STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. Classical Philology Volume VIII April, IQIJ Number 2 THE INTERSTATE USE OF THE GREEK DIALECTS By Carl Darling Buck Not only in earlier times, but also, in most parts of Greece, long after Attic had become the norm of literary prose, each state employed its own dialect, both in private and public monuments of internal concern, and in those of a more external or interstate char- acter, such as decrees in honor of foreigners, decisions of interstate arbitration, and, in general, communications between different states. This brief statement of the general situation, which I have given elsewhere and illustrated with some examples, 1 I repeat here by way of introduction to a more detailed consideration, with special reference to peculiar and exceptional cases, of the question of dialect in "interstate" inscriptions. This term is retained, in default of a better, but is to be understood here in the broadest sense, to include, for example, dedications set up away from the home of the dedicator. There is no doubt as to what is the general practice, namely the use of the native dialect 2 of the authors of inscriptions, regardless of the parties to whom they are addressed, the subject-matter, or the place of publication. But there are some real or apparent excep- tions, and such have now and then misled competent epigraphists 1 Introduction to the Study of the Greek Dialects, pp. 154 ff. 2 Likewise the native type of alphabet. But it is to be remembered that the native alphabets were given up in favor of the Ionic long before most of the dialects were given up in favor of the Attic /nuKi). [Classical Philology VIII, April, 1913] 133 134 Carl Darling Buck into unnecessarily vague statements or even directly false assertions as to what is usual. 1 It is obvious that certain forms of dialect mixture which occur have no bearing on the question before us. The influence of the literary epic, which shows itself in varying degrees in the metrical dedications and epitaphs, is no more common in the inscriptions set up abroad than in those set up at home. Nor is the admixture of Attic forms in the later Doric or other dialects generally any different in the interstate inscriptions from that which the same dialect shows in the inscriptions found within its own territory. It is only the substitution of the Attic kolvt) for the local dialect with Attic color- ing that would concern us. In the period of fluctuation between Attic and the local dialect the choice of the former may in some given instances be due to the interstate character of a document. But this should not be assumed without weighing the question in each case whether the difference is really between interstate and other documents, or between those of a formal public character, whether interstate or local, and others, as for example in the Nicareta inscription (IG VII, 3172), where the formally drawn financial con- trast is in Attic, the other portions in Boeotian. Certain aberrations from the native dialect, in the nature of elimination of some specific local peculiarity, have often been pointed out as due to the interstate character of the inscriptions in which they occur, and no doubt correctly in some instances. Thus Solmsen KZ XXXII, 539, speaking of the inscription on the serpent column from Delphi, which is in the Laconian alphabet, but nevertheless has i-rroXefieov without the regular Laconian change of antevocalic e to i, remarks : Es sollte das nationale geschenk aller Hellenen darstellen, soweit sie sich an den freiheitskampfen beteiligt hatten, und so ist es nur naturlich, dass die Lakedaemonier, wenn sie die weihinschrift auch in ihrem alphabet eingraben liessen, doch in der sprache die speciellen eigenheiten ihrer mundart ver- 1 Thus Hoffmann Griech. Dial. I, p. 17, regards the use of Attic in an Arcadian decree in honor of an Athenian as natural, "weil Ehrendekrete fur Angehorige eines fremden Staates schon seit altester Zeit nicht im einheimischen Dialekte, sondern im Dialekte des Geehrten gehalten zu werden pflegten." This statement can only be understood as a momentary lapse on the part of a scholar who must on reflection be well aware that, whatever the explanation of this particular case, the general practice is precisely the opposite. See below, pp. 145 ff . Interstate Use op the Greek Dialects 135 mieden und das den meisten dorischen stammen gemeinsame zu grunde legten; man kann darin einen ansatz zu einer dor. gemeinsprache sehen. Auch Pindar, der fur den ganzen hellenischen adel dorischen stammes dichtete, hat die in rede stehenden besonderheiten des lakonisehen und seinen heimatlichen dialekts durchaus vermieden. Similarly Meister Dorer und Achder I, 9 ff . : Bei Verwendung dieser von Spartanen verfassten Inschriften ftir die Erkenntniss des spartanischen Dialekts sind die bekannten Tatsaehen in Rechnung zu ziehen, dass in den Texten, die ausserhalb der heimischen Land- sohaften aufgestellt ftir andere Greichen und fur den internationalen Verkehr bestimmt waren, besonders exzentrische Eigentiimlichkeiten des Dialekts wie des Alphabets gewohnlich unterdriickt zu werden pflegten. The "gewohnlich" he qualifies in a footnote by "nicht immer." I have no doubt that such elimination of certain specific peculiarities is to be recognized, but I believe that "occasionally, but not usually," instead of "usually, but not always," would be nearer the mark, and further that this is not entirely confined to interstate inscriptions. 