Alloglо̄ssoi: Multilingualism and Minority Languages in Ancient Europe 9783110779684, 9783110779523

Table of contents :
Preface
Contents
List of Figures
List of Tables
Priscian, the divisio graeca and the History of Word-formation in Graeco-Roman Grammar
Dream Language and Dream Ideology: Echoes from the Memphis Serapeum
A Tale of Coins and Suffixes: Syracusan Greek ἑξᾶς, Latin sextāns, and Congeners
Dialects in Contact in the Ancient Kingdom of Macedon 
Onomastic Formulae from N. Epirus and S. Illyria: Lingustic and Sociocultural Connotations
Dialect Contact and Koineization: The Case of the Greek Colonies of Aegean Thrace
The Cappadocian Phrasal Compound παιρ-παίνω [per-péno] “Take Away” as an Example of Turkish Pattern Replication
Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire
Motion and Posture Verbs in Multiverb Constructions: Evidence from the New Testament
Contact and Interaction between Greeks and Messapians
The Etymology and Semantics of Oscan pukam
The Messapic Inscription from Grotta Poesia MLM 3 Ro: Analysis with Frame Semantics
Latin uncia à la Heron
List of Contributors
General Index

Citation preview

Alloglо̄ ssoi

Trends in Classics – Greek and Latin Linguistics

General Editors Franco Montanari and Antonios Rengakos Series Editors Albio Cesare Cassio, Georgios K. Giannakis and Michael L. Weiss Advisory Board Luz Conti, Emilio Crespo, Wolfgang de Melo, Panagiotis Filos, Giovanbattista Galdi, Giuseppe Pezzini, Olga Spevak, Olga Tribulato and Andreas Willi

Volume 2

Alloglо̄ ssoi

Multilingualism and Minority Languages in Ancient Europe Edited by Albio Cesare Cassio and Sara Kaczko

ISBN 978-3-11-077952-3 e-ISBN (PDF) 978-3-11-077968-4 e-ISBN (EPUB) 978-3-11-077978-3 ISSN 2940-6374 Library of Congress Control Number: 2023942993 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2023 Walter de Gruyter GmbH, Berlin/Boston Cover image: littleclie / iStock / Getty Images Plus Printing and binding: CPI books GmbH, Leck www.degruyter.com



To the memory of Enzo Lippolis

Preface This volume collects the essays presented at a conference, held at the University of Rome “La Sapienza” in June 2019, that marked the conclusion of a three-year European HERA - JRP “Uses of the Past” Research on Multilingualism and Minority Languages in Ancient Europe (acronym MuMiL-EU; HERA.29.015 | Cassio). This research was aimed at emphasizing that the problems of refusal/integration of speakers of different languages and bearers of different cultures, all too familiar from present-day Europe, did not arise from scratch in relatively modern times, but are deep-rooted in behaviors and practices amply attested in the Ancient World. Scholars from diverse nations took part in this many-sided European research; special attention was paid to the linguistic relationship between the dominant languages, Greek and Latin, and the indigenous languages bound to be cancelled over time except for a number of (often short and barely comprehensible) texts that have come down to us. Yet the classical languages themselves were far from being impenetrable boulders: they could influence each other at both a popular and a sophisticated level, and in their turn be influenced by the very indigenous languages they were on their way to obliterating; not to speak of the impact of Hebrew and Aramaic on the Greek translations of the Old Testament and the text of the New one. In many areas different Greek dialects intermingled with each other and were ultimately obliged to surrender to the koine, “common Greek”. One of the case studies in the volume takes us to medieval/modern Asia Minor and heavy Turkish interference with Cappadocian Greek. For a variety of reasons (colonization, military conquest, cultural prestige) from a certain point onwards Greek was spoken and written outside Greece proper, e.g., in Sicily, Magna Graecia, Thrace, Illyria, Macedonia, Egypt; various essays in this volume deal with literary texts or inscriptions originating from these areas, and special attention has been paid to linguistic interference of some Greek dialects with Italic languages (Oscan, Messapian, Sikel). An outstanding achievement of the HERA research was the acquisition of 3D images of the more than 900 Messapian, Greek and Latin inscriptions found on the vault of the Grotta della Poesia (Melendugno, Lecce), which were otherwise at risk of becoming illegible in a relatively short time. Unfortunately, because of the difficulties encountered in approaching the cave in question and numerous other technical problems, the 3D images were available only in the latest phases of our research; as a consequence, apart from one study contained in this volume, serious examination of most of the texts made available by the new images https://doi.org/10.1515/9783110779684-203

VIII  Preface became impossible for reasons of time. It is to be hoped that study of both the site and the inscriptions will resume before long. In our opinion the vital feature of our HERA research, as mirrored in the essays of this volume, was precisely the inhomogeneity of the linguistic/cultural situations examined, and of the different types of modern approach to them. Classicists are in the habit of relying on comfortable Greek and Latin printed texts, often originating in an alleged “golden age” of the language, conveying the idea of a self-sufficient and self-contained linguistic world. We hope that the case studies found in this volume will contribute to emphasize to what extent the dominant languages of the ancient western world were heavily indebted to each other and to the “barbarian” languages with which they came into contact through the passage of time. The often despised linguistic and cultural Otherness provided fresh stimuli for time-honored civilizations, ultimately turning out beneficial to those who had refused it at first: an important lesson we had better learn quickly in our troubled times. A.C. Cassio; S. Kaczko

Contents Preface  VII List of Figures  XI List of Tables  XIII Luca Alfieri Priscian, the divisio graeca and the History of Word-formation in Graeco-Roman Grammar  1 Carla Bruno Dream Language and Dream Ideology: Echoes from the Memphis Serapeum  29 Albio Cesare Cassio A Tale of Coins and Suffixes: Syracusan Greek ἑξᾶς, Latin sextāns, and Congeners  53 Emilio Crespo Dialects in Contact in the Ancient Kingdom of Macedon  59 Panagiotis Filos Onomastic Formulae from N. Epirus and S. Illyria: Lingustic and Sociocultural Connotations  83 Paloma Guijarro Ruano Dialect Contact and Koineization: The Case of the Greek Colonies of Aegean Thrace  115 Mark Janse and Johan Vandewalle The Cappadocian Phrasal Compound παιρ-παίνω [per-péno] “Take Away” as an Example of Turkish Pattern Replication  139 Sara Kaczko Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  163 Felicia Logozzo and Liana Tronci Motion and Posture Verbs in Multiverb Constructions: Evidence from the New Testament  185

X  Contents Mario Lombardo and Giovanni Boffa Contact and Interaction between Greeks and Messapians  215 Marco Mancini The Etymology and Semantics of Oscan pukam  251 Simona Marchesini The Messapic Inscription from Grotta Poesia MLM 3 Ro: Analysis with Frame Semantics  283 Michael Weiss Latin uncia à la Heron  299 List of Contributors  313 General Index  317

List of Figures Fig. 1: Fig. 2: Fig. 3: Fig. 4: Fig. 5: Fig. 6:

Logical structure of the divisio graeca (in Priscian’s version).  13 Logical structure of the divisio latina.  15 Onomastic formulae in N. Epirus and S. Illyria.  109 East Asia Minor Greek dialects.  158 Geographical distribution of Cappadocian and Pharasiot.  159 Lucanian Pelike Trendall 1967: no. 282, side A (from Pugliese Carratelli et al. 1983: fig. 512).  166 Fig. 7: Lucanian Pelike Trendall 1967: no. 282, side A, detail (from Pugliese Carratelli et al. 1983: fig. 514).  167 Fig. 8: Lucanian Pelike Trendall 1967: no. 282, side B (from Pugliese Carratelli et al. 1983: fig. 514).  167 Fig. 9: Strabo’s Messapia.  228 Fig. 10: Chronological and typological distribution of Messapian inscriptions (from Lombardo 2013b).  229 Fig. 11: Grotta Poesia di Rocavecchia (Melendugno, Lecce): view of the interior.  230 Fig. 12: Grotta Poesia di Rocavecchia (Melendugno, Lecce): detail of a portion of the wall with a Messapian inscription (MLM 23 Ro).  231 Fig. 13: Grotta Poesia di Rocavecchia (Melendugno, Lecce): cast of a portion of a wall with ‘palimpsest’ of inscriptions: Messapian (red boxes), Greek (blue box), Latin (yellow boxes).  231 Fig. 14: Percentage of funeral inscriptions on the total known inscriptions: about 230 out of about 640 (= about 38%) and their chronological distribution (from Lombardo 2013b).  232 Fig. 15: Tarentine and Messapian alphabets with values of graphemes (from Lombardo 2014).  233 Fig. 16: Messapia: cults and cult sites (from Lombardo 2013b).  234 Fig. 17: Names and terms referable to the religious-cultic sphere present in the Messapian inscriptions and their percentages in the funeral inscriptions (from Lombardo 2013b).  235 Fig. 18: Oria (Brindisi), Episcopio. Pebble mosaic (ca mq. 2.50). Last quarter of the fourth century B.C. (from Giannotta 2015).  235 Fig. 19: The column from Patù: grooves and inscriptions.  236 Fig. 20: The column from Patù: damaged sides.  237 Fig. 21: The column from Patù: the drawing of the ships.  238 Fig. 22: The column from Patù: detail of the letters near the ships.  238 Fig. 23a-b: The column from Patù: inscriptions within grooves. a) photo of the object; b) epoxy resin cast.  239 Fig. 24: The column from Patù: the greek abecedarium.  240 Fig. 25: Dexylos’ inscription from Lizzano (Taranto). Mid-sixth century B.C. (from Ferrandini Trosi 2015).  240 Fig. 26: Archaic inscription from Torricella (Taranto). First half of the sixth century B.C. (from Ferrandini Troisi 2015).  241 Fig. 27: The column from Patù: Messapic inscription.  241 Fig. 28: The column from Patù: Messapic inscription.  241 https://doi.org/10.1515/9783110779684-205

XII  List of Figures Fig. 29: Fig. 30: Fig. 31: Fig. 32: Fig. 33: Fig. 34: Fig. 35:

The column from Patù: Messapic inscription.  242 Thick grid intentionally drawn.  242 The column from Patù: isolated letters.  242 The column from Patù: grid with Messapic inscription.  242 The column from Patù: drawing of a thick grid.  242 The column from Patù: alphabetic signs.  242 Salve (LE), Masseria Fani: partial abecedarium scratched on an imported lekane. Second half of the 6th cent. BCE  243 Fig. 36: The Vaste abecedarium.  243 Fig. 37: Muro Leccese (Lecce): fluted pillar with Messapic inscriptions.  244 Fig. 38: Muro Leccese (Lecce): fluted pillar with Messapic inscriptions.  245 Fig. 39: Ancient Vereto (Lecce): Fragmentary ten-sided column with an archaic Messapic inscription.  246 Fig. 40: Ancient Vereto (Lecce): fluted column with two archaic Messapic inscriptions.  247 Fig. 41: Photo of MLM 3 Ro. Courtesy of Riccardo Guglielmino.  286 Fig. 42: Detail of the inscription MLM 3 Ro. Courtesy of Teodoro Scarano.  286 Fig. 43: 6th phase of Messapic epigraphy. Drawing S. Marchesini.  287 Fig. 44: Drawing of MLM 3 Ro from a photo taken after polishing (drawing: S. Marchesini).  288 Fig. 45a-c: MLM 3 Car. Drawings from Mommsen (1850), Minervini (1848), and Andriani (1889).  291 Fig. 46: MLM 1 Di (photo: S. Marchesini for MLM).  292

List of Tables Tab. 1: Tab. 2: Tab. 3: Tab. 4: Tab. 5: Tab. 6:

Dialects in the inscriptions found at Aigeai and Pella.  67 Dialects in the Golden Orphic Lamellae found in the ancient Kingdom of Macedon.  68 Subgrouping of the Cappadocian dialects.  159 Types of PCCs.  190 PACs and PCCs with motion and posture verbs in the Gospels.  192 Distribution of PACs and PCCs in the four Gospels.  211

https://doi.org/10.1515/9783110779684-206

Luca Alfieri

Priscian, the divisio graeca and the History of Word-formation in Graeco-Roman Grammar Abstract: This paper focuses on the theories and techniques employed by ancient grammarians to describe complex — especially derived — words. Its aim is twofold. On the one hand, it suggests that Greek-Latin grammar can be considered a “bilingual theory”, that is, a single theory of grammar expressed in two languages, rather than the sum of two theories, both of which are expressed in one language. On the other hand, it describes the theoretical bases for the study of word-formation in antiquity, namely the theory of the πάθη τῆς λέξεως; it distinguishes two strongly different taxonomies in the study of accidentia nominum, namely the divisio graeca and the divisio latina, and explains why and how the techniques employed by Greek and Latin grammarians to describe wordformation differ from the contemporary standard. More in general, the paper aims to reconsider the position of the study of word-formation in respect to the ideal line that divides “synchrony and “diachrony” in Late Antiquity and, especially, in Priscian.

 Introduction In 2004, Kaltz lamented that knowledge was scarce on the history of the notion of word-formation in pre-modern times (Kaltz 2004: 23). In the last few years, a few works on the topic have been published and some of these are specifically devoted to classical antiquity.1 However, many aspects of the history of word-

 The present paper was written thanks to the project MuMiL-EU, Multilingualism and Minority Languages in Ancient Europe (HERA.29.015 | Cassio) funded by the EU through the HERA — JRP program. I sincerely thank M. Benedetti, A. Cassio, and A. Del Tomba for their insightful comments on a former draft of the paper. All the remaining mistakes are, obviously, my own.  1 On the history of word-formation in general, see Forsgren/Kaltz (2004); Kaltz (2004; 2008); Kaltz/Leclerq (2015); Kastovsky (2006); Biondi (2014; 2018); Lindner (2015); Alfieri (2018; 2019). On the Alexandrine grammarians, see Matthaios (2004; 2008) and Lallot (2008); on the Roman grammarians, see Kircher (1988; 1999) and Flobert (1989); on the Roman rhetoricians, see Fögen (2008); for a general discussion, see Vaahtera (1998). https://doi.org/10.1515/9783110779684-001

  Luca Alfieri formation theories in antiquity are still poorly known. The present contribution is devoted to one of these aspects, specifically the word-formation theories of Greek and Roman grammarians. This topic has been selected for two reasons. On the one hand, it shows how structurally bilingual the Roman world was: that is to say, Greek-Latin bilingualism (see Adams 2003; Mullen/James 2012) affected almost all aspects of Latin intellectual life, the theories of grammar being a paramount case. As Vaahtera (1998: 60) puts it: “it can quite correctly be argued that that there was no real division between the Roman and the Greek grammatical traditions”. In other words, the history of word-formation is a perfect case for showing that Greek-Latin grammar is a “bilingual theory”, that is, a single theory of grammar expressed in two languages, rather than the sum of two theories, both of which are expressed in one language.2 On the other hand, our knowledge of word-formation theories in antiquity is not terribly scarce, but can be improved. In her fundamental book, Vaahtera (1998) has laid down the general framework for studying the topic (see § 2). However, her book fails to answer a question that, in my view, is crucial for the history of word-formation. It is well known that the lack of distinction — or, from a retrospective point of view, the confusion — between synchrony and diachrony typical of so-called “traditional linguistics” revolved around the notion of word-formation, which is a chiefly synchronic notion to us, but a chiefly diachronic-ontogenetic notion to all scholars active from Plato to Saussure (Alfieri 2018; 2019). Therefore, from a post-Saussurean perspective, the history of the notion word-formation is, by and large, also the history of the confusion between the modern notions of synchrony and diachrony, but Vaahtera’s work does not say where, why or how has this confusion originated or, more simply, what the relation is between word-formation and the dimension of time in antiquity. Obviously, such a question is too broad to be answered directly. Vaahtera has shown that the data on word-formation were cast in four different fields of study in Greek-Roman times, and they were treated in a slightly different manner in each of them. For reasons of space, only one of these fields will be analysed here, namely the science of grammar. Two factors are behind this choice. First, grammar is the field in which the data on word-formation are described most abundantly. Second, our question enables us to reconsider some of the philological data quoted but underplayed by Vaahtera and to bring into the

 2 An enormous amount of literature has been written on the topic. See Desbordes (1988), Rochette (1997), Basset et al. (2007), Fortes (2009) and Benedetti (2018) for a first discussion.

Priscian, the divisio graeca and the History of Word-formation  

discussion other data that she overlooked or that were not available when her book was published.

 Word-formation in Classical Antiquity Before addressing the grammarians’ theories of word-formation, it is worth recalling the general framework in which the notion of word-formation was cast in antiquity. There is a structural difference between our contemporary notion of derivation and the Greek-Latin one. We divide synchronic derivation, that is, productive word-formation rules which are active in the mind of the speakers (Engl. play + -er → player), and diachronic derivation, that is, etymology, which is known to specialists though inaccessible to speakers (Engl. father < ProtoGerm. *fader < PIE *ph2-tér). To Greek-Latin scholars, on the other hand, derivation was a single ontogenetic notion: that is, it referred to the whole process of creation of Greek or Latin words from the πρῶται φωναί in the first human language, which is chiefly the same process that presided over the creation of neologisms by poets, who could somehow draw on the same magical power of the original νομοθέτης. In antiquity, therefore, the notion of word-formation was neither fully synchronic nor fully diachronic; it rather spread over both dimensions of language study or, more precisely, it represented the exact point in which their division blurred. The lack of distinction between synchrony and diachrony that characterised the ancient notion of word-formation is clearly shown in the following passage by Aristophanes of Byzantium reported by Eustathius 1761, 34 ad ξ 350 (SGLG 6, fr. 24 AB, 18–9): ἔφη δὲ καὶ ὅτι ἡ μαγίς ἀπό τῆς μάζης ἤ τοῦ μαστεύειν [μάττειν, conj. Nauck 1848: 205, LII] ῥηθεῖσα παράγει τὸν μάγειρον, ὅς οὕτω, φησί, λέγεται παρὰ τὸ μαγίδας αἴρειν ἤγουν προσφέρειν.

The term μαγίς “kneaded mass” — Aristophanes says — is named after μᾶζα “barley-cake” or from μάττειν “to knead”, while the noun μάγειρος “cook” is named after the act of serving the kneaded mass (παρὰ τὸ μαγίδας αἴρειν). In other words, modern linguists consider μάγειρος as noun of uncertain etymology which, however, is likely to be formed through the derivational suffix -ρο-, but Aristophanes considers it the result of a lexicalised compound build on the noun μαγίς and the verb αἴρειν, the “cook” being “the one who serves the

  Luca Alfieri kneaded mass” (**μαγίς + αἴρειν (in the sense of προσφέρειν) > **μάγ-αιρ-ος > μάγειρος).3 Passages of this sort, in which a derivational suffix is interpreted as the result of later grammarians would consider a “corrupted” compound’s constituent, are not uncommon in antiquity and they clearly show the continuity between word-formation and etymology in the earliest phases of the science of grammar (on the metaphor of noun corruption, see below, especially fn. 5 and § 3.1).4 The continuity between etymology and word-formation is perfectly comprehensible if it is cast in the theory of the πάθη τῆς λέξεως (Lat. accidentia nominum). After the mythical phase of the νομοθεσία (Lat. impositio nominum), the πρῶται φωναί of the first language did not remain unchanged for good, they collided with one another and built compounds; the single parts of such protocompounds could get blurred, deformed and “corrupted” by different “affections” in the course of time, being eventually transformed — we would rather say grammaticalized — into a παραγωγή (Lat. terminatio). In this view, wordformation represents the most recent phase of the ontogenetic process of language evolution, whereas etymology sheds some light on its most remote phases, which are close to the very origin of language. Such a theory is not discussed explicitly in Greek and Roman sources, but it is referred to by Plato (Crat. 398c), who uses the verb παράγω to refer to the transformations which words undergo in their evolution and the term ὀνόματα συμπλακέντα “compound nouns” (Thaet. 202b); moreover, it is assumed in almost all works on etymology and/or

 3 Frisk, following Schwyzer (1939: 275) and Wackernagel (1909: 326 ff.), considers μάγειρος as a loan from Dor. μάγῑρος with an irregular spelling for /i:/ (GEW: s.v.). The link with the root *μαγ- “to knead” found in μάττειν < *μαγ-jo- and μαγίς < *μαγ-ίδ-ς is formally possible, but odd semantically (DELG: s.v.), since the μάγειρος is rather a “butcher”, than a “baker” in Greek sources. Schwyzer (1939: 471, fn. 12), thus, connects the word with the root *mak- found in Lat. mactare and Pisani (1934) considers it a Macedonian loan taken from the root *makh- found in μάχαιρα “butcher knife” (with Macedonian -g- from *-kh-). However, Beekes ventures that it could also be a non-PIE, possibly pre-Greek, word derived from *mak-ary- (EDG: s.v.). 4 The same situation in which a morpheme is explained as a corrupted compound’s constituent is found in Plato (Crat. 412c6 and 415b1), where the suffixes -σύνη and -ία are traced to σύνεσις “understanding” and ἰέναι “to go”, in Aristarchus of Samothrace (Schironi 2004: 120, fr. 11; 399, fr. 51; 413, fr. 53, and 510 frr. 30–33), Apollonius Dyscolus (GG II/2.254.5 ff. and 260.1 ff.) and Gellius (Noct. 3.19.1–5, 7.12.1–6, 10.5.1–3 and 12.10.1). Note that etymology is usually trusted as a relevant heuristic tool in antiquity, but some authors rejected it (e.g., Sextus Empiricus, Adv. math. 1.241–7; partly Quintilian, Inst. Or. 1.6.32; and the skeptical G.A. Cotta, see Cic., De nat. deo. 2.24–28). On Greek etymology in general, see recently Zucker/Le Feuvre (2021).

Priscian, the divisio graeca and the History of Word-formation  

derivation, representing the common ground between the grammatical and the philosophical approach to the study of language.5 In this framework, the process of derivation as a whole is not the main subject of any specific branch of language science. Nonetheless, some parts of it fall within the interest of four different disciplines, each of which shows its own approach. Having distinguished these four approaches is one of the main merits of Vaahtera’s work (1998). The poets who created new words shared the mythical power of the νομοθέτης to some extent, thanks to which they were able to move the audience and its emotions. Rhetoricians and scholars in poetics, therefore, were interested in the techniques of coining new words and in their effect on the audience. Vaahtera (1994: 19–51) describes this research line quoting Aristotle’s rhetorical works (i.e., Poet. 1457a18–33 and 31–35; Rhet. 1410a34–5 and De int. 16a19–26, but see also Cat. 1a12) and, in Roman times, Cicero (De Or. 3.149–152; Part. Or. 16) and Quintilian (Inst. Or. 1.5.71, 3.3.30 and 36, 7.3.15, 26 and 30–38). However, in the Knights 1374–1380, Aristophanes makes fun of the sophists, who taught the coining of new words, and of the young Athenians who frequented their schools and filled up their discourses with ridiculous new formations in -ικός.6 Some attention to the techniques of coining new words, therefore, should be traced back to the sophistic schools in Athens in the last quarter of the 5th cent. BCE. Still, in Roman times, such a line of study was not particularly original, attesting almost the same classification system of derived words that was better detailed among the grammarians.

 5 The phrase πάθη τῆς λέξεως is first found in Aristotle (Poet. 1460b12), but Plato had already used the perfect πέπονθεν (from πάσχω) to refer to the changes of the noun ἄνθρωπος (Crat. 399b7). The theory of the “corruption” (Gk. παραφθορά) of words is referred to by Plato (Crat. 418a5 ff., 431e9 ff.), Varro (Ling. Lat. 5.7, 7.12), Quintilian (Inst. Or. 1.5.68–9), Dionysius Thrax (GG I/1.30.1–4), Apollonius Dyscolus (GG II/1.164.14), Charisius (GL I.16.29 ff.), Priscian (GL II.177.20 ff.), Probus (GL IV.54.1 ff.), Servius (GL IV.408.22) and Cledonius (GL V.11.9). Apollonius and Herodian, probably following Tryphon and Didyme’s work Περὶ παθῶν, used the same theory to account for the dialectal variation of Greek words (Wackernagel 1979 [1876]: 1428 ff., 1432 ff. and 1448 ff.). 6 See Neil (1909: 180–181): τὰ μειράκια ταυτὶ λέγω, τἀν τῷ μύρῳ,/ ἃ στωμυλεῖται τοιαδὶ καθήμενα· / σοφός γ᾽ ὁ Φαίαξ δεξιῶς τ᾽ οὐκ ἀπέθανεν./ Συνερτικὸς γάρ ἐστι καὶ περαντικός, /καὶ γνωμοτυπικός καὶ σαφὴς καὶ κρουστικός, /καταληπτικός τ᾽ ἄριστα τοῦ θορυβητικοῦ. Similar mockery is found in Nub. 317 ff. (Dover 1968 ad loc.) and Dait. fr. 1 and 28 (Cassio 1977: 32–36). Moreover, in Nub. 658–694 Aristophanes ridiculed Protagoras, who stated specific rules for the gender of derivatives (fr. A28 DK). I am very grateful to A.C. Cassio for having brought these passages to my attention.

  Luca Alfieri The grammarians were not interested in the process of word creation, which concerned the rhetoricians on the one hand, and the philosophers-etymologists on the other, but they described the different types of words (simple, compound and derived) once they are stored in the lexicon (see § 3). The ancient etymologists did not develop a special interest in wordformation: no reference to derivatives is found in the SVF. However, Plato’s Cratylus and the Aristophanes passage show that etymology and wordformation appeared as contiguous notions, and Varro explored such contiguity in a fairly original way. Varro’s etymology extends its domain to the whole process of formation of Latin words, from their remote origin ubi est adytum et initia regis — whatever the exact meaning of the formula — to their actual functioning in Latin. This process of language formation is referred to through the term declinatio (a calque of the Stoic term κλίσις) and is divided into two parts: the declinatio voluntaria, which refers to the creation of new words (primary or derived), and the declinatio naturalis, which refers to the inflection of words once they are formed and stored in the lexicon (Ling. Lat. 8.21–23, 9.34–35, 10.14–15, 10.51–53).7 Moreover, in book 8, Varro supplies two classifications of derived words, either on the base of their referents (Ling. Lat. 8.14–18) or the base of their form/meaning (Ling. Lat. 8.52–57). Finally, he discusses some types of derivatives such as agent nouns in -tor, place names in -ile etc. (Ling. Lat. 8.15, 18, 56, 61–62, 83 and 10.15–16, 39, 74), and a few specific cases of derived and compound words (Ling. Lat. 5.7, 21, 92–93, 105, 119, 141; 6.36, 46; 8.61–62 etc.). Varro’s approach is a unique case in antiquity, but it garnered some success between the 16th and the 18th cent. CE, inspiring Scaliger’s De causis (1540), Sanctius’ Minerva (1587) and, generally, the diachronic-ontogenetic grammars of the German language written in the 17th–18th cent. CE, which are based on a similar idea (Alfieri 2018; 2019). Since the main adequacy criterion to establish the correct form of a word in antiquity was its etymology, lexicographers and antiquarians could also be interested in derivation. In his Noctes Atticae, Gellius studied some types of  7 The term κλίσις is not used technically in the SVF, where its derivatives ἐγκλίνω and ἔγκλισις still mean “to refuse” and “refusal”. In the 1st cent. BCE, Nigidius Figulus employs inclinamentum (Gell. Noct. 4.9.2) calquing κλίσις or ἔγκλισις. Varro prefers declinatio to refer to any modification of a noun, be it an inflectional modification of the “analogical” type (lego, legis, legit, Ling. Lat. 6.37), or a derivational change of the “anomalistic” type (processit et recessit, accessit et abscessit, Ling. Lat. 6.38). Yet the term declinatio is not totally technical in Varro, who also uses, though more rarely, inclinatio (Ling. Lat. 9.1 and 114), transitus (Ling. Lat. 8.39, 9.70 and 10.51), incrementum (Ling. Lat. 8.17 and 9.66), or a periphrasis based on the verb propago, calquing παράγω “to derive” (Ling. Lat. 9.74, cf. Flobert 1989: 744–745).

Priscian, the divisio graeca and the History of Word-formation  

derived words and analysed a few single cases of derived or compound words, offering some of the most insightful views on word-formation in antiquity (Noct. 2.20.1–8 and 21.6–8, 3.19.1–5, 4.9.12–14, 7.5.1–4 and 12.1–6, 9.12.1–2, 10.5.1–3, 11.15.1–8, 12.10.1 and 15.1–2, 13.3.1–2 etc.; see Vaahtera 1994: 94–114). Similarly, lexicographers such as Flavius Caper (2nd cent. CE), Orus (5th cent. CE) and Paulus ex Festus (8th–9th cent. CE) discussed some types of derivatives (Vaahtera 1998: 147–167). Yet, Gellius’ reflections, brilliant as they were, were highly unsystematic and remained isolated, while the lexicographical accounts supplied a fundamental inspiration for the lexicography of the Middle Ages (e.g., Uguccio’s Derivationes, see Biondi 2014; 2018) and, generally, for the German ontogenetic lexica written in the 17th–18th cent. CE (e.g., Stieler, Becher, Wachter, Fulda, and Mäkze).

 The grammarians’ way to word-formation Vaahtera begins her chapter on word-formation in the grammarians with the work of Dionysius Thrax, for whom she accepts the traditional dating to the 2nd cent. BCE, at least in practice (Vaahtera 1998: 52–53). However, the paternity and the dating of the work have been debated in the last decades following Di Benedetto (1958–1959).8 Moreover, recent publications have shed some light on the very initial phases of the science of grammar (see Slater 1986; Matthaios 1999; 2008; Schironi 2004), so that our knowledge of this period of the history of word-formation can be improved. Aristotle employed the term πτῶσις “falling” to refer to any modification of a word, be it inflectional (Φίλωνος “Philon (nom.)” vs. Φίλωνι “Philon (dat.)”, see De int. 16a33) or derivational (ἀνδρεία “courage” vs. ἀνδρεῖος “courageous”, see Cat. 1a12). He also used the term παρώνυμος to refer to the denominative noun (γραμματικός “grammarian” vs. γραμματική “grammar”, see Cat. 1a12), and distinguished various types of nouns (εἶδη), such as the simple noun (ἀπλοῦν), the double noun (διπλοῦν) and the compound noun (πεπλεγμένος or συμπεπλεγμένος, see Poet. 1457a31 ff. and De int. 16a19 ff.).9  8 For a critical assessment of the problem of dating and authenticity of Dionysius’ grammar, see Lallot (1989: 19–26); Robins (1995); Law/Sluiter (1995); and Pagani (2010). 9 For further cases of πτῶσις meaning “inflection”, see De int. 16b16 and Poet. 1457a18–23; for further cases of πτῶσις meaning “derivation”, see Rhet. 1410a34–5. On the notion of πτῶσις in Aristotle, see Belardi (1975: 43, 62; 1985: 121–123); Vaahtera (1998: 117–118); and Lallot (2008: 55 ff.); on the notion of πτῶσις in general, see De Mauro (2005 [19651]: III, 34 ff.), with the speci-

  Luca Alfieri Aristophanes of Byzantium used the term παράγω “to derive” and παράγωγον “derivative” to address the relation between μαγίς “kneaded mass” and μάγειρος “cook” (see § 1) and that between μολοβρός “wild swine” (henceforth “greedy fellow”) and μολοβρίτης “young of the wild swine”.10 In his fragments, however, the terms πτῶσις, κλίσις and σύνθεσις are not found (although he could have used the term ἄκλιτα with reference to indeclinable words, see Callanan 1987: 61). Aristarchus of Samothrace employed the Aristotelian term πτῶσις for (nominal) inflection, but preferred παραγωγή and παράγωγον for derivation, and used σύνθεσις for composition (and σύνθετος σχηματισμός for the compound), following Theophrastus and the Peripatetic tradition. Moreover, he employed συζυγία for any kind of relation between two linguistic forms, the hallmark of which was, in fact, verbal inflection (Lat. coniugatio).11 However, he did not use the terms εἶδος and σχῆμα in their grammatical meaning, to which he preferred the generic label of σχηματισμός (Matthaios 1999: 257–258).

 fications in Belardi (1985: 167 ff.) and Belardi/Cipriano (1990), and Ronzitti (2014). The notion of παρώνυμος can be traced to Speusippus (fr. 45 Isnardi Parente, and Simplicius, in Ar. Cat. 38.19–29.16, cf. Ebbesen 1981: 5) and is particularly complex: it may refer to any denominative noun (as in Dionysius Thrax, GG I/1.29.1 and Apollonius Dyscolus, fr. ζ, GG II/3.46.11), including the feminine ἀσβέστη from ἄσβεστος “inextinguishable” (Herodian. ad Π 123c); to any derived noun (as in Priscian’s book IV de denominativiis); or to any specific noun derived “semantically” (see fn. 34), without reference to its linguistic form (e.g., velox from velocitas), as in the Middle Age philosophers (Ebbesen 2009; Biondi 2014). The notion of “compound noun” is found also in Plato’s Thaet. 202b, where it is said that notions formed by more than one element are συμπλακέντα “compound (or complex)”, thus they can be addressed by “compound (or complex) terms (or nouns)” (ὀνόματα συμπλακέντα). 10 See SGLG 6, fr. 197 AE, 65: ἰστέον δὲ ὅτι μολοβρὸν καὶ παράγωγον αὐτοῦ μολοβρίτης δοκεῖ ἐπὶ συὸς λέγεσθαι (comment in Callanan 1987: 42 ff.). Aristophanes also employed παρώνυμος in relation to the couple νύμφη “bride”: νυμφίος “bridegroom” (SGLG 6.93, fr. 281: ὄτι παρώνυμος τῇ νύμφῃ ὁ νυμφίος), but Slater (1989) has shown that the term is not used in a fully technical sense here. 11 See Mattahios (1999: 78–80, frr. 21–22 for composition and σχηματισμός; 80, fr. 23 for derivation; 93, fr. 48–49 for inflection; 126, fr. 103 for conjugation; and the comment on pp. 254– 259, 261, 293, 354–355, 451). The specialisation of πτῶσις with reference to declension was probably due to Chrysippus’ lost work, Περὶ πέντε πτώσεων (SVF II, fr. 14, 183), although in the Stoic lore πτώσεις refers to the meanings of word-forms rather than to their forms, as was the case in the Peripatetic tradition (Brandenburg 2020). Aristarchus might also have used the term κλίσις for inflection, if Herodian’s words are to be trusted (Εrbse, ad Δ 235); and Charisius confirms that Aristarchus first distinguished between derivation and composition, if his class of simplicia include also derived nouns, as usually thought: Aristarchus discipulus eius [sc. of Aristophanes] illud addidit, ne umquam simplicia compostis aptemus (GL I.117.4–5).

Priscian, the divisio graeca and the History of Word-formation  

Between Pliny the Elder and Cicero — probably thanks to Tryphon’s activity12 — the grammarians’ technical terminology entered Rome. Pliny used derivativus and derivatio, calquing παράγωγον and παραγωγή (GRF M fr. 94 and 95 = Pompeius, GL V.144.14 ff.), and Cicero distinguished between neologisms obtained through inflexio “derivation” (a calque of κλίσις) and those obtained through adiectio verborum “compounding” (a calque of σύνθεσις, see Part. Or. 16). However, the division between derivation and inflection was not yet stable at this time, or at least it was not stable in the “philosophical-etymological” tradition mirrored by Varro (see fn. 7) and, perhaps, Cicero. Between Apollonius Dyscolus and Quintilian the grammarians’ terminology arrived at a standard (Vaahtera 1998: 52 ff.). Quintilian distinguishes derivare, flectere and coniungere (Inst. Οr. 3.3.30 and 36) and the same division is found in Apollonius Dyscolus, Dionysius Thrax, Probus, Donatus, Charisius, Diomedes and Priscian. As a rule, the grammarians use casus (Gk. πτῶσις) or declinatio (Gk. κλίσις) to refer to nominal inflection, coniugatio (Gk. συζυγία) or, rarely, declinatio for verbal inflection (Gk. κλίσις), derivatio for derivation (Gk. παραγωγή) and compositio (Gk. σύνθεσις) for composition.13 On cases, the term declinatio refers to the sum of inflection and derivation, following the Stoic use, and the term derivatio is employed as an umbrella label meaning “formation”, regardless of the difference between inflection, derivation and composition, but the cases of this type are infrequent.14 Similarly, the verbs orior, nascor and

 12 Tryphon discussed derived nouns in a few fragments (Velsen 1853: 58–62, frr. 83–93) and distinguished the personal pronouns, which are in facts primary (πρωτότυπα), from the possessive adjectives, which are derived (παράγωγα, Velsen 1853: 26–27, fr. 28). Moreover, he wrote a Περὶ τῆς ἐν κλίσεσιν ἀναλογίας and a Περὶ τῶν παρωνύμων (Velsen 1853: 3, 58). 13 For the contrast coniugatio/declinatio, see Dionysius Thrax, who defined συζυγία as ἀκόλουθος ῥημάτων κλίσις “regular inflection of verbs” (GG I/3.53.6), and Probus: quidam sic esse nominum declinationes numero quinque voluerunt, sicut verborum coniugationes tres (GL IV.3.3–5). A similar use is found in Charisius (GL I.18.10. vs. 164.15), Donatus (GL IV.373.7 vs. 381.26), Diomedes (GL I. 301.32 vs. 359.10), Priscian (GL II.183.20 vs. II.369.16). For the use of declinatio/κλίσις with reference to verb inflection, see Apollonius Dyscolus (GG II/2.397.7), Dionysius Thrax (GG I/1.53.6), Probus (GL IV.174.23), Charisius (GL I.169.11), Diomedes (GL I.359.10) and Priscian (GL II.442.18). For further examples, see Schad (2007: s.v.). 14 For the use of declinatio meaning “derivation”, see Pompeius (invenimus enim varias declinationes, fons fontius fontanus, mons montius montanus, GL V.144.12). For the use of derivatus meaning “inflected”, see Charisius (sunt quaedam perfecta a diversis instantibus derivata, GL I.371.27), Diomedes (GL I.377.21) and Priscian (GL II.510.23 and 522.19). For derivatio meaning “composition”, see Probus (“lucifer”…“frugifer”…“signifer” et siqua a “ferendo” fuerint derivata, GL IV.14.13). On cases, συζυγία can refer to (pro)nominal derivation (Apollonius Dyscolus, GG II/2.137.1).

  Luca Alfieri (de)duco (chiefly in the passive voice), as well as their derivatives, can refer to the formation of words, be it inflectional or derivational in nature, but their use is not particularly frequent (Schad 2007: s.v.).

. The classification of the accidents Apart from the terminology, which is fairly similar to the modern one, the way in which ancient grammarians look at word-formation sharply differs from ours. To them, the word was not only the most important unit of language, it was also a minimum, indivisible unit.15 Greek-Latin grammar, as a result, is a “grammar of entire words”, that is, a grammar which does not describe the process of word-formation, which is the subject of etymology, but describes the forms of the words stored in the lexicon, which are the synchronic, lexicalised products of the diachronic-ontogenetic process of derivation (Vaahtera 1998: 170). In other words, to ancient scholars, all words displayed some formal or semantic properties which were produced during their process of formation ab origine linguae, as claimed in the theory of πάθη τῆς λέξεως, and the description of these accidental properties of words (Gk. παρεπόμενα, Lat. accidentia), which had been produced during their process of formation, is the major task of the science of grammar as a whole. The accidents of words are mainly of three types. Some are peculiar: they are the accidents which affected specific words during their process of formation from the original tongue, “corrupting” their form. Donatus says that compound words can be formed by intact nouns, e.g., respublica “republic” which is from res “thing” and publica “common”, but also by words which had been corrupted in the course of time, e.g., municeps “citizen” which is built from munus “duty” and capio “take” (GL IV.377.6, similarly also Dionisios Thrax GG I/1.30.1–4). This type of accident is peculiar to each word, so it does not need to be described in grammars and is rather to be discussed in works on etymology. Other accidents are more general and their description indeed concerns grammar. Still, not all general accidents are so general as to affect all words, none excluded. Some of them affect only a specific group of words, such as case for the noun, and tense for the verb. The accidents of this sort are the most im-

 15 See Varro: verbum dico orationis vocalem partem, quae sit minima et indivisa (Ling. Lat. 10.77). For a comment, see Robins (1970 11967]: 70), Taylor (1974: 9 ff.) and Auroux (1994: 174). The Latin grammarians interpret the indivisibility of the word more strictly than the Greeks (Matthaios 1999: 264 fn. 289, 267; Dickey 2007: 122).

Priscian, the divisio graeca and the History of Word-formation  

portant, since they are used to divide words into classes (the partes orationis) and, as is well known, the description of word-classes and their inflections represents the most important part of any Greek or Latin grammar. Besides specific accidents, which affect single words, and “classificatory” accidents, which are common only to the words of a given class, absolutely general accidents that are common to all (major) word classes can also be found. However, in a recent book, Luhtala (2005: 38 ff.) has shown that the description of this type of accident follows two different taxonomies, whose contrast was already clear to the ancients. In a passage (GL V.338.20 ff.), Consentius contrasted a divisio graeca “Greek taxonomy” to what we may label as divisio latina “Latin taxonomy” on their treatment of derivatives and on their definition of the noun (that is, on the difference between the proper noun, the common noun and the epithet).16 Luhtala has examined this contrast, focusing mainly on their definition of the noun, on the different accounts of homonyms and synonyms, and more generally on the different influx of Stoic and NeoPlatonic philosophy in the two divisiones. In the following, therefore, only the classification of derived nouns will be focused on specifically.17

. The divisio graeca The divisio graeca is found in the Τέχνη γραμματική usually attributed to Dionysius Thrax and in Priscian and can also be ascribed to Apollonius Dyscolus and his son Herodian;18 however, it was probably not totally fixed at the time of

 16 See Consentius (GL V.338.20): Divisio graeca magis sequenda est, quae fit in primam positionem et derivationem. Prima positio dividitur in nomen proprium et appellativum et epitheton. derivativorum species sunt novem et specie manant propria ab individuis. Consentius’ view is confirmed by Pompeius (GL V.139.33–5): in proprio nomine quattuor sunt species, in appellativo habemus vigιnti et semptem. has Graeci rettulerunt ad semptem speciem. habetis diligentissime in Apollonio tractatus de speciebus istis. 17 The contrast between the divisio graeca and the divisio latina in the treatment of derived nouns is mentioned by Vaahtera (1994: 60 fn. 226, 92–93) and Luhtala (2005: 42–43), but none of the sources quoted in fn. 1 has focused on it specifically. 18 This view has been traditional since Schneider (GG II/3.39.11 ff.), followed by Lallot (1989: 31–34) and Vaahtera (1998: 76 ff.). In fact, Apollonius recognised εἶδος and σχῆμα as accidents of the noun (as Choeroboscus said, GG IV/1.107.2) and of the verb (GG II/3.72.14 ff.); he treated species before the gender according to the scholiasts to Dionysius Thrax (GG I/3.524.21), as Priscian also did (GL II.57.8); he employed the notion of “termination” no differently from Priscian (see e.g., fnn. 28, 32, 33), and Pompeius confirms that Apollonius treated the species in the Greek way excellently (see fn. 16).

  Luca Alfieri Aristarchus, who knew the εἴδη of patronymics, denominatives (παρώνυμα) and comparatives, though not the other types of derived nouns that would be common later (Matthaios 1999: 293–294). This taxonomy, as a result, is attested earlier, dating to the 2nd–1st cent. BCE, if the traditional dating of the Τέχνη is followed, or to the 2nd cent. CE, if Di Benedetto’s dating of the Τέχνη is preferred (1958–1959, 2000), as in Luhtala (2005), and Priscian followed the taxonomy he read in Apollonius, as is usually assumed. In either case, in the divisio graeca the accidents common to the noun and the verb are of two types: species (Gk. εἶδος) and figura (Gk. σχῆμα).19 The species fall into two groups: verba primitiva or primae positiones (Gk. πρωτότυπα), such as mons “mountain”, and verba derivata (Gk. παράγωγα), such as montanus “mountain (adj.), concerning a mountain”.20 The derived nouns, in this case, build up a super-ordered category that is opposed to simple nouns, and include seven sub-species of derived nouns in Dionysius Thrax (GG I/1.24.6 ff.) and nine in Priscian (GL II.60.5 ff.), namely: patronymicum (Gk. πατρωνυμικόν), possessivum (Gk. κτητικόν), comparativum (Gk. συγκριτικόν), superlativum (Gk. ὑπερθετικόν), deminutivum (Gk. ὑποκοριστικόν), denominativum (Gk. παρώνυμον) and verbale (gr. ῥηματικόν), to which Priscian added the nomen participiale and the nomen adverbiale.21 These seven or nine species are mainly semantic in nature, but they can also be defined formally, at least in part. Alongside with these species, a varying number of further species nominum defined only semantically are found: these latter species lack a hyperonymical label and may include derived as well as simple words. That is, the nomina corporalia, incorporalia, homonyma and synonyma in the case of nomina adpellativa; and the nomen gentile, patrium, interrogativum, infinitum, demonstrativum and many others in the case of the nomina adiectiva. The formae of the nouns fall into three classes: verba simplicia (Gk. ἀπλᾶ), such as mons “mountain”; verba composita (Gk. σύνθετα), such as magnanimus

 19 The grammatical usage of the term σχῆμα was first established by Theophrastus (Novokatko 2020). 20 The species verborum are parallel to the species nominum, thus they include verba primitiva, such as lego, and verba derivativa, such as lecturio, see Priscian (GL II.427.10). 21 Priscian’s terminology is problematic. In book IV, de denominativis, Priscian defines denominativum as “denominative” (denominativum appellatur a voce primitivi sic nominatum, non ab aliqua speciali significatione, sicut suprae dictae species. habent igitur generalem nominationem omnium formam, quae a nomine derivatur, see GL II.117.20). However, in the exposition he describes all types of derivatives, including deverbal nouns, as if denominativum and derivatum were synonyms (Kircher 1988: 202 and 1999, who takes up the criticism of Dionysius’ scholiasts of Priscian, cf. GG I/3.228.4 ff.).

Priscian, the divisio graeca and the History of Word-formation  

“magnanimous” or inimicus “enemy”, since prepositions and pre-verbs are equated with full words in ancient grammar inasmuch as prepositions can be independent words in the sentence (see Apollonius Dyscolus, GG II/2.476.2 ff.); and decomposita (Gk. παρασύνθετα) that is, nouns derived from compounds which, as a consequence, turn out to be both derived and compound, such as magnanimitas “magnanimity”.22 Since decomposita nomina represent the sum of compound and derived words, while simplicia nomina are necessary also primitiva (and vice versa), the logical architecture of the divisio graeca is based on threefold opposition distributed on two classes (species and figurae), that between simple (or primitive) words, derived words and compound words. In this case, therefore, the notions of derivation and composition are fundamental: they represent the first-order properties employed to classify the general accidents of the words and the basic notions around which the whole taxonomy of the types of words revolves (Fig. 1):

primitiva

patronymicum

species

possessivum comparativum derivata

superlativum deminutivum denominativum

General

verbale

accidents

participiale adverbiale simplicia figurae

composita decomposita

Fig. 1: Logical structure of the divisio graeca (in Priscian’s version).

 22 The category of decompounds is probably to be traced to Theophrastus (Matthaios 1999: 262). The equation between compounds and derivatives formed with a preposition is already found in Aristarchus (Schironi 2004: 115, fr. 10).

  Luca Alfieri

. The divisio latina The second taxonomy can be labelled as divisio Latina, although Consentius did not use such a label, because it is found in all Roman grammarians, bar Priscian. Specifically, it is found in Charisius (GL I.153.6), Donatus (GL IV.373.11), Servius (GL IV.429.15), Cledonius (GL V.35.5), and Pompeius (GL V.139.33). This taxonomy, as a result, seems to date back to the 3rd cent. CE, if it is not to be attributed to some lost works. In this taxonomy, the general accidents common to both nouns and verbs are three, genus, qualitas and forma. The genus (Gk. γένος) refers to gender in the case of the noun, but to the difference between finite and non-finite verbal forms in the case of the verb.23 The qualitas (a calque on ποιότης) concerns the distinction between proper and common nouns in case of the noun, but diathesis in case of the verb. The figura refers to the difference between verba simplicia and verba composita, while the class of decomposita is not included in the classification.24 As a result, in this model the difference between simple and derived words does not represent one of the general accidents of the noun. Nonetheless, the grammarians list a fairly large number of different species (usually twentyseven) among the qualitates of the common noun. Such types of nouns are mainly semantic in nature, but they can also be formal, at least in part. For instance, Donatus quotes the following species (GL IV.373.11 ff.): nomina corporalia, incorporalia, primae positionis, derivata, deminutiva, graecae declinationis, Greek nouns with Latin declension, Greek nouns with mixed declension, homonyma, synonyma, possessiva, qualitatem significantia, quantitatem significatia, gentis, patriae, numeri, ordinis, relational nouns (ad aliquid), quasirelational nouns (quasi ad aliquid), generalia, specialia, verbalia, nouns similar to participles, and nouns similar to verbs. Since most of these species are mainly semantic in nature, they can include both simple and derived nouns without distinction.25  23 See Charisius (GL I.164.22), Donatus (GL IV.359.6 and IV.381.14), Servius (GL IV.411.26), Sergius (GL IV.503.5), Cledonius (GL V.16.9), Pompeius (GL V.231.36), and Consentius (GL V.367.8). In the divisio graeca, on the other hand, the genus refers only to nominal gender. 24 Diomedes (GL I.344.27) and Servius (GL IV.411.26) use forma as a calque of σχῆμα instead of figura, following the divisio graeca (see below). Moreover, Macrobius uses forma as a synonym of species, rather than of figura (de formis sive specibus verborum in GL V.625.24). 25 For a similar treatment of species, see Servius (GL IV.429.15 ff.), Cledonius (GL V.35.5) and Pompeius (GL V.139.33). The semantic definition of the nomina derivata clearly paved the way for a certain ambiguity in the definition of diminutives, which is not found in the divisio graeca:

Priscian, the divisio graeca and the History of Word-formation  

In this taxonomy, the contrast between simple and derived nouns is not ignored, but it is downplayed, being cast among second-level properties of the noun, so derived and simple nouns appear only as two among the many possible types of nouns (Fig. 2):

genus

gender diathesis nomina propria nomina corporalia nomina incorporalia

qualitas

nomina primae positionis

General nomina

accidents

communia

nomina derivata nomina deminutiva nomina graecae declinationis …. …. …

figura

simplicia composita

Fig. 2: Logical structure of the divisio latina.

Obviously, interferences between the two models are possible. Probus follows the divisio graeca for the most part but uses ordo instead of species as a calque of εἶδος and adds the nomina deminutiva to the primitiva and the derivata among the species of the noun, as Donatus also did (GL IV.73.34, see fn. 24 for a similar case). In the same vein, Diomedes follows the divisio latina and casts the derivata nomina among the qualitates nominum, but then he lists the same seven types of derived nouns that are found in Dionysius Thrax (GL I.323.19 ff.). And the scholiasts, on cases, merge the two taxonomies, since the Scholia Marciana to Dionysius Thrax list 24 εἴδη of nouns which are almost identical to  Cledonius says that the verbs such as sugillo “bruise” and sorbillo “sip” (from sugo “suck” and sorbeo “suck in”) may appear as diminutive, but are in fact derivatives (GL V.54.33), and a similar problem is discussed by Pompeius (GL V.143.19–25).

  Luca Alfieri those usually found in the divisio latina (GG I/3.384.22 ff.).26 However, the logical contrast between the two taxonomies is clear: the divisio graeca is structurally based on the difference between derivatio and compositio; in the divisio latina, compositio plays a similar role, while derivatio represents only a lower-layer accident.

. Technical terms and procedures of analysis Despite the different taxonomies employed, the way in which the grammarians look at the internal structure of words is quite similar in both cases. As we said, the grammarians do not describe the formation of words in our contemporary sense; they rather describe the forms of the words stored in the lexicon (see § 3.1). However, this is not to say that they ignore the same formal and semantic regularities that we use to describe the internal structure of words in the present day; they confine themselves to describing such regularities differently. In both taxonomies, the words may show a παραγωγή “derivation”, a κατάληξις “closing”, or τέλος “ending”, all terms that were calqued as terminatio or, more rarely, as extremitas and finis in Latin.27 However, the grammarians — especially the Latin ones — use terminatio as a substantive only very rarely; they prefer to say that a given noun desinit or terminat in a certain way, or employ desinentia as a participle (e.g., the nomina in “us” desinentia in Priscian, GL II.119.5).28 Moreover, even if the existence of terminationes is accepted, con 26 Moreover, the sequence in which the accidents are listed is specific to each grammar. Priscian describes species and figurae nominum before case, but species and figurae verborum after verbal conjugation, while Donatus and Charisius both describe species after inflection (GL IV.373.7 and 381.16; GL I.153.6 and 164.13). Dionysius Thrax, on the other hand, describes species and figurae (both nominum et verborum) before inflection (GG I/1.25.3 and I/1.50.1), whereas Probus and Diomedes describe figurae nominum before case, but figurae verborum between person and tense (GL I.301.24 and 359.10; GL IV.51.21 and 155.21). 27 Each term is found in Apollonius Dyscolus, though not in his forerunners. As for τέλος, see GG II/2.218.1 and 217.10 with reference to the nominal ending -ος of the direct case; as for κατάληξις, see GG II/2.387.2 with the verbal ending -ω of the 1sg.pres.ind.; as for παραγωγή, see II/2.259.1 ff. and 261.2 ff., with reference to cases such as -νος in ὀψινός “tardive” from ὀψέ “late”, to -της in ποιητής “poet” from ποιέω “to make” etc. For terminatio, see Charisius (GL I 235.29), Priscian (GL II.117.20, 122.17, 184.5) and Eutyches, who also uses finalitas (GL V.455.21, 460.24); for extremitas, see Priscian (GL II.3.16), Diomedes (GL I.492.28), Charisius (GL I.50.7) and Terentianus Scaurus (GL VII.13.6) etc. For finis, see Priscian (GL II.453.19), Servius (GL IV.409.1), Pompeius (GL V.222.19). 28 See also: in ‘es’ vero desinentibus, quorum genetivus Graecus in ‘ου’ desinit; and in ‘do’ desinentia derivativa tam a verbis quam a nominibus vel a participiis veniunt (Priscian, GL II.66.6

Priscian, the divisio graeca and the History of Word-formation  

ceptually terminatio does not refer to the ending as distinct from the stem, nor does it refer to an autonomous unit which has an existence independent from that of the words in which it appears; rather it represents the rightmost part of the word; the part that derives from a primitive word that attached to another word after its prima positio and got “corrupted” in the course of language evolution. In practice, such a final part of the word may coincide with an ending, such as a in rosa “rose (nom.)” and is in consulis “consul (gen.)”, or with the whole of a suffix and an ending, such as tus in amatus (Priscian, GL II.569.12, 77.9, 284.12 and 106.19–20); on cases, the terminatio can also refer to a part of the word that does not coincide with any significant unit of analysis, such as ar in lacunar “ceiling” from lacus “lake” and as in civitas “city” from civis “citizen” (Priscian, GL IV.118.1 ff.).29 However, the passages in which the notion of terminatio is employed to inspect the internal structure of words are not frequent, especially in the divisio latina. Similarly, the grammarians knew the terms thema (a borrowing of Gk. θέμα) and prima positio (a calque on Gk. πρώτη θέσις, which is found, for instance, in Dionysius Thrax, GG I/3.25.4). Strictly speaking, however, these terms do not refer to the stem as distinct from the ending; rather, they refer to the original word imposed on the object by the first impositor nominum.30 Such original words can also be the source from which all derived and inflected words descend but, in practice, it is a full word inflected in the nominative singular, in the case of a noun, or in the (third person of the) present indicative active in the case of a verb, as Priscian stated (GL II.421.20, similarly Donatus, GL IV.553.3).31  and 122.10); quamvis forma sit possessivae haec, quae in ‘rius’ desinit, tamen diversae significationis nomina in hac quoque inveniuntur; and alia vero in ‘mentum’ desinunt (Priscian, GL II.75.2 and 125.12); id est in ‘us’ masculinum, in ‘a’ femininum, in ‘um’ neutrum terminat (Priscian, GL II.556.26). The term desinentia as a substantive is unattested before the Middle Ages (Lindner 2012: 140). The hyphen signaling morphemic analysis is uncommon before Bopp (1816). 29 The relation between the termination and an ex-word is clear in Apollonius, who says that each derivative (παρηγμένον), be it nominal or verbal, can be analysed into a πρωτότυπον added to a word that means the same as the παραγωγή used (καθόλου πᾶν παρηγμένον ἀπό τινος ἀνάλυσιν ἔχει τὴν πρὸς τὸ πρωτότυπον μετὰ λέξεως τῆς σημαινούσης ταὐτὸν τῇ παραγωγῇ, GG II/2.326.11 ff. and Householder 1981: 38, 163). 30 See Priscian (ipsius positio nominationis, GL III.119.18), who takes up a similar passage in Apollonius Dyscolus (ἡ…θέσις τοῦ ὀνόματος, GG II/24.2). 31 The idea of the nominative as a general case or as non-case is found in Aristotle (An. pr. 48b14), who distinguishes κλῆσις τῶν ὀνομάτων “nominative” and πτῶσις τῶν ὀνομάτων “oblique case”; in the scholiasts to Dionysius Thrax (GG I/3.546.6, 548.4), who attributed it to Apollonius Dyscolus and Herodian (GG I/3.546.10 and 16–7); in Charisius (GL I.195.13), Probus (GL IV.56.35), and Priscian, according to whom the nominative abusive casus nominatur, since

  Luca Alfieri On cases, especially though not only, when the different stems of a pronoun are looked at, a more technical use of the notion of θέμα or positio can be found.32 But passages of this sort are fairly uncommon, especially before Priscian (see § 3.3). Cases in which the internal structure of words is focused on more precisely, or in which something similar to what we would now call a word-formation rule is proposed, can be found. However, even when they are found, the input form for any word-formation rule is a full word, rather than a stem in the modern sense. Diomedes, for instance, says that nouns in a build their diminutive adding ol: galea → galeola (GL I.325.25). Similarly, Servius thinks that nouns in tor are built by substituting tor for the last syllable of the supine, ut amatu amator (GL IV.430.27–9).33 And discussing the nomina gentis, Pompeius proposes a general rule according to which derived nouns display a larger number of syllables than the simple nouns from which they are built, as is shown in diminutives such as fons → fonticulus, where the number of syllables increases although the meaning decreases (quando sensus minuitur saepe crescit numerus syllabarum, GL I.325.25).34 However, such detailed rules are fairly rare before Priscian and are usually used to establish the correct form of the nouns stored in

 it is a casus generalis that facit alios casus, falling in all the positiones found in a paradigm (GL II.172.6, 185.13). See Belardi/Cipriano (1990: 117 ff. and 120 ff.) and Brandenburg (2020) for discussion. In fact, the use of positio meaning “full word” is traditional: Gk. θέμα is used in the same way by Apollonius Dyscolus (GG II/2.65.1, cf. Lallot 1997: II, 45, but also GG II/2.137.9) and Herodian (Erbse, ad Ξ 171 and Σ 76, cf. Wackernagel 1979 [1876]: 1438). 32 See Apollonius Dyscolus (αἱ δὲ προκείμεναι πρωτότυποι ἀντωνυμίαι θεματικώτερον ἐκλίθησαν, GG II/2.140.9–10, but see also GG II/2.139.2, 229.14), followed by Priscian (diversae sunt pronominis positiones in tertia persona, GL III.144.15). Similarly, Apollonius says that the παραγωγή in nouns is called εἶδος, and the noun acquires something similar to a εἶδος in the modification of an ending (ἡ μὲν ἐν τοῖς ὀνόμασι παραγωγὴ εἶδος καλεῖται. εἴδει γὰρ παραπλήσιόν τι παραδέχεται κατὰ τὴν ἑτεροίωσιν τοῦ τέλους, δι’ οὗ γνωρίζεται, GG II/1.18.14); and he adds that the παραγωγή may change the grammatical category of a word, since the noun ποιητής is derived from the verb ποιέω (GG II/2.259.1). 33 A similar rule is also found in Cledonius (GL V.37.7–16) and Pompeius (GL V.149.1; V.148.8– 18), according to whom, however, this type of noun is formed through the suffix -or, rather than -tor. 34 The grammarians often lamented the form-meaning mismatches in diminutives, which as a rule “increase” in their number of syllables, but “decrease” in the meaning. Similarly, a few nouns such as fabula, tabula etc., are said to be diminutive sono, though not sensu. On the contrast between formal and semantic derivation (κατὰ φωνήν and κατὰ σημασίαν), see Apollonius Dyscolus (GG II/3.39.31 ff.); the scholiasts to Dionysius Thrax (GG I/3.527.27 ff.); Diomedes (GL I.325.27 f.), Sergius (GL IV.429.22 ff.), Servius (GL IV.429.23 ff.), Pompeius (GL V.143.21 and 145.2 ff.), and Consentius (GL V.340.25 ff.). For a critique, see Biondi (2014: 144 and fn. 31, 2018) and Klink (1970: 26).

Priscian, the divisio graeca and the History of Word-formation  

the lexicon so as to avoid barbarisms, rather than to describe the rules that the speakers use to build new words (Vaahtera 1998: 92). In sum, the grammarians describe some of the empirical data whereby we exemplify the notion of word-formation at the present, but they use the notions of positio and terminatio only limitedly; moreover, they consider them either as a part of a word, or as ex-words, though not as significant units that differ from full words and exist as such in the mind of the speakers who manipulate them productively. In other terms, they do not ignore word-formation completely, but consider it a fairly peripheral topic with respect to the overall architecture of grammar, which is mainly concerned with the study of inflection. As a result, the description of derivativa nomina is completely absent from the simplest regulae, it is addressed only briefly in the classical Schulgrammatiken (e.g., Dionysius’ Tékhnē, Donatus’ Ars minor and Probus’ Instituta artium, where types of words are exemplified though not discussed at all), and it adds up to about 20 lines in the whole of Donatus’ Ars maior.35 Only Priscian seems to play a special attention to the topic.

. The peculiar position of the ars Prisciani Priscian describes the form of the words stored in the Latin lexicon, rather than their formation, as any other grammarian. Nonetheless, he focuses on the description of derivativa nomina more than all the others. As a confirmation, Priscian lists among his concerns that of explaining […] quomodo derivantur, ex quibus primitivis patronymic nouns, the terminationes of possessive nouns, ex qua ratione formantur, ex quibus positivis the comparatives and their extremitates, and quomodo nascuntur, ex qua primitivis denominative nouns (GL II.3.14 ff.). Moreover, all of book IV of his ars36 (de denominativis) is devoted to the study of derived words, and also in the smaller works, such as the Partitiones, he focuses on the internal structure of words more than usual. In the same vein, he exploits the notion of terminatio more than his forerunners, both to identify the inflectional classes of nouns and verbs in books VI and VII, and to describe the forms of derived nouns; he reviews the most common terminationes of possessive and diminutive nouns (GL II.69.20 ff. and II.102.5 ff.);37 he  35 On the classification of the various types of Latin grammars, see Law (1987: 191 ff.). 36 The traditional title Institutiones grammaticae comes from the frontispiece of Keil’s edition, but it is not original (De Nonno 1988: 279; 2009: 251). 37 The same concern is found in Dionysius Thrax, who lists the patronymic nouns in -δης, -ων and -αδιος (e.g., Ἀτρείδης, Ἀτρείων and ῾Υρράδιος), the comparatives in -τερος, -ων and -σσων

  Luca Alfieri discusses the word-formation rules of some types of derived nouns, such as the comparatives and the adverbs in ra, which change ra to er before adding ior (extra → exterior, GL II.90.7);38 he focuses on the meaning of derived nouns, such as the nouns in bundus vero desinentia, which similitudinem habere significant (GL II.137.16 ff., which seems a reply to Gell. Noct. 9.15.1–8); he discusses the mismatches between form and meaning in derivatives quite frequently and not only with reference to diminutive nouns (e.g., Euripides, which is a patronymic noun formally though not semantically, or the nouns in rius, which do not display a common meaning, since mensarium “tablecloth” means quod in mensa est, while frumentaria is a law concerning wheat, GL II.63.4 and 75.1);39 he often analyses the formal relation between derivatives and the bases from which they are built (e.g., the contrast between the long ā in stāre “to stay” and the short ă in its derivative stăbulum “stall”, and the variable length of the vowel before the termination nus, cf. Sullă but sullānus, GL II.124.9 and 77.9).40 Priscian’s theoretical beliefs on word-formation are different from ours. Like any other grammarian, he thinks that the input form for word-formation rules is the full word that is most similar to the output, disregarding stem vowels and morpho-phono-tactical adjustments. To him, therefore, the possessive nouns in cus such as Cypricus or the abstract nouns in tas such as bonitas are formed from the genitive singular, given the -i- before the suffix; and the iterative verbs such as lecturio from lego are built from the past participle or from the supine, so as to justify the -tu- (GL II.69.25, 127.26 and 429.10). On cases Priscian is still hindered by philosophical argumentations of a Stoic flavour. In the Partitiones, he asks himself if verbs pre-exist nouns or nouns pre-exist verbs, and answers that it depends on ipsa significatione et natura rerum. Therefore, he says that the verb armo “to arm” comes from arma “weap (e.g., ὀξύτερος, βελτίων, and κρείσσων) and the superlatives in -τατος and- τος (e.g., ὀξύτατος, ἄριστος), see GG I/1.25.3 ff. 38 Ab adverbiis vero derivata in ‘ra’ desinentibus mutant ‘ra’ in ‘er’, et accepta ‘ior’, faciunt comparativum, ut extra exterior… (GL II.90.7). 39 Inveniuntur quaedam nomina formae patronymicorum significatione denominativa, ut Euripides, non Euripi filius sed a Euripo sic nominatus (GL II.63.2); et, quod iam supra diximus, sciendum, quod, quamvis forma sit possessiva haec, quae in ‘rius’ desinit, tamen diversae significationis nomina in hac quoque inveniuntur (GL II.75.1). 40 In a desinentia antecedente aliqua consonante seu vocali producta eadem a et accepta ‘nus’ faciunt derivativum supra dictae formae […], ut Sulla Sullānus (GL II.77.9); alia enim in ‘bulum’, quae forma assimilis est deminutivis, et veniunt a nominibus vel a verbis, ut ‘cuna cunabulum’ […]. et attendendo quod paenultima in omnibus corripitur, antepaenultima vero si i habeat, corripitur: vestĭbŭlum […], sin vero a, producitur: cunābŭlum […]; excipitur stăbŭlum, quod corripit a (GL II.124.9 ff.).

Priscian, the divisio graeca and the History of Word-formation  

ons”, since nobody can get armed without weapons (GL III.462.35),41 but elsewhere he claims that the nouns cursus, cogitatio, cura and lector have been taken from the verbs currere, cogitare, curare and legere, since a run, a thought, a cure or a lector cannot exist without the act of running, thinking, taking care and reading (GL III.480.5 and 463.4).42 And on cases, he inclines towards “creative” derivations, such as fere “almost” from ferus “cruel” (GL III.71.6–8, following Varro, Ling. Lat. 8.92). However, Priscian’s attention to word-formation represents an exceptional case among the grammarians. One may, thus, wonder to what extent attention to the study of wordformation was a specific feature of Priscian or a hallmark of the divisio graeca in general. It is well known that Priscian did not invent the grammatical theory that he used. He rather adapted to Latin a body of doctrines that had been developed by Apollonius and were circulating especially among the Greek grammarians. The attention to word-formation, therefore, is probably a feature of the divisio graeca in general, although this feature is absent from Dionysius’ work for reasons of space. Still, besides his appreciation for Apollonius, maximus auctor artis grammaticae (GL II.548.6), and his son Herodian (GL II.1.3 ff), and besides the convenience of adopting a Greek model for teaching Latin in Byzantium, Priscian might have followed the divisio graeca for another reason.43 In late antiquity, interest in etymology increased to a great extent and wordformation had always been considered the most recent part of etymology: Priscian, who was the grammarian most attentive to word-formation, was also the

 41 Unde certum est, quando verbum a nomine et quando nomen a verbo nascitur? ex ipsa significatione et natura rerum. nam non possumus dicere armo nisi prius sint arma, quibus armemus aliquem (GL III.462.35). 42 Et quaeritur verbum ex re an res ex verbo nascitur? et potius antecedit verbi actus, in hoc quo sine res esse non potest. nec enim cursus intellegi potest sine currente, nec cogitatio sine cogitante: sic ergo nec cura sine curante (GL III.480.5); contra autem a verbo nascitur nomen, quod non potest esse in aliquo, nisi prius actus verbi in eo intellegatur, ut, si dicas lector, intellego eum prius legere et sic ei id nomen do (GL III.463.4). Derivations based on the ontological priority of the referents, rather than on the forms of the linguistic signs referring to them (that is, derivations sensu rather than litteratura, as Biondi 2018 put it), can be traced to Aristotle’s notion of παρώνυμον (see fn. 9 and 34) and are found also in Varro (nobiles from nobilitas, Ling. Lat. 8.15), Quintilian (velox from velocitas, Inst. Or. 1.6.38) and Isidore (a bonitate bonus et a malitia malus, but also homo ab humanitate and sapiens a sapientia, Or. 10.26.4 and 10.1). 43 Priscian was born in a Latin-speaking family in Mauretania; due to the Vandalic invasion, he moved to Byzantium, where he studied and worked as a professor of Latin in the university founded by Theodosius II. On Priscian’s life, see Ballaira (1984: 19, 29, 36). On his linguistic environment, see Fortes (2009). On his relation with the Greek models, see Lallot (2009) and Schmidhauser (2009).

  Luca Alfieri grammarian who employed the argumenta ab etymologia most frequently (see, e.g., coniugatio propter coniugatas consonantes, GL II.442.24).44 Priscian, therefore, might have adopted, and perhaps further elaborated on, the divisio graeca also so as to bring the science of grammar closer to the etymological interest typical of his times.

 Concluding remarks Two conclusions can be drawn from the discussion above. The first concerns Greek-Latin bilingualism. It is well known that the theory of grammar employed by Roman grammarians is of Greek origin, and the theory of word-formation is no exception. Still, it is striking that the best witness of the divisio graeca is Priscian, the most famous Roman grammarian, who taught Latin in a Greekspeaking environment and passed that divisio to the vernacular grammars of European languages as one of the hallmarks of the Latin, rather than the Greek, tradition. The second conclusion concerns the history of ancient grammar and the history of word-formation. A strong impression of uniformity comes from the reading of the ancient grammarians, and this is for a reason. The grammarians consider the word as the minimum and most significant unit of language. Moreover, they share the theory whereby words are described in grammars (the theory of the accidentia); and they are all convinced that the most fundamental properties of words are those that affect one single class of words and enable them to be divided into classes (the partes orationis). However, the way in which the grammarians classify the accidents that are common to the major wordclasses and, more specifically, the accidents that concern the species and the qualitas of words, among which derivativa nomina are described, may follow two different taxonomies.

