of Rhodian Doric is the infinitive in -μειν: δοῦναι, &c. (= Attic δοῦναι), which passed also to Gela and Agrigentum. The inscriptions from Cos are numerous, but too late to represent the earliest form of the dialect.

(b) The dialects of N.W. Doric, Locrian, Phocian, Aetolian, with which go Elean and Achaean, present a more uncouth appearance than the other Doric dialects except perhaps Cretan. Only from Locris and Phocis come fairly old inscriptions; later a κοινή was developed, in which the documents of the Aetolian league are written, and of which the most distinctive mark is the dative plural of consonant stems in -οις: ἀρχόντοις (= Attic ἄρχουσι), ἀγώνοις (= Attic ἀγῶσι), &c. Phocian and the Locrian of Opus have also forms like Aeolic in -εσσι. In place of the dative in -ῳ, locatives in -οι are used in Locrian and Phocian. Generally north of the Corinthian gulf the middle present participle from -εω-verbs ends in-ειμενος; similar forms are found also in Elean. Locrian changed ε before ρ into α: πατάρα for πατέρα; cf. English Kerr and Carr, sergeant and Sargeaunt. στ appears for σθ, and ϙ and ϝ are still much in use in the 5th century B.C. Many thousands of inscriptions were found in the French excavations at Delphi, but nothing earlier than the 5th century B.C. In the older inscriptions the Aeolic influence—datives in -εσσι, ὄνυμα for ὄνομα—is better marked than later. In the Laws of the Labyad phratry (about 400 B.C.) the genitive is in ου, but a form in -ω is also found, ϝοίκω, which seems to be an old ablative fossilized as an adverb. The nom. pl. δεκατέτορες is used for the acc.; similar forms are found in Elean and Achaean.

The more important of the older materials for Achaean come from the Achaean colonies of S. Italy, and being scanty give us only an imperfect view of the dialect, but it is clearly in its main features Doric. Much more remarkable is the Elean dialect known chiefly from inscriptions found at Olympia, some of which are as early as the beginning of the 6th century. The native dialect was replaced first by a Doric and then by the Attic κοινή, but under the Caesars the archaic dialect was restored. Many of its characteristics it shares with the dialects north of the Corinthian gulf, but it changes original ē to α: μά = μη, &c.; δ was apparently a spirant, as in modern Greek (= th in English the, thine), and is represented by ζ in some of the earliest inscriptions. Final -σ became -ρ; this is found also in Laconian; -ty- became -σσ-, but was not simplified as in Attic to -σ-: ὄσσα = Attic ὄσα.