1 dedications The matter of the dialect employed in dedications has recently been recalled to notice by the Cleobis and Biton dedication at Delphi, apropos of which Premerstein Oest. Jhrh. XIII, 48 remarks : "In aller Regel sind die Inschriften archaischer Weihungen im Alphabet und Dialekt der Dedikanten gehalten; doch gibt es Ausnahmen, indem auswartige Stifter mitunter der ortsublichen Schrift und Sprache sich bedienen, so z. B. die Sohne des Pariers Charopinos auf ihren delphischen Anathemen." Examples in plenty of the normal practice, the use of the alpha- bet and dialect of the dedicators, regardless of where the objects are set up, are furnished by the dedications from all parts of the Greek world found at the Panhellenic sanctuaries of Olympia, Delos, Delphi, and Dodona, as well as at other sites, e.g., a Naxian dedica- tion at Boeotian Orchomenus (SGDI 5422; Roberts I, No. 28). The artist's signatures are likewise normally in the dialect and alphabet of the artist. Yet now and then the artist working abroad may adapt himself to the local practice. 1 See Introduction to the Study of the Greek Dialects, § 275, and the remarks on vabs, below, p. 137, and on U.v0ay6pas, below, p. 138. 136 Carl Darling Buck In those cases where both dedication and artist's signature belonging to the same work are preserved, and dedicator and artist are not compatriots, we expect and find, as a general rule, the appropriate differentiation. But there are also some instances in which the dedication shows the form appropriate to the artist, indicating that the matter was intrusted to his hands; and others in which, con- versely, the artist's signature is adapted to the dedication; and still others in which the local influence of the place of dedication is a third element. Some illustrations of the normal differentiation are given first, followed by examples of adaptation or local influence. Inschriften von Olympia 271; Loewy Inschriften griechischer Bild- hauer 33: o) [rXaWJai /ue KaXov y£V£[a! F]aA.eiop iiroie. b) [rXjavKiip 6 AvkkiSco [to]i 'Epfirji 'P[rj]yivos. The signature of the Elean artist is in the Elean dialect and alphabet; the dedication in the Ionic dialect and in the alphabet then current at Rhegium. Inschriften von Olympia 143; Loewy 28: a) TeXov Aavo/ieVeos r«\oi]os avtOeice. b) TAavKtas Aiytvaras efVJoieo'e. There is nothing here significant for the dialect. But the contrast between S in the dedication and £ in the artist's signature is not accidental. For that the use of 5 was established at both Gela and Syracuse considerably earlier than in many parts of Greece, for example at Athens where it is rare before the middle of the fifth century, we know from other sources, e.g., the coin legends of Gela and Syracuse, the Delphian dedication of Gelon, quoted below, the Olympian dedication of Hieron (Inschr. von Olympia 249), and the early inscriptions on the steps of the temple of Apollo at Syracuse (Roberts I, No. 110). Inschriften von Olympia 162: Loewy 91 : Of the original inscription, the later restoration of which reads Tlv0oic\.f)$ 'AXeto?. [IIoXv]«;\«to? eirokt. 'Apyeios, enough is pre- served to show clearly the difference in alphabet, the signature of Interstate Use of the Greek Dialects 137 the artist showing the characteristic Argive form of lambda, while the dedication shows the usual form current in Elis. Dedication of Gelon of Syracuse at Delphi: Homolle BCH XXI, 589; Melanges Weil 212 ff.; Hicks 2 16; Ditt. Syll. 2 910: o) Tekov 6 Aeivo/uci^eos] | aveOeKC roiroWovi | SvpapoVios. 6) Tov TptiroSa kcu. rev vutlv lpywra.ro | Bi'ov Aiooooo vios MiAlaios. The dedication is Syracusan in alphabet and dialect. The article without spiritus asper has already been attested for Syracuse, among other places (cf. my Greek Dialects, p. 50), by the Olympian dedication of Hieron, beginning ~H.id.pov 6. The artist's signature is in the Ionic dialect (riv vUev) as is natural, but the alphabet is not the Milesian Ionic (note E=^, 0=g>). It cannot be a case of continuation of the alphabet used in the dedication, for it has h , £ in contrast to C, 5 = 7, o". There seems to be left only one possibility, namely, that the alphabet is Delphian. Yet Delphian inscriptions of about the same date show T =7 and § =a. Inschriften von Olympia 259; Loewy 49: a) Maro-avioi ml Natrcra/crioi av£6ev Au | 'OAv/uiriitoi Sexarav airb rmv iroAe/itW. b) Ilcuuivios iiroirjcre Mev&uos | koi TOKpiOTrjpia iroubv «rl tov vaov iviKa. The alphabet of both a and b is Ionic. This may be thought of as the use of the artist's alphabet in the dedication as well. But the date of the monument is not long before the Ionic alphabet came into general use. In dialect the dedication and artist's inscription are still distinct, the former in Doric, the latter in Ionic, except for vaov. This use of vaos is not due merely to the fact that the particular temple referred to was so called (Dittenberger- Purgold, loc. tit.), but is an early example of the tendency to replace the Attic-Ionic vew by the more cosmopolitan form, which is also observed in dedications from Ionic soil and which results in the establishment of vcuk in the koivyj} 1 Cf. Michel 1209: paaCKds 'AX^Jai»5pos iviBriKe rhv ra6v'A$T)valrii TloTudSi. (Priene, 334 B.C.); and Michel 1134: SevonXeldris lUcrios iviBriKe rbv vattv 'ApriiuSi 'Aypo- rtpat kt\. (Panagoria, second half of the fourth century.) va6s prevails in Attic inscriptions after about 250 B.C. (Meisterhans-Schwyzer 128), and is the Modern Greek form (similarly always Xa6s). 