 44 The success of etymology in Late Antiquity is on the one hand the result of the etymological practices of Hellenised Judeans such as Origen, Eusebius from Caesarea and Saint Jerome (Amsler 1989: 123 ff., 141 ff.), on the other a consequence of the isomorphism between nouns and things typical of Middle Age philosophy (Alfieri 2014: 154 ff.). On the etymologies of the grammarians, and specifically on Priscian’s etymologies, see Amsler (1989: 59 ff., 63 ff., 71 ff., 120 ff.) and Maltby (2009). On Priscian’s “Stoicism”, see Ebbesen (2009) and Luhtala (2009). Note that Priscian quotes 49 times Varro, the Roman etymologist par excellence, while Donatus quotes Varro 19, Charisius 18 and Probus only 3 (Collart 1945: 343–344).

Priscian, the divisio graeca and the History of Word-formation  

In the divisio latina, which by and large keeps away from philosophical influences and finds its champion in Donatus’ Ars maior, word-formation is relatively unimportant: various semantic and formal types of words are listed, and derived words represent only one of the many possible types of the qualitates of the nouns; in this case, the notions of terminatio and positio are exploited very limitedly and for the most part are used to clarify inflection. In the divisio graeca, which is more prone to philosophical (that is, mainly Stoic and NeoPlatonic) influences and finds its champion in Priscian’s Ars, the description of general accidents is grounded on the difference between simple, derived and compound words; in this divisio, the data on word-formation are described more abundantly, and the notions of positio and terminatio are more frequent. Probably, to ancient scholars the difference between the two divisiones appeared first as a “geographical” contrast (Greek vs. Latin), then as a philosophical contrast (stronger vs. weaker philosophical attitude) and finally as a practical contrast on some specific problems (the division between the common noun, the proper noun and the epithet; the treatment of homonymy and synonymy; the classification of “general” accidents, among which nomina derivativa are cast). The grammars that adopted the divisio graeca, in other words, appeared as more Greek-styled, more ancient and more comprehensive — that is, as grammars that included some of the philosophical cues which were at the very base of grammar and, thanks to such cues, could better describe the data on word-formation. However, from a retrospective point of view, this contrast has a fundamental side-effect. In antiquity, word-formation was the most recent part of etymology: therefore, the more a grammar described word-formation, the more it drew some diachronic-ontogenetic (that is, etymological) aspects into the synchronic description of language, blurring the synchrony/diachrony divide. The contrast between the divisio latina and the divisio graeca, in other words, can be reframed as a contrast between two ways of interpreting the contrast between the synchronic nature of grammar and the diachronic-ontogenetic nature of wordformation: the divisio latina was less informative, since it soft-pedalled the data on word-formation that were useful in practice, but more consistent in its positioning on the axis of time, since it excluded the etymological cues as much as possible; quite to the contrary, the divisio graeca was more informative, since it described the data on word-formation more generously, but less coherent in its positioning on the axis of time, since it merged synchronic and diachronicontogenetic data, at least in part. It is hard to see if Priscian perceived such a contrast mainly as a geographical, a philosophical and a practical one, rather than a contrast between the

  Luca Alfieri synchronic nature of grammar and the diachronic-ontogenetic nature of wordformation. What is certain is that the whole history of word-formation, from the Middle Ages onwards, is rooted in the last of these contrasts (Alfieri 2018; 2019). The grammaticae particulares from the Middle Ages onwards followed the divisio latina in the quantity of data supplied, but the divisio graeca in terms of the taxonomy used. The grammaticae speculativae treated the data on word-formation only limitedly, mainly focusing on the derivations sensu (e.g., a bonitate bonus, see fnn. 19 and 34) but disregarding those based on linguistic form. The general grammars written in France in the 17th and 18th centuries excluded data on wordformation totally, projecting them into the lexicon. As a result, data on wordformation had been progressively excluded from all the synchronic grammars, while they were taken up in the diachronic-ontogenetic grammars written in Germany in the 17th and 18th centuries, where they were viewed as empirical proof of the relation between the German language, which was particularly rich in word-formation, and the ultimate origin of language. Modern linguists are fully aware that the contrast between the diachronicontogenetic nature of word-formation and the synchronic nature of grammar is illusory: productive word-formation rules pertain to synchrony. The divisio graeca, therefore, does not really blur the synchrony-diachrony divide, not even from a purely retrospective standpoint; it blurs the divide from the perspective of the pre-modern scholars, who were convinced that word-formation was a diachronic-ontogenetic problem. To put it simply, when the German grammarians of the 17th and 18th centuries looked at Priscian’s grammar and its Greekstyled treatment of word-formation, they probably considered it an antecedent (or at least a partial antecedent) for their new type of grammar, which was philosophical, diachronic-ontogenetic and described the data on word-formation more abundantly than any other. Their view was not entirely correct from a strictly philological point of view, but there was a reason for it. The whole system of classification of words underling Greek-Latin grammar arose from the theory of the πάθη τῆς λέξεως, that is, a philosophical theory of a diachronic-ontogenetic nature. One of the main efforts of ancient grammarians was that of disentangling the philosophicaletymological aspects of this theory from its empirical application in the description of Greek and Latin. Their effort was well-directed for the most part, but word-formation was the part of grammar in which the disentanglement was less successful. As a result, the linguistic data on derived nouns and the technical terms used to describe them had always had a certain diachronic-ontogenetic flavor, which has always put them in contrast with the synchronic nature of grammar, at least until Saussure’s work. Ancient grammarians may or may not

Priscian, the divisio graeca and the History of Word-formation  

have been fully aware of this contrast, but the whole history of word-formation risks being misinterpreted, or at least seriously impoverished, if the contrast is not highlighted from its beginnings.

Bibliography Adams, J.N. (2003), Bilingualism and the Latin Language, Cambridge. Alfieri, L. (2014), “The birth of a grammatical category: the case of the adjective class”, Studi e Saggi Linguistici 52.1, 141–175. Alfieri, L. (2018), “La storia della derivatio, le grammatiche rationales e la confusione tra le moderne nozioni di sincronia e diacronia tra il ‘600 e il ‘700”, Atti del Sodalizio Glottologico Milanese 13, 7–22. Alfieri, L. (2019), “La storia della derivazione, la confusione tra sincronia e diacronia e le grammatiche “filosofiche” tra il XIII e il XVIII secolo”, Linguistica e Filologia 29, 63–106. Amsler, M.E. (1989), Etymology and Grammatical Discourse in Late Antiquity and the Middle Ages, Amsterdam. Auroux, S. (1994), La révolution technologique de la grammatisation: Introduction à l’histoire des science du langage, Liège. Ballaira, G. (1989), Prisciano e i suoi amici, Torino. Basset, L./Biville, F./Colombat, B./Swiggers, P./Wouters, A. (eds.) (2007), Bilinguisme et terminologie grammaticale gréco-latine, Louvain/Paris/Dudley. Belardi, W. (1975), Il linguaggio nella filosofia di Aristotele, Roma. Belardi, W. (1985), Filosofia, grammatica e retorica nel pensiero antico, Roma. Belardi, W./Cipriano, P. (1990), Casus interrogandi. Nigidio Figulo e la teoria stoica della lingua, Viterbo/Roma. Benedetti, M. (2018), “Conversione di suffissi nel metalinguaggio grammaticale: latino -tīvus e greco -ικός nell’opera di Dositeo”, Archivio Glottologico Italiano 103.2, 207–221. Biondi, L. (2014), “Litteratura e sensus: alcune considerazioni”, in: F. Bognini (ed.), Meminisse iuvat. Studi in memoria di Violetta de Angelis, Pisa, 135–162. Biondi, L. (2018), “Espressioni metaforiche e derivatio nei grammatici latini medievali: prime annotazioni”, Archivium Latinitatis Medii Aevi 76, 39–68. Bopp, F. (1816), Über das Conjugationssystem der Sanskritsprache in Vergleichung mit jenem der griechischen, lateinischen, persischen und germanischen Sprache, Frankfurt a.M. Brandenburg, Ph. (2020), “Case (πτῶσις), ancient theories of”, in: G.K. Giannakis et al. (eds.), Encyclopedia of Ancient Greek Language and Linguistics, vol. 1 A-F, Leiden, 268–271. Callanan, C.K. (1987), Die Sprachbeschreibung bei Aristophanes von Byzantium, Göttingen. Cassio, A.C. (1977), I Banchettanti, Pisa. Collart, J. (1954), Varron grammairien latin, Paris. De Nonno, M. (1988), “Prisciano (Priscianus)”, in: Enciclopedia Virgiliana, Roma, t. IV., 279–281. De Nonno, M. (2009), “Ars Prisciani Cesariensis: problemi di tipologia e di composizione”, in: M. Baratin/B. Colombat/L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire. De l’Antiquité aux Modernes, Lyon, 249–278. DELG = Chantraine, P. (1968–1980), Dictionnaire étymologique de la langue grecque. Histoire de mots, Paris.

  Luca Alfieri Desbordes, F. (1988), “La fonction du grec chez les grammairiens latins”, in: I. Rosier (ed.), L’héritage des grammairiens latin de l’antiquité aux Lumières. Actes du colloque de Chantilly (2–4 septembre 1987), Louvain, 15–26. [repr. Id. (2007), Idées grecques et latin sur le langage. Travaux d’histoire et d’épistémologie, Lyon, 107–119]. Di Benedetto, V. (1958–59), “Dioniso Trace e la techne a lui attribuita”, Annali della Scuola Normale Superiore di Pisa II.27, 169–210; II.28: 87–118. Di Benedetto, V. (2000), “Dionysius Thrax and the Tékhnē”, in: S. Auroux/E.F.K. Koerner et al. (eds.), History of Language Science. An International Handbook of the Evolution of the Study of Language from the Beginnings to the Present, Berlin/New York, vol. I, 394–400. Dickey, E. (2007), Ancient Greek Scholarship, Oxford. DK = Diels, H./Kranz, W. (1959–609 [19031]), Die Fragmente der Vorsokratiker, 3 vols., Berlin [It. trans. G. Reale (ed.), I presocratici. Testo greco a fronte, Milano]. Dover, K.J. (1968), Aristophanes Clouds, Oxford. Ebbesen, S. (2009), “Priscian and the philosophers”, in: M. Baratin/B. Colombat/L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire. De l’Antiquité aux Modernes, Lyon, 86–107. Ebbesen, S. (1981), Commentators and Commentaries on Aristotle’s Sophistici Elenchi, Vol. 1, Leiden. EDG = Beekes, R. (2010), Etymological Dictionary of Greek, Leiden/Boston. Erbse, H. (1969–1988, ed.), Scholia graeca in Homeri Iliadem (scholia vetera), VII vols., Berlin. Flobert, P. (1989), “La dérivation nominale chez les grammairiens romains”, Latomus 48, 741–752. Fögen, Th. (2008), “La formation des mots et l’enrichissement de la langue vue par quelques auteurs latins”, in: B. Kaltz (ed.), Regards croisés sur le mots non-simple, Lyon, 65–84. Forsgren, K.-Å./Kaltz, B. (2004) (eds.), Studien zur Geschichte der Wortbildungstheorien, Münster. Fortes, F. (2019), “Prisciano y el bilingüismo grecoromano: un ejemplo de análysis interlingüístico en las De constructione (Inst. gramm. XVII) de Prisciano”, Revista de Estudios Clássicos 47, 25–35. GEW = Frisk, H. (1960–1972), Griechisches etymologisches Wörterbuch, 2 vols., Heidelberg. GG = Schneider, R./Uhling, G./Hilgard A., (1878–1910), Grammatici graeci, IV vols., Lipsiae. GL = Heinrich, K. et al. (1961–1981 [1855–18801], eds.), Grammatici latini, VII vols., Hiledesheim/New York [repr. 2009–2010, Cambridge]. GRF M = Mazzarino, A. (1955), Grammaticae Romanae Fragmenta aetatis caesareae, Augustae Taurinorum. Householder, F.W. (1981), The Syntax of Apollonius Dyscolus, Amsterdam. Kaltz, B. (2004), “Zur Herausbildung der Wortbildungstheorie in der deutschen Grammatikographie”, Beiträge zur Geschichte der Sprachwissenschaft 14.1, 23–40. Kaltz, B. (2008) (ed.), Regards croisés sur les mots non-simple, Lyon. Kaltz, B./Leclercq, O. (2015), “Word-formation: from its beginnings to the 19th century”, in: P. Müller/I. Ohnheiser/S. Olsen/F. Rainer (eds.), Word-formation. An International Handbook of the Languages of Europe, Berlin/Boston, 22–37. Kastovsky, D. (2006), “Morphology as word-formation in the 20th century”, in: S. Auroux/E.F.K. Koerner et al. (eds.), History of Language Science. An International Handbook of the Evolution of the Study of Language from the Beginnings to the Present, Berlin/New York, Vol. II, 2324–2340.

Priscian, the divisio graeca and the History of Word-formation  

Kircher, Ch. (1999), “Le paramètre sémantique dans l’analyse des dérivés chez Priscien”, in: M. Baratin/C. Moussy (eds.), Conceptions latines du sens et de la signification, Paris, 155–166. Kircher, Ch. (1988), “La formation des noms en latin d’après Priscien Institutiones grammaticales livres II, III et IV”, Ktèma 13, 195–203. Klinck, R. (1970), Die lateinische Etymologie des Mittelalters, München. Lallot, J. (1989), La Grammaire de Denys le Thrace, Paris. Lallot, J. (1997), Apollonius Dyscole De la construction (syntaxe), 2 vols., Paris. Lallot, J. (2008), “De Platon aux grammairiens: regards grecs sur la structure de mots non simple”, in: B. Kaltz (ed.), Regards croisés sur le mots non-simple, Lyon, 51–63. Lallot, J. (2009), “Entre Apollonius et Planude: Priscien passeur”, in: M. Baratin/B. Colombat/ L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire. De l’Antiquité aux Modernes, Lyon, 153–166. Law, V. (1987), “The Latin grammarians in the early Middle Ages. A typological history”, in: D.J. Taylor (ed.), The History of Linguistics in the Classical Period, Amsterdam/Philadelphia, 191–206. Law, V./Sluiter, I. (eds.) (1995), Dionysus Thrax and the Techne Grammatike, Münster. Lindner, Th. (2015), “Word-formation in historical-comparative grammar”, in: P. Müller/ I. Ohnheiser/S. Olsen/F. Rainer (eds.), Word-formation. An International Handbook of the Languages of Europe, Berlin/Boston, 38–51. Lindner, Th. (2012), Indogermanische Grammatik. Band IV/1. Komposition, Lief. 2, Heidelberg. Luthala, A. (2005), Grammar and Philosophy in Late Antiquity, Amsterdam/Philadelphia. Luthala, A. (2009), “Priscian’s philosophy”, in: M. Baratin/B. Colombat/L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire. De l’Antiquité aux Modernes, Lyon, 109–124. Maltby, R. (2009), “Priscian’s etymologies: sources, function and theoretical basis «Graeci quibus in omnia doctrinae auctoribus utimur»”, in: M. Baratin/B. Colombat/L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire. De l’Antiquité aux Modernes, Lyon, 239–246. Matthaios, S. (1999), Untersuchungen zur Grammatik Aristarchs: Texte und Interpretation zur Wortartenlehre, Göttingen. Matthaios, S. (2004), “Die Wortbildungstheorie in der alexandrinischen Grammatik”, Beiträge zur Geschichte der Sprachwissenschaft 14.1, 5–22. Matthaios, S. (2008), “Théories des grammairiens alexandrins sur la formation des mots”, in: B. Kaltz (ed.), Regards croisés sur le mots non-simple, Lyon, 35–50. Melanchthon, Ph.S. (1558), Grammatica latina, Augusta Vindelicorum. Mullen, A./James, P. (eds.) (2012), Multilingualism in the Graeco-Roman Worlds, Cambridge. Nauck, A. (1848), Aristophanis Byzantii grammatici alexandrini fragmenta, Halis. Neil, R.A. (1909), The Knights of Aristophanes, Cambridge. Novokatko, A. (2020), “Figures (σχήματα): ancient theories of”, in: G.K. Giannakis et al. (eds.), Encyclopedia of Ancient Greek Language and Linguistics, vol. 1 A-F, Leiden, 589–593. Pagani, L. (2010), “La Tekhne Grammatike attribuita a Dioniso Trace e la nascita della grammatica nell’antichità greca”, Rivista di filologia e istruzione classica 138.3–4, 390–409. Pisani, V. (1934), “Una parola greca di probabile origine macedone: μάγειρος”, Revue internationale des études balkaniques 1, 255–259. Robins, R.H. (1970 [19671]), A Short History of Linguistics, London.

  Luca Alfieri Robins, R.H. (1995), “The authenticity of the Technē. The status quaestionis”, in: V. Law/I. Sluiter (eds.), Dionysus Thrax and the Techne Grammatike, Münster, 13–26. Rochette, B. (1997), Le latin dans le monde grec. Recherches sur la diffusion de la langue et des lettres latines dans le provinces hellénophones de l’Empire romaine, Bruxelles. Ronzitti, R. (2014), Due metafore del caso grammaticale: a.ind. víbhakti- e gr. ptōsis; preistoria e storia comparata, Innsbruck. Sanctius, F. (1587), Minerva seu de causis linguae latinae commentarius, cui inserta sunt, uncis inclusa, quae addidit Gasp. Schoppius, Salamanca. Scaliger, J.C. (1540), De causis linguae latinae in libri XIII, Lugduni. Schad, S. (2007), A Lexicon of Latin Grammatical Terminology, Pisa/Roma. Schironi, F. (2004), I frammenti di Aristarco di Samotracia negli etimologici bizantini: Etymologicum Genuinum, Magnum, Simeonis, Megalē Grammatikē, Zonarae Lexicon; introduzione, edizione critica e commento, Göttingen. Schmidhauser, A.U. (2009), “Le De pronomine de Priscien et son modèle grec”, in: M. Baratin/ B. Colombat/L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire. De l’Antiquité aux Modernes, Lyon, 167–180. Schwyzer, E. (1939), Griechische Grammatik, 2 vols., München. SGLG = Alpers, K./Erbse, H./Kleinlogel, A. (eds.), Sammlung griechischer und lateinischer Grammatiker, vol. 6: Slater, W.J. (1986), Aristophanis Byzantii fragmenta, Berlin/New York. Simplicius = Kalbfleisch, C. (1907), In Aristotelis Categorias commentarium, Commentaria in Aristotelem Graeca, vol. VIII, Berlin. Slater, W.J. (1989), review of Callanan (1987), Kratylos 34, 5–9. SVF = Von Arnim, H. (1905–1924), Stoicorum veterum fragmenta, 4 vols., Leipzig [It. trans. Radice, R. (1998), Stoici antichi. Tutti i frammenti secondo la raccolta di Hans Von Arnim, Milano]. Taylor, D.J. (1974), Declinatio. A Study of the Linguistic Theory of Marcus Terentius Varro, Amsterdam. Vaahtera, J. (1998), Derivation. Greek and Roman Views on Word Formation, Turku. De Velsen, A. (1853), Tryphonis grammatici Alexandrini fragmenta, Berolini. Wackernagel, J. (1876), De pathologiae veterum initiis, diss. Basel [repr. Id. (1979), Kleine Schriften, vol. 3, Göttingen, 1427–1486]. Wackernagel, J. (1909), “Attische Vorstufen des Itazismus”, Indogermanische Forschungen 25, 326–337. Zucker, A./Le Feuvre, C. (2021), Ancient and Medieval Greek Etymology. Theory and Practice, Berlin.

Carla Bruno

Dream Language and Dream Ideology: Echoes from the Memphis Serapeum Abstract: This paper explores the language of dream reports of Ptolemaios, katochos “recluse” in the Serapeum at Memphis in the 2nd cent. BCE, particularly focusing on the pattern regularly introducing the dream: it includes ἐνύπνιον (as the word for the dream), a predicate of visual perception, and οἴομαι, which marks the transition to the dream narration. As testified by a long-standing tradition of literary and documentary testimonies, the contrast between sight and inference is typical of the Ancient Greek dream mode, from which Ptolemaios accounts nonetheless diverge in the regular use of οἴομαι (instead of δοκέω) for the dreamer’s elaboration of the oneiric vision. Whether such usage may conceal an innovative attitude towards dreams among Memphis Serapeum katochoi will be discussed, respectively considering the lexical choices of the author as well as their morphosyntactic expression.

Whether a dream is a thought. Whether dreaming is thinking about something. L. Wittgenstein, Conversations on Freud, 1943

 Dreams and dream reports: language and cultural patterns from the writings of the katochos Ptolemaios It is through the words of dreamers that dreams become available — beyond the Self — to the Other. Dreams are in fact nothing but their narratives:1 their verbal recollections are the only empirical evidence of their symbolic imaginary, which is otherwise bound to remain “hidden” and “unknown” — Benveniste (1966) observes — outside the language framework. Language not only provides the primary access to the dream, but it shapes the dream itself according to its categories (Benveniste 1966: 72): from the millenial Ramesside dream manuals of Egypt to Freud’s Traumdeutung, words — in particular, puns and assonances —  1 Cf. Kirtsoglou (2010) for a discussion of the asymmetry between dream experience and its narrative from an anthropological perspective. https://doi.org/10.1515/9783110779684-002

  Carla Bruno are acknowledged as crucially revealing the dream’s inner meaning. The uniqueness of the dream experience contrasts sharply with the uniformity of the linguistic means through which every language shapes its dream narratives, which are “receptive” to the beliefs and experiences of the community in which the dreams occur (cf. Heijnen/Edgar 2010). “Dream patterns” reflect “culture patterns” (cf. Dodds 1951: 102) and, accordingly, dream patterns vary across languages and cultures. Languages that involve a system of evidential grammatical markers may possess a specific series of suffixes for evidence from dreams (cf. Kracke 2010), as, e.g., in Kwakiutl, the language of a North American Indian community where Boas (1910: 496) identified specific dream-discourse markers, which — according to Jakobson (1957) — are currently labelled as “revelative” evidentials. On the other hand, reporting a dream may require different evidential markers across languages (Aikhenvald 2003: 22–23; 2004: 343–344): Yukaghir people, in the Russian Far East, mark dreams as “indirect”, i.e., “non-eyewitness” (cf. Jochelson 1905: 400), experiences belonging to supernatural dimensions; the Jarawara people, an ethnic group in Brazil, encode dreams as “direct”, i.e., “visual” events.2 Languages without specific evidentiality markers also display linguistic regularities in the recounting of dreams: the features singled out by Kratschmer/Heijnen (2010) in the synchrony and diachrony of the languages of Europe exhibit significant parallels with the categorization of languages in which the source of information is regularly specified. Dreams are, again, at the crossroads between direct (i.e., visual) and indirect (i.e., inferential) experiences: their narratives — especially in the early stages of European languages — are for instance characterized by the vocabulary of visual perception alongside inferential expressions, often conveying epistemic reserve. Against the background of the unsettled debate on the criteria for distinguishing grammatical evidentiality from other expressions about the source of information available in all languages (cf. Aikhenvald 2004: 2), and the related discussion on the relationship between evidential and epistemic values, this paper focuses on some recurrent features in one of the most ancient collections of first-person dream accounts in Greek (Renberg 2007: 658), preserved among the private papyri of the Hellenistic Archive of the Katochoi of the Serapeum in Memphis.  2 Both attitudes are attested among the Shipibo-Conibo, an indigenous tribe of the Amazon rainforest, where dreams of common people are marked as indirect experiences, while shamans’ dreams are encoded as direct experiences, because of their privileged access to the supernatural realm (Aikhenvald 2003: 22).

Dream Language and Dream Ideology  

The author of these dream reports was Ptolemaios, the son of the Macedonian soldier Glaukias, ἐγκάτοχος “recluse”3 in the sanctuary for about twenty years (presumably from 172 until 152 BCE, cf. Thompson 1988: 234). The reports are included in the archive of his papers, which contains about a hundred documents, such as petitions, accounts, letters and texts related to the education of Apollonios, Ptolemaios’ younger brother,4 whose hand has been identified alongside the more uncertain handwriting of the elder brother.5 Through their writings, the two brothers outline a vivid picture of the troubled co-existence of Greeks and Egyptians in the Serapeum,6 where, under the syncretic cult of Sarapis, the Egyptian Osiris and Apis merge with the Hellenic universe. Among their papers, three papyrus sheets — UPZ I 77, 78, and 79 according to the edition of Wilcken (1927) — preserve about fifteen dream reports: two of these are included in the body of the letter that Ptolemaios addressed to a certain Damoxenos (cf. UPZ I 78), while the others appear to be personal notes. In particular, Ptolemaios does not note down only his own dreams, but also those of others: the dream of Thaues, one of the two Egyptian twins serving during the rites and ceremonies for the mourning of Apis, is reported in UPZ I 77, while UPZ I 79 contains the seven dreams of the Egyptian Nektembes, maybe a katochos.7 These reports are not the only documents from the Serapeum that testify to the relevance of the oneiric experience for its people. Their fascination for the  3 The causes behind reclusion in temples in Lagide Egypt remain uncertain: were the recluses mystics “held by the god” (cf., e.g., Wilcken 1927: 55–62) or trangressors detained for their crimes (cf., e.g., Rolandi 2005: 263)? The κατοχή has been on the other hand related also to asylum laws and practices (cf., e.g., Legras 2011) with an overview of the debate (Legras 2011: 14–21). 4 Apollonios spent a shorter period of time as a recluse in the temple over 158 BCE (cf. Thompson 1988: 199). Perhaps due to Ptolemaios’ repeated appeals to the royals, from the summer of 156 BCE, Apollonios was enlisted as chief of the police in service on the necropolis, where “he kept an eye on the changing population of the temple enclosures, noting those who took asylum in the Sarapieion and reporting back to his superior” (cf. Thompson 1988: 233). 5 Since Apollonios was entrusted by Ptolemaios with writing letters on his behalf, his handwriting is well documented within the archive (cf. also Vierros 2020). On the acquaintance of the two brothers with Greek writing, cf. Del Corso (2014), where the evolution in Apollonios’ scripts is taken as evidence that his basic education took place in the sanctuary. 6 During his detention, Ptolemaios reports in his letters attacks suffered by the temple personnel, which was mostly Egyptian, because of his Greek ethnicity (cf., e.g., ἐμέ τε παρὰ τὸ Ἕλληνα εἶναι in UPZ I 7, 4; 161 BCE or ἕνεκα τοῦ] Ἕλληνά με εἶν]αι in UPZ I 15, 16–17; 156 BCE). Cf. Thompson (1988: 213–214) with a reconstruction of the circumstances of these attacks, but also Lewis (1986, 84–87), who interprets UPZ I 7, 4 as “even though I am Greek”. 7 UPZ 80, maybe a list of dreams, which is generally associated to these papers (cf. Wilcken 1927: 369), is not taken into account here, since it does not include any dream narrative.

  Carla Bruno dream dimension is firstly attested by the pieces of literature selected by the two brothers: particularly relevant is the presence of the opening of the Dream of Nektanebos, a piece of Hellenistic Egyptian literature containing a prophecy on the demise of Nectanebo II, the last Egyptian ruler.8 The text was copied by Apollonios, and ends abruptly after the description of the pharaoh’s dream, as “this κάτοχος of the Serapeum at Memphis was only interested in the dream” (Koenen 1985: 193). Then, other dreamers inhabited the Memphis Serapeum during the 2nd cent. BCE: they recorded their dreams in Demotic, as documented not only by a group of ostraca and papyri, whose attribution to the Ptolemaios’ archive is still debated,9 but also by the accounts of Hor of Sebennytos, a minor Egyptian priest, who lived in the Serapeum complex during the same years as the two sons of Glaukias (cf. Ray 2002).10 To what extent such a varied and rich set of documents about dreams and dreamers has to do with the practice of incubation in the sanctuary is still under discussion.11 Similarly to incubatory dreams, Ptolemaios’ dreams are dated and  8 Among the literary excerpts selected by the two brothers there is also a Menander fragment with the reference to Asclepius and the practice of incubation (Legras 2011: 202; Renberg 2017: 117–118), and an astronomical treatise whose use may be connected to the interpretation of dreams (cf. Thompson 1988: 235). 9 These are the Bologna papyri 3171 and 3173, and the Petersburg ostracon 1129. Their attribution to the Katochoi Archive has been argued on the basis of the recurrence of anthroponyms such as “Apollonios” and “Thaues” in Pap. Dem. Bologna 3173 (the pertinence of ostracon is less clear, however). Accordingly, Ptolemaios’ younger brother is assumed to be the author of these demotic texts (cf., e.g., Thompson 1988: 230 or Bresciani 2005: 136), which are taken as evidence of his writing skills in Demotic. Due to the intense relationship with the Egyptian community of the Serapeum, the two brothers were probably bilingual (cf. Vierros 2014: 236), but there is no certain evidence of their ability to read or write Demotic (cf. Clarysse 2010: 65). Legras (2011) opts for the hypothesis that “Ptolémaios et Apollonios n’avaient qu’une diglossie orale” (Legras 2011: 279). Doubts on the pertinence of these demotic texts to the Katochoi Archive are also expressed by Wilcken (1927: 350–351); cf. also Bresciani et al. (1978), who exclude the possibility that the same hand was at work in the demotic papyri at issue. See also Kidd (2011) and Prada (2013), who debate on the status of Demotic within the oneiromantic practices of Ptolemaic Egypt. 10 It is presumably the translation (or the free adaptation) of an Egyptian original (cf. Koenen 1985). Whether Apollonios is also the author of the translation is still debated (cf. Del Corso 2014: 311). Cf. also Ryholt (2002), who discusses the possible influence of a Greek version of the novel on the Egyptian model of the text. 11 According to Renberg (2007: 659), “the evidence for incubation at Saqqâra is less abundant and more complex than previously recognized”. In particular, only the archive of Hor offers clear evidence of incubation at the Serapeum (cf. also Renberg 2017: 402, ft. 26), while the documents of the Ptolemaios’ archive are less decisive. Ptolemaios’ interest in dreams is assumed to be evidence for incubation by Weber (1998: 30) and Dunand (2006: 11).

Dream Language and Dream Ideology  

numbered, and the fact itself that each of the three sheets of the archive contains the dreams dreamt over the same month — i.e., Pachon in UPZ I 77, Tybi in UPZ I 78, Pharmouti in UPZ I 79 — has increased the esoteric aura of these accounts over the years.12 Nonetheless, Wilcken (1927: 349) excludes their incubatory source in view of their narratives: none of the dreams of Ptolemaios include either an invocation of the deity or the subsequent revelation of the god, which are typical of both incubatory iatromantic practices such as those documented by the Asclepian ἰάματα at the Epidaurus temple, and of oracular dreams, such as Nektanebos’ dream of the novel copied by Apollonios, where the prophetic dream follows the dreamer’s votive offerings.13 Whatever their source and their intent, the accounts of Ptolemaios offer a unique insight into dreams from the ancient world, “wie jeder Mensch sie jederzeit haben kann” (Wilcken 1927: 350). Indeed, Ptolemaios’ dreams represent timeless experiences: they are far from the “epiphany” dream pattern (Harris 2009: 23), which is recurrent in the Homeric world, where the supernatural reveals itself to the dreamer through a personified entity, such as the demon Oneiros in the Iliad, “gods in person or in the guise of the dreamer’s relatives or friends” (Del Corno 1982: 57). They rather correspond to the modern representation of dreams as “episodes”, i.e., sequences of scenes, where the dreamers themselves may be involved as “actors”, not just recipients of a message (cf. Harris 2009: 46). In particular, in Greek literary sources, episode dream representations became increasingly frequent between the 6th and 5th centuries BCE (cf. Del Corno 1982: 58), when the “primitive” concept of dreams as divine revelations opens to alternative views, which are the prelude to the concept of dreams as “psychic activity of the sleeper” (cf. Freud 1900: 2).14 In classical antiquity, episode dreams may also convey messages to the dreamer. But, unlike epiphany dreams, where the message is clearly uttered by the entity visiting the dreamer, episode dreams express their message through symbols and allegories, whose meaning is not directly accessible to the dreamer, so that interpretation is needed. All the dreams from the Serapeum were presumably transcribed for their interpretation: oneiromantic practices in the  12 Various hypotheses have arisen to explain such a circumstance, also involving the Egyptian zodiac. Cf. Wilcken (1927: 349 and references therein), who also suggests a link with Ptolemaios’ date of birth. 13 Ptolemaios’ accounts also include occasional invocations of deities, but they all occur in Ptolemaios’ dreams: there is hence no causal relationship between invocation and dream revelation (cf. Wilcken 1927: 349). 14 Cf. Del Corno (1982: 58), who relates the spread of the episode pattern to Eastern and Egyptian influences (as also suggested by the Herodotus reports).

  Carla Bruno Serapeum complex are, moreover, attested by the founding of a colorful signboard through which a Cretan dream interpreter advertises his business (cf. Prada 2013: 98). Ptolemaios himself occasionally comments on his dreams (cf. ἃ εἶδον … εὖ εἴη μοι “may that which I saw turn out well for me”, UPZ I 77, Col. II, 31) and the letter to Damoxenos may contain an implicit request for the advisory opinion of a friend on the dreams transcribed (cf. Legras 2011: 256). Ptolemaios then produces these texts under the urgent need to capture dream details holding hidden meanings before forgetting everything. In accordance with this practical purpose, and with the personal tenor of these annotations (Bentein 2015: 480), the accounts display an uncontrolled use of the language, which is characterized by a large amount of orthographic and morphosyntactic variation as well as a very poor syntax (cf. Bentein 2015; Vierros 2020).15 Wilcken (1927: 350) assumes that they were “Musterbeispiele der vulgären Sprache”, more open to vernacular strategies avoided in other — more formal — communicative situations.16 Not all dreams are represented with the same amount of information: those transcribed for Damoxenos are the most detailed, others are more concise, such as the quick note closing UPZ I 77, where after the date — Παχὼν κ̣γ “Pachon 24th” — Ptolemaios just adds the word γενεσί[ο]ις “birthday”. In spite of text length discrepancies and morphosyntactic variation, the choice of the same recursive elements to mark the dream content has a strong cohesive effect on these dream collections, such as the recurrence of the word ἐνύπνιον for the dream, its association with a predicate of visual perception (e.g., εἶδον in 1) and the introduction of the narrative by οἴομαι. All these elements may converge in the formulaic pattern shown in (1), taken from one of the dreams dreamt by Ptolemaios in Pachon, which is often encountered at the beginning of the narrative, marking the transition from one dream to the following one.

 15 On the other hand, since Ptolemaios is the main author of these accounts a lesser degree of linguistic accuracy is here expected: he is acknowledged to be less skilled in writing than his brother Apollonios, whose hand among the dreams is identified only in the letter to Damoxenos presumably dictated by Ptolemaios. Cf. Vierros (2020), for an overview of some idiolectal tendencies in the writings of the two katochoi brothers. 16 Cf. Bruno (forthcoming), where, among clausal complements, an increased incidence of ὅτι clauses, which are mostly associated with more popular levels of expression (cf. Wilcken 1927: 346), is encountered in Serapeum dream reports.

Dream Language and Dream Ideology  

(1)

UPZ I, 77, 18–21; 161–158 BCE τὸ ἐνύπνιον, ὃ εἶδον Παχὼν | κ. οἴομαι ἀρειθμεῖν με | λέγων ὅτι Θῶυθ (ἔτους) κ | ἥως κ. “The dream that I (Ptolemaios) saw on Pachon 20. I seem [lit. “I think”] to be counting (the days of the month) Thoth of the year 20 up to the 20th day” (transl. by Rowlandson 1998: 103).

The extent to which this pattern may conceal a coherent and peculiar dream ideology circulating among the Memphis Serapeum Katochoi will be discussed in Sections 2 and 3, respectively focusing on the lexical choices of the author (cf. Section 2) as well as their morphosyntactic expression (cf. Section 3).

 Dream narratives from the Serapeum: dream ideology through lexical choices Ptolemaios and Apollonios regularly employ the neuter ἐνύπνιον for the dreams recounted, which are accordingly presented as something “in their sleep” (ἐν ὕπνῳ) (cf. Casevitz 1982: 72). This sense is mostly available to the word after Homer, presumably following its use as attribute of the more common ὄνειρος (cf. Od. 14.495),17 which is on the other hand also attested within the archive, but just once, i.e., in the literary dream of Nektanebos (cf. UPZ I 81, 49). Conversely, ἐνύπνιον is the word currently in use for the dream in diverse circumstances: it also occurs in Apollonios’ letters to his brother (cf. ἐγὼ γὰρ ἐνύπνια | ὁρῶ πονηρά “I had bad dreams”, UPZ I 68, 5–6; πιστεύοντες | τὰ ἐνύπνια “trusting dreams”, UPZ I 70, 29–30) and the Cretan interpreter put it on the limestone stele advertising his activity (ἐνύπνια κρίνω “I interpret dreams”), cf. Thompson 1988: Plate vii). In Ptolemaios’ dream accounts, its use may therefore reflect “die lebendige Sprache des praktischen Lebens” (Wilcken 1927: 355), in which ὄνειρος may have sounded obsolete.18 As illustrated by the passages selected in (2)–(7), it is due to its regular occurrence alongside two sets of predicates that ἐνύπνιον — the generic word for  17 The early Greek word for a dream is the neuter ὄναρ, to which is associated the masculine ὄνειρος as its less common neuter variant ὄνειρον, cf. Casevitz (1982) for an overview of their distribution. On the gender variation of the variants, cf. Lazzeroni (2002). 18 According to Artemidorus’ Oneirocritica (2nd cent. CE), ἐνύπνιον and ὄνειρος define different types of dreams: the former applies to non-predictive dreams, and the latter to prophetic dreams. However, since Artemidorus himself occasionally ignored this distinction, Kessels (1969: 392) suggested that the difference was ignored by the language of the time, as the older data from the Serapeum also confirm.

  Carla Bruno dreams among the inhabitants of the Serapeum — takes its value: on the one hand, the set of visual perception predicates also available for ordinary sight (such as εἶδεν/εἶδον in 2–6) to which it relates as direct object (also via the anaphora of the relative pronoun in 2–5),19 on the other hand, the inferential predication expressed by the οἴομαι verb form, which introduces the dream content (cf. 2, 3, 4, and 7). Predicates of perception are particularly recurrent in the heading of the account, besides the date of the dream (cf. 2–5), while the cognitive inferential predicate marks the incipit of the narrative (cf. 2, 3, 4, and 7). These predicates are, for instance, both included in (2)-(4), from the Pachon sheet (UPZ I 77), but elsewhere, as in (5) and (6), taken from the dreams of Nektembes on Pharmouthi 5th, there is only the visual predicate and in (7), from the dreams sent to Damoxenos, οἴομαι is enough to identify what follows as a dream. Even the phrase ἐν τῷ ὕπνῳ “in sleep”, which elsewhere, e.g., in (2), refers to the subject’s elaboration of the dream dimension, is here lacking. (2)

UPZ I 77, 1–13; 161–158 BCE τὸ ἐνύπνειον, | ὃ εἶδεν Ταγῆς | δυδίμη Παχὼν | ιζ. οἴετο ἐν τῷ ὕ- | πνῳ καταβαίνου- | σα δειὰ τοῦ ἀμφόδου | ἀρειθμοῦσα οἰκείας | θ. εἰβουλόμην | ἐπειστρέψαι, ἔλεγον | ὅτι ταῦτα πάντα | τὰ πολλὰ ἐννήα εἰσί. | λεγουσει· ν⟦α̣ί⟧, λύει ἀπελ- | θεῖν. ὠψέ μοί ἐστειν. “The dream that Thaues the twin saw on Pachon 17. She seemed [lit. “she thought”] in the dream to be walking down the street, counting nine houses. I wanted to turn away and I said: All these are no more than nine (houses). They say: You are free to go. I said: It is too late for me” (transl. by Rowlandson 1998: 102).

(3) UPZ I 77, 14–20; 161–158 BCE τὸ ἐνύπνειον, ὃ εἶδεν Πτολε- | μαῖος Σεληνειήοις Παχὼν κε. | οἴομαι τὴν Ταγῆν εὔφονον | οὐσα καὶ ἡδυτέραι τῇ φωνῇ | καὶ εὖ δειακειμενη καὶ ὁρῶ | τὴν Ταοῦν γελωσα καὶ τὸν πό- | δα αὐτῆς μέγαν καθαρόν. “The dream that Ptolemaios saw at the festival of the Moon on Pachon 25. I seem [lit. “I think”] to see Thaues rather sweetly singing (?), and well disposed, and I see Taous laughing, and her foot is big and pure” (transl. by Rowlandson 1998: 103). (4) UPZ I 79, 13–18; 159 BCE Φαρμοῦθι κδ. τὸ ἐνύπνιον, ὃ εἶδεν Νεκ- | τομβῆς περὶ τῆς κατοχῆς τῶν διδυμῶν καὶ ἐμ αὐτοῦ, ὅτι ἀτυχῶ | καὶ τῆς οἰκίας. οἴετο ἄνθρωπον λέγιν μοι· φέρε τὸ δέρμα τοῦ ποδός σου καὶ | ἐγὼ δώσω σοι τὸ δέρμα τοῦ ποδός μου. λέγω· οὐ θέλω. τὸ δεύτερον·| ᾤμην οἰκίαν κα 19 Predicates for ordinary sight are also involved in ancient Egyptian dream representations, as testified by the dream manual on P. Chester Beatty III.13 (13th cent. BCE), cf. Szpakowska (2011) and Prada (2013), the latter with a comparison with Demotic and Hellenistic Greek evidence.

Dream Language and Dream Ideology  

θαίρεσθαι καλλύνοντος αὐτήν. τὸ τρίτον· ᾤμην | περιστερίδʼ ἡ δʼ ἐκπέφευγεν ἐκ τῆς χερός μου καὶ κατατρέχω αὐτὴν. “Pharmouti 24th. The dreams that Nektembes saw about the divine custody of the twins and myself, that I have had bad luck with my house. [The first dream.] A man seemed [lit. “he thought a man”] to be saying to me: Bring me the leather (sandal) of your foot and I will give you the leather of my foot. I say: I don’t want to. The second. It seemed to me [lit. “I thought”] someone was sleeping a house, sweeping it clean. The third. A dove seemed to me [lit. “I thought a dove”] fly out of my hand and I run after it” (transl. by Lewis 1986: 84). (5) UPZ I 79, 1–3, 159 BCE (ἔτους) κβ Φαρμοῦθι ε. | τὸ πρῶτον ἐνενύπνιον, ὃ εἶδεν Νεκθονβῆς πρεὶ τοον διδυμῶν καὶ ἐμ αὐτοῦ· | Ἀπολλώνιον εἶδον, προσπορεύεταί μοι. λέγι· χαῖρε, Νεκθεμβῆς, καλῶς. “Year 22 Pharmouti 5th. The first dream, that Nektembes saw involving the twins and myself. I saw Apollonios coming towards me. He says: A good greeting to you, Nektembes.” (transl. by Lewis 1986: 84). (6) UPZ I 79, 5–7, 159 BCE τὸ τρίτον· | εἰδον Πτολεμαῖον ἔχοντα μάχαιραν ἐν τῇ χερεὶ πορευμε- | νον διὰ τῆς ῥύμης. “The third. I saw Ptolemaios with a knife in his hand, walking in the street” (transl. by Lewis 1986: 84). (7) UPZ I 78, 1–8; 159 BCE Π[τ]ολεμ[αῖος] Δαμοξέ[νωι χαίρειν.] | (ἔτους) κβ Τῦβ[ι] ιβ εἰς [τ]ὴν ιγ ᾤμη[ν με ἐν Μέμφει] | βατίζειν με ἀπὸ λειβὸς ἕως ἀ[πηλ]ιώτου | καὶ ἀναπίπτομαι ἐπʼ ἄχυρον κ[αὶ ἄν]θρωπο[ς] | ἀπὸ λιβός μου ἐχόμενός μου [ἀν]απιπτι καὶ αὐτὸς καὶ ὥσπερ κεκλειμ[ένοι] μου | ἦσαν οἱ ὀφθαλμοί μου καὶ ἐξαί[φνης] ἀνύγωι | τοὺς ὀφθαλμούς μου. “Ptolemaios to Damoxenos, greeting. Year 23 in the night of Tybi 12th to 13th. I seemed [lit. “I thought”] to be walking in Memphis from west to east, and I come upon a pile of chaff and a man coming to me from the west also comes upon it and as for me — my eyes were sort of closed and suddenly I open my eyes” (transl. by Lewis 1986: 82).

Accordingly, the dreams — ἐνύπνια — are classified as direct visual experiences, whose objectivity is however challenged by the expression of epistemic reserve by the dreaming subject. In particular, from the combined use of the predicates for ordinary sight and οἴομαι, the dream appears as an “untrustworthy vision”.20 The formula introducing the dream accounts of Ptolemaios thus summarizes the

 20 Note that perception and inference predicates are also opposed according to the tenseaspectual stem, since οἴομαι is inflected only in imperfective stems. On the use of imperfective verb forms in dream accounts, cf. Kratschmer/Heijnen (2010).

  Carla Bruno two poles of the transition from the archaic conception of the dream as an external phenomenon experienced by the dreamers (shown by the persistence of the lexicon of visual perception) to an ideology of the dream as the result of psychic activity (conveyed by the inferential predicate).21 This ideological tension between an external and internal dream is a leitmotif in the Greek dream narrative. It is shown by the selection of passages in (8)– (11), where the same association between visual (i.e., direct) and inferential (i.e., indirect) evidence is encountered across different textual types: from Herodotus’ Histories (cf. 8) to Plato’s dialogues (cf. 9), from Hippocrates’ treaties (cf. 11) to the Epidaurian ἰάματα (cf. 12). However, unlike Ptolemaios’ accounts, here, beside the visual perception predicate, it is δοκέω, not οἴομαι, that invariably introduces the dream content. (8) Hdt. 6.131; 5th cent. BCE ἔγκυος ἐοῦσα εἶδε ὄψιν ἐν τῷ ὕπνῳ, ἐδόκεε δὲ λέοντα τεκεῖν, καὶ μετ᾽ ὀλίγας ἡμέρας τίκτει Περικλέα Ξανθίππῳ. “when she was pregnant, she saw in her sleep a vision in which she thought she gave birth to a lion. In a few days she bore Xanthippus a son, Pericles” (transl. by A.D. Godley). (9) Plato Cri. 44 a-b; 5th–4th cent. BCE ΣΩ. […] τεκμαίρομαι δὲ ἔκ τινος ἐνυπνίου ὃ ἑώρακα ὀλίγον πρότερον ταύτης τῆς νυκτός: καὶ κινδυνεύεις ἐν καιρῷ τινι οὐκ ἐγεῖραί με. ΚΡ. ἦν δὲ δὴ τί τὸ ἐνύπνιον; ΣΩ. ἐδόκει τίς μοι γυνὴ προσελθοῦσα καλὴ καὶ εὐειδής λευκὰ ἱμάτια ἔχουσα, καλέσαι με καὶ εἰπεῖν: ὦ Σώκρατες, ἤματί κεν τριτάτῳ Φθίην ἐρίβωλον ἵκοιο. “Socrates: […] And my reason for this is a dream which I had [lit. “I saw”] a little while ago in the course of this night. And perhaps you let me sleep just at the right time. Crito: What was the dream? Socrates: I dreamed that a beautiful, fair woman, clothed in white raiment [lit. “a … woman … seemed”], came to me and called me and said: Socrates, on the third day thou wouldst come to fertile Phthia” (transl. by H.N. Fowler). (10) Hippocr. Vict. 93; 5th cent. BCE Ἢν δὲ ἐν τῷ ὕπνῳ ἐσθίειν δοκέῃ ἢ πίνειν τῶν συνήθων σιτίων ἢ πομάτων, ἔνδειαν σημαίνει τροφῆς καὶ ψυχῆς ἐπιθυμίην.

 21 On the evolution of Greek oneiric imagery, and the decline of the more archaic pattern of the “visitation” of the deity, see Harris (2009). On the coexistence of the two patterns in Homer, where Penelope’s dream anticipates the subjective dream, cf. D’Agostino (2008). Regardless of the archaicity of the pattern, the prominence of the visitation type in Homer has presumably also to do with its role as a literary device, “indispensable […] pour mettre en branle la fable” (cf. Björck 1946: 311).

Dream Language and Dream Ideology  

“Whenever in his sleep a man thinks he is eating or drinking his usual food or drink, it indicates a want of nourishment and depression of the soul” (transl. by W.H.S. Jones). (11) IG IV2 1, 121, 10–12; ca. 350–300 BCE (II) Ἰθμονίκα Πελλανὶς ἀφίκετο εἰς τὸ ἱαρὸν ὑπὲρ γενεᾶς. ἐγ[κατα]- | [κοι]μαθεῖσα δὲ ὄψιν εἶδε· ἐδόκει αἰτεῖσθαι τὸν θεὸν κυῆσαι κό- | [ραν]. “Ithmonika of Pellene came to the sanctuary to have children. Sleeping in the shrine, she saw a vision. It seemed [lit. “she thought”] that she asked the god if she could conceive a daughter” (transl. by L.R. Lidonnici). IG IV2 1, 121, 120–122; ca. 350–300 BCE (XVIII) Ἀλκέτας Ἁλικός· οὗτος τυφλὸς ἐὼν ἐνύπνιον εἶδε· ἐδόκει οἱ ὁ θεὸς ποτελθὼν τοῖς δα- | κτύλοις διάγειν τὰ ὄμματα καὶ ἰδεῖν τὰ δένδρ̣η πρᾶτον τὰ ἐν τῶι ἱαρῶι. ἁμέρας δὲ γε- | νομένας ὑγιὴς ἐξῆλθε. “Alketas of Halieis. This man being blind, saw a dream. It seemed to him that the god came towards him and drew open his eyes with his fingers, and he first saw the trees in the sanctuary. When day came, he left well.” (transl. by L.R. Lidonnici).

What emerges is a Greek tradition which uniformly converges on the verb forms of δοκέω to express the dreamer’s cognitive processing of the oneiric vision. The regular contrast with the perception predication (cf. 8, 9, 11) presumably follows a conventional, formulaic pattern, as suggested by its recursive occurrence across the forty-three Epidaurian healing dreams.22 Due to its recurrence in dream representations, δοκέω is assumed by Orlandini/Poccetti (2015: 508) to be a marker of revelative evidentiality, which qualifies the dream as an inferential process. The inference is the “the semantic core” of the evidential values of δοκέω (cf. Van Rooy 2016: 37),23 which in particular “s’applique à une élaboration mentale ou intellectuelle de ce qui se présente à la vue” (Orlandini/Poccetti 2015: 508), so that it is particularly fitting to the context of oneiric perception. Nonetheless, the katochos Ptolemaios avoids it and opts for οἴομαι, which is, similarly to δοκέω, available for inferential reasoning.24 The fixed formulas ὡς δοκεῖ ἑμοί and ὡς οἴομαι offer an example of the  22 Cf. Lévy (1983: 151), who pointed out the singular concordance between the lexicon of tragedy and the Epidaurian ἰάματα, where the use of δοκέω is “quasi-formulaire”. 23 Cf. Van Rooy (2016: 37) for other evidential values of δοκέω, such as the “reportative” and “presumptive” values. 24 Cf. Orlandini/Poccetti (2015: 512), who consider that the οἴομαι middle forms express an “attenuation de la part du locuteur par rapport à son assertion” in contrast with the active ones. In papyri, active forms are extremely rare: middle inflection is apparently applied in most cases, so that a functional opposition between the two sets of endings is excluded.

  Carla Bruno alternation of two predicates in the expression of the speaker’s stance.25 Unsurprisingly, in the dream narratives, as elsewhere, these two predicates, both expressing the subjectivity of a statement, alternate. But, apart from the Serapeum writings, οἴομαι only occasionally occurs in dreams, as in (12), where Aristotle uses it alongside δοκέω for the dreamer’s impressions. (12) Aristot. Div. Somn. 463a; 4th cent. BCE ἐν δὲ τῷ καθεύδειν τοὐναντίον· καὶ γὰρ αἱ μικραὶ μεγάλαι δοκοῦσιν εἶναι. δῆλον δ’ ἐπὶ τῶν συμβαινόντων κατὰ τοὺς ὕπνους πολλάκις· οἴονται γὰρ κεραυνοῦσθαι καὶ βροντᾶσθαι μικρῶν ἤχων ἐν τοῖς ὠσὶ γινομένων. “In sleep the opposite takes place, for then even trifling movements seem considerable. This is plain in what often happens during sleep; for example, dreamers fancy that they are affected by thunder and lightning, when in fact there are only faint ringings in their ears” (transl. by J.I. Beare).

On the other hand, the Greeks of the Serapeum were not unaware of the conventional dream representation through δοκέω, which occurs in the incipit of the literary Dream of Nektanebos transcribed by Apollonios (cf. 13). (13) UPZ I 81, 2–7; 2nd cent. BCE Νεκτοναβὼς | τοῦ βασιλέως καταγινομένου ἐν Μέμφει καὶ θυσίαν | ποτὲ συντελεσαμένου καὶ ἀξιώσαντος τοὺς θεοὺς | δηλῶσαι ἀυτῶι τὰ ἐνεστηκότα ἔδοξεν | κατ᾽ ἐνύπνιον πλοῖον παπύρινον, ὅ καλεῖται | ἀγυπτιστεὶ ῥωψ, προσορμῆσαι εἰς Μέμφιν. “When Nektanebo was king in Memphis, one day after making a sacrifice and asking the gods to reveal to him the future, he dreamt that he was on a papyrus boat, which is called hrops in Egyptian, setting out to Memphis” (transl. by Kidd 2011: 120).

The variants for the elaboration of the dream — δοκέω and οἴομαι — therefore coexist within the Serapeum writings, but with a different distribution, since the former is limited to a formal, literary text, while the latter occurs in the more spontaneous personal notes of Ptolemaios, which feature an unconventional dream narrative,26 maybe indicative of a new, emerging dream ideology.  25 Here, as pointed out by Orlandini/Poccetti (2015: 511), the epistemic reserve on the statement also has a mitigating effect, which may trigger a politeness judgement. 26 As argued in Bruno (2020), the only contexts that show comparable uses of οἴομαι instead of δοκέω are Joseph’s dreams in the Septuagint Genesis. In view of the singular consonance with Ptolemaios’ dreams, the use, which excludes a Hebrew model, may reveal a local formulation of the dream narrative, which like other details in Joseph’s story contributes to giving an Egyptian (i.e., Greek-Egyptian) flavor to the novel. Cf., e.g., LXX Ge. 40, 16: Κἀγὼ εἶδον ἐνύπνιον καὶ ᾤμην τρία κανᾶ χονδριτῶν αἴρειν ἐπὶ τῆς κεϕαλῆς μου “I also saw a dream, and

Dream Language and Dream Ideology  

While δοκέω generally applies to inferences based on visual perception (i.e., “inductions” in Chafe’s 1986 terms),27 οἴομαι is also available for purely mental processes, based “on something other than evidence alone” (i.e., “beliefs” according to Chafe 1986). As a result, in Ptolemaios’ dream accounts, the use of οἴομαι instead of δοκέω may reflect people’s changing attitude to the dream experience, where visual perception becomes gradually less relevant and the subject’s psyche activity rises to prominence.

 Dream narratives from the Serapeum: dream ideology through morphosyntactic choices The use of οἴομαι instead of δοκέω does not involve only a lexical variation, but also a different syntax of the narrative. In particular, while οἴομαι imposes the dreamer as clausal subject, with δοκέω the dreamer may appear either — similarly to οἴομαι — as subject (as in 8, 10, and 11/II) or as indirect object, as in (9) and (11/XVIII), where it is associated to the dative case. In the latter cases, which involve a raising syntactic configuration, it is a thematic argument of the embedded infinitival clause that appears as δοκέω subject, like the woman (τίς … γυνή) calling Socrates (καλέσαι με) in his dream in (9) or Asclepius (ὁ θεός) opening the eyes (διάγειν τὰ ὄμματα) of the blind from the miracle inscription at Epidaurus in (11/XVIII). The dream content — whose actors are raised from the complement to the main clause — is then foregrounded, while the dreamers — excluded from the subject position — become less prominent.28 In fact, they may be omitted here, as in (12) above, where they are left unspecified in presence of δοκέω (cf. αἱ μικραὶ μεγάλαι δοκοῦσιν εἶναι), but not when the author, Aristoteles, switches to οἴομαι, which imposes the dreamer as subject (cf. οἴονται … κεραυνοῦσθαι). Apparently, Ptolemaios’ reports converge on the lexical variant — οἴομαι — that through its syntax defines the dream in relation to the dreaming subject.  methought I took up on my head three baskets of mealy food” (transl. by L.C.L. Brenton) or LXX Ge. 41, 17: ’Εν τῷ ὕπνῳ μου ᾤμην ἑστάναι παρὰ τὸ χεῖλος τοῦ ποταμοῦ “In my dream methought I stood by the bank of the river” (transl. by L.C.L. Brenton). 27 Cf. also Orlandini/Poccetti (2015), who find evidence of the different attitude towards the inferential process involved by δοκέω in the basic meaning of its root, i.e., *déḱ- “(an-, auf-) nehmen, wahrnehmen” (cf. LIV2: s.v.). 28 Cf. Benedetti/Gianollo (2020) for an analysis of the dative item in terms of “non-canonical subjecthood”.

  Carla Bruno Remarkably, in later texts dreamers tend with increasing frequency to occur as subjects also in δοκέω narratives: in the catalogue of dreams from the fifth book of Artemidorus’ Oneirocritica (2nd cent. CE), for instance, δοκέω associates almost exclusively to this syntactic pattern (cf. 14 below). (14) Artemid. Onir. 5.11; 2nd cent. CE Ἔδοξέ τις ἀπὸ τῆς σελήνης λύχνον ἅπτειν. τυφλὸς ἐγένετο. ὅθεν γὰρ οὐκ ἠδύνατο ἅψαι, ἐκεῖθεν ἐλάμβανε τὸ φῶς. ἄλλως τε καὶ τὴν σελήνην φασὶν οὐκ ἔχειν ἴδιον φῶς. “A certain man imagined that he lit a lamp using the moon. He became blind. For he received his light from a source by which one cannot light anything. And, differently put, they say that the moon does not have its own light.” (transl. by D.E. Harris-McCoy).

Through the extended use of the synchronic variant οἴομαι among katochoi, and the spread of δοκέω “nominative” dreamers in diachrony, dreamers then move towards dream narratives where they establish themselves as subjects of the dreaming process. The tendency to avoid dative dreamers in raising structures is on the other hand consistent with the treatment of infinitival complements in post-classical Greek, where the raising pattern decreases in frequency (cf. Bruno forthcoming on a corpus of Ptolemaic documentary papyri). In accordance with the developmental tendencies of the language, Ptolemaios then selects the lexical variant involving more stable syntactic patterns. Nevertheless, the use of οἴομαι in these accounts exhibits singular anomalies, when compared with other sections of the archive, where it regularly involves infinitives controlled by the matrix subjects or combined with its own (disjoint) accusative subject.29 Remarkably, in these dream reports, οἴομαι excludes control patterns, since its infinitival complement displays accusative subjects also when identical to its subject. As illustrated by (1) and (7) above and (15) and (16) below, this particularly occurs when Ptolemaios annotates his dreams, where redundant accusative pronominal items (i.e., the first singular person με) overtly expressing the subject of the infinitive appear.30

 29 On the diachronic persistence of the infinitive in control patterns, cf. Horrocks (2010: 93– 94), especially in the case of subject control as also shown by the remains of the infinitive in Modern Greek (cf. Joseph 1983: 69). Control is on the other hand by far the most common complementation strategy with οἴομαι among papyrus documentation, cf. also Mayser (1970 [1926]: 356), who refers to occasional participle complements “in vulgären Traumberichten”. 30 Cf. Bruno (forthcoming), where the pattern is taken as a vernacular strategy, through which the writer eases the tracking of the infinitival subject by avoiding cross-clausal reference.

Dream Language and Dream Ideology  

(15) UPZ I 77 r. 2, 21–27; 161–158 BCE (ἔτους) κγ Παχὼν δ. ᾤμην | ἐν τῷ ὕπνῳ ἐπεικαλεῖν με τὸν | μέγιστον Ἄμμωνα ἔρχεσθαι ἀ[πὸ] | βορρᾶ μου τρίτος ὤν, ἥως παραγ[ί]νηται. | ᾤμην με βοῦν ἐν τῷ τόπῳ εἶναι κα[ὶ]| οὐδεινουσα. “Year 24, Pachon 4. I (Ptolemaios) seemed [lit. “I thought”] in the dream to be calling upon the very great god Ammon, calling upon him to come to me from the north with two other (gods). I seemed [lit. “I thought” (to see) a cow in the compound and she is in labor” (transl. by Rowlandson 1998: 103). (16) UPZ I 78, 28–31; 159 BCE τῇ ιδ ᾤμην με ἐν Ἀλεξαν- | δρήᾳ με εἶναι ἐπάνω πύργου μεγάλου. εἶχον | πρόσοπον καλόν {εἶχον} καὶ οὐκ ἤθελον οὐθενεὶ δῖξαί μου τὸ πόρσωπον διὰ τὸ καλὸν αὐτὸν | εἶν[α]ι. “On the 14th I seemed [lit. “I thought”] [in a dream] to be on a big tower in Alexandria. I had a handsome face and I didn’t want to show my face to anyone because it was so beautiful” (transl. by Lewis 1986: 83).

In Ancient Greek, emphasis is one of the circumstances under which an embedded infinitive may display subjects that are coreferential with the matrix verb.31 Dreams are Ego-experiences and the redundancy of this syntactically marked construction actually increases the incidence of first-person singular pronouns. The pervasive presence of the dreamer is further amplified, on the one hand, by passages in which με is repeated more than once (cf. ᾤμη[ν με ἐν Μέμφει] | βατίζειν με in 7 and ᾤμην με ἐν Ἀλεξαν- | δρήᾳ με εἶναι in 16), and on the other hand, by the “förmlich pleonastisch” use of μου genitives (Mayser 1970 [1926]: 63), as in (7), taken from the dream shared with Damoxenos.32 The first person prevails even when Ptolemaios reports the dreams of others, i.e., Thaues and Nektembes, where the third person alternates with the first person, as in (2) (cf. οἴετο, i.e., Thaues, at line 3 and then εἰβουλόμην at line 8) and (4) (cf. οἴετο, i.e., Nektembes, at line 15 and then ᾤμην at line 17).33  31 Contrast is the other factor triggering infinitival coreferential subjects (cf., e.g., Chantraine 1953: 312 or Kühner/Gerth 1955, II, 30–31 among traditional grammars and Luraghi 1999 for discussion). Since Homer, clitic personal pronouns may also appear in these contexts, cf., e.g., Il. 20.361 οὔ μ᾽ ἔτι φημὶ μεθησέμεν οὐδ᾽ ἠβαιόν “in no wise, methinks, shall I be slack, nay, not a whit”. 32 According to Mayser (1970 [1926]: 63), such a lavish use of personal pronouns — often encountered in post-classical Greek where a Semitic model is at work (cf. Conybeare/Stock 1905]1995: 65) — is revealing of “ein charakteristischer Zug der Umgangssprache” more likely to emerge in the uncontrolled and spontaneous style of Ptolemaios. 33 Redundancy and repetition are typical of oral discourse, where “both the difficulty in planning […] and the characteristic lack of permanence of the oral channel […] triggers the repeat-

  Carla Bruno Apparently, like other singularities of Ptolemaios’ wording, these redundant accusative subjects are a function of the inherent subjectivity, to which not only dreams, but also their narratives are subordinated. However, after Wilcken (1927), scholars tend to link these pleonastic accusatives to οἴομαι rather than to the infinitive, particularly when they occur just after the verb (cf. ᾤμην με in UPZ I 77, 26; 78, 2; 78, 28). Quack (2008: 375), e.g., who takes it as an idiolectal feature of Ptolemaios’ writing, translates as “Ich meine mir”. Wilcken (1927)’s argument is mostly based on UPZ I 77, 26 (cf. 15), where it excludes that Ptolemaios may dream of being a cow (sic). He also excludes με as subject of the infinitive, since elsewhere (as, e.g., ἀρειθμεῖν με in 1) accusative subjects follow the infinitive, in accordance with later developmental tendencies leading to verb-subject word order (cf. Horrocks 2007: 173; Janse 1993). The argument is not decisive, since, crucially, in Ptolemaios’ accounts, με subjects do not necessarily follow the infinitive (cf. UPZ I 78, 28 με εἶναι ἐπάνω πύργου μεγάλου and 78, 36 με λέ[γειν]). However, Wilcken’s (1927) suggestion of linking με to οἴομαι opens up other interpretations: in particular, the pattern may involve an external — Egyptian — model. In Papyrus Chester Beatty 3 book of dreams (13th cent. BCE), the formula introducing the dream content (i.e., ἰr mꜣꜣ sw s m rsw.t “if a man sees himself in a dream”) similarly features a (dependent) personal pronoun (i.e., sw “him”) coreferential with the clausal subject (i.e., s “man”). Also Ptolemaios’ comment after the dream about the cow (εὖ εἴη μοι in UPZ I 77, Col. II 31) may possibly echo ancient Egyptian oneirocritical formulas: each of the dreams of Papyrus Chester Beatty 3 is marked as nfr “good” or ḏw “bad” (cf. Prada 2012: 169–70). Accordingly, οἴομαι με may represent an idiosyncrasy of Ptolemaios’ style shaped on an Egyptian dream pattern.34 The lexicon of ordinary sight prevails in these ancient Egyptian narratives, and a further set of anomalies encountered within the οἴομαι complements concerns their contamination with perception predicate patterns, since participle constructions here alternate with the expected infinitival complements, as in Thaues’ dream in (2) (cf. οἴετο … καταβαίνουσα … ἀρειθμοῦσα) and in the  ing of relevant information […])” (Bazzanella 2011: 246). For elements of orality in Ptolemaios’ accounts, cf. Wilcken (1927: 350), who comments: “an manchen Stellen glaubt man den Betreffenden zu hören”. 34 The Demotic dreams from the Petersberg Ostraca — co-eval with those of Ptolemaios — provide similar data, since in O. Pet. 1129, literally, “die Träumende ‘sieht sich’ in Memphis” (cf. Wilcken 1927: 351). The relationship between this Egyptian formula and the phrase in Ptolemaios is also acknowledged by Ray (1987: 87–88), who finds these Greek texts crucial for a better understanding of the Egyptian data.

Dream Language and Dream Ideology  

25th Pachon, Ptolemaios’ dream in (3), where οἴομαι is then coordinated to ὁράω which is associated to the expected participle pattern (οἴομαι τὴν Ταγῆν εὔφονον οὐσα … καὶ ὁρῶ τὴν Ταοῦν γελῶσα). Sometimes, infinitive and participle complements occur alongside in the same context as in (15) (cf. ᾤμην με βοῦν … εἶναι κα[ὶ] οὐδεινουσα) or in (17), where the matrix verb is lacking. (17) UPZ I 77, 21–25; 161–158 BCE κθ. δυω ἄνθρωποι ἐργά- | σζοντος ἐν τῇ προστάτι | καὶ Ταοῦν ἐπεὶ τῆς κλείμα- | κος καθῆσθαι καὶ προσπαι- | ζουσα αὐ[τ]οῖς. “Two men are working in the vestibule, and Taous is sitting on the steps and joking with them.” (transl. by Rowlandson 1998: 103).

Due the decline of the non-finite complementation system, similar extensions of the participial syntax into infinitival environments are not uncommon in postclassical Greek (cf., e.g., Bentein 2017). Nevertheless, outside these passages from Ptolemaios’ dream accounts, οἴομαι only occasionally associates to participles in papyri.35 The recursive use of this pattern within these documents may therefore not only attest a non-standard syntactic pattern avoided in more controlled writings, maybe here triggered by contamination by the recurrent visual predicates, but may also testify to a peculiar dream pattern emerging among the Serapeum katochoi. It is acknowledged that participle complements are opposed to infinitive clauses in terms of factivity (cf., e.g., Cristofaro 2008; 2012), since the former are typical of complements presupposing the truth of the event, while the latter are common in contexts where “the propositional content of the complement clause is not presented as positively true” (Cristofaro 2012). Participles then combine not only with sensory perception predicates, but also with knowledge predicates (e.g., γιγνώσκω), whereas infinitives express opinions and are expected with verba putandi, such as οἴομαι. Participles and infinitives differ according to the degree of speaker commitment, which is greater with participles and lesser with infinitives. Accordingly, in Ptolemaios’ dream narratives, the use of the participle instead of the infinitive may rest on a different attitude of the dreamer to the oneiric experience,  35 Cf., e.g., οὐκ αἶμαι μὲν σὲ | εἰδότα in P. Stras. 7 625, 4–5, in the fragment of the last part of a letter from the 2nd cent. BCE, where the participle replaces the infinitive in this variant of the most common disclosure formula οὐκ οἶμαι μέν σε ἀγνοεῖν (P. Mich. Zen. I 6, 1). Other possible evidence is a piece of correspondence from the Serapeum from Apollonios to the brother Sarapion (cf. σ̣ε̣ οἶμαι τετευχὼς | [τῆ]ς παρὰ σ]οῦ βοη[θί]ας, P. Mil. II 28, 11–12). However, the reading is not certain, since there is a gap in the text, cf. Calderini (1933) for a different conjecture.

  Carla Bruno which moves from the tradition of deceitful vision to the conception of dreams as a source of knowledge.