As we have seen, Ionians, Aetolians and Dorians tended to level local peculiarities and make a generally intelligible dialect in which treaties and other important records were framed. The language of literature is always of necessity to some extent a κοινή: with some Greek writers the use of a κοινή was especially necessary. The local dialect of Boeotia was not easily intelligible in other districts, and a writer like Pindar, whose patrons were mostly not Boeotians, had perforce to write in a dialect that they could understand. Hence he writes in a conventional Doric with Aeolic elements, which forms a strong contrast to that of Corinna, who kept more or less closely to the Boeotian dialect. For different literary purposes Greek had different κοιναί. A poet who would write an epic must adopt a form of language modelled on that of Homer and Hesiod; Alcaeus and Sappho were the models for the love lyric, which was therefore Aeolic; Stesichorus was the founder of the triumphal ode, which, as he was a Dorian of Sicily, must henceforth be in Doric, though Pindar was an Aeolian, and its other chief representatives, Simonides and Bacchylides, were Ionians from Ceos. The choral ode of tragedy was always conventional Doric, and in the iambics also are Doric words like δράω, λάω, &c. Elegy and epigram were founded on epic; the satirical iambics of Hipponax and his late disciple Herondas are Ionic. The first Greek prose was developed in Ionia, of which an excellent example has been preserved to us in Herodotus. Thucydides was not an Ionian, but he could not shake himself free of the tradition: he therefore writes πράσσω, τάσσω, &c., with -σσ-, which was Ionic, but is never found in Attic inscriptions nor in the writers who imitate the language of common life—Aristophanes (when not parodying tragedy, or other forms of literature or dialect), Plato and the Orators (with the partial exception of Antiphon, who ordinarily has -σσ-, but in the one speech actually intended for the law-courts -ττ-). Similarly Hippocrates and his medical school in Cos wrote in Ionic, not, however, in the Ionic of Herodotus, but in a language more akin to the Ionic κοινή of the inscriptions; and this dialect continued to be used in medicine later, much as doctors now use Latin for their prescriptions. The first literary document written in Attic prose is the treatise on the Constitution of Athens, which is generally printed amongst the minor works of Xenophon, but really belongs to about 425 B.C. From the fragment of Aristophanes’ 499 Banqueters and from the first speech of Lysias “Against Theomnestos” it is clear that the Attic dialect had changed rapidly in the 6th and 5th centuries B.C., and that much of the phraseology of Solon’s laws was no longer intelligible by 400 B.C. Among the most difficult of the literary dialects to trace is the earliest—the Homeric dialect. The Homeric question cannot be discussed here, and on that question it may be said quot homines tot sententiae. To the present writer, however, it seems probable that the poems were composed in Chios as tradition asserted; the language contains many Aeolisms, and the heroes sung are, except for the Athenians (very briefly referred to), and possibly Telamonian Ajax, not of the Ionic stock. Chios was itself an Ionicized Aeolic colony (Diodorus v. 81. 7). The hypothesis of a great poet writing on the basis of earlier Aeolic lays (κλέα ἀνδρῶν) in Chios seems to explain the main peculiarities of the Homeric language, which, however, was modified to some extent in later times first under Ionic and afterwards under Athenian influence.

Of Dorian literature we know little. The works of Archimedes written in the Syracusan dialect were much altered in language by the late copyists. The most striking development of the late classical age in Doric lands is that of pastoral poetry, which, like Spenser, is “writ in no language,” but, on a basis of Syracusan and possibly Coan Doric, has in its structure many elements borrowed from the Aeolic love lyric and from epic.