138 Carl Darling Buck Inschriften von Olymjria 144; Loewy 22: Ev0u;u.os Aoxpos 'AoruKAeos rpls 'OXv/mtl' IviKtov exKova 8' ?<7Tijo-£v njvSc /Jporois iuopav. HvOvp.oi Aoxpos djro Ztpo>v i\£e7roirjv xjai lauTp/s]. 'Evos ] dirap^ev. | KpeeuAas | eiroetrev. Attic alphabet and dialect. IG I, 403; Loewy 46; Hoffmann Syll. Ejrigr. Graec. 269; from Athens: TovBt IIvpijs] aveOtKt HoXvfiLvicrTo <£i'Ao[s vlo;, Attic metrical inscriptions, until the Alexandrian period, maintain the Attic forms with a high degree of con- sistency (for examples, cf. Meisterhans-Schwyzer, p. 17). Specifically, 'Aflijrafrj occurs only here and in the dedication of Hegelochus, who was clearly an Ionian resident of Athens (see below, p. 143), as against nearly a dozen examples of *A0ijra£a in early Attic metrical inscriptions; and where a special explanation of this, and also of the use of the Ionic alphabet throughout, is available, it is not proper to ignore it. 140 Cakl Darling Buck an Attic metrical inscription, and that there is no occasion to assume with Kirchhoff and others, that KuSowera? is an error. For the alternation of Ion. rj with Dor. d retained in the final syllable, Mess compares 'ASpirjva 1 ; in Eur. Hipp. 736. One may also recall the similar alternation of Att.-Ion. rj and Dor. a in Kv/3epvrJTa<{ Bacchy- lides 5. 47 and other like cases. Roberts II, 197; from Athens: "Apxqo/ios iiroiiaev 6 Xlbs. li8iK£ n' aviBtKev Adivauu iroXio^ot. The dedication, as well as the signature, is from the hand of the Chian artist. He intended to follow the Attic style of writing, and, after correcting eiroiHo-ev to iwoiE&ev, avoids the use of i? or to. But his native habit shows through in the omission of the spiritus asper (o not ho), and in the forms of certain letters which are at variance with the Attic usage. Passing now from such special cases, where dedication and artist's signature by citizens of different states are juxtaposed or combined, we may consider other dedications or artists' signatures with reference to the question of local influence. A stock example for the substitution of the local alphabet is the one cited by Premerstein, the Delphian dedication, or as we now know two identical dedications, by the sons of the Parian Charopinus : TOIYAPOriNOPAIDESANEGESANTOPAPIO toI Xap/8is Kal Bi'Jtov Tav ftardpa &) tayayoi' toi SVyoi | . . . jteSfS eiroiee hapyaos. The alphabet had been called Argive by Homolle, but Premerstein points out that the characters which occur here are such as have substantially the same form in both the Argive and the Phocian alphabets. The dialect, he thinks, can only be Phocian (Delphian). But this question of dialect is more complicated. The form pardpa, the reading of which is declared certain, is without doubt Delphian, ranging itself beside Del ph. (pdpep, Bdpfiara, and wevrafiapiTevav, with ap from ep, as regularly in Locrian and Elean. This is enough to show that the dedication was not inscribed at Argos, but at Delphi. But did the Delphian stonecutter use his own dialect throughout, or did just this one form of his native dialect slip in, while he otherwise followed correctly an Argive copy ? This latter, I believe, is correct. The form of crasis 1 seen in hapyehs (o 'Apyelo?) is so well attested in early Argolic inscriptions as to demand recogni- tion as one of the characteristics of the dialect. So Arg. rapyeloi (rol 'Apyeloi), Inschr. von Olymjria 250, 251; Arg. HayeXaiSa rapyeCo (6 'AyeXatSa rov 'Apyetov), ibid. 631; Epid. TaltTKXairiel (ra> A-), IG IV, 1203 (the editor reads hauricXaTnel, but Kabbadias' reading is probably correct; cf. Solmsen Inscr. select., p. 45, footnote). From other dialects, leaving Attic out of account, there are a few scatter- ing examples, namely, so as far I know, one each in Corinthian, Arcadian, Rhodian, and several in Cyprian. But none such occurs in Delphian, which has only TwiroXXwvi etc., like most of the other dialects. Where, as in this dedication, it is a question of Delphian or Argive dialect, hapyelos is as distinctive of the latter as is fiardpa of the former. The verb form eiroke Premerstein says cannot stand for the Argive aorist eVoi/rehe, and must therefore be regarded as Phocian and an imperfect. But there is no ground for the inference that, 1 In reality, I believe, elision instead of true crasis; of. my Greek Dialects, p. 73 with footnote. But this does not affect the present argument. 142 Cabl Darling Buck taken as an imperfect, it must be Phocian. The form is equally strange in either dialect. Although intervocalic f lasted longer in Argolic than in Phocian, in a sixth-century inscription it is to be expected in Phocian also (cf. «\efo?, aipel on the altar of Crissa). The absence of contraction in eiroke, which Premerstein calls "das erste urkundliche Zeugniss fiir die unkontrahierte Vorstufe zu eTroiei," but which can be explained only in the manner suggested to him by Kretschmer, namely, as due to the analogical influence of forms like itroieov with regularly uncontracted eo, is just as likely to be Argive as Delphian. Or better, it is no more unlikely. We have some few examples of verb forms with uncontracted er/, as Locr. BoKeei, etc., which are probably due to analogical influence (cf. my Greek Dialects, §§ 42. 4 and 45. 5), but nothing parallel to iiroiee, outside the pages of Herodotus. Accordingly I prefer to believe that the first of the two E's is intended for f (cf. EOINON= foivov in the Delphian stadium inscription, and other similar cases 1 ) and that the correct reading is iiroife. This form might be either Delphian or Argive. In Svyoi = &ya>, if Premerstein's reading is accepted, 2 the B = £ is new for either Delphian or Argive. But in view of the special relations between the Argive and Laconian dialects, the occurrence of f=S in an early Argive inscription, and certain other facts (cf. especially Meister Dorer und Achaer 52 ff . ; in connection with Rhod. to^ =to£', note that Aew has now turned up on a Rhodian vase), it may be said that the eventual appearance of S = £ in Argive has been looked for. Dedication of land to the Muses of Helicon, by Philetaerus, son of Attalus of Pergamum. Two identical copies found in the neigh- borhood of Thespiae. Homolle BCH VIII, 158; 10 VII, 1789; Michel 1103: &t\erripos 'XttoXw IIepya/x,eus dvcfcucc rliv yov rrjs Maio-ijs Trjs "EA.ik \ oXtfTTTOs pa& | atai wp.6pas in 210, 211, 214, eXefe in 219, dire- [yp]d(pa>>9o in 214. Plainly it was not unusual to employ Boeotian forms at this time, and at most only their more nearly consistent use in No. 207 can be attributed to the fact that it was in honor of a Boeotian town. 1 The only trace of the native dialect is the retention of ?jMrai\ov tov 'A9av[at]ov, MeAavoVo hww, Trpo£evov ko\ cvepyirav tov 'Aktibv ypd? persisted longest. INTERSTATE ARBITRATIONS Decisions of interstate arbitration were regularly rendered in the dialect of the arbitrators, and copies in this form were set up by the states involved in the dispute, at home and often also in one of the religious centers as Olympia or Delos. The majority of the extant inscriptions of this class, 1 as it happens, involve arbitrators and disputants who have the same dialect anyway, or had come to use the same at the given date, by the spread of one of the forms of Koivtf. So, for example, besides those most numerous cases in which the use of Attic tcoivq was common to all parties, a decision of unknown arbitrators in a dispute between Sparta 1 Conveniently grouped in Berard De arbitrio inter liberas Graecorum cimtatea. Many of the texts there cited are not copies of the decisions proper, but decrees of parties to a dispute either authorizing arbitration or carrying out its terms. Interstate Use op the Greek Dialects 151 and the Achaeans acting in behalf of Megalopolis, copy found at Olympia (Inschr. von Olympia 47), is in the Doric icoivrj, which at the time, second century B.C., was used throughout the Peloponnesus; a decision of the Aetolians in a dispute between two towns of Phthiotis, Melitea and Perea, copy found at Melitea (Ditt. Syll* 425; SGDI 1415), is in the Northwest Greek koivtj, which was regularly used by the Aetolians, but also at this time, about 200 B.C., in Phthiotis. But there remain some examples of decisions rendered by arbi- trators whose dialect was not the same as that of the disputants. Decision of the Argives in a dispute between the islands of Melos and Cimolos, copy set up at Cimolos, whence it was carried in modern times to Smyrna (IG XII, iii, 1259; Ditt. Syll. 2 428; SGDI 3277; fourth century B.C.). — This is in the Argive dialect; note especially the specifically Argive word dptfreve. Another, recently discovered, Argive inscription (BCH XXXIV, 331 ff.; cf. also Class. Phil. VI, 219; fifth century b.c.) regulates the relations of Cnossus and Tylissus in Crete. If we had the copy set up at Tylissus, to which reference is made in the text, it would show the same Argive dialect. Decision of the Megarians in a dispute between Corinth and Epidaurus, copy found at Epidaurus (IG IV, 926; Ditt. Syll* 452; SGDI 3025; Michel 20; between 242 and 234 b.c.).— Here, it must be admitted, nothing occurs, except in the personal names subscribed, which could not be Epidaurian as well as Megarian, and one feature, the psilosis in en* lapew, is due to the Epidaurian stonecutter (cf. my Greek Dialects § 586). Decision of the Rhodians in a dispute between Samos and Priene, copy found at Priene (Inschr. von Priene 37; SGDI 3758; early second century b.c). — This is in the Rhodian dialect, in the mixed form which appears in all Rhodian inscriptions of the time. An earlier decision of Lysimachus in the same dispute is of course in the Attic Kotvrj, as regularly used by the Macedonians. Decision of the Milesians in the dispute between the Spartans and Messenians, copy found at Olympia (Inschr. von Olympia 52; Ditt. Syll. 2 314; Michel 31; about 135 b.c.).