 Concluding remarks: dream ideology through lexicon and grammar Dreams are not their narratives, but through their narratives they become empirical facts and objects of interpretation. The language at use then provides a suggestive insight into the attitude of the dreamers to the oneiric experience. Some languages force dreamers to express their stance about dreams through the grammar of evidentiality; in others, devoid of similar grammatical means, dreamers do not nonetheless renounce expressing themselves and force the language to convey their impressions in words. Ptolemaios, a katochos in the Memphis Serapeum during the second half of the 2nd cent. BCE, reported in Greek about fifteen dreams experienced by himself and others from his circle of acquaintances. They are mostly personal notes: only two are addressed to a recipient, since they are included in the letter to a certain Damoxenos. All of the accounts display a very personal tenor and are characterized by uncontrolled and emotional writing, which embodies a large amount of orthographic and morphosyntactic variation. They are thus generally taken as an inventory of substandard linguistic features, pervaded by phenomena typical of the oral mode, such as repetition and redundancy. Ancient Greek did not possess an evidentiality system, but there are regularities in Ptolemaios’ unstructured and ungrammatical dream accounts that may signify a peculiar attitude to the dream experience. Accordingly, beside their personal tenor and spontaneous mode, the nature of the content (i.e., the “field” in Halliday’s 1978 terms) may be a variable on which the linguistic anomalies of the writings may depend. Ptolemaios’ dream reports are introduced by the same near-formulaic expression: it includes ἐνύπνιον as the word for the dream, a predicate of visual perception and οἴομαι, which projects the dream into the dreamers’ cognitive universe. The contrast between perception and inference corresponds to a dream pattern that is well-documented in the Greek world, from which Ptolemaios’ narratives nonetheless diverge in their regular use of οἴομαι instead of the forms of δοκέω for the dreamer’s elaboration of the oneiric vision.

Dream Language and Dream Ideology  

The use of οἴομαι then follows a drift where lexicon and syntax converge into an alternative view of dreams, whose “external”, i.e., visual, nature is challenged by the prominence of the dreamer as source of the dream. As regards its meaning, unlike δοκέω, the lexical root of οἴομαι does not imply a cognitive process triggered by a sensory perception, thus preluding an “inner” conception of the dream as the product of the dreamer’s imagination. This view is promoted by the οἴομαι syntax in which only the subject position is available for the dreamer, whereas δοκέω also allows dative dreamers in raising constructions that foreground the dream content. Moreover, the prevalence of nominative dreamers, which in diachrony becomes increasingly common in δοκέω dream narratives as well, correlates with a greater commitment of the dreamer to his statement. This trend may be confirmed also by the irregular syntactic behaviour of οἴομαι in these accounts, which alternates participle constructions with the expected infinitives. Besides Ptolemaios’ poor literacy and possible contamination by the syntax of the visual predicates, which occur in the same contexts, the feature may correspond to a change of attitude by the dreamer to dreams, since participles are expected not only with perception predicates, but also with knowledge predicates. In Ptolemaios’ dream accounts, the variant οἴομαι as well as its incoherent syntax provide the clues to an innovative attitude towards dreams among katochoi, who acknowledged the prophetic value of dreams, which reveal to the dreamers hidden truths about their future. What underlies Ptolemaios’ lexical and syntactic choices is conversely manifested by Apollonios in a letter addressed to his elder brother after his reclusion, from which the two excerpts in (18) are taken. (18) UPZ I 70, 2–13; 152 BCE ὀμνύ- | ο τὸν Σάραπιν, ἰ μὴ μικρόν | τι ἐντρέπομαι, οὐκ ἄν με | ἶδες τὸ πρ\ό/σωπόν μου | πόποτε, ὅτι ψεύδηι | πάντα καὶ οἱ παρὰ σὲ | θεοὶ ὁμοίως, ὅτι ἐν- | βέβληκαν ὑμᾶς εἰς ὕλην | μεγάλην καὶ οὗ δυνάμε- | θα ἀποθανεῖν καὶ κα\ν/ ἴδῃς | ὅτι μέλλομεν σωθῆναι, | τότε βαπτιζώμεθα. “I swear by Sarapis that if I had a little compunction, you would never have seen my face again; for you utter nothing but lies and your gods likewise, for they have plunged us into a deep mire in which may die, and when you have a vision that we are to be rescued, then we sink outright.” (transl. by A.S. Hunt/C.C. Edgar). (19) UPZ I 70, 23–30; 152 BCE οὐκ ἔστι ἀνακύ- | ψα\ί με/ πόποτε ἐν τῇ Τρικομίαι | ὑπὸ τῆς αἰσχύνης, ἰ καὶ | αὑτοὺς δεδώκαμεν | καὶ ἀποπεπτώκαμεν | πλανόμενοι ὑπὸ τῶν | θεῶν καὶ πιστεύοντες | τὰ ἐνύπνια.

  Carla Bruno “Never again can I hold my head in Tricomia for shame we have given ourselves away and been deluded, misled by gods and trusting in dreams.” (transl. by A.S. Hunt/C.C. Edgar).

Apollonios’ tone is harsh: the letter conveys his disillusionment with dreams and their predictions. In the verso, alongside the name of the recipient, he sarcastically adds the words πρὸς τοὺς τὴν ἀλήθειαν λέγοντες “to those who tell the truth”; in the body of the letter, he then deplores his brother’s lies (ὅτι ψεύδηι | πάντα, ll. 6–7) and the gods’ deceits (καὶ ἀποπεπτώκαμεν | πλανόμενοι ὑπὸ τῶν | θεῶν, ll. 27–28). He complains about the illusionary dreams of his brother (καὶ κα\ν/ ἴδῃς | ὅτι μέλλομεν σωθῆναι, | τότε βαπτιζώμεθα, ll. 11–13) and regrets having trusted in them (καὶ πιστεύοντες | τὰ ἐνύπνια, ll. 29–30). While he distances himself from his brother’s stance, Apollonios confirms what Ptolemaios himself suggested through the lexicon and syntax, i.e., the truth-value (cf. ἀλήθειαν) attributed to his dreams and the consequent trust in them (cf. πιστεύοντες). For Apollonios, the refusal of the values to which the elder brother introduced him at the Serapeum implies then a return to the traditional theme of the dream as δόξα, an untrustworthy vision.

Bibliography Aikhenvald, A.Y. (2003), “Evidentiality in typological perspective”, in: A.Y. Aikhenvald/ R.M.W. Dixon (eds.), Studies in Evidentiality, Amsterdam/Philadelphia, 1–31. Aikhenvald, A.Y. (2004), Evidentiality, Oxford. Bazzanella, C. (2011), “Redundancy, repetition, and intensity in discourse”, Language Sciences 33, 243–254. Benedetti, M./Gianollo, C. (2020), “Criteria for subjecthood and non-canonical subjects in Classical Greek”, in: B. Drinka (ed.), Historical Linguistics 2017. Selected papers from the 23rd International Conference on Historical Linguistics, San Antonio, Texas, 31 July – 4 August 2017, Amsterdam/Philadelphia, 29–48. Bentein, K. (2015), “The Greek documentary papyri as a linguistically heterogeneous corpus: the case of the Katochoi of the Sarapeion archive”, Classical Word 108, 461–484. Bentein, K. (2017), “Finite vs. νon-finite complementation in post-classical and early byzantine Greek. Towards a pragmatic restructuring of the complementation system?”, Journal of Greek Linguistics 17, 1–34. Benveniste, É. (1966), “Remarques sur la function du language dans la découverte freudienne”, Problèmes de linguistique générale 1, Paris, 75–87. Björck, G. (1946), “ONAP IΔEIN. De la perception de la réve chez les anciens”, Eranos 44, 306–314. Boas, F. (1910), Kwakiutl. An Illustrative Sketch, Washington. Bresciani, E. (2005), La porta dei sogni. Interpreti e sognatori nell’Egitto antico, Torino. Bresciani, E. et al. (1978), “Una rilettura dei Pap. Dem. Bologna 3173 e 3171”, Egitto e Vicino Oriente 1, 95–104.

Dream Language and Dream Ideology  

Bruno, C. (2020), “Sospesi tra due mondi. Modulazioni del racconto onirico nei papiri del recluso Tolomeo”, Atene e Roma 14, 103–116. Bruno, C. (forthcoming), “Infinitives at work. Competing patterns in early Ptolemaic papyri letters”, in: S. Dahlgren et al. (eds.), Act of the Scribe: Interfaces Between Scribal Work and Language Use, Cambridge. Calderini, A. (1933), “Un nuovo papiro del Serapeo di Memfi nella raccolta Milanese”, Aegyptus 13, 674–689. Casevitz, M. (1982), “Les mots du rêve en grec ancient”, Ktema 7, 67–73. Chafe, W.L. (1986), “Evidentiality in English conversation and academic writing”, in: W.L. Chafe/ J. Nichols (eds.), Evidentiality: The Linguistic Coding of Epistemology, Nordwood, New Jersey, 261–272. Chantraine, P. (1953), Grammaire homérique, Tome II: Syntaxe, Paris. Clarysse, W. (2010), “Bilingual papyrological archives”, in: A. Papaconstantinou (ed.), The Multilingual Experience in Egypt, from the Ptolemies to the Abbasids, London/New York, 42–72. Conybeare, F.C./Stock, G.W.J. ([1905] 1995), Grammar of Septuagint Greek: With Selected Readings, Vocabularies, and Updated Indexes, Peabody, MA. Cristofaro, S. (2008), “A constructionist approach to complementation: Evidence from Ancient Greek”, Linguistics 46, 571–606. Cristofaro, S. (2012), “Clausal complementation in Ancient Greek: the role of form-function reanalysis in the development of argument encoding patterns”, in: H. Diessela/V. Gast (eds.), Clause Linkage in Cross-Linguistic Perspective, Berlin/New York, 335–362. D’Agostino, E. (2008), “Ho visto un sogno: Io assente ed es-terno in Omero”, Quaderns d’Italià 13, 11–28. Del Corno, D. (1982), “Dreams and their interpretation in Ancient Greece”, Bulletin of the Institute of Classical Studies 29, 55–62. Del Corso, L. (2014), “I figli di Glaucia e i papiri del Serapeo. Tra produzione scritta e identità etnica”, in: D. Bianconi (ed.), Storia della scrittura e altre storie, Roma, 285–336. Dodds, E.R. (1951), The Greeks and the Irrational, Los Angeles/Berkeley. Dunand, F. (2006), “La guérison dans les temples (Égypte, époque tardive)”, Archiv für Religionsgeschichte 8, 4–24. Freud, S. (1900), Die Traumdeutung, Leipzig/Wien. Halliday, M.A.K. (1978), Language as a Social Semiotic: The Social Interpretation of Language and Meaning, London. Harris, W.V. (2009), Dreams and Experience in Classical Antiquity, Cambridge, MA/London. Heijnen, A./Edgar, I. (2010), “Imprints of dreaming”, History and Anthropology 21, 217–226. Horrocks, G. (2010), Greek: A History of the Language and its Speakers, Oxford. Jakobson, R. (1957), Shifters, Verbal Categories, and the Russian Verb, Harvard. Janse, M. (1993), “La position des pronoms personnels enclitiques en grec néo-testamentaire à la lumière des dialectes néo-helléniques”, in: C. Brixhe (ed.), La Koiné Grecque Antique, Nancy, I, 83–121. Jochelson, W. (1905), “Essay on the grammar of the Yukaghir language”, American Anthropologist Supplement 7, 369–424. Joseph, B.D. (1983), The Synchrony and Diachrony of the Balkan Infinitive: A Study in Areal, General, and Historical Linguistics, Cambridge. Kessels, A.H.M. (1969), “Ancient systems of dream-classification”, Mnemosyne 22, 389–424.

  Carla Bruno Kidd, S. (2011), “Dreams in bilingual papyri from the Ptolemaic Period”, The Bulletin of the American Society of Papyrologists 48, 113–130. Koenen, L. (1985), “The dream of Nektanebos”, The Bulletin of the American Society of Papyrologists 22, 171–194. Kracke, W.H. (2010), “Dream as deceit, dream as truth: the grammar of telling dreams”, Anthropological Linguistics 51, 64–77. Kratschmer, A./Heijnen, A. (2010), “Revelative evidentiality in European languages: linguistic coding and its anthropological background”, in: G. Diewald/E. Smirnova (eds.), The Linguistic Realization of Evidentiality in European Languages, Berlin, 331–368. Kühner, R./Gehrt, B. (1904), Ausführliche Grammatik der griechischen Sprache, vol. II.2: Satzlehre, Hannover/Leipzig. Lazzeroni, R. (2002), “Il nome greco del sogno e il neutro indoeuropeo”, Archivio glottologico italiano 87/2, 145–162. Legras, B. (2011), Les reclus grecs du Sarapieion de Memphis. Une enquête sur l’Hellénisme égyptien, Leuven/Paris/Walpole MA. Lévy, E. (1983), “Le théâtre et le rêve: le rêve dans le théâtre d’Eschile”, in: Théâtre et spectacles dans l’antiquité. Actes du Colloque de Strasbourg, 5–7 novembre 1981, Leiden, 141–168. Lewis, N. (1986), Greeks in Ptolemaic Egypt, Oakville CT. LIV 2 = Kümmel, M./Rix, H. (eds.) (2001), Lexikon der indogermanischen Verben: die Wurzeln und ihre Primärstammbildungen, Wiesbaden. Luraghi, S. (1999), “The subject of complement clauses with the infinitive”, in: B. Jacquinod (ed.), Les complétives en grec ancient, Saint-Etienne, 199–213. Mayser, E. ([1926]1970), Grammatik der griechischen Papyri aus der Ptolemäerzeit, Band 2: Satzlehre, Teil 1: Analytischer Teil (Erste Hälfte), Berlin. Orlandini, A./Poccetti, P. (2015), “Specie — re uera: deux mondes en parallèle”, in: G.V.M. Haverling, (ed.), Latin Linguistics in the Early 21st Century. Acts of the 16th International Colloquium on Latin Linguistics (Uppsala, June 6th–11th, 2011), Uppsala, 502–516. Prada, L. (2012), “Classifying dreams, classifying the world: Ancient Egyptian oneiromancy and demotic dream books”, in: H. Abd El Gawad et al. (eds.), Current Research in Egyptology 2011, Proceedings of the twelfth annual symposium (Durham University, March 2011), Oxford/Oakville. Prada, L. (2013), “Dreams, bilingualism, and oneiromancy in Ptolemaic Egypt: remarks on a recent study”, Zeitschrift für Papyrologie und Epigraphik 184, 85–101. Quack, J.F. (2008), “Demotische magische und divinatorische Texte,” in: B. Janowski/ G. Wilhelm (eds.), Omina, Orakel, Rituale und Beschwörungen, TUAT, n.s. 4, München, 331–385. Ray, J.D. (1987), “Phrases used in dream texts”, in: S.P. Vleeming, Aspects of Demotic Lexicography, Acts of the Second International Conference for Demotic Studies (Leiden, 19–21 September 1984), Leuven, 85–93. Ray, J.D. (2002), Reflections of Osiris: Lives from Ancient Egypt, New York. Renberg, G.H. (2007), Incubation at Saqqâra, in: Proceedings of the Twenty-Fifth International Congress of Papyrology, Ann Arbor, 649–662. Renberg, G.H. (2017), Where Dreams May Come: Incubation Sanctuaries in the Greco-Roman World, Leiden/ Boston. Rowlandson, J. (1998), Women and Society in Greek and Roman Egypt. A Sourcebook, Cambridge.

Dream Language and Dream Ideology  

Ryholt, K. (2002), “Nectanebo’s dream or the prophecy of Petesis”, in: A. Blasius (ed.), Apokalyptik und Ägypten: Eine kritische Analyse der relevanten Texte aus dem griechischrömischen Ägypten, Leuven/Paris/Sterling VA, 221–241. Szpakowska, K. (2011), “Dream Interpretation in the Ramesside Age”, in: M. Collier/S. Snape (eds.), Ramesside Studies in Honour of K. A. Kitchen, Bolton, 509–517. Thompson, D.J. (1988), Memphis under the Ptolemies, Princeton/Oxford. UPZ = Wilcken, U. (1927), Urkunden der Ptolemäerzeit. (Ältere Funde), Erster Band: Papyri aus Unterägypten, Berlin/Leipzig. Van Rooy, R. (2016), “The relevance of evidentiality for Ancient Greek: some explorative steps through Plato”, Journal of Greek Linguistics 16, 3–46. Vierros, M. (2013), “Bilingualism in Hellenistic Egypt”, in: G. Giannakis (ed.), Encyclopedia of Ancient Greek Language and Linguistics, 234–238. Vierros, M. (2020), “Idiolect in focus: two brothers in the Memphis Sarapieion (II BCE)”, in: K. Bentein/M. Janse (eds.), Varietes of Post-classical and Byzantine Greek, Berlin/Boston, 39–74. Weber, G. (1998), “Traum und Alltag in hellenistischer Zeit”, Zeitschrift für Religions- und Geistesgeschichte 50, 22–39. Wilcken, U. (1927), Urkunden der Ptolemäerzeit (Ältere Funde), Erster Band: Papyri aus Unterägypten, Berlin/Leipzig.

Albio Cesare Cassio

A Tale of Coins and Suffixes: Syracusan Greek ἑξᾶς, Latin sextāns, and Congeners Abstract: Ἑξᾶς, -ᾶντος, τετρᾶς, -ᾶντος, and other fractions of the λίτρα are amply attested in ancient Sicily in literary texts and on coins. This study provides an explanation for the rare -ᾶς, -ᾶντος suffix, and shows that the old doctrine, still accepted in recent publications, that it was formed after Latin -āns, -antis, is untenable; the reverse is certainly the case. In the Roman world -ᾶς, -ᾶντος was assimilated to such adjectives as comans stellans that inflected like participles, which explains why τριᾶς became triens.

 The litra and its fractions This is a brief story of ancient weights, coins, their values and their names. It will touch upon the intricate contacts between the western Greeks and the ancient Italic world, and the complicated linguistic mechanisms at work in the transition from the indigenous Italic languages to Greek, and from Greek specifically to Latin. Λίτρα was the Greek rendering of the name of an indigenous weight unit (bronze pound) divided into twelve ὀγκίαι (ounces, on which see Weiss in this volume). The duodecimal system is Italic in origin (Heinrich Chantraine 1962; Parise 1989): it is attested in Sicily, Etruria, Umbria and Latium (Roman lībra divided into twelve unciae, Parise 1989: 586 ff.). The voiceless dental in Greek (λίτρα) it is commonly explained by admitting that the word *līdhra reached Syracuse through a language of ancient Italy, probably Sikel, in which the aspirated voiced dental *-dh- became aspirated voiceless *-th- at a first stage, then unaspirated -t- (Leumann 1977: 171; Lejeune 1993: 11). Syracuse adopted Attic denominations (1 drachm equal to 6 obols), but around 465 BCE (and probably earlier) a new silver denomination, the λίτρα, corresponding in value to the indigenous bronze pound, was added to the old ones, causing concurrent circulation of obols and litrai (Fischer-Bossert 2012: 144 ff.). The names of pounds and ounces are attested at an early date both in Epicharmus (10.2 K.-A. πεντόγκιον; 9.3 λίτραν; late 6th – early 5th cent. BCE) and in a Sicilian inscription (Dubois 1989: nr. 20 δέκα λίτρας ἀποτεισάτω from Megara Hyblaea, early 6th cent. BCE). In a fragment of Sophocles (1065 Radt, unfortunately without context), we read λιτροσκόπος, glossed with ἀργυραμοιβός https://doi.org/10.1515/9783110779684-003

  Albio Cesare Cassio “money exchanger”, which might imply circulation of Sicilian coins outside Sicily. At Syracuse the two systems, (a) 1 drachm = 5 litrai; 1 litra =12 ounces, and (b) 1 drachm = 6 obols; 1 obol = 10 ounces, were made to interact by creating the ἑξᾶς, gen. -ᾶντος “6th part of a litra”, that is two ounces, a fractional denomination (implied by Epicharmus 10.2 ἑξάντιον) which allowed easy conversion from litrai to obols (6 ἑξᾶντες = 1 litra, 5 ἑξᾶντες = 1 obol); later Latin sextāns, -antis (1/6 of an as equals 2 unciae). Other fractions of the litra in -ᾶς, -ᾶντος are attested, along with corresponding formations in Latin: τετρᾶς, -ᾶντος “fourth part of a λίτρα” = 3 ὀγκίαι (lat. quadrāns, -antis, 1/4 of an as = 3 unciae), and τριᾶς, -ᾶντος “third part of a litra” = 4 ὀγκίαι: lat. triēns, -entis). Note the different ending in Latin, triēns, not *triāns, to which I’ll return shortly. Accent and vowel length of -ᾶς, -ᾶντος are guaranteed by ancient grammarians, [Herodian.] GG III/1.56.21f. τὸ δὲ Μητρᾶς Ἡρᾶς τετρᾶς βορρᾶς περισπᾶται, τὸ δὲ τετράς τετράδος συστέλλεται. These names, ἑξᾶντες, τετρᾶντες etc. are attested, often in abbreviated form, on coins of Eryx, Himera, Croton and elsewhere in the Western Greek world (H. Chantraine 1962: 52–53).

 Meaning and function of the -ᾶς, -ᾶντος suffix The interpretation of -ᾶς in the names of these fractions has been marred by a long-standing and serious misunderstanding. According to Schwyzer 1939: 599 and Leumann 1977: 583 -ᾶς, -ᾶντος was formed after Latin -āns, -antis, an opinion still subscribed to in recent years by Robert Beekes (2010: 433 s.v. ἕξ) but this is impossible because the beginnings of coinage in Rome go back to virtually 300 BCE, while the Greek forms are attested in Sophron and Epicharmus in the mid 5th cent. BCE and possibly earlier (H. Chantraine 1962: 54 and fn. 14). Also not much thought has been given to an important detail, namely that Latin triēns, -entis would have yelded in Greek something like *τρίεις, *-εντος, not τριᾶς, -ᾶντος; and if Latin sextans (from sext-[-a pars]) had been the starting point, from ordinal ἑκτός Greek would have created *ἑκτᾶς, not ἑξᾶς. It is clear that a derivation from Latin is excluded for more than one reason, and -ᾶς, -ᾶντος should be explained in a different way. Greek has two problematic declension types, (1) -ᾶς, -ᾶντος and (2) -ᾶς, -ᾶ.

A Tale of Coins and Suffixes  

(1) -ᾶς, -ᾶντος was more of a “conglomeration” than a suffix, the outcome of *-ᾱ-ϝεντς, *-ᾱ-ϝεντος, and as such found in Doric literary texts, e.g., Bacchylid. 13.45 ποταμοῦ... δινᾶντος1 “whirling”, < * δῑνᾱ́-ϝεντ- and Pind. Ol. 2.85 βέλη... φωνᾶντα2 “sounding” < *φωνᾱ́-ϝεντα. However, very rare cases of an -ᾶς, -ᾶντος declension are found in Attic, too, e.g., πελεκᾶς, -ᾶντος “woodpecker”. (2) -ᾶς, -ᾶ was mainly used for “des mots abrégés, de caractère populaire” (Chantraine 1933: 31), e.g., μαχαιρᾶς, gen. -ᾶ = μαχαιροποιός “knifesmith”, λαχανᾶς gen. -ᾶ = λαχανοπώλης “greengrocer”), as well as for shortened forms of personal names: Ἀλεξᾶς (< Ἀλέξανδρος) gen. Ἀλεξᾶ, Ἐπαφρᾶς (< Ἐπαφρόδιτος) gen. Ἐπαφρᾶ (rarely found in literary texts: however see the long list in Threatte 1996: 85). In rare cases we find ᾶς, -ᾶντος along -ᾶς, -ᾶ (Μαρικᾶς has an accusative τὸν Μαρικᾶν in Sch. Aristoph. nub. 552 but an -ᾶντα accusative in Eup. fr. 203 K.-A. Μαρικᾶντα μὴ κολάζειν). Heinrich Chantraine 1962: 57 was dissatisfied with the explanation of Syracusan -ᾶς, -ᾶντος as a Greek suffix, and hypothesized “eine sikulische Endung”; more recently Andreas Willi argued for a special Italic lengthening of an original *-ant.3 Obviously nothing certain can be said of Sikel suffixes, so we are left with (1) Greek -ᾶς, -ᾶντος and (2) Latin -āns, -antis. (1) -ᾶς, -ᾶντος usually means “provided with” (as in δινᾶντος etc. just quoted); more rarely it simply means “related to” (like, e.g., πελεκᾶς, -ᾶντος “woodpecker” from the bird who cuts wood like an axe, πέλεκυς). (2) -āns, -antis is especially interesting since it appears not only as the participle of -āre verbs, but is also applied to denominal “possessiv-Adjective” (Leumann 1977: 583), e.g., comāns “with long hair”, stellāns “starry”, inflecting exactly like participles of -āre verbs. Given the possessive meaning of the -āns adjectives Leumann (1977: 583) regarded quadrāns = “a quarter of” as rätselhaft:4 he obviously expected it to mean something like “containing a square” or “square in shape”. But is quadrāns really as enigmatic as Leumann thought?  1 Attested in this form in the Bacchylides Papyrus (British Library Papyrus 733), cf. Hom. Od. 11.242 ποταμοῦ...δινήεντος. 2 T. Mommsen restored φωνάεντα with -αε- in synizesis, but the Bacchylides papyrus shows that the contraction is likely to be ancient. 3 2008: 35 f. “Da in den entsprechenden lateinischen Wörtern nun ein genau gleich lautendes Suffix -ant- auftritt, liegt die Annahme einer Suffixdehnung von -ᾶς unter italischem Einfluss nahe”. 4 The meaning of this quadrāns has nothing to do with its homophonous quadrāns the participle of quadrare “to fit in with”, still used in Italian, i conti non quadrano, meaning “the accounts do not balance”.

  Albio Cesare Cassio For a start, we should not forget that some Greek formations based on numbers could be remarkably ambiguous. For instance, τετράς, -άδος is a group of four units in Hippocr. Prognost. 20, but can also mean one unit out of four, as in the Thessalian τετράδες, each representing a μέρος of Thessaly as a whole;5 and τετράρχης means “king of a fourth part”. Πεμπάς means “the fifth part” in the famous Molpoi inscription (Miletus, 5th cent. BCE, Herda 2006: 10, lines 36, 39), but “a group of five” in Xenophon Cyrop. 2. 1, with πεμπάδαρχος “commander of five men”. But when it comes to money, oscillations are out of place, and we should not forget that both in form and in meaning the -ᾶς, -ᾶντος coins are an artificial creation of the Syracusan Greeks. In practice, some numbers were “blocked” in a fractional value and provided with one specific suffix to indicate one specific value. Formally, it may have happened what Anna Morpurgo Davies suggested to me many years ago (ap. Cassio 2002: 68 fn. 56), namely that in Syracusan Doric -ᾶς, -ᾶντος from *-ᾱ-ϝεντς, *-ᾱ-ϝεντος was at some stage resegmented as a suffix in its own right and used in a way similar to -ᾶς, -ᾶ to create Kurzformen of compounds meaning “consisting of x parts”, e.g., τετρᾶς from τετραμερής “quadripartite”, ἑξᾶς from ἑξαμερής “divided into six parts” etcetera. At the start they may have been masculine adjectives of an implicit νόμος meaning “coin”, very frequent in Sicily (Kassel/Austin 2001: 250 on Sophron fr. 161), but they soon became autonomous nouns.

 Roman borrowings of -ᾶς, -ᾶντος and their meanings Contrary to what we have been told in recent and less recent bibliography, Roman quadrāns (one quarter of an as), sextāns (one sixth of an as) etc. depend on the Greek forms and not the other way around. Leumann deemed quadrāns (one quarter of an as) “rätselhaft” because he did not regard it as dependent on Greek τετρᾶς (one quarter of a λίτρα), but its model; yet if we start from the fractional Greek form, the Latin meaning is easily explicable. The Roman forms were obvious accomodated to the Latin vocabulary and morphology: sext-āns was not built on the number sex (as it was in Greek ἑξ-ᾶς) but on sext-(-a pars), a logical reinterpretation; in the same vein quadrāns was built on quadr- of quad 5 Hellanic. FGrHist 4 F 52 τεττάρων μερῶν ὄντων τῆς Θετταλίας, ἕκαστον μέρος τετρὰς ἐκαλεῖτο, καθά φησιν Ἑλλάνικος ἐν τοῖς Θετταλικοῖς.

A Tale of Coins and Suffixes  

ratus (in the meaning “divided into four parts”) or quadripertitus. This is a very special chapter in the history of Latin word-formation, and one full of contradictions. Quadrāns and its congeners were in many ways cognate with comāns, stellāns, gemmāns, described by Leumann (1977: 583), as we have seen, as “denominative Adjectiva auf -ant, formal Partizipia”; they were special formations based on nouns (there is no verb *stellāre or comāre), yet some of them were completely homophonous with real verbal participles, but in this case the meaning was quite different: the verb gemmare is used of plants putting out buds, while gemmāns means “adorned with precious stones” (Ov. Met. 3.264 f. gemmantia ... sceptra). Besides, in Latin comans, stellans and congeners were adjectives, while quadrans and the like were nouns, and masculine ones at that (uno quadrante Hor. Sat. 2.3.43), simply because their models were masculine in Greek; they inflected like Latin participles. And one of these forms is very special indeed. Nobody seems to have cared about triēns, although it does deserve some attention. To my mind triēns can be explained only by admitting that -ᾶς ᾶντος was perceived by the Romans as cognate with their participles in -āns antis from, say, portāre or cantāre (portāns, cantāns probably already pronounced portās, cantās Leumann 1977: 112 f.; Sydney Allen 1978: 65 ff.); τριᾶς was the only Greek form where -ᾶς was preceded by [i] and the Latin version became triēns on the analogy of the participles faciēns, veniēns, dormiēns.

 Conclusions This story is a good illustration of the complications inherent to multiculturalism and multilingualism. The Greek colonists originally used the monetary system of their native land, drachmes and obols, but were soon confronted with an Italic system of weights, and later coins, completely different and using local names, λίτρα and ὀγκία. In order to ensure easy conversion from litrai to obols, new fractional denominations were created ending in -ᾶς, -ᾶντος, probably originally a Doric “conglomerate” which at some point became an autonomous suffix used exclusively for fractions. Ἑξᾶς, τετρᾶς, and τριᾶς were borrowed by the Romans but underwent various formal modifications. Greek -ᾶς, -ᾶντος became -āns, -antis on the model of the “special” possessive adjectives of the comāns and stellāns type; at the same time Latin emphasized the fractional value of the stem by turning original ἑξᾶς into sextāns. On the other hand, comāns and stellāns inflected like participles, and the “participial interpretation” of sextāns and quadrāns was so strong that τριᾶς was dragged into the

  Albio Cesare Cassio Latin system to such an extent that it became triēns, formally the participle of a fictional verb *tríĕre. Granted that the well-known patrii sermonis egestas was real, and not invented by a poet, the Romans knew all too well how to remedy its shortcomings.

Bibliography Beekes, R. (2010), Etymological Dictionary of Greek, Leiden. Cassio, A.C. (2002), “The Language of Doric Comedy”, in: A. Willi (ed.), The Language of Greek Comedy, Oxford, 51–83. Chantraine, P. (1933), La formation des noms en grec ancien, Paris. Chantraine, H. (1962), “Bemerkungen zum ältesten sizilischen und römischen Münzwesen”, Jahrbücher für Numismatik und Geldgeschichte 12, 51–64. Dubois, L. (1989), Inscriptions grecques dialectales de Sicile, Rome. Fischer-Bossert, W. (2012), “The coinage of Sicily”, in: W.E. Metcalf, The Oxford Handbook of Greek and Roman Coinage, Oxford, 142–156. Herda, A. (2006), “Der Apollon-Delphinios Kult in Milet und die Neujahrsprozession nach Didyma”, Milesische Forschungen, Band 4, Mainz. Kassel, R./Austin, C. (2001), Poetae comici graeci 1. 1. Comoedia Dorica, mimi, phlyaces, Berolini et Novi Eboraci. Lejeune, M. (1993), “Le nom de mesure λίτρα: essai lexical”, REG 106, 1–11. Leumann, M. (1977), Lateinische Laut- und Formenlehre, München. Parise, N.F. (1989), “Libbre e bronzo librale nell’Italia antica”, in: G. Pugliese Carratelli (ed.), Italia omnium terrarum parens, Milano, 581–599. Schwyzer, E. (1939), Griechische Grammatik I, München. Sydney Allen, W. (1978), Vox Latina, Cambridge. Threatte, L. (1996), The Grammar of Attic Inscriptions. 2. Morphology, Berlin/New York. Willi, A. (2008), Sikelismos. Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.-5. Jh. v. Chr.), Basel.

Emilio Crespo

Dialects in Contact in the Ancient Kingdom of Macedon Abstract: The purpose of this chapter is to identify the linguistic variations found in documents from the central part of Lower Macedonia and to offer a provisional historical interpretation thereof. The inscriptions show that at least a variety of Northwest Doric and another of the Attic-Ionic type coexisted in contact since the beginning of the 4th cent. BCE. In addition, there are hybrid inscriptions featuring diverse mixtures of both dialectal groups. Such inscriptions in mixed dialect bear witness to linguistic contact among speakers of different dialects, to dialectal accommodation on the part of at least one of the participants in the exchange, and to subsequent levelling and dialectal convergence. Once a dialectal mixture became first routine and then permanent through a process of koineization, a new variety emerged which was eventually to evolve into the Attic-Ionic koine.

 Introduction Many of the most prominent developments that we have seen in the past few decades in the study of Greek antiquity are from the geographical area of the ancient Kingdom of Macedon. Well-known milestones include: the finding and subsequent editio princeps of the Derveni papyrus;1 the excavations of the royal tombs in the great tumulus of Vergina and other discoveries in the same archaeological area;2 the more recent excavations at Pella, which brought to light the famous curse tablet (κατάδεσμος)3 and the cist-grave “of the philosophers”  Previous versions of this chapter were read at the Aristotle University of Thessaloniki, at the Seventh International Colloquium on Ancient Greek Dialects held at Basel and Lausanne, at the Complutense University of Madrid and at the XV FIEC/CA Congress (London). Thanks are due to the audiences for their enlightening comments. I am also grateful to Paloma Guijarro Ruano (Complutense University) for her comments on a previous draft and to Albio Cesare Cassio for his sharp queries on my penultimate version.  1 Tsantsanoglou-Parassoglou-Kouremenos (2006); see also Kotwick (2017). 2 Andronikos (1984; 1994); Roisman/Worthington (2010); Lane Fox (2011). 3 Voutiras (1998). https://doi.org/10.1515/9783110779684-004

  Emilio Crespo featuring painted decorations, named after the men who adorn its walls, one of whom is depicted with a papyrus roll;4 the new findings in the so-called “tomb of Judgement” at ancient Mieza near Naousa, which depicts a deceased warrior who, accompanied by Hermes psychopompós, appears before the judges of the underworld, Aeacus and Rhadamanthys, as in Plato’s Gorgias (523a–524a); the recent findings in the well at Methone, namely ceramic pots and shards — some of them bearing inscriptions — which were brought to this northern site as early as around 700 BCE;5 and more recently the great tomb at Amphipolis, among other discoveries (see Hatzopoulos 2011a; 2018; 2020). These important findings in the ancient Kingdom of Macedon — and more generally in northern Greece if we include the recent editio princeps of 4.216 oracular epigraphs (χρηστήρια ἐλάσματα) from the Dodona sanctuary6 — have greatly expanded our knowledge but have also prompted new questions. In the field of dialect studies, the two-volume edition of Επιγραφές Κάτω Μακεδονίας Β’ (EKM 2) by Gounaropoulou/Paschidis/Hatzopoulos (2015) consisting of 632 texts, 108 of which have been published for the first time, the new digital supplements to the inscriptions of Upper Macedon (Martín González/ Paschidis 2020a) and of Beroia (Martín González/Paschidis 2020b) and the two supplements (IG X, 2 1 suppl. 1–2) to the inscriptions from Thessaloniki, edited by Nigdelis (2017) and by Papakonstantinou-Diamantourou/Martín González/ Hallof (2021), respectively, along with the previously published editions of the inscriptions from Beroia (EKM 1) and from the sanctuary of Leukopetra (I.Leukopetra), represent the entire body of published Greek inscriptions found in the ancient regions of Bottia and Almopia, which lie to the north of Pieria and between the Mt. Bermion to the west and the Axios River to the east, and Upper Macedon.7 These publications offer new prospects in the study of the ancient Greek dialects spoken in the area where Aigeai and Pella, the successive capitals of the ancient Macedonian kingdom, were located. Unfortunately, most of the documents are of Hellenistic or Roman date and merely consist of personal names inscribed in slabs of stone set on graves or in other hard surfaces. There are not long classical (let alone archaic) texts that would enable a full description of the Macedonian dialect of Greek. These new publications, however, provide us with fresh evidence of local and perhaps chronological and social varieties.

 4 Lilibaki-Akamati (2007). 5 Besios/Tziphopoulos/Kotsonas (2012); Strauss Clay/Malkin/Tzifopoulos (2017). 6 Dakaris/Vokotopoulou/Christidis (2013). 7 For the studies in the ΜΕΛΕΤΗΜΑΤΑ series, see Hatzopoulos (2011a: 40–41; 2020: 73–74).

Dialects in Contact in the Ancient Kingdom of Macedon  

This chapter aims to identify the linguistic variations found in documents from this part of Lower Macedonia and to offer a provisional historical interpretation thereof. As we will see, the inscriptions show that at least a variety of Doric dialect and another of the Attic-Ionic type coexisted in contact since the beginning of the 4th cent. BCE. In addition, there are hybrid inscriptions featuring diverse mixtures of both dialectal groups. Such inscriptions in mixed dialect bear witness to linguistic contact among speakers of different dialects, to dialectal accommodation on the part of at least one of the participants in the exchange, and to subsequent levelling and dialectal convergence. Once a dialectal mixture became first routine and then permanent through a process of koineization, a new dialect emerged which at the end was eventually to evolve into the Attic-Ionic koine. We are fully aware that several difficulties undermine the certainty of any conclusions that can ultimately be drawn, most notably: – Unlike the Athenian state from the mid-5th cent. BCE onwards, the Temenid kingdom lacked the “epigraphic habit” of publishing public and private documents carved in stone until the reign of Alexander III. This explains the scarcity and brevity of inscriptions that have come down to us. – Documents from before the mid-4th cent. BCE are extremely short and fragmentary. We thus cannot ascertain the dialectal subgroup in which they were written, but only whether they were composed in an undifferentiated dialectal variety common to Attic, Ionic and koine or common to Doric and to the Northwest Doric dialect group. – Since most inscriptions merely consist of personal names inscribed on graves, the dialect they were written in is either difficult to identify or is — in many cases — controversial, given that personal names travel with their bearers and are subject to trends that do not necessarily coincide with the borders between dialects.

 Differences between Aigeai and Pella In spite of the scanty evidence, there is a crucial difference between the inscriptions from Aigeai and those from Pella as regards the most prominent feature of the Macedonian dialect of Greek: the occasional use of the graphemes instead of or of , respectively, in inscriptions found in Mace-

  Emilio Crespo donia and in the neighboring areas of Thessaly.8 While no instances of such graphic alternatives occur in the 149 inscriptions from Pella,9 there are eleven sure or highly likely examples — all of them in personal names or in epithets of gods — in the eighty inscriptions from Aigeai. As seen in the lists of examples given in (1) and (2), three of the instances cited in (1) are dated to the 4th cent. BCE and the other eight appear in texts written in koine and dated to the period of the Roman Empire: (1) instead of , respectively, in inscriptions from Aigeai: – Ἀμένανδρος Κέββα (EKM 2. 16, Aigeai, 400–375 BCE), short for Κέβαλος instead of Κέφαλος. – Βερεννὼ Φιλίστου (EKM 2. 22, Aigeai; ca. 350 BCE), a hypocoristic personal name formed with the first element Φερε- seen in Φερενίκα (< *bher-) and the second element “victory” seen in Βερενίκη (see Masson 1984). – Ἐπικράτης Στρώφακος Στρωφακίδου Βερνίκα (EKM 2. 27, Aigeai, ca. 300 BCE) corresponding, as Βερεννώ, to Φερενίκα (< *bher-). – ἐν Βλαγάνοις (EKM 2, 64, 6; 66, 5, Aigeai, 216/7 CE) and Βλαγανῖτις (EKM 2, 67, Aigeai, 227/8 CE), probably related to βλαχάν· ὁ βάτραχος (Hesych.); see Hatzopoulos (1987a; 2018: 305).10 Further instances of instead of , respectively, appear in other areas of the ancient core of the Macedonian kingdom: at Methone11 and Pydna12  8 There are similar examples in personal names on inscriptions from Thessaly concentrated in the northern region of Pelasgiotis and in Perrhaebia, adjacent to the Macedonian cities of Aigeai, Beroia and to the region of Pieria (see Hatzopoulos 2007: 171; 2018: 312–313). Such proper names might designate Macedonians who settled in Thessaly or speakers of a Thessalian variety who shared this dialectal feature with Macedonians. In either case, they document linguistic contact and accommodation between speakers of different dialects. 9 There is only one example from Tyrissa, to the north of Pella, if the identification of the place name by the editors of EKM 2 is correct: [Β]άλακρος (EKM 2. 425. 23, Tyrissa, 3rd cent. BCE), a personal name formed from the adjective φαλακρός “bald”. 10 We should add ΒΙΛΑΒΡΑΤΕΛΛΟΥ (EKM 2, 23, Aigeai, 350–325 BCE) if the interpretation of the beginning as *Βιλάγρα (instead of *Φιλάγρα; Φίλαγρος is relatively common in Macedonia) or as *Βιλάβρα (instead of *Φιλάβρα) is correct (Hatzopoulos 1987b; see also the commentary on EKM 2, 23). 11 Βουλομάγα with the initial corrected to or the other way around (the letter O is smaller in size than the other letters; see Tzifopoulos 2010–2013) on an Orphic gold lamella (SEG 40.541, Methone, ca. 325–300 BCE); the form of this personal name corresponds to AtticIonic Φυλομάχη (see Masson 1984); its second element probably also reappears in Λαμάγα (SEG 50, 629, Pydna; cf. Hatzopoulos 2000). 12 Βίλιστος (IG X,2 2, 326 II, 73 and 77, 90/1 CE; Mélanges Daux 51, 10, Pydna, Alonia, 251 CE) corresponding to Φίλιστος in the remaining Greek dialects.

Dialects in Contact in the Ancient Kingdom of Macedon  

in Pieria; at Mieza and Beroia in Bottia; and in Eordaia, which was geographically part of Upper Macedon, but had always belonged to the Temenid kingdom.13 However, thus far there are only few extant sure examples of such alternative spellings in inscriptions carved in stone from the plains of Lower Macedonia on either side of the Axios River (Pella and Ichnai in eastern Bottia; Chalastra, Kalindoia, Sindos, Therme and Lete in Mygdonia; and Anthemous to the southwest and Crestonia to the northeast of Mygdonia), which were conquered by the Macedonians in the period between the end of the 6th cent. BCE and the end of the Persian wars. There are only two mentions of a priestess called Βερενίκα — a name of several queens and princesses of the Ptolemaic dynasty who are often mentioned in literary sources (see LGPN s.v.) — in two inscriptions from Lete (SEG 44.535) and a further single instance, which is in all likelihood secondary, of a lady called Βιλίστα (instead of Φιλίστα) on an inscription found on the western coastline of the peninsula of Sithonia.14 Turning now to the geographical areas conquered by the ancient Kingdom of Macedon at a later period, there are also examples of voiced plosives instead of aspirates from the region of Linkestis, to the north of Eordaia, which was conquered by the Macedonian kingdom at the beginning of the 4th cent. BCE and became part of Upper Macedonia.15 As to the form βεφαίως (Hatzopoulos 1991: 38–43, fig. 11; Hatzopoulos 1996: vol. II, no. 88, Amphipolis, 352–350 BCE), attested farther to the east, in the lands conquered by Philip II to the east of the Strymon River, it is likely to be a hypercorrection for βεβαίως “firmly”, “certainly” due to linguistic contact between speakers of different dialects and accommodation to each other for communication to succeed.16 The most likely interpretation of such graphic alternations is that instead of , respectively, reflect the spirantization of the inherited voiced plosives and of the voiceless aspirated plosives in Macedonian

 13 Mieza: Βα[λάκρου] (ΕΚΜ 2, 93 A’ 7, 350–325 BCE); Beroia: Βίλος (EKM 1, 179, 2nd cent. BCE); Eordaia: Βίλος (EAM 87, 12, 181–180 BCE = [ΜΕΛΕΤΗΜΑΤΑ 22, 2, no. 17, 181/180 BCE]); corresponding to Φίλος. 14 SEG 38.600, Macedonia: Neos Marmaras, late Hellenistic period. 15 Βίλου (EAM 179, Lynkestis, Lyke, Hellenistic period), corresponding to Φίλου in the remaining Greek dialects; cf. φίλος (see Masson 1984). 16 The personal name Κεβάλιος in a Dodona tablet (ΧΕΔ 2493A) probably refers to a Macedonian since the form of the name provides a further example of instead of (cf. κεφαλή “head”; Hsch. κεβαλή· κεφαλή), but his exact origin is unknown (cf. Méndez Dosuna 2007, 2018: 292–293). According to Méndez Dosuna (2018: 293), Jaime Curbera may be right in interpreting the personal name Ἄλβειος which appears in another Dodona tablet (ΧΕΔ 94, 400–350 BCE) as Ἀλφειός, a name attested in Thessaly (SEG 56.638, ca. 160–150 BCE) and elsewhere.

  Emilio Crespo (Hatzopoulos 2007; Méndez Dosuna 2012; 2014). We cannot determine if such evolution started in all or only in certain contexts, but the latter possibility seems more probable (Crespo 2018: 344). (2) (and < Β, Δ>?) instead of < Κ> (and ?), respectively, in inscriptions from Aigeai:17 – Διγαῖα (along with Δειγαῖα, Δειγέα) for Δικαία (EKM 2, 62, 1; 63, 3; 69, 10; 71, 2; 72, Aigeai, all of them between 189 and the first half of the 3rd cent. CE); see Hatzopoulos (1987a; 2018: 305). There are further instances of instead of in late inscriptions from Bottia and from other areas of Macedonia and elsewhere undated or dated to the 2nd and 3rd cent. CE, as in the pairs Κλεοπάτρα – Γλευπάτρα (EKM 2, 143, 3–5, Edessa, 215 CE), Βάλακρος – Βάλαγρος (IG XI, 4, 585, 3, Delos, 300–250 BCE; see LGPN s.v.), Κερτίμμας – Κερδίμμας (Arr., An. 2, 13, 7, 4th cent. BCE; cf. LGPN s.v.), Κυδίας – Γυδίας (Hatzopoulos, Meletemata 22, 2, no. 62, Kalindoia, c. 316 BCE; cf. SEG 36.626), Κραστωνία – Γραστωνία, Γορτυνία – Γορδυνία (Hatzopoulos 2018: 305). An isolated and late inscription found close to Lake Ohrid bears witness to instead of .18 Finally, there are inverse examples, which seem to be hypercorrections, of instead of 19 at Leukopetra, and of instead of in inscriptions undated or dated to the 2nd or 3rd cent. CE.20 The graphic alternations between (and < Β, Δ>?) and < Κ> (and ?), reflect the fricativization or spirantization of the voiced plosives inherited from Indo-European. The fricativization of the inherited voiced plosives did not start in all contexts but, as in other ancient and modern languages, only between vowels or between a vowel and a resonant in the spoken string and perhaps in other contexts. This circumstance would explain the alternations of voiceless and voiced plosives (Crespo 2018: 344–345). Nevertheless, one should not exclude the possibility that all or some later examples such as Γλευπάτρα instead  17 If ΔΑΠΙΝΑ in the Pella curse tablet represents ταπεινά (see Dubois 1995; Voutiras 1998), we would also have a single instance of for . Giannakis (2020) reads ΔΑΓΙΝΑ and interprets it as δαFινα, an adjectival formation meaning “burnt, wretched; abandoned, deserted”. 18 ὑβό for ὑπό (IG X,2 2. 396, 2, Dassaretia — Lychnis (Ohrid) —Monastery of Saint Naum — 3rd cent. CE). 19 For ὑπροσσθῆνε for ὑβρισθῆναι (I.Leukopetra, 5, 2nd cent. CE) and ὑπρείσῃ (I.Leukopetra, 10, 2nd cent. CE) as against ὑβρείσῃ (I.Leukopetra, 5) see Brixhe (2010: 65). 20 κλυκυτάτῃ instead of γλυκυτάτῃ is widespread over Macedonia; τάκρυν instead of δάκρυν is attested in inscriptions from Beroia of the 3rd and 4th cent. CE (Hatzopoulos 2018: 305).

Dialects in Contact in the Ancient Kingdom of Macedon  

of Κλεοπάτρα are due to a secondary (hypercorrect) imitation of a peculiarity in the pronunciation of the ancient Macedonian dialect of Greek. In the epigraphic evidence brought to light so far, most alternative spellings of instead of or of , respectively, are documented at Aigeai and in the western core of the ancient Kingdom of Macedon. It is true, however, that since the evidence is scant and originates from personal names, there is only limited basis for the hypothesis that this feature distinguishes two geographical varieties. Nevertheless, in view of the extant evidence, said hypothesis is the most likely. Hopefully, new inscriptions will eventually confirm or refute it. The existence of two varieties in Aigeai and Pella, distinguished by the presence or absence of alternative spellings instead of or of , respectively, is congruent with the historical reconstruction of the chronology and the reasons that the Macedonian capital was moved from Aigeai to Pella. Historians date the relocation of the capital during the reign of Archelaus I (413–399 BCE; Roisman 2010: 156) or of Amyntas III (393 and 392–370 BCE; Hatzopoulos 1996, I 174–179; I 471–472). The tragic poets Euripides and Agathon, as well as Zeuxis, the most famous painter of his time, spent time in Macedonia, having been invited by Archelaus to his court. Such invitations show that this king made a concerted effort to open his kingdom to a wider world. Shifting the capital from Aigeai, in the ancient core of his realm, to Pella, close to the lands that had been definitively conquered by Macedonia to the east of the Axios at the time of the Persian wars and probably connected to the sea by the lake and the mouth of the Loudias River, is consistent with the goal of opening the country to foreign influences. In this respect, it is probably no coincidence that whereas only one foreigner appears in the inscriptions from Aigeai,21 a total of eight foreigners are documented in the inscriptions from Pella.22 The new

 21 Εὐθίας [Λ]ουσίο[υ] Ἀργεῖος (EKM 2, 28, Aigeai, 350–300 BCE). EKM 2, 51, Aigeai, dated to the Roman period, has pre-Hellenic personal names, one female and another male. EKM 2, 55, Aigeai, has a man named Amadokos in a funerary epigram written in a dialectal form that may be interpreted as Attic or as Ionic (Hatzopoulos 2011b: 61). 22 Until Hellenistic times foreigners at Pella came not only from other Macedonian cities such as Lete (EKM 2, 466), Sirrae (EKM 2, 509) and probably Herakleia in Mygdonia (EKM 2, 476), but also from other cities or regions of the Greek mainland such as Phalanna (EKM 2, 542) and Pherai (EKM 2, 502) in Thessaly, Thaumakós in Achaia Phthia (EKM 2, 486), Boeotia (EKM 2, 513), Corinth (EKM 2, 544) or from more distant places: Delos (EKM 2, 507), Magnesia on the Maeander (EKM 2, 445) and Soloi (EKM 2, 480) (cf. EKM 2, vol. II, p. 607). Romans (EKM 2, 439) and Italians arriving from Greek-speaking cities (EKM 2, 548) are attested from the late 2nd cent. BCE onwards. A relatively large number of citizens from Pella are known from external literary

  Emilio Crespo capital soon became the most important city, and in 380 BCE Xenophon (HG 5, 2, 13) was already calling it μεγίστη τῶν ἐν Μακεδονίᾳ πόλεων “the biggest of the cities in Macedonia”. In addition, the transfer of the capital was intended to enable control of the cities and regions lying to the east of the Axios River (Mygdonia, Crestonia, Bisaltia, and Anthemous), which had been definitively annexed to the Macedonian kingdom after the Persian wars and probably held the status of allied cities, not of ethne like the ancient core of the kingdom. It is therefore likely that the absence of alternative spellings instead of or of , respectively, at Pella and its scarcity in the regions to the east of the Axios are a consequence of the progressive decline of a pronunciation that was closely associated to the more archaic areas of the Macedonian kingdom. The opening of Pella to influences from abroad also explains the fact that, while in Pella there are three inscriptions in the Attic dialect (EKM 2, 434, ca. 300 BCE; 481, ca. 300 BCE; 493, 300–100 BCE) and two in a dialectal form that may be interpreted as Doric or koine (EKM 2, 535, 150–100 BCE; 547, 350–325 BCE), no inscription from Aigeai can be understood as being unquestionably written in Attic or in an undifferentiated form that could be Doric or koine. This suggests that the Attic-Ionic koine penetrated the new capital at an earlier date, and more profoundly.

 Dialectal features common to Aigeai and Pella A closer look at the inscriptions published in the recently released volumes of EKM 2, and at other non-literary documents such as curse tablets and Orphic gold lamellae found in the ancient Temenid kingdom, reveals several features that are common to its two successive capitals. First, both cities exhibit a lack of dialectal unity from the earliest inscriptions up to the progressive yet never fully complete predominance of the AtticIonic koine. While every Greek political centre located to the south of Macedonia — excluding the pan-Hellenic sanctuaries of Olympia, Delphi, and Dodona — displays a single dialect, only threatened by the emergence of the koine, the ancient Temenid kingdom lacks dialectal unity right from the earliest docu and epigraphic sources (cf. EKM 2, vol. II, p. 60734). Some individuals bear non-Hellenic names: one of the two women mentioned in 454 has an Illyrian name, the name of the women mentioned in 514 and 564 is pre-Hellenic (see EKM 2, vol. I, p. 95), and the name of the father referenced in EKM 2, 523 is probably Thracian.

Dialects in Contact in the Ancient Kingdom of Macedon  

ments that have come down to us, dated to the late 5th cent. BCE (see Bufalis 2020: 855). Even the inscription that seems to be the most ancient epigraph from Aigeai, a public document dated to the end of the 5th or the beginning of the 4th cent. BCE (EKM 2, 6), bears witness to the Attic-Ionic sequence τόσημ μέν. Many epigraphs are written in a variety that could be called Attic-Ionickoine, a designation intended to reflect that the document in question cannot be assigned to any one of these particular dialects, or in a variety that could be labelled Doric-Northwest Doric in the sense that its dialectal form lacks the particular features that distinguish all of the Doric and Northwest Doric varieties. Some other inscriptions are written in koine, in Attic, in Ionic and in what we call Attic-koine or Doric-koine. Furthermore, several inscriptions are hybrid and attest features that belong to different dialectal varieties. Finally, the dialect of other inscriptions is undetermined and is referred to as uncertain. The results are presented in Table 1: Tab. 1: Dialects in the inscriptions found at Aigeai and Pella. Dialect Attic-Ionic-koine Attic-koine Ionic

Aigeai

Pella













Koine

23

24

Doric-NW Doric

25





26

Thessalian Hybrid





Uncertain





Total inscriptions published in EKM 





Such dialectal diversity occurs not only in inscriptions carved on stone, but also in texts written on other surfaces. A case in point are the golden Orphic lamellae  23 Eight of them show in lieu of or of (see § 2 above). 24 EKM 2, 535 (Ἀρ]χελάου ἥρωι, 150–100 BCE) and 547 ([----]άσιον λαοῖς ἰατρόν, 350–325 BCE), which might also be taken as undifferentiated forms either of Doric or of koine, are classed as koine. 25 Three of them have instead of (see § 2 above). 26 The inscription on Tomb B from Derveni is also written in the Thessalian dialect of Larisa: Ἀστιούνειος Ἀναξαγοραίοι ἐς Λαρίσας.

  Emilio Crespo found in the ancient Kingdom of Macedon, which show the varieties displayed in Table 2:27 Tab. 2: Dialects in the Golden Orphic Lamellae found in the ancient Kingdom of Macedon. Aigeai Attic-Ionic-koine

Methone



Pella

Mygdonia

Amphipolis







Attic Koine Doric-NW Doric Hybrid Uncertain

28

  

The lack of dialectal unity is partly due to the fact that some areas conquered by the Macedonian kings were ancient Greek colonies — most of them dispatched from Chalcis or Eretria in Euboea or from various Cycladic islands — which spoke the dialect that the colonists had brought from their metropolis much earlier. Such conquests were a comparatively recent phenomenon, dating from the late 6th cent. BCE until after the Persian wars. Another wave of subsequent conquests occurred in the mid-fourth cent. BCE, when Philip II extended the Macedonian kingdom to the Chalcidice peninsula, where most cities had been founded by Euboean colonists, and farther out, to more eastern areas previously dominated by Thracian peoples. The historical causes of this lack of dialectal unity will be discussed in § 4 below. Secondly, most personal names documented by the inscriptions from Aigeai and Pella belong to the pool of personal names found all over Macedonia (see Hatzopoulos 2000; EKM 2 passim; Boufalis 2020: 855–856). Most inscriptions are on graves and consist solely of personal names. As we know, such forms are generally unhelpful in identifying the dialect spoken in a geographical area, because personal names tend to travel and to cross linguistic borders.

 27 Orphic gold lamellae from Macedonia are brought together in Fragment 496 of Poetae Epici Graeci. Orphicorum et Orphicis similium Testimonia et Fragmenta. Pars II, fasc. 1–2, ed. A. Bernabé, Monachii et Lipsiae (2004–2005). For their dialectal diversity see Crespo (2011). For the lamella from Methone, see Tzifopoulos (2010–2013). 28 The text has a corrected into or the other way round; see fn. 12 above.

Dialects in Contact in the Ancient Kingdom of Macedon  

Thirdly, the inscriptions from Aigeai and Pella, regardless of the dialect in which they are written, attest to alternative spellings among which stand for front mid vowels and, more rarely, between which represent mid or close back vowels. A list of examples from Aigeai and Pella is given in (3) and (4): (3) Alternative spellings between : – Ἀντ[ι]γήνης Φιλλ[έ]α (EKM 2, 18, Aigeai, 350–300 BCE): H for E – Μακηδον[ίη] (EKM 2, 57, Aigeai, ca. 350 BCE): H for E metri causa?29 – Ἀντιπατρίδης Ἀγαθοκλήους (EKM 2, 459, Pella, ca. 250 BCE): EI corrected to H – κῖμαι (EKM 2, 471, Pella, ca. 350 BCE): I for EI – [Ἡ]ρακλέδης (EKM 2, 494, Pella, 350–300 BCE): E for EI – ἐσστε (for ἐστί), Ἰφεκράτους (EKM 2, 542, Pella, 375–350 BCE): E for I – Ἡρακλίδης (EKM 2, 574, Pella, undated): I for EI – διελέξαιμι for διελίξαιμι (SEG 43.434, Pella, 375–350 BCE): E for I – ἰμέ if written for εἰμί (SEG 43.434, Pella, 375–350 BCE): I for EI and E for I – πάλειν (πάλ{}ιν according to Bufalis 2020, 248) for πάλιν (SEG 43.434, Pella, 375–350 BCE): EI for I – ΔΑΠΙΝΑ for ταπεινά? (SEG 43.434, Pella, 375–350 BCE): I for EI? (4) Alternative spellings between : – Πάτων Λαάνδρο (EKM 2, 42, Aigeai, 350–325 BCE; Boufalis 2020, 195 Λαάνδρου͎̣ ): O for ΟΥ. – ἀνορόξασα for ἀνορύξασα (SEG 43.434, Pella, 375–350 BCE): O for Y.30 The interpretation of such alternative spellings is controversial (see Méndez Dosuna 2014: 396). They have been explained as the outcome of the raising of /e/ into /i/ (early iotacism; see EKM 2, vol. II, pp. 819–820) and of the corresponding lowering of the close back /u/ into /o/, or as the result of the phonetic lowering of /i/ and /u/ into /e/ and /o/, respectively, with occasional graphic hypercorrections in both hypotheses.31 In any event, given the number of instances, at least the alternative spellings are unlikely to reflect mere

 29 Also, in EKM 1, 38, Beroia. 30 The genitive Εὐκλέος in Ἀππελῆς Εὐκλέος (EKM 2, 460, Pella, 4th cent. BCE) probably does not belong to this group. The genitive -κλέος is the most frequent form in most dialects (Buck 1955: 42) and particularly in Northwest Doric Dialects (Méndez Dosuna 1985: 79–86). 31 According to Lhôte (2006: 431), such alternative spellings reflect an isogloss linking Macedonia with Thessaly, Epirus and Boeotia, regions where parallel instances occur.

  Emilio Crespo graphic errors. Regarding the alternative spellings , since they are rarer, they may represent simple graphic mistakes. But if they reflect a phonetic shift, they will indicate that was pronounced as a close back vowel in the Macedonian dialect (not a front one as in Attic-Ionic-koine) which is neutralized with a close-mid vowel in contact with /r/. The famous curse tablet found at Pella displays instances of alternative spellings between and : διελέξαιμι for διελίξαιμι, ἰμέ for εἰμί, πάλειν for πάλιν, ΔΑΠΙΝΑ for ταπεινά, if this is the correct interpretation (see Dubois 1995; Voutiras 1998), and ἀνορόξασα for ἀνορύξασα. Overall, it shows a Doric variety with a cluster of features which is not coincident with those that are attested by the other Doric and North-West Doric varieties.32 Thence one can infer that Thetima’s curse reflects a Doric dialect not imported but proper to Macedonia. Besides this, the ending of dative in δαίμοσι and, in particular, the spelling γίνο[μαι with total assimilation of /g/ may reflect the influence of the Attic-Ionic or of the koine dialect on this Doric variety; in this case the Pella curse dialect would be hybrid, which is frequent in other documents found in Macedonia (see below). Τhe dating of the tablet in the first half of the 4th cent. BCE makes this influence plausible. However, even if this were the case, the argument would remain that the combination of Doric features found on the curse plate is not documented anywhere else. Based on the current evidence, we can safely conclude that the inscriptions on stone are likely to reflect the same variety of Northwest Doric dialect as the Pella κατάδεσμος. Finally, some inscriptions from both Aigeai and from Pella present a hybrid dialect, something that occurs only rarely and at a much later date in other Greek-speaking areas. Several examples are given in (5): (5) Examples of dialectally hybrid documents: – EKM 2, 34, Aigeai, 350–300 BCE: Ḳα̣λ̣λ̣ί̣α̣ς̣ Ἀ̣ρτεμιδώρου, Αρτεμίδωρ̣ο̣ ς̣ Κ̣α̣λλία, Δημητρία Καλλία (Attic-Ionic-koine Δημητρία vs. Doric Καλλία genitive, masculine). – EKM 2, 56, Aigeai, 350–325 BCE: Ἁρπάλο Σκύτας. ἀδελφή με ἀνέθηκε Παγκάστα (Attic-koine ἀδελφή vs. Doric -ᾱ in Παγκάστα). – EKM 2, 484, Pella, 300–250 BCE: Εὐβούλα Λύσωνος (the root of Εὐβούλα is Attic-Ionic-koine, but the ending is not).

 32 See Brixhe (1999: 45–51); Panayotou (2007); Méndez Dosuna (2012; 2014); Hatzopoulos (2018: 310–314).

Dialects in Contact in the Ancient Kingdom of Macedon  

–

EKM 2, 495, Pella, 3rd cent. BCE: Καλλίας Δημητρίω, Δημήτριος Καλλίω, Ὀαδίστη Δημητρίου (If Ὀαδίστη is an approximate spelling of Ϝαδίστη,33 its beginning is not Attic-Ionic-koine, but its ending -η and that of γυνή is. The form of Δημήτρι- is also Attic-Ionic-koine as against Doric Δᾱμᾱ́τηρ. The genitive endings of the personal names referred to present difficulties. On the first line the editors of EKM 2 edit Δημητρίω, assuming a graphic error due to the merger of /ɔ:/ and /o:/. This reading seems to go against the reading Δημητρίου on the third line which is undoubted. On the second line Καλλίω (if this is the right reading) could be Attic-Ionic (with ancient -ιέω contracted into -ίω) or, more probably, koine, with -ω being also a graphic mistake for -ου due to the merger of /ɔ:/ and /o:/).

The inscriptions that show a hybrid dialect may be manifestations of long-term contact among speakers of different dialects and of the ensuing dialectal convergence between the vernacular Northwest Doric and the Attic-Ionic, which had greater prestige in written usage (see § 4 below).34 It is remarkable that the Derveni papyrus also exhibits a hybrid dialect which has been described as northern eastern Ionic (from Chios) with an Attic overlay (Kouremenos/Parássoglou/Tsantsanoglou 2006) or as Attic with northern eastern Ionic features (Willi 2014).35 The following alternative Ionic and Attic spellings are attested: – αἰτίην (col. III 9) / αἰτίαν (col. V 7) and further examples of ᾱ/η after ε, ι, ρ. – κυκλοειδέα (col. XXIV 2) / ἰσομελῆ (col. XXIV 2) and similar instances. – οὐδέκοτε (col. ΙΙΙ 6) / ὁπότε (col. III 6), οὐδέποτε (col. XII 10). – a single uncertain instance of -oισιν (col. XXI 11) / -οις in many cases. – ἑωυτοῦ (col. XIV 2) / ἑαυτοῦ (col. XX 14) and αὑτ- (col. IX 3). – ἐπείτ(ε) (twice; cf. col. XVII 5) / ἐπεί (twice; cf. col. XVIII 10). – Ζᾶνα (col. IX 3 and at least a further instance) / Ζῆνα (col. XVIII 9 and a further example). – ἐών / ὤν (both forms are common in different cases and numbers).

 33 See also EKM 2, 516, 3rd cent. BCE Ὀαδεία (Ϝᾱδεία); EKM 2, 455, ca. 350 BCE Ἁδέα; EKM 2, 491, 325–300 BCE and 492, 275–225 BCE Ἡδίστη. The text is quoted according to the edition in EKM 2, 495. On Ὀαδίστη as an approximative spelling of Ϝαδίστη, see Cassio (1991). 34 Thucydides 6.5 reports that most of the settlers who founded Himera were Chalcidians who were joined by some exiles from Syracuse and that the language of the colony was a mixture of Doric and Chalcidian, although their institutions were those of Chalcis. 35 Lulli (2011: 98) defines the dialect as “una mistione di ionico e attico capace di imporsi come lingua letteraria transregionale”. See also Crespo (2019).

  Emilio Crespo Furthermore, the following Ionic spellings are consistently employed in the preserved fragments of the Derveni papyrus and do not alternate with the corresponding Attic forms: – -σσ- instead of Attic -ττ-: e. g. ἅσσα (e.g., col. X 13), not Attic ἅττα. – ἤν (col. XXIV 5) as against as Attic ἐάν. – ending -έων in the genitive plural of -ᾱ-noun-stems (ὡρέων in XXIV 11), but no sure instance of Attic -ῶν. – γίνεσθαι and γινώσκειν, not Attic γίγνεσθαι and γιγνώσκειν, respectively. – ἕνεκεν (not εἵνεκεν), not Attic ἕνεκα. – preposition σύν, not archaic Attic ξύν. Conversely, the following Attic features are consistently used in the preserved text and do not alternate with the corresponding Ionic forms: – -ου- as the outcome from -εο- and -οο-: e.g., νοῦς (col. XVI 15). – -εῖον, not Ionic -ήϊον. – ἀεί, not Ionic αἰεί. – μόνος, not Ionic μοῦνος. – ἀδελφή, not Ionic ἀδελφεή. – θαυμάζειν, not Ionic θωμάζειν. – κοιν-, not Ionic ξυν-. – οὖν, not Ionic ὦν.36 The Derveni papyrus also attests the Doric, Epic, and tragic personal pronoun νιν in the author’s commentary (col. XI 3) and in a quotation from Heraclitus (col. IV 9) that is also quoted by Plutarch (Mor. 604A, De exil. 11) with the expected Ionic form μιν instead of νιν. Dialectally hybrid texts are extremely rare amongst non-literary documents from other areas of the Greek-speaking world around 400 BCE but seem to be commonplace in Macedonia. The use of a hybrid dialect in the Derveni papyrus probably shows that the phenomenon of dialectal hybridization existed both in oral communication and in artistic literary prose because exchanges between speakers of different dialects were common until the configuration of the koine.

 36 Many forms found in the Derveni papyrus may be Attic and Ionic. At least two cases are debatable: the ᾱ in ἀήρ is Attic and old Ionic and Homeric, and Ionic ἠήρ is secondary. The verb ὀνομάζω is Attic and Ionic, but the noun οὔνομα, systematically employed by Herodotus, may have been also Ionic.

Dialects in Contact in the Ancient Kingdom of Macedon  

 Vernacular dialect and adopted dialect Let us turn now to the question of how to account for the lack of dialectal unity in the Macedonian kingdom before Alexander the Great. As we have seen in § 3 above, the Pella curse tablet exhibits a cluster of features that do not appear in any other vernacular or literary Doric subdialect. In view of that, the most likely hypothesis is that it reflects a vernacular dialect spoken at Pella and that the inscriptions found in Macedonia that we have described as Doric-Northwest Doric represent the same or a very similar vernacular dialect. The brevity of the preserved documents, however, makes it impossible to identify other features of this variant of Doric. In contrast, the inscriptions that show the dialectal variety that we conventionally call Attic-Ionic-koine would not reflect a vernacular dialect in the ancient core of the Macedonian kingdom, but rather vernacular dialects spoken by successors of the colonists who had come from abroad to areas that were later conquered by the Temenid kingdom (Bufalis 2020: 8). As we know, the first official documents issued by the Macedonian kingdom were found outside of the old Macedonian kingdom and are written in koine (Horrocks 2010: 80). The rapid conquests in the geographical area situated beyond the ancient eastern border of the Macedonian kingdom along the Axios River (Mygdonia, Anthemous, Crestonia and Bisaltia in the period between the end of the 6th cent. BCE and the end of the Persian wars and, farther to the southeast, Chalcidice, Olynthos, the peninsulas of the Pallene, Sithonia and Akte, Amphipolis and the western Thrace in the first half of the 4th cent. BCE) and the subsequent movements of people to the east surely played a crucial role in the linguistic development of the Macedonian kingdom.37 These conquests probably provoked the migration of many people and the foundation of new towns with populations who came from diverse provenances and spoke different dialectal varieties. From a dialectal point of view, the diversity of settlers in each settlement would

 37 The ancient Temenid kingdom displays a wide dialectal diversity ranging from North-West Ionic in the Euboean colonial foundations (Olynthos and Thorone founded from Chalcis; Dicaea, Mende in the Pallene and Methone on the Pierian coast, founded from Eretria) to Attic (Amphipolis was founded as a colony from Athens in 437 BCE), to insular Ionic in the colonies dispatched by the Cycladic islands (Acanthus, Stagira, Ennea Hodoi, Skione) and Doric in the Corinthian colony of Potidaea. The Attic dialect was imposed in the cities that were members of the Attic-Delic League (Argilos, Berge, Methone, Herakleion in Mygdonia, Aphytis etc.). After 424, Thorone, Acanthus, Mende, Skione and Amphipolis were not part of the alliance (koinón) of the Chalcidians, but they cooperated with cities that were hostile to Athens.