From the latter part of the 5th century B.C. Athens became ever more important as a literary centre, and Attic prose became the model for the later κοινή, which grew up as a consequence of the decay of the local dialects. For this decay there were several reasons. If the Athenian empire had survived the Peloponnesian War, Attic influence would no doubt soon have permeated the whole of that empire. This consummation was postponed. Attic became the court language of Macedon, and, when Alexander’s conquests led to the foundation of great new towns, like Alexandria, filled with inhabitants from all parts of the Greek world, this dialect furnished a basis for common intercourse. Naturally the resultant dialect was not pure Attic. There were in it considerable traces of Ionic. In Attica itself the dialect was less uniform than elsewhere even in the 5th century B.C., because Athens was a centre of empire, literature and commerce. Like every other language which is not under the dominion of the schoolmaster, it borrowed the names of foreign objects which it imported from foreign lands, not only from those of Greek-speaking peoples, but also from Egypt, Persia, Lydia, Phoenicia, Thrace and elsewhere. The Ionians were great seafarers, and from them Athens borrowed words for seacraft and even for the tides: ἄμτωτις “ebb,” ῥαχία “high tide,” an Ionic word ῥηχίη spelt in Attic fashion. From the Dorians it borrowed words connected with war and sport: λοχαγός, κυναγός, &c. A soldier of fortune like Xenophon, who spent most of his life away from Athens, introduced not only strange words but strange grammatical constructions also into his literary compositions. With Aristotle, not a born Athenian but long resident in Athens, the κοινή may be said to have begun. Some characteristics of Attic foreigners found it hard to acquire—its subtle use of particles and its accent. Hence in Hellenistic Greek particles are comparatively rare. According to Cicero, Theophrastus, who came from as near Attica as Eretria in Euboea, was easily detected by a market-woman as no Athenian after he had lived thirty years in Athens. Thoucritus, an Athenian, who was taken prisoner in the Peloponnesian War and lived for many years in Epirus as a slave, was unable to recover the Athenian accent on his return, and his family lay under the suspicion that they were an alien’s children, as his son tells us in Demosthenes’ speech “Against Eubulides.” In the κοινή there were several divisions, though the line between them is faint and irregular. There was a κοινή of literary men like Polybius and of carefully prepared state documents, as at Magnesia or Pergamum; and a different κοινή of the vulgar which is represented to us in its Egyptian form in the Pentateuch, in a later and at least partially Palestinian form in the Gospels. Still more corrupt is the language which we find in the ill-written and ill-spelt private letters found amongst the Egyptian papyri. Not out of the old dialects but out of this κοινή arose modern Greek, with a variety of dialects no less bewildering than that of ancient Greek. In one place more rapidly, in another more slowly, the characteristics of modern Greek begin to appear. As we have seen, in Boeotia the vowels and diphthongs began to pass into the characteristic sounds of modern Greek four centuries before Christ. Dorian dialects illustrate early the passing of the old aspirate θ, the sound of which was like the final t in English bit, into a sound like the English th in thin, pith, which it still retains in modern Greek. The change of γ between vowels into a y sound was charged by the comic poets against Hyperbolus the demagogue about 415 B.C. Only when the Attic sound changes stood isolated amongst the Greek dialects did they give way in the κοινή to Ionic. Thus the forms with -σσ- instead of -ττ- won the day, while modern Greek shows that sometimes the -ρρ- which Attic shared with some Doric dialects and Arcadian was retained, and that sometimes the Ionic -ρσ-, which was also Lesbian and partly Doric, took its place. In other cases, where Ionic and Attic did not agree, forms came in which were different from either: the genitives of masculine ā stems were now formed as in Doric with ᾱ, but the analogy of the other cases may have been the effective force. The form ναός “temple,” instead of Ionic νηός, Attic νεώς, can only be Doric.1 In the first five centuries of the Christian era came in the modern Greek characteristics of Itacism and vowel contraction, of the pronunciation of μπ and ντ as mb and nd and many other sound changes, the loss of the dative and the confusion of the 1st with the 3rd declension, the dropping of the -μι conjugation, the loss of the optative and the assimilation of the imperfect and second aorist endings to those of the first aorist.2 There were meantime spasmodic attempts at the revival of the old language. Lucian wrote Attic dialogue with a facility almost equal to Plato; the old dialect was revived in the inscriptions of Sparta; Balbilla, a lady-in-waiting on Hadrian’s empress, wrote epigrams in Aeolic, and there were other attempts of the same kind. But they were only tours de force, κῆποι Ἀδώνιδος, whose flowers had no root in the spoken language and therefore could not survive. Even in the hands of a cultivated man like Plutarch the κοινή of the 1st century A.D. looks entirely different from Attic Greek. Apart from non-Attic constructions, which are not very numerous, the difference consists largely in the new vocabulary of the philosophical schools since Aristotle, whose jargon had become part of the language of educated men in Plutarch’s time, and made a difference in the language not unlike that which has been brought about in English by the development of the natural sciences. It is hardly necessary to say that these changes, whether of the κοινή or of modern Greek, did not of necessity impair the powers of the language as an organ of expression; if elaborate inflection were a necessity for the highest literary merit, then we must prefer Cædmon to Milton and Cynewulf to Shakespeare.

The Chief Characteristics of Greek.

As is obvious from the foregoing account of the Greek dialects, it is not possible to speak of the early history of Greek as handed down to us as that of a single uniform tongue. From the earliest times it shows much variety of dialect accentuated by the geographical characteristics of the country, but arising, at least in part, from the fact that the Greeks came into the country in separate waves divided from one another by centuries. For the history of the language it is necessary to take as a beginning the form of the Indo-European language from which Greek descended, so far as it can be reconstructed from a comparison of the individual I.E. languages (see Indo-European Languages). The sounds of this language, so far as at present ascertained, were the following:—

(a) 11 vowels: a, ā, e, ē, i, ī, o, ō, u, ū, ǝ (a short indistinct vowel).

(b) 14 diphthongs: ai, au, ei, eu, oi, ou, āi, āu, ēi, ēu, ōi, ōu, ǝi, ǝu.

(c) 20 stop consonants.

Labials: p, b, ph, bh (ph and bh being p and b followed by an audible breath, not f and v).

Dentals: t, d, th, dh (th and dh not spirants like the two English sounds in thin and then, but aspirated t and d).

Palatals: ǩ, ǧ, ǩh, ǧh (kh and gh aspirates as explained above).

Velars: q, g, qh, gh (velars differ from palatals by being produced against the soft palate instead of the roof of the mouth).

Labio-velars: qṷ, qṷ, qṷh, gṷh (these differ from the velars by being combined with a slight labial w-sound).

(d) Spirants—

Closely akin to w and y and often confused with them were the semi-vowels ṷ and ḭ.

(e) Liquids: l, r.

(f) Nasals: m (labial), n (dental), ñ (palatal), ɲ (velar), the last three in combination with similar consonants.

(a) As far as the vowels are concerned, Greek retains the original state of things more accurately than any other language. The sounds of short e and short o in Attic and Ionic were close, so that e + e contracted to a long close e represented by ει, o + o to a long close o represented by ου. In these dialects u, both long and short, was modified to ü, and they changed the long ā to ē, though Attic has ᾱ after ε, ι and ρ. In Greek ǝ appeared regularly as α, but under the influence of analogy often as ε and ο.

(b) The short diphthongs as a whole remained unchanged before a following consonant. Before a following vowel the diphthong was divided between the two syllables, the ι or υ forming a consonant at the beginning of the second syllable, which ultimately disappeared. Thus from a root dheu- “run” comes a verb θέω for θε-ϝω, from an earlier *θευ-ω. The corresponding adjective is θοός “swift,” for θο-ϝο-ς, from an earlier *θου-ο-ς. The only dialect which kept the whole diphthong in one syllable was Aeolic. The long diphthongs, except at the ends of words, were shortened in Attic. Some of these appear merely as long vowels, having lost their second element in the proethnic period. Apparent long diphthongs like those in λῃτουργία, σᾡζω arise by contraction of two syllables.

(c) The consonants suffered more extensive change. The voiced aspirates became unvoiced, so that bh, dh, ḡh, gh, gṷh are confused with original ph, th, ǩh, qh, qṷh: I.E. *bherō (Skt. bharāmi) is Gr. φέρω; I.E. *dhūmos (Skt. dhūmas), Gr. θῡμος; I.E. *ǧhimo- (Skt. 500 hima-), Gr. (δυσ)-χιμο-ς; I.E. *stigh- (Skt. stigh-), Gr. στίχες; I.E. gṷhen- (Skt. han-), Gr. θείνω (probably), φόνος. The palatal and velar series cannot be distinguished in Greek; for the differences between them resort must be had to languages of the satem-group, such as Sanskrit, Zend or Slavonic, where the palatals appear as sibilants (see Indo-European Languages). The labio-velar series present a great variety of forms in the different Greek dialects, and in the same dialect before different sounds. Thus in Attic before o vowels, nasals and liquids, the series appears as π, β, φ; before e and i vowels as τ, β (δ), θ; in combination with u, which led to loss of the ṷ by dissimilation, κ, γ χ. Thus ἕπομαι corresponds to the Latin sequo-r, apart from the ending; βοῦς to Latin bos (borrowed from Sabine), English cow; φόνος “slaughter,” ἕπεφνον, old Irish gonim, “I wound.” Parallel to these forms with p are forms in the Italic languages except Latin and Faliscan, and in the Cymric group of the Celtic languages. The dental forms τ, δ, θ stand by themselves. Thus τις (from the same root as ποῦ, ποῖ, πόθεν, etc.) is parallel to the Latin quis, the Oscan pis, old Irish cía, Welsh pwy, “who?” “what?”; Attic τέτταρες, Ionic τέσσερες “four” is parallel to Latin quattuor, Oscan πετορα, old Irish cethir, old Welsh petguar; τίσις is from the same root as ποινή. For the voiced sound, β is much more common than δ before e and i sounds; thus βίος “life,” from the same root as Skt. jīvas, Latin vīvus; βιός “bowstring,” Skt. jyā, &c. In Arcado-Cyprian and Aeolic, π and β often precede e and i sounds. Thus parallel to Attic τέτταρες Lesbian has πέσσυρες, Homer πίσυρες, Boeotian πέτταρες; Thessalian βέλλομαι, Boeotian βείλομαι alongside of Attic βούλομαι, Lesbian βόλλομαι, Doric βώλομαι and also δήλομαι. In Arcadian and Cyprian the form corresponding to τις was σις, in Thessalian κις, where the labialization was lost (see the article on Q).

A great variety of changes in the stopped consonants arose in combination with other sounds, especially ḭ (a semivowel of the nature of English y), ṷ (w) and s; -τḭ-, -θḭ- became first -σσ- and later -σ- in Attic Greek, -ττ- in Boeotian (the precise pronunciation of -σσ- and -ττ- is uncertain): Attic ὁ-πόσος, earlier ὁ-πόσσος, Boeotian ὁ-πόττος, from the same stem as the Latin quot, quotiens; Homeric μέσσος, Attic μέσος from *μεθιος, Latin medius; -κḭ-, -χḭ- became -σσ-, Attic -ττ-: πίσσα “pitch,” Attic πίττα from *πίκḭα, cp. Latin pix, picis, ἐλάσσων, Attic ἐλάττων comparative to ἐλαχύς. δḭ and γḭ became ζ: Ζεύς (Skt. Dyāuṣ) ἐλπίζω from ἐλπίς, stem ἐλπιδ- “hope,” μαστίζω from μάστιξ, stem μαστῑγ- “lash.”

(d) The sound ṷ was represented in the Greek alphabet by ϝ, the “digamma,” but in Attic and Ionic the sound was lost very early. In Aeolic, particularly Boeotian and Lesbian, it was persistent, and so also in many Doric dialects, especially at the beginning of words. When the Ionic alphabet was adopted by districts which had retained ϝ, it was represented by β: βρόδον Aeolic for ῥόδον, i.e. ϝρόδον. In Attic it disappeared, leaving no trace; in Ionic it lengthened the preceding syllable; thus in Homer ὑποδείσας is scanned with ο long because the root of the verb contained ϝ: δϝει-. Attic has ξένος, but Ionic ξεῖνος for ξένϝος. Its combination with τ became -σσ-, Attic and Boeotian -ττ-, in τέσσερες, τέτταρες, πέτταρες for I.E. gṷetu-.

But the most effective of all elements in changing the appearance of Greek words was the sound s. Before vowels at the beginning, or between vowels in the middle of words, it passed into an h sound, the “rough breathing.” Thus ἑπτά is the same word as the Latin septem, English seven; ἅλ-ς has the same stem as the Latin sal, English sal-t; εὕω for εὐhω is the same as the Latin uro (*eusô). Combined with i or ṷ also it passes into h; ὑμήν, Skt. syūman, “band”; ἡδύς, Doric ἆδύς, Latin suā(d)vis, English sweet; cp. οἴκοιο for *ϝοικοḭο, νηός, Lesbian ναῦος “temple,” through ναϝός from *νασϝο-ς connected with ναίω “dwell.” Before nasals and liquids s was assimilated: μει-δάω, Latin mi-ru-s, English smile; νίφα, Latin nivem, English snow; λήγω, Latin laxus, English slack; ῥέω from *sreu-ō of the same origin as English stream (where t is a later insertion), imperfect ἔῤῥεον for *esreṷom; cp. also φιλομμείδης, ἀγάννιφος, ἄλληκτος.

After nasals s is assimilated except finally; when assimilated, in all dialects except Aeolic the previous syllable is lengthened if not already long: Attic ἔνειμα, ἔμεινα for the first aorist *enemsa, *emensa; but τόνς, τάνς, &c., of the accusative pl. either remained or became in Aeolic τοίς, ταίς, in Ionic and Attic τούς, τάς, in Doric τώς, τάς; cp. τιθείς for *τιθέντς, βάς for *βάντς, είς “one” for *sem-s, then by analogy of the neuter *sens. Assimilation of σ to preceding ρ and λ is a matter of dialect: Ionic θαρσέω, but Attic θαρρῶ, and so also the Doric of Thera: ἔκελσα, but ἔστειλα for *ἔστελσα. With nasals ḭ affected the previous syllable: τεκταίνω (*τεκτṋḭω), where ṋ is the nasal of the stem τέκτων, itself forming a syllable (see the article N for these so-called sonant nasals). Before ḭ original m becomes n; hence βαίνω with n, though from the same root as English come. Original ḭ does not survive in Greek, but is represented by the aspirate at the beginning of words, ἁγνός = Skt. yajnas; medially after consonants it disappears, affecting the preceding consonant or syllable where a consonant precedes; between vowels it disappears. A sound of the same kind is indicated in Cyprian and some other dialects as a glide or transition sound between two vowels.

(e) The most remarkable feature in the treatment of the nasals is that when n or m forms a syllable by itself its consonant character disappears altogether and it is represented by the vowel α only: τατός, Latin tentus, α- negative particle, Latin in, English un; ἁ-πλόος has the same prefix as the Latin sim-plex (sṃ). The liquids in similar cases show λα or αλ and ρα or αρ: τέ-τλα-μεν, πέ-παλται; ἔδρακον, θρασύς, θάρσος.

The ends of words were modified in appearance by the loss of all stop-consonants and the change of final m to n, ἔδειξε, Latin dixit; ζυγόν, Latin iugum.

Accent.—The vowel system of Greek has been so well preserved because it shows till late times very little in the way of stress accent. As in early Sanskrit the accent was predominantly a pitch accent (see Accent).

Noun System.—The I.E. noun had three numbers, but the dual was limited to pairs, the two hands, the two horses in the chariot, and was so little in use that the original form of the oblique cases cannot be restored with certainty. Ionic has no dual. The I.E. noun had the following cases: Nominative, Accusative, Genitive, Ablative, Instrumental, Locative and Dative. The vocative was not properly a case, because it usually stands outside the syntactical construction of the sentence; when a distinctive form appears, it is the bare stem, and there is no form (separate from the nominative) for the plural. Greek has confused genitive and ablative (the distinction between them seems to have been derived from the pronouns), except for the solitary ϝοίκω = οἴκοθεν in an inscription of Delphi. The instrumental, locative and dative are mixed in one case, partly for phonetic, partly for syntactical reasons. In Arcadian, Elean, Boeotian, and later widely in N. Greece, the locative -οι is used for the dative. The masculine ā-stems make the nom. in most dialects in -ᾱς. The genitive is in -ᾱο (with ο borrowed from the o-stems), which remains in Homer and Boeotian, appears in Arcado-Cyprian as -αυ, and with metathesis of quantity -εω in Ionic. The Attic form in -ου is borrowed directly from the o-stems. In the plural the ᾱ and -o stems follow the article in making their nominatives in -αι and -οι instead of the original -ās and -ōs. The neuter plural was in origin a collective singular, and for this reason takes a singular verb; the plural of ζυγόν “yoke” was originally *iugā, and declined like any other -ā stem. But through the influence of the masculine and feminine forms the neuter took the same oblique cases, and like its own singular made the accusative the same as the nominative. In the plural of -ā and -ō stems, the locative in -αισι, -οισι was long kept apart from the instrumental-dative form in -αις, -οις.

The Verb System.—The verb system of Greek is more complete than that of any of the other I.E. languages. Its only rival, the early Vedic verb system, is already in decay when history begins, and when the classical period of Sanskrit arrives the moods have broken down, and the aorist, perfect, and imperfect tenses are syntactically confused. Throughout the Greek classical period the moods are maintained, but in the period of the κοινή the optative occurs less and less and finally disappears. The original I.E. had two voices, an active and a middle, and to these Greek has added a third, the passive, distinguished from the middle in many verbs by separate forms for the future and aorist, made with a syllable -θη-, τιμηθήσομαι, ἐτιμήθην, though in this instance, τιμήσομαι, the future middle, is often used with a passive sense. Other forms which Greek has added to the original system are the pluperfect—in form a past of the perfect stem with aorist endings. It merely expressed the perfect action in past time, and, except as derived from the context, did not possess the notion of relative time (past at a time already past), which attaches to the Latin forms with the same name. The future optative was also a new formation, betraying its origin in the fact that it is almost entirely limited to Oratio Obliqua. The aorist imperatives were also new; the history of some of them, as the second sing. act. παῦσον, is not very clear. The whole verb system is affected by the distinction between -ō and -mi verbs; the former or thematic verbs have a so-called “thematic vowel” between the root and the personal suffix, while the -mi verbs attach the suffixes directly to the root. The distinction is really one between monosyllabic and disyllabic roots. The history of the personal endings is not altogether clear; the -ō verbs have in the present forms for the 2nd and 3rd person in -εις and -ει, which are not yet elucidated. In the middle, Greek does not entirely agree with Sanskrit in its personal endings, and the original forms cannot all be restored with certainty. The endings of the primary tenses differed from those of the secondary, but there has been a certain amount of confusion between them.

The syntax of the verb is founded on the original I.E. distinction of the verb forms, not by time (tense), but by forms of action, progressive action (present and imperfect), consummated action (aorist), state arising from action, emphatic or repeated action (perfect). For the details of this see Indo-European Languages.

Bibliography.—(i.) A grammar of Greek, which will deal fully with the whole material of the language, is at present a desideratum, and is hardly possible so long as new dialect material is being constantly added and while comparatively so little has been done on the syntax of the dialects. The greatest collection of material is to be found in the new edition of Kühner’s Griechische Grammatik, Laut- und Formenlehre, by Blass (2 vols., 1890-1892); Syntax, by Gerth (2 vols., 1896, 1900). Blass’s part is useful only for material, the explanations being entirely antiquated. The only full historical account of the language (sounds, forms and syntax) at present in existence is K. Brugmann’s Griechische Grammatik (3rd ed., 1900). 501 Gustav Meyer’s Griechische Grammatik (nothing on accent or syntax), which did excellent pioneer work when it first appeared in 1880, was hardly brought up to date in its 3rd edition (1896), but is still useful for the dialect and bibliographical material collected. See also H. Hirt, Handbuch der griech. Laut- und Formenlehre (1902). Of smaller grammars in English perhaps the most complete is that of J. Thompson (London, 1902). The grammar of Homer was handled by D. B. Monro (2nd ed., Oxford, 1891). The syntax has been treated in many special works, amongst which may be mentioned W. W. Goodwin, Syntax of the Greek Moods and Tenses (new ed., 1889); B. L. Gildersleeve and C. W. E. Miller, Syntax of Classical Greek from Homer to Demosthenes, pt. i. (New York, 1901—and following); J. M. Stahl, Kritisch-historische Syntax des griechischen Verbums (1907); F. E. Thompson, Attic Greek Syntax (1907). (ii.) The relations between Greek and the other I.E. languages are very well brought out in P. Kretschmer’s Einleitung in die Geschichte der griechischen Sprache (Göttingen, 1896). For comparative grammar see K. Brugmann and B. Delbrück, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen (the 2nd ed., begun 1897, is still incomplete) and Brugmann’s Kurze vergleichende Grammatik (1902-1903); A. Meillet, Introduction à l’étude comparative des langues indo-européennes (2nd ed., 1908). Greek compared with Latin and English: P. Giles, A Short Manual of Comparative Philology for Classical Students (2nd ed., 1901, with an appendix containing a brief account and specimens of the dialects); Riemann and Goelzer, Grammaire comparative du Grec et du Latin (1901), a parallel grammar in 2 vols., specially valuable for syntax. (iii.) For the dialects two works have recently appeared, both covering in brief space the whole field: A. Thumb, Handbuch der griechischen Dialekte (with bibliographies for each dialect, 1909); C. D. Buck, Introduction to the Study of the Greek Dialects, Grammar, Selected Inscriptions, Glossary (Boston, 1910). Works on a larger scale have been undertaken by R. Meister, by O. Hoffmann and by H. W. Smyth. For the κοινή may be specially mentioned A. Thumb, Die griech. Sprache in Zeitalter des Hellenismus (1901); E. Mayser, Grammatik der griechischen Papyri aus der Ptolemäerzeit: Laut- und Wortlehre (1906); H. St J. Thackeray, A Grammar of the Old Testament in Greek, vol. i. (1909); Blass, Grammar of New Testament Greek, trans. by Thackeray (1898); J. H. Moulton, A Grammar of New Testament Greek. I. Prolegomena (3rd ed., 1906). (iv.) For the development from the κοινή to modern Greek: A. N. Jannaris, An Historical Greek Grammar, chiefly of the Attic Dialect, as written and spoken from Classical Antiquity down to the Present Time (1901); G. N. Hatzidakis, Einleitung in die neugriechische Grammatik (1892); A. Thumb, Handbuch der neugriechischen Volkssprache (2nd ed. 1910). (v.) The inscriptions are collected in Inscriptiones Graecae in the course of publication by the Berlin Academy, those important for dialect in the Sammlung der griech. Dialektinschriften, edited by Collitz and Bechtel. The earlier parts of this collection are to some extent superseded by later volumes of the Inscr. Graecae, containing better readings and new inscriptions. A good selection (too brief) is Solmsen’s Inscriptiones Graecae ad inlustrandas dialectos selectae (3rd ed., 1910). A serviceable lexicon for dialect words is van Herwerden’s Lexicon Graecum suppletorium et dialecticum (2nd ed., much enlarged, 2 vols. 1910). (vi.) The historical basis for the distribution of the Greek dialects is discussed at length in the histories of E. Meyer (Geschichte des Altertums, ii.) and G. Busolt (Griechische Geschichte, i.); by Professor Ridgeway, Early Age of Greece, i. (1901), and P. Kretschmer in Glotta, i. 9 ff. See also A. Fick, Die vorgriechischen Ortsnamen (1905). (vii.) Bibliographies containing the new publications on Greek, with some account of their contents, appear from time to time in Indogermanische Forschungen: Anzeiger (Strassburg, Trübner), annually in Glotta (Göttingen, Vandenhoeck und Ruprecht), and The Year’s Work in Classical Studies (London, Murray).

(P. Gi.)

1 Thumb, Die griechische Sprache im Zeitalter des Hellenismus (1901), pp. 242-243.

2 Thumb, op. cit. p. 249.