— This is in the Attic koivt), which had long since been adopted by the Ionians, but not yet by 152 Carl Darling Buck either Spartans or Messenians. This decision Is preceded on the stone by a decree of the Eleans, in Doric icoivrj, granting the request of the Messenians that they be allowed to publish the decision at Olympia, and followed by a Roman decree, this last of course in Attic koivt). Of a decision of the Magnesians in a dispute between Itanus and Hierapytna we have parts of two copies, one from Magnesia and one from Itanus (Inschr. von Magnesia 105; Ditt. Syll. 2 929; about 138 b.c). Where the same portion is preserved in both copies, they agree except in some insignificant matters of spelling. The decision is of course in the Attic icoivrj, but various Cretan documents, such as boundary records, decrees, and letters, are incorporated, in their original dialect. TREATIES Of the thirty-five or so epigraphical texts of Greek treaties which are given by R. von Scala Staatsvertrage des Altertums, erster Teil (down to 338 B.C.), all but four are in Attic. That this is the case is due simply to the fact that they are, except the four, treaties between Athens and various other states and are the Athenian versions of the same, found at Athens. The other parties to the treaties are of dif- ferent speech, Boeotians, Thessalians, Locrians, Phocians, Arcadians, Corcyraeans, Eretrians, Mytilenaeans, Argives, Byzantians, etc., but there is no trace of this in the form employed, not even when the oath to be taken by the other party is quoted — with one exception : In a treaty with Corcyra (Scala 143; Ditt. Syll. 2 84; Hicks 2 106; Michel 9; 374 b.c.) there is a feeble attempt, not consistently carried out, to give its proper Doric form to the oath to be taken by the Corcyraeans (a? «a Tt?=eaz> t« of the Athenian oath, though the proper Doric order is ai rk ica, also [ff\d\aaa[av], [Ad]/j,aT[pa], but [y]vv, [8\rjfMoi). In a treaty with the Thessalians (Scala 176; Ditt. Syll. 2 108; Hicks 2 123; Michel 11; 361/0 b.c.) the oath to be taken by the Thessalians is given, like all of the rest of the text, in Attic. This is what is to be expected, not the touch of linguistic realism seen in the Corcyraean oath, which is rare. 1 1 Just as linguistic realism is almost unknown among the Greek prose-writers Xenophon being a notable exception in his fondness for a bit of dialectic color here and there. The puzzle in Thucydides is not that speeches of foreigners and treaties are Interstate Use of the Greek Dialects 153 The four treaties in other dialects which are included in v. Scala's collection are: treaty found at Olympia, and in the Elean dialect, between two towns otherwise unknown, but doubtless, at least one of them, Elean (Scala 33; Inschr. von Olympia 10; Michel 2; SGDI 1150; sixth or fifth century B.C.); another from Olympia, and in the Elean dialect (Scala 27; Inschr. von Olympia 9; Hicks 2 9; Michel 1; SGDI 1149; sixth century b.c.) between the Eleans and the Heraeans of Arcadia ( ? this is the usual view, but is doubted by Niese Gene- thliakon Carl Robert, p. 20) ; treaty between Oeanthea and Chaleion, found at Oeanthea, in the Locrian dialect (Scala 58; IG IX, i, 333; Hicks 2 44; Michel 3; SGDI 1479; fifth century b.c); treaty between the Olynthian Confederacy and Amyntas of Macedon, found at Olynthus, and in the Euboean dialect (Scala 107; Ditt. Syll. 2 77; Hicks 2 95; SGDI 5285; 389-383 b.c). Another, not given in v. Scala's collection, though falling within its chronological limits, is the treaty between Eretria and Histiaea, found at Eretria, and in the local Eretrian variety of Euboean (Michel 7; SGDI 5307; early fourth century b.c). Cf. also the monetary agreement between Mytilene and Phocaea, found at Mytilene and in the Lesbian dialect (IG XII, ii, 1 ; Hicks 2 94; Michel 8; SGDI 213; first half of fourth century b.c); and a customs agreement, between Aegae in Asiatic Aeolis and certain 'OXvixir-qvoi (revised text by Keil-Premerstein Denkschriften Wien. Akad. LIII, 2, 112 ff.; probably late fourth century B.C.), found near the site of Aegae, and in Aeolic, but with some apparently Ionic forms (cf. Kretschmer Glotta III, 301). Of treaties of later date and not in Attic, the most numerous are those of the third and second centuries between various Cretan cities, as : of Drerus with Cnossus (SGDI 4952) ; of Gortyn with Cnossus (SGDI 5015), Lappa (SGDI 5018), Elyrus (SGDI 5014); of Gortyn regularly given in Attic form, but why an exception is made in the case of the two treaties between the Spartans and the Argives (V, 77 and 79). There is no probability in the view of Kirchhoff that all the other texts of treaties, for example of the three between Sparta and Persia (VIII, 18, 37, 58), were copied from actual versions in Attic and that the translation into Attic is never due to Thucydides himself ("deren Urheber auf keinen Fall Thukydides sein konnte, da dieser, wie die beiden letzten Urkunden des filnften Buches beweisen, hellenischen Urkunden nicht attischer Fassung ihre dialektische Fassung zu belassen pflegt" — Kirchhoff Ueber die von Thukydides benutzten Urkunden, p. 153). 154 Carl Darling Buck and Hierapytna with Priansus (SODI 5024); of Hierapytna with Priansus (SGDI 5040), Lyttus (SGDI 5041), Cnossus (SGDI 5073), Praesus (Rev. e%. gr. 1911, 379 ff.), Itanus (Rev. M. gr. 1911, 415 ff.); of Malla with Lyttus (SGDI 5100) ; of Olus with Lyttus, copy found at Athens (SGDI 5147); of Olus with Lato (SGDI 5075); cf. also two inscriptions from Delos containing agreements of Olus and Lato to submit their differences to the arbitration of Cnossus (SGDI 5149 and BCH XXIX, 204 ff.). While the same type of modified Cretan was common to both parties in the case of many of these treaties, e.g., to Gortyn and Cnossus, to Olus and Lato, this is not true of all. That between Drerus and Cnossus, copy found at Dre- rus, is in a mixed type which differs much from that seen in con- temporaneous inscriptions of Cnossus. At Hierapytna a Doric icoivrj, lacking the special characteristics of central Cretan, was employed. It is this which appears in the treaty with Lyttus, the copy of which, though transported to Venice, must have come from Hierapytna, as is shown also by the fact that the same stone contains two other treaties of Hierapytna; likewise in the treaty with Priansus, the copy of which, though transported to Oxford, undoubtedly came from Hierapytna. But the treaty between Hierapytna and Cnossus is in a dialect which indicates that the copy, the actual provenance of which is unknown, represents the Cnossian version, and it is therefore so classed by Blass (SGDI 5073). So the dialect of the treaty of Gortyn and Hierapytna with Priansus is such as to indicate that our copy, transported to Venice, is either the Gortynian or the Priansian version, probably the former and so classed by Blass (SGDI 5024). There are also from Hierapytna, and in the Doric Koivrj, a treaty with the Magnesians on the Maeander (SGDI 5042), and one with Rhodes. But the latter is rightly classed as a Rhodian inscription (SGDI 3749), not because of anything in its dialect which might not, at this time, be Hierapytnian as well as Rhodian, but because it is in form a decree of Rhodes, a copy of which was brought to Hierapytna and there set up, embodying the terms of agreement. The situation is the same in the case of the joint-citizenship agree- ment, made at the behest of the Aetolian League, between the Messenians and the Phigalians (Ditt. Syll? 234; SGDI 4645; Michel 187; between 250 and 222 B.C.). The text is a copy of the Messenian Interstate Use of the Greek Dialects 155 decree, sent to Phigalia, there being added a decree of Phigalia accepting the terms of the former. Among other later treaties in dialect may be mentioned that between certain Lesbian towns, found at Delos, and in the Lesbian dialect (BCH XXIX, 210 ff.; third or second century B.C.); one between the Arcadian Orchomenus and Euaemon, in the Arcadian dialect (Athen. Mitt. XXXIV, 237 ff . ; Solmsen Inscriptiones Seledae 3 2; about 300 b.c.) ; a joint-citizenship agreement between the two Phocian towns of Stiris and Medeon, found at Stiris, and in the Phocian dialect with some peculiarities from the neighboring Boeotian (JGIX,i,32; Ditt. Syll* 42&; Michel 24; SGDI 1539; early second century B.C.); alliance between the Aetolians and Acarnanians, in Northwest Greek icoivr, ('E<£. 'Apx- 1905, 55 ff.; cf. Swoboda Klio X, 397, who dates it between 272 and 265 B.C.). This last was ordered published on bronze stelae at Actium, Thermum, Olympia, Delphi, and Dodona. The copy we have is from Thermum. But Wilhelm ('E<£. 'Apx- 1910, 147) has acutely recognized in a small fragment found at Olympia (Inschr. von Olympia 40) a portion of the copy set up at Olympia. In all of the above-mentioned treaties, whenever the parties to a given treaty are such as differ in their native dialect, the dialect employed is the one appropriate to that party in whose territory the text was found, so far as its provenance is actually known. That is, we have the home versions in the home dialect. We naturally conclude that the corresponding version which was set up in the home of the other party (or in the home of each of the several other parties) was likewise in its dialect — that, for example, if we had the version of the monetary agreement between Mytilene and Phocaea which was set up at Phocaea, it would not be in Lesbian, like the copy we have from Mytilene, but in Ionic. And this I take to be the prevail- ing view, so far as scholars have considered the question at all. 1 1 It is stated in the most positive form by Kirehhoff Ueber die von Thukydides benutzten Urkunden, p. 99, in discussing the treaty quoted in Thuc. v. 47, copies of which were ordered published at Athens (the one of which part has been found), Argos, Mantinea, and Olympia: "Zwar sind die Vertrage selbst in Athen geschlossen worden und die originale Formuliring ihres Wortlautes ist darum sicher in attischer Sprachform erfolgt; daran ist aber gar nicht zu denken, dass die in der Peloponnes publicirten Exemplare diese Fassung beibehalten haben sollten. Im Gegentheil ist fur sie ebenso gewiss die epichorische Sprachform zur Anwendung gelangt, wie fur die 156 Carl Darling Buck Yet we must look to future discoveries to furnish a concrete demonstration of this, for unfortunately there is no treaty of this kind of which we now possess more than one of the versions. 1 Nor, with all the explicitness which we often find in the directions as to where the different copies were to be published, and whether they were to be inscribed on stone or bronze, is there ever any mention of different dialects to be employed. This, however, is merely a lack of specific confirmation and not to be used as an argument in disproof of the employment of different dialects. For direct allusions to the existing differences in dialect are unknown in epigraphical records, as they are rare in literature. Pending the discovery of satisfactory concrete evidence, I hold in Athen publicirten Texte des zu Sparta abgeschlossenen Nikiasfriedens und des ihn erganzenden Bundesvertrages die attische. Selbst das in Olympia aufgestelltes Exemplar, obwohlimNamenundAuftrage aller amAbschlusse des Vertrages Betheilig- ten veroffentlicht, kann doch, da es an Stelle einer Sonderpublication der Eleer zu treten hatte und auf elischem Gebeite jedenfalls durch Vermittelung der Eleer errichtet worden ist, keine andere als Elische Sprachform gezeigt haben." While this statement of Kirchhoff's represents what I believe to be the correct view to take in general of the dialect of different versions, he went too far in denying the possibility of any exceptions and representing it as inconceivable that in this case the formal draft made at Athens was followed in the other copies. See below, p. 157. I do not, however, agree with those who wish to explain the. verbal discrepancies between the text of Thucydides and the extant fragment of the Athenian version by assuming that the former was copied from one of the other versions, namely the Elean; and, still less, with Kirchhoff's own view that the discrepancies are due to corruption of the text of Thucydides. This would imply a degree of textual corruption which, if assumed for the text as a whole, would render it folly for anyone to discuss the "language of Thucydides." Thucydides gave the text of the treaty in all essential agreement with the Attic version, but without seeking verbal accuracy, which is a comparatively modern requirement. Inscriptions have taught us, as Wilhelm has pointed out, that official duplicates were not always precisely identical in form. 1 Two separate versions of a treaty between Olus and Lato found their way to Venice, but, while one came to light only recently, the other long since disappeared and is known only through a transcript of a transcript. Cf. Blass SGDI 5075 and especially Dieters De Cretensium titulis publicis quaestiones epigraphicae 27 ff. It is evident that the two versions differed here and there in the fulness of detail, in phrase- ology, in syntax (once at 5£ ris riva aditcrfirai versus at Si rls ica riva abiK-fyrrii), and in the form of individual words. On this last point one must of course be cautious, for the transcript of the lost stone is full of obvious inaccuracies, but the following differences of this kind between the transcript (A) and the extant stone (B) are signifi- cant: K^sA = /coi^sB; ttotI A = iroprt B ; ttX/ows A. = w\iovev B; 'T&aTlav A = hrrla\v\ B; j>v8ixi^0VT(s A = ^i/ff/n(TTOt[r«] B; irbnirav, irepiafiirh-i^ A= ir(nnra.v, Trepmiriririi B; (but also 7repia/«r£rif B); *OX6n-i, '0\6vtu>v, 'OXovriws, 'OXovrtav A= BoXo'«(v)ti, BoXcfvrioe, BoXocrios, Bo\otvrl(av) B (11. 39, 69, 84, elsewhere 'OXopti, etc.). Except in the case of icijs =Kal is, the A forms quoted show elimination of the local Interstate Use of the Greek Dialects 157 to the general truth of the view that the versions of the several parties to a treaty were published each in its appropriate dialect — but that on the other hand there might well be exceptions to this owing to the special circumstances of the formulation of a given treaty. The terms of a treaty might be incorporated in a decree passed by one party and copies of this decree be sent to the other party or parties. In this case the decree would be published in its original form, like any other decree sent abroad (cf. above, p. 146). We have already noted from the later period two examples of pre- cisely this procedure (p. 154, bottom). And even where the formula- tion was not in the form of a decree, it might now and then happen that copies of a formal draft, made by the recording official of that peculiarities of the B forms. To attribute this to the transcriber, to suppose, for example, that he corrected iroprl into irorl or j>vdpirrovres into j>v8p.l£ovTes, is to give him far too much credit. It is just such variations between local Cretan and Doric (or Attic) Koirfi forms that are now f amiliar to us in Cretan inscriptions of this period. But, while the inscriptions of different towns are distinguished by varying degrees of koiHi influence, we have too little material from Olus or Lato separately, to warrant us in deciding, upon this basis, their respective claims to these versions (a third copy, moreover, was ordered set up at Cnossus). Note also that B itself varies between BoX6«[>']ti, etc., and '0\6vrt, etc., between irepiamr£ri{ (and -tis) and weptap.- ir£ri{, that both versions vary between al and el, and that the Olus-Lato arbitration agreement (SGDI 5149) has both iroprl and irorl, both iyypd^ovr[e]v (nom. pi. like irXlovcv) and irpd%avres. Yet one may hazard the guess that B, which retains in part the old forms Bo\6e(v)ri, etc. (with /3 for earlier f), which occur elsewhere only in an Olus-Lyttus treaty {SGDI 5147), was the copy set up at Olus. It is nothing against this that Lato precedes Olus in the introductory lines, for the order is the same in both versions. In general, the order in which the parties to a treaty are mentioned in the heading or text has no such unqualified significance as has sometimes been assumed, e.g., by Boeckh, who, in commenting on this Olus-Lato treaty in the only version then known, asserts (CIG II, p. 406): "nostra quippe tabula hand dubie a Latiis posita est, quemobrem Latii primo loco nominari solent." So recently, but with more reserve, H. J. Reinach Rev. des Et. Grec. 1911, 381, remarks: "Le fait que, dans l'en-t&te comme dans tout le cours du document, les Hierapytniens sont nommes avant les Praisiens incline a croire que la copie a 6t6 faite sur le texte conserve aux archives d'Hi6rapytna," and points to other Cretan treaties {SGDI 5015, 5018, 5024, 5040) as conforming. But while the precedence of the home city as the "party of the first part" is the natural order, and is the one followed in perhaps the majority of cases, it is by no means an invariable practice upon which we can rely. Note, for example, from Athenian treaties in the Athenian versions found at Athens [o-vp]a%la Boiu\r&v Kal ' ABtivalav] (Hicks 2 84) ; 'Eperpte'a[v o~vp.p.a%lii (Hicks 2 106). As already noted, we have a small fragment of the Olympian copy, besides the nearly complete text from Thermum, of a treaty between the Aetolians and Acarnani- ans. But, aside from the fact that the fragment is too small to permit any satisfactory comparison, all versions were doubtless in the Northwest Greek Koivfi, then employed by both parties, and nothing more than slight accidental variations could be expected. 158 Carl Darling Buck party at whose home the conference of the envoys arranging the terms took place, were sent to the other parties and published without transfer to the local dialect. A notable case, mention of which has been deferred until this point, is the inscription, IG IV, 556, Hicks 2 120, pertaining to a peace agreement between the Greek states, probably one made soon after the battle of Mantinea in 362 B.C. It is not the text of the original agreement, but a resolution to announce this agreement and the attitude of the states toward Persia to the representative of the Persian satraps. This copy was found at Argos, but is in the Attic dialect. 1 Upon this Wilhelm Oest. Jhrh. Ill, 147 and 159, remarks : Der Dialekt ist der attische, wie Dittenberger gelegentlich bemerkte; an sich fur Ort und Zeit auffallig, erklart er sich durch den Umstand dass die Urkunde nicht Sondergelegenheiten der Argeier betrifft Die In- schrift ist nur verstandlich als Beschluss der an der koivjj dpr/vr) betheiligten Hellenen oder vielmehr als Beschluss ihrer bevollmachtigten Vertreter. Es ist anzunehmen dass die Urkunden solcher gemeinsamen Abmachungen attisch abgefasst, und dass sie von den einzelnen Staaten offentlich auf- gestellt, mindestens im Archiv hintergelegt wurden. But this is a dangerously broad conclusion. Attic was certainly not a generally recognized medium for interstate proceedings until a very much later period. It is possible that at just this time the situation in Argos and some of the other states with reference to the use of Attic was peculiar (see above, p. 148, on Arcadia). The most probable conclusion, however, is that of Frankel RhM LVI, 34 (cf . also Hicks, loc. cit.) : " Und zwar notigt die Wahl dieser Mundart zu dem Schlusse, dass unter den verbundeten Staaten Athen eine fuhrende Stelle eingenommen hat." The resolution was drafted by Athens, as the leading state, Sparta holding aloof, and copies were published elsewhere in this form. Another question is, in assuming the existence of two or more home versions, each in its appropriate dialect, what was the form employed in those versions which, as was as often the case, were set up in the name of all parties (.icoivrj) in some center like Olympia ? 1 Doric forms occur in 1. 18. But with 1. 17 begins a new decree (cf. Frankel RhM LVI, 244), and, for all that we can tell from the little that is left of these lines, this may have been wholly in Doric. Interstate Use of the Greek Dialects 159 If we could have before us, as Pausanias once had, the Olympian copy of the treaty of Athens with Argos, Mantinea, and Elis, we should note its dialect with the keenest interest. Yet it would not be a test case. For, as pointed out by Kirchhoff (see above, p. 156, footnote), this copy served a double function, and if it was in the Elean dialect, as assumed by Kirchhoff, and also by Grundy Thu- cydides 55, it was solely on account of its second function as the home version of the Eleans. The copies of such common versions found at Olympia, Delos, etc., are enough to show, what might be taken for granted, that the local dialect was not substituted. But, again unfortunately, they all, like those which have been mentioned above (passim), pertain to parties whose dialect was the same, or had come to be the same, for most of them are comparatively late. We may conjecture that, if we had treaties between parties of different dialect in the copies which were published in their com- mon name, they would show now the dialect of one, now of another, according to the channels through which they reached publication. But this whole matter of the dialect in treaties is the one, of all the questions which have been raised in this article, upon which more light from new discoveries is most needed. Note to p. 142. Similar skepticism of the reading Iwoiit is expressed in the posthumous article of the late Professor Solmsen, Idg. Forsch., XXXI, 448 ff. (see p. 473, footnote), which has just reached me.