  Emilio Crespo have given rise to the diversity of dialectal varieties observed in the inscriptions from Aigeai and Pella. Later, the progression of long-term contacts between speakers of different dialects that were mutually intelligible led to the mixing of their varieties. Some mixing of dialects is evidenced by the hybrid documents found in the Macedonian kingdom. The last steps of progressive dialectal convergence in this milieu characterized by a frequent and increasing contact among speakers should have been the predominance of a given form among the varieties that were in use in every context and, finally, the generalization or leveling of one or other form of a given variety with loss of the other alternatives. This never-ending development gave rise to the fixation of koine (Bubenik 2014).38 Although we lack the evidence on the reasons behind the dialectal diversity observed and the early appearance of dialectally hybrid documents, this historical hypothesis finds some support in what happened in Alexandria two or three generations after the conquest of Egypt by Alexander the Great, and much later in some dormitory towns founded during the 20th cent. CE, whose populations lost their ancestral dialect and created a new one based on various provenances (see Kerswill/Trudgill 2005). The papyri of Alexandria show that the Greek dialects of the Macedonian conquerors who arrived in Alexandria from Macedon and other geographical areas of the Greek-speaking world quickly disappeared, leaving behind only a small number of tokens attesting to a transitory mixing; in a few generations such dialectal elements were replaced by the Attic-Ionic koine. Among such vestiges, there are remains of the use of the Doric dialect among the Macedonians who arrived in Egypt with Alexander's expedition or shortly thereafter.39 Thus, the earliest dated Greek papyrus from Egypt (P. Elephantine 1, 310 BCE), which contains a Macedonian marriage contract before six witnesses and is

 38 The process by which a new dialect was formed had likely started by the beginning of the 4th cent. BCE in the Chalcidic paeninsula before its conquest by Philip II. An agreement between the Macedonian king Amyntas III and the Chalcidians that has come down to us in an inscription from Olynthos (Hatzopoulos 1996 II 2, 393/2 BCE) offers a hybrid dialect with a majority of North-Ionic features (among them Μακεδονίης, φιλίην), but also with Attic μιᾶς. 39 In Alexandrian literature there are terms that present unmistakably Macedonian phonetic features: for example, κεβλήγονος (Euphorion 108), κεβλή (Calimachus fr. 657) and κεβλήγονος (Nicander Alexiphármaka, 433), which correspond to Attic-Ionic κεφαλή “head”, document the characteristic Macedonian β instead of φ in intersonorant contexts (i.e., in intervocalic position or after a vowel and before a resonant). See also Callimachus (fr. 203.17–18 Pf.) τοῦτ᾿ ἐμπ [έ]πλεκται καὶ λαλευσ|[ . . ] . [ | Ἰαστὶ καὶ Δωριστὶ καὶ τὸ σύμμικ|το̣ ν̣ .] “is intertwined and they speak... in Ionic, in Doric and in a mixture”.

Dialects in Contact in the Ancient Kingdom of Macedon  

written in koine, bears witness to the masculine plural nominative of the article in Doric and in other dialects τοί instead of οἱ, and to the non-koine forms εἰάν (twice) instead of ἐάν and μάρτερες in the interior writing, which alternates with μάρτυρες in the duplicate.40 In this regard, there is a telling passage from Theocritus (15.90–93) in which two Syracusan women who are attending a procession rebel against other attendants who ask them not to speak, replying that they have the right to speak Doric (see Willi 2012).41 In addition, various sociolinguistic studies on dialects spoken in cities that were newly created in the 20th cent. with immigrants from areas where different dialects of the same language were spoken have shown that the first generation of immigrants created a peculiar dialect that retained traits of the dialectal diversity brought by such immigrants from their places of origin, and that the inhabitants of these cities, after a short period of mixing and leveling, create a stable dialect from the third generation onward (see Trudgill 1986; Kerswill 2002). Similarly, it is plausible that the population movements caused by the Macedonian conquests favoured the emergence of local hybrid dialects that soon suffered competition from Great Attic (i.e., international use of Attic) and the incipient Attic-Ionic koine, which ultimately gained ground. The use of a non-vernacular form from the early 4th cent. BCE and in the first official inscriptions of the kingdom of Macedonia from the last third of the 4th cent. BCE suggests that the linguistic form that would develop into the AtticIonic koine and was gaining ground was adopted at some stage and became the vehicle of formal communication. The Macedonian court is likely to have adopted a form based on Great Attic as its official vehicle at least for written formal communication. As Brixhe/Panayotou (1988) persuasively suggested, the Macedonian kingdom played a key role in the configuration and diffusion of the foreign Attic which evolved into a form of Greek that over time would come to be called (Attic-Ionic) koine; at the end of the never ended process of its configuration, the verb μακεδονίζειν would become the common way to say, “to speak in koine”. The main motivation for the Macedonian court’s adoption of the AtticIonic koine for official communications was likely the fact that the vernacular

 40 As in Macedonian inscriptions from the late 4th cent. BCE, the poet Posidippus of Pella combined Doricisms with epic, Ionic, Attic and koine forms in some of his poems, e.g., Posidippus 139 Austin-Bastianini (8 GP = AP XII 131); see Sens (2004). This literary phenomenon has its roots in the literary tradition but may also have been fostered by the dialectal diversity of the Greek spoken in the Macedonian kingdom. 41 In Greek papyri from Egypt of the Ptolemaic era, some personal names are attested with final alpha as in Doric rather than with eta, as in normal koine forms (see Clarysse 1998: 4–6).

  Emilio Crespo Northwest Doric dialect group had relatively little overt prestige and was mainly used orally.

 The Macedonian court’s adoption of Attic-Ionic for official communication We can hardly determine the date when the ancient Macedonian court adopted for official communications the form of Greek that over time would become the Attic-Ionic koine. However, it is remarkable that IG II2 236 (338/7 BCE), which was found at Athens and records the oath sworn by the Athenians as members of the Greek synédrion under Macedonian hegemony commonly called the League of Corinth (338/7 BCE), where Philip II was elected as commander-inchief of the Greek general alliance, attests θάλασσαν in the oath and Θεσσαλῶν in the list of allied states, with σσ, not ττ:42 (6) IG II2 236 (338/7 BCE), Fragment A, col. I, 7–8: [οῖς ὅρκοις οὔτε κατὰ γῆν] οὔτε κατὰ [θ]άλασ[σαν· οὐδὲ πόλιν οὐδὲ φρο]ύριον καταλήψομFragment B, 2–5:

[— — — — — — — — — — ∶ Θεσ]σαλῶν ∶Δ [— — — — — — — — — — — ῶ]ν ∶ΙΙ [— — — — — — — — — Ἐλειμ]ιωτῶν ∶Ι [— — — — Σαμοθράικων καὶ] Θασίων ∶ΙΙ

Thus, the coexistence of several dialects in some cities of the ancient Temenid kingdom and the lack of a written form of the Macedonian vernacular dialect could have encouraged the early adoption of the international form of Attic that nowadays is called Great Attic for the kingdom’s official documents. The date of the creation of the League of Corinth is thus the earliest attestation of -σσ-, one of the distinctive features of what would later be called koine,

 42 It would make no difference to the argument should it be true that the treaty refers to the peace between Philip and Athens after the war of 340–338 BCE, as Worthington (2009) suggests.

Dialects in Contact in the Ancient Kingdom of Macedon  

in a public document found at Athens. As in Thucydides’ history, but unlike the rest of the public inscriptions, both contemporary and later, written in the Attic dialect, the presence of -σσ- instead of -ττ- avoids an Attic oddity to adhere to a feature that is common to most Greek dialects and that would be fixed in the koine. It is highly plausible that the political circumstances after the defeat of Chaeronea advised the Athenians to accommodate themselves to the common behaviour by avoiding peculiarities. It is likely that the Macedonian court, who imposed the content of the agreement at the conference held at Corinth, where the Greek states signed a common peace (κοινὴ εἰρήνη) under the leadership of Philip II, also fixed a dialectal trait of what over time would come to be known as the koine.

 Conclusions The conclusions drawn from the previous discussion can be summarized as follows: 1. The inscriptions from Aigeai, the ancient capital of the Macedonian kingdom, and from the other regions (Pieria, Beroia and Eordaia) that formed the core of the ancient Temenid kingdom show occasional instances from the 4th cent. BCE in which the graphemes appear instead of or of , respectively. Similar examples are also attested in northern Thessaly. Such alternative spellings, however, do not occur in Pella, the new capital of the kingdom, and are rare in the eastern areas of Macedonia. 2. Although the inscriptions and other texts from Aigeai and Pella exhibit dialectal variations in even the earliest documents, they share a common dialectal ground. This common ground is seen in the lack of dialectal unity (Doric-Northwest Doric dialect group and Attic-Ionic-koine being the most common dialects in inscriptions), in the existence of an inherited pool of personal names, in the early rise of alternative spellings among and perhaps among , and in the early frequency of dialectally hybrid documents, a feature shared by the Derveni papyrus. 3. While a variety of the Northwest Doric dialect group was the vernacular dialect for most of the population in the ancient core of the Macedonian kingdom, forms of Attic-Ionic and of international Attic (Great Attic) were vernacular in the eastern areas and cities conquered by the Macedonian kingdom. Such varieties, after a process of increasing long-term

  Emilio Crespo

4.

communicative exchanges between speakers of different dialects, linguistic convergence, dialectal mixing, and subsequent levelling at a not only Macedonian, but also Hellenic scale, gave rise ultimately to the Attic-Ionic koine. The Macedonian court adopted a form of Attic-Ionic as the vehicle of public communication later. The adoption of Great Attic (which was to evolve into Attic-Ionic koine) as its vehicle of formal communication took place before 338/7 BCE and is probably related to the relocation of the capital to Pella during the reign of Archelaus I (413–399 BCE) or of Amyntas III (393, 392–370 BCE).

Bibliography Andronikos, M. (1984), Vergina. The Royal Tombs, Athens. Andronikos, M. (1994), Βεργίνα ΙΙ Ὁ τἀφος τῆς Περσεφόνης, Athens. Buck, C.D. (1955), The Greek Dialects, Chicago/London. Besios, M./Tzifopoulos, Y./Kotsonas, A. (2012), Μεθώνη Πιερίας I: Επιγραφές, χαράγματα και εμπορικά σύμβολα στη γεωμετρική και αρχαϊκή κεραμική από το ‘Υπόγειο’ της Μεθώνης Πιερίας στη Μακεδονία, Thessaloniki. Bufalis, A. (2020), Οι επιγραφές της αρχαϊκής και κλασικής εποχής στη Μακεδονία (doctoral dissertation), Aristotle University of Thessaloniki. Brixhe, C. (1999), “Un ‘nouveau’ champ de la dialectologie grecque : le macédonien”, in: A.C. Cassio (ed.), KATA DIALEKTON. Atti del III Colloquio Internazionale di Dialettologia Greca, (= AION 19, 1997), 41–71. Brixhe, C. (2010), “Quelle koiné en Macédoine au début de notre ère? Le test de Leukopétra”, in: F.C. Gabaudan/J.V. Méndez Dosuna (eds.), Dic mihi, Musa, virum. Homenaje al profesor Antonio López Eire, Salamanca, 61–68. Brixhe, C./Panayotou, A. (1988), “L’atticisation de la Macédoine: L’une des sources de la koiné”, Verbum 11, 245–260. Bubenik, V. (2014), “Koine, origins of”, in: G.K. Giannakis (ed.), Encyclopedia of Ancient Greek Language and Linguistics, Leiden, 277–285. Clarysse, W. (1998), “Ethnic diversity and dialect among the Greeks of Hellenistic Egypt”, in: A.M.F.W. Verhoogt/S.P. Vleeming, The Two-faces of Graeco-Roman Egypt. Greek and Demotic and Greek-Demotic Texts and Studies presented to P.W. Pestman, Leiden, 1–13. Clay, J.S./Malkin, I./Tzifopoulos, Y.Z. (eds.) (2017), Panhellenes at Methone, Berlin. Crespo, E. (2011), “OF fr. 496: dialectal diversity in Macedonia at the end of the fourth century B.C.”, in: M. Herrero de Jáuregui et al. (eds.), Tracing Orpheus: Studies of Orphic Fragments in Honour of Alberto Bernabé, Berlin/New York, 227–230. Crespo, E. (2018), “The softening of obstruent consonants in Macedonian”, in: G.K. Giannakis/ E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, Berlin/New York, 329–348.

Dialects in Contact in the Ancient Kingdom of Macedon  

Crespo, E. (2019), “El papiro de Derveni en su contexto dialectal”, in: J. Piquero/P. De Paz/ S. Planchas (eds.), Nunc est Bacchandum. Homenaje a Alberto Bernabé, Madrid, 359– 368. Dubois, L. (1995), “Une tablette de malédiction de Pella: s’agit-il du premier texte macédonien?”, REG 108, 190–197. Dakaris, S.I./Vokotopoulou, I./Christidis, A.-F. (2013), Τα χρηστήρια ελάσματα της Δωδώνης των ανασκαφών Δ. Ευαγγελίδη, ed. S. Τselikas, indices by G. Papadopoulos, 2 vols., Athens. EAM: Rizakes, Th./Touratsoglou, Y. (1985), Επιγραφές Άνω Μακεδονία: Ελίμεια, Εορδαία, Νότια Λυγκηστίς, Ορεστίς: τόμος Α' κατάλογος επιγραφών, Athens. EKM 1: Gounaropoulou, L./Hatzopoulos, M.B. (1998), Επιγραφές Κάτω Μακεδονίας (Μεταξύ του Βερμίου όρους και του Αξιού ποταμού). Τεύχος Α’. Eπιγραφές Βεροίας, Athens. EKM 2: Gounaropoulou, L./Paschidis, P./Hatzopoulos, M.B. (2015), Επιγραφές Κάτω Μακεδονίας, Τεῦχος Β’, Μεταξύ του Βερμίου όρους και του Αξιού ποταμού, Μέρος Α’ - Β’, Ἀθήνα. Giannakis, G.K. (2020), “A note on the Pella curse tablet”, in: M. Fritz/T. Kitazumi/Veksina (eds.), Maiores philologiae pontes. Festschrift für Michael Meier-Brügger, Ann Arbor/New York, 62–67. Gounaropoulou, L./Hatzopoulos, M.B. (1998), Επιγραφές Κάτω Μακεδονίας (Μεταξύ του Βερμίου όρους και του Αξιού ποταμού). Τεύχος Α’. Eπιγραφές Βεροίας, Athens. Gounaropoulou, L./Paschidis, P./Hatzopoulos, M.B. (2015), Επιγραφές Κάτω Μακεδονίας, Τεῦχος Β’, Μεταξύ του Βερμίου όρους και του Αξιού ποταμού, Μέρος Α’ - Β’, Athens. Hatzopoulos, M. (1987a), “Artémis Digaia Blaganitis en Macédoine”, BCH 111, 397–412. Hatzopoulos, M. (1987b), ‘‘ΒΙΛΑΡΡΑ ΤΕΛΛΟΥ?’’, ZPE 68, 237–240. Hatzopoulos, M. (1991), Actes de vente d’Amphipolis, ΜΕΛΕΤΗΜΑΤΑ 14, Athens. Hatzopoulos, M. (1996), Macedonian Institutions under the Kings: A Historical and Epigraphic Study, I-II, ΜΕΛΕΤΗΜΑΤΑ 22, Athens. Hatzopoulos, M. (2000), “‘L’histoire par les noms’ in Macedonia”, in: S. Hornblower/E. Matthews (eds.), Greek Personal Names. Their Value as Evidence, Oxford, 99–117. Hatzopoulos, M. (2007), “La position dialectale du macédonien à la lumière des découvertes épigraphiques récentes”, in: I. Hajnal (ed.), Die altgriechischen Dialekte. Wesen und Werden, Innsbruck, 157–176. Hatzopoulos, M. (2011a), “Macedonian studies”, in: R.J. Lane Fox (ed.), Brill’s Companion to Ancient Macedonia. Studies in the Archaeology and History of Macedonia, 650 B.C.-300 A.D., Leiden/Boston, 34–42. Hatzopoulos, M. (2011b), “Macedonia and Macedonians”, in: R.J. Lane Fox (ed.), Brill’s Companion to Ancient Macedonia. Studies in the Archaeology and History of Macedonia, 650 B.C.-300 A.D., Leiden/Boston, 43–49. Hatzopoulos, M. (2018), “Recent research in the Ancient Macedonian dialect: consolidation and new perspectives”, in: G.K. Giannakis/E. Crespo/P. Filos, Studies in Ancient Greek Dialects. From Central Greece to the Black Sea, Berlin/Boston, 299–328. Hatzopoulos, M. (2020), Ancient Macedonia, Berlin/Boston. I.Leukopetra: Petsas, Ph./Hatzopoulos, M.B./Gounaropoulou, L./Paschidis, P. (2000), Inscriptions du sanctuaire de la Mère des Dieux autochthone de Leukopetra (Macedoine), MEΛETHMATA 28, Athens. Horrocks, G. (2010), Greek. A History of the Language and its Speakers, 2nd ed., Wiley. Kerswill, P. (2002), “Koineization and accommodation”, in: J.K. Chambers/P. Trudgill/ P. Schilling-Estes (eds.), The Handbook of Language Variation and Change, Oxford, 669– 702.

  Emilio Crespo Kerswill, P./Trudgill, P. (2005): “The birth of new dialects”, in: P. Auer/F. Hinskens/P. Kerswill (eds.), Dialect Change. Convergence and Divergence in European Languages, Cambridge, 196–220. Kotwick, M.E. (2017), Der Papyrus von Derveni (Introduction, German translation, and Commentary by M.E. Kotwick, and Text by R. Janko), Berlin/Boston. Kouremenos, Th./Parássoglou, G.M./Tsantsanoglou, K. (2006), The Derveni Papyrus, edited with Introduction and Commentary, Studi e Testi per il Corpus dei Papiri Filosofici 13, Florence: Leo S. Olschki. LGPN = Fraser, P.M./Matthews, E., A Lexicon of Greek Personal Names, IV Macedonia, Thrace, Northern Regions of the Black Sea, Oxford 2005. Lhôte, É. (2006), Les lamelles oraculaires de Dodone (Hautes études du monde gréco-romain 36), Geneva. Lilibaki-Akamati, L. (2007), Κιβωτιόσχημος τάφος με ζωγραφική διακόσμηση από την Πέλλα (= Cist-grave with painted decoration from Pella), Θεσσαλονίκη. Lulli, L. (2011), “La lingua del papiro di Derveni”, in: Papiri Filosofici. Miscellanea di Studi. VI Studi e testi per il corpus dei papiri filosofici greci e latini (STCPF) 16, Firenze, 91–104. Martín González, E./Paschidis, P. (2020a), A Supplement to Ἐπιγραφὲς Ἄνω Μακεδονίας, Athens. Martín González, E./Paschidis, P. (2020b), A Supplement to Ἐπιγραφὲς κάτω Μακεδονίας Α’: Επιγραφές Βέροιας, Athens. Masson, O. (1984), “Quelques noms de femmes en Macédoine”, ZPE 55, 133–136 (= Onomastica Graeca Selecta (OGS), vol. II, Paris, 1990, 417–420). Méndez Dosuna, J. (1985), Los dialectos dorios del Noroeste. Gramática y estudio dialectal, Salamanca. Méndez Dosuna, J. (2007), “Παρατηρήσεις στις νέες μαντειακές πινακίδες της Δωδώνης”, in: Studies in Greek Linguistics 27. Proceedings of the Annual Meeting of the Department of Linguistics, School of Philology, Faculty of Philosophy, Aristotle University of Thessaloniki, May 6–7. 2006. In Memoriam A.-F. Christidis, Thessaloniki, 277–285. Méndez Dosuna, J. (2012), “Ancient Macedonia as a Greek dialect: a critical survey on recent work”, in: G.K. Giannakis (ed.), Ancient Macedonia: Language, History, Culture, Thessaloniki, 133–145. Méndez Dosuna, J. (2014), “Macedonian”, in: G.K. Giannakis (ed.), Encyclopedia of Ancient Greek Language and Linguistics, Leiden, 392–397. Méndez Dosuna, J. (2018), “The language of the Dodona oracular tablets: the non-Doric inquiries”, in: G.K. Giannakis/E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, Berlin, 265–296. Nigdelis, P. (2017), Inscriptiones Thessalonicae et Viciniae: Supplementum Primum: Tituli inter A. MCMLX et MMXV Reperti, Berlin. Panayotou, A. (2007), “The position of the Macedonian dialect”, in: A.-F. Christidis (ed.), A History of Ancient Greek from the Beginnings to Late Antiquity, Cambridge, 433–443. Panayotou-Triantaphyllopoulou, A. (2007), “Macedonia among ancient literary tradition, 19th century debates and the new texts”, in: I. Hajnal (ed.), Die altgriechischen Dialekte. Wesen und Werden, Innsbruck, 385–392. Papakonstantinou-Diamantourou, D./Martín González, E./Hallof, K. (2021), Inscriptiones Thessalonicae et Viciniae: Supplementum Alterum: Addenda, Indices, Tabulae, Berlin.

Dialects in Contact in the Ancient Kingdom of Macedon  

Petsas, Ph./Hatzopoulos, M.B./Gounaropoulou, L./Paschidis, P. (2000), Inscriptions du sanctuaire de la Mère des Dieux autochthone de Leukopetra (Macedoine), MEΛETHMATA 28, Athens. Rizakes, Th./Touratsoglou, Y. (1985), Επιγραφές Άνω Μακεδονίας: Ελίμεια, Εορδαία, Νότια Λυγκηστίς, Ορεστίς : τόμος Α' κατάλογος επιγραφών, Athens. Roisman, J. (2010), “Classical Macedonia to Perdiccas III”, in: J. Roisman/I. Worthington (eds.), A Companion to Ancient Macedonia, Chichester, 145–165. Sens, A. (2004), “Doricisms in the new and old Posidippus”, in: B. Acosta-Hughes/E. Kosmetatou/M. Baumbach (eds.), Labored in Papyrus Leaves: Perspectives on an Epigram Collection Attributed to Posidippus (P.Mil.Vogl. VIII 309), Washington DC. Trudgill, P. (1986), Dialects in Contact, Oxford. Tzifopoulos, Y. Z. (2010–2013), “Χρυσά ενεπίγραφα επιστόμια, νέα και παλαιά”, HOROS 22–25, 351–356. Voutiras, E. (1998), Διονυσοφῶντος γάμοι. Marital Life and Magic in Fourth Century Pella, Amsterdam. Willi, A. (2012), “We speak Peloponnesian: tradition and linguistic identity in post-classical Sicilian literature”, in: O. Tribulato (ed.), Sikelia: Multilingualism and Cultural Interaction in Ancient Sicily, Cambridge, 265–288. Willi, A. (2014), “Creating ‘classical’ Greek: from fourth-century practice to atticist theory”, RFIC 142, 44–74. Worthington, I. (2009), “IG II2 236 and Philip’s common peace of 337”, in: L. Mitchell/L. Rubinstein (eds.), Greek History and Epigraphy. Essays in honour of P. J. Rhodes, Swansea, 213– 223. ΧΕΔ = Dakaris, S.I., Vokotopoulou, I. & Christidis, A.-F., Τα χρηστήρια ελάσματα της Δωδώνης των ανασκαφών Δ. Ευαγγελίδη, ed. by S. Τselikas, indices by G. Papadopoulos, 2 vols., Athens 2013.

This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 649307

Panagiotis Filos

Onomastic Formulae from N. Epirus and S. Illyria: Linguistic and Sociocultural Connotations Abstract: The numerous corpora of Greek inscriptions from the Greek colonies in S. Illyria (Epidamnos-Dyrrhachium, Apollonia), urban centers and other sites in N. Epirus, especially Bouthrotos, but also from a number of locations across the supposed “border zone” between N. Epirus and S. Illyria, provide us with invaluable linguistic information. In particular, the evidence demonstrates that Greek onomastic formulae (: “Greek name + Greek patronymic”) make up well over half of the total number of the data across these areas (ca. 69–94%), while onomastic formulae that include non-Greek (“Illyrian”) names, either exclusively or along with a Greek name/patronymic (“mixed”, i.e., bilingual formulae), are considerably fewer, especially in N. Epirus. The rather limited linguistic interplay between Greek and non-Greek onomastics, not only in the form of “mixed” onomastic formulae but also in morphological respects (e.g., suffixation, compounding, hybrid forms etc.) points to a more “static” picture of onomastic bilingualism in these areas in comparison to other bilingual regions of the Hellenistic world. However, the lack of non-Greek (“Illyrian”) texts, which would otherwise have facilitated some useful comparisons, points to the potential importance of various extra-linguistic factors (political, social, ethnic, cultural) that could be examined in future studies.

 Parts of this paper were presented in the context of various conferences and workshops over the past few years: Thessaloniki (2016), Madrid (2018), Rome (2019), London (2019). On the last three occasions I participated in my capacity as a member of the Madrid research group (PI: Prof. E. Crespo) of the EU-funded (HERA) MuMil project (www.mumil.eu) headed by Prof. A.C. Cassio. I take the opportunity here to thank the respective audiences of my talks for their helpful comments. Many more thanks are due to M. Benedetti, S. Kaczko, L. Conti, P. Guijarro, and above all, G.K. Giannakis, E. Crespo, and A.C. Cassio for amicable collaboration. https://doi.org/10.1515/9783110779684-005

  Panagiotis Filos

 Introduction Epirus and S. Illyria, two neighboring areas on the NW periphery of the classical world were marked by intersecting historical courses, notwithstanding the fact that they belonged to different political, sociocultural and linguistic spheres for a considerable period of time. The historically known populations of the two areas had probably come into contact since the pre-classical period.1 But in fact, both areas, particularly S. Illyria, largely remained outside the spotlight of ancient history until the Hellenistic period.2 In that respect, it is hardly surprising that modern scholarship often relies on external, notably (southern) Greek literary accounts, alongside some later, local epigraphic evidence written almost exclusively in Greek.3 The ethnic and linguistic makeup of these two areas, especially as regards the “border zone” between them was apparently not uniform, while modern research has often been marked by diverging views for both scholarly and nonscholarly reasons. In the absence of adequate historical accounts of the earliest periods, onomastics, despite its inherent shortcomings which often prevent it from providing clear answers to a number of (extra-)linguistic issues, can cast some light on various matters, such as (the perceived) linguistic identity, language contacts between the populations of the two regions etc.; in addition, onomastic evidence can also be informative about a number of concomitant issues, e.g., political relationships, cultural interactions etc.4

 1 Some sparse ancient historical information but also some modern archaeological evidence would suggest that the earliest contacts date back to the archaic period or even earlier (cf. Hammond 1967: 423–424; CAH III.1 (1982), especially Hammond (653–656) and — with some caution — Crossland (839–843); Wilkes 1992: 91–180). 2 As the historical accounts indicate, and some additional epigraphic and archaeological evidence seems to corroborate, Epirus became integrated in earnest into the historical timeline of the classical Greek world from the 4th cent. BCE onwards; on the other hand, S. Illyria remained a largely alien, yet not literally unfamiliar land given both its proximity to Corfu and Epirus (cf. Dodona, Ambracia) and, particularly, the presence of Greek colonies alongside its coastline (e.g., Epidamnos-Dyrrhachium, Apollonia) from about the late 7th cent. BCE onwards. Warfare between Illyrians on the one hand, and Epirotes and/or Macedonians on the other, did occur at times despite the fact that long-term peaceful contacts (e.g., commerce) seem to have been probably the norm. 3 On ancient accounts of Epirus see Filos (2018a: 216–219); note also Kokoszko/Witczak (2009) as well as Malkin (2001) and Davies (2002). On Illyria, see in particular Wilkes (1992). 4 Obviously, any linguistic characterization ought to be distinguished from ethnic identity/ affiliation since these two concepts do not necessarily overlap, especially in the context of

Onomastic Formulae from N. Epirus and S. Illyria  

 Epirus and S. Illyria Epirus and S. Illyria were two contiguous areas — the latter lying to the north of the former — located in the broader SW region of the Balkan peninsula, not too far away from the NW extremity of the classical Greek world proper.5 The geographic boundary between the two neighboring areas was not firmly established in antiquity; nonetheless, from about the 4th cent. BCE onwards the socalled Keraunia or Akrokeraunia mountain range, situated near the modern-day town of Ch(e)imara/Himara (ancient Chimaira?) in S. Albania, marked, according to most accounts, their conventional border.6 The population makeup of Epirus, at least in its northern part, was probably not entirely Greek if the (southern) Greek colonies are taken out of the overall picture.7 Ancient historians (e.g., Strabo 7.7.8) mention three major Epirote tribes (Molossians, Thesprotians, Chaonians), and alongside them, a number of minor tribes (e.g., Athamanians, Kassopaioi, Paroraioi, Tymphaioi). In addition, there were various epichoric tribes and clans in the north, notably Chaonia, some of which were of questionable linguistic affinity and/or ethnicity, e.g., Ἀτιντᾶνες, Βυλλίονες etc. The rather conflicting descriptions by ancient Greek historians may relate to this complexity, even though they often seem to be largely the product of indirect knowledge.8 Nonetheless, the current epigraphic evidence points to a Greek linguistic community in Epirus that spoke a NorthWest Doric (or North-West Greek) variety akin to those varieties spoken in neighboring areas (e.g., Aetolia, Acarnania, and partly, Macedonia).9  onomastics; hence, terms like “Greek” or “Illyrian” — the latter being a cover term anyway — will be used in a largely linguistic sense, without any ethnic implications necessarily. 5 S. Illyria largely corresponds to central and (parts of) N. and S. Albania nowadays, while Epirus forms the NW part of modern-day Greece, barring its northern area, notably Chaonia, which is part of modern-day S. Albania. 6 The geographical boundaries of Epirus on the other three sides were the Ambracian Gulf to the south, the Pindus mountain range to the east, and the Ionian(-Adriatic) Sea to the west (cf. Strabo 7.7.5; see Hammond 1967: 414 f.; Filos 2018a: 216–219). 7 See Filos (2018a: 216–220); cf. also Restelli (1969: 815 f.); Kokoszko/Witczak (2009: 10–16). 8 Some classical Greek authors with no immediate knowledge of the area like Thucydides seem to challenge the Greek character of the Epirotes; but in fact, they may simply refer to the more archaic institutions and customs, and particularly, to the less refined Greek variety of the Epirotes, as is also the case with other adjacent areas, e.g., Amphilochia, Acarnania, Aetolia, but also Macedonia (see Hammond 1967: 423). On the other hand, Strabo, refers to “bilingual” people to the west of Macedonia (cf. Wilkes 1992: 69; see also Filos 2018a; 2018b). 9 NW Doric was phonologically a Doris mitior variety with 7 long and 5 short vowels (see Filos 2018a: 227–230). Note also the presence of Greeks from the south (e.g., Corinthians, Eleans) in

  Panagiotis Filos S. Illyria, on the other hand, corresponds to the SW part of a broader and rather vaguely defined “Illyrian” territory that largely covered the western and central Balkan peninsula (cf. Strabo 7.5.1).10 Nonetheless, the linguistic data, particularly onomastics, suggest that different areas (“onomastic regions”) must be subsumed under the cover term “Illyrian territory”.11 In general, the word “Illyrian” is seen nowadays as a generic term that designated different populations and languages of the Balkans, but perhaps also people located — in later times? — in the Italian peninsula: note Messapic in the south and Liburnian in the northeast (see Katičić 1976: I.154–165; Wilkes 1992: 70 f.; Hammond/Wilkes 2012; De Simone 2018: 1868–1869).12 “Illyria proper” was more specifically defined by some Greek authors as a region lying to the northwest of Macedonia and to the north of Epirus (cf. Katičić 1976: I.154–157; Wilkes 1992: 70); its core part corresponded to the (broader) hinterland of the ancient Greek colony of Epidamnos (Lat. Dyrrhachium) and the surrounding areas (cf. Plin. NH 3.144: proprie dicti Illyrii et Taulantii et Pyraei; Pomponius Mela 2.56), but also to areas farther up to the north — cf. the last Illyrian capital located by Lake Scodra — as well as to the south.

 the colonies along the southern and western coastal areas of Epirus as well as the possible existence of some non-Greek people (“Illyrians”?) in the northern part of Epirus (Chaonia). 10 According to some ancient authors (cf. Appian Ill. 1), Illyria covered a vast area from the northern extremes of Epirus and Macedonia up to the recess, i.e., innermost part of the Adriatic Sea, on the one hand, and from the north-western boundaries of Thrace up to the Danube River, on the other; cf. also Hammond/Wilkes (2012), and particularly Wilkes (1992: 68–69): “from the head of the Adriatic Sea to the hinterland of the Gulf of Valona and ...] northwards as far as the eastern Alps and the Danube [...] and eastwards into some districts beyond the Balkan range”. Note also Dzino (2014). 11 On the basis of previous works, Wilkes (1992: 70 f.) groups together five different onomastic areas under the cover term “Illyrian”; “Illyria proper” represents just one of all these onomastic areas. On the other hand, De Simone (1993: 36–40, 50 ff.) distinguishes the “south-east onomastic region”, which is closer to “Illyrian proper”, from “middle-Delmatian”. Note also Rendić-Miočević (1993). 12 Krahe’s old “Illyrian” theory, which inter alia argued for an extensive “Illyrian” area that covered not only a large part of the Balkans but also large swathes of Italy where interrelated, yet different languages were spoken (cf. “Messapic” in Calabria and Apulia, Liburnian in the northeast) has been discredited over the past decades. Note also some relevant terms for the designation of different “Illyrian” onomastic regions or varieties: “southeast” (covering most of modern-day Albania as well as much of the Adriatic coast and its hinterland farther up to the north), “local” etc; and more recently, the so-called “onomastic regions”: Liburnian, central Delmatian, southeast Delmatian (cf. Wilkes 1992: 68–72; De Simone 1993: 36–37; 2018: 1868– 1869).

Onomastic Formulae from N. Epirus and S. Illyria  

S. Illyria was apparently the most prominent part of Illyria from a cultural point of view, partly thanks to the existence of nearby Greek colonies. S. Illyria was probably inhabited by different populations and was in contact with both Epirus and Macedonia, but also with the Greek colonies alongside its coast: note Apollonia and Epidamnos (Dyrrhachium). Bilingualism, and probably multilingualism at times, was a commonplace in those urban areas, especially in the Hellenistic and Greco-Roman periods.13 Onomastics points inter alia to people with Messapic names, be it Illyrians or Messapians proper (see De Simone 1993: 37–39); nonetheless, there also occur many Greek names as well as others, namely Celtic (e.g., Gaizatos/Gaizatorix), Latin (e.g., Marcus), (possibly) Thracian (Ταρουλας), (possibly) Semitic (Βαρναιος) etc. (cf. Masson 1993; Lhôte 2011); see also 6.4–6.7 below.14

 Greek and Illyrian Greek and Illyrian were cognate Indo-European languages, but it is hard to detect many linguistic affinities between them despite their geographical proximity in historical times. Ιn fact, Greek was an unmistakably centum language (cf., e.g., *ḱṃtom > Gk. ἑ-κατόν “100”, *g΄enh1– > Gk. γέν-ος “clan, stock, kin etc.”), whereas the status of Illyrian within Indo-European remains a moot point, although it might well have belonged to the same PIE sub-branch as Greek; note e.g., the name Gentius/Genthios which may relate to Lat. gens, gentis, but also to Gk. γένος. Nonetheless, the lack of adequate evidence — barring onomastics and a few attested glosses — renders any corroboration or refutation of this argument difficult.15  13 I shall not refer here to cases involving Latin names since the focus is on Greek and “Illyrian” in the context of onomastic formulae. Besides, Latin-related formulae must be seen in a different context, namely in relation to the expansion of Latin and the concomitant large-scale Greco-Latin bilingualism. On the use of Latin gentilicia in these inscriptions see, for instance, Wilkes (2019). 14 De Simone (1993: 38–39) soberly remarks that some names (e.g., Δάζιος) were probably of Messapic origin but came to be considered “Illyrian” over time, in the same way as everyone today considers Germ. Müller (cf. also Engl. miller) as a German noun without taking into account its ultimate Latin original form molinarius. 15 Primary evidence for Illyrian includes onomastics and glosses. (1) Onomastics: (i) Place names, e.g., Trieste (Tergeste) which may be related to Alb. tregë “market”(?). Note also a possibly relevant name (Pannonia): P. Domatius P. f. Tergitio negotiator, i.e., Tergitio “Merchant” (?), with its meaning assumed because of the presence of Lat. negotiator “merchant” next to it;

  Panagiotis Filos Besides their common Indo-European origin, Greek and Illyrian also shared some other linguistic features due to their prolonged, multilateral contacts (political, economic, cultural, but also warfare-related). On a linguistic level, there was a two-way influence, even though not on an equal basis: the epigraphic evidence indicates increasing Greek influence across S. Illyria from the Hellenistic times onwards, which must have led to widespread bilingualism.16 Greek influenced Illyrian in multiple ways, particularly in onomastics. There were also Greek loanwords into Illyrian, even though we can only figure them out indirectly; for instance, some early Greek loans in Old Albanian may be indicative of those ancient contacts, e.g., Alb. mokërë “millstone” < Dor. μᾱχανᾱ́ (?) (cf. Joseph 2018: 203). On the other hand, Illyrian too exerted some influence on Greek, especially in the lexicon: for instance, words like κράβατος “bed” (cf. also Lat. grabatus) may have derived from Illyrian (< *grab- “beech”).17 Moreover, there has been some speculation that the presence of voiced stops in certain words in Macedonian, e.g., γαβαλά “head” (gloss in Hesychius), and Epirote, e.g., δράμιξ “bread”,18 among other Greek dialects, could possibly be due to adstrate/substrate influ-

 however, this is not a firm interpretation (cf. Katičić 1976: I.172); (ii) Personal names, e.g., Teuta (“queen”?), Teutis etc.; cf. cognates like Germ. Deutsch, Eng. Tudor (< PIE *teut- “people”?); Deipaturos “Jupiter” (cf. Gk. Ζεῦ πάτερ (voc.), Lat. Jup(p)iter < PIE *Di̯eṷ- ph2ter-); Vescleves < PIE *u̯ esu “good” + *ḱleu̯ os “glory” (cf. Gk. εὖ + κλέος, i.e., Εὐκλεής, Skt. (va)su-śravas etc. (2) Glosses: there is a handful of safer Illyrian glosses: Δευάδαι, οἱ σάτοι ὑπ᾽ Ἰλλυριῶν (Hesychius); σάτοι may be related to σάτυροι “satyrs” or Σάτραι (Thracian tribe). Another gloss, ῥινός “mist, haze” (Illyrians), “cloud” (Oenotrians) may be related to Alb. (Geg) re (< ren) “cloud” and/or AGk. καλαρῖνες, ὀχετοί; cf. also ῥινοῦχος “underground channel” (< ῥίς “channel”). Moreover, sabaia/sabaium “a kind of beer” (cf. Germ. Saft, Eng. sap) and sybina “hunting spear” (unsafe): cf. AGk. συβίνη > Lat. sybina, Arm. səvīn (cf. Katičić 1976: I.165 f.; Wilkes 1992: 67–87; Tzitzilis 2007; 2014; Fortson 2010: 464–465; De Simone 2018). 16 Note also the characteristic Roman name of the province, i.e., Epirus noua, in contrast to the name reserved for Epirus (proper), i.e., Epirus uetus. Nonetheless, (most of) Illyria normally belonged to the province of Illyricum. 17 Cf. also the epithet Grabovius used for Jup(p)iter and Mars in Umbrian (e.g., voc. sg. crabouie in the Tabulae Iguvinae (VIa, l. 27)); in fact, this epithet seems to correspond semantically to the Greek epithet φηγωναῖος “of the beech, oak” (< φηγός “beech, oak”). Note also the form γράβιον “firebrand, torch” (cognate ?) which seems to have survived into some Modern Greek dialects, e.g., γράβος in Epirus, γάβρος in Arcadia (“hop-hornbeam”) etc. But this may alternatively/also relate to a Slavic loanword grabă into medieval Greek; see Tzitzilis (2007: 751). 18 These words are also considered by some scholars as potential Illyrian loanwords, even though this is far from certain.

Onomastic Formulae from N. Epirus and S. Illyria  

ence by some indigenous Balkan language(s), perhaps in older times (see Filos forthcoming).19

 Epirote and “Illyrian” onomastics Onomastics is a special field of the lexicon,20 primarily comprising personal and place names (toponyms), but also various other sub-categories.21 As far as ancient languages are concerned, onomastics is a primary source of information, especially in the case of poorly attested languages (Restsprachen) like Illyrian since there is often no other adequate evidence for a proper study of their linguistic makeup. By contrast, Greek onomastics has been very well-studied, even though in an uneven manner (e.g., Attic vs. NW Greek) regarding its various dialects. Illyrian is one of the most characteristic examples of ancient “rudimentary” languages. Notwithstanding the fact that are no surviving Illyrian inscriptions proper, there are at our disposal numerous “Illyrian” names, particularly from inscriptions — Greek, and to a lesser extent, Latin — and coin legends; cf. also some literary texts where a few historical Illyrian names occur too, e.g., Gentius/ Genthios, Teuta (royal names) etc.  19 The extent of the phonological impact of PIE Balkan languages on Greek is questionable, especially as regards the presence of apparent voiced stops in lieu of voiceless aspirates, e.g., Mac. Βερ(ε)νίκα in place of Gk. Φερενίκᾱ/-η; some of these cases may largely be due to later, internal developments (fricativization), while Pre-Greek words are also characterized by “unexpected” voiced stops, e.g., δαλάγχα alongside θάλαττα/-σσα “sea”. 20 Grammatical norms may often manifest themselves in a special manner in onomastics, especially as far as morphology, word-formation, morphosyntax and, particularly, semantics are concerned. In fact, “names are nouns with unique denotation, they are definite, have no restrictive relative modifiers, and occupy a special place in anaphoric relations” (van Langendock/van de Velde 2016: 38). For instance, proper names may not be modified by an article (determiner) in some languages, e.g., Eng. London (but The Hague), Paris (but The Paris you knew no longer exists). Moreover, compound names may display more often an occasional interchange between their two parts (e.g., Πάτρο-κλος vs. Κλεο-πάτρα, Θεόδωρος vs. Δωρόθεος) (cf. Schmitt 2018: 300). Above all, the semantics of proper names is often difficult or impossible to etymologize. For (ancient) Greek onomastics, in particular, see some general points in Neumann (2007), Forssman/Untermann (2008) and Filos (2021: 16–17). 21 Note also ethnics and other relevant terms (e.g., demonyms), patronymics and similar words (e.g., matronymics, andronymics), clan/family names of various forms (e.g., gentilicia), nicknames and affectionate names (including those for inanimate nouns), geographical names of different kinds etc. (cf. Hough 2016).

  Panagiotis Filos Illyrian onomastics is often divided into different categories which correspond to different geographic zones: south-eastern, which is considered Illyrian proper, central and northern (Tzitzilis 2007: 748). On the other hand, De Simone (1993) pointed out that one ought to differentiate between “local” and “nonlocal” names when it comes to (Greek) epigraphic material from S. Illyria and adjacent areas.22 In the former onomastic category, which he calls “southeastern”, belong Illyrian “proper” names, clearly or probably (e.g., Epikados, Genthios, Teuta(na), Tatta etc.). In the latter onomastic set belong names that are definitely, or most likely, of “non-local” provenance, namely Greek, Latin, but also “middle Delmatian” etc.;23 however, it is often difficult to distinguish this last category from Illyrian names proper (cf. De Simone 1993: 36 f.; 2018; see also Katičić 1976: I.178–184; Wilkes 1992: 71, 74 f.).24 On the other hand, Epirote personal names are still lacking a proper study from a linguistic point of view which would take into account all major corpora, notably Dodona (Molossia), Bouthrotos (Chaonia), Ambracia (from the classical period onwards).25 As a matter of fact, Epirote names seem to lack any major “idiosyncrasies”, especially in the context of NW Doric.26 On the other hand, there were also names borrowed from other dialects, including Macedonian (e.g., Βερε/Φερε-νίκᾱ), while koine features, for instance, endings, are still not much common until the late centuries BCE — early centuries CE (cf. Filos 2021). Finally, there  22 Note also some important points by De Simone (1993: 37): a. non-Greek names ought to be distinguished, on the grounds of adequate linguistic evidence, between “local” (“south-east”) and “non-local” ones (e.g., “mid-Delmatian” etc.); b. a “local” name is not necessarily “local” etymologically (e.g., Ἄβα, Ἄν(ν)α); c. a poorly known ancient language cannot be reconstructed exclusively on the basis of onomastics; d. the borders between different onomastic regions do not necessarily represent linguistic/dialectal borders. 23 These names occur in the western, central and northern parts of the Balkans, namely as far north as the Danube river but also around the innermost part of the Adriatic sea, i.e., close to modern-day Trieste and Venice. Venetic and Liburnian as well as the languages of Dalmatia and Pannonia are often co-listed with Illyrian. 24 “Illyrian” names are attested even in the Dodona oracular corpus, even though their number is meager, i.e., hardly a dozen names out of a total of ca. 1200 names (see Curbera 2013: 420); moreover, Πλάτωρ occurs in Ambracia too (Katsadima 2019: 170). 25 Personal names from Epirus amount to 3330 entries according to LGPN III.A (1997) (online: http://www.lgpn.ox.ac.uk/publications/vol3a/places.html) accessed 20.10.2021; but in fact, they correspond to fewer names since there are often multiple occurrences (: normally, different individuals) per name. Around 80% of the names are male and ca. 20% female (Matthews, in Cabanes et al. 1993: 177). 26 But note names like Θεύδοτος or Κλευπάτρα (Bouthrotos), with the characteristic mid-vowel “height dissimilation” (or, in an alternative phonological interpretation (J. Méndez Dosuna), “right-edge glide formation”).

Onomastic Formulae from N. Epirus and S. Illyria  

seems to be a preference for certain suffixes, especially in the formation of ethnic names: e.g., -ωπός (e.g., Κοιλωπός), -τός (e.g., Θεσπρωτός, Φοινατός), -αιος (e.g., Φαρναῖος), and especially in -ᾱνες, e.g., Ἀτιντᾶνες, Ἀθαμᾶνες, Ταλαιᾶνες, Ἀρκτᾶνες etc.; and, more generally about NW Greece, note Ἀκαρνᾶνες, Εὐρυτᾶνες etc. (see Filos 2018a: 240; cf. also Restelli 1969: 821–822; Schaffner 2014).

 Greek-“Illyrian” onomastic interaction . General Language contact and bi-/multi-lingualism phenomena in onomastics are harder to interpret properly than is usually the case in linguistics. The mere use of “foreign” proper names is not a safe criterion for the existence of bilingualism, since personal names tend to spread across languages and peoples: note, for instance, English names like Jane or Harry being in use around the globe nowadays. In fact, speakers of a given variety may adopt names from another language or dialect even if their overall attitude towards the source language and/ or its speakers is not really positive (cf. also Colvin 2004: 46). Obviously, there inevitably arise some issues when one attempts to study, in the context of a Greek epigraphic corpus the onomastics of an ancient “rudimentary” Balkan language like Illyrian. One may, inter alia, bring up the following points (cf. also De Simone 1993: 36–37): 1) How much legitimate is it to compare Greek vs. non-Greek names found in Greek texts, given inter alia the (possible) impact of Greek linguistic norms on “foreign” onomastic material (cf. also the use of the Greek alphabet)? 2) What are the appropriate criteria for the etymology of an “obscure” name? In other words, should an “alien” proper name be called “non-Greek” merely because it can hardly be etymologized within Greek? At any rate, Louis Robert demonstrated long time ago that this should not be the case (see Colvin 2004: 49; cf. also Parker 2013: 5–10). 3) What can the fluctuation in the use of Greek names among an indigenous language group tell us about socio-/extra-linguistic matters, such as the degree of bilingualism, ethnic identity issues, and also possible political, cultural and/or socioeconomic motives?27  27 Note also the role of local geography in the different predilections of the local population: for instance, Colvin (2004: 51–52) has pointed out that central Lycia was less amenable to Greek onomastic influence, at least in earlier times.

  Panagiotis Filos 4) How do the different types of adopted names relate to different ways of borrowing, i.e., should Greek “popular” / short names be linked to “bottomup” borrowing patterns, but literary Greek names to “top-down” bilingualism? Apparently not, since there are different possibilities.28 It goes without saying that the above points are too broad to address here; but they can be taken into account in terms of the Greco-“Illyrian” onomastic contacts as long as the available evidence can provide some clues. It is obvious that Greek, be it the Doric dialect of the Corinthian colonies along the S. Illyrian coastline or the NW Doric variety of Epirus, or even the Attic-Ionic koine at a later time (late Hellenistic period onwards), exerted a gradual, but clear influence on “Illyrian” onomastics (see § 6 below).

. “Illyrian” names in Greek context: interaction and linguistic norms The impact of Greek on “Illyrian” onomastics is more evident on the level of morphology, derivational and inflectional alike, but also in word-formation (cf. De Simone 1993: 42–44):29 1) Inflection: “Illyrian” names normally decline according to the Greek rules of inflection since they appear within a Greek text: for instance, Ἐπικάδου (gen. sg.) inflects like a Greek thematic (o-stem) form. In general, Doric inflectional patterns are prominent, obviously due to the Doric (Corinthian) character of the Greek colonies of S. Illyria, but also because of the local idiom of neighboring Epirus: (a) masc. gen. sg. -ᾱ (ᾱ-stem nouns), e.g., Λαιδία; (b) several feminine ᾱ-stem nouns end in -ᾱ (nom. sg.), even after a consonantal stem, although this may (also) have to do with the original “Illyrian” form, e.g., Ἀννίκα, Μαλλίκα, Τάτα. Over the course of time, namely in the last centuries BCE, Attic-Ionic (koine) endings in -η too make their appearance: note cases like Γενθιανη, Δαρδανη etc.30 Interestingly enough, indeclinable names are not a common phenomenon.

 28 For instance, a Greek name may be (mis-)associated, e.g., via folk-etymology and/or mythology, with a local name regardless of its course of borrowing. 29 Note also some possible onomastic interaction at the level of semantics, e.g., speaking names, folk etymology, puns; cf. also extra-linguistic factors, such as sociopolitical and/or cultural prestige etc. 30 The actual position of the accent is in fact unclear.

Onomastic Formulae from N. Epirus and S. Illyria  

2)

Derivation: the combination of “Illyrian” nominal stems with potentially Greek suffixes is very common: e.g., -ιώ (Δαζ-ιώ), -εας (Γενθ-έας), -ιας (Λαιδίας), -ις (Γενθ-ίς) etc. But obviously, one can hardly argue with certainty that all these suffixes were Greek exclusively. Moreover, there was some possible overlapping with Latin, i.e., we come across “Illyrian” names with potentially Latin (?) suffixes: e.g., -ανος (Γενθιανος), -ατος (Γαιζατος), -ηνα /-εννα (Μαδηνα, Βρικεννα), -ικος/α, -ινος/α (Μαλλίκα, Ἀμμίνα) etc. (cf. De Simone 1993: 42–45).31 3) Compounding: despite the presence of some (real or apparent) Illyrian compound names, e.g., Τευτίαπλος, Τριτύμαλλος, Τεμιταύτα, Τρίτευτα, there is very little evidence for cross-linguistic interaction in terms of compounding; note, however, a few possible cases like Tεισι-λαίδας/-λαΐδας (cf. Gk. Tεισίδᾱμος vs. “local” Scerdi-laedus/Sceru-ladus).32

 Onomastic formulae in N. Epirus and S. Illyria . General Onomastic formulae may simply be defined as the way(s) in which a person is referred to in a certain context (communicative, textual, sociolinguistic etc.). Despite the fact that the term “formula” may imply a rather stereotypical form of identification and/or address, onomastic formulae may be differentiated within any given linguistic and/or social framework according to a number of factors,

 31 On the suffixation of Greek personal names in general, but also in the context of language contact between Greek and other languages see Alonso Déniz et al. (2017). 32 Note, however, that alternative explanations cannot be dismissed a priori: for instance, the partial overlapping of Tεισιλαίδας with genuinely Greek names starting with Tεισι- (Tεισίδᾱμος, but also Τεισίμαχος, Τεισικράτης etc.) could in theory also be attributed to phonetic coincidence (plus folk-etymology perhaps ?): e.g., Πλάτωρ < Platur after Greek nouns in -ωρ (gen. -ορος), but also due to Gk. Πλάτων probably, with the three names forming a continuum: Πλάτ(o)υρ (“Illyrian”) – Πλάτωρ (adopted/hybrid?) — Πλάτων (Greek) (see De Simone 1993: 39, 60–61, 72–73; Curbera 2013: 420); cf. also fn. 61. On the other hand, the second member of Tεισιλαίδας, i.e., -λαΐδας/-λαίδας, which is often considered the adaptation of an Illyrian morpheme, could also have been (mis-?)associated with Greek names such as Τεισιλαΐδας, an augmented form of Τεισίλας (Thera) (cf. De Simone 1993: 40); note, similarly, names like Ἐχελαίδας, Μνᾱσιλαίδας (/-λαΐδας), Περιλαίδας etc.

  Panagiotis Filos namely class, status, sex/gender, age, period of time, social (or other) circumstances etc.33 In classical antiquity, there were many types of onomastic formulae, but normally, the bare minimum was the use of one, at least, “core” personal name, e.g., Gk. Σωκράτης, Λεωνίδας, Lat. Claudius. On most occasions, however, there were some additional names and/or accompanying modifiers/designations which facilitated the identification of a person, and its concomitant differentiation from other people with (quasi)-similar names: a second/third name, a patronymic,34 a family/clan/tribal/ethnic name, a demonym, a sobriquet, a person’s main profession, but also some title(s) and various other forms of modifiers could play this role; frequently, a combination of the above was also possible: e.g., Gk. Διόδωρος Θεοκρίτου Κεῖος, βασιλεὺς Σέλευκος Ἀντιόχου Νικάτωρ, Lat. Quintus Pompeius Rufus etc.35 Greek personal names, in accordance with the norms present in other languages of Indo-European descent, were normally modified by patronymics (normally in gen. sg., but also in nom. sg. in the form of patronymic adjectives), e.g., Ἀλέξανδρος Φιλίππου (Macedonia, and elsewhere), Πηληϊάδης Ἀχιλ(λ)εύς (cf. Homer), Δίκαιος Μνασιμάχειος (Thessaly). On many occasions, some other form of broader affiliation, such as demonyms / tribal names / ethnics could follow suit as well, e.g., Θουκυδίδης Ὀλόρου Ἀλιμούσιος (Attica) etc.

 33 Note, for instance, that in many modern societies it is far more common for a married woman to acquire her husband’s last name, in addition to or in replacement of her own family name, rather than the other way round. 34 Note, however, languages like Etruscan, which favored matronyms (mother’s names) next to personal names, at least from a certain period onwards, and Lycian where one may even find a grandfather’s name next to his grandson’s. 35 Latin naming patterns were different; Roman names were normally tripartite, at least as far as male names are concerned; in their standard form, Roman names included a praenomen, a nomen gentile/gentilicium (clan name), and a cognomen (tria nomina), e.g., Caius Iulius Caesar. In fact, Latin names shared their tripartite pattern with other ancient languages of Italy, such as Faliscan, the Sabellic languages, Etruscan etc. However, a full Roman onomastic formula for a man could also include some other elements, e.g., a father’s name, the name of one’s tribe etc.: e.g., M(arcus) Tullius M(arci) f(ilius) Cor(nelia tribu) Cicero; sometimes, the name of one’s original clan name (as an adjective) can be found next to one’s new name as an adoptee, e.g., Publius Cornelius Scipio Aemilianus Africanus, namely the son of a certain person from the Aemilia clan adopted later by a member of the Cornelia clan, and with the full name being complemented by the agnomen (“nickname”) Africanus. In later times, gentilicia went out of fashion while cognomina became more important (cf. Horrocks/Clackson 2007: 42–45; Maras 2020: 24–27).

Onomastic Formulae from N. Epirus and S. Illyria  

. Onomastic formulae in N. Epirus and S. Illyria: language, form, and other Onomastic formulae in Epirus may occur in different types of inscriptional texts: decrees, honorary texts, funerary inscriptions, manumissions etc.36 Names are often modified by a patronymic (in gen. sg.), e.g., Μαχάτας Άριστοφάνεος (Bouthrotos).37 On other occasions, an ethnic (or demonym) is also present, e.g., Ἀντάνωρ Εὐθυμίδου Χάων (Delphi), Κλεόστρατος Δαμάρχου Δυρραχῖνος (EpidamnosDyrrhachium); more rarely, this may be followed by as many as three different “ethnics”, i.e., a (major) tribe name, a minor tribe name (phylonym), and finally, a clan name, e.g., Σαβυρ[τίου Μολο]σσῶν Ὀνοπέρνου [Καρτά]του (all forms in gen.) (Dodona). On the other hand, a (male) person outside Epirus would normally have his name followed by the generic ethnic, normally in a non-Doric form, Ἠπειρώτης “Epirote”, e.g., Ἀλκέμαχος Χάροπος Ἠπειρώτης (Athens). On the other hand, Illyrian naming formulae were simpler, i.e., a name plus a patronymic (gen. sg.), e.g., Γένθιος Δαζαίου, Iettus Dasantis. Tripartite formulae (name, gentilicium, patronymic) appear mostly in Roman times, e.g., Aplis Staticus Aplinis f(ilius). In S. Illyria, in particular, onomastic formulae normally appear in funerary inscriptions, and as a consequence, they are accompanied by a salutation χαῖρε “(lit.) rejoice; hello”. On the other hand, in the west of the “Illyrian” continuum, i.e., Pannonia, Liburnia and elsewhere, we note a preference for a two-name formula (plus a patronymic and/or family name), e.g., Avita Suioca Vesclevesis, Cato Stataria, Tu(r)i f(ilia) etc. (cf. Wilkes 1992: 75 f.; Rendić-Miočević 1993: 122). The advent of the Roman era (ca. 2nd cent. BCE onwards) brought about a gradual change in the onomastics of both Epirus and S. Illyria. Apart from the use of Roman names as such, e.g., Μάρκος, Φλαβία Μαρκία etc., one may also notice the use of two- or three-name, mixed-language formulas, according to the Roman model, e.g., Φλάβιος Ἀθηναγόρας (Epirus), Ἰουνία Τιμ[ο]ν[ό]η, Κλαύδιος Γένθιος (Illyria). Obviously, the same phenomenon is much more noticeable in Latin inscriptions from the same areas, e.g., Memmi[us Ph]ronimus (Thesprotia, Epirus), Insteius Saturninus (Byllis, Illyria) etc. (cf. Wilkes 2019).

 36 A similar picture is presented in the epigraphic corpora from other parts of Epirus, notably Molossia and Thesprotia, including the Corinthian colony of Ambracia and the oracle of Dodona. Note, however, that in the oracular inquiries from Dodona, names belonging to non-local inquirers are sometimes modified by demonyms/ethnics, while patronymics may be absent. 37 Cf. the characteristic lack of contraction in NW Doric.

  Panagiotis Filos In general, the use of two or even three names in Greek inscriptions from Epirus is a rather rare phenomenon as regards both men and women, while in Illyria this is conversely a more common trend. Nevertheless, double (or triple) names descending from different languages can point to cross-cultural influences and onomastic trends, in the context of a Hellenized environment or a GrecoRoman multicultural milieu.

. Onomastic formulae of the type “name + patronymic” in N. Epirus and S. Illyria Onomastic data of linguistically mixed character may provide evidence for potential language contact, sociocultural interaction and convergence, or even for the possible development of multiple identities (ethnic, cultural, sociopolitical, and other). Obviously, such indications are more concrete on a larger scale rather than on an individual basis. In fact, possible cross-linguistic interaction in the field of onomastics may be highlighted in various ways: (a) hybrid forms; (b) double names, with either name having a different linguistic descent (e.g., Ἰουνία Τιμ[ο]ν[ό]η); (c) sequences of names of different linguistic background denoting different people; (d) alternative forms of the same name relating to different varieties; for instance, an Illyrian name occurring also in a more Hellenized/Latinized form (e.g., Πλάτυρ – Πλάτωρ); (e) in particular, onomastic formulae, especially of the type “name + patronymic” (and/or with ethnic designation(s) of different background(s)). Undoubtedly, onomastic formulae are the most pronounced manifestation of possible bilingualism and/or sociocultural and ethnic interaction since they may also bring to the fore the inter-generation parameter.38 In addition, onomastic formulae of this type may provide important evidence for possible intermarriages, sociolinguistic and cultural trends, ethnic attitudes, and other traits, with the proviso that the onomastic data are not taken at face value.39 In other words, the adoption of a name from a different lan 38 Ethnics may also provide some useful clues, especially when they point to a different linguistic background from that of the names (+ patronymics) that precede them. However, formulae of this type are far fewer in numbers, while the evidence they may provide for linguistic, cultural or other forms of possible interaction and convergence is less affirmative. 39 On the other hand, plain names can also show traits of language contact (cf., e.g., hybrid compound names, combinations of stems with suffixes from different languages); nonetheless, this is a rather “static” picture which cannot offer many other indications beyond some clues pertaining to bilingual morphology and word-formation.

Onomastic Formulae from N. Epirus and S. Illyria  

guage, which is normally given by (either of) the parents, might be an indication, but no proof, particularly in a multicultural society, of shifting identities, ethnic and/or social, even if the family as a whole were not bi-/multi-lingual.40 For instance, a Greek name accompanied by a patronymic in a native language (“Illyrian”), e.g., Ἀνδρίσκος Πρευράτου (in Klos, a town of the Koinon of Bylliones), could be only an indication of ongoing Hellenization; other interpretative options, however, are also available, i.e., a bilingual family, or even a rather positive, however uncertain as to its depth and motives, stance towards Greek culture etc.41 On the other hand, a native (“Illyrian”) name followed by a father’s name also in a native (“Illyrian”) language, e.g., Πλάτωρ Κλεβέτου, or even in Greek, e.g., Ἐπίκαδος Λύκου (both formulae from Epidamnos-Dyrrhachium), could be an indication of the opposite trend, i.e., growing ethnic self-confidence, in line with increased linguistic and cultural self-awareness and pride (cf. Colvin 2004: 54). It goes without saying, though, that none can rule out the possibility that nothing of the above may be true, at least on some occasions, and that it is all about current naming trends in the context of non-monolingual societies. One need not stress the fact that such phenomena were by no means unique to this broader area. The coexistence of Greek-speaking people with indigenous populations in the Balkans, Asia Minor, Italy, Sicily, and elsewhere, gave rise to similar phenomena in the form of onomastic formulae, and beyond. Hellenistic Lycia, for instance, is another interesting example of the confluence of languages, the formation of multiple identities, and the possible impact of those phenomena on onomastic data.42 Colvin (2004) examined inter alia the onomastic formulae from Lycia and drew some interesting conclusions which can be taken into account here, at least to some extent, with regard to the analysis of the data from S. Illyria, N. Epirus and the “border zone” between them.43  40 Obviously, the names of slaves and freedmen must be examined in a different context since these were often given by their (former) masters. 41 As Colvin (2004: 46) argues, “[o]ne can adopt foreign names for a variety of reasons before one has embraced the language they are associated with; and conversely, if a particular community is obliged to adopt a foreign language for all or part of the time, there is no guarantee of positive social attitudes towards the new idiom. […] the community can make a positive statement of its separate identity […] by retaining names redolent of its own national history or culture. […] this was not necessarily the same as retaining an indigenous name in Hellenized form”. 42 Lycia (SW Asia Minor) was home to a neo-Luwian language; cf. also neighboring languages, such as Carian, Pisidian. 43 Besides the linguistically mixed onomastic formulae, features of Greek-Lycian onomastic interaction may also be found in a number of phenomena and forms: (i) phonetic coincidence, which relates to phenomena of folk etymology, e.g., Πύρρος ~ Puresi, Ὀρθαγόρας ~ Urtaqija,

  Panagiotis Filos First of all, onomastic formulae in Lycia appear predominantly in funerary inscriptions as is also the case in S. Illyria (Apollonia, Epidamnos-Dyrrhachium), and to some extent, in N. Epirus and the “border zone” between Epirus and S. Illyria. On the other hand, there is a crucial difference between the two areas: Lycian names occur in both Lycian and Greek inscriptions found on Lycian territory, while in S. Illyria the epigraphic data originate exclusively from Greek inscriptions in Greek colonies. As far as the onomastic formulae are concerned, Lycian family naming patterns are similar to Greek: a single/compound name accompanied by a father’s name (both in gen. sg. and with a special suffix). In terms of statistics (numbers per category), Colvin draws a line between earlier (i.e., early Hellenistic period) and later (i.e., Hellenistic proper/early Roman periods) Lycian texts, some of which are bilingual (Lycian-Greek): (a) Early Lycian texts: ca. 80% Lycian-named parents with Lycian-name offspring; 1% Greek-named parents with Greek-named offspring; 9% Greek-named parents with Lycian-named offspring (in more “Hellenized” areas); 8% Lyciannamed parents with Greek-named offspring; 2% other (e.g., Persian); (ii) Hellenistic (proper) Lycian texts: 74% Greek-named parents with Greeknamed offspring; 9% Lycian-named parents with Lycian-named offspring; 12% Greek-named parents with Lycian-named offspring (in more “Hellenized” areas); 5% Lycian-named parents with Greek-named offspring. In later texts (late Hellenistic) figures are similar except for the category “Lycian-named parents with Greek-named offspring”, which rises from 5% to 11%, possibly due to awaken ethnic self-confidence (Colvin 2004: 54). Other factors, such as micro-geography, sex, and various extra-linguistic aspects may have played some role too. The analysis of the onomastic formulae from Lycia provides some important insights for the analysis of the data from S. Illyria (Apollonia, EpidamnosDyrrhachium), N. Epirus (Chaonia, including Bouthrotos) and the “border zone” between them, despite the differences between the two areas, but also between the respective corpora. But before I proceed to the examination of the epigraphic data from these areas some preliminary remarks are in order: (a) Personal names just like glosses cannot form a solid basis for reconstructing in earnest Restsprachen like Illyrian; even more, they cannot (dis)prove any

 Ὕπατος ~ Upazi, Ὄβριμος ~ Uppara. In addition, some neo-Luwian theophoric names beginning with Arm-/Erm-, i.e., Αρμα- (lunar god) are often confused with the Greek Ἑρμῆς; cf. also Artima- (and the like) vs. Ἄρτεμι-; (ii) possible calques: Μa ~ Ἀθήναιος, Νatrbbiyẽmi ~ Απολλόδοτος; (iii) adaptation to Greek rules of inflection: Purihimeti → Πυριματις (-ιος) (Colvin 2004: 58 f.; cf. also Brixhe 2013; Parker 2013: 5–6).

Onomastic Formulae from N. Epirus and S. Illyria  

facts about ethnic identity, especially when the data originate from a text written in another language. (b) Inscriptional evidence, especially when it is poorly preserved, is often a matter of editorial reading (and, possibly, restoration); in this respect, I have relied on the onomastic data provided by Cabanes et al. (1995–2020), but have left out incomplete forms, including names that are not fully restored. (c) It is occasionally difficult, with respect to feminine names in particular, to tell in the case of a pair of names whether the second name is a second feminine name, a patronymic (cf. Doric a-stem masc. gen. sg. -α), a gentilicium or even an obscure ethnic.44 Note also the presence of a few matronymics accompanying both male and female names (e.g., Ἀνδρείας Ἀριστοῦς), which may cause additional confusion. Finally, in some manumission texts the names of slaves and freedmen may sometimes be modified by the names of their (former) masters and mistresses rather than of their parents (cf. also De Simone 1993).45 (d) The linguistic background of a name may also be a thorny issue, especially when the name is rare and/or badly attested: Greek (normally (NW) Doric), “Illyrian”/“local”/“native” (i.e., non-Greek) or other (normally Latin, e.g., Μάρκος, Φλαβία; but rarely, something different, e.g., Iranian Φαρνάκης etc.). For the purpose of this study, I have relied, by and large, on Cabanes et al. (1995–2020), in conjunction with De Simone (1993), particularly his analysis and criteria for determining the non-Greek names from Apollonia and Epidamnos-Dyrrhachium; note, however, that these two viewpoints are not always congruent. In addition, I have also taken into consideration Wilkes (1992: 67– 87) and the studies in Cabanes (1993),46 despite any deviating approaches at times, while I have also consulted LGPN online, with an aim to find out whether a rare and obscure name occurs in Greek epigraphic texts from other areas too; in that case, one may normally be more confident that the name in question is probably not “local” (: non-Greek). On the other hand, I have not followed De Simone’s strict tripartite classification of the non-Greek names into “local south-east”, “possibly local southeast” and “non-local south-east” names since this is, on the one hand, a somewhat cumbersome categorization for the purpose of this study, and on the oth 44 The opposite case is also possible, namely a masculine name followed by another name, normally in -ας which need not be a second name or an “Aeolic” patronymic adjective in nom. sg.; other options, such as a matronymic gen. sg. in -ας or -ιος (e.g., fem. gen. sg. Εὐταξίας, Ἐπικτήσιος) as well as an ethnic (nom. sg. Ἀργεία) or a gentilicium (e.g., nom. sg. Γρανία), are possible at times. 45 Note also one probable occasion of a female name in neuter (a servant ?): Ἐλευθέριν. 46 Note also Katičić (1976: I.178–85).

  Panagiotis Filos er, problematic with regard to the third sub-category: in this group one may find, for instance, non-Greek names like Γέτα(ς), Πλάτ(o)υρ/Πλάτωρ alongside apparently Macedonian names like Ἀμύντας (/Ἀμέντας) as well as Latin(ate) ones like Βένετος. Instead, I have opted for only one, comprehensive group of non-Greek names, from which Macedonian, Latin and other “non-local” names have been excluded. On the other hand, the vague term “Illyrian” is used here in a broader sense rather than in too strict a manner. Last but not least, linguistically “mixed” onomastic formulae containing Latin names have also been excluded from the study since their presence would distort the main picture. The endpoint is that no absolute accuracy in numbers is really feasible. A statistical inaccuracy (ca. 10%, or even less, normally) for groups of some considerable size, i.e., with more than 10 forms, is possible, while for smaller groups such a divergence may be even higher. Nevertheless, statistical deviations of this size are of lesser importance in comparison to the significance of the emerging broader picture. (e) Finally, two otherwise noteworthy parameters, namely the role of both the type of document and the dating (period) of the inscriptions, have been relegated in importance. Document type and date have not been examined comprehensibly since there is a considerable qualitative discrepancy between the epigraphic material from the two Greek colonies in S. Illyria (Apollonia, Epidamnos-Dyrrhachium) and the other areas, in the sense that the former group is largely comprised of funerary inscriptions/epitaphs — in fact, this is almost entirely the case in Epidamnos-Dyrrhachium — with no reliable dating being available on many occasions;47 by contrast, in the other group (Chaonia, “border zone”), there is a larger variety of texts (honorary, funerary, decrees, epitaphs) which date to the (later) Hellenistic-Roman periods (particularly the last two centuries BCE — first two centuries CE).48 In that respect, a more detailed, comprehensive analysis that would attempt to bring into the equation the additional aspect of sub-periods would make the statistics of the study look more complicated, and the analysis more cumbersome, but still, without some noticeable benefits probably.

 47 Note, however, that in Cabanes et al. (1995–2020) there have also been included epigraphic texts from other areas (e.g., Dodona, Delphi, Olympia etc.) which refer to S. Illyria and N. Epirus (including “border territory”). 48 Note that some later names (Byzantine period) featuring in Cabanes et al. (1995–2020) have been excluded.

Onomastic Formulae from N. Epirus and S. Illyria  

. Apollonia Apollonia was a Greek colony in S. Illyria founded in ca. 600 BCE by Corinthian settlers. The corpus of the Greek inscriptions from Apollonia spans through the archaic, classical, Hellenistic, and Roman periods, but is practically confined to the last two periods. The Greek epigraphic texts (Cabanes et al. 1997 (vol. Ib)) include a large number of Greek names but also feature a sizeable set of non-Greek names, most of which are apparently or potentially “Illyrian”, e.g., Γένθιος, Τρίτος, Τευτέα (fem.) etc. In rough estimates, “non-Greek” names — in the sense of entries rather than of tokens — are around 5% of the total (vs. less than 15% in EpidamnosDyrrhachium, and below 5% in Bouthrotos and Chaonia), a number which is low, but not negligible. Onomastic formulae of the type name + patronymic (+ other designation) normally follow a monolingual pattern: alongside the numerous “Greek + Greek patronymic” formulae (e.g., Ἀνδρίων Ποδᾶ Μαχιάδας, Κλεινώ Φίλωνος) there are also “non-Greek + non-Greek patronymic” formulae (e.g., Kορέτα Πλάτορος, Τρίτος Πλάτορος), but also some mixed onomastic formulae, e.g., Νεικάνωρ Γενθίου, Kλέα Βαντίου, Ἡρακλείδᾱς Βαντίου. The survey of the inscriptional corpus from Apollonia renders a total of nearly 100 (more precisely, 97(+?)) onomastic formulae that fulfill the criteria for proper analysis as set out above.49 The classification of these formulae into separate sub-categories, according to the apparent linguistic provenance of the name and its modifier(s) (normally in gen. sg.),50 is as follows: 1) 80(+?) cases of the type “Greek name + Greek patronymic”; 2) 5(+?) cases of the type “non-Greek name + non-Greek patronymic”; 3) 3(+?) cases of the type “Greek name + non-Greek patronymic”; 4) 2 (+1) cases of the type “non-Greek name + Greek patronymic”.51

 49 In addition to the formulae (name + patronymic in gen. sg.) there are also a few cases in which the name and the patronymic appear in a different inflectional case and/or in different morphosyntactic roles (marked: +?). A few forms are difficult to classify firmly into any subcategory, and that is why there is an apparent numerical discrepancy between the total (97) and the added up numbers of the individual sub-categories. 50 The two names may also be found in different cases and in different morphosyntactic roles. 51 This small number of cases must be considered somewhat conventional since it essentially refers to a “local” name plus a Greek patronymic (Τρίτος Μουσαίου, Δρείγαλος (?), son of Χαρίτων). Nonetheless, note also the case of Ταρουλας Σωτα, which seems to be made up of a Thracian (?) name and a modifying Greek patronymic. Cf. also the following footnote.

  Panagiotis Filos Finally, there are more than half a dozen formulae which are hard to classify properly due to the obscure etymology of the names. The numbers show that over 80% of the formulae belong to type 1, even though a few names may in fact be the Hellenized forms of names from some other language, which need not be “Illyrian”.52 The prevalence of Greek names in the onomastic formulae is hardly a surprise for Greek texts from a Greek colony in S. Illyria.53 From a linguistic point of view, Greek names show, to a large extent, features of (NW) Doric morphonology; note, for instance, the prevalence of Doric endings (e.g., nom. sg. Ἀμφαλκίδας, Ἡρακλείδας; gen. sg. Γλαυκία, Αἰσχίνα) for a-stem masculine names; similarly, note Doric a-stem feminine names like Νικαρέτα. Nonetheless, there are also a few Attic-Ionic forms (e.g., Νεαγένης Τεισιδάμου, Στρατονίκη Ἀνδρονίκου, etc). At first sight, this Doric preponderance seems rather expected given both the Corinthian origins of Apollonia and the broader Doric-speaking area (Corcyra, Epirus) to the southern vicinity of the city. In addition, the slow advance of Attic-Koine into the broader area, including neighboring Epirus, down to the Roman period is also reflected in Apollonia. On the other hand, the other three categories (2–4) offer some additional indications, despite their apparently meager numbers. In particular, one may note the coexistence of Greek and non-Greek names within the same family: for instance, the “local” name Τρίτος, which features together with a Greek patronymic Μουσαίου in a formula (twice), reappears as a patronymic in another formula, in which the modified name is now Greek (Ἀγαθίων Τρίτου).54 According to the available extra-linguistic evidence, these onomastic formulae denote closely related people (see Cabanes et al. 1997 (vol. Ib): 89, 91). This example demonstrates the coexistence of Greek and non-Greek names within the same family. The role of sex/gender is also of some interest: formulae with male names represent slightly less than 70% of the total (75/97), with the rest (22/97) featuring a female name, a rather expected disproportion for an ancient society.  52 Note also the possibility that some names may originate from a substrate language preceding Greek and Illyrian (cf. Hatzopoulos 1993: 84). 53 It goes without saying that any large-scale study would require, beyond the detailed etymological analysis of obscure names -and still without any guarantee for indisputable etymological conclusions- the examination of extra-linguistic factors, including a prosopographic study that would lead to establishing possible family trees. 54 In fact, the formula appears twice, once on an urn and once more on a cippus. They both belong in the same sepulchral place from the Hellenistic period and obviously refer to the same person.

Onomastic Formulae from N. Epirus and S. Illyria  

Nonetheless, almost all of the formulae with a feminine first name belong to sub-category (1), except for the formula Τευτέα Ἐπικάδου (“Illyrian” name + “Illyrian” patronymic), which appears in a 2nd cent. BCE epitaph featuring names of a non-Greek family. Conclusions are not easy to draw here, since the numbers for the corresponding masculine forms are equally poor, however higher. Suffice it to say perhaps that in the social context of a Greek colony, people with native names, especially female, and probably also of “native” origin, were less likely to have risen to a high(er) social status that would allow them to feature in a (Greek) inscription. In conclusion, the onomastic stock of Apollonia presents an overwhelming Greek character and indicates a minimally cross-linguistic/-cultural picture. Note also the (near-)lack of mixed forms, e.g., forms comprising a Greek stem + “Illyrian” suffix. But overall, one may argue that this “monolingual” picture is probably a matter of exclusively Greek sources.

. Epidamnos-Dyrrhachium Epidamnos was co-founded by Corinth and its colony Corcyra in ca. 625 BCE. The city flourished thanks to its harbor, which virtually monopolized commerce with the Illyrian hinterland. From the late 3rd cent. BCE onwards it came under Roman control, with its new Roman name, Dyrrhachium, apparently the Latinized form of an alternative Greek name (or of a nearby place name), gradually prevailing over the theretofore name Epidamnos.55 The epigraphic corpus from Epidamos-Dyrrhachium is comprised almost entirely of funerary inscriptions (Cabanes et al. 1995 (vol. Ia)). The corpus is larger than that of Apollonia, and so is its onomastic stock, nearly 25% larger in terms of names (forms), but considerably more, i.e., roughly 50% in terms of overall attestations (tokens). As expected, the bulk of the names featuring in the Greek inscriptions are predominantly Greek. Nonetheless, the number of non-Greek names is higher here than in Apollonia, namely close to 15% of the total (but cf. the caveat of unsafe etymology for some names). Onomastic formulae are once again abundant, namely slightly fewer than 300 (in fact, ca. 280 if one excludes formulae that do not conform with the criteria

 55 Epidamnos-Dyrrhachium and Apollonia served from the mid-2nd cent. BCE as the two western terminal points of the via Egnatia, which stretched thence through the southern part of the Balkans all the way to Byzantium (and later, Constantinople) in the east.

  Panagiotis Filos of our study, e.g., formulae that contain Latin names). Nonetheless, hardly 250(+) of them can be classified with relative certainty: 1) 166(+) formulae of the type “Greek name + Greek patronymic”;56 2) 28(+) formulae of the type of “non-Greek name + non-Greek patronymic”; 3) 23(+) formulae of the type of “Greek name + non-Greek patronymic”; 4) 26(+) formulae of the type of “non-Greek name + Greek patronymic”. In statistical terms, forms per category are as follows: (1) 68.5%, (2) ca. 11.5%, (3) 9.5%, (4) ca. 10.5%. At first glance, there is a distribution that is (quasi-) similar to that in Apollonia, even though the numbers for categories (2)-(4) in Apollonia are too small to allow any proper conclusions. On the other hand, it is clear that the percentage for category (1) in Epidamnos-Dyrrhachium is clearly smaller than in Apollonia (68.5% vs. 80%), even though it represents more than half of the total. Nonetheless, “Illyrian” names followed by “Illyrian” patronymics are not a negligible group (11.5%), even though one ought to note here that the majority of them (15 out of a total of 28 names) are feminine names (see below). On the other hand, mixed-language onomastic formulae, which include one non-Greek name, are almost evenly split between those that have a nonGreek patronymic after a Greek name (9.5%, class (3)) and those that have a non-Greek name followed by a Greek patronymic (ca. 10.5%, class (4)). The considerable number of non-Greek names in general is reflected in their presence in three different categories (2, 3, 4), but appears more pronounced in group (4) (“non-Greek name + a Greek patronymic”) which includes (at least) 10 cases of male non-Greek names plus a Greek patronymic as well as (at least) 16 female non-Greek names plus a Greek patronymic. Uniformity in the type of documents (funerary inscriptions) and the lack of any (more) precise chronological indications for many of the inscriptions prevents us from exploring in earnest the possible role of these factors in the statistical makeup. On the other hand, there is some other lead of investigation, namely the role of sex/gender, which must eventually relate to the type of the documents (funerary) too: 141 formulae with male names vs. 102 with female names, i.e., ca. 58% vs. 42%. It is clear that the percentage of formulae with female names is considerably higher than in Apollonia (ca. 70% vs. 30%) but, the actual reason may simply be the uniform type of the documents (funerary) in Epidamnos-Dyrrhachium, in contrast to a larger variety of document types in Apollonia.  56 Greek names show Doric morphonological traits, at least to some degree; note in particular names belonging to the characteristic paradigm -ην, gen. -ῆνος.

Onomastic Formulae from N. Epirus and S. Illyria  

However, there is a more interesting aspect here, i.e., the ratio between male- and female-name formulae per category: (1) 105 vs. 61; (2) 13 vs. 15; (3) 13 vs. 10; (4) 10 vs. 16. It is obvious that female “Illyrian” names, regardless of the linguistic background of the modifying patronymic (classes (2), (4)), are more numerous than male “Illyrian” names. Numbers are neither too high to be taken at face value nor too low to be ignored; a prompt recourse to serendipity would probably sound as an effort to dodge the problem. It is obvious than non-Greek female names (Τευτα etc.) show a greater tenacity or at least popularity and that the discrepancy between them and non-Greek male names can be interpreted in several possible ways, which may largely be due to extralinguistic reasons: conservatism, limited adaptation on behalf of non-Greek women to Greek social and cultural norms, less advanced cultural and/or social mobility, local cultural prestige, possible intermarriages, superficial similarity to Greek names (including adaptation) etc.

. Chaonia (and Bouthrotos) Chaonia represented the northern part of ancient Epirus, which extended beyond the river Thyamis (modern-day Kalamas) up to the Keraunia mountains. Among its major urban centers were the capital Phoenice and the regional center of the Koinon of Prasaiboi, the city of Bouthrotos (or Bouthroton), opposite Corcyra, which in the late 1st cent. BCE became a Roman colony (Buthrotum). The epigraphic material from Chaonia (Cabanes et al. (2016 (vol. III)), including Bouthrotos (Cabanes et al. 2007 (vol. II)), includes a considerable set of Greek names. Non-Greek names — excluding Roman names — are few, but still, more numerous than in most other regions of ancient Epirus (Molossia, Thesprotia etc.). The same holds true for ethnics since all but one minor tribal name (cf. Ill. Τέμουοι) seem to be Greek (cf. Lhôte 2011: 110). The epigraphic corpus from Bouthrotos is remarkably large and diverse, and includes various types of documents, among which numerous manumissions. Funerary texts here do not overshadow the other types of documents in the corpus, which stands in stark contrast to the corpora from Apollonia and Epidamnos-Dyrrhachium. This fact may also account for some of the qualitative differences observed in the onomastic data, alongside the fact that Bouthrotos was an Epirote city rather than a remote Corinthian colony in S. Illyria. As far as onomastic formulae are concerned, there are around 400(+) onomastic formulae, which almost in their entirety are of the type “Greek name + Greek patronymic”. Formulae with non-Greek names (excluding Latin names) are extremely rare, e.g., Λέων Δέρδα (Greek + “Illyrian” patronymic).

  Panagiotis Filos The other parts of Chaonia have yielded considerably fewer inscriptions, even though in a variety of types: decrees, manumissions, dedications, official texts, funerary inscriptions etc. Consequently, the onomastic evidence is less abundant, while the number of (well-attested) onomastic formulae is meager, namely 33. These can be classified as follows: 1) 31 cases (22 masculine + 9 feminine) of the type “Greek name + Greek patronymic”; 2) 0 cases of the type “non-Greek name + non-Greek patronymic”; 3) 1 case of the type “Greek name + non-Greek patronymic”; 4) 1 (possible) case of the type “non-Greek name + Greek patronymic”. The overall picture points to a clearly onomastic Greek character for this area: formulae of the type “Greek name + Greek patronymic” amount to ca. 94% of the total. From an onomastic point of view, some names show a local, or at least “non-Attic-Ionic” character, even though their morphonological features are not necessarily Doric: e.g., Ἀλκέμαχος Χάροπος Ἠπειρώτης (in lieu of Dor. Ἀπειρώτας), Ἀντάνωρ Εὐθυμίδου (instead of Dor. -α) Χάων, Κλεοφά[νης Ἀλυ]πήτου Χάων Πευκεστός, Λυσίμαχος Φιλίππου, Ἀλκινόα Γλαυκία (Dor. -α endings), Φιλίστα Σωστρίωνος (note -α, but also , instead of an expected Macedonian and (an alleged) Epirote) ) etc. Overall, the Greek onomastic formulae represent a linguistic mixture of epichoric and Attic-Ionic koine features. On the other hand, it is remarkable that there is no well-preserved formula of the type “non-Greek name + non-Greek patronymic”. The very few formulae that feature non-Greek names are [Νίκ]ανδρος Ἐπικάδου (Greek name + “Illyrian” patronymic), and perhaps, Πραυγισσος Νικοδάμου (“Illyrian” (?) name + Greek patronymic).57 Finally, it is worth mentioning that out of a total of 34 formulae only 9 include feminine names, which must once again be related to the type of the documents (decrees, manumissions, and the like rather than exclusively funerary texts).

 57 The name Πραυγισσος (uncertain position of accent) occurs only 4 times in the LGPN (online) corpus: 3 times in this (broader) area and once more in Lissos farther up to the north, in “Illyria proper”. The name seems to be morphonologically adapted to Greek but its origin is most likely non-Greek.

Onomastic Formulae from N. Epirus and S. Illyria  

. The “border zone” between N. Epirus and S. Illyria The linguistic border between N. Epirus and S. Illyria was neither fixed over the course of time nor clearly defined, as is normally the case in matters of language contact generally. An intermediate “border zone” characterized by the coexistence of Greek and non-Greek (“Illyrian” and perhaps other) speakers, even in unequal numbers, was most probably the reality until the late Hellenistic period. I will refrain here from discussing in detail the extent of bilingualism in this area or the (probable) picture of its linguistic microgeography. Instead, I will adopt, for practical reasons too, the proposal by Cabanes et al. (2016 (vol. III)) without necessarily subscribing in full to their views, especially as regards sites to the south of Orikos.58 The multitude of towns/cities (Amantia, Antipatreia, Byllis, Dimale, Himara, Orikos, Olympe, Aulon, the port at modern-day Grammata bay situated at the foot of Mt. Akrokeraunia), but also the number of Koina/Leagues (Bylliones, Balaiitai), both on the Ionian coast and in the hinterland, have offered us a rather small but important corpus of inscriptions, part of which has survived down to modern times. The number of (well-attested) onomastic formulae is once again limited, but there are almost twice as many inscriptions as in Chaonia, namely 63. These can be classified as follows: 1) 53(+1) formulae (45 masculine, 8(+1) feminine) of the type “Greek name + Greek patronymic”; 2) 2 formulae (both masculine) of the type “non-Greek name + non-Greek patronymic”; 3) 3 (possible) formula(e) (all masculine) of the type “Greek name + non-Greek patronymic”;

 58 Note that in Cabanes et al. (2016 (vol. III)) there is also a separate section on Greek inscriptions found in central and north Albania (e.g., Tirana, Pogradec etc.) which is not examined here since they fall outside the geographical area proper of this study. Suffice it to say perhaps that the small number (19) of the onomastic formulae would hardly allow any firm conclusions. In fact, there are 9(+1) formulae (7(+1) masculine, 1 feminine) of the type “Greek name + Greek patronymic”, 7 formulae (4(+1) masculine, 2 (possible) feminine) of the type “non-Greek name + non-Greek patronymic”, 2 formula(e) (both masculine) of the type “Greek name + non-Greek patronymic”, and finally, no formulae of the type “non-Greek name + Greek patronymic”.

  Panagiotis Filos 4) 1(+2) formula(e) (all masculine) of the type “non-Greek name + Greek patronymic”.59 The formula type “Greek name + Greek patronymic” is by far the most common once again, namely ca. 85% (cf. a quasi-similar percentage (80%) in Apollonia (6.4)), i.e., slightly less than in nearby Chaonia (94%). As far as categories (2)(4) are concerned, absolute numbers (occurrences) per category are larger than those in Chaonia, but rather insignificant in terms of percentage rates. From an onomastic point of view, the picture is similar to that in Chaonia: names are akin to the area of Epirus, i.e., not (necessarily) “Attic-Ionic”, even though they do not always have Doric morphonological features: e.g., Ἀρχέλαος Ἀλεξάνδρου, Αἰσχίνας Σιμία (cf. Doric endings -ας, -α), Κράτιλλος (sic) Μαχάτα, Κόσμιλλα Ἀλεξάνδρου, Πανταρέτα Δεξάνδρου; but note also some forms with Attic-Ionic features, e.g., Σωτίχη Ὀνησίμου. Overall, Doric-flavored names prevail but are not the exclusive members of the set. As far as formulae with non-Greek names (types (2)-(4)) are concerned, numbers are once again low and concern male names exclusively: two formulae of the type “non-Greek name + non-Greek patronymic” (Πλάτωρ Ἐπικάδου, [Πρ]αυγαῖος Πραυγίσ[σου]), 2 (possible) formulae of the type “Greek name + non-Greek patronymic” (Ἀρνέας Πραούγου, Κλέανδρος Πραυγίου),60 and 1(+2) formulae of the type “non-Greek(?) name + Greek patronymic” (Τράσος Ἱέρωνος, and possibly [Πρα]υγίμμας Βοΐ[σκου], [Πρα]υγίμμας Νικαίου).61

 59 Note that the formula Κρίσων Σαβυρτίου, which occurs in an honorary decree at Dodona (3rd cent. BCE), has been left out of all these categories since on the one hand, it is difficult to be classified with regard to the linguistic provenance of the patronymic, while on the other, it refers to a Molossian person, Κρίσων. 60 The patronymic names in this group had better be dubbed as potentially “non-Greek”; none of them appears in other areas according to the LGPN (online) corpus. Unfortunately, I have not been able to take advantage of the addenda-corrigenda in Cabanes et al. (2020 (vol. IV)) since I was unable to consult it in time; in any case, addenda-corrigenda have not been exploited systematically in previous volumes either. I also regret the fact that LGPN-ling (https://lgpn-ling.huma-num.fr/), the fraternal site of LGPN, which provides morphological/ semantic analysis of Greek (and not only) names, was launched too late to be used for the purpose of this study. 61 Note here the interesting case of the Greek name Πλάτων (followed by the patronymic Ἀνθρωπίσκου), which is the name of a prytanis featuring in a Hellenistic inscription from Byllis. This is a rather rare occurrence of the name which resembles the “Illyrian” masculine names Πλάτωρ (as a Hellenized hybrid?) and Πλάτ(ο)υρ, but also the cognate feminine name Πλάτουρα. These “Illyrian” names occur both in this area and in Chaonia, but also in Apollonia and Epidamnos-Dyrrhachium. The name Πλάτων occurs in an oracular inquiry at Dodona, and it has

Onomastic Formulae from N. Epirus and S. Illyria  

Percentage %

Finally, it is worth mentioning that out of a total of 63 formulae only 8(+1) include feminine names, which must once again be ascribed to the various types of the documents they appear in, namely decrees and other types of inscriptions rather than exclusively funerary texts.

EpidamnosDyrrhachium Apollonia ‘Border zone’ Chaonia

Language-related types of onomastic formulae (NB: Gk = Greek, Ill. = ‘Illyrian’, n. = name, pat. = patronymic Fig. 3: Onomastic formulae in N. Epirus and S. Illyria.

 Conclusions The study of the onomastic formulae of the type “name + patronymic” that feature in the Greek epigraphic corpora from N. Epirus, S. Illyria, and the “border zone” between them allows some general remarks.62 First of all, both Greek names and Greek onomastic formulae (: “Greek name + Greek patronymic”) dominate as one would perhaps expect; Greek onomastic formulae make up more than half of the total number of the attested formulae across the areas examined above (ca. 69–94%). On the other hand, onomastic formulae including non-Greek (“Illyrian”) names, either exclusively or together with a Greek

 been speculated that it may be the Hellenized version of an Illyrian name Πλάτ(ο)υρ/ Πλάτωρ (cf. Curbera 2013: 420); cf. also fnn. 24, 32. 62 A much lengthier and more detailed study ought to examine the evidence from the onomastic formulae in conjunction with a comprehensive linguistic analysis of the individual personal names, Greek and non-Greek alike, ideally in a way that would also take into consideration prosopographic, historical and other aspects.

  Panagiotis Filos name/patronymic (i.e., “mixed” formulae), are much fewer and normally appear in numbers (percentages) that are not sharply different. The evidence suggests that the farther away from the Greek world one moves (cf. the north-most colony of Epidamnos-Dyrrhachium) the smaller the number of Greek epigraphic formulae, and conversely, the larger the number of the formulae with non-Greek names, in full or partly. But one may wonder whether this is really the case, or it is, perhaps, a matter of false appearances, at least to some extent. Obviously, there is hardly a firm answer: the makeup of a corpus, i.e., the prevailing type(s) of documents must also play some role; note that in the Greek epigraphic corpus from Epidamnos-Dyrrhachium, which is almost entirely made up of funerary inscriptions, both formulae with non-Greek names and formulae with feminine names appear in much higher numbers than anywhere else. In a broader manner, one may ask with reason whether the rather limited interplay between Greek and non-Greek onomastics, be it in the sense of “mixed” onomastic formulae or in other respects (suffixation, derivation, compounding, calques, hybrids, folk etymology etc.) testifies to a somewhat more “static” picture of onomastic bilingualism in this broader area in comparison to other areas (e.g., coastal Asia Minor). Of course, we lack Illyrian texts here (cf. Lycian texts in Lycia), which would offer us a better picture. In any case, population dynamics, the language equilibrium, but also political, social and cultural aspects, including identity issues must have played a crucial role too. Finally, it is worth adding a more general remark about a potential discrepancy between the Hellenistic and the Roman periods with regard to the use of the two “world languages”, i.e., Greek and Latin, in the sphere of onomastics: “in the Latin-speaking provinces it appears to have been the practice of many local communities assimilated to the Roman way of life at first to record their local nomenclature but after a generation or so they would discard these in favor of Latin names, in particular when the prospect of Roman citizenship existed” (Wilkes 2019: 175). The adoption of Roman names points inter alia to extralinguistic motives too. On the other hand, one must also take note of a different tendency, namely that “...non-Greek names massively survive long after the adoption of Greek as a communication medium by a non-Greek population group” (Hatzopoulos 2020: 72). Perhaps, we should recall here Fishman’s (1967) classical term “bilingualism with diglossia” and (re-)consider it now with special regard to onomastics, notably to different spheres of usage.63

 63 Note also the studies in Parker (2019) and in Clackson et al. (2020).

Onomastic Formulae from N. Epirus and S. Illyria  

Bibliography Alonso Déniz, A. et al. (eds.) (2017), La suffixation des anthroponymes grecs antiques, Genève. Brixhe, C. (2013), “Anatolian anthroponymy after Louis Robert…and some others”, in: R. Parker (ed.), Personal Names in Ancient Anatolia, Oxford, 15–30. Cabanes, P. (1976), L’Épire de la mort de Pyrrhos à la conquête romaine (272–167 av. J.C.), Paris. Cabanes, P. (ed.) (1993), Grecs et Illyriens dans les inscriptions en langue grecque d’Épidamnos-Dyrrachion et d’Apollonia d’Illyrie, Paris. Cabanes, P. et al. (eds.) (1987–2004), L’Illyrie méridionale et l’Épire dans l’antiquité, 4 vols. (1987, 1993, 1999, 2004), Clermont-Ferrand/Paris. Cabanes, P. (1995–2020), Corpus des inscriptions grecques d’Illyrie méridionale et d’Épire, 4 vols. (1995–1997 (Ia-Ib), 2007 (II), 2016 (III), 2020 (IV)), Athens/Paris. CAH: The Cambridge Ancient History. 1970–2005. 14 vols., 2nd/3rd ed., Cambridge. Clackson, J. et al. (eds.) (2020), Migration, Mobility and Language Contact in and around the Ancient Mediterranean, Cambridge. Colvin, S. (2004), The Greco-Roman East: Politics, Culture, Society, Cambridge. Curbera, J. (2013), “The personal names”, in: S. Dakaris/I. Vokotopoulou/A. Christidis (eds.), Τα χρηστήρια ελάσματα της Δωδώνης των ανασκαφών Δ. Ευαγγελίδη, vol. II, Athens, 419–432. Davies, J.K. (2002), “A wholly non-Aristotelian universe: the Molossians as ethnos, state and monarchy”, in: R. Brock/S. Hodkinson (eds.), Alternatives to Athens: Varieties of Political Organization and Community in Ancient Greece, Oxford, 234–258. De Simone, C. (1993), “L’elemento non greco nelle iscrizioni di Durazzo ed Apollonia”, in: P. Cabanes (ed.), Grecs et Illyriens dans les inscriptions en langue grecque d’ÉpidamnosDyrrachion et d’Apollonia d’Illyrie, Paris, 35–75. De Simone, C. (2018), “Illyrian”, in: J. Klein/B. Joseph/M. Fritz, in cooperation with M. Wenthe (eds.), Handbook of Comparative and Historical Indo-European Linguistics, vol. III, Berlin/ Boston, 1867–1872. Dzino, D. (2014), “Illyrians in ancient ethnographic discourse”, Dialogues d’histoire ancienne 40.2, 45–65. Filos, P. (2018a), “The dialectal variety of Epirus”, in: G.K. Giannakis/E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects. From Central Greece to the Black Sea, Boston/Berlin, 215–247. Filos, P. (2018b), “Linguistic aspects of epirote ethnics”, in: A.J. Domínguez (ed.), Politics, Territory and Identity in Ancient Epirus, Pisa, 283–301. Filos, P. (2021), “Reflexes of koineization in ancient epirote feminine names”, in: G.K. Giannakis et al. (eds.), Synchrony and Diachrony of Ancient Greek: Language, Linguistics, and Philology. Essays in Honor of Emilio Crespo, Berlin/Boston, 11–29. Filos, P. (forthcoming), “Some remarks on ancient epirote glosses”, in: G.K. Giannakis et al. (eds.), Classical Philology and Linguistics: Old Themes and New Perspectives, Berlin/ Boston. Fishman, J.A. (1967), “Bilingualism with and without diglossia. Diglossia with and without bilingualism”, Journal of Social Issues 23.2, 29–38. Forssman, B./Untermann, J. (2008), “Onomastics”, in: M. Landfester (ed.; Eng. ed.: Gentry, F. G.), Brill’s New Pauly (Classical Tradition), vol. IV, Leiden/Boston. Fortson, B.W. (2010), Indo-European Language and Culture. An Introduction, 2nd ed., Oxford/

  Panagiotis Filos Malden, MA. Hammond, N.G.L. (1967), Epirus. Τhe Geography, the Ancient Remains, the History and the Topography of Epirus and Adjacent Areas, Oxford. Hammond, N.G.L./Wilkes, J. (2012), “Illyrii”, in: S. Hornblower/A. Spawforth/E. Eidinow (eds.), The Oxford Classical Dictionary, 4th ed., Oxford/New York. Hatzopoulos, M.B. (1993), “Anthroponymie grecque et non-grecque dans les inscriptions de Dyrrhachion et d’Apollonia”, in: P. Cabanes et al. (eds.) (1987–2004), L’Illyrie méridionale et l’Épire dans l’antiquité, 4 vols. (1987, 1993, 1999, 2004), Clermont-Ferrand/Paris, 81–87. Hatzopoulos, M.B. (2020), Ancient Macedonia, Berlin/Boston. Horrocks, G./Clackson, J. (2007), The Blackwell History of the Latin Language, Oxford/Malden, MA. Hough, C. (ed.) (2016), The Oxford Handbook of Names and Naming, Oxford. Joseph, B.D. (2018), “Linguistic contact in the ancient Balkans: a Sprachbund, or something else?”, in: G.K. Giannakis/E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, Berlin/Boston, 197–211. Katičić, R. (1976), Ancient Languages of the Balkans, 2 vols., The Hague – Paris. Katsadima, I. (2019), “Ambracian onomastics: problems and perspectives”, in: J. Lamboley/ L. Përzhita/A. Skenderaj (eds.), vol. I, 165–172. Kokoszko, M./Witczak, K.T. (2009), “Stosunki etnolingwistyczne w starożytnym Epirze [= Ethnolinguistic Relations in Ancient Epirus]”, in: R. Sajkowskiego/M. Wolnego (eds.), Grecja, Kartagina, Rzym [= Greece, Carthage, Rome], vol. I, Olsztyn, 9–26 [in Polish, with an English summary; an earlier, English version (1991): “Ancient Epirus and its inhabitants”, Linguistique Balkanique 34.1–2, 41–49]. Lamboley, J./Përzhita, L./Skenderaj, A. (eds.) (2019), L’Illyrie méridionale et l’Épire dans l’antiquité VI (Αctes du Ve Colloque internationale de Tirana, 20–23 mai 2015), 3 vols., Paris. Lhôte, É. (2011), “Les ethniques de Bouthrôte: étude linguistique”, in: L. Lamboley/M.P. Castiglioni (eds.), L’Illyrie méridionale et l’Épire dans l’antiquité V (Αctes du Ve Colloque internationale de Grenoble, 8–11 octobre 2008), vol. I, Paris, 105–112. LGPN III.A: Fraser, P.M./Matthews, E. (eds.) 1997. A Lexicon of Greek Personal Names: The Peloponnese. Western Greece, Sicily, and Magna Graecia, vol. III.A, Oxford. LGPN online: https://www.lgpn.ox.ac.uk/. Malkin, I. (2001), “Greek ambiguities: “ancient Hellas” and “barbarian Epirus””, in: I. Malkin (ed.), Αncient Perceptions of Greek Ethnicity, Washington D.C., 187–212. Maras, D.F. (2020), “Interethnic mobility and integration in pre-Roman Etruria: the contribution of onomastics”, in: J. Clackson et al. (eds.), Migration, Mobility and Language Contact in and around the Ancient Mediterranean, Cambridge, 23–52. Masson, O. (1993), “Encore les noms grecs et noms illyriens à Apollonia et Dyrrhachion”, in: P. Cabanes et al. (eds.) (1987–2004), L’Illyrie méridionale et l’Épire dans l’antiquité, 4 vols. (1987, 1993, 1999, 2004), Clermont-Ferrand/Paris, 77–80. Neumann, G. (2007), “Onomastics”, in: H. Cancik/H. Schneider (eds.; Eng. ed.: C.F. Salazar), Brill’s New Pauly (Antiquity), vol. X, Leiden/Boston. Parker, R. (ed.) (2013), Personal Names in Ancient Anatolia, Oxford. Parker, R. (2013), “Introduction”, in: R. Parker (ed.) (2013), Personal Names in Ancient Anatolia, Oxford, 1–14. Parker, R. (ed.) (2019), Changing Names. Tradition and Innovation in Ancient Greek Onomastics, Oxford. Rendić-Miočević, D. (1993), “Épidamnos-Dyrrhachion, Rider-Municipium Riditarum (Dalmatie)

Onomastic Formulae from N. Epirus and S. Illyria  

et leurs fonds épigraphiques comme sources de l’onomastique illyrienne”, in: P. Cabanes et al. (eds.) (1987–2004), L’Illyrie méridionale et l’Épire dans l’antiquité, 4 vols. (1987, 1993, 1999, 2004), Clermont-Ferrand/Paris, 119–125. Restelli, G. (1969), “Epirotico γνώσκω e il problema della posizione linguistica dell’antico epirotico”, Studi linguistici in onore di Vittore Pisani, vol. II, Brescia, 813–830. Schaffner, S. (2014), “Die slavischen Ethnonyme des Typs *poljáne ‚Feldbewohner‘ und die griechischen Ethnonyme auf -ᾱνες”, in: N. Oettinger/T. Scherr (eds.), Das Nomen im Indogermanischen. Morphologie, Substantiv versus Adjektiv, Kollektivum. Akten der Arbeitstagung der Indogermanischen Gesellschaft vom 14. bis 16. September 2011 in Erlangen, Wiesbaden, 360–383. Schmitt, R. (2018), “Greek personal names with Philo- as first or -philos as second element”, in: L. Alfieri et al. (eds.), Linguistica, filologia e storia culturale. In ricordo di Palmira Cipriano, Rome, 293–307. Shpuza, S. (2016), La romanisation del’Illyrie méridionale et de la Chaônie, Paris. Tzitzilis, Chr. (2007), “Greek and Illyrian”, in: A.-F. Christidis (ed.), A History of Ancient Greek: From the Beginnings to Late Antiquity, Cambridge, 745–751, 603–607. Tzitzilis, Chr. (2014), “Greek and Illyrian”, in: G.K. Giannakis et al. (eds.). Encyclopedia of Ancient Greek Language and Linguistics, Leiden/Βoston, vol. II, 56–59. Van Langendock, W./Van de Velde, M. (2016), “Names and grammar”, in: C. Hough (ed.), 17–38. Wilkes, J. (1992), The Illyrians, Oxford/Cambridge, MA. Wilkes, J. (2019), “Cohabitation or coexistence: Roman gentilicia in Greek and Roman epitaphs in Epirus and Illyris”, in: J. Lamboley/L. Përzhita/A. Skenderaj (eds.), L’Illyrie méridionale et l’Épire dans l’antiquité VI (Αctes du Ve Colloque internationale de Tirana, 20–23 mai 2015), 3 vols., Paris, vol. I, 173–180.

Paloma Guijarro Ruano

Dialect Contact and Koineization: The Case of the Greek Colonies of Aegean Thrace Abstract: The objective of this paper is the analysis of cross-dialectal interaction among the Greek colonies established in the northern Aegean Sea, alongside the Greek-Thracian border. Aside from the linguistic exchanges between Greek and non-Greek speakers (referred to as “Thracians” by the Greeks), I will focus on contacts between the various epichoric Greek dialects, largely from a diachronic perspective, including the process of koineization attested in this region.

 Introduction: dialect contact in the Northern coast of the Aegean Sea During the last decades, we have witnessed a renewed multidisciplinary interest in the northern coast of the Aegean Sea, greatly motivated by the systematic research of the National Hellenic Research Foundation on northern Greece and by new publications.1 Very importantly, these recent works have also boosted linguistic research. The relatively recent edition of the epigraphic corpus from Aegean Thrace (I.Thrake Aeg.) provides us now with the entire body of inscriptions (many of them unpublished) found in this peripheral area until 2005.2 New numismatic evidence from Orthagoreia, Abdera and Maroneia came to light thanks to Chryssanthaki-Nagle (2004; 2007) and Psoma et al. (2008).  This paper is framed within two projects, namely the Spanish national project Modos de contacto interdialectal en los textos epigráficos del griego antiguo (II) (FFI2017–82590–C2–1–P) and the European HERA project Uses of the Past: Multilingualism and Minority Languages in Ancient Europe, funded by the EU (grant agreement no. 649307), and, more specifically, within the Madrid-Research-Unit project (Dialect contacts and Dialect Levelling in Ancient Northern Greece). Its focus lies on the analysis of documents produced in the geographical area of modern northern Greece, Bulgaria, Albania and North Macedonia. In this sense, my work covers the northern coast of the Aegean Sea, in other words, the eastern part of this west-east geographical axis.  1 A work that also comprises Ancient Macedonia. An account of recent archaeological findings can be found in Fournier/Hammond/Parissaki 2015; on the linguistic interest in the area, see Guijarro Ruano 2019). 2 Loukopoulou/Parisaki/Psoma/Zournatzi 2005. See Thoneman’s review 2006. https://doi.org/10.1515/9783110779684-006

  Paloma Guijarro Ruano Furthermore, in 2015 the whole collection of inscriptions from the sanctuary of Apollo in Zone was finally published in Αρχαία Ζώνη I.3 Elaborating partly on new evidence, onomastic studies have also benefited from this research, due to the publication of the LGPN IV (2005),4 Dana’s Onomasticon Thracicum (2014),5 and Parissaki’s (2007) prospographical analysis. As far as language is concerned, cross-linguistic interaction between the northern local population(s), designated as Thracians by the Greeks,6 and Greek settlers must have been a reality alongside the Greek-Thracian strip, located between the rivers Strymon and Nestos and bordered by the Rhodope Mountains in the north. Their traces, however, are difficult to recover. In this work, I will therefore focus on cross-dialectal contacts. The existence of Aeolian and Dorian colonies in the neighbouring areas of the Black Sea and in Potidea7 and the intense trade activity alongside the northern Aegean coast, ensured the interaction among local communities and, consequently, dialect contact, despite the meagre material evidence.8 Two main Ionic varieties emerge in this area: Cycladic Ionic (in Thasian colonies such as Neapolis, Berge, Galepsos, Stryme etc.) and Asia Minor Ionic (in Abdera, Maroneia and in the Samothracian peraea, as in Zone, Drys, Sale etc.).9 Nevertheless, we have to deal with the difficulty of locating many of these settlements (Dikaia, Mesembria, Stryme etc.), while the identification of others with modern places or cities is either debated or still unconfirmed.10 From the period of the Attic control over this northern coast after the Persian wars (475–450 BCE) onwards, contact with the Attic dialect must have emerged too (and obviously later during the spreading of the  3 First notices in Brixhe 2006: 124; 2008: 21; later references in Brixhe 2018: 1852–1853. 4 It also comprises Macedonia, Thrace and northern regions of the Black Sea. 5 See also the online supplement (OnomThracSuppl, version 8, December 2021) available here: https://www.hisoma.mom.fr/production-scientifique/ressources-numeriques/onomasticonthracicum-supplementum [accessed in June 2023]. See also Dana’s 2014 review by Hălmagi 2015. 6 Up-to-date accounts of the Thracian language in Brixhe/Panayotou 1997; Brixhe 2006; Panayotou 2007; Tzitzilis 2014; Brixhe 2018; Yanakieva 2018; and Sowa 2020. On the new epigraphic and linguistic research into this specific area, see Guijarro Ruano (2019: 69–76). 7 A Corinthian colony created ca. 600 BCE. 8 Aeolian colonies were founded in Ainos (on the mouth of the Nestos river) and in the Propontis. An account of the possible Aeolic origins of the first Greek settlers of Samothracia can be seen in del Barrio 2018a: 491–496, with earlier bibliography. 9 On Ionic and its varieties, see del Barrio 2014. On Ionic, see also Bechtel 1924; Knitl 1938; Thumb/Scherer 1959; Stüber 1996; and del Barrio 2018a (an updated version of her previous work of 2015). Alonso Déniz’ 2018 study on Archilochus and the dialect of Thasos is also revealing. 10 For more extensive references to particular areas, see Loukopoulou 2004a; 2004b and Flensted-Jensen 2004 and the introductory parts devoted to geographical regions in I.Thrake Aeg.

Dialect Contact and Koineization  

Attic-Ionic koine). Not in vain, neighboring Imbros and Lemnos were Atticspeaking areas due to the establishments of cleruchies. Furthermore, inland Thasian colonies achieved their independence from Thasos at the same time as Athens benefited from the occasion. Our research is mainly based on the epigraphic corpus of I.Thrake Aeg., which includes 499 inscriptions,11 though only ca. 50 are dated between the 6th and late-5th/early-4th cent. BCE (only one is a public document). They mainly come from Abdera, the classical site of cape Molyvoti, and Zone. Until the 3rd cent. BCE, we find around 132 texts, and between the 2nd and the 1st cent. BCE, when the region came under Roman control (first as a client region and then as a province in 46 CE), we count ca. 80 inscriptions. Most of the epigraphic evidence comes from the 1st cent. CE, since Greek remained the primary language of communication over Latin.12 Updated material from the SEG,13 Αρχαία Ζώνη and later numismatic publications are also considered, although there are no new documents of significant linguistic interest.14 Such documents were mostly written in the koine, yet this region, not included in Bubenik’s (1989) study on koineization, has not received as much scholarly attention as other neighboring areas: see Panayotou’s (1998) on Macedonia and Tokhasiev (2009), Slavova (2009) and Bechet (2009) on the western coast of the Black Sea. Thanks to del Barrio’s study of the extant epigraphic evidence from the northern Aegean (2018a), we now have at our disposal the linguistic description of the use of the Ionic dialect between the rivers Strymon and Hebrus.15 Thus, it seems compelling to explore possible interferences among dialects and the later koineization of Aegean Thrace in comparison with other surrounding areas. To do so, we will examine diachronically the extant evidence, starting from the linguistic representation of Ionic and Attic in the  11 We have excluded Christian texts, for which see I.Thrace orientale byz. 12 I.Thrake Aeg. 149. The reverse situation in Lower Moesia, cf. Dana 2015: 244. 13 The latest SEG volume included in I.Thrake Aeg. was n. 50 [2000]. Νew texts are short and/or fragmentary, and are not linguistically revealing (cf. SEG 52.657bis-1162; 54.640; 59.720; 62.485–488, 495–496; 64.606); some readings have been corrected (cf. SEG 54.640; 61.557; 62.503; 63.492 etc.). Following del Barrio 2018a, we incorporate Thasian colonies located between the rivers Strymon and Nestos. 14 From the previously unpublished Greek material in Αρχαία Ζώνη, only three texts can be read: IV, ΑΓΚ 640 + 6143, A, (late-5th cent. BCE / early-4th cent. BCE); V, ΑΓΚ 6139a+b (ca. 411–400/ 375 BCE) and VI, ΑΓΚ 6152 (late-5th cent. BCE /early-4th cent. BCE). They have no dialectal relevance. For Thracian documents, see Αρχαία Ζώνη I, 209–564. Inscriptions contained in Αρχαία Ζώνη II, IIIa and IVa where already published in I.Thrake Aeg. 15 Her work includes I.Thrake Aeg. but also adds the region between the rivers Strymon and Nestos to the East.

  Paloma Guijarro Ruano inscriptions. We will then analyse the emergence and development of the AtticIonic koine in the whole area.

 Dialect contact in the prekoine stage . Ionic dialect In the epigraphic corpus of I.Thrake Aeg., safe diagnostic features of Ionic can be traced from the 5th cent. BCE onwards, as del Barrio (2018a) has revealed.16 Beyond graphic interferences,17 however, the analysis is exclusively confined to the realm of phonetics and morphology, since older documents are quite short and mainly funerary, with simple and repetitive sentence constructions. Syntax and vocabulary remain an unknown field in the archaic and classical periods. Among the salient traits of Ionic prior to ca. 350 BCE we can highlight the use of for ā in all contexts, also after /r/, /e/ and /i/.18 It could also be the case of the foreign ethnics Αἰγινήτης (E165, Molyvoti, ca. 450–425 BCE, fun., see also Φρασιδήμō) and Κυρηναῖος (E60, Abdera, ca. 350 BCE, fun.), although they could also be Attic and point to wider (trading?) contacts. We also find unaltered -ηιο- sequences in Ποσιδήϊος,19 as well as forms in ἱρο- (cf. Ἰρομνήμων,

 16 Inscriptions from I.Thrake Aeg. are quoted E1, E2 etc., as in del Barrio 2018a. When the term Molyvoti is used, it refers to the Molyvoti Peninsula. 17 I.Thrake Aeg. editors date inscriptions in the epichoric alphabet according to Pouilloux’ study on the Thasian alphabet (which fell into disuse from ca. 430–420 BCE onwards, cf. Pouilloux 1954, 443), but often no references to the alphabet are expressed, so that we are not certain about the use of local alphabets. The influence of the Parian-Thasian alphabet (use of for /o/, /o:/ and for / ɔː/) is observed in two documents found in Abdera (E30, mid6th cent. BCE, funerary; E13, mid-5th cent. BCE, ex-voto), on which editors cast doubt regarding its possible Thasian origin; also possible in another document owing to the rendering of rho (E1, ca. 480–470 BCE). The Parian-Thasian alphabet spread up to Zone, as Brixhe (2015: 283– 284) infers from the lunar beta and from the exchange between omicron and omega as they appear in Ἀπώλ(λ)οο and in Χωιρῖνως. It must have been imported by Thasian settlers (or by Parians, cf. Brixhe 2006: 126). 18 Ἀρτεμισίης (E32, Abdera, ca. 475–450 BCE), πρωτέρηι, ὑστέρηι (SEG 51.797, Berge, ca. 470– 460 BCE), Ἰστίης, Πυθείης, E14, Abdera/Thasos, late-5th cent. BCE), Ἀθηνάης (08, Molyvoti, late-5th cent. BCE), Ἀθηναγόρης, ἡμέρης (SEG 52.711, Pistiros, late-5th cent. BCE), Ἡγησαγόρης (May 1996, nn. 228–232, Abdera, ca. 415–395 BCE, coin) and Πολυαρήτου (E141, Molyvoti, ca. 375–350 BCE, fun.). 19 Ποσιδήιō (E113, Molyvoti, ca. 450–400 BCE, fun.), Ποσιδηίου ((S)GDI 5685.18, Maroneia, 450–400 BCE, coin), but later Ποσιδείου ((S)GDI 5685.19, Maroneia, 400–350 BCE).

Dialect Contact and Koineization  

LGPN IV, Abdera, ca. 360–350 BCE, coin) and in ἱστ- (Ἰστίης, E14, Abdera, late 5th cent. BCE, dedication). Normally, hiatuses do not contract. Thus, εο and εω were graphically preserved until the 4th cent. BCE in common words (cf. [ὑ]πνώδεος, ὠνεομέν[οις] in E3, Abdera, bef. 350 BCE, law), in personal names (Ἀ̣ρ̣θ̣έων, E222, Maroneia, 4th cent. BCE, fun.)20 and in Θεοo and Νεοo compounds. The genitive singulars of astems and s-stems display non-contracted -εω21 and -εος22 respectively (mostly from Molyvoti, Zone and Maroneia, but see infra Θεογένōς and Ἀντιφάνōς̣),23 along with the genitive plural forms of a-stems in -εων, mainly in ethnics.24 Ιndeed, alleged contracted forms -εω > -ω are dubious. In E39 (Abdera, ca. 400– 350 BCE, fun.) we read Μεγαδο[ ], assumed to be a genitive Μεγάδω by the first editor, and thus rendered in SEG 47.1029.25 Μεγάδω implies a nominative Μεγάδης and the contraction -εω > -ω after a consonant (del Barrio 2018a: 501). However, another possibility comes into play: nom. Μεγᾶς, gen. Μεγᾶδο[ς], as proposed in I.Thrake Aeg., and suggested by Parissaki (2007: 209). This second

 20 Ἀ̣ρ̣θ̣έων has been interpreted as an Illyrian name, although a Greek origin seems more probable, cf. Parissaki (2007: 136–137). 21 Ἱππέω (E33, Abdera, ca. 475–450 BCE), τ̣ õ Μένεω (E163, Molyvoti, mid-5th BCE, fun.), Νυμέω (E123, Molyvoti, ca. 425–400 BCE, fun.), Θ̣ε̣ο̣ ξ̣ενίδεω, (E222, Maroneia, 4th BCE, fun.), Ἑ̣ρμαγόρεω (E38, Abdera, early-4th cent. BCE), Ἀπελλέω (E136, Molyvoti, ca. 400–375 BCE, fun.), Ἀπελλέω, Ἀριστόλεω, Ἡγησαγόρεω, Ἡρακλείδεω, Ἀθήνεω ((S)GDI 5686, Maroneia, 400– 350 BCE, coins), Ἀθήνεω (Schönert-Geiss, Maroneia 152, ca. 432–423 BCE, coin). The patronymic of Πατροκλῆς Παυσανία (E138, Molyvoti, ca. 400–375 BCE, fun.), devoid of context, can be a foreign (Doric?) form (we cannot dismiss the Spartan influence/presence in this coast before the general participation of these poleis in the Second Athenian League in 378 BCE, cf. Brixhe 2015: 275) or a koine genitive. On the rapid spread of the declension -ίας, -ία from the early 4th cent. BCE onwards, even in Ionic territory, see Brixhe/Panayotou 1988: 251; Brixhe 1963; del Barrio 2014, 264. 22 Ἀ̣ριστομένεος (E112, Molyvoti, ca. 450–400 BCE, fun.), Διονυσοφάνεος (E137, Molyvoti, ca. 400–375 BCE, fun.), Π̣ολυχάρεος (E414, Zone, ca. 400–350 BCE, fun.), Πολυκράτεος (E420, Zone, mid-4th cent. BCE, fun.), Καλλικράτεος, Μητροφάνεος ((S)GDI 5685, Maroneia, 400–350 BCE, coins). 23 An overview of the interferences between a-stem and s-stem genitives in Thasos and Archilochus in Alonso Déniz 2018: 546–550. 24 Ἀβδεριτέων ((S)GDI 5644, nos. 2, 5, 14, 17, 20, 21, 500–350 BCE, coins), Μαρωνειτέων ((S)GDI 5685.15, 400–350 BCE, coins), Νεοπολιτέων ((S)GDI 5488, 350–300 BCE), Ἀπολλωνιητέων (SEG 49.911, Pistiros, ca. 359–339, regulation), Ὀρθαγορέων (Chryssanthaki-Nagle 2004: 52–53, ca. 350–300 BCE), Μαρωνηιτέων/Μαρωνειτέων/Μαρωνιτέων (cf. Bechtel 1924: 49) and ἐμποριτέων (SEG 49.911, Pistiros, ca. 359–339 BCE). 25 Cf. ΜΕΓΑΔΩ Triantaphyllos AD 1992 [1997]: 499–500 BCE; cf. [ ]πὼ Μεγάδω γυνή in SEG 47.1029, and [Ἀ]σ̣πὼ Μεγαδο[ ] γυνή in I.Thrake Aeg. E39.

  Paloma Guijarro Ruano hypothesis has the advantage of presenting more parallels of these hypocoristic names (cf. Μεγᾶτος P. Petrie 3.43, Egypt, 240 BCE),26 against the unique example of Mεγάδης, which appears in Homer but as a patronymic (Πέριμον Μεγάδην, Il. 16.695).27 Besides, it matches with the characteristic Ionic declension -ᾶς, -ᾶδος attested and in the same region (cf. Mολπᾶδος, Abdera, 450–425 BCE, coin) and in Zone (cf. Βοστᾶδος E404, mid-5th cent. BCE; E411 and E412, ca. 400– 350 BCE, fun.) in the same period.28 Additionally, a contracted Μεγάδω is not represented in Aegean Thrace until the mid-4th cent. BCE, unless we understand other forms such as Ἀλκεβιάδō/Ἀλκιβιάδō (E405–407, Sapli Dere, mid-5th cent. BCE fun.) in the same way, though these are probably Attic forms (see infra).29 The interpretation of Πρωτόφαος (E130, Molyvoti ca. 400 BCE, fun.) is debated. For the I.Thrake Aeg. editors it is in the nominative case, while for del Barrio (2018a: 488, 506) it must be a contracted genitive in -φαōς (=φάους < -φάεος), built upon names in -φάης, documented in Thasos, its peraea and in the neighbouring Aeolis. With regard to E30 (Abdera < Thasos?, mid-6th cent. BCE), written in the Parian-Thasian alphabet, Gauthier and Dubois’ (BÉ 118 [2005] 436) read ΠΡΟΤΙΟ as Πρωτίω. However, Del Barrio (2018a: 501) and Alonso Déniz (2018: 549–550, fn. 66) point to the possibility of a graphic Parian-Thasian variant of the omega with two oblique strokes (ΠΡΟΤΙΩ). While the former interprets Πρωτίō as Πρωτίου, an Attic(-koine) genitive of Πρωτίας or Πρώτιος, Alonso Déniz proposes two different readings: a genitive Πρωτίο with vowel loss -εo > -o,30 or

 26 For these genitives, see Dubois 2017: 317–321. ΜΕΓΑ on a coin from Abdera (May 1966: III 100–101) does not represent a telling occurrence. 27 Μεγάδης derives from Μέγας (attested in Mycenean, cf. Mycenean gen. me-ka-o PY Na 571), cf. Eustathius ad loc. (I thank the anonymous referee for this reference). See also both accusatives Μέγητα (Il. 19.239) and Μέγην (Il. 15.302) in Homer. Ιt only exists Μεγγίδας in Boeotia (IG VII 585, Tanagra, 426 BCE). Αll the LGPN forms (14 entries) appear under a nominative Μέγας (cf. Bechtel 1917: 299–300), according to Μέγης (around 10 examples among which we find the dental inflection in -ητος, the earliest in IG XII,9 944, Euboea, mid-2nd BCE). But gen. -ᾶδος/ ᾶτος forms point to Μεγᾶς. 28 Βechtel 1924: 129; del Barrio 2014: 264; Curbera 2013: 115. Further development in Guijarro Ruano 2023. 29 εω > ω is mostly documented in s-stems and in gen. pl. of a-stems (also in literary Ionic). Nevertheless, synizesis of /o/ after /e/ is attested in metrical inscriptions from Ionic regions (cf. Méndez Dosuna 1993: 100), such as Naxos, Paros and Thasos (for the latter, see Alonso Déniz 2018: 531 fn. 3, 532 fn. 4). Concerning literary Ionic, it applies to iambic authors, not to elegiacs (cf. Méndez Dosuna 1993: 99–101 and Kaczko 20162: 292). 30 He also raises a similar case of Homeric imperative *ἀποαιρέεο > ἀποαίρεο “take away”. Aside from the possible metrical use of this form, this poetic example belongs to a different

Dialect Contact and Koineization  

Πρωτίō, with contraction -εω > -ω. Alonso Déniz’ hypothesis is based on an analogical phenomenon attested in 4th-cent. BCE Thasian anthroponyms, by which -εο genitives replace genitives in -εω in -a stems (cf. Thasos Λεωνίδευ, SEG 31.782) and extended secondarily to s-stems (cf. Σοκράτευ, Pouilloux 1954: 313 fn. 108). An analogous case could be found in ΦΡΑΣΙΗΡΙΔΩ (CEG 162, Thasos, 500–490 BCE), in which the epichoric Parian-Thasian alphabet permits a double reading Φρασιηρίδō (= Φρασιηρίδω < Φρασιηρίδεω) or Φρασιηρίδo (< Φρασιηρίδεο). In Alonso Déniz’ opinion, this phenomenon could be related to the replacement of -εω with -ω “under the influence of -ίω and -έω (< *έεω) for which vowel dropping or contraction is regular” in other Ionic dialects.31 Be as it may, it is a localized feature still open to debate. Other distinctive morphological Ionic features are mainly visible in some anthroponyms, such as i-stem anthroponyms, which exhibit genitives in -ιδος, but also in -ιος.32 Further potential Ionic traits are late enough to attribute them to the koine (see infra). Concerning the sketching out of local Ionic sub-varieties across this geographical border, very few traces have survived, and salient features occur mostly in anthroponyms. This important caveat makes it difficult to draw conclusive observations. Regarding the Eastern Ιonic variety, the aforementioned Ἰρομνήμων (LGPN IV, ca. 360–350 BCE, coin) from Abdera fits in well with ἰρός, but is at variance with the almost sure form ἱ[ερ]ῆς (E3, before 350 BCE), from the same polis too.33 In this Ionic variety ἰερός is also documented in the Dodecapolis, although ἰρός appears in the northernmost cities of Asia Minor too.34 However, it is impossible to know whether word-initial aspiration was pronounced in this word.  morphological category and contradicts his hypothesis of a substitution of -εω genitives with -εο in some Ionic areas. 31 See for instance Ἀριστίδω, Ποσεδεωνίω or Ξανθίω (IG XII,6 2, 996, Samos, ca. 550 BCE); cf. Bechtel 1924: 140. 32 Πάρμιδος (E43, Abdera, mid-4th cent. BCE), Πόσιδος (E224, Nea Maroneia, ca. 325–300 BCE, fun.) and Φάσιδος (E212, Maroneia, 2nd–1st cent. BCE, cat.), but Νέσ̣ τιος (E114, Molyvoti, ca. 450–400 BCE, fun.), Εὐξιθέμιος, Εὐπόλιος, Κλεοπόλιος ((S)GDI 5685, Maroneia, 400–350 BCE, coins) and Τείσιος (E51, Abdera, 3rd–2nd cent. BCE). Examples of common names are fairly late (cf. τῆς πόλεως, E7, Τeos < Abdera, 2nd cent. BCE, proxeny decree). Hellenised Thracian names are not consistent with their adaptation to this declension; see, for instance, the rendering in gen. -πορεως (Ῥασκου]πόρεως, E458, Plotinopolis, 42–31 CE, dedication; Μουπορεως, E273, Maroneia, 2nd cent. CE, fun.; Μοκαπορεως E387, Dyme, 1st cent. CE, fun.), completely assimilated to Greek -πολεως (cf. Ῥασκου]π̣όλεως, E207, Maroneia, 1st cent. BCE/1st cent. CE), but -por in Egypt, or -ποριος, -ποριδος outside Aegean Thrace. 33 Masson 1984: 52; del Barrio 2018ª: 500. 34 Cf. del Barrio 2014: 263.

  Paloma Guijarro Ruano The so-called 3rd compensatory lengthening is not well represented in our corpus aside from anthroponyms, where we find non-lengthened forms in *ksenu̯- (cf. Θ̣ε̣ο̣ ξ̣ενίδεω, E222, Maroneia, 4th cent. BCE, fun.; Ξ̣ενοφῶντος, E146, Molyvoti, ca. 400–350 BCE, fun.; Ξεναῖος, E238, Maroneia, 3rd cent. BCE, fun.)35 and contradictory outcomes from *koru̯- in two contemporary inscriptions from Molyvoti Διοσκορίδ[ης] (E129, ca. 400 BCE, fun.) and Δ̣ιοσκουρ̣[ίδου] (E143, ca. 375–350 BCE, fun.).36 The issue of aspiration in the northern Aegean colonies is far from clear. In general lines, East Ionic is psilotic while Central Ionic is not. In fact, only Thasian data permits serious discussion of aspiration: on different grounds, del Barrio (2018a: 479–484) and Alonso Déniz (2018: 538–542) have attempted to prove the existence of psilosis in Thasos, in contrast with what happens with its metropolis Paros. The absence of a written sign for initial /h/ in Thasos and in its colonies indicates that was used for /εː/ and does not necessarily imply psilosis. This could be the case of ὑστέρηι (SEG 50.575; SEG 51.797, Berge ca. 470–460 BCE).37 In the Thasian colonies, where the Parian-Thasian alphabet was imported, the data are contradictory: see in Galepsos hώρως vs. wordboundary psilosis in τὠκατωνπέδω (SEG 43.400, ca. 500 BCE, boundary). No occurrences from Molyvoti show attempts to render an initial h-.38 Traditionally, this place has been identified with Stryme, a Thasian colony, but more recently with the site of pre-Hellenistic Maroneia, originally founded by Chios.39 If the examples were conclusive, psilosis could be attributed to either of the two possible metropolises, Thasos or Chios. To this end, we have to either accept that the Thasian dialect was psilotic or support the identification Molyvoti-Maroneia. Even so, however, there is no direct evidence of Maroneian aspirated/psilotic

 35 There are no conclusive forms from Abdera: νόσου (E2, before 350 BCE) and προξέν]ων (E7, found in Teos, 2nd cent. BCE, proxeny decree). 36 In I.Thrake Aeg. no indications concerning the alphabet are provided. For later examples, see μόνους (E205, Maroneia, mid-2nd/early 1st cent. BCE), Ξενόκριτος, Ξενοκρίτου, Ξένωνος, Θεόξενο̣ ς], Θεοξένου, Φιλόξενος and Φιλοξένου vs. Διοσκουρίδης and Διοσκουρίδου in E212 (Maroneia, 2nd–1st cent. BCE, catalogue). 37 del Barrio 2018a: 488–489. 38 Ὁρ̣ίō (E107, ca. 450–400 BCE), Ἠραΐννη (E109, ca. 450 BCE), Ἡροβṓλō (E119, ca. 450–425 BCE), Ἠ̣ροδώρō (E163, mid-5th cent. BCE), Ἡράϊννα (E127, ca. 400 BCE), Ἠρεσστράτō (E130, ca. 400 BCE) etc. 39 Cf. Loukopoulou and Psoma 2008: 55–86. See the whole report in Arrington et al. 2016: 4, who prefer to set this place under Maroneia’s influence, rather than identify it with pre-Hellenistic Maroneia.

Dialect Contact and Koineization  

forms in this period. As del Barrio (2018a: 502) states, the absence and lack of dialectal features “provide us with little information on its dialect”. According to del Barrio (2018a: 500), in the absence of conclusive proofs of psilosis in Abdera, we may hypothesize that it existed, as in Teos. There are, indeed, some forms without h- (cf. article ὀ, E15, late 5th cent. BCE), but others remain problematic. First, (E14, late 5th cent. BCE), rendered Ἱστ- by the editors but Ἰστ- by del Barrio (ib.), appears in an inscription whose left part is broken so that the aspiration could have been marked in the lost part.40 Similarly, Ἰρομνήμων (LGPN IV, ca. 360–350 BCE, coin) has no indication of h- in del Barrio’s analysis in contrast to Masson’s reading Ἱρο- (1984: 52 = OGS II, 431). In addition to the explanation of (see supra), (E30, mid6th cent. BCE, fun.) has also been the subject of discussion. For del Barrio (2018a: 502), is as an abbreviated writing of , likely documented in Paros too. Notwithstanding the use of local stone from Abdera, she solves the problem by attributing to the alleged Parian origin of the deceased, following the editors’ remarks. Contrasting the non-aspirated article ὠ (= ὁ) with the initial /h/ in , Alonso Déniz (ib.: 542) sees a starting process of lenition of word-initial aspiration, first rendered in grammatical words. Other distinctive East Ionic traits are not preserved in our texts, such as the demonstrative pronoun κεῖνος or the plural datives in -οισι(ν)/-η(ι)σι(ν), aside from later occurrences in (semi-)metrical texts.41 The same applies to the outcome εο, εου > ευ and to the interchangeable use of αο and εο and αυ and ευ (attested in Thasos and other regions).42 Before ca. 350 BCE πῶς/κῶς-type adverbs do not appear, and later only πῶς occurs.43

. Attic dialect The Athenian presence in Aegean Thrace is known from historical accounts, but also from some of the extant epigraphic material. Furthermore, historical and archaeological sources, which report the incorporation of some colonies into the First Maritime League and their inclusion in the Athenian tribute lists and  40 The cult of Ηestia is first documented in the area (Thasian epigraphic evidence dates from the 4th cent. BCE, see I.Thrake Aeg. p. 216). 41 οἷσι δόμοις (E436, Traianoupolis, late-2nd/early-3rd cent. CE, sundial), ἐ̣ ν Μαρωνίτ]αισιν (E217, Maroneia, 2nd–3rd cent. CE, epigram). 42 del Barrio 2014. 43 In official Ionic inscriptions we also find πῶς-type, cf. Horrocks 2010: 60–61. Similarly, ἐσθλός appears late (cf. E88, Topeiros, 3rd cent. CE).

  Paloma Guijarro Ruano commercial trade, confirm a two-way relationship between Athens and the colonies of the north-Aegean coast. The inscriptions bear testimony to Athenians buried across this area and trade exchanges44 and, conversely, to the passing of Agean-Thrace people in Athens from the mid-4th cent. BCE. Owing to the nature of the anthroponyms (cf. Θρᾶιξ and Θρᾶι[ττ]α and the common Thracian name Κότυος) and patronymics (Γέτου “Geta”, Ταλούλου, both “Attic” genitives), these were probably Thracian slaves buried there.45 Attic evidence from the 5th cent. BCE on local institutions and cults is also detected in the Molyvoti documents. Here the worship of Athena and Zeus Horios, imported from Athens, is attested for the first time in Ἀ̣θ̣ ε̄ ν[άε̄ ]ς̣ Δὸς Ὀρ̣ίō (E107, ca. 450–400 BCE).46 Besides, the addition of Athena to the three healing divinities Podaleirius, Machaon and Periesto (Ποδαλιρίō Μαχάōνος Περιεστõς Ἀθηνάης) in E108 (Molyvoti, late5th cent. BCE), all of them Asclepios’ children, points to an Attic loan: probably, this Athena has to be related with Athena Ὑγιεία, another of Asclepius’ children.47 However both texts, dated to the 5th cent. BCE, when Athenian influence in the area was more intense, display Ionic traits (cf. Ἀθηνάης < Ἀθηναίης). For this reason, linguistic interferences between Attic and Ionic dialects in our inscriptions can be expected from this time onwards. Attic features found in our epigraphical corpus in the prekoine stage are mainly limited to the presence of the so-called Attic retroversion, which dates from the mid-4th cent. BCE (cf. Ἀ̣ν̣ τ̣ α̣γ̣ό̣ρας, E149, Molyvoti, mid-4th cent. BCE, fun.). Α troublesome feature is the -ō rendering of thematic genitives instead of the Ionic -εω in the mid-5th cent. BCE, documented in Samothracian colonies, in Ἀλκεβιάδō/Ἀλκιβιάδō (E405–407, Zone, fun.) and Ἀνταγόρō (E109, Molyvoti, fun., epigram). Around 350 BCE we already encounter in the area of Molyvoti Ἡγησιππίδου (E151, fun.), Δάδου (E153, fun.)48 and Ἀντιπατρίδου̣ (E161, fun.). Supposedly, these could be Attic forms in -oυ or, in the light of the oldest

 44 Θαλίαρχος Ἀθε̄ ναῖος Κ̣ Ε․․․․]ὺς Θ̣α̣λ̣ ί̣ ν̣ọ̄ (E58, Abdera, ca. 475–450 BCE). On trade, see, for instance, E3 (Abdera, before 350 BCE). 45 Κότυος Μαρω]νίτης (IG II² 9289, post ca. 350 BCE), Θρᾶιξ Γέτου Μαρωνίτης (IG II² 9287, III BC), Γλαυκίας Ταλούλου Μαρωνίτης (IG II² 9286, Roman period) and Θρᾶιττ]α Ἀνδράβυδος Μρωνῖτις (IG II² 9288, undated). 46 See the editors’ references in I.Thrake Aeg. to other testimonies to this cult in Athens, Abdera (E14) and Propontis Thrace. 47 See Masson (BÉ 104 [1991], 210) pace I.Thrake Aeg. 288, 290 (also concerning the assimilation Περιεστώ - Ἴασις). On the cult in Attica of Ἀθηνᾶ Ὑγεία as Asclepios’ daughter, see Paus. 1.23.4. 48 A Thracian name in Dana 2014 s.v. Dada, Δαδας; see also Parissaki 2007: 155 with earlier bibliography.

Dialect Contact and Koineization  

examples, contracted Ionic genitives in (-εω >) -ω. However, given the absence of later genitives in -ω, together with the rendering of /ɔː/ with in other mid-5th cent. BCE examples (cf. Πυθόδωρος, E409, Zone, ca. 450–400 BCE, fun.; Ἀ̣πολ̣ λ̣ ω̣νίδου, E152, Molyvoti, mid-4th cent. BCE, fun.; see infra), del Barrio (2018a: 498–499) claims that “they belong to Attic-Ionic koine”. After all, -oυ genitives finally predominate, even in Hellenised Thracian names.49 Indeed, Parissaki (2007: 126) attempts to link individuals from Zone named Ἀλκιβιάδης with Athens, bearing in mind the influence of Athens in this place and, particularly, the regional distribution of these anthroponyms, since the majority of them come from Attica.50 To this end, we must compare the whole ’Ηγησίλαος Ἀλκιβιάδ̣ọ̄ (E407), and its final -λαος, with the proper Ionic Ἡγησίλεως (cf. Schönert-Geiss, Maroneia 162, ca. 400–378 BCE, coin). Tellingly, these forms start to appear in a period in which the so-called Great Attic emerges, which develops in the 5th cent. BCE and is especially prominent during ca. 425–336 BCE.51 It coincides with the inclusion of most of north-Aegean cities in the Athenian tribute lists and among the members of the Delian League.52 In connection to this, we are confronted with two analogous interpretations when analysing the contracted s-stem genitives Θεογένōς (E409, Sapli Dere, ca. 450–400 BCE, fun.),53 and Ἀντιφάνōς̣ (E126, Molyvoti, ca. 400 BCE, fun.). These could be Attic or early contracted Ionic forms (del Barrio 2018a: 498),54 but there is no reason to deny an Attic influence, given other non-contracted occurrences from Sapli Dere until the mid-4th cent. BCE (Π̣ολυχάρεος, E414, ca. 400–350 BCE, fun.; Πολυκράτεος, E420, mid-4th cent. BCE, fun.) and the historical Athenian influence over the region. Indeed, both forms are accompanied by Πυθόδωρος (E409) and Ἀπο[λ]λ̣όδωρος̣ (E126), where a graphic distinction between original /ɔː/ and /o/ is made. Thus, in E126 /o/ and /ɔː/ are graphically differenti-

 49 Despite the issue of establishing their declension and their graphic variants, cf. Ῥοιμητάλκου, Καπρουβήβου (E387, Dyme, 1st CE, fun.), Δωλέου̣ (E249 Maroneia, 2nd cent. CE, fun.), Τήρου (E384, Agiasma, 3rd CE, dedication), Δολεοῦ, Βίζου (E386, Asomatoi, 3rd cent. CE?, fun.), some of them seem to have been clearly reinterpreted as a-stem masculine names (cf. Τηρης, Τηρου; -αλκης, -άλκου). On the Hellenization of Thracian names, see Dana 2017. 50 Parissaki brings forth E401 (ca. 425–400 BCE) and E403 (ca. 425–400 BCE); on the strong Athenian influence in classical Maroneia, see Psoma et al. 2008: lii–liii too. 51 Horrocks 2010: 73–78; Teodorsson 2014. 52 In the case of Zone, see the assessment decree of 422/421 BCE (IG I3 77.5.27–28). 53 Θεοο remains unaltered, without contraction nor diphthongization (see § 1.1). 54 Both inscriptions are very short and without further linguistic clues to clarify the dialect used. See fn. 88 for later outcomes in -ους.

  Paloma Guijarro Ruano ated from secondary [oː], rendered by ō < /eo/ (Ἀπο[λ]λ̣όδωρος̣ vs. Ἀντιφάνōς̣).55 Indeed, a little later we find Ἀ̣πολ̣λ̣ω̣νίδου (E152, Molyvoti, mid-4th cent. BCE, fun.). The analysis of the genitive forms in -κλεος attested in the region (cf. Θεοκλέoς, E121, Molyvoti, ca. 425–400 BCE, fun.; Προκλέoς, E135, Molyvoti, ca. 400–375 BCE, fun.; Σωκλέoς, E418, Zone, mid-4th cent. BCE, fun.; Πατροκλέoς, Schönert-Geiss Maroneia, 492–493, 365–336 BCE, coin; [Π]αντακλέoς̣, E223, mid-4th cent. BCE, fun.) presents a similar difficulty. They can render the expected Ionic genitive (-κλέ(ε)ος >) -κλέος56 or early cases of Attic contraction -κλέōς (-κλέους). In I.Thrake Aeg. no clear indication to their alphabets is made, but and are used in the same inscriptions (cf. Σωκλέoς in E418, Διονυσοφάνης in E121, Ἀθηναΐς in E135 or Ἡραΐς in E223). Unfortunately, there are no parallels of secondary /oː/ until the 3rd c. BC, when we do come across Ἀθην̣ ο̣ κ̣λέ̣ ους (E234, Maroneia, fun.) and Χαρικλέους (E235, Maroneia, fun.), Πατροκλέους (E254, Maroneia, Hellenistic period, fun.). This means that the abovementioned -κλεος genitives might well be interpreted as Ionic -klĕŏs, but they could also be contracted -κλέōς forms not yet rendered -κλέους due to the lack of other dialectal features and to the early Attic influence over the region.57

 A turning point, the 4th cent. BCE Because of the continuous contact between Ionic and Attic speakers and the gradual growth of a supradialectal variety in the Greek world, the 4th cent. BCE is the most suitable period to detect the influence of the koine and phenomena of dialect levelling in our corpus.58 However, few inscriptions seem to point to dialect contact thus far. In some regions such as Zone, the earliest documents only exhibit Attic (-koine) features. As we can see in Αρχαία Ζώνη I, ΑΓΚ 640 + 6143 (late-5th/early4th cent. BCE, decree?), a clear Attic [ἀ]συλίαν stands out with other forms like ἐκγόνοις, [πο]λέμωι, [εἰρήν]ηι, which can also be Ionic.59 Nonetheless, in E3, a

 55 del Barrio 2018a does not report examples of for original /ɔː/ nor for secondary /ɔː/. 56 Bechtel 1924: 66. 57 If some extra documentary evidence proved the kinship of the Διονυσοφάνης Θεοκλέoς of E121 with the one in Πατροκλῆς Διονυσοφάνεος (E137, Molyvoti, ca. 400–375 BCE), as proposed in I.Thrake Aeg., 304, we could have further arguments to claim the Ionic origin of these individuals. 58 A similar approach in Brixhe/Panayotou 1988 (see also Brixhe 2010); Crespo 2009; 2010. 59 The inscription is labelled as bilingual, but it does not seem a bi-version text (in Greek and Thracian), but two different texts (cf. Αρχαία Ζώνη I, 257–269, 358–372).

Dialect Contact and Koineization  

decree from Abdera dated before 350 BCE, specific Ionic features such as /εː/ (ἰ[ερ]ῆς, δυσεντερίης and [πο]δαλγίης), non-contracted ὠνεομεν[ and [ὑ]πνώδεος and the numeral τέσσερας coexist with Attic/Attic-Ionic koine forms in grammatical words, such as the dative τοῖς or ἐάν (vs. the expected τοῖσι and ἐήν in Abdera).60 The reconstructed forms [ἀγορ]ᾶς and — more securely — [ἡμ]έρας point to the same direction.61 Additionally, the genitive Διονυσᾶδος (E2, Abdera, ca. 350–300 BCE) and probably the ethnic Κυρηναῖος (E60, Abdera, ca. 350 BCE, fun.) also exhibit an Ionic layer62 that contrasts with later Σωτάδας (E263, Maroneia, 3rd cent. BCE, fun.) and [Π]τ̣ ολέμα[ς] (E264, Maroneia, 3rd cent. BCE, fun.) who come from Crete (Κρής) and Philippoi (Φιλιππεύς) respectively.63 Another interesting case is the famous Pistiros inscription, found in Vetren (Bulgaria) and first published in IGBulg V 5557(3), which is dated after 359 BCE.64 Here salient Ionic traits are retained: /εː/ in πρήσσειν, φρουρήμ, Μαρωνείης, Πιστιρηνῶμ and Ἀπολλωνιητέων; /eɔː/ in the genitives ἐμποριτέωμ and Ἀπολλωνιητέων, along with ἀποκτενέω (/ea/ is also uncontracted in τέλεα) and ταõτα and αὀτόμ instead of ταῦτα and αὐτόν, which could reflect East Ionic or Thasian influence (but αὐτός, αὐτούς, as well). Furthermore, the ethnic suffix -ηνός (cf. Πιστιρηνῶμ) is also common in Propontis and Asia Minor.65 On the other hand, the datives τοῖς ἐμπορίταις and the aspiration at word-boundary (ἀφαιρεῖ[σθ]αι, ἀφαιρήσομαι) could also be attributed to Attic(-koine).66 When it comes to analysing these forms, we are faced with the much-debated identification between the Pistiros (Πίστιρος) mentioned in the inscription, a trading centre in modern Bulgaria, and a Thasian colony called Pistyros (Πίστυρος), located by Herodotus (7.109.2) in the northern coastline of the Aegean.67 The widespread  60 Plural dative in -οις/-ηις is documented in East Ionic from the 4th cent. BCE (cf. del Barrio 2014: 264); in Thasos it spread out even in lexical forms (cf. Alonso Déniz 2018: 533). 61 Del Barrio 2018a: 499–500. For γῖνος “small mule”, cf. Hesychius s.v. γίννος (LatteCunningham 567). 62 Κῖκος vacat Φρασιδήμō Αἰγινήτης vacat. According to Threatte 1980: 52, 539, foreign toponyms and ethnonyms used to appear adapted to Attic in Attic inscriptions, although from the 5th cent. BCE foreign forms prevail even in the tribute lists. 63 According to LGPN IV, there is one more example in Macedonia (ILeukopetra 17, 1, Beroia, 184 CE). 64 We follow Chankowski and Domaradzka’s final edition 1999 (= SEG 49.911) with later corrections of SEG 63.492. 65 del Barrio 2018a: 491. 66 In Thasos dat. -οις/-ηις is expected, cf. Alonso Déniz 2018: 533 fn. 9. 67 Herodotus describes Pistyros as παραθαλασσία and Ἑλληνίς. Other occurrences of Pistiros, in combination with our epigraphic testimony, appear in Stephanus of Byzantium 524.11 (Πίστιρος, ἐμπόριον Θρᾴκης), distinguished as emporion from ib. 171.6 (Βίστιρος, πόλις Θρᾴκης

  Paloma Guijarro Ruano assertion that the Aegean Pistyros is a Thasian colony is based on its identification with modern Pontolivado, where some Thasian material evidence has been found. In light of the resemblance between both toponyms, some authors have identified the two settlements as the same, while the majority favour the idea that the emporion of Pistiros was founded by citizens from the Aegean Pistyros.68 Be as it may, the text shows the interaction with the Ionic dialect of its metropolis, already displayed in earlier graffiti,69 mixed with Attic(-koine) forms. In conclusion, the statement that this text is in Ionic dialect is reasonable in linguistic terms but its connection with a particular region is a matter of dispute. After ca. 350 BCE, we deal with the ever-growing difficulty of differentiating Attic (or Great Attic) from Attic integrated within the Attic-Ionic koine.70 This is mainly the case with the rendering of a-stem masculine genitives in -ου and of the using of ᾱ instead of η after palatal vowels and /r/. The first testimonies to ᾱ date from the second half of the 4th cent. BCE (cf. Ἀ̣ν̣ τ̣ α̣γ̣ό̣ρας, Ε149, Molyvoti and Μ̣αιανδρία, Ε422, Zone). Similarly, in E400 (late-4th/early-3rd cent. BCE), a decree-copy found in Molyvoti issued by Drys, we find non-Ionic /aː/ in Πολυάρατος and other lexical words (προξενίαν, εὐεργασίαν, πολιτείαν). The local Ionic origin of Πολυάρατος is confirmed by his patronym Ἱστιαῖος, common in Ionicspeaking regions,71 and by other occurrences of Πολυαρητο- in Molyvoti (E141, ca. 375–350 BCE, fun.), in some coins from Maroneia (Schönert-Geiss, Maroneia 494–8, ca. 365–336 BCE) and in a contemporary inscription from Thasos (IG XII, 8 267, early-3rd cent. BCE). The latter (Πολυά̣ρητος ὁ Ἱστιαίου), probably the same person as in our inscription, received the Thasian citizenship.72 The form Πολυάρατος in E400 is due to the strong influence of the koine in this official text (as opposed to the Thasian inscription, in which local features — mainly in anthroponyms — coexist with koine features in common words). Along with participle ὤν or the datives ἐκγόνοις and Δρυΐταις, the document also exhibits other koine-forms, such as an analogical aorist εἶπαν (for εἶπον)73 and the standardized formula καὶ εἴσπλουν καὶ ἔκπλουν καὶ ἐμ πολέμωι καὶ ἐν εἰρήνηι.  ὡς Πίστιρος τὸ ἐμπόριον). On the whole, see Loukopoulou 2004a: 866. Due to the mention of Maronites in the inscription, del Barrio (2018: 490) does not rule out the possible influence of their dialect in the text (or, failing this, the dialect of its metropolis, Chios). 68 Velkov/Domaradzka 1999: 7; Loukopoulou (2004a: 867); and Hatzopoulos 2013: 15–16. 69 Ἀθηναγόρης, ἡμέρης (SEG 52.711, late-5th cent. BCE), Δημέα (SEG 64.606, 5th–3rd cent. BCE). 70 See the seminal works of Bubenik 1989; 1993; López Eire 1993; 1996; Brixhe/Hodot 2001. 71 Parissaki 2007: 193, 232 fn. 597. That Πολυάρητος is the Ionic form is assured by other occurrences in Samos, Paros and Erythrae, as well. 72 Parissaki 2007: 231–232; I.Thrake Aeg. 504. 73 These analogical aorists are also attested in Ionic prose, cf. Vessella 2016: 356.

Dialect Contact and Koineization  

Other standard koine forms such as πραγμo74 or νᾱός75 start to appear from the 2nd cent. BCE onwards, whereas the mixing of η and ᾱ arises at a later period (cf. τῇ ἱερίῃ but μνείας, Λολλίᾳ in E93, Abdera, imp. Rom.) and in two late epigrams from Maroneia.76 Among non-Greek anthroponyms, there is no fixed tendency in the rendering of some concrete names, as for instance, those Hellenised with the ending -αλκας/αλκης, such as Ῥοιμητάλκας (E84, Topeiros, 46–54 CE [x2]) or Σιτάλκας (E212, Maroneia, 2nd/1st cent. BCE), or Σεύθης/ας (cf. Σεύθης, E255, Maroneia, 1st cent. BCE, fun.).77

 Koineization From the 2nd cent. BCE, we detect other conventional phonetic features of the koine. Among these are the use of for (/iː/), which is overall attested in Maroneia from the late-3rd cent. BCE78 (and from the 1st cent. BCE in other regions).79 Later, more changes associated with vowel quantity and the reduction of diphthongs emerge. This is the case of instead of before a vowel

 74 Πράγμασιν (E6.16, Teos < Abdera, 166–160 BCE, decree), πράγματα (E5.17, Teos < Abdera, 166–160 BCE, decree; and E5.180, Β13, Samothracia < Maroneia, 41/42 or 46 CE, decree). 75 τοὺς ναούς but γαίης (E6, Teos < Abdera, 166–160 BCE, decree), ναυσίν (E168, Maroneia, 167 BCE, treaty, x2]). 76 Cf. γαίης and μοίρης vs. φοβ]ερᾶς, ἱερᾶς̣, ἡλικίαν̣ (E221, Nea Maroneia, 3rd–4th cent. CE) and Εὐδικίης in the metrical part vs. εὐνοίας in prose (E217, 2nd–3rd cent. CE). 77 More forms in fn. 49. The geographical distribution of Ῥοιμητάλκας/ης does not display significant differences (the previous, attested in other parts of Thrace, Attica, Scythia Minor, Chios, Mysia and Egypt, while the latter also occurs in Thrace, Moesia Superior, Chios and Mysia, cf. Dana 2014: ss.vv). Interestingly, there are no cases of Ῥοιματαλκo. Frequently, the choice of η/ᾱ is arbitrary (and there is no certainty about the long quantity of the vowel), as in Βεισούλας (E64, Abdera, 1st BCE/1st cent. CE). 78 Common words: πολειτῶν (E180, found in Samothrace, 41/42 or 46 CE, decree), σφραγεῖδι (E180, fr. A, found in Samothrace, 41/42 or 46 CE, decree, but σφραγιάμενοι), ἡμεῖν (E180, found in Samothrace, 41/42 or 46 CE, decree), τειμῆς (E181, 1st/2nd cent. CE, honorific decree). In Greek onomastics: Ἀφροδεισία (E240, Nea Maroneia, 3rd/2nd cent. BCE, fun.), Εἴσιδι (E199, 2nd/1st BCE?), Εἴσιδος, Δειφίλου, Νεικία, ἱερονείκης] (E212, 2nd/1st BCE, cat.), Ν]εικονέμεισιν (E204, 1st BCE, dedication), Βερονείκη (E303, Nea Maroneia, 1st cent. BCE/1st cent. CE, fun.), and in Latin names: Πρεῖσκα (E276, Nea Maroneia, 2nd CE, fun.), Πρείσκου] (E196, Nea Maroneia, 3rd CE). 79 Maroneia: Ποσιδῶνος (E204, 1st BCE, dedication); Topeiros: Δῖνις (E85, 46–54 CE), Ἀβδ̣ηρ]είτης, παροδεῖτα (E103, 2nd cent. CE), χείλια, δώσι (E88, 3rd CE); Abdera: παροδεῖτα (E104, 3rd cent. CE).

  Paloma Guijarro Ruano from the 1st cent. CE,80 or instead of ,81 and for ,82 both dated to the 2nd–3rd cent. CE onwards (and even later for ).83 Among other vocalic changes, contracted forms prevail from the 3rd cent. BCE,84 and ἑαυτός is rendered with ἁυ- or ἑα- forms from the 3rd cent. BCE.85 Koine (and Ionic) -σσ- and -ρρ- for Attic -ττ- and -ρσ- start to appear in the 3rd/2nd cent. BCE, although only -σσ- is attested in later examples.86 Once more, examples come chiefly from Maroneia, Topeiros and Abdera. Two 2nd-cent. BCE decrees from Abdera found in Teos document the first examples of μηθείς and οὐθείς.87 An important grammatical development is the analogical processes of levelling occurring in s-stems. No Ionic genitives in -εος have survived after the mid4th cent. BCE. We find, instead, Attic/koine -ους in three inscriptions (previously rendered with -ōς, cf. §§ 1.1, 1.2).88 However, most of the s-stems nouns (nom. -φάνης, -γένης, -μενης, -σθένης, -άλκης, -πείθης), dated between the mid-4th cent. BCE and the Roman Imperial period, present -ου genitives by analogy with masculine a-stems.89 A third analogical genitive in -κλήους appears from the

 80 Cf. ἀπολήας, πρεσβήα, συνεδρήαν (E180, Samothrace < Maroneia, 41/42 or 46 CE, decree), θῆναι (E312, Maroneia, 3rd cent. CE, fun.). 81 ὑ]ποκίμενον] (E100, Topeiros, 3rd cent. CE), oὐδίς (E466, Plotinopolis, 3rd cent. CE, epigram), κ̣ῖτε (E440, Traianoupolis, 3rd/4th cent. CE, fun.) and Σωκλείους (E183, Maroneia, 2nd cent. BCE, honorific decree). 82 κέ τε͂ς γυνεξίν, κέ [x3] (E396, Dione-Kokkinochomata, 235–238 CE, fun.), κεῖμ̣ε (E466, Plotinopolis, 3rd cent. CE, epigram) κ̣ῖτε (E440, Traianoupolis, 3rd/4th cent. CE, fun.). 83 κυμητή̣ριον (E96, Topeiros, 5th/6th cent. CE). 84 See, for instance, Ἀβδηριτῶν (E4, Abdera, mid-3rd cent. BCE). 85 αὑτóν (E451, Traianoupolis-Doriskos, ca. 250–200 BCE), καθ’ αὑτóν (E7, Teos < Abdera, 2nd cent. BCE), αὑτόν (E176 Maroneia, ca. 150–100 BCE, honorary decree), αὑτῇ (E438, Traianoupolis, 2nd CE, fun.), ἑατῷ (E321, Nea Maroneia, 3rd/4th CE, fun.), αὑτῆς] (E 475, Plotinopolis, 501 CE, fun.). 86 κρεῖσσον [x2], διαλλάσσον, ἧσσον (E205, Maroneia, mid-2nd/early 1st cent. BCE, aretalogy), διαφυλάσσειν] (E6, Teos < Abdera, 166–160 BCE, decree), καὶ κατὰ γῆν καὶ κατὰ θάλασσαν, τασσομένοις (E168, Maroneia, 167 BCE, treaty), φυλάσσηται (E180 B9, Samothrace < Maroneia, 41/42 or 46 CE, decree), τεσσεράκοντα (E88, Topeiros, 3rd cent. CE, fun.) etc. For -ρρ-, cf. θαρρῶν (E205, Maroneia, mid-2nd/early 1st cent. BCE, aretalogy) and NP ․․]ρρίας (E236, Maroneia, 3rd cent. BCE, fun.). 87 οὐθέν (E6, 166–160 BCE), μηθέν (E5, 166–160 BCE). 88 Διογένους, Προκλέους (E251, Maroneia, Hellenistic period?, fun.), Πλουτογένους (E290, Nea Maroneia, Rom. Imp. period, fun.), Διογένους (E446, Traianoupolis-Alexandroupolis, 2nd cent. CE, letter). 89 Εὐπείθου (E41, Abdera, ca. 350–300 BCE, fun.), Διογένου (E257, Maroneia, 3rd cent. BCE, fun.), Διογέ̣ν̣ο̣υ̣, Διογέ]νου (E259, Maroneia, Hellenistic period, fun.), Mητρο̣φάνου (E199,

Dialect Contact and Koineization  

2nd cent. BCE (cf. Ἀριστοκλήους, E62, Abdera; Μενεκλήους, E269, Maroneia), which contrasts with those in -κλέους (or -κλέōς, see supra). Finally, we find Ἀριστῆος (E241, Nea Maroneia, 3rd–2nd BCE, fun.). According to Bechtel (1923: 486; 1924: 153), -ηος derives from analogical levelling influenced by the nominative case,90 but for Nieto Izquierdo (2009: 123–124) it results from a hesitation between and due to the loss of contrastive distinctions in vowel quantity between /εː/ and /e/ before a vowel.91 In a similar way, other examples of analogical levelling emerge in the accusative and dative of s-stems: cf. Σωκλῆν (E183, Maroneia, 2nd cent. BCE, hon. decree), [Πυ]θογένην (E297, Nea Maroneia, 1st cent. BCE, fun.) and Ἁρφοκράτηι (E199, Maroneia, 2nd–1st cent. BCE?).92 An analogous process is observed in the 3rd-plural active and middle imperative verbal forms -τωσαν and -σθωσαν (for -των and -σθων respectively). The earliest occurrence is dated to the mid3rd cent. BCE (cf. [ἀπο]λογισάσθωσαν, [ἐπι]μελείσθωσαν, E4, Abdera, honorary decree), and they occur until one century later in official documents, mostly from Abdera.93 In a longer perspective, some Ionic features resist during the Hellenistic period and even later in anthroponymy. Despite the unstoppable linguistic colonization of the Attic-Ionic koine, some names still exhibit Ionic genitive forms in -εω, as we see in Maroneia, as late as the 2nd cent. BCE: cf. Π̣όρκης Πόρκεω94 (E184, Nea Maroneia, decree) and Ἀθήνεω (E168, Maroneia, 167 BCE, treaty). However, in the latter example, Ἀθήνεω is combined with other koine forms (cf. Μητροφάνου, φιλία, συμμαχία, ναυσίν etc.). The problem of distinguishing  Maroneia, 2nd/1st cent. BCE?), Μητροφάνου, Λεωσθένου, Διογένου, Καλλιγένου, Λεοντομένου, Διονυσιφ]άνου, Ἐπ̣άλκου (E212, Maroneia, 2nd/1st cent. BCE, cat.), Πλουτογένου (E286, Nea Maroneia, Rom. Imp. period, fun.). 90 Οn the influence of the vocative case on the hesitation between s-stems and a-stems, see Schulze 1893: 195–197; 1924. 91 Discussed forms in Bechtel 1924: 153, from Pantikapaion (Αὀτοκλῆος, IosPE II 140, 174, 4th cent. BCE); Delphi (Διοκλῆος, SGDI II 2259, 50/51 BCE); Hermion (Λακράτηος, IG IV, 78.6, 3rd cent. BCE; Κλεομήδηος, IG IV, 731, 2nd–1st cent. BCE?) and Crete (Δαμοκάδηος, IC II iii 6, Aptera, early-2nd BCE; Θυραιγένηος, IC II xi 4, Diktynnaion, 1st cent. BCE); there are also new examples (cf. Διοκλῆος, IK Byzantion 365, undated). 92 In the inscription there is also a koine form Mητρο̣ φάνου. 93 ἀν]αγραψάτωσαν, στησά]τωσαν, ἑέσθωσαν, κομισ̣ ά̣σ]θ̣ ωσαν (E5, Teos < Abdera, decree), ἀναγραψάτωσαν, δότωσαν, ἑλέσθωσαν (E7, Teos < Abdera, proxeny decree), ἀναγραψάτωσαν (E8, Abdera, honorary decree), ἀναθέτωσαν (E9, Abdera, honorary decree), διιέτωσαν, χορηγείτωσαν (E168, Maroneia, 167 BCE, treaty). 94 Πόρκης is considered a Greek name by Parissaki 2007: 233, while Detschew 1976 and Brixhe 2006: 139 suggest a possible Thracian origin; included in Dana 2014, s.v. PORK.], Πορκης.

  Paloma Guijarro Ruano between koine and Ionic forms emerges concerning genitives like Σ̣ουσᾶ (E482, prov. inc., aet. rom., dedication) or Ἀσκλᾶ (E483, prov. inc. [Komotini Mus.], 2nd/3rd cent. CE, dedication). A similar situation is reflected in the final outcome of γιγν- > γιν-, which is attested from the 2nd cent. BCE.95 An artificial Ionic veneer is observed in Κλεοπάτρη (E154, Molyvoti, ca. 350 BCE, fun.), which contrasts with Κλεοπάτρα (E420, Zone, ca. 350 BCE, fun.) and Σωσιπάτρα (E229, Maroneia, early-3rd cent. BCE, fun.).96 Finally, it is possible that some late Thracian names could have maintained Ionicisms in the light of some genitive forms, such as non-contracted genitives in -εος, attested in Αυλ̣ ο̣ υ̣[ζ]ε̣ ν̣ εος (E478, Komara, 2nd–3rd cent. CE) or Διασενεος, Αὐλουζενεος (E104, Abdera, 3rd cent. CE).97 However, their adaptation to a particular Greek declension is not consistent so that we find alternative renderings of the same names, cf. Διασενεως (E387, Dyme, 1st cent. CE, fun.) or Αὐλουζειδος (E212, Nea Maroneia, 2nd–1st cent. BCE). Be that as it may, these graphic inconsistencies and variations when writing similar names98 and the loss of vowel-quantity distinction should be treated cautiously, since they can lead us to wrong hypotheses. While it is true that factors like chronology or geographical distribution can play an important role in the analysis of these names, the extant evidence from our region is not sufficient enough to draw any definite conclusions.99

 95 γ]ινώσκ̣ωσιν (E5, Teos < Abdera, 166–160 BCE, decree), γινόμενος (E10, Abdera, 2nd cent. BCE, honorary decree), γίνεται (E8, Abdera, 2nd cent. BCE, honorary decree), γεινόμενον, ἐπιγινώσκωσιν (E183, Maroneia, 2nd cent. BCE, honorary decree), γινώσκων (E205, Maroneia, mid-2nd–early 1st cent. BCE), παρα]γινόμενοι, παραγίνεσθαι (E185, Maroneia, 131 CE, edict); but cf. γίνομαι in Thasos (IG XII Suppl. 349, late 5th cent. BCE). 96 There is no context apart from names. According to del Barrio 2018a: 484, 487, Κλεοπάτρη and nom. πάτρη, acc. πάτρην (in Thasos and Samos) are hyper-Ionicisms since this -a was originally short. 97 In both examples the nominative Αὐλούζενις also appears in the text. Compare with gen. Αυλουζενεω (IGBulg III,2 1794, Hadrianopolis, 2nd cent. CE date after Dana 2014: 19). On the variation of this name (nom. -ζηνεις/ζενις/ζηνης/σανις/ζανος), see Dana 2014: 18–22. 98 See Dana 2014: XCIV–XCV. 99 See Dana’s 2014 hypothesis about the existence of four main Thracian onomastic regions, and Guijarro Ruano (forthcoming) for its application into Aegean Thrace.

Dialect Contact and Koineization  

 Conclusions As we have seen, due to the scarcity of the available material, linguistic analysis is mainly limited to anthroponymy in pre-Hellenistic times, and to some concrete phonetic and morphological categories in later periods. Distinctive Ionic features arise from the period spanning the mid-6th cent. BCE (in inscriptions whose Thasian origin is still questionable) and the 5th and 4th centuries, while during the 4th cent. BCE they coexist with Attic(-koine) traits. The Ionic dialect residually survives in personal names until the 2nd cent. BCE (leaving aside possible remnants of Hellenised anthroponyms). Furthermore, isolating intradialectal differences in this area remains a challenging task, due to the lack of extensive parallels and to extra-linguistic factors, such as the early influence of Athens and the technical complexity of identifying some colonies. Indeed, Attic influence seems to emerge at an early date, around the mid-5th cent. BCE, in contracted genitives such as Ἀλκιβιάδō (E405–407, Zone, fun.) and Ἀνταγόρō (E109, Molyvoti, epigram). These forms, along with the reading of Μεγαδο[ ] (E39, Abdera, ca. 400–350 BCE, fun.) as Μεγάδω and as Πρωτίω (E30, Abdera < Thasos?, mid-6th cent. BCE), could be taken as examples of Ionic contraction -εω > -ω in the genitives of masculine a-stems. However, all these forms admit alternative interpretations (Μεγᾶδο[ς], Πρωτίου/Πρωτίο), so that they do not constitute definitive arguments in favour of this change. As things stand and in view of the available evidence, Attic traits in inscriptions before the arrival of the Hellenistic koine can easily be explained by the strong interest and presence of Athens (and therefore of Great Attic) in the region after the Persian wars: many of these cities were members of the Delian League and were included in the Athenian tribute lists. Ionic-Attic convergence appears hesitantly in E3 (Abdera, bef. 350 BCE, law), where the Ionic η or uncontracted forms (δυσεντερίης, [ὑ]πνόδεος) are found side by side with other Attic features visible in grammatical words (τοῖς, ἐάν). An early diffusion of the Attic dialect in the northern Aegean coast matches with previous accounts of its rapid spread in Macedonia, from the early 5th cent. BCE,100 and with its later expansion toward the western and northern coasts of the Black Sea, where already koine features appear shortly after, from the 4th cent. BCE onwards.101 Overall, these were neighbouring regions without clear political borders and with blurry geographical limits.

 100 Brixhe/Panayotou 1988. 101 Bechet 2009: 107–111; Slavova 2009: 209; Tokhasiev 2009.

  Paloma Guijarro Ruano In general lines, we see how the patterns of dialectal convergence and dialect levelling led to the adoption of the koine as the sole common linguistic vehicle for communication. These phenomena relate to a process of simplification of the grammar that gradually gave rise, through koineization, to a new variety, the emerging Attic-Ionic koine.

Bibliography Alonso Déniz, A. (2018), “The dialect of Thasos and the transmission of Archilochus’ fragments”, in: G.K. Giannakis/E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects. From Central Greece to the Black Sea, Berlin/Boston, 531–560. Arrington, N.T./Terzopoulou, D./Tasaklaki, D./Lawall, M.L./Brellas, D.J./White, C.E. (2016), “Molyvoti, Thrace, Archaeological Project: 2013 Preliminary Report”, Hesperia 85, 1–64. Αρχαία Ζώνη I = Tsatsopoulou-Kaloudi, P./Brixhe, C./Pardalidou, C., Αρχαία Ζώνη I: Το Ιερό του Απόλλωνα, Vol. 1, Komotini 2015. Bechet, F. (2009), “Les colonies grecques du Pont gauche: histoire et langue”, in: G. Vottéro (ed.), Le grec du monde colonial Antique: Le N. et N.-O. de la Mer Noire. Actes de la table Ronde de Nancy, 28–29 septembre 2007, Nancy, 81–118. Bechtel, F. (1917), Die historischen Personennamen des Griechischen bis zur Kaiserzeit, Halle (repr. Hildesheim 1964). Bechtel, F. (1923), Die griechischen Dialekte II. Die westgriechischen Dialekte, Berlin. Bechtel, F. (1924), Die griechischen Dialekte III. Der ionische Dialekt, Berlin. Brixhe, C. (1963), “Les noms de personnes en-ᾶς/-ᾶ, -ᾶς/-ᾶτος et-ης/-ητος dans le dialecte pamphylien”, RÉG 76, 10–33. Brixhe, C. (2006), “Zôné et Samothrace: lueurs sur la langue thrace et nouveau chapitre de la grammaire comparée?”, CRAI 150, 1, 121–146. Brixhe, C. (2008), “Un phénomène aréal: la substitution de< E> à < O> en finale en Thrace, à Thasos et en Thessalie”, in: Μ. Θεοδωροπούλου (ed.), ΘΕΡΜΗ ΚΑΙ ΦΩΣ. LICHT UND WÄRME. Αφιερωματικός Τόμος στη μνήμη του Α.-Φ. Χριστίδη, Thessaloniki, 215–223. Brixhe, C. (2010), “Quelle koiné en Macédoine au début de notre ère? Filetest de Leukopétra”, in: J. Méndez Dosuna/F. Cortés Gabaudán (eds.), Dic mihi, Musa, virum. Homenaje al profesor Antonio López Eire, Salamanca, 61–68. Brixhe, C. (2015) = Αρχαία Ζώνη I. Brixhe, C. (2018), “Thracian”, in: J. Klein/B. Joseph/M. Fritz (eds.), HSK 41.3, 1850–1854. Brixhe, C./Hodot, R. (2001), “De l’attique à la koiné”, MSLP nouvelle série 11, 29–46. Brixhe, C./Panayotou, A. (1988), “L’atticisation de la Macédoine: l’une des sources de la koiné”, Verbum 11, 245–260. Brixhe, C./Panayotou, A. (1997), “Le Thrace”, in: F. Bader (ed.), Langues indo-européennes, Paris, 179–203. Bubenik, V. (1989), Hellenistic and Roman Greece as a Sociolinguistic Area, Amsterdam. Bubenik, V. (1993), “Dialect contact and Koineization: the case of Hellenistic Greek”, IJSL 99, 9–24.

Dialect Contact and Koineization  

Chankowski, V./Domaradzka, L. (1999), “Réédition de l'inscription de Pistiros et problèmes d’interprétation”, BCH 123.1, 246–258. Chryssanthaki-Nagle, K. (2004), “Le monnayage d’Orthagoreia, Revue numismatique 160, 49–62. Chryssanthaki-Nagle, K. (2007), L’histoire monétaire d’Abdère en Thrace (VIe s. av. J.-C. – IIe s. ap. J.-C.), Athens. Crespo, E. (2009), “La difusión temprana del dialecto ático en el Peloponeso”, in: A. Martínez Hernández (ed.), Estudios de epigrafía griega, La Laguna, 137–144. Crespo, E. (2010), “El proceso de configuración y fijación de la koiné en el siglo IV”, in: J.V. Méndez Dosuna/F. Cortés Gabaudán (eds.), Dic mihi, mvsa, virvm, Salamanca, 139–146. Curbera, J. (2013), “Simple names in Ionia”, in: R. Parker (ed.), Personal Names in Ancient Anatolia, Oxford, 107–143. Dana, D. (2014), Onomasticon Thracicum: répertoire des noms indigènes de Thrace, Macédoine Orientale, Mésies, Dacie et Bithynie, Athens. Dana, D. (2015), “Inscriptions”, in: J. Valeva/E. Nankov/D. Graninger (eds.), A Companion to Ancient Thrace, Malden, MA/Oxford, 243–264. Dana, D. (2017), “Hellénisation par suffixation: noms non grecs et suffixes grecs”, in: A. Alonso Déniz et al. (eds.), La suffixation des anthroponymes grecs antiques (SAGA). Actes du colloque international de Lyon, 17–19 septembre 2015, Université Jean-Moulin-Lyon 3, Genève, 201–223. Del Barrio = Del Barrio Vega, M.L., “Ionic”, in: G.K. Giannakis et al. (eds.), Encyclopedia of Ancient Greek Language and Linguistics, vol. 2, Leiden/Boston 2014, 260–267. Del Barrio Vega, M.L. (2015), “Η ιωνική του βόρειου Αιγαίου (από τον Στρυμόνα μέχρι τον Έβρο”, in: G.K. Giannakis (ed.), Ο γλωσσικός χάρτης της κεντρικής και βόρειας Ελλάδας κατά την αρχαιότητα, Thessaloniki, 153–182. Del Barrio Vega, M.L. (2018a), “The Ionic of North Aegean (from the River Strymon to the River Hebrus)”, in: G.K. Giannakis/E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects. From Central Greece to the Black Sea, Berlin/Boston, 473–510. Del Barrio Vega, M.L. (2018b), “The Greek language in the Black Sea”, in: G.K. Giannakis/ E. Crespo/P. Filos (eds.), Studies in Ancient Greek Dialects. From Central Greece to the Black Sea, Berlin/Boston, 511–530. Detschew, D. (1976), Die thrakischen Sprachreste, mit Bibliographie 1955–1974 von Živka Velkova, Viena. Dubois, L. (2017), “Des noms en -ᾶς”, in: A. Alonso Déniz et al. (eds.), La suffixation des anthroponymes grecs antiques (saga). Actes du colloque international de Lyon, 17–19 septembre 2015, Université Jean-Moulin-Lyon 3, Genève, 299–332. Flensted-Jensen, P. (2004), “Thrace from Axios to Strymon to Nestos”, in: M.H. Hansen/ T.H. Nielsen, (eds.), An Inventory of Archaic and Classical Poleis, Oxford, 810–853. Fournier, J./Hamon, P./Parissaki, M.G. (2015), “Recent epigraphic research in Thasos, Aegean Thrace & Samothrace (2005–2015)”, Archaeological Reports 61, 75–93. Guijarro Ruano, P. (2019), “The northern coast of the Aegean Sea: a linguistic overview”, Linguarum Varietas 8, 63–83. Guijarro Ruano, P. (2023). “¿Μέγας, Μεγᾶς ο Μεγάδης? I.Thrake Aeg. E39 y los antropónimos derivados de μέγα(ς) en el norte del Egeo”, ZPE 225, 82–88. Guijarro Ruano, P. (forthcoming), “Onomastics and linguistic contact in the Greek colonies of Aegean Thrace”, in: A. Alonso Déniz et al. (eds.), Contacts linguistiques en Grèce ancienne: diachronie et synchronie. Actes du colloque international Coliga (7–9 avril 2021), Lyon, 35–51.

  Paloma Guijarro Ruano Hălmagi, D. (2015), “Dan Dana, Onomasticon Thracicum. Répertoire des noms indigènes de Thrace, Macédoine Orientale, Mésies, Dacie et Bithynie“, Studia Antiqua et Archaeologica 21.2, 259–271. Hatzopoulos, M. (2013), “Τὰ τῶν ἐμποριτῶν φιλάνθρωπα: observations on the Pistiros inscription (SEG XLIII 486)”, in: P. Martzavou/N. Papazardakas (eds.), Epigraphical Approaches to the Postclassical Polis: 4th Century BC to 2nd Century AD, New York, 13–21. Horrocks, G. (20102), Greek: A History of the Language and its Speakers, Oxford/Malden, MA. I.Thrace orientale byz. = Asdracha, C., Inscriptions protobyzantines et byzantines de la Thrace orientale et de l’ile d’Imbros (IIIe-XVe siecles): présentation et commentaire historique, Athens 2003. I.Thrake Aeg. = Loukopoulou, L.D./Parissaki, M.G./Psoma, S./Zournatzi, A., Ἐπιγραφὲς τῆς Θράκης τοῦ Αἰγαίου μεταξὺ τῶν ποταμῶν Νέστου καὶ Ἕβρου (Νομοὶ Ξάνϑης, Ροδόπης καὶ Ἕβρου), Athens 2005. Kaczko, S. (20162), “Il giambo”, in: A.C. Cassio (ed.), Storia delle lingue letterarie, Milano, 289–304. Knitl, E. (1938), Die Sprache der ionischen Kykladen nach den inschriftlichen Quellen, München. LGPN IV = Fraser, P.M./Matthews, E./Catling, R.W.V., Lexicon of Greek Personal Names IV: Macedonia, Thrace, Northern Regions of the Black Sea, Oxford 2005. López Eire, A. (1993), “De l’attique à la koiné”, in: C. Brixhe (ed.), La koiné grecque antique: I. Une langue introuvable?, Nancy, 41–57. López Eire, A. (1996). “L’influence de l’ionien-attique sur les autres dialectes épigraphiques et l’origine de la koiné”, in: C. Brixhe (éd.), La koiné grecque antique. II La concurrence, Nancy, 7–42. Loukopoulou, L.D. (2004a), “Thrace from Strymon to Nestos”, in: M.H. Hansen/T.H. Nielsen (eds.), An Inventory of Archaic and Classical Poleis, Oxford, 854–869. Loukopoulou, L.D. (2004b), “Thrace from Nestos to Hebros”, in: M.H. Hansen/T.H. Nielsen (eds.), An Inventory of Archaic and Classical Poleis, Oxford, 870–884. Loukopoulou, L.D./Psoma, S. (2008), “Maroneia and Stryme revisited: some problems of historical topography”, in: L. Loukopoulou/S. Psoma (eds.), Thrakika Zetemata I (Meletemata 58), Athens, 55–86. Masson, O. (1984), “Quelques noms de magistrats monétaires grecs”, Revue de Numismatique 6 (24), 48–60 [= OGS II, 427–439]. May, J.M. (1966), The Coinage of Abdera (540–345 B.C.), London. Méndez Dosuna, J. V. (1993), “Metátesis de cantidad en jónico-ático y heracleota”, Emerita 61, 95–134. Nieto Izquierdo, E. (2009), Gramática de las inscripciones de la Argólide, Tesis doctoral inédita, Madrid. Panayotou, A. (2007), “Greek and Thracian”, in: A.-P. Christidis (ed.), A History of Ancient Greek. From the Beginnings to Late Antiquity, Cambridge, 745–775. Parissaki, M.G. (2007), Prosopography and Onomasticon of Aegean Thrace, Paris. Pouilloux, J. (1954), Recherches sur l’histoire et les cultes de Thasos. I: de la fondation de la cité à 196 avant J.-C., vol. 1, Paris. Psoma, S./Karadima, C./Terzopoulou, D. (2008), The Coins from Maroneia and the Classical City at Molyvoti: A Contribution to the History of Aegean Thrace, Athens. Schönert-Geiss, E. (1987), Griechisches Münzwerk: Die Münzprägung von Maroneia, 2 vols., Berlin.

Dialect Contact and Koineization  

Schulze, W. (1893), “Koseformen im Griechischen”, ZVS 32, 195–1977 (= Kleine Schriften 1966, 310–312). Schulze, W. (1924), “Zur Bildung des Vokativs im Griechischen und im Lateinischen”, in: ᾽Aντίδωρον. Festschrift J. Wackernagel, Göttingen, 240–254 (= Kleine Schriften 1966, 82–96). Slavova, M. (2009) “The Greek language on the Bulgarian coast of the Black Sea in antiquity (6th cent. BC-1st cent. BC) ”, in: G. Vottéro (ed.), Le grec du monde colonial Antique: Le N. et N.-O. de la Mer Noire. Actes de la table Ronde de Nancy, 28–29 septembre 2007, Nancy, 195–220. Sowa, W. (2020) “Thracian”, Palaeohispanica 20, 787–817. Stüber, K. (1996), Zur dialektalen Einheit des Ostionischen, Innsbruck. Teodorsson, S.-T. (2014), “Attic”, in: G.K. Giannakis et al. (eds.), Encyclopedia of Ancient Greek Language and Linguistics, vol. 1, Leiden/Boston, 194–195. Threatte, L. (1980), The Grammar of Attic Inscriptions Ι: Phonology, Berlin. Thonemann, P.J. (2006), “Review of Loukopoulou, L.D./Parissaki, M.G./Psoma, S./Zournatzi, A. (2005), Ἐπιγραφὲς τῆς Θράκης τοῦ Αἰγαίου μεταξὺ τῶν ποταμῶν Νέστου καὶ Ἕβρου (Νομοὶ Ξάνϑης, Ροδόπης καὶ Ἕβρου), Αϑήνα”, CR 56, 2, 458–459. Thumb, A./Scherer, A. (1959), Handbuch der griechischen Dialekte, vol. 2 (2nd ed. by A. Scherer), Heidelberg. Tokhasiev, S. R. (2009), “De nouvelles données sur l’histoire des formes les plus anciennes de la koiné dans le nord de la Mer Noire”, in: G. Vottéro (ed.), Le grec du monde colonial Antique: Le N. et N.-O. de la Mer Noire. Actes de la table Ronde de Nancy, 28–29 septembre 2007, Nancy, 33–49. Tzitzilis, C. (2014), “Greek and Thracian”, in: G.K. Giannakis et al. (eds.), Encyclopedia of Ancient Greek language and Linguistics, vol. 2, Leiden/Boston, 83–89. Velkov, V./Domaradzka, L. (1999), “Kotys I (383/382–359 av. J.-C.) et l’emporion Pistiros de Thrace”, BCH 118, 1–15. Vessella, C. (2016), “La prosa”, in: A.C. Cassio (ed.), Storia delle lingue letterarie, Milano, 354– 384. Yanakieva, S. (2018), “The Thracian language”, ORPHEUS 25, 26–68.

Mark Janse and Johan Vandewalle

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno] “Take Away” as an Example of Turkish Pattern Replication Abstract: In this chapter we investigate παιρ-παίνω [per-péno], a unique phrasal compound attested in Cappadocian, but not in any other Modern Greek dialect. It is very different from the productive Modern Greek verbal dvandva compounds: it lacks a compound marker and, most importantly, the first member is inflected for TMA as opposed to verbal dvandvas like Modern Greek μπαίν-οβγαίνω [benovʝéno], where the verbal stem of the first member is always imperfective. By contrast, the first member of the perfective indicative πηρ-πήγα [pirpíɣa] and subjunctive παρ-πάγω [par-páɣo] have the ablaut of the independent forms πήρα [píra] and πάρω [páro]. Research on the basis of the Turkish National Corpus revealed that a parallel structure exists in Turkish, consisting of the verbal- “take” and a following deictic verb. The most frequent combination is alıp götür- “take and carry away”, a converb construction with a transitive deictic verb παιρ-παίνω [per-péno], however, is a phrasal compound with suspended affixation and an intransitive deictic verb. Turkish has a perfect match: al-git“take and go”, with an intransitive deictic verb and a construction very similar to παιρ-παίνω [per-péno]. As such, παιρ-παίνω [per-péno] is an example of Turkish pattern replication (PAT) in Cappadocian.

 The bulk of the research for this chapter was done within the framework of a collaborative research project funded by the HERA network entitled “Multilingualism and Minority Languages in Ancient Europe” (HERA.15.029). A preliminary version of this chapter was presented at the 8th International Conference on Greek Linguistics in Ioannina (Janse 2007). The authors would like to thank Eline Daveloose for her careful reading of the text. https://doi.org/10.1515/9783110779684-007

  Mark Janse and Johan Vandewalle

 Introduction The history of Cappadocian has been told many times, so we will restrict ourselves to what is strictly relevant for the purpose of the present chapter.1 Cappadocian is an East Asia Minor Greek dialect (Fig. 4), which developed after the Hellenization of Cappadocia in the Hellenistic and Roman periods from the Asia Minor Koine.2 After the defeat of the Byzantine army by the Seljuk Turks in the battle at Manzikert in 1071, Cappadocia was cut off from the rest of the Greekspeaking world and Cappadocian became increasingly turkicized during the Ottoman period. The result of this Turkicisation is what has come to be known in the literature as a bilingual mixed language.3 The Greek component of Cappadocian is essentially an East Asia Minor variety of Late Medieval Greek.4 The Turkish component is a variety of Central Anatolian Turkish, with features of both Western and Eastern Anatolian Turkish.5 The Turkicisation of Cappadocian was such that Thomason and Kaufman presented it as the first case study in their classic study of language contact, labelling it “an excellent example of heavy borrowing” (1988: 215), category 5 (the highest) in their borrowing scale (1988: 74 ff.).6 In this chapter we will focus on a hitherto poorly understood case of heavy borrowing, labelled pattern replication (PAT) by Matras and Sakel.7 It is defined as follows: “PAT describes the case where only the patterns of the other language are replicated, i.e., the organization, distribution and mapping of grammatical or semantic meaning, while the form itself is not borrowed” (Sakel 2007: 15). The case in question is the compound verb παιρπαίνω [perpéno] “take away”, which at first sight looks like a coordinative or co-compound, called dvandva in the Pāṇinian tradition,8 but on closer inspection turns out to be something fundamentally different.  1 The classic work on Cappadocian and other varieties of East Asia Minor Greek is Dawkins (1916); for recent accounts with further references see Janse (2002: 347–359; 2020a; 2023: §1), Karatsareas (2011: 10–64). 2 On the definition of East Asia Minor Greek see Janse (2008: 190; 2020b: 202f.); cf. Dawkins (1916: 213). 3 See Janse (2009: 104) with further references. 4 See Janse (2020b) for a detailed study of Late Medieval Greek features in Cappadocian. 5 On the problematic classification of the Anatolian Turkish dialects, see Kowalski (1931); Korkmaz (1975–76; 1990); Boeschoten (1991); Brendemoen (1998). 6 Cappadocian is also assigned a prominent position in Winford (2005: 401–409). 7 See Matras and Sakel (2007); Sakel (2007); Matras (2009: 234–74). 8 Sanskrit द्वंद्व dvaṃdvá (Pāṇini, Aṣṭādhyāyī 2.4.2).

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

 The Cappadocian data The verb παιρπαίνω [perpéno] “take away” is uniquely attested in Cappadocian. It is not found in the other East Asia Minor dialects Pharasiot, Pontic, and Silliot (Lycaonian), nor in any other Greek dialect for that matter. It is quite obviously some sort of compound composed of the verbs παίρω [péro] “take” and παίνω [péno] “go” (Dawkins 1916: 631).9 Mavrochalyvidis, Kesisoglou and Phosteris gloss the verb as “παίρνω κάτι καὶ τὸ πηγαίνω”,10 with the latter verb used transitively, as is very common in Modern Greek, which immediately raises an interesting question: who needs a mysterious compound verb meaning “take away”, if you already have the two constituent parts as separate verbs to express the same meaning? Consider the following example from Delmeso:11 ()

πατισ̌άχ-ϊς

τσ̌η

ναίκα-τ

πήρ-εν

do

patiʃáx-ɯs

tʃi

néka-t

pír-en

do

king-nom.sg.

the.f.acc.sg.

wife.acc.sg.sg.poss.

take.pfvind.sg.

sg.acc.

“the king took his wife” και

πήγ-εν

do

σ-ο

μεɩ̯dέν

ce

píʝ-en

do

s-o

meɪdén

and

go.pfv.-ind.sg.

sg.acc.

to-the.n.acc.sg.

public space

“and he took her to the public space” (Delmeso, D)

The verb πήγεν [píʝen] is here used transitively, but this use is extremely rare in Cappadocian and (1) is actually the only example in the folktales recorded by Dawkins (1916: 304–464), Kesisoglou (136–160), Mavrochalyvidis & Kesisoglou  9 Alektoridis (1883: 501), Archelaos (1899: 258 s.v. παγαίνω), Krinopoulos (1889: 59), Pharasopoulos (1895: 122) and Karphopoulos (2008: 125) write περπαίνω, deriving it from an otherwise unattested περιπηγαίνω, an etymology rightly rejected for semantic reasons by Dawkins (1916: 631) with reference to Krinopoulos and Pharasopoulos. 10 Cf. Mavrochalyvidis/Kesisoglou (1960: 115); Phosteris/Kesisoglou (1960: 37); Mavrochalyvidis (1990: 636). 11 For the geographical distribution and subgrouping of the Cappadocian dialects see Figure 2 and Table 1. Abbreviations follow the Leipzig glossing rules (www.eva.mpg.de/lingua/ resources/glossing-rules.php), with the following additions: AUGM = augment, CM = compound marker (Ralli 2008), CONV. = converb, IPFV = imperfective, PERS = person, PFV = perfective, PROJ = projection suffix (van Schaaik 2020: 271), TMA = tense-mood-aspect.

  Mark Janse and Johan Vandewalle (1960: 186–220) and Phosteris & Kesisoglou (1960: 98–126). In other cases where παίρω [péro] and παίνω [péno] are combined, the latter is used intransitively, as in the following example, where all the verbs have an overt object clitic pronoun except πήγεν [píʝen]: ()

πήρ-εν

τσ̌ην

και

πήγ-εν

σ-ο

χαμάμ

pír-en

tʃin

ce

píʝ-en

s-o

xamám

take.pfv.-ind.sg.

sg.f.acc. and

go.pfvind.sg.

tohamthe.n.acc.sg. mam

“he took her and went to the hammam”

και

λούσ-εν

τσ̌ην

και

έ-πλυν-έν

τσ̌ην

ce

lús-en

tʃin

ce

é-pl.in-én

tʃin

and

bath.pfv-ind.sg. sg.f.acc.

and

augm.-wash.pfv-ind.sg.

sg.f.acc.

“and he bathed her and washed her” (Delmeso, Df.)

Instead of transitive παίνω [péno], παιρπαίνω [perpéno] is used throughout. The difference between the two verbs, the former intransitive and the latter transitive, is particularly evident in the following example, where (3b) is the girl’s reply to the boy’s initial statement (3a): (a)

να

σε

πάρ-ω

na

se

pár-o

prt

sg.acc.

take.pfv-subj.sg.

“I will take you”

να

πέγ-ω12

σ-o

βαβά-μ

na

péɣ-o

s-o

vavá-m

prt

go.pfv.-ind.sg.

to-the.m.acc.sg.

father-sg.poss.

“I will go to my father” (Delmeso, D)

 12 The raising of [a] to [e] in πέγω [péγo] instead of πάγω [páγo] at Delmeso originated in the third person singular as a result of palatalization (Dawkins 1916: 65 f.): πάγει [páʝi] → πάɩ̯χ [páɪç] → πέχ [péç], which resulted in paradigmatic levelling (Hock/Joseph 2009: 152): 1sg. πέγω [péɣo], 2sg. πές [pés], 1pl. πέμ [pém], 2pl. πέτε [péte], 3pl. πέν [pén] (Dawkins 1916: 63).

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

(b)

εμένα

αν

με

παρ-σ̌ά-ς13

eména

an

me

par-ʃá-s

sg.acc.

if

sg.acc.

take-go.pfv.subj.-sg.

“as for myself, if you take me away” να

με

πάρ-ουν

άσ-α

χέρɩ̯-α-ς

na

me

pár-un

ás-a

çérj-a-s

prt

sg.acc.

take.pfv.-subj.pl.

from-the.n.acc.pl.

hand-pl.-sg.poss.

“they will take me from your hands” (Delmeso, D)

The following examples further illustrate the use of παιρπαίνω [perpéno]: ()

σ-το

τσ̌εσ̌μέ

ποτισμέν-ο

παιρ-παίν-ω

σε

s-to

tʃeʃmé

potizmén-o

per-pén-o

se

to-the.n.acc.sg.

fountain

watered-acc.sg. take-go.ipfv.-ind.sg.

sg.acc.

“watered I take you to the fountain”

και

λιψασμέν-ο14

φέρ-ω

σε

ce

lipsazmén-o

fer- o

se

and

thirsty-acc.sg.

bring.ipfv.-ind.sg.

sg.acc.

“and thirsty I bring you back” (Axo, MK)

()

παιρ-παίν-ισ̌κ-εν

dα

σ-ο

λουτρ-ό

per-pén-iʃc-en

da

s-o

lutr-ó

take-go.ipfv.-ipf.-ind.sg.

pl.acc.

to-the.n.acc.sg.

bath-acc.sg.

“he used to take them to the bath” (Aravan, PK)

 13 On the peculiar form of the perfective subjunctive 1sg. παρ-σ̌ά-ω, 2sg. παρ-σ̌ά-ς etc. see Dawkins (1916: 137 f.) and especially Janse (2020b: 217 f.). 14 Medieval Greek διψασμένος.

  Mark Janse and Johan Vandewalle

()

πηρ-πήγ-αν

do

σ

ενα

τζ ̌ αdΐ

καρά

pir-píɣ-an

do

s

ena

dʒadɯ́

kará

take-go.pfv.-ind.pl.

sg.acc.

to

a.acc.sg.

witch

wife

“they took her to a witch-wife” (Ghurzono, D)

()

βασ̌ιλɩ̯-ός

κίρεψ-εν15

καμbηλɩ̯-ού

κιρɩ̯άς16

vaʃiʎ-ós

kíreps-en

kambilʎ-ú

cirjás

king-nom.sg.

want.pfv.-ind.g.

camel-gen.sg.

meat

“the king asked for camel-flesh” και

να

do

παρ-πά-ω

ce

na

do

par-pá-o

and

prt

sg.acc.

take-go.pfv.-subj.sg.

“and I will take it [to him]” (Floïta, D)

()

ας

γɩ̯ομώσ-ω

ενα

σανdΐχ

λίρ-ες

as

ʝomós-o

ena

sandɯ́x

lír-es

prt

fill.pfv.-subj.sg.

a.acc.sg.

chest

lira-pl.

“let me fill a chest with liras” και

ας

παρ-πά-με

τα

ce

as

par-pá-me

ta

and

prt

take-go.pfv.-subj.pl.

pl.acc.

“and let us take it away [to her]” (Axo, D)

()

έπαρ-Ø

το

χαɩ̯βάνι-ς

épar-Ø

to

xaɪváni-s

take.pfv-imp.sg.

the.acc.sg.

animal-sg.poss.

“take your animal”

 15 Medieval Greek γυρεύω. 16 Medieval Greek κρέας, κρεάς, κριάς.

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

και

παρ-πάγ-ε

το

ας

το

λουτρό

οπίσω

ce

par-páʝ-e

to

as

to

lutró

opíso

and

take-go.pfv.-imp.sg. sg.acc.

from

the.acc.sg.

bath

behind

“and take it away from behind the bath” (Aravan, PK)

In a characteristically lucid and concise account, Dawkins explains the formation of παιρπαίνω [perpéno] as follows: “The word is […] likely to be a compound of παίρω and παίνω in its transitive sense “I take and carry off” like such words as μπαινοβγαίνω, ἀνοιγοκλείνω etc. with the -o- after παιρmissing owing to the influence of the use of the Turkish aləp gel “bring (i.e., taking come)”, or, still closer, the Turki alip barmaq […] “to take away”, i.e., alip “having taken” and barmaq “to go” […], and apkitmak (= Turkish aləp gitmek) with the same meaning” (Dawkins 1916: 631 s.v. παιρπαίνω)17

We will discuss the Greek and Turkish comparanda in the next sections.

 The Greek perspective: verbal dvandva compounds At first sight, παιρ-παίνω [per-péno] looks indeed very similar to Modern Greek verbal dvandva compounds like ανοιγ-ο-κλίνω [aníɣ-o-klíno] “open-CM-close”, τρωγ-ο-πίνω [troɣ-o-píno] “eat-CM-drink”, μπαιν-ο-βγαίνω [ben-o-vʝéno] “go inCM-go out” or πηγαιν-ο-έρχομαι [piʝen-o-érxome] “go-CM-come”, which have the following properties c.q. constraints (Ralli 2009: 53 ff.):18 (i) Phonologically and morphologically, they behave like true compounds in that they have a single stress, are inflected at the right-hand edge, and its members are linked by the compound marker [o]: πηγαιν-o-έρχ-εται *πηγαίν-ει-έρχ-εται

[piʝen-o-érç-ete] [piʝén-i-érç-ete]

“go.ipfv-cm-come.ipfv.-ind.3sg.” ~ “go.ipfv-3sg.-come.ipfv.-ind.3sg.”

 17 We have adapted Dawkins’ use of italics and quotation marks to the style of the present chapter. 18 On verbal dvandva compounds in Modern Greek see Kiparsky (2009); Manolessou and Tsolakidis (2009); Nicholas/Joseph (2009); Ralli (2009); Ralli/Karasimos (2010); Janse (forthcoming).

  Mark Janse and Johan Vandewalle (ii) Semantically, the two members express compatible, often synonymous, or opposite meanings, have parallel argument structures, and the meaning of the compound is a conjunction of the meanings of its subparts: ανοίγ-ο-κλίν-ει μπαιν-ο-βγαίν-ει τρωγ-ο-πίν-ει

[aniɣ-o-klín-i] [ben-o-vɣén-i] [troɣ-o-pín-i]

“open-cm-close.ipfv.-ind.3sg.” “go in-cm-go out.ipfv.-ind.3sg.” “eat-cm-drink.ipfv.-ind.3sg.”

Ralli mentions a few other properties which are not directly relevant for our purposes, with the exception of the “bare-stem constraint” (Ralli/Karasimos 2010), which “requires the first constituent of a compound to be a bare stem” (Ralli 2009: 54). The phenomenon is described in similar terms by Kiparsky: “The first member must be a bare stem unspecified for any functional features”, which in the case of verbal dvandva compounds means that it has “[n]o (perfective) Aspect, hence invariant stem in first member” (2009: 190). Likewise, Nicholas and Joseph note that “the first member is morphologically unmarked” in the sense that “there is no derivational morphology […] and no tense-moodaspect marking” (2009: 176). The details of the bare-stem constraint are irrelevant for the purpose of the present chapter and will be discussed elsewhere (Janse forthcoming). What is important to note, however, is that the bare stem in the case of the Modern Greek verbal dvandva compounds is always the imperfective stem, which remains invariable, even if the perfective stem is used in the second member:19

(a)

(b)

πηγαιν-ο-ήρθ-ε

[piʝeno-írθ-e]

“go.ipfv-cm-come.pfv-ind.sg.”

*πηγ-ο-ήρθ-ε

[piɣ-o-írθ-e]

“go.pfv-cm-come.pfv-ind.sg.”

~ πήγ-ε

~ [píʝ-e]

~ “go.pfv-ind.sg.”

πηγαιν-ο-έρθ-ει

[piʝeno-érθ-i]

“go.ipfv-cm-come.pfv-subj.sg.”

*πα-ο-έρθ-ει

[pa-o-érθ-i]

“go.pfv-cm-come.pfv-subj.sg.”

~ πά-ει

~ [pá-i]

~ “go.pfv-subj.sg.”

 19 On apparent counterexamples such as ανεβ-ο-κατεβαίνω anev-o-katevéno] instead of *ανεβαιν-ο-κατεβαίνω [aneven-o-katevéno] “go up-CM-go down” see Ralli/Karasimos (2010: 10 ff.) and Janse (forthcoming).

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

It now becomes clear that the inflection of παιρπαίνω [perpéno] is very different from the inflection of Modern Greek verbal dvandva compounds like πηγαινοέρχομαι [piʝenoérxome] (10a-b): (a)

παιρ-παίν-ω

[per-pén-o]

“take-go.ipfv-ind.sg.”

(b)

παιρ-παίν-ισ̌κ-α

[per-pén-iʃk-a]

“take-go.ipfv-ipf-ind.sg.”

(c)

πηρ-πήγ-α

[pir-píɣ-a]

“take-go.pfv-ind.sg.”

(d)

παρ-πάγ-ω

[par-páɣ-o]

“take-go.pfv-subj.sg.”

More specifically, the first member is not inflected for person, but it is inflected for tense, aspect and perhaps even for mood. Compare, for instance, πηρ-πή-α [pir-píɣa] (11c) with πήρα [píra] “take.pfv-ind.1sg.” and πήγα [píɣa] “go.pfvind.1sg.” and παρ-πάγω [par-páɣo] (11d) with πάρω [páro] “take.pfv.(subj.)subj.1sg.” and πάγω [páɣo] “go.pfv.(subj.)-subj.1sg.”.20 Two other points are noteworthy. First of all, the word stress does not fall on the antepenultimate syllable in cases like πηρ-πήγαν [pir-píɣan] (6), as opposed to *πήρ-πηγαν [pír-piɣan], and παρ-πάγε [par-páʝe] (9), as opposed to *πάρ-παγε [pár-paʝe]. The second point concerns the parallelism between the members of a verbal dvandva compound: “the two basic members are of the same grammatical category, display parallel argument structures, and the meaning of the construction is a conjunction of the meanings of its subparts” (Ralli 2009: 55).21 The two members of πηγαιν-ο-έρχομαι [piʝen-o-érxomai] “goCM-come” are both intransitive – or to be more precise: the first member is here used in the intransitive sense of “go”, as opposed to πηγαιν-ο-φέρνω [piʝen-oférno] “take-CM-bring”, where both members are transitive. As already remarked above, there is thus a crucial difference between Standard Modern Greek πηγαίνω [piʝéno], which is a labile verb, and Cappadocian παίνω [péno], which is (almost) exclusively used intransitively.

 20 Τhe perfective stems °πηρ- [pir-] / °παρ- [par-] and °πηγ- [piɣ-] / °παγ- [paɣ-] may be considered allomorphs of the same perfective stem. 21 Kiparsky also emphasizes the parallel argument structure as a distinctive feature of what he prefers to call “co-compounds” (2009: 187, 190).

  Mark Janse and Johan Vandewalle

 The Turkish perspective: suspended affixation In light of the conclusion of the preceding section, it is remarkable that Dawkins, in his above quoted account of παιρπαίνω [perpéno], takes παίνω [péno] in the transitive sense of “take” which, as we have seen, is attested only once in Dawkins’ folktales. The examples of the Turkish and Turki constructions to which he compares παιρπαίνω [perpéno], however, all contain intransitive motion verbs. Dawkins’ use of the words “still closer” does not refer to expressions with transitive motion verbs, but to intransitive motion verbs like bar- “go” and kit- “go” (Turkish git-), which express the same direction as παίνω [péno], i.c. “away from the deictic centre”, as opposed to to gel- “come”, which expresses the opposite direction, i.c. “towards the deictic centre”. Curiously, Dawkins does not immediately compare παιρ-παίν-ω [per-pén-o] to the also existing Turkish compound verb al-ıp git-, litt. “having-taken go”, which he only gives between brackets as the Turkish equivalent of Turki ap- kit-. In order to explain the meaning of παιρπαίνω [perpéno], it is not necessary to take παίνω [péno] in the transitive sense, as Dawkins does. The original semantics of both παιρπαίνω [perpéno] and al-ıp git- can be reconstructed as follows: the first verb expresses the act of taking possession of the object that will be moved, whereas the second verb, if used intransitively, expresses the subsequent motion of the agent who took possession of the object. In this case, the motion of the agent immediately implies the motion of the object as well. Another possibility is the combination of al-ıp with a transitive motion verb, directly expressing the caused motion which the object undergoes. Now consider the following example from a folktale recorded by Dawkins at Ghurzono: (a)

το

τσ̌υφλ-ό

πήρ-εν

το

κορίσ̌-Ø

to

tʃifl-ó

pír-en

to

koríʃ-Ø

the.n.nom.sg. blind-man.nom.sg. take.pfv-ind.sg. the.n.acc.sg.

girl.acc.sg.

“the blind man took the girl” (b)

πηρ-πήγ-εν

dο

σ-ο

χωρɩ̯-ό

pir-píʝ-en

do

s-o

xorj-ó

take-go.pfv-ind.pl.

sg.acc.

to-the.n.acc.sg.

village-acc.sg.

“he took her to the village”

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

(c)

ως

το

παιρ-παίν-ισ̌κ-ε

os

to

per-pén-iʃc-e

as

sg.acc.

take-go.ipfv-ipf-ind.pl.

“as he was taking her [to the village]”

καρσ̌ουλάτσ̌-ε

το

κλέφτσ̌-ης

karsulátʃ-e

to

kléftʃ-is

meet.pfv-ind.sg.

sg.acc.

thief-nom.sg.

“the thief met him” (Ghurzono, D)

It is clear that the original semantics as described above has been altered. The fact that two clauses containing παιρπαίνω [perpéno] (12b-c) can be preceded by a clause containing παίρω [péro] (12a) with referentially the same direct object, indicates that the notion of “taking possession” has disappeared from the meaning of παιρ-παίν-ω [per-pén-o]: the same object cannot be taken in possession a second time by someone if it is already in her or his possession. Dawkins compares παιρπαίνω [perpéno] with Turkish and Turki constructions, all involving the -(y)Ip or -(i)b converb form of the verb al- “take” as the first constituent, followed by a motion verb as the second. Semantically, the relationship between the clauses resembles one of coordination. Johanson stresses the fact that in this use, the -(y)Ip or -(i)b converb does not encode temporal, causal or other relationships between the clauses: “As a matter of fact, the relation between the propositions is just as open as in European juxtapositions and coordinations such as My horse got lost; [and] I had to walk. It is left to the addressee to reconstruct a possibly intended logical relation (“therefore”, “thus”, “then” etc.); but no such relation is encoded in the verb forms” (1995: 328). As Johanson (1995: 314) points out, constructions in Turkic languages consisting of two verbs linked by the -(y)Ip or -(i)b converb may undergo grammaticalization: the second verb loses its lexical content and turns into an auxiliary merely expressing aspect or Aktionsart. The combination becomes semantically uncompositional as in (13a): the meaning of the converb construction is no longer a function of the meanings the two verbs express when used on their own. In (13) the verb dur-, meaning “stand, stop” when used independently, has turned into an auxiliary expressing duration. Lewis (2000: 190 f.), Kornfilt (1997: 478), and van Schaaik (2020: 709) note that there is a semantically identical, parallel, construction with some of these grammaticalized

  Mark Janse and Johan Vandewalle converb constructions, which is based on an asyndetic succession of two identically inflected verb forms: (a)

söylen-ip

dur-ur

grumble-conv.

aux.-aor.

“he keeps grumbling”

(b)

söylen-ir

dur-ur

grumble-aor.

aux.-aor.

“he keeps grumbling”

As illustrated above, the Turkish converb constructions with al-ıp with which Dawkins compares παιρπαίνω [perpéno] are semantically compositional. At first sight constructions such as al-ıp gel- “having taken come” and al-ıp git- “having taken go” seem less grammaticalized in Turkish and in other Turkic languages than the semantically uncompositional ones. However, Vandewalle (2016) has demonstrated that Uzbek converb constructions expressing motion events constitute a system in which the participating verbs expressing different aspects of the motion event obey strict rules as to their combinability and order. Vandewalle (2018) argues that the Uzbek equivalents of Turkish alıp gel- and alıp git- show signs of reanalysis and grammaticalization. For this reason, no sharp dividing line should be drawn between the semantically uncompositional and the semantically compositional converb constructions. Searches in the online Turkish National Corpus (TNC), developed by Aksan et al. (2012), further reveal that the alternative construction in (13b) also exists for Turkish converb constructions combining al-ıp with a motion verb. Compare, for instance, the following minimal pair: (a)

onu

da

gizlice

al-ıp

götür-dü-m

sg.acc.

prt

secretly

take-conv.

carry away-pfv-sg.

“I took it secretly and carried it away”

(b)

al-dı-m

götür-dü-m

araba-yı

servis-i-ne

take-pfv-sg.

carry away-pfv-sg.

car-acc.

service station-poss.sg.-dat.

“I took the car and took it to its service station”

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

Before arguing that this alternative construction is morphologically more similar to παιρπαίνω [perpéno] than the converb construction, we need to explain some aspects of verbal inflection in Turkish and the phenomenon of “suspended affixation”. The term was probably coined by Lewis in the first edition of his Turkish Grammar (1967) and defined as follows: “when one grammatical ending serves two or more parallel words” (2000: 34).22 As shown in Lewis (2000: 93), Turkish finite verb forms are mainly constructed by adding a suffix expressing tense, mood or aspect to the verb stem. In the case of the simple tenses, this suffix is immediately followed by another suffix marking person and number of the verb form or, in the case of compound tenses, by a “projection suffix” (van Schaaik 2020: 271), expressing, among other things, past, reportativeness/inferentiality and conditionality, to which a personal suffix may be attached.23 In all of these cases two types of personal suffixes are used the choice of which depends on the preceding suffix (a zero suffix being used for the third person singular):24 simple tenses: v.-TMA-pers.

compound tenses: v.-TMA-proj.-pers.

götür-ür-üm götür-dü-m carry away-aor.-1sg. carry away-pfv-1sg. “I use to/will carry away” “I have carried away”

götür-ür-dü-m carry away-aor.-pst-2sg. “I used to/would carry away”

In the alternative to the converb construction, two identically inflected verb forms follow each other without any intervening conjunction. In accordance

 22 The term has gained wide currency in Turkish linguistics (e.g., Kabak 2007; Broadwell 2008; Kharytonava 2012; Kornfilt 2012; Akkuş 2015) and has been applied to similar phenomena in other languages (e.g., Despić 2017 [Serbian]; Yoon 2017 [Korean]; Erschler 2018 [Ossetic]; Kuritsyna 2018 [Tocharian]). It may be noted that Kornfilt does not use the term in her grammar (1997), as opposed to Göksel and Kerslake (2005: 457–461). Van Schaaik prefers the term “postponed suffixation” (2020: 307). 23 The projection suffix is historically derived from forms of the Turkish copula i-. 24 The main exception to this principle is the imperative, for which specific personal endings are directly attached to the verb stem and a zero suffix is used in the second person singular instead of the third person singular (Lewis 2000: 138 f.; Kornfilt 1997: 369 f.; van Schaaik 2020: 201 f.).

  Mark Janse and Johan Vandewalle with the principle of suspended affixation personal suffixes are expressed in the simple tenses only on the second verb if they are Type I (pers./I, related to the personal pronouns), but on both verbs if they are Type II (pers./II, related to the possessive suffixes). In the compound tenses, the projection and personal suffixes are attached to the second verb only: simple tenses: al-TMA + v-TMA-pers./I

al-TMA-pers/ıı + v-TMA-pers./II

compound tenses: al-TMA + v-TMA-proj.-pers.

al-ır götür-ür-üm take-aor. carry away-aor.-1sg. “I use to/will take and carry away” al-dı-m götür-dü-m take-pfv-1sg. carry away-pfv-1sg. “I took and carrried away”

al-ır götür-ür-dü-m take-aor. carry away-aor-pst-1sg. “I used to/would take and carry away”

An important difference between the Turkish and the Cappadocian inflection is that in the Cappadocian TMA is marked by ablaut in the verb stem, as can be seen in (11a-d), whereas in Turkish, a prototypical agglutinative language, a TMA suffix is attached to the phonologically invariable verb stem. As a consequence, an invariable verb stem + TMA suffix in Turkish may be equivalent to a variable verb stem in Greek, to which the personal endings are attached in both cases. Such phonological changes are observed in the stems of both παίρω [péro] and παίνω [péno] in παιρπαίνω [perpéno], as was shown above. It is now clear that παιρπαίνω [perpéno] is not a morphological compound like the Modern Greek verbal dvandva compounds discussed in section 3. It is in fact a “phrasal compound” in the sense of Bağrıaçık and Ralli (2019), modelled on the Turkish construction with suspended affixation al-ır götür-ür-üm. The first member °παιρ- [per-] is marked for TMA, but not for person and number, the affixation of which is suspended and applied only to the second member °παιν- [pen-]. In other words, the first member °παιρ- [per-] behaves exactly like its Turkish equivalent, the verb al-, the only difference being that the Turkish verb stem is invariable and TMA is marked by suffixation, whereas the Cappadocian verb stem is variable and TMA is marked by ablaut.

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

 Cappadocian παιρπαίνω as an example of Turkish pattern replication Remarkably, παιρπαίνω [perpéno] is the only phrasal compound with suspended affixation of its kind in Cappadocian. It expresses the same direction as its second member παίνω [péno]: “go” (away from the deictic centre). Apparently, no other instances developed, e.g., *παιρέρχουμαι [*perérxume] with έρχουμαι [érxume] “come” for the opposite direction in the sense of “bring” (towards the deictic centre), for which φερίσ̌κω [feríʃko] is generally used (Dawkins 1916: 656). In Turkish, more than one motion verb is used as the second member of this pattern and they may be intransitive as well as transitive. As noted above, the construction itself exhibits two variants: one with the converb al-ıp from al“take” as its first part, and one with the verb al- inflected identically as the second verb, with the possibility of suspended affixation in this last case. In order to determine which one of the variants is used most and which motion verbs (transitive or intransitive) occur most in the Turkish construction, we performed searches on the basis of the TNC (Aksan et al. 2012).25 These searches show that the construction consisting of al-ıp and a following motion verb occurs 1383 times in this corpus and the construction with two inflected verbs 539 times.26 This means that of a total of 1922 constructions with al- as the first verb and a motion verb as the second, 72% follow the first construction, 28% the second construction. Clearly, the converb construction is used two and a halve times more than the construction with two identically inflected verbs, with or without suspended affixation. Concerning the motion verbs used in the Turkish construction, the four motion verbs which scored highest are the following:

 25 The TNC can be accessed at: https://v3.tnc.org.tr/. 26 This is an estimate: the online version of the TNC does not allow for two-word-searches using wildcard symbols which would immediately yield the required results. Instead, the search has to be performed in two successive steps during the first of which the total number of hits for only one of the two words in the corpus is determined. The number of viewable contexts corresponding to these hits, which have to be searched for the presence of the other word, however, is restricted to a certain maximum and frequently not all contexts can be visualized. The proportion between the total number of hits and the number of visualised contexts was used to extrapolate the results of the second search to an estimate for the entire corpus.

  Mark Janse and Johan Vandewalle

Construction :

Construction :

al-ıp + götür-

“carry away”



%

git-

“go away”



%

gel-

“come”



%

getir-

“bring”



%

other



%

total



%

“carry away”



%

git-

“go away”



%

gel-

“come”



%

getir-

“bring”



%

al- + götür-

other



%

total



%

It is noteworthy that the four verbs are the same for both constructions and exhibit the same order of frequency. These four verbs can be identified as the deictic verbs, transitive and intransitive, which are used in Turkish. The transitive verb expressing “motion away from the deictic centre”, götür-, is clearly in the lead, with nearly two times as many occurrences in the constructions as the second verb. When we compare the deictic directions which are expressed, we get the following results: Construction :

al-ıp + götür-/gitgel-/getir-

Construction :

al- +

“away from deictic centre”



%

“towards deictic centre”



%

other



%

total



%

götür-/git-

“away from deictic centre”



%

gel-/getir-

“towards deictic centre”



%

other



%

total



%

It appears that “motion away from the deictic centre” is expressed in more than half of the constructions combining al- “take” with a following motion verb in

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

the TNC. In addition, this direction is more than three times as frequent as the opposite (“motion towards the deictic centre”) in construction 1 and more than two times in construction 2. It is therefore not surprising that the “pattern” expressing “motion away from the deictic centre” was the first one, and at the same time also the only one, to be “replicated” in Cappadocian, thus constituting a clear case of PAT in the sense of Matras and Sakel.27 Cappadocian παιρπαίνω [perpéno] shares with alıp götür-, the most frequent Turkish combination, the fact that the verb “take” (παιρ-/al-) is followed by a deictic verb expressing motion away from the deictic centre (παίνω/götür-). However, it differs from alıp götür- in selecting an intransitive deictic verb (παίνω) instead of a transitive one (götür-) and applying construction 2 with two inflected verbs and the possibility of suspended affixation instead of construction 1. Rather than a calque of alıp götür-, παιρπαίνω [perpéno] seems to be a calque from al- gitwhich has an intransitive deictic verb (git-) and applies construction 2, but according to our findings on the basis of the TNC is less frequent than alıp götür-.

 Predecessors and alternatives to παιρπαίνω The verb παίνω [péno], together with its variants πααίνω [paéno] and παγαίνω [paʝéno],28 is derived from (ὑ)παγαίνω [(i)paʝéno], itself an innovative present based on the imperfect (ὑ)πῆγα [(i)píɣa], which came to be used as the aorist as well in Late Medieval Greek (Holton et al. 2019: 1391). Being originally a transitive verb, the intransitive use of the original present ὑπάγω in the sense of “go” (away from the deictic center) is already well attested in Classical Greek,29 and was generalized in Post-classical Greek, where it came to be used as the antonym of ἔρχομαι “come” (towards the deictic center).30 In Late Medieval Greek, (ὑ)πάγω [(i)páɣo] and (ὑ)παγαίνω [(i)paʝéno] again developed a transitive use

 27 See fn. 7. 28 Dawkins recorded παγαίνω at Potamia, but marked it as “possibly non dialectic” (1916: 634). It should be noted, however, that παγαίνω and its variants πααίνω and παίνω are also found at Sinasos (Archelaos 1899: 258) and Anaku (Costakis 1964: 93), as well as in Pharasiot (Dawkins 1916: 634; Andriotis 1948: 30). 29 Cf., e.g., ὕπαγε “go away!” (Aristoph. Nub. 1298, Τhesm. 956, Vesp. 290 bis), ὑπάγοιμί γ’ ἆρ’ ἄν “I’ll be going then” (Aristoph. Av. 1017). 30 In New Testament Greek, for instance, ὑπάγω is only used in the sense of “weggehen, hingehen, gehen” (Bauer et al. 1988: 1667), cf. ἦσαν γὰρ οἱ ἐρχόμενοι καὶ οἱ ὑπάγοντες πολλοί “for there were many people coming and going” (Mc. 6.31).

  Mark Janse and Johan Vandewalle alongside the intransitive one, as in the following examples from a sixteenthcentury prose version of the Alexander Romance: (a)

ὑπῆγ-εν

εἰς

τὴν

ἐκκλησίαν

ipíʝ-en

is

tin

eklisían

go.pfv-ind.sg.

to

the.acc.sg.f.

church

“he went to the church” (rec. E, .)

(b)

ὑπῆγ-εν

τον

εἰς

τὴν

ἐκκλησίαν

ipíʝ-en

ton

is

tin

eklisían

go.pfv-ind.sg.

sg.acc.m.

to

the.sg.acc.f.

church

“he took him to the church” (rec. E, .)

The Standard Modern Greek equivalents πάω [páo] and πηγαίνω [piʝéno] have firmly established both uses: (a)

πήγ-ε

σ’

την

εκκλησία

píʝ-e

is

tin

eklisía

go.pfv-ind.sg.

to

the.acc.sg.f.

church

“he went to the church”

(a)

τον

πήγ-ε

σ’

την

εκκλησία

ton

píʝ-e

is

tin

eklisía

sg.acc.m

go.pfv-ind.sg.

to

the.acc.sg.f.

church

“he took him to the church”

The question is now what to make of the unique example of transitive παίνω in Cappadocian. It is tempting to think that (1) is a remnant of the Late Medieval Greek use illustrated in (15b), but then why was the phrasal compound παιρπαίνω [per-péno] created in the first place? There is a reason to believe that example (1) is perhaps not dialectal or at least not traditional. The distribution of παιρ-παίνω [per-péno] is restricted to the dialects of Aravan, Ghurzono, Fertek, Delmeso, Axo, Floïta and Malakopi (Dawkins 1916: 631). Geographically speaking, these dialects are roughly located in the western part of Cappadocia (Figure 2). The remaing dialects, which are roughly located in the eastern part of

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

Cappadocia with the exception of Silata, use another verb to express the meaning “take” (away from the deictic centre), i.c. πα[γ]άζω [paɣázo],31 which is also used in Pharasiot (Dawkins 1916: 630).32 Compare, for instance, the following examples from Ulağaç: ()

πάασ-Ø

με

βαβά

páas-Ø

me

vavá

take.pfv-imp.sg.

sg.acc.

father.voc.sg.

“take me there, father”

do

βαβά-τ

πάασ-εν

dο

do

ton

páas-en

do

sg.acc.m

father.nom.sg.-sg.poss.

take.pfv-ind.sg.

sg.acc.

“his father took him there” (Ulağaç, D)

It is very likely that πα[γ]άζω [pa[ɣ]ázo] was a shared innovation of Cappadocian and Pharasiot, as the verb is not attested in the other East Asia Minor Greek varieties.33 Apparently, the verb was later replaced by παιρ-παίνω [per-péno] in most of the western Cappadocian dialects. It is remarkable that παιρ-παίνω [perpéno] is used at Axo (Dawkins 1916: 631), whereas πη(γ)άζου [pi(ɣ)ázu] is used in the closely related Central Cappadocian dialect of the neighbouring village Misti (Dawkins 1916: 630).

 Conclusions In this chapter we investigated παιρ-παίνω [per-péno], a unique phrasal compound attested in Cappadocian, but not in any other Modern Greek dialect. We concluded that it is very different from the productive Modern Greek verbal dvandva compounds (§3): it lacks a compound marker and, most importantly, the first member is inflected for TMA as opposed to verbal dvandvas like μπαίν-

 31 The variants πηγάζω [piɣázo] and πεγάζω [peɣázo] are attested at Silata and Misti (Dawkins 1916: 630). 32 Cf. Andriotis (1948: 70) for Pharasiot and Kesisoglou (1951: 85) for Ulağaç. 33 Or, indeed, in any other Greek variety, whether Medieval or Modern.

  Mark Janse and Johan Vandewalle ο-βγαίνω [ben-o-vʝéno-] where the verbal stem of the first member is always imperfective. This appears in the forms of the perfective indicative πηρ-πήγα [pir-píɣa] (11c) and the perfective subjunctive παρ-πάγω [par-páɣo] (11d), where the first member has the ablaut associated with the independent verb forms πήρα [píra] and πάρω [páro]. Dawkins already noted the Turkish connection and research on the basis of the Turkish National Corpus (TNC) revealed that a parallel structure exists in Turkish, consisting of the Turkish verb al- “take” and a following deictic verb. The findings show that the most frequent combination is al-ıp götür- “take and carry away”. Whereas al-ıp götür- is a converb construction with a transitive deictic verb, παιρ-παίνω [per-péno] is a phrasal compound with suspended affixation and an intransitive deictic verb. However, Turkish has a perfect match: al- git- “take and go”, with an intransitive deictic verb and a construction very similar to παιρ-παίνω [per-péno]. In this way, the relation between al- git- and παιρ-παίνω [per-péno] can be described as an example of Turkish pattern replication (PAT).

Figures Asia Minor Koine

EAST ASIA MINOR GREEK

Proto-Cappadocian

Proto-Pontic

Pontic

Cappadocian

Lycaonian

Silliot

Pharasiot

Fig. 4: East Asia Minor Greek dialects. Cf. Janse (2008: 191; 2020b: 203). The dotted lines indicate contact zones.

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

Fig. 5: Geographical distribution of Cappadocian and Pharasiot. Tab. 3: Subgrouping of the Cappadocian dialects. Northwest Cappadocian

Northeast Cappadocian

• Floïta • Silata • Anaku • Malakopi

• Sinasos • Potamia • Delmeso

Central Cappadocian • Axo • Misti South Cappadocian Southwest Cappadocian

Southeast Cappadocian

• Aravan • Ghurzono • Fertek

• Ulağaç • Semendere

  Mark Janse and Johan Vandewalle

Bibliography D Dawkins (1916) MK Mavrochalyvidis & Kesisoglou (1960) PK Phosteris & Kesisoglou (1960) Akkuş, F. (2015), “Suspended affixation with derivational suffixes and lexical integrity”, in: J. Audring/F. Masini/W. Sandler (eds.), Quo Vadis Morphology: On-line Proceedings of the Tenth Mediterranean Morphology Meeting (MMM10) Haifa, Israel, 7–10 September 2015, 1–15. [doi.org/10.26220/mmm.2720]. Aksan, Y. et al., (2012), “Construction of the Turkish National Corpus (TNC)”, in: N. Calzolari et al. (eds.), Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC 2012), İstanbul – www.lrec-conf.org/proceedings/lrec2012/papers.html Alektoridis, A.S. (1883), “Λεξιλόγιον τοῦ ἐν Φερτακαίνοις τῆς Καππαδοκίας γλωσσικοῦ ἰδιώματος”, Δελτίον τῆς Ἱστορικῆς καὶ Ἐθνολογικῆς Ἑταιρείας τῆς Ἑλλάδος 1, 480–508. Andriotis, N.P. (1948), Τὸ γλωσσικὸ ἰδίωμα τῶν Φαράσων, Athens. Archelaos, I.S. (1899), Ἡ Σινασός: ἦτοι θέσις, ἱστορία, ἠθικὴ καὶ διανοητικὴ κατάστασις, ἤθη, ἔθιμα καὶ γλῶσσα τῆς ἐν Καππαδοκίᾳ κωμοπόλεως Σινασοῦ, ἐν ἐπιμέτρῳ δὲ καὶ σύντομος περιγραφὴ τῶν ἐν ταῖς ἐπαρχίαις Καισάρειας καὶ Ἰκονίου Ἑλληνικῶν Κοινοτήτων, ὡς καὶ τῶν ἐν αὐταῖς σῳζομένων ἑλληνικῶν διαλέκτων ἐν σχέσει πρὸς τὴν ἐν Σινασῷ λαλουμένην, Athens. Bağrıaçık, M./Ralli, A. (2015), “Phrasal vs. morphological compounds: insights from modern Greek and Turkish”, Sprachtypologie und Universalienforschung 68, 323–357. Bağrıaçık, M./Göksel, A./Ralli, A. (2017), “Copying compound structures: the case of Pharasiot Greek”, in: C. Trips/J. Kornfilt (eds.), Further Investigations into the Nature of Phrasal Compounding, 185–231, Berlin. Bağrıaçık, M./Göksel, A./Ralli, A. (2019), “Two Turkish suffixes in Pharasiot: constraints against phrasal bases”, in: A. Ralli (ed.), The Morphology of Asia Minor Greek, 116–147, Leiden. Bauer, W./Aland, K./Aland, B. (1988), Griechisch-deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur, 6th ed., Berlin. Boeschoten, H. (1991), “Dialectology”, in: H. Boeschoten/L. Verhoeven (eds.), Turkish Linguistics Today, 150–165, Leiden. Brendemoen, Brent (1998), “Turkish dialects”, in: L. Johansson/E. Csa (eds.), 236–241, London. Broadwell, G.A. (2008), “Turkish suspended affixation is lexical sharing”, in: M. Butt/T.H. King (eds.), Proceedings of the LFG08 Conference, 198–213, Stanford, CA. Costakis, A.P. (1964), Le parler grec d’Anaku, Athens. Dawkins, R.M. (1916), Modern Greek in Asia Minor: A Study of the dialects of Sílli, Cappadocia and Phárasa with Grammar, Texts, Translations and Glossary, Cambridge. Despić, M. (2017), “Suspended morphology in Serbian: clitics vs. affixes”, Glossa 2(1), 12. DOI: http://doi.org/10.5334/gjgl.130 Erschler, D. (2018), “Suspended affixation as morpheme ellipsis: evidence from Ossetic alternative questions”, Glossa 3(1), 12. DOI: http://doi.org/10.5334/gjgl.501 Göksel, A./Kerslake, C. (2005), Turkish: A Comprehensive Grammar, London. Hock, H.H./Joseph, B.D. (2009), Language History, Language Change, and Language Relationship: An Introduction to Historical and Comparative Linguistics, 2nd ed., Berlin.

The Cappadocian Phrasal Compound παιρ-παίνω [per-péno]  

Holton, D./Horrocks, G./Janssen, M./Lendari, T./Manolessou, I./Toufexis, N. (2019), The Cambridge Grammar of Medieval and Early Modern Greek, Cambridge. Janse, M. (2002), “Aspects of bilingualism in the history of the Greek language”, in: J.N. Adams/ M. Janse/S. Swain (eds.), Bilingualism in Ancient Society: Language Contact and the Written Text, 332–390, Oxford. Janse, M. (2007), “Coordinative verbal compounds in Dutch, English, Greek and Cappadocian”, paper presented at the 8th International Conference on Greek Linguistics, Ioannina, 2007. Janse, M. (2008), “Clitic doubling from Ancient to Asia Minor Greek”, in: D. Kallulli/L. Tasmowski (eds.), Clitic Doubling in the Balkan Languages, 165–202, Amsterdam. Janse, M. (2009), “Watkins” law and the development of agglutinative inflections in Asia Minor Greek”, Journal of Greek Linguistics 7, 32–48. Janse, M. (2019), “Agglutinative noun inflection in Cappadocian”, in: A. Ralli (ed.), The Morphology of Asia Minor Greek, 66–115, Leiden. Janse, M. (2020a), “Ἑλληνιστὶ γινώσκεις; (Acts 21:37): The survival of Cappadocian Greek”, Cursor 16, 49–57. Online version: www.amici-online.eu/Cursor_16.pdf Janse, M. (2020b), “Back to the future: akritic light on diachronic variation in Cappadocian (East Asia Minor Greek)”, in: K. Bentein/M. Janse (eds.), Variation in Post-classical and Byzantine Greek, 201–239, Berlin. Janse, M. (2023), “Η Καππαδοκική διάλεκτος”, in: Chr. Tzitzilis (ed.), Νεο-ελληνικές διάλεκτοι, Thessaloniki. Janse, M. (forthcoming), “Medieval and Modern Greek verbal dvandvas and the bare root constraint”. Johanson, L. (1995), “On Turkic converb clauses”, in: Haspelmath/E. König (eds.), Converbs in Cross-Linguistic Perspective, 313–347, Berlin. Kabak, B. (2007), “Turkish suspended affixation”, Linguistics 45, 311–347. Karatsareas, P. (2011), A Study of Cappadocian Greek Nominal Morphology from a Diachronic and Dialectological Perspective, Cambridge: University of Cambridge dissertation. Karphopoulos, P. (2008), “Λεξιλόγιον”, in: E.G. Tzioutzia (ed.), Μαλακοπὴ Καππαδοκίας: Mέσα ἀπό πηγές, ἔγγραφα καὶ μαρτυρίες, 97–146, Thessaloniki. Kharytonava, O. (2012), “Word minimality and suspended affixation”, Studia Uralo-ALtaica 49, 279–290. Kiparsky, P. (2009), “Verbal co-compounds and subcompounds in Greek”, MIT Working Papers in Linguistics 57, 187–195. Korkmaz, Z. (1975–1976), “Anadolu ağızları üzerindeki araştırmaların bugünkü durumu ve karşılaştığı sorunlar”, Türk Dili Araştırmaları Yılığı Belleten 23–24, 143–172. Korkmaz, Z. (1990), “Anatolian dialects”, in: G. Hazai (ed.), Handbuch der türkischen Sprachwissenschaft, 388–413, Budapest. Kornfilt, J. (1997), Turkish, London. Kornfilt, J. (2012), “Revisiting ‘suspended affixation’ and other coordinate mysteries”, in: L. Brugé/A. Cardinaletti/A. Giusti/N. Munaro/C. Poletto (eds.), Functional Heads, 181–196, Oxford. Kowalski, T. (1931), “Osmanisch-türkische dialekte”, in: Enzyklopädie des Islam, Bd. ΙV, 991– 1010, Leiden. Krinopoulos, S. (1889), Τὰ Φερτάκαινα: Ὑπὸ ἐθνολογικὴν καὶ φιλολογικὴν ἔποψιν ἐξεταζόμενα, Athens: Δημ. Φέξης. Kuritsyna, A. (2018), “Suspended affixation with Tocharian adjectival suffix A -şi / B -şşe and its possible parallel in Old Uighur”, Voprosy Jazykovogo Rodstva 16, 265–276.

  Mark Janse and Johan Vandewalle Lewis, G. (2000), Turkish, 2nd ed., Oxford. Manolessou, I./Tsolakidis, S. (2009), “Greek coordinated compounds: synchrony and diachrony”, Patras Working Papers in Linguistics 1, 23–39. Matras, Y./ Sakel, S. (2007), “Investigating the mechanisms of pattern replication in language convergence”, Studies in Language 31, 829–865. Matras, Y. (2009), Language Contact, Cambridge. Mavrochalyvidis, G.P./Kesisoglou, I. (1960), Τὸ γλωσσικὸ ἰδίωμα τῆς Ἀξοῦ, Athens. Mavrochalyvidis, G.P. (1990), Ἡ Ἀξὸ Καππαδοκίας, Αthens: Ἑστία Νέας Σμύρνης. Nicholas, N./Joseph, B.D. (2009), “Verbal dvandvas in Modern Greek”, MIT Working Papers in Linguistics 57, 171–186. Orgun, C.O. (1995), “Flat versus branching morphological structures: the case of suspended affixation”, Proceedings of the Berkeley Linguistic Society 21, 252–261. Orgun, C.O. (1996), “Suspended affixation: a new look at the phonology-morphology interface”, in: U. Kleinhenz (ed.), Interfaces in Phonology, 251–261, Berlin. Pharasopoulos, S.S. (1895), Τὰ Σύλατα: Μελέτη τοῦ νομού Ἰκονίου ὑπό γεωγραφικήν, φιλολογικὴν καὶ ἐθνολογικήν ἔποψιν, Athens. Phosteris, D./Kesisoglou, I.I. (1960), Λεξιλόγιο τοῦ Ἀραβανί, Athens. Ralli, A./Karasimos, A. (2010), “The bare-stem constraint in Greek compound formation”, Gengo Kenkyu 135, 29–48. Ralli, A. (2008), “Compound markers and parametric variation”, Language Typology and Universals 61, 19–38. Ralli, A. (2009), “Modern Greek V V dvandva compounds: a linguistic innovation in the IndoEuropean languages”, Word Structure 2, 48–68. Sakel, J. (2007), “Types of loan: matter and pattern”, in: Y. Matras/J. Sakel (eds.), Grammatical Borrowing in Cross-Linguistic Perspective, 15–29, Berlin. Schaaik, G. van (2020), The Oxford Turkish Grammar, Oxford. Thomason, S.G./Kaufman, T. (1988), Language Contact, Creolization, and Genetic Linguistics, Berkeley. Vandewalle, J. (2016), “On Uzbek converb constructions expressing motion events”, Bilig 78, 117–149. Vandewalle, J. (2018), “On the Uzbek converb construction starting with olib, its reanalysis, and its grammaticalisation”, Central Asiatic Journal 61, 1–47. Winford, D. (2005), “Contact-induced changes: classification and processes”, Diachronica 22, 373–427. Yoon, J.H.S. (2017), “Lexical integrity and suspended affixation in two types of denominal predicates in Korean”, Glossa 2(1), 45. DOI: http://doi.org/10.5334/gjgl.248

Sara Kaczko

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire Abstract: This paper explores the importance of 5th-4th cent. BCE South Italian red-figure vases, which are the outcome of the interaction between the Greek citizens of the colonies and the local populations in southern Italy and Sicily, as a source of information on multiculturalism and multilingualism of ancient Magna Graecia. It focuses on vessels that display Greek inscriptions and specifically on a Lucanian pelike from Heraclea, dating to 410–390 BCE ca., which features the name-label ΟΣΔΑΝ (Ποσδαν), an otherwise unattested form of the name of the god Poseidon. It argues that the variant Ποσδαν is due to the contact between Greeks and local populations, specifically to the interference of the local language spoken by the Lucanians, and that it is the earliest example of the similar phonological phenomenon, probably generated under similar conditions of linguistic interference, found in Παιστανο (ΑΙΤΑΝΟ), the legend on the monetary issues of the South Italian colony of Poseidonia/Paestum from the 3rd cent. BCE.

 Introduction A significant cultural legacy of ancient Magna Graecia — along with philosophy, rhetoric, poetry, medicine, and literature — is the impressive number (20– 25,000) of South Italian red-figure vases, not rarely of equally impressive quality, dating from the middle of the 5th to the late 4th cent. BCE. These artifacts are of great importance from the perspectives of art, archeology, and history as well as multiculturalism and multilingualism, which are the main topics of the present paper. As far as multiculturalism is concerned, these vases are the product of contact between Greeks and local populations in southern Italy and Sicily. They reflect a world that was shaped by both the presence of affluent, dynamic, and cultured Greek colonies (e.g., Cumae, Naples, Tarentum, Heraclea, Metapontion, Thurii, and Syracuse) and contact with local populations (Romans,

 I am grateful to A.C. Cassio, C.A. Ciancaglini, and P. Poccetti for their useful suggestions. The research for this paper was supported by funding from the HERA - JRP “Uses of the Past” (Project “Multilingualism and Minority Languages in Ancient Europe, MuMiL-EU”). https://doi.org/10.1515/9783110779684-008

  Sara Kaczko Umbrians, Campanians, Bruttians, Lucanians, Messapians, Daunians, Peucetians etc.) that often resulted in conflict but also, eventually, in reciprocal influence. Interestingly for the present discussion, a number of these South Italian red-figure vases feature Greek inscriptions that, like the red-figure technique itself, were inspired by practices in mainland Greece, specifically Attica. On the other hand, the inscriptions display some differences from, and less variety than, those found on Attic vases, the majority taking the form of name labels of heroes or gods, though a type featuring nouns inscribed on objects, such as ΤΕΡΜΩΝ and ΝΙΚΑ on pillars, is observable as well.1 On a general level, these inscriptions deserve attention as supplements to the decoration (perhaps as an additional Greek element perceived as valuable by the patrons)2 and, obviously, as regards their relationship to the scene depicted on the vase and their language, dialect, and script. In this paper, I focus on these inscriptions from the point of view of multilingualism. Inscribed South Italian red-figure vases conveyed their messages through the interplay of visual cues and written language, i.e., through a combination of image (shape and iconography, naturally in relation to the context) and text. Most of these vessels were crafted for the Greek citizens of the colonies or the Hellenized indigenous élites, whose L1 was, in many cases, an epichoric Italic language. Multilingualism was at play also with regard to the Greek clientele, since various dialects were spoken in the Greek colonies depending on the founding city (e.g., Laconian in Tarentum, Achaean in Metapontion, and Ionic and, later, Doric in Rhegion). The period from the late 5th through the 4th centu 1 Red-figure vases featuring inscriptions were crafted in all five of the major Italic red-figure fabrics (Lucanian, Apulian, Campanian, Paestan, and Sicilian) in roughly the same proportions (1% of the total), the majority coming from Apulia and Lucania. According to Roscino 2012, there are 236 of these inscriptions, of which 20 are artisans’ marks and thus of no concern here, and 216 relate to the decorations. Most of the 216 inscriptions are painted close to the figures to which they refer; more rarely, they appear on figurative elements, such as stelai or pillars. An interesting case is the abovementioned ΤΕΡΜΩΝ and stelai or pillar types supporting inscriptions that identify a nearby figure (e.g., ΟΡΕΣΤΑΣ on Trendall 1983: 70, no. BB 43) or the recipient of the tomb associated with the stele or the pillar (e.g., ΑΓΑΜΕΜΝΟΝ on Trendall 1967: 115 no. 597). Overall, then, most of the inscriptions are name labels in the nominative (though see the genitives ΤΡΩΙΛΟ on a funerary monument and ΕΡΜΗΝΟΣ on a pillar, cf. Moret 1979: 245); the kalos type is sparsely attested, as are signatures, which are essentially limited to one workshop in Paestum (that of Assteas and Python). Longer sentences are rare, around half a dozen in total, including a couple of dialogues (cf. RVAp Suppl. I 174–175, no. 86 f.; Trendall 1967: 31, no. 305), a katapygon inscription (RVAp: 501–511, no. 131), and an epigram (on which see Kaczko 2021). On the topic, see the useful surveys in RVAp: lii–liii; Roscino 2012; Oakley 2016. 2 Cf. Roscino 2012: 341.

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

ries BCE is one of transition from these dialects to “common Greek,” the koine, though, in some poleis, such as Himera, the transition occurred before the koine prevailed. Further, it was precisely in this period that the old local Greek alphabets underwent a dramatic change, moving toward the standardized form based on the East Ionic script that was to prevail in Hellenistic and later times. My specific focus here is on a Lucanian pelike from the late 5th cent. BCE Heraclea that is a noteworthy example of the sort of information about multilingualism that South Italian red-figure vases offer. As we shall see, this vessel features the inscription ΟΣΔΑΝ (Ποσδαν),3 an otherwise unattested form of the name of the god Poseidon, a variant likely due to the contact between Greeks and local populations. Moreover, Ποσδαν is also possibly the earliest example of the similar phonological phenomenon, probably generated under similar conditions of linguistic interference, found in Παιστανο (ΑΙΤΑΝΟ), the legend on the monetary issues of the South Italian colony of Poseidonia/Paestum from the 3rd cent. BCE.

 The Lucanian pelike reading Ποσδαν The pelike, h. 45 cm, dm. 33.5 cm, again, of Lucanian manufacture, dates to ca. 410–390 BCE and is possibly the work of one of the painters of the PalermoKarneia-Policoro Group (associated with the Pisticci-Amykos workshop), i.e., either a Painter close to the Karneia Painter or the Policoro Painter.4 It was unearthed in 1963 near Policoro/Heraclea in the tomb of an important individual that yielded 19 pots, including a dozen fine large red-figure vases,5 likely from the same workshop. Several of these pots can be attributed to well-known Lucanian painters, such as the Amykos Painter, the Creusa Painter, the Karneia

 3 Incidentally, Ποσδαν is written in the standardized form based on the East Ionic alphabet. 4 Trendall 1967: 55 ad no. 282 (also Todisco 2003: L 6 and Todisco 2012: 7) associated the pelike with the Karneia Painter (along with another pelike found in the same tomb, i.e., Trendall 1967: 55 no. 283), cf. also Zuchtriegel 2018: 82, whereas Degrassi 1965: 17 attributed it to the Policoro Painter. The pelike seems to be a special commission, on which also there is debate among scholars (see below). 5 On the exceptional nature of the tomb (especially a possible parallel with a late 6th cent. BCE heroon in Poseidonia), its decoration, its owner (an individual with close ties to Athens, or a member of the local elite, or an important figure in the foundation period of Heraclea etc.), and the grave goods, see, e.g., Degrassi 1965; Trendall 1967: 50–51; Pianu 1989; Zuchtriegel 2018: 87; Todisco 2012: II.V. 89–90 with previous bibliography (e.g., Gadaleta 2003a; 2003b).

  Sara Kaczko Painter, the Policoro Painter,6 and are decorated with mythical scenes (according to some interpretations, these scenes may be influenced by the specific versions of the myths that appear in Greek plays).7

Fig. 6: Lucanian Pelike Trendall 1967: no. 282, side A (from Pugliese Carratelli et al. 1983: fig. 512).  6 Trendall (1967: 50–60) identified, along with the hands of the already known Karneia Painter, also those of a new artist whom he labeled the Policoro Painter; the resulting PalermoKarneia-Policoro (PKP) Group shows a link to the late work of the Amykos Painter and the early work of the Creusa Painter. On the workshop and its possible location in Heraclea, cf. Degrassi 1965: 19. On the other hand, Todisco 2012: 7 argued for production of the pelike in Metapontion or Tarentum. The Karneia Painter, if he crafted the pelike, is tied to Metapontion and to the Amykos workshop located there (cf. Denoyelle 2002), whereas there is debate over the personality (that of an Athenian according to some scholars), area of activity, and workshop of the Policoro Painter, for which a local dimension is of course possible (see Denoyelle 2002a: 110–112). 7 Degrassi 1965: 11; see, e.g., Trendall 1967: 58 no. 286 (figg. 26–27) = P&P 117–121, no. 34 (= Todisco 2003: L 9), the well-known large hydria portraying Medeia, also identified by the name-label ΜΕΔΕΙΑ, escaping in the chariot of the Sun and Trendall 1967: 52 no. 283 = P&P 127, no. 37 by the Karneia Painter; Trendall 1967: 57 no. 285 and figg. 26.2, 27.2 (= Todisco 2003: L7, cf. also P&P: 73–74), a hydria by the Policoro Painter representing Thanatos and Hypnos carrying off the body of Sarpedon (labeled ΣΑΡΠΕΔΟΝ).

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

Fig. 7: Lucanian Pelike Trendall 1967: no. 282, side A, detail (from Pugliese Carratelli et al. 1983: fig. 514).

Fig. 8: Lucanian Pelike Trendall 1967: no. 282, side B (from Pugliese Carratelli et al. 1983: fig. 514).

  Sara Kaczko The pelike of interest here displays, on side A (cf. Fig. 6), two figures, Poseidon on horseback and brandishing the trident, identified by the name-label Ποσδαν (cf. Fig. 7), and a young warrior, also riding a horse, with a keryreion tattooed on its thigh. On side B (cf. Fig. 8), Athena rides in a chariot (beneath which figures a small olive tree) driven by another deity in a scene unanimously interpreted as alluding to the well-known and often portrayed myth of the struggle between Poseidon and Athena over the possession of Attica (such as, e.g., the one found on the pediment of the Parthenon). Scholars disagree, however, about a variety of other issues relating to the scene, including the identity of the minor figures on each side of the vase (Hermes or Eumolpos and Iris, Nike or Zeuxippe, the daughter of Erechtheus, respectively) and the precise meaning and context of the scene, including patronage. Aside from Weidauer 1969, who favors an interpretation of the scene as inspired by Euripides’ Erechtheus, with the young warrior representing Eumolpos, most interpretations connect the scenes depicted on the vase with the politics of its find-spot, namely Heraclea. Thus, for Degrassi 1965, the young warrior is Hermes, the other female figure Iris, and the scene, which he compares to the aforementioned one on the Parthenon, depicts a moment before strife erupts between the deities, thereby lending to an originally Attic theme a Doric tinge, with Poseidon’s name in the latter dialect referencing the co-foundation of Heraclea by Tarentum in addition to Thurii. According to Pianu 1997, the patronage is tied to Athens, but the presence of both Athena and Poseidon emphasizes the importance of harmony in the recently founded colony. According to Roscino 2012: 338–339, the namelabel located close to an already well-characterized Poseidon emphasizes Poseidon over Athena and more broadly configures the scenes depicted on the vase overall as a revisitation of the traditional strife between the two deities over Athens. In her opinion, this revisitation involved inverting the traditional roles of the two deities — which had Athena prevailing over Poseidon — to match the political situation of Heraclea, which was founded to sanction a truce between the rivals Thurii (ergo Athens) and Tarentum but, in fact, soon became subject to Tarentine hegemony. By contrast, Zuchtriegel (2018: 87–88) argued that the scene could be read as anti-Dorian propaganda, since in the myth Athena prevailed over Poseidon, and that the name-label in the Doric form (Ποσδαν) is instrumental to this end.

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

 The name Poseidon from PIE to Greek Notably, the vessel, aside from being an example of the possible interactions between image and text, also provides a variant of the name Poseidon, Ποσδαν, not attested elsewhere in Greek so far. Since the numerous Greek forms of the name of the deity are normally trisyllabic (see below for the attestations) and are supposed to derive from *Ποτειδά̄ σων, the disyllabic variant Ποσδαν is surprising at first glance. The etymology of the various ancient Greek names of Poseidon is commonly traced back to an -ων enlargement of the compound *Ποτει-δᾱς, its former member (Ποτει-) being an old full-grade vocative of *πότις “lord” < PIE *potis, attested in Greek as πόσις, and the latter a genitive of *dā “earth”.8 Thus, the compound is supposed to mean “lord of the earth”, a vocative also usable as a nominative, precisely as Latin Iuppiter.9 The form *Ποτει-δᾱhων (< *-δᾱσ-ων) underwent various modifications according to the dialect (and type of text). It was left uncontracted in Homeric Ποσειδάων (with Ionic /s/ assibilated < /t/ , see below), Cretan Ποτειδάων, and Corinthian Ποτειδάϝων with a labiovelar /w/ glide,10 but appears as Ποτειδᾱ́ν (with /aː/ contracted from /aːɔː/) in Rhodian and other West Greek dialects, as Ποσειδέων (/-eɔːn/ < /εːɔːn/) in Ionic, and Ποσειδῶν (contracted) in Attic. However, in some securely attested forms, such as voc. Ποτιδᾶ, acc. Ποτιδᾶν (in all likelihood with /iː/, see below), -ων has been replaced by the suffix -ας, possibly one of the well-known hypocoristic -ᾶς formations.11  8 On the etymology and formation of the theonym, see Ruijgh 1967; cf. Kretschmer 1909: 27– 28; Risch 19742: 57, fn. 50; Mayrhofer 1952: 59–61; Schwyzer 1939: 271; DELG s.v. Ποσειδῶν; GEW, s.v. Ποσειδῶν. The majority of scholars agree on the meaning “earth” of the second element of compound, *dā, and therefore on the interpretation of *Ποτειδά̄σων (> Ποσειδῶν) as “lord of the earth” (in addition to the scholars mentioned at the beginning of this footnote, see also Willi 2007; Oettinger 2020), whereas others interpret *dā as “water” and, as a consequence, Ποσειδῶν as “lord of the water” (vd. Littleton 1973; Nikolaev 2019: 794–795; Janda 2000: 292–296). 9 See Kretschmer 1909: 28, stressing that the vocative of the *i stems ended in *ei, for which he quotes Old Indian vocatives such as pátē (e grade) vs. pátiḥ nominative (zero grade) and Lit. naktë̃, OCSl. nošti. 10 On Ποτειδά̄ϝων featuring a glide /w/, probably a secondary extension analogous to personal and divine names such as *Μαχά̄ϝων, Ἰά̄ϝων, Παιά̄ϝων, see Heubeck 1959: 234; Ruijgh 1967: 12–13. 11 For the hypothesis of Ποτιδᾶς as a hypocoristic see DELG: s.v. Ποσειδῶν; Ruijgh 1967: 11 proposed explaining it from *Ποτειδάhας, i.e., as a “doublet morphologique” of *Ποτειδάhων, but with the suffix -ᾱς, cf. Dubois 2017: 305 fn. 23.

  Sara Kaczko The most ancient form of the adjective meaning “of Poseidon” is a derivation from the same compound without final -ων (which must have been a pervasive but secondary enlargement), its first member appearing in the zero grade /i/ (vs. the e grade, /ei/, of the noun) plus the -ιος suffix, i.e., *Ποτιδάhιος, cf. Rhodian Ποτίδαιον (Carpath.).12 The form Ποτῐ-, with the dental stop and the zero grade expected in the derived forms, appeared assibilated in certain dialects, chiefly Attic-Ionic and Arcado-Cyprian, yielding Ποσῐ-; see po-si-da-i-jo (the sanctuary of Poseidon) in Mycenaean (PY Tn 316), Ποσῐδήιον in Homer (with short /i/ metrically guaranteed, cf. Ποσιδήιον ἄλσος in Hom. Il. 2.506), Ποσίδειος in Attic, also the month name Ποσιδεών (< Ποσιδει-ών with the added final -ών), standard in Attic, and Ποσιδάϊος (the name of a month) at Epidauros.13 The sibilant of the adjective was analogically extended to the forms of the god’s name (with e grade, /ei/) in dialects that feature assibilation, giving rise to such Ποσει- forms as Mycenaean po-se-da-o-ne,14 Hom. Ποσειδάων, AtticIonic Ποσειδῶν, and Arcadian Ποσοιδᾶν with -οι- resulting from vowel assimilation.15 In the West Greek dialects, as expected, the forms are usually found with a retained inherited dental stop, e.g., Ποτειδάν in Corinth16 and Ποτ- in Cretan, Rhodian, Coan, Delphian, Boeotian, and Thessalian.17 However, the assibilated forms are also found in some West Greek dialects, e.g., Laconian Ποhοιδάν,18 Argolic Ποσειδάν, and Achaean Ποσειδανία, the original form of the name of the southern Italian colony known as Poseidonia (Ποσειδωνία)/Paestum,19 an -ία derivation of the god’s name.20 All of these assibilated forms in Doric dialects

 12 Buck 1955: 45. 13 Cf. also the personal name Ποσίδειος (vd., e.g., Athens, IG I3 744), Ποσίδηος (and later Ποσίδεος); Buck 1955: 45. 14 Note that the analogical process already occurred in Mycenaean, in which, as just seen, both the derivative po-si-da-i-jo and the name of the god po-se-da-o-ne are found. 15 It has also been proposed that the difference in the vocalism of the diphthong among the forms Ποτει-, Ποσει-, and Ποσοι- should be explained with reference to differences in apophonic grade; cf. Ruijgh 1967: 7. 16 Most forms in Corinth are written with Ε = /eː/ (cf., e.g., the dative ΠοτΕδᾶνι for Ποτειδᾶνι); οn this graphic solution, see § 6. 17 See again Buck 1955: 38 and 45 (also for the variant Ποτιδάν, on which see below). 18 Cf. the Laconian derivative Ποhοίδαια, “festivals of Poseidon” (CEG 378, Sparta, 450–431 BCE) and Ποσοίδαια, a tribe at Mantinea, IG V,2 271. 19 Ποσειδωνία from the end of the 5th cent. BCE; Paestum is the Latin name, via an Italic toponym; cf. the ethnonym attested in the form ΑΙΤΑΝΟ, ΑΙ on coins from the 3rd cent. BCE, on which see § 8 (also on the unclear and debated /ai/). 20 See Dubois 2002: 53, who also mentions Ποσείδιον, a toponym close to Pellene, in Achaia (Paus. 7.27.8), as well as the historical, cultural, and religious background of the cult of Posei-

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

are usually attributed to a Pre-Doric substrate.21 As a consequence, it may be significant that Ποσδαν, however apparently “irregular”, points to a borrowing from a West Greek dialect that featured assibilation of /t/, something expected in the region around Heraclea and in the nearby areas.22

 The disyllabic Ποσδαν As mentioned, the disyllabic Ποσδαν, rather than the normally trisyllabic name of the god in the various Greek dialects (Ποτειδάν, Ποσειδῶν etc.), comes as a surprise. The main question is whether the form is attributable to an inadvertent painter who wanted to write, perhaps, *Ποσειδαν or *Ποσιδαν23 and omitted the

 don in Achaia, Helike, and its colony Sybaris (cf. Hom. Il. 8.204; 20.404 and Diod. 15.48–49). Therefore, the assibilated variant Ποσει- was the one used by the West Greeks in Achaia before they established colonies in southern Italy in the 8th cent. BCE. In this respect, the consonantism of Ποσειδανία is to be opposed to Ποτειδαία (Potidaea), the name of the colony of Corinth in the Calchidian Peninsula, which is also a formation in -ία of the name of the god but with Ποτει- retained, as expected. 21 See Buck 1955: 58 (cf. ML 42 Ποσειδάν, cf. Ποσίδαον, a deme in Argo); Lejeune 1972: 64 fn. 8. This is possibly the case of forms in Ποσ- in later inscriptions in Cretan and Rhodian, though Buck 1955: 58 argued that some of them may have been influenced by Ποσειδῶν, the standard (i.e., koine) form, as, for example, is Ποτειδοῦνι in Thessalian with inherited native /t/, but also -ουνι, the local rendering of Attic -ωνι. Laconian Ποhοιδάν features intervocalic aspiration that developed from intervocalic /s/ attested from the 5th cent. BCE, which, as is well known, is not found in Tarentum (see, e.g., νίκαhας CEG 378, Sparta, 450–431 BCE vs. ξαίνōσα LSAG: 280, 283, pl. 53.1, Tarentum, 540–530 BCE), probably because the process -s- > -h- postdates the colonization, cf. Thumb 1901: § 102; Schwyzer 1939: 93; Uguzzoni 1968: 40; Cassio 2002: 442–446. 22 The vessel has been connected not only to Heraclea, but more broadly to an area comprising Heraclea, Tarentum, and Metapontion (see above fn. 6), where are attested West Greek dialects featuring Ποσ- in the first member of the name of the god, starting with Heraclea, a polis in which the West Greek component soon prevailed over the Attic(-Ionic) one, the dialect of which was ultimately derived from a Laconian one; this is of course the case also of Metapontion, an early Achaean colony. On the contrary, an alternative explanation of /s/ in Ποσδαν would be the influence of the Attic-Ionic in use in Thurii (also based on the fact that Heraclea was a co-foundation of Thurii and Tarentum), though this does not seem really necessary. It should be added that, since the vessel dates to the end of the 5th cent. BCE, it is unlikely that /s/ in Ποσδαν is due to the influence of the koine. 23 Theoretically *Ποσοιδάν would also be possible; yet this form, attested only in Arcadian inscriptions, is never found in Magna Graecia. Moreover, at both an older and a more recent chronological stage direct importation from Laconia (Ποhοιδάν) is doubtful and would obviously imply the restoration of intervocalic /s/ (see above at fn. 21). To assume a Ποσειδάν or

  Sara Kaczko second diphthong or vowel by (a mechanical) mistake or to phonology, specifically, involving the complex relationships among the Greek dialects and the Italic languages in Lucania. Support for the latter option includes, as mentioned and discussed in detail presently, the fact that the name of Ποσειδανία, an early colony of Sybaris, appears as Paestum in Latin.24 That is, a form in which /s/ immediately precedes a dental, as is the case with Ποσδαν, which in turn implies the elimination of , probably /eː/ (or perhaps even /iː/), from the Greek word. Looking more closely at the data, the context in which the pelike was crafted and inscribed with the god’s name likely played an important role. The scenario is the typical milieu of 4th cent. BCE Magna Graecia: the pelike, found at Heraclea in Lucania, is of Lucanian manufacture, the product of an artist who was active in an area of linguistic contact between Greeks, speaking mainly West Greek dialects (though there was an important Attic-Ionic component in Thurii) and Italic peoples.

 The vocalism of the first member of the theonym Most of the forms of the name Poseidon feature an /ei/ diphthong (e.g., Ποτει-, Ποτειδάν in Corinthian, Thessalian etc.; Att. Ποσειδῶν; Arg. Ποσειδάν); as we have seen, /oi/ is well attested but less common (e.g., Arc. Ποσoιδᾶν, Lac. Ποhοιδάν), and, lastly, there are securely attested forms, such as Ποτιδᾶ, Ποτιδᾶν, and also Ποτῑδά̄ [ν], Ποτῑδά̄ ϝο̄ νι featuring (certainly /iː/ for metrical reasons, see below). Aside from the variants featuring /oi/, both the type in /ei/, quite obviously, and the type in /i/ derive from an original form featuring an inherited /ei/ diphthong. However, this diphthong, which was, of course, originally and for quite some time pronounced as such, in all likelihood soon underwent monophthongization so that most forms written with feature conservative writing, in that was, in fact, pronounced as a single closed vowel /eː/ or even /iː/. This process must have affected also /οι/, though in later times.25 Returning to Ποσδαν, one must then ponder whether the form attests  *Ποσιδάν as the preform of Ποσδαν seems much safer. In any case, both /ei/ and /oi/ were monophthongized at an early date in the spoken language, see below at fn. 25. 24 Cf. fn. 19 above and, in more detail, § 8 below. 25 Monophthongization is common in several Greek dialects. For example, the original diphthong /ei/ began to be pronounced /eː/ in Attica around the 7th–6th cent. BCE (and subsequently the digraph EI (), used for the original diphthong, came into use for /eː/ as the outcome of

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

that the elimination of , (for /iː/), or had already occurred in at least one of the Greek dialects spoken in Magna Graecia. In Greek, syncope occurs in words of at least three syllables and involves an unaccented short vowel between two consonants, one of which is usually a liquid, a nasal, or the palatal glide /j/.26 The latter of these two conditions seems not to be the case for the preform of Ποσδαν; more importantly, syncope is sparsely attested and by no means systematic. Indeed, most scholars27 regard syncope in ancient Greek as a rare phenomenon tied to prosodic conditions that are both specific and unusual in the language, involving forms with a particular “Sprechtempo” or an initial stress accent. The clearest examples of the former are “mots accessoires” subject to phonetic reduction such as οἶμαι < οἴομαι in Attic with the meaning “in my view”, which developed a function similar to that of a particle and “Allegroformen” such as τίπτε < τίποτε, imperatives such as ἐλθέ < *ἐλυθέ (which probably triggered syncope in other forms of the verb, such as aor. ind. ἦλθον < ἤλυθον),28 and, perhaps, vocatives in personal names in -ώνδας (e.g., that of the poet Ἡρώνδας) < -ωνίδας (more precisely from *-ώνιδᾰ).29 Regarding the type of accent, Thessalian is the only Greek dialect in which syncope is well attested (cf. the name of the god Ἄπλουν and personal names in Ἀστο-, e.g., Ἀστόδαμος < Ἀριστόδαμος),30 a phenomenon that has been attributed to the early development of a stress accent on the initial syllable in that dialect.31  contractions and compensatory lengthening), in Corinth already in the earliest written evidence (see in detail § 6), and early in Megara, as well; cf. Buck 1955: 31; Lejeune 1972: 229–230. In some dialects, the diphthong /ei/ was pronounced as a very closed long vowel /eː/ and subsequently evolved toward an even more closed pronunciation, i.e., /iː/; examples include forms such as τελῑτο̄ (vs. τελείτω) in Argos and Πί̃θαρχος (vs. Πείθαρχος) in Boeotia in the 5th cent. BCE (see Buck 1955: 30; Lejeune 1972: 229–230). As regards /oi/, there is early evidence from Boeotia of its monophtongization into a long back vowel, though Teodorsson 1974: 204 argued that the process began sporadically also in Attica in the 6th cent. BCE and then was completed by the first half of the 3rd cent. BCE (cf. also Gignac 1975: 201; Mayser 19232: 108–112, quoting cases such as δέσπονα for δέσποινα). 26 On such conditions and syncope in Ancient Greek, see Szemerényi 1964 (esp. pp. 258 ff.). 27 Cf., e.g., Rix 19922: 18 and 58; Lejeune 1972: 223. 28 Both forms are already found in Homer, see, e.g., Il. 13.250 and 252. 29 Szemerényi 1964: 36–37 (post Solmsen 1909) advanced the hypothesis that the original suffix of personal names in -ώνδας was -ίδας and that -ώνδας originated in an *-ώνιδᾰ vocative, but it has not been universally accepted. Thus, according to other scholars, starting from Neumann 1908 and more recently Rau 2007, -ώνδας was formed by adding the ending -δας of the patronymic to names in -ων; cf. del Barrio Vega 2017, who argued that personal names in -ώνδας originated in Thessalian and then spread to nearby regions. 30 However, some instances of possible syncope have also been explained differently, particularly, Ἀστο- < Ἀριστο- in personal names with the absorption of /i/ after /r/ in Aeolic (Lejeune

  Sara Kaczko A phonological solution in Greek terms — in other words, accepting that the name of Poseidon suffered syncope already in a Greek dialect in southern Italy — would, then, involve Ποσῐ- with short /i/ as the expected first member of the preform of Ποσδαν, i.e., with the ποσῐ- stem of the adjective used for the god’s name. This would be the reverse process of what is observed elsewhere, i.e., with the stem of the noun in the first member of the adjective (e.g., Ποσει- in the name of the colony Ποσειδανία).32 However, this supposition comes up against the serious difficulty that, in practice, all of the attested forms of Poseidon’s name show either /iː/ or a diphthong in the first member.

 Variants in Ποτι- in Greek There are, in fact, a few instances of Ποτι- in the god’s name, but neither the epigraphic nor the literary occurrences attest to a short /i/ in the first member; quite the contrary, the metrically guaranteed forms show /iː/. Many epigraphical instances of the name of the god in West Greek dialects come from Corinth, which is only natural given the importance of Poseidon there. Numerous inscriptions in which the god is named were executed by artisans, e.g., on the vase Wachter 2001: COR 97, and, especially, on the pinakes from Penteskouphia. In Corinthian inscriptions, the form is, as a rule, Ποτειδάν (cf. § 3), with the vocalic element of the second syllable mostly rendered with E, which in the Corinthian script stands for /eː/ already from the first written records — both for /eː/ as the outcome of vocalic contraction and compensatory lengthening and

 1972: 223) and Ἄπλουν as simply showing the zero grade rather than the e grade, cf. Lac. Ἀπέλλων (cf. Buck 1955: 46). 31 Chadwick 1992. Greek eventually developed the stress accent (perhaps already in the 2nd cent. BCE) found in modern Greek, in which syncope and apheresis are frequent (e.g., φέρτε < φέρτε, μέρα < ἡμέρα, δέν < οὐδέν). Some dialects may have anticipated this process; aside from Thessalian, Mendéz Dosuna 1994 argued that the development of stress accents in Heraclea Pontica from late classical times may explain some differences in the phonological treatment of the group /-rw-/. On the other hand, the nature of the pitch accent in ancient Greek (cf., e.g., Allen 1973; Devine Stephens 1994, esp. 215), the opposition between “pitch” accent and “stress” accent, and the relationship between accentual type and syncope are all subjects of debate; see Szemerényi 1964: ix; Clackson 2007: 77–78; cf. Probert 2006: 55–58. 32 The names of the festivals of Poseidon in Laconia, Ποhοιδάια (CEG 378, Sparta 5th cent. BCE, vs. Ποσιδεῖα, the festivals of Poseidon in Delos), and of the Corinthian and Achaean colonies, Ποτείδαια and the abovementioned Ποσειδανία respectively (vs. Ποτιδανία, a city in Aetolia), were built on the name of the god.

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

for /eː/ as the outcome of the early monophthongization of /ei/, for instance in COP 4, 73E, and COR 97.33 Interestingly for the present discussion, there are also two instances of Ποτι-, both written with  (= /i/), the peculiar form of iota typical of Corinthian script, i.e., Ποτῑδά̄ ϝο̄ νι in COP 2A = DGE 123.4 and Ποτῑδά̄ [ν] in COP 73L. The former is a dactylic dedication in which a long second syllable in Ποτῑδά̄ ϝο̄ νι is required by the meter. These Ποτι- forms almost certainly presuppose Ποτει-, in the sense that /iː/ is the outcome of the monophthongization of /ei/, for which there are parallels such as Φῑδίας for Φειδίας and, vice-versa, Ἀ(μ)φιτρΕταν (with Ε) for Ἀμφιτρί̄ταν in Corinthian inscriptions.34 At any rate, at least COP 2A attests to a long ī (/iː/) in the name of Poseidon, and possibly COP 73L as well. A number of literary instances relevant to the present discussion are found in the treatise On Anomalous Words (GG III/2.916.16–917.5) by Herodianus, who mentions forms of the name Poseidon that are rare from his perspective, which is to say, in comparison with the more common Attic-Ionic Ποσειδῶν. These instances include some variants attested either in Doric authors or contexts tied to the Doric world (such as Eupolis’s Heilotes). Herodianus lists the Doric forms as Ποτειδάν and Ποτιδᾶς. As an example of the former, he quotes Epich. 186 K.A. τοῦ Ποτειδᾶνος δὲ χρηστὸν υἱὸν ἱπποκάμπιον. On the other hand, the ancient scholar cites as a parallel for Ποτιδᾶς Κερκιδᾶς,35 in that the form differs from Ποτειδάν both in featuring an iota and in the accentuation of the final syllable, since the form Ποτιδᾶς has a circumflex accent, whereas Ποτειδάν has an acute accent. For this reason, Herodianus adduces examples from comic authors, two of them, notably, from Magna Graecia: first, the accusative from Epicharmus’ Cyclops (fr. 70 K.A.) ναὶ τὸν Ποτιδᾶν, κοιλότερος ὁλμοῦ πολύ and then a vocative in a passage from Sophron fr. 125 K.A. Ποτιδᾶ δραστοχαῖτα and a genitive from Eupolis’s Heilotes fr. 149 K.A. τέμενος Ποτιδᾶ ποντίω.36 All of the

 33 Cf. also forms such as ΔϝEνία (LSAG: 131, no. 6 = DGE 124, middle of the 7th cent. BCE) and φεύγΕ (COR 131c, late 6th cent. BCE). Turning back to Ποτειδάν, also attested on the Corinthian pinakes are a few other graphic solutions regarding the vocalic element of the second syllable, most of which point toward an original Ποτειδάν. These include sporadic examples of the digraph in the form  (with  usually representing /e/, /εː/ in the Corinthian alphabet) to render the vocalic element of the second syllable and, more surprisingly, of ; cf. also the digraph  in COR 38B(a); see Wachter 2001: 268, also with previous bibliography. 34 See Lejeune 1945: 110; Wachter 2001: 243. 35 Cf. GG III/2.917.1: εἴρηται δὲ καὶ Ποτιδᾶς ὡς Κερκιδᾶς.

36 Note that the manuscripts agree in offering Ποτι- in Herodianus’s quotation of Epicharmus fr. 70 K.A.; Sophron fr. 125 K.A.; Eupolis fr. 149 K.A., though in the parallel with Κερκιδᾶς H reads Ποτιδᾶς ὡς Κερκιδᾶς (more precisely, Ποτίδας ὡς Κερκίδας, sic), whereas V reads Ποτει-

  Sara Kaczko relevant instances feature the second syllable in a position of the line in which a long syllable is required. Therefore, at least some of these forms likely attest to a variant of the name of the god in Ποτι- with ī (/iː/), along the more familiar Doric form Ποτειδάν, whereas a short /i/ seems unsupported.37 To conclude, the available evidence indicates that the first member of the preform of Ποσδαν should have been Ποσει- or *Ποσῑ-.38 This conclusion, along with the fact that, in Greek, syncope occurs only with short vowels under specific conditions (and not those of Ποσδαν) and is generally rare, militate against the hypothesis that Ποσδαν can be explained through a phonological process occurring in the Greek language.

 Interference from Italic languages The situation is quite different in the Italic languages, in which syncope is universally explained as one consequence of an initial stress accent. The phenomenon is extensive in Latin (e.g., caldus vs. calidus, rettuli < *retetuli), Etruscan (Atlnta < Ἀταλάντα, Αχle < Ἀχιλλεύς, Hercle < Ἡρακλῆς),39 and, importantly for

 δᾶς ὡς Κερκιδᾶς (more precisely, Ποτείδας ὡς Κερκίδας, sic). As is well known, the writing 〈ει〉 for /iː/ and also of 〈ι〉 instead of 〈ει〉 are common from the Roman period onward (Threatte 1980: 198–199; Gignac 1976: 189–190), therefore, this reading of MS V is not surprising; however, it may have caused variation in the interpretations of the spelling by modern scholars (for instance Kassel/Austin print Ποτει- also at Epich. 70; Sophr. 125; Eupol. 149, whereas Kaibel in his CGF I chose Ποτι- for Epich. fr. 81 and Sophr. fr. 131); see Kaczko (forthcoming). 37 Ahrens 1843: 243–246; Id. 1894 argued that Ποτῑδάν and Ποτῑδᾶς were the genuine Doric forms on the basis of ancient grammatical sources, such as the abovementioned Hdn. GG III.2.916.16–917.5 (with the quotations of comic authors), Et. Magn. 136.40, and Et. Gud. 530.37. On the other hand, LSJ quotes both Ποτῐδᾶς and Ποτειδᾶς as possible because of variants in manuscripts, although the form with short /i/ is unsupported. 38 This does not mean that we should imagine Ποσδαν on the Lucanian pelike discussed here as derived from a hypocoristic *Ποσιδᾶς that corresponds to the hypocoristic Ποτιδᾶς, because in this scenario Ποσδαν (< Ποσῑδαν) would be the accusative from *Ποσιδᾶς and on South Italian red-figure vases the name-labels of gods are almost exclusively in the nominative (see § 1 and fn. 1), unless one imagines that Ποσδαν (< Ποσῑδαν) was an accusative used as a nominative (cf., e.g., Ζάν at Aristoph. Av. 470; cf. Chantraine 19612: 99). Ποτιδᾶς simply attests to the fact that there were forms of the name Poseidon featuring /iː/ in the first member, possible owing to the monophthongization of /ei/, as seen in the instances Ποτῑδά̄ϝο̄νι and Ποτῑδά̄[ν] in COP 2A and COP 73L, from Corinth, i.e., from a nominative Ποτῑδάν. 39 See, e.g., Wallace 2008: 37–40 (cf. also variants of the name of Achilles showing vowel weakening such as Aχale, Aχile, Aχele, Aχule).

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

the present discussion, in the Sabellic languages. There, medial syncope is a common Sabellic phenomenon occurring from ca. 500 BCE, if not earlier,40 with dozens of attestations, e.g., Osc. actud, Umbr. aitu < *agĕtōd; SPi. aitúpas, cf. Umbr. eitipes < *ag-ĕto-; Osc. minstreis, < *-is-tero; Osc. destrst, Umbr. destre < *dexiteros (cf. Greek δεξίτερος, but Lat. dexter); Osc. cebnust, Umbr. benust < *ke-gu̯ ĕnust, cf. Lat. venerit; Umbr. struhc̹ la < *struu̯ i-kelā; Osc. prúffed < *profefed; Osc. teremenniú, teremníss, cf. Lat. terminibus; Osc. úpsannam, cf. Lat. operandam; Osc. núvlanúis < *nuvela.41 All of these forms attest to the syncope of short internal vowels in reconstructed preforms of the Italic language, but the case of Ποσδαν vs. Ποσειδάν42 or *Ποσῑδάν differs markedly in that it involves a non-Italic language. In essence, it seems possible that the quantity of the medial unaccented vocalic element was neutralized in the Greek word Ποσειδάν or *Ποσῑδάν as pronounced by local speakers of Oscan, the distinctive vowel length having been lost as a phonemic trait. This process may have paved the way for a further step, namely, the elimination of the vocalic element in question (see below).43 Regarding vowel length and its neutralization, notably, in Oscan, the wellknown graphic doubling of the vocalic sign indicating vowel length occurs only in the first syllable of a word, e.g., aasas, cf. Lat. āra, Fluusaí, cf. Lat. Flōra, paam, cf. Lat. quam.44 Moreover, as already mentioned, in Oscan and other Italic languages, a number of modifications involving vocalic sounds were cer 40 Meiser 1986: 131–132; Rix 1998: 254–255; Nishimura 2012: 387–388. 41 See Benediktsson 1960: 178 ff.; cf. Bottiglioni 1954: 49–53; Buck 1904: 57–60; Nishimura 2012. 42 On the actual pronunciation of ει (i.e., likely as a closed long vowel), see above, § 5. 43 Regarding a pre-form Ποσι-, it must be noted that the “high vowels” /i/ and /u/, possibly because they have a shorter duration, are more prone to modifications, such as deletion or desonorization, in a variety of languages. Cf. in Greek the aforementioned ἐλθέ, ἦλθον < *ἐλυθέ, ἤλυθον and -ωνδας < -ωνίδας, probably from the vocative *-ωνιδᾰ (perhaps also χρυσός < * χυρυσός, from Phoenician ḥurōṣ or ḥurūṣ, according to Szemérenyi 1964: 53–54, on the basis of Mycenaean ku-ru-so and the forms in Akk. ẖurāṣu, Hebrew ḥāruṣ < ḥurāṣu, Phoenician ḥrṣ, but see DELG: s.v.); see more generally Szemérenyi 1964: 3–100. Cf. in Latin, e.g., tegumen, tegimen, tegmen, also Lat. ocŭlus, vecŭlus, virĭdus > Pre-Romance Latin (cf. the Appendix Probi) oclu-, veclu-, virdu- (hence Ital. occhio, vecchio, verde, Rom. ochiu, vechiu, verde), see Belardi 1979: 78–79. The phenomenon is widespread also in modern languages (including those without opposition of quantity) ranging from the Romance tongues, e.g., French coude (< Anc. Fr. code) < Lat. cubĭtu-, Prov. and Ibero-Romance senda, French sente < Lat. semĭta etc., to Japanese (cf. Yuen 2000) and Bantu dialects (cf. Myers 2005). 44 See, e.g., Bottiglioni 1954: 19–20; Buck 1904: 22 (the apparent exception of trístaamentud has been explained either with reference to the influence of paam written in the line immediately above trístaamentud in the same inscription or to the influence of Lat. testāmentum, see Lejeune 1975: 243 fn. 40; Poccetti 1982–1983).

  Sara Kaczko tainly due to the presence of a stress accent on the first syllable — not only syncope in medial and final syllables, but vowel reduction as well, a phenomenon also apparently indicating that the vowels in medial unaccented syllables were reduced (i.e., were short or weak).45 Vowel reduction is often regarded as a fairly limited phenomenon in the Sabellic languages compared with Latin.46 However, the question seems to have been considered answered too quickly, for there have been convincing counter-arguments to the traditional view showing that there are, indeed, numerous instances of vowel reduction dating to a very early stage of Sabellic, and in Oscan the process was operating synchronically as well, e.g., Osc. prupukid < *prō-pakii̯o,47 amprufid < Proto-Sabellic *amprofēd (cf. Lat. improbē), praefucus vs. facus, possibly also aflukad vs. aflakus, and cf. Umbr. atripusatu, atropusatu < *-pod- (cf. Lat. pēs, pedis), comoltu vs. kumaltu, Paleo-Umbr. setums < *septomos etc.48 The fixation of the accent on the first syllable is usually dated at least to the Proto-Sabellic period (and an even earlier dating, ante 550–500 BCE, has recently been proposed) and the loss of short vowels in word internal syllables to ca. 500 BCE.49 Therefore, both phenomena antedate the Lucanian pelike discussed here. As a rule, loan-words take on the accentuation rules of the languages that borrow them,50 and this must also have been the case with Greek loan-words in  45 Cf. Lejeune 1975: 245 (esp. 243–245). 46 See, e.g., von Planta 1892: 235; Buck 1904: 55. 47 Weiss 1993: 36–39. 48 Bottiglioni 1954: 45–46; Buck 1904: 55–57; Nishimura 2012: 381 ff.; Zair 2016: 301–302. Nishimura, emphasizing the fact that vowel reduction was particularly noticeable in the preand post-labial positions, argued that the process was already operative at a very early stage of Sabellic, 550–500 BCE or earlier (see also above in the text), which would mean, of course, that the initial stress accent is to be dated even earlier. Zair, on the other hand, hypothesized that vowel weakening resulted from independent developments in Sabellic languages and was a late phenomenon in Oscan, i.e., to be dated not much before 300 BCE and post-dating loss of vowel length, which he also proposed dating relatively late, namely in the second half of the 4th cent. BCE. 49 See, e.g., Meiser 2017: 748. 50 Modern examples include the accentuation of foreign words according to the accentuation rules of French regardless of the original accent of the word; cf. Canepari 1979: 252. Ancient examples include Latin Aristoteles, Thucydides, and Plutarchus vs. the original Greek Ἀριστοτέλης, Θουκυδίδης and Πλούταρχος; on the Latin prosody of Greek names, cf. also observations regarding the usual pronunciation of Romans vs. the Greek pronunciation as an educated affectation by Quint. Inst. Or. 12.10.57; cf. Serg. GL IV.526.14–19 and, e.g., Adams 2003: 108– 110. In Greek, the original accent of the Latin loan-word was maintained only when it was compatible with the Greek rules of accentuation, otherwise it was changed to conform with them, cf. e.g., centuria, κεντουρία, Traianus, Τραιανός, along with the adjectives in -icus, -anus,

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

Oscan, which were pronounced with an initial stress accent, often triggering, in turn, the phenomena of vowel reduction and syncope mentioned above. Examples include the vowel reduction in Arentika[i, cf. Hesych. Ἀράντισιν· Ἐρινύσι, Μακέδονες and the syncope in Osc. culchna, culcfnam, a borrowing of West Greek κυλίχνα (Att.-Ion. κυλίχνη; cf. Lat. culigna, Etr. culcna)51 and in Herekleís for Ἡρακλῆς.52 Importantly, Herekleís features syncope of /a/ (cf. also Vest. Herclo), which created the conditions for the shortening of the vowel before /r/ plus-consonant, which, in turn, caused the vocalic anaptyxis (hence Here-). Similarly, it is likely that the proto-form of Ποσδαν, after being borrowed in Oscan, was pronounced with an initial stress accent that subsequently triggered the syncope of the unaccented vocalic element in the medial syllable, the quantity of which had previously been neutralized.53 If so, this peculiar variant of the name of the god Poseidon is attributable to the influence of the language spoken by Lucanian populations.

 The parallel of ΑΙΤΑΝΟ The idea that the lack of an /e/ or /i/ between the sibilant and the dental in Ποσδαν is due to a special linguistic development of an Italic language, and neither to a phonological modification occurred in a Greek dialect nor to a mechanical error in the inscription, receives powerful support from the fact, alluded to earlier, that the same phenomenon is attested in ΑΙΤΑΝΟ, the legend found on the earliest monetary issues of old Ποσειδανία. This southern Italian colony of Sybaris became a Roman colony in 273 BCE and was called Paestum in Latin. As mentioned (cf. § 3), the original name of the colony, Ποσειδᾱνία, was engraved  -enus turning into -ικός, -ανός, -ινός. So also in case of Egyptian names in Greek, the accent depends on the declension to which they belong (e.g., -ις, -ης, -ευς) and the Greek rules of accentuation (cf. Choerob. GG IV/2.196; GG IV/2.205, and the Persian name Ἀρταβάζης and Ἀρτάβαζος etc.; see Clarysse 1997). 51 Cf. Untermann 2002: 407–408. 52 Buck 1904: 20–21; Poccetti 2012: 82–88 (see also p. 85 on the independent borrowing of the name of Heracles in Etruscan, Latin, and Oscan). On Greek loan-words in Oscan see also Lazzeroni 1972; Prosdocimi 1976; 1978. 53 In many languages, at any rate, pre-tonic vowels behave exactly as post-tonic vowels with regard to deletion or reduction: see, e.g., Lat. *bonĭtāte > Ital. bontà, Lat. *bellĭtāte > Ital. beltà; Russian /sadovód/ as /sәdavót/, /gorodók/ as /gәradók/ (see Belardi 1979: 79; Maiden 1995; Crosswhite 2001: 60–61; Id. 2004). In some languages, even a long /i/ is dropped under certain conditions, e.g., Lat. venīre habeo, Ital. verrò, Lat. vestīmentu > Rum. veşmînt; vd. Belardi 1979: 79.

  Sara Kaczko also on the monetary legends from the earliest examples to the 3rd cent. BCE.54 On the other hand, the coinage of the Roman colony reads ΑΙΤΑΝΟ in the Greek alphabet, which corresponds unmistakably to the Latin Paestum. Notably, the first issue has been dated even before the foundation of the colony, though there is disagreement on this point among scholars.55 At any rate, it seems likely that the Lucanians knew the toponym of the southern Italian colony before they conquered the city around the 4th cent. BCE56 and that the form Παιστανο is a genitive referring to the (Lucanian) inhabitants of the colony (cf. also the monetary legends ROMANO, SVESANO, ΛΟΥΚΑΝΟΜ etc.), that is, an ethnonym comparable to the toponym attested in Greek literary sources in the forms Παιστός and Παῖστον. Scholars indeed agree that Παιστανο derives from Ποσειδανία with syncope of the second syllable.57 The Ποσδαν on the Lucanian pelike, an artifact datable to the end of the 5th cent. BCE, is, then, the earliest known available instance of the kind of syncope found in Παιστανο. As regards Παιστανο, a “sorte d’adaptation étrusco-italique du toponyme grec Ποσειδ(αν-ία)”58 may have been at work (-sd- > -st- being obvious in Etruscan), however, given that the sequence -sd- is unstable because the two sounds differ in voicing and that assimilations in voicing often occur,59 modification in -st- may not have been an exclusively Etruscan phenomenon, so that a solution in “Lucanian terms” seems possible. At any rate, Lucanian interference seems very likely in the case of Ποσδαν.

 54 There is also an example of a coin reading Ποσιδαν(ιατας) dated between 410 and 325 BCE, for which see Dubois 2002: 60. 55 The first coins inscribed with Παιστανο have been dated to the first quarter of the 3rd cent. BCE, and, in any case, ante 280 BCE, by Pontrandolfo 1983, both on the basis of the dating of the archaeological context of the coins and on the hiatus between the Greek coinage of Poseidonia and the Παιστανο coins; cf. also Wonder 2002: 40 fn. 2. On the other hand, Burnett — Crawford 1998 argued for a dating post 269 BCE, possibly in the 250s; see recently also Dubois 2002: 60 and Hornsaes 2004 (with previous bibliography). Afterward, the coins begin to be engraved with the legend in the Latin alphabet PAISTANO/PAIS and eventually PAES in the 1st cent. BCE. 56 For a dating of the Lucanian takeover of Poseidonia to the end of the 5th cent. BCE, see Wonder 2002: 40 fn. 1; for a dating in the 4th cent. BCE, see Musti 2009. 57 See, e.g., Dubois 2002: 61. 58 Dubois 2002: 61 (also mentioning the uncertainty regarding the origin of /ai/). 59 Cf., e.g., *h1ḗdsi > *hḗtsi (Hitt. ēzzi /étsi/, skr. átsi); *leĝ-tó- > lek-to (Gr. λεκτός); the reverse assimilation is found in *ni-sd-o > ni-zd-o, see Fortson 2004: 63; cf. Meillet 19657: 132; Leumann 19775: 204 ff.

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

 Conclusions The phonological modification of Ποσειδ- into Παιστ- (cf. the toponym Ποσειδανία and the ethnonym Παιστανο) and the syncope in Herekleís < Ἡρακλῆς, therefore, offer parallels for Ποσδαν. The finely-crafted pelike — of Lucanian manufacture, found in Heraclea, produced for local consumption in an area of contact between Greeks and local populations, decorated with a Greek myth, and reading the name of a Greek god in the standardized Greek alphabet but in the form Ποσδαν, a phonological variant attributable to the interference of the local language spoken by the Lucanians — represents yet another remarkable example of multiculturalism and multilingualism in 4th cent. BCE Magna Graecia.

Bibliography Adams, J.L. (2003), Bilingualism and the Latin Language, Cambridge. Ahrens, H.L. (1843), De graecae linguae dialectis, II. De dialecto dorica, Göttingae. Ahrens, H.L. (1894), “Über den Namen des Poseidons”, Philologus 23, 1–27. Allen, T.W. (1973), Accent and Rhythm, Cambridge. Belardi, W. (1979), Dal latino alle lingue romanze. I. Il vocalismo, Roma. Benediktsson, H. (1960), “The vowel syncope in Oscan-Umbrian”, Norsk Tidsskrift for Sprogvidenskap 19, 157–297. Bottiglioni, G. (1954), Manuale dei dialetti italici, Roma. Buck, C.D. (1904), A Grammar of Oscan and Umbrian, Boston. Buck, C.D. (1955), The Greek Dialects, Chicago. Burnett, A./Crawford, H. (1998), “Overstrikes at Neapolis and coinage in Poseidonia-Paestum”, in: R. Ashton/S. Hurter/G. Le Rider/R. Bland (eds.), Studies in Numismatic Method in Memory of Martin Jessop Price, London, 55–57. Canepari, L. (1979), Introduzione alla fonetica, Torino. Cassio, A.C. (2002), “Il dialetto greco di Taranto”, in: Taranto e il Mediterraneo. Atti del XLI Convegno di Studi sulla Magna Grecia, Taranto, 435–465. CGF I = G. Kaibel (ed.), Comicorum Graecorum Fragmenta, I, Berolini 1899. Chadwick, J. (1992), “The Thessalian accent”, Glotta 70, 2–14. Chantraine, P. (19612), Morphologie historique du Grec, Paris. Clackson, J. (2007), Indo-European Linguistics. An Introduction, Cambridge. Clarysse, W. (1997), “Greek accents on Egyptian names”, ZPE 119, 177–184. Crosswhite, K. (2001), Vowel Reduction in Optimality Theory, New York. Crosswhite, K. (2004), “Vowel reduction”, in: B. Hayes/R. Kirchner/D. Steriade (eds.), Phonetically Based Phonology, Cambridge/New York, 191–231. Degrassi, N. (1965), “Il Pittore di Policoro e l’officina ceramica protoitaliota di Heraclea lucana”, BdA 50, 5–37. del Barrio Vega, M.L. (2017), “Remarques sur le suffixe -ωνδας”, in: A. Alonso Déniz/L. Dubois/ C. Le Feuvre/S. Minon, La suffixation des anthroponymes grecs antiques. Actes du col-

  Sara Kaczko loque international de Lyon, 17–19 september, Université Jean-Moulin-Lyon-3, Genève, 579–590. DELG = Chantraine, P., Dictionnaire étymologique de la langue grecque, Paris 1968–1980. Denoyelle, M. (2002), “Style individuel, style local et centres de production: retour sur le cratère des «Karneia»”, MEFRA, 114/2, 587–609. Denoyelle, M. (2002a), “Il mito greco in Occidente nel V secolo: Metaponto ed Herakleia”, in: M.L. Mava/M. Osanna (eds.), Immagini e mito nella Basilicata antica, Venosa, 104–112. Devine, A.M./Stephens, L.D. (1994), The Prosody of Greek Speech, New York/Oxford. DGE = E. Schwyzer, Dialectorum Graecarum exempla epigraphica potiora, Lipsiae 19233. Dubois, L. (2002), Inscriptions Grecques Dialectales de Grande Grèce, II, Colonies Achéenes, Genève. Dubois, L. (2017), “Des noms en -ᾶς”, in: A. Alonso Déniz/L. Dubois/C. Le Feuvre/S. Minon, La suffixation des anthroponymes grecs antiques. Actes du colloque international de Lyon, 17–19 september, Université Jean-Moulin-Lyon-3, Genève, 299–332. Fortson, B.W. (2004), Indo-European Language and Culture: An Introduction, Oxford/Malden, MA. Gadaleta, G. (2003a), “La ceramica italiota e siceliota a soggetto tragico nei contesti archeologici delle colonie e dei centri indigeni dell’Italia meridionale e della Sicilia”, in: L. Todisco, La ceramica figurata a soggetto tragico in Magna Grecia e in Sicilia, Roma, 133–223. Gadaleta, G. (2003b), “Catalogo dei contesti (vasi italioti e sicelioti)”, in: L. Todisco, La ceramica figurata a soggetto tragico in Magna Grecia e in Sicilia, Roma, 533–572. GEW = Frisk, H., Griechisches etymologisches Wörterbuch, Heidelberg 1960–1972. GG = Grammatici Graeci, Leipzig 1867–1910 (repr. Hildesheim 1965). Gignac, F.T. (1976), A Grammar of the Greek Papyri of the Roman and Byzantine Periods, I, Milano. Heubeck, A. (1959), “Poseidon”, IF 64, 225–240. Hornsaes, H.W. (2004), “Romanization at Paestum in the 3rd c. BC: a note on the chronology of the ΠΑΙSTANO coins and the interpretation of the wall-paintings from the Spinazzo cemetery”, JRA 17, 305–311. Janda, M. (2000), Eleusis: das indogermanische Erbe der Mysterien, Innsbruck. Kaczko, S. (2021), “Tra epigrafia e letteratura: circolazione, variazioni e doppie redazioni di epigrammi. Il caso di SEG 41. 855 vs. Aristot. fr. 644 Rose”, Scienze dell’Antichità 27.1, 259–274. Kaczko, S. (forthcoming), Commedia dorica, edizioni moderne e varianti del nome di Poseidone: alcune considerazioni su Epich. fr. 70 K.A., Sophr. fr. 125 K.A. ed Eupol. fr. 149 K.A. Kretschmer, P. (1909), “Zur Geschichte der griechischen Dialekte”, Glotta 1, 9–59. Lazzeroni, R. (1972), “Contatti di lingue e di culture nell’Italia antica: elementi greci nei dialetti italici”, SSL 12, 1–24. Lejeune, M. (1945), “En marge d’inscriptions grecques dialectales, II”, REA 47, 97–115. Lejeune, M. (1972), Phonologique historique du mycénien et du grec ancien, Paris. Lejeune, M. (1975), “Reflexions sur la phonologie du vocalisme osque”, BSL 70, 233–251. Leumann, M. (19775), Lateinische Laut- und Formenlehre, München. Littleton, C.S. (1973), “Poseidon as a reflex of the Indo-European ‘Source of Waters’ God”, JIES 1, 423–440. LSAG = Jeffery, L.H., The Local Scripts of Archaic Greece (rev. edition), Oxford 1990. Maiden, M. (1995), “Evidence from the Italian dialects for the internal structure of prosodic domains”, in: J.C. Smith/M. Maiden (eds.), Linguistic Theory and the Romance Languages, Amsterdam, 115–131. Mayrhofer, M. (1952), “Zum Namen Poseidon”, AAHG 5, 59–61.

Ποσειδῶν, Ποσδαν, Paestum, and a Greek God in Lucanian Attire  

Mayser, E. (19232), Grammatik der griechischen Papyri aus der Ptolemäerzeit I: Laut- und Wortlehre, Berlin/Leipzig. Meillet, A. (19758), Aperc̹ u d’une histoire de la langue grecque, Paris. Meiser, G. (1986), Lautgeschichte der umbrischen Sprache, Innsbruck. Meiser, G. (2017), “The phonology of Italic”, in: J. Klein/B. Joseph/M. Fritz (in cooperation with M. Wenthe), Handbook of Comparative and Historical Indo-European Linguistics, Berlin/ Boston, 743–751. Méndez Dosuna, J. (1994), “Contactos silábicos y procesos de geminación en greco antiguo. A propósito de las variantes dialectales ορρος (át. ὅρος) y κορρα (át. Κόρη), Die Sprache 36, 103–127. ML = R. Meiggs/D.M. Lewis, A Selection of Greek Historical Inscriptions. To the End of the Fifth Century B.C., Oxford 1969. Moret, J.M. (1979), “Un ancêtre du phylactère: le pilier inscrit des vases Italiotes. II”, RA, n.s. 1, 235–294. Musti, D. (2009), “Greci e indigeni fra Campania e Lucania nel IV sec. a.C.: il quadro storico”, in: M. Osanna (cur.), Verso la città: forme insediative in Lucania e nel mondo italico fra IV e III sec. a.C.: atti delle giornate di studio, Venosa, 13–14 maggio 2006, Venosa, 13–24. Myers, S. (2005), “Vowel duration and neutralization of vowel length contrasts in Kinyarwanda”, Journal of Phonetics 33.4, 427–446. Neumann, G. (1908), De nominibus Boeotorum propriis, diss. Königsberg. Nikolaev, A. (2019), “Δάειρα, the Lady of the Waters”, Indo-European Linguistics and Classical Philology 23, 790–795. Nishimura, K. (2012), “Vowel reduction and deletion in Sabellic”, in: B.N. Whitehead/T. Olander/ B.A. Nielssen/J.A. Rasmussen, The Sound of Indo-European. Phonetics, Phonemics, and Morphophonemics, Copenhagen, 381–398. Oakley, J.H. (2016), “Inscriptions on Apulian red-figure vases: a survey”, in: D. Yatromanolakis, Epigraphy of Art: Ancient Greek Vase-Inscriptions and Vase-Paintings, Oxford, 121–133. Oettinger, N. (2020), “Poseidon”, MSS 73/2, 97–108. Pianu, G. (1989), “Riflessioni sulla cosiddetta tomba del Pittore di Policoro”, in: AA.VV., Studi su Siris-Eraclea, Roma, 85–94. Pianu, G. (1997), “Eraclea Lucana”, Ostraka 6, 161–165. von Planta, R. (1892), Grammatik des Oskischen-Umbrischen Dialekte, I, Einleitung und Lautlehre, Strassburg. Poccetti, P. (1982/1983), “Osco trístaamentud: un problema di interferenza linguistica”, IncLing 8, 150–155. Poccetti, P. (2012), “Reflexes of variations in Latin and Greek through neither Latin nor Greek documentation: names of Greek religion and mythology in the languages of Ancient Italy”, in: M. Leiwo/H. Halla-aho/M. Vjerros, Variation and Change in Greek and Latin, Helsinki, 71–96. Pontrandolfo, A. (1983), “Per una precisazione della cronologia delle monete a leggenda Paistano”, AnnIstItalNum 30, 63–81. P&P = O. Taplin, Pots and Plays, Interactions between Tragedy and Greek Vase-Painting of the Fourth Century B.C., Los Angeles 2007. Probert, P. (2006), Ancient Greek Accentuation, Oxford. Prosdocimi, A. L. (1976), “Sui grecismi nell’osco”, in: Scritti in onore di Giuliano Bonfante, Brescia, 781–866.

  Sara Kaczko Prosdocimi, A.L. (1978), “Contatti e conflitti di lingue nell’Italia antica”, in: A.L. Prosdocimi (cur.), Popoli e civiltà dell’Italia antica, VI. Lingue e dialetti, Roma, 1031–1088. Pugliese Carratelli, G. et al. (eds.) (1983), Megale Hellas: storia e civiltà della Magna Grecia, Milano. Rau, J. (2007), “The derivational history of the Greek stems in -ád”, Harvard Working Papers in Linguistics, 12, 137–173. E. Risch, Wortbildung der homerischen Sprache, Berlin/New York 19742. Rix, H. (19922), Historische Grammatik des Griechischen, Darmstadt. Rix, H. (1998), “Eine neue frühsabellische Inschrift und der altitalische Präventiv”, HS 111, 247–269. Roscino, C. (2012), “Iscrizioni”, in: L. Todisco, La ceramica a figure rosse della Magna Grecia e della Sicilia, II, Roma, 337–342. Ruijgh, C.J. (1967), “Sur le nom de Poséidon et sur les noms en -ᾱ-ϝον-, -ῑ-ϝον-”, REG 80, 6–16. RVAp = Trendall, A.D., The Red-figured Vases of Apulia, I, Early and Middle Apulian, Oxford 1978. Schwyzer, E. (1939), Griechische Grammatik, I, München. Solmsen, F. (1909), Beiträge zur griechischen Wortforschung, I, Strassburg. Szemerényi, O. (1964), Syncope in Greek and Indo-European and the nature of Indo-European Accent, Napoli. Teodorsson, S.T. (1974), The Phonemic System of the Attic Dialect, 400–340 B.C., Göteborg. Threatte, L. (1980), The Grammar of Attic Inscriptions, I. Phonology, Berlin/New York. Thumb, A. (1901), Die griechische Sprache im Zeitalter des Hellenismus, Strassburg. Todisco, L. (2003), La ceramica figurata a soggetto tragico in Magna Grecia e in Sicilia, Roma. Todisco, L. (2012), La ceramica a figure rosse della Magna Grecia e della Sicilia, I-III, Roma. Trendall, A.D. (1967), The Red-figured Vases of Lucania, Campania and Sicily, Oxford. Trendall, A.D. (1983), The Red-figured Vases of Lucania, Campania and Sicily, III Supplement, (BICS Suppl. 41), London. Uguzzoni, A. (1968), “Analisi linguistica”, in: A. Uguzzoni/F. Ghinatti, Le tavole greche di Eraclea, Roma, 29–79. Untermann, J. (2000), Wörterbuch des Oskisch-Umbrischen, Heidelberg. Yuen, C.L.-K. (2000), “The perception of Japanese devoiced vowels”, Proceedings of the Chicago Linguistic Society 36, 531–547. Wachter, R. (2001), Non-Attic Greek Vase Inscriptions, Oxford. Wallace, R.E. (2008), Zikh Rasna: a Manual of the Etruscan Language and Inscriptions, Ann Arbor/New York. Weidauer, L. (1969), “Poseidon und Eumolpos auf einer Pelike aus Policoro”, AK 12, 91–93. Weiss, M. (1993), Studies in Italic Nominal Morphology (Cornell University Dissertation), Ithaca, NY. Willi, Α. (2007), “Demeter, Gê, and the Indo-European word(s) for ‘earth’”, HS 120, 160–194. Wonder, J. (2002), “What happened to the Greeks in Lucanian-occupied Paestum? Multiculturalism in Southern Italy”, Phoenix 56, 1/2, 40–55. Zair, N. (2016), “Vowel weakening in the Sabellic languages as language contact”, IF 121, 295– 315. Zuchtriegel, G. (2018), Colonization and Subalternity in Classical Greece. Experience of the Nonelite Population, Cambridge/New York.

Felicia Logozzo and Liana Tronci

Motion and Posture Verbs in Multiverb Constructions: Evidence from the New Testament Abstract: This study deals with two types of multi-verb constructions in New Testament Greek, analysing them from a syntactic, semantic, and textual point of view. In both constructions, a verb of movement or posture (V1) combines with the main verb of the sentence (V2), providing deictic information. The difference between the two constructions lies in the form of V1, which is a participle in one of the constructions and a verb coordinated with the main verb in the other. The study shows that the two constructions have a different distribution based on the following parameters: (1) opposition between story and discourse, (2) lexical type of V1. Finally, the two constructions do not have equal distribution in the four Gospels.

 Introduction In the Greek of the New Testament (henceforth NT), there are several multiverb constructions (henceforth MVCs) formed by two verbs: the first is either a verb of motion, e.g. ἔρχομαι “I come”, or a verb of posture, e.g. ἀνίστημι “I arise”, and the second does not impose any semantic restrictions on the event. The two verbs of a MVC are not equivalent in their syntactic and semantic functions. In particular, the first verb (V1) is not a full verb from a syntactic point of view in that it does not govern any argument. Semantically, the motion or posture conveyed by V1 is considered to be part of the action conveyed by the second verb (V2), which is the full verb of the clause. V1 hence provides the clause with some deictic information, concerning the place and the posture of the subject. By the V1 ἔρχομαι we learn that the subject approaches the interlocutor or another participant in order to accomplish the main action of the clause, while by the V1

 This research was carried out as part of the project “Multilingualism and Minority Languages in Ancient Europe” [HERA.29.015| CASSIO], funded by the Hera Joint Research Programme “Uses of the Past”, Horizon 2020 – 649307. This article is the result of joint work by the two authors. However, for academic purposes, Felicia Logozzo is responsible for Sections 1, 2, 3.2, 3.4, and Liana Tronci for Sections 3.1, 3.3, 3.5, 4. https://doi.org/10.1515/9783110779684-009

  Felicia Logozzo and Liana Tronci ἀνίστημι we learn that the subject was sitting and that he stands up in order to carry out the action conveyed by V2. As we will show later, further deictic meanings are added to spatial deixis in imperatival MVCs. The two verbs of the MVC may occur in two different morpho-syntactic configurations. The first one is the coordinating type, where the two verbs are inflected in the same mood, person and number and are linked by the coordinator καί. If the verbs are imperatives, they may also be merely juxtaposed (without any overt coordinator). The second configuration is the subordinating type, where the first verb is a participle and the second one is a finite form (or an infinitive in infinitival constructions). Hereafter, we give one example1 of each type, with the verb ἔρχομαι as V1: (1)

καὶ ἀκούσαντες οἱ μαθηταὶ αὐτοῦ ἦλθαν καὶ ἦραν τὸ πτῶμα αὐτοῦ, καὶ ἔθηκαν αὐτὸν ἐν μνημείῳ (Mk 6.29). “When his disciples heard about this, they came and took away his body and laid it in a tomb”.

(2)

εἴπατε ὅτι οἱ μαθηταὶ αὐτοῦ νυκτὸς ἐλθόντες ἔκλεψαν αὐτὸν ἡμῶν κοιμωμένων (Mt 28.13). “You are to say, ‘His disciples came by night and stole Him away while we were asleep’”.

In both constructions, the first verb, be it a finite verb or a participle, does not govern any complement of place and the movement it denotes is not independent from the action meant by the second verb. This is the case, conversely, when the verb is employed as a full verb, e.g. (3), with a finite motion verb, and (4) with a participial motion verb: (3) οὗτος ἦλθεν πρὸς αὐτὸν νυκτὸς καὶ εἶπεν αὐτῷ (Jh 3.2). “This man came to Jesus by night and said to Him”. (4) καὶ ἐλθὼν ὁ Ἰησοῦς εἰς τὴν οἰκίαν Πέτρου εἶδεν τὴν πενθερὰν αὐτοῦ βεβλημένην καὶ πυρέσσουσαν (Mt 8.14). “When Jesus came into Peter’s home, He saw his mother-in-law lying sick in bed with a fever”.

 1 Greek texts are taken from TLG http://stephanus.tlg.uci.edu/ (last accessed March 2021); English translations of Gospel examples are given according to the New American Standard Bible, as reported on https://www.biblestudytools.com/ (last accessed March 2021), with adjustments.

Motion and Posture Verbs in Multiverb Constructions  

One might think that the lack of arguments governed by V1 in MVCs is an optical illusion caused by the co-reference between the complement governed by the motion verb and the complement governed by the second verb. Against this hypothesis, example (3) makes it clear that not only does co-reference not inhibit biclausality, but also biclausality is a necessary condition for co-reference. A complement of place, though, may occur in MVCs, as in the imperatival construction with the V1 ὑπάγω “I go” in (5). In this case, the complement of place relates to the MVC in its entirety. This type of occurrence is rare in our corpus. (5) ὕπαγε νίψαι εἰς τὴν κολυμβήθραν τοῦ Σιλωάμ (Jh 9.7). “Go, wash in the pool of Siloam”.

The goal of this paper is to give an account of MVCs with motion and posture verbs in the NT, paying special attention to the distribution of the coordinating and the subordinating types. We focused on investigating the following aspects: (1) whether there are some syntactic, semantic and pragmatic differences between the two types; (2) which verbs occur as V1 in the two types; (3) how the two types are distributed in the four Gospels. The structure of the paper is as follows. In Section 2 we give a brief survey of the literature; Section 3 is devoted to the analysis of data; finally, in Section 4 we draw some conclusions.

 A brief survey of the literature As we already argued in Logozzo/Tronci (2022), the MVCs in which the two verbs are inflected in the same mood, person and number and are either coordinated by καί or juxtaposed can be analysed as pseudocoordinated constructions lato sensu. In them, the two verbs seem to have the same syntactic and semantic “weight”, but this is in fact not the case, since the first verb is not a full verb and functions as a modifier of the second verb. The function of modifier is particularly evident in imperatival MVCs, where injunction is more urgent and immediate than in simple imperatives. Compare, for instance, the serialized imperatives ὑπάγετε εἴπατε in (6) and the simple imperative εἴπατε in (7). (6) ἀλλὰ ὑπάγετε εἴπατε τοῖς μαθηταῖς αὐτοῦ καὶ τῷ Πέτρῳ ὅτι προάγει ὑμᾶς εἰς τὴν Γαλιλαίαν (Mk 16.7). “But go, tell his disciples and Peter that he is going ahead of you to Galilee”.

  Felicia Logozzo and Liana Tronci (7) καὶ ὅπου ἐὰν εἰσέλθῃ εἴπατε τῷ οἰκοδεσπότῃ ὅτι ὁ διδάσκαλος λέγει (Μk 14.14). “And wherever he enters, say to the owner of the house, ‘The Teacher says’”.

The relationship between pseudocoordinated constructions with motion verbs and imperatives is not surprising. In Homeric and Classical Greek, imperatives of motion verbs may occur before another imperative without any linker, e.g. (8): (8) μηδενός σοι, ἔφη, μελέτω, ὦ Σώκρατες, ἀλλ᾽ ἴθι διαλέγου αὐτῷ (Pl. Lys. 211c6). “Take no heed of anyone, Socrates, he said; just go on and have a talk with him”.

The status of these constructions is debated. Some scholars analyse them as serialized constructions and find evidence for serialized imperatives in several ancient Indo-European languages (cf. Hock 2002; 2014; Yates 2014). Serialization appears to be restricted to imperatives in most ancient Indo-European languages, excepting Hittite and Classical Armenian, where serialization is also attested with other moods (cf. van den Hout 2003; 2010 for Hittite and Kölligan (2021) for Classical Armenian). The restriction to imperatives in most ancient Indo-European languages led scholars to argue that serialization cannot be reconstructed as a grammatical pattern in Proto-Indo-European, as is the case, on the contrary, for many other language families (cf. Aikhenvald 2018: 208– 211). Therefore, the label of Quasi Serial Verb Constructions (QSVC) was created to account for the Proto-Indo-European state of affairs, which is reconstructed according to data from Latin, Sanskrit, Ancient Greek, Classical Armenian, and Hittite (examples follow from Yates 2014: 238). (9) a. age abduce hasce intro quas mecum adduxi, Stiche (Pl. St. 418). “Go take these (women) I’ve brought with me indoors, Stichus”. b. idáṃ te ánnaṃ yújiyam sámukṣitaṃ tásyéhi prá dravā píba (RV 8.4.12cd). “Here is your food, ready for yoking, fully sprinkled: come run drink of it”. c. ἀλλ᾽ ἴθι οἱ νέκτάρ τε καὶ ἀμβροσίην ἐρατεινὴν στάξον ἐνὶ στήθεσσ᾽, ἵνα μή μιν λιμὸς ἵκηται. (Il. 19.347–348). “But go pour nectar and lovely ambrosia into him, in (his) breast, so that hunger will not reach him”.

Motion and Posture Verbs in Multiverb Constructions  

d. ertʿ cʿoycʿ zanjn kʿo kʿahanayin (Lk 5.14). “Go show yourself to priest”. e. īt=war ašta pargamuš ḪUR.SAG.DIDLI.ḪI.A-ašaš šāḫ (KUB 17.10 i 24–5). “Go search the high mountains”.

Ancient Greek is one of the Indo-European languages in which serialization appears to be restricted to imperatives. However, there is no general agreement among scholars on the status of the first imperative in constructions such as (9c). In the recent Cambridge Grammar of Classical Greek by van Emde Boas et al. (2019: 482), imperatives such as ἴθι in (9c), ἄγε(τε), φέρε(τε) etc. are considered to be interjections meaning “come (on), go (ahead)”. They are assumed to be the result of a process of lexicalization. This is in accordance with the remark by Kühner/Gerth (1976 [1898–1904]: § 397) who already stressed the semantic function of “encouraging” (Aufmunterung) of these imperatives. The idea that they underwent lexicalization and became pragmatic markers is also suggested by Biraud (2010) in her study on interjections in Classical Greek theatre (see in particular Biraud 2010: 160–169). This topic is discussed in detail by Denizot (2011: 205–211), who, although arguing for lexicalization, points out that the different imperatives (e.g., ἄγε, φέρε, ἴθι, ἰδού) are not lexicalized to the same degree, making it difficult in many cases to decide whether an imperative is lexicalized or not (see examples in Denizot 2011: 206–210). In summary, Ancient Greek appears to accept serialization only with imperatives. True serialized imperatives, however, turn out to be very rare, given that several modifier verbs have lexicalized into interjections. Unlike serialization, pseudocoordination has not yet received much attention in studies on ancient Indo-European languages (cf. Ross 2016 for an overview). Pseudocoordination is similar to serialization as regards (a) the relationship of the two verbs to a single event, and (b) the modifier function of the first verb, which belongs to a lexically closed class (motion verbs, posture verbs etc.) and does not govern any complement. The main difference between pseudocoordination and serialization concerns the occurrence of the overt linker καί between the two verbs. Because of the overt coordinator, pseudocoordinated constructions resemble asymmetric coordinated constructions and may occur outside dialogical interaction, namely in narrative contexts. In their paper on coordination in Latin, Orlandini/Poccetti (2008: 102) argue that all verbal moods, not only imperatives, occur in Latin serialized and pseudocoordinated constructions:

  Felicia Logozzo and Liana Tronci (10) a. Si sapias, eas ac decumbas domi (Plaut. Merc. 373). “If you were prudent, you’d go home and lie down”. b. Ibo, adloquar (Plaut. Merc. 366). “I’ll go and accost him”. c. Abi, nuntia (Liv. 1.16.7). “Go, tell [the Romans]…”.

In their approach, pseudocoordination and serialization are considered to be functionally similar, even though they are slightly different in form, in accordance with the statement of Ross (2016) that “[t]he emphasis on form in both pseudocoordination and serialization may be exaggerated, as both form multiverb complex predicates with similar properties”.2 In this paper, we adopt a unified approach to pseudocoordination and serialization and we consider them as two realizations of the same functional pattern. In both of them there are two verbs, the first of which is lexically restricted to the classes of motion and posture verbs. The two verbs are usually contiguous and relate to a single event. The difference between pseudocoordination and serialization consists in the overt (i.e., syndetic, by καί) vs covert (i.e., asyndetic, by juxtaposition) coordination of the two verbs. Henceforth, the acronym PCCs (PseudoCoordinated Constructions) relates to both the syndetic type (pseudocoordination stricto sensu) and the asyndetic one (serialization), as summarized in Table 4. Tab. 4: Types of PCCs. PCCs Syndetic PCCs

 2 Ross 2016: 229.

Asyndetic PCCs

↓

↓

V καί V

V V

Motion and Posture Verbs in Multiverb Constructions  

As already mentioned in Section 1, PCCs appear to alternate with another type of construction, in which V1 is a participle and V2 the main verb of the clause. In this case too, the two verbs are contiguous and relate to a single event. Henceforth, the label of PAC (Participial Construction) is used to refer to this type of structure. PACs have not received special attention in the literature, with the exception of several studies by Pompei (2006; 2012; 2015) devoted to investigating Ancient Greek participles as converbs. According to Pompei, the participles of PACs are “cosubordinative participles”, together with other types of participle. Cosubordinative participles are “non-modifying forms, i.e., they do not answer questions such as how/when/under what circumstances the main predication occurs”,3 can be replaced by a clause coordinated to the main clause (and not by a subordinated clause) and provide foreground (and not background) information. In Logozzo/Tronci (2019) and (2019 [2020]) we suggested a different analysis of PACs, based on all the occurrences of the participle of ἔρχομαι plus another verb in the Gospels. We argued that PACs function as a syntactic unit and relate to a single event. The observation that the participle of a PAC can be replaced by a verb coordinated to the main verb is not a clue that the two verbs are “co-ordinated”; rather, it is evidence of pseudocoordination, as we will discuss in detail in this paper. In fact, PACs appear to share with PCCs some typical features, such as the contiguity of the two verbs and the lack of arguments governed by V1, be it a participle or a verb coordinated to V2 (cf. Logozzo/Tronci 2019 [2020] for further details on PACs). In Section 3, we analyse the distribution of PACs and PCCs in the four Gospels. Given that they do not seem to be in free alternation, we investigate which parameters determine the occurrence of one construction instead of the other. A crucial role is played, in our view, by textual and pragmatic features, such as spatial and temporal deixis and the types of speech act, especially imperatival directives. We also suggest that, within a more general perspective, the difference between dialogical and narrative contexts is relevant (discourse vs narrative story according to Benveniste 1966 [1959]; cf. § 4).

 3 Pompei 2015: 146.

  Felicia Logozzo and Liana Tronci

 Analysis of data 3.1. The corpus for this research was constituted by searching for all occurrences of verbs of motion and posture in the online TLG4 and by selecting the relevant constructions for our investigation. For the sake of uniformity, we limited the analysis to the Gospels. Table 5 summarizes the quantitative data, providing the list of motion and posture verbs that occur as V1 and the number of occurrences for each verb in both PACs and PCCs. Tab. 5: PACs and PCCs with motion and posture verbs in the Gospels. V

PACs

PCCs

Total

ἀνίστημι “I get up”







ἀπέρχομαι “I go away”













ἐξέρχομαι “I go out”







ἔρχομαι “I come”







ἵστημι “I stay up”



















ἐγείρω “I wake up, I get up”

καταβαίνω “I go down” πορεύομαι “I go” προσέρχομαι “I go to, I approach”







ὑπάγω “I go”













Total

Some verbs, e.g., ἔρχομαι and ἐγείρω, are well attested in both types of construction, whilst other verbs either occur in only one type, e.g., ἀνίστημι and ὑπάγω, or are mostly attested in one type, e.g., ἀπέρχομαι and πορεύομαι. The distribution of the verbs and the constructions in the four Gospels is not even: in Luke’s Gospel, the participial construction is preferred, so no occurrence of MVCs with ὑπάγω is attested; conversely, the participial construction is not usual in John’s Gospel, hence the V1s that only occur in the participial type are not attested there. The following Sections are devoted to analysing how the two constructions are distributed in the four Gospels and in the different textual units. Overall,

 4 (last accessed March 2021).

Motion and Posture Verbs in Multiverb Constructions  

PCCs are rarer in Luke’s Gospel than in the other three Gospels. Matthew’s and Mark’s Gospels make use of both PACs and PCCs, whilst in John’s Gospel we found mainly PCCs. The largest number of PACs and PCCs occurs in Matthew’s Gospel. Another aspect to stress is that both constructions host many occurrences of imperatives; however, most imperatives occur in PCCs. We counted 24 occurrences out of 57 PCCs against 21 occurrences out of 106 PACs (11 of which with the V1 πορεύομαι), i.e., a ratio of 42% in PCCs and 20% in PACs. In the following pair, one PAC and one PCC occur with the same main verb in the imperative. The motion verbs are the aorist participle of ἀπέρχομαι in the PAC and the present imperative of ὑπάγω in the PCC. Note that the PAC is attested in the Gospel of Luke and the PCC in that of Matthew. (11) καὶ αὐτὸς παρήγγειλεν αὐτῷ μηδενὶ εἰπεῖν, ἀλλὰ ἀπελθὼν δεῖξον σεαυτὸν τῷ ἱερεῖ (Lk 5.14). “And He ordered him to tell no one, ‘But go and show yourself to the priest’”. (12) καὶ λέγει αὐτῷ ὁ Ἰησοῦς, ὅρα μηδενὶ εἴπῃς, ἀλλὰ ὕπαγε σεαυτὸν δεῖξον τῷ ἱερεῖ (Mt 8.4). “And Jesus said to him, ‘See that you tell no one; but go, show yourself to the priest’”.

The correlation with the imperative is interesting since the imperative is the typical verbal mood of dialogical situations and, particularly, of directive speech acts. As is well known, the content of the speech cannot be detached from the deictic categories of person, time and space in dialogical situations. This is even truer in directives, where the interlocutor is requested to perform what the speaker asks him. As argued in Logozzo/Tronci (2022), motion verbs in PCCs encode spatial deixis (approaching the speaker or the addressee, departure from the speaker etc.) and also provide the construction with additional nuances of immediacy and urgency (temporal deixis). 3.2. Let us begin with the verb ἔρχομαι, which occurs in both types of construction and is the most frequently attested verb (38 total occurrences). The following couple of clauses includes one PAC (13) and one PCC (14) with the same main verb προσέπεσεν. (13) ἀλλ’ εὐθὺς ἀκούσασα γυνὴ περὶ αὐτοῦ, ἧς εἶχεν τὸ θυγάτριον αὐτῆς πνεῦμα ἀκάθαρτον, ἐλθοῦσα προσέπεσεν πρὸς τοὺς πόδας αὐτοῦ (Mk 7.25). “But after hearing of Him, a woman whose little daughter had an unclean spirit immediately came and fell at His feet”.

  Felicia Logozzo and Liana Tronci (14) ἡ δὲ γυνὴ φοβηθεῖσα καὶ τρέμουσα, εἰδυῖα ὃ γέγονεν αὐτῇ, ἦλθεν καὶ προσέπεσεν αὐτῷ καὶ εἶπεν αὐτῷ πᾶσαν τὴν ἀλήθειαν (Mk 5.33). “But the woman fearing and trembling, aware of what had happened to her, came and fell down before Him and told Him the whole truth”.

As regards the distribution in the Gospels, we observed that PACs with ἔρχομαι occur in the Gospels of Matthew, Mark and Luke, but not in that of John, whilst PCCs occur in the Gospels of Mark, Luke and John, but not in that of Matthew. We also observed some differences from one Gospel to another in the morphosyntactic features of the modifier verb. As a participle, ἔρχομαι is usually inflected in the aorist, independently from the tense-aspectual inflection of V2: (15=2) εἴπατε ὅτι οἱ μαθηταὶ αὐτοῦ νυκτὸς ἐλθόντες ἔκλεψαν αὐτὸν ἡμῶν κοιμωμένων (Mt 28.13). “You are to say, ‘His disciples came by night and stole Him away while we were asleep’”. (16) ἰδοὺ πολλοὶ τελῶναι καὶ ἁμαρτωλοὶ ἐλθόντες συνανέκειντο τῷ Ἰησοῦ καὶ τοῖς μαθηταῖς αὐτοῦ (Mt 9.10). “Behold, many tax collectors and sinners came and were dining with Jesus and His disciples”.

The occurrence of the aorist participle is not problematic. As Blass, Debrunner, and Funk (1961 [1896]: § 339) suggest, the aorist inflection can be explained in terms of “relative past time”: before the action denoted by the participle is completed and, then, the action of the finite verb is done (see also discussion in Porter 1989: 380–381). In Matthew’s and Mark’s Gospels, all PACs with V1 ἔρχομαι exhibit the aorist participle. In Luke’s Gospel, conversely, only 2 PACs (Lk 5.7, 7.3) display the aorist participle, against 4 occurrences with the present participle (Lk 13.14, 15.25, 16.21, 18.5). This distribution between aorist and present participle appears to match the aspectual inflection of the main verb, the only exception being Lk 15.25. Thus, the participle is aorist when the V2 is an aorist, e.g. (17); conversely, the participle is present when the V2 is imperfective (a present or an imperfect), e.g. (18). (17) ἀκούσας δὲ περὶ τοῦ Ἰησοῦ ἀπέστειλεν πρὸς αὐτὸν πρεσβυτέρους τῶν Ἰουδαίων, ἐρωτῶν αὐτὸν ὅπως ἐλθὼν διασώσῃ τὸν δοῦλον αὐτοῦ (Lk 7.3). “When he heard about Jesus, he sent some Jewish elders asking Him to come and save the life of his slave”.

Motion and Posture Verbs in Multiverb Constructions  

(18) εἰ καὶ τὸν θεὸν οὐ φοβοῦμαι οὐδὲ ἄνθρωπον ἐντρέπομαι, διά γε τὸ παρέχειν μοι κόπον τὴν χήραν ταύτην, ἐκδικήσω αὐτήν, ἵνα μὴ εἰς τέλος ἐρχομένη ὑπωπιάζῃ με (Lk 18.4–5). “Even though I do not fear God nor respect man, yet because this widow bothers me, I will give her legal protection, otherwise by continually coming she will wear me out”.

Our hypothesis is that Luke’s choice of the aorist vs present participle depends on the aspectual features exhibited by V2. The tendency of the participle to match the aspectual features of the main verb, on the one hand, makes it clear that the action meant by the participle is less autonomous aspectually from that of the main verb and, on the other hand, points to a greater compactness of the PAC. In Luke’s system, the “relative past time” of the participle is not the relevant feature, since the participle is an aspectual marker of the event in its entirety, no less than the main verb. If this hypothesis is correct, the sequence of the present participle ἐρχόμενος and the aorist ἤγγισεν in the following instance remains to be explained: (19) ἦν δὲ ὁ υἱὸς αὐτοῦ ὁ πρεσβύτερος ἐν ἀγρῷ· καὶ ὡς ἐρχόμενος ἤγγισεν τῇ οἰκίᾳ, ἤκουσεν συμφωνίας καὶ χορῶν (Lk 15.25). “Now his older son was in the field, and when he came and approached the house, he heard music and dancing”.

This example comes from the parable of the prodigal son. Luke is reporting here the scene in which the older son, who had remained at home when his brother left, is coming home after his day’s work. In the first clause of (19), Luke introduces the new referent, i.e., the older son, and says that he was not at home, where a great party had been prepared for the younger son, but in the field. The second clause of (19) reports what happened when the older son came back home: as he was approaching home, he heard some music and dancing. The subordinate clause introduced by ὡς provides the background information for the action of the main clause. From the aspectual viewpoint, the two actions of the subordinate clause and the main clause are opposed, in that the first one is imperfective and progressive (ἐρχόμενος ἤγγισεν) while the second one is punctual and perfective (ἤκουσεν). Within this perspective, the present participle of the PAC ἐρχόμενος ἤγγισεν serves the purpose of providing the verbal sequence with an imperfective value, given that the imperfect of the verb ἐγγίζω is rare and never combines with human subjects in the NT.5 Indirect evidence for this  5 The imperfect ἤγγιζεν is attested only twice in the NT, once in the Gospel of Luke (Lk 22.1) and once in the Acts of Apostles (Acts 7.17). In both cases, the subject of the verb is an event

  Felicia Logozzo and Liana Tronci analysis comes from the Latin translation in Jerome’s Vulgate. As is frequently the case in the Latin Vulgate, Jerome renders the Greek PAC by a Latin PCC. What is relevant for us is that he translates the Greek aorist by the Latin imperfect subjunctive (et cum veniret et adpropinquaret domui audivit symphoniam et chorum),6 thus pointing to the progressive aspectual value of the Greek PAC. Turning to PCCs with V1 ἔρχομαι, they are all of the syndetic type and basically occur in narrative contexts (cf. Logozzo/Tronci 2022). The verbs are inflected in the present, e.g. (20), the aorist, e.g. (21), and the future, e.g., (22). (20) καὶ ἦσαν οἱ μαθηταὶ Ἰωάννου καὶ οἱ Φαρισαῖοι νηστεύοντες. καὶ ἔρχονται καὶ λέγουσιν αὐτῷ (Mk 2.18). “And there were John’s disciples and the Pharisees that were fasting; and they came and said to Him”. (21) καὶ ἀκούσαντες οἱ μαθηταὶ αὐτοῦ ἦλθαν καὶ ἦραν τὸ πτῶμα αὐτοῦ, καὶ ἔθηκαν αὐτὸν ἐν μνημείῳ (Mk 6.29). “When his disciples heard about this, they came and took away his body and laid it in a tomb”. (22) τί ποιήσει ὁ κύριος τοῦ ἀμπελῶνος; ἐλεύσεται καὶ ἀπολέσει τοὺς γεωργούς, καὶ δώσει τὸν ἀμπελῶνα ἄλλοις (Mk 12.9). “What will the owner of the vineyard do? He will come and destroy the vine-growers, and will give the vineyard to others”.

About half of the PCCs with V1 ἔρχομαι are attested in the Gospel of John, who also makes use of PCCs with imperatives, unlike Mark and Luke: (23) καὶ εἶπεν· ποῦ τεθείκατε αὐτόν; λέγουσιν αὐτῷ· κύριε, ἔρχου καὶ ἴδε (Jh 11.34). “And he said, ‘Where have you laid him?’ They said to Him, ‘Lord, come and see’”.

To sum up, PACs and PCCs with ἔρχομαι as V1 are not distributed equally in the Gospels. Only one type is admitted in Matthew’s and John’s Gospels, namely the PAC in Matthew and the PCC in John. This distribution cannot be explained by  noun (ἡ ἑορτὴ τῶν ἀζύμων “the Feast of unleavened bread” in Luke’s Gospel and ὁ χρόνος τῆς ἐπαγγελίας “the time of the promise” in the Acts). 6 According to “Biblia Sacra Vulgata. Editio quinta” edited by Robert Weber and Roger Gryson; Deutsche Bibelgesellschaft, https://www.academic-bible.com/en/online-bibles/aboutthe-online-bibles/ (last accessed April 2021).

Motion and Posture Verbs in Multiverb Constructions  

the fact that PCC is dispreferred in Matthew and PAC in John, since they make use of both constructions with other verbs. Concerning Matthew’s Gospel in particular, we can claim that pseudocoordination is a very common syntactic construction with other verbs than ἔρχομαι (see below). It can be assumed, then, that the different picture depends on Matthew’s desire to model his language on Ancient Greek tradition, in which combinations of ἐλθών plus a finite verb had been attested since Homer onwards (cf. Logozzo/Tronci 2019 [2020]). Moreover, we should not forget that Matthew tends to make greater use of PCCs in dialogues than in narrative contexts. John’s Gospel, conversely, does not make extensive use of PACs, whatever the modifier verb, and definitely prefers PCCs. The preference for PCCs is consistent with the typical paratactic style of John (Burrows 1930: 104), as well as the “rarity of the aorist participle describing action anterior to the finite verb”7, which Burney (1922) put forward as evidence for the Aramaic origin of John’s Gospel. 3.3. Let us turn now to MVCs formed with the verbs of going ἀπέρχομαι, πορεύομαι, and ὑπάγω. The first remark is that there are three verbs for “going” as opposed to only one verb for “coming”. The three verbs are not exact synonyms in Classical Greek. The verb ἀπέρχομαι is intransitive and denotes departure from one place; as a full verb, it may govern a complement of motion from place (ἀπό, ἐκ + genitive) and a complement of motion to place (εἰς + accusative). The Classical Greek πορεύω/πορεύομαι is a transitive/intransitive verb, which means “to carry” in the active form and “to be driven, to walk, to go” in the middle form (LSJ, s.v.). In Hellenistic Greek, only the middle and intransitive form is preserved, cf. the entry πορεύομαι in the Dictionary of the Septuagint by Muraoka (2009, s.v.). In the NT, outside MVCs, it is usually combined with a complement of motion to place (εἰς, πρός + accusative) or is used absolutely to mean departure. Finally, the verb ὑπάγω is basically inflected in the active voice and can be both transitive (“lead or bring under”) and intransitive (“go away, withdraw, retire”) in Classical Greek. It is seldom attested in Classical Greek but the number of its occurrences started to increase from the 1st cent. CE onwards. In the NT, it is only inflected in the present tense, with the exception of two occurrences in the imperfect (Jh 6.21, 12.11). Unlike ἀπέρχομαι and πορεύομαι, which are attested in both PACs and PCCs, ὑπάγω only occurs in PCCs.

 7 Burney 1922: 56.

  Felicia Logozzo and Liana Tronci Our investigation of MVCs with verbs of going begins by analysing PCCs with V1 ὑπάγω.8 The first remark concerns the correlation between the PCCs with ὑπάγω and the imperative mood. Out of the total number of 15 PCCs, only two occurrences do not display the verbs in the imperative: (24) ὁμοία ἐστὶν ἡ βασιλεία τῶν οὐρανῶν θησαυρῷ κεκρυμμένῳ ἐν τῷ ἀγρῷ, ὃν εὑρὼν ἄνθρωπος ἔκρυψεν, καὶ ἀπὸ τῆς χαρᾶς αὐτοῦ ὑπάγει καὶ πωλεῖ πάντα ὅσα ἔχει καὶ ἀγοράζει τὸν ἀγρὸν ἐκεῖνον (Mt 13.44). “The kingdom of heaven is like a treasure hidden in the field, which a man found and hid again; and from joy over it he goes and sells all that he has and buys that field”. (25) οὐχ ὑμεῖς με ἐξελέξασθε, ἀλλ’ ἐγὼ ἐξελεξάμην ὑμᾶς καὶ ἔθηκα ὑμᾶς ἵνα ὑμεῖς ὑπάγητε καὶ καρπὸν φέρητε καὶ ὁ καρπὸς ὑμῶν μένῃ (Jh 15.16). “You did not choose Me but I chose you, and appointed you that you would go and bear fruit, and that your fruit would remain”.

The second remark concerns the constraint on the tense-aspectual inflection of ὑπάγω. In all occurrences, ὑπάγω is inflected in the present stem, regardless of the tense-aspectual inflection of the main verb. That means that in most cases V1 and V2 do not match in tense and aspect, as V2 is inflected in the aorist 12 times. One example is given below: (26) εἰ θέλεις τέλειος εἶναι, ὕπαγε πώλησόν σου τὰ ὑπάρχοντα καὶ δὸς πτωχοῖς, καὶ ἕξεις θησαυρὸν ἐν οὐρανοῖς, καὶ δεῦρο ἀκολούθει μοι (Mt 19.21). “If you wish to be complete, go sell your possessions and give to the poor, and you will have treasure in heaven; and come, follow Me”.

As is well known, in the imperative more than in other moods, the opposition between present and aorist does not code any temporal contrast, only an aspectual one (Denizot, 2011: 216 and references therein). The aspectual opposition between perfective and imperfective injunctions accounts for the aorist imperatives πώλησον and δός, on the one hand, and the present imperative ἀκολούθει, on the other. The actions of selling all possessions and giving them to poor people are aoristic since they are to be completed suddenly. Conversely, the action of following Jesus is coded by a present imperative, since this action is for life,

 8 For a more in-depth analysis of ὑπάγω in Ancient and Biblical Greek, also from a diachronic point of view, see Logozzo/Tronci 2022: §§ 3.2.1 and 5.

Motion and Posture Verbs in Multiverb Constructions  

without any endpoint. The present imperative ὕπαγε remains outside of this aspectual system of oppositions. It goes without saying that the present stem of ἀκολούθει is by no means comparable with the present stem of ὕπαγε. Evidence for this is given by the fact that the present imperative does not alternate with any aorist in the case of ὑπάγω; conversely, the present imperative of ἀκολουθέω in (26) alternates with the aorist imperative in (27). The comparison between (26) and (27) brings to light the aspectual difference: the action is punctual and includes an endpoint in (27), whilst it is required to be imperfective and endless in (26). (27) ὑπάγετε εἰς τὴν πόλιν, καὶ ἀπαντήσει ὑμῖν ἄνθρωπος κεράμιον ὕδατος βαστάζων· ἀκολουθήσατε αὐτῷ, καὶ ὅπου ἐὰν εἰσέλθῃ εἴπατε τῷ οἰκοδεσπότῃ (Mk 14.13–14). “Go into the city, and a man will meet you carrying a pitcher of water; follow him; and wherever he enters, say to the owner of the house”.

The inability of the present imperative ὕπαγε to alternate with the aorist can be explained by its function as a modifier verb of a PCC. Modifier verbs not only lose their syntactic property to govern complements but may also undergo attrition in their inflectional paradigms. This occurrence is not rare crosslinguistically. Ross (2016: 216) mentions the case of syndetic pseudocoordination in Frisian, where the modifier verb is exclusively inflected in the infinitive, whatever the inflection of the main verb. Ledgeway (2016) argues for attrition in person and number agreement, in his study on PCCs with the verbs for “stand” and “go” in the dialects of Apulia. He remarks that attrition in the inflectional paradigms “can be taken to represent a weakening in their defining verbal characteristics and, at the same time, to signal a concomitant change in their category from lexical verb (V) to functional predicate (Aux)”. Accordingly, Aikhenvald (2018: 218) correlates inflectional attrition with grammaticalization: “[t]he minor verb in asymmetrical serial verb constructions can gradually lose its verbal status and develop into an exponent of a grammatical category, undergoing grammaticalization”. Turning back to ὑπάγω, attrition is evident in the category of the aspect. However, even though ὑπάγω cannot be inflected for aspect, aspectual markers are exhibited by the main verb of the PCC. As already said, most PCCs with V1 ὑπάγω are aspectually aoristic against only 3 imperfective PCCs. Besides the occurrences in (24) and (25) above, where the construction occurs in a narrative context and the verbs are expected to match for tense-aspectual inflection, there is only one imperatival PCC:

  Felicia Logozzo and Liana Tronci (28) τέκνον, ὕπαγε σήμερον ἐργάζου ἐν τῷ ἀμπελῶνι (Mt 21.28). “Son, go work today in the vineyard”.

Concerning the distribution in the four Gospels, it is noteworthy that the Gospel of Luke is the only one not to host PCCs with ὑπάγω. However, the fact that the verb ὑπάγω is attested outside PCCs in Luke’s Gospel makes it clear that there are no lexical constraints and that the lack of PCCs with ὑπάγω is connected with the tendency of Luke to avoid PCCs. In the following pair of clauses, taken from Matthew’s and Luke’s Gospels respectively, the same scene is depicted: Jesus has cleansed someone of leprosy and tells him to show himself to the priest and present his offering. The same injunction is coded by a PCC in the Gospel of Matthew and by a PAC in that of Luke: (29) καὶ λέγει αὐτῷ ὁ Ἰησοῦς, ὅρα μηδενὶ εἴπῃς, ἀλλὰ ὕπαγε σεαυτὸν δεῖξον τῷ ἱερεῖ, καὶ προσένεγκον τὸ δῶρον ὃ προσέταξεν Μωϋσῆς, εἰς μαρτύριον αὐτοῖς (Mt 8.4). “And Jesus said to him, ‘See that you tell no one; but go, show yourself to the priest and present the offering that Moses commanded, as a testimony to them’”. (30) καὶ αὐτὸς παρήγγειλεν αὐτῷ μηδενὶ εἰπεῖν, ἀλλὰ ἀπελθὼν δεῖξον σεαυτὸν τῷ ἱερεῖ, καὶ προσένεγκε περὶ τοῦ καθαρισμοῦ σου καθὼς προσέταξεν Μωϋσῆς, εἰς μαρτύριον αὐτοῖς (Lk 5.14). “And He ordered him to tell no one, ‘But go and show yourself to the priest and make an offering for your cleansing, just as Moses commanded, as a testimony to them’”.

Further evidence comes from the comparison of (31) below, from Luke’s Gospel, with (26) above, from Matthew’s Gospel. In this case too, Matthew and Luke depict the same scene, but they make different syntactic choices. While Matthew uses the PCC ὕπαγε πώλησον “go sell”, Luke prefers the simple imperative πώλησον “sell”. (31) ἔτι ἕν σοι λείπει· πάντα ὅσα ἔχεις πώλησον καὶ διάδος πτωχοῖς, καὶ ἕξεις θησαυρὸν ἐν οὐρανοῖς, καὶ δεῦρο ἀκολούθει μοι (Lk 18.22). “One thing you still lack; sell all that you possess and distribute it to the poor, and you shall have treasure in heaven; and come, follow Me”.

As we have already remarked, PCCs are not abundant in the Gospel of Luke. We found only the following two instances, in both of which the verb of going πορεύομαι is combined with another verb inflected in the same mood, tense, person and number:

Motion and Posture Verbs in Multiverb Constructions  

(32) εἶπεν δὲ αὐτῷ ὁ Ἰησοῦς· πορεύου καὶ σὺ ποίει ὁμοίως (Lk 10.37). “Then Jesus said to him, ‘Go and do the same’”. (33) τότε πορεύεται καὶ παραλαμβάνει ἕτερα πνεύματα πονηρότερα ἑαυτοῦ ἑπτά (Lk 11.26). “Then it goes and takes along seven other spirits more evil than itself”.

Both instances are somehow problematic. The analysis of (32) as a PCC may be questioned because of the non-contiguity of the two verbs. The two imperatives “go and do the same” could be analysed as an asymmetric coordination. Nevertheless, we suggest analysing (32) as a PCC, since the two actions are conceived as a single event and the verb πορεύου, far from pointing to a departure from one place to another, appears, rather, to modify the action meant by the second verb, by adding temporal nuances. The moment of leaving (πορεύου) overlaps with the starting point for performing the main injunction ποίει ὁμοίως, given by Jesus to his interlocutor. We argue, therefore, that πορεύου points to a temporal qualification of the action conveyed by V2, so that the PCC could be rephrased as “from the moment you leave (= from now on), do the same”. Turning to (33), the same clause occurs in Mt 12.45, so probably there has been some influence between the two Gospels. We limit ourselves to this note, without looking further into this issue. Outside (32) and (33), the verbs of going only occur in PACs in the Gospel of Luke. As V1, πορεύομαι is more frequent than ἀπέρχομαι (9 occurrences against 3)9 and usually occurs as an aorist participle, excepting Lk 8.14, where it is inflected in the present, in line with the present tense of the main verb. The main verb can be an imperative or another mood and is inflected in the aorist, excepting Lk 8.14 (34). (34) τὸ δὲ εἰς τὰς ἀκάνθας πεσόν, οὗτοί εἰσίν οἱ ἀκούσαντες, καὶ ὑπὸ μεριμνῶν καὶ πλούτου καὶ ἡδονῶν τοῦ βίου πορευόμενοι συνπνίγονται καὶ οὐ τελεσφοροῦσιν (Lk 8.14). “The seed which fell among the thorns, these are the ones who have heard, and as they go on their way they are choked with worries and riches and pleasures of this life, and bring no fruit to maturity”.

The participle πορευόμενοι does not mean here any real motion; rather, it points to the metaphorical going on of life. This nuance is not common within the PACs with V1 πορεύομαι, since the participle usually points to a motion from the  9 Already Robertson 1919: 122 stressed that Luke “loves πορεύομαι”.

  Felicia Logozzo and Liana Tronci place where the speaker and the interlocutor are (but see the PCC in (32) above). The place towards which the interlocutor is asked to go and perform the action required is either known by the context, e.g. (35), or remains undetermined, e.g. (36). (35) ἐν αὐτῇ τῇ ὥρᾳ προσῆλθάν τινες Φαρισαῖοι λέγοντες αὐτῷ· ἔξελθε καὶ πορεύου ἐντεῦθεν, ὅτι Ἡρώδης θέλει σε ἀποκτεῖναι. καὶ εἶπεν αὐτοῖς· πορευθέντες εἴπατε τῇ ἀλώπεκι ταύτῃ (Lk 13.31–32). “Just at that time some Pharisees approached, saying to Him, ‘Go away, leave here, for Herod wants to kill You’. And He said to them, ‘Go and tell that fox’”. (36) καὶ ἀπέστειλεν Πέτρον καὶ Ἰωάννην εἰπών· πορευθέντες ἑτοιμάσατε ἡμῖν τὸ πάσχα ἵνα φάγωμεν (Lk 22.8). “And Jesus sent Peter and John, saying, ‘Go and prepare the Passover for us, so that we may eat it’”.

Besides Luke, Matthew makes frequent use of PACs with V1 πορεύομαι. Of the total 8 occurrences, 6 display the main verb in the imperative and 2 in the indicative. Unlike Luke’s Gospel, in Matthew’s Gospel PACs with V1 πορεύομαι appear to alternate with PCCs with V1 ὑπάγω, so that the analysis cannot neglect to take into account the distribution of the two types of construction. As already shown above, ὑπάγω provides the PCC with a nuance of immediacy and urgency for the action conveyed by the main verb. This is not the case for the PAC with V1 πορεύομαι, where the stress, rather, is put on the motion required to perform the injunction conveyed by V2. The difference clearly emerges in the following pair of clauses, where one PAC and one PCC with the same V2 ἀπαγγείλατε occur. (37) καὶ ἀποκριθεὶς ὁ Ἰησοῦς εἶπεν αὐτοῖς, πορευθέντες ἀπαγγείλατε Ἰωάννῃ ἃ ἀκούετε καὶ βλέπετε (Mt 11.4). “Jesus answered and said to them, ‘Go and report to John what you hear and see’”. (38) τότε λέγει αὐταῖς ὁ Ἰησοῦς· μὴ φοβεῖσθε· ὑπάγετε ἀπαγγείλατε τοῖς ἀδελφοῖς μου ἵνα ἀπέλθωσιν εἰς τὴν Γαλιλαίαν, καὶ ἐκεῖ με ὄψονται (Mt 28.10). “Then Jesus said to them, ‘Do not be afraid; go and take word to My brethren to leave for Galilee, and there they will see Me’”.

In (37) Jesus replies to a question that John the Baptist, who is in prison, asked him through his disciples. By the PAC πορευθέντες ἀπαγγείλατε, Jesus stresses

Motion and Posture Verbs in Multiverb Constructions  

the fact that John is not there and that the disciples must go to where he is in order to perform the injunction. In (38), conversely, the motion verb ὑπάγω points to the urgency to perform the request: Jesus is asking the women to announce to his disciples that he is resurrected and that they must go immediately to Galilee to see him. The two constructions are not in free variation: the first one (PAC with πορεύομαι) highlights spatial deixis, whilst the second one (PCC with ὑπάγω) correlates with temporal nuances of immediacy. Another relevant aspect concerning the distribution of the two constructions in Matthew’s Gospel emerges from the comparison between the following two instances, which roughly relate to the same event, but depict it in a very different way. (39) ἔφη αὐτοῖς ὁ Πειλᾶτος· ἔχετε κουστωδίαν· ὑπάγετε ἀσφαλίσασθε ὡς οἴδατε (Mt 27.65). “Pilate said to them, ‘You have a guard; go, make it as secure as you know how’”. (40) οἱ δὲ πορευθέντες ἠσφαλίσαντο τὸν τάφον, σφραγίσαντες τὸν λίθον μετὰ τῆς κουστωδίας (Mt 27.66). “And they went and made the grave secure, and along with the guard they set a seal on the stone”.

In the first occurrence, Pilate replies to the request of the Pharisees by two imperatival clauses, namely ἔχετε κουστωδίαν (a concession) and ὑπάγετε ἀσφαλίσασθε (an injunction). In the latter, the presence of ὑπάγω stresses not only the immediacy and the urgency of the action, but also Pilate’s wish that the Pharisees should leave quickly and not involve him in their affairs. We should not forget, incidentally, that the Pharisees had ordered Pilate, without mincing words, to give orders for the grave to be made secure until the third day (Mt 27.64); it comes as no surprise, therefore, that Pilate was annoyed by the Pharisees and answered them by the PCC in (39). Let us turn now to the instance in (40), which immediately follows (39) in Matthew’s Gospel. Here, Matthew is reporting that the Pharisees performed the action ordered by Pilate. Expectedly, his report displays a PAC with the verb πορεύομαι. The nuance of immediacy and urgency conveyed by ὑπάγω is not relevant here; rather, the participle πορευθέντες highlights that the Pharisees left Pilate’s residence and moved towards the grave to make it secure. Summarizing, the two MVCs ὑπάγετε ἀσφαλίσασθε and πορευθέντες ἠσφαλίσαντο basically relate to the same event and depict it through a combination of two verbs, the first of which is a motion verb and functions as a modifier of the second one. The two types of construction (PCC vs PAC) and the two motion

  Felicia Logozzo and Liana Tronci verbs (ὑπάγω vs πορεύομαι) qualify two different ways of “viewing” the event. On the one hand, the viewpoint is that of the speaker who gives an order to his addressee and makes use of the PCC instead of the simple imperative to mean that his request must be performed immediately and urgently (against the simple imperative which codes a more general order). On the other hand, the PAC codes a narrative report: the viewpoint is that of the narrator, who describes the events as they happened. The sequence of the motion verb plus the main verb points here to the real sequence of the actions (the Pharisees went and made the grave secure). Unlike PAC, the sequence of motion verb plus main verb in PCC is not merely a description of the actions, since the actions have not yet been performed. They only exist in the speech act of the speaker. In other words, the two constructions manifest the narrative levels of discourse and story, respectively. Let us turn now to the third verb of going that occurs as V1 in MVCs, ἀπέρχομαι. It is attested 17 times in PACs and 6 times in PCCs. The distribution of the two constructions in the Gospels is uneven: 8 PACs and 1 PCC in Matthew, 5 PACs and 2 PCCs in Mark, 3 PACs and no PCC in Luke, 1 PAC and 3 PCCs in John. PACs are typical of narrative contexts in Matthew and Mark; in Luke, conversely, 2 occurrences out of the total number of 3 have the main verb in the imperative. In reporting some similar events, Mark’s Gospel makes use of PACs with V1 ἀπέρχομαι, e.g. (41) and (43), whilst Luke’s Gospel unsurprisingly prefers PACs with V1 πορεύομαι, e.g. (42) and (44). (41) καὶ λέγουσιν αὐτῷ· ἀπελθόντες ἀγοράσωμεν δηναρίων διακοσίων ἄρτους, καὶ δώσομεν αὐτοῖς φαγεῖν; (Mk 6.37). “And they said to Him, ‘Shall we go and spend two hundred denarii on bread and give them something to eat?’”. (42) οἱ δὲ εἶπαν. οὐκ εἰσὶν ἡμῖν πλεῖον ἢ ἄρτοι πέντε καὶ ἰχθύες δύο, εἰ μήτι πορευθέντες ἡμεῖς ἀγοράσωμεν εἰς πάντα τὸν λαὸν τοῦτον βρώματα (Lk 9.13). “And they said, ‘We have no more than five loaves and two fish, unless perhaps we go and buy food for all these people’”. (43) λέγουσιν αὐτῷ οἱ μαθηταὶ αὐτοῦ· ποῦ θέλεις ἀπελθόντες ἑτοιμάσωμεν ἵνα φάγῃς τὸ πάσχα; (Mk 14.12). “His disciples said to Him, ‘Where do You want us to go and prepare for You to eat the Passover?’”.

Motion and Posture Verbs in Multiverb Constructions  

(44) καὶ ἀπέστειλεν Πέτρον καὶ Ἰωάννην εἰπών· πορευθέντες ἑτοιμάσατε ἡμῖν τὸ πάσχα ἵνα φάγωμεν (Lk 22.8). “And Jesus sent Peter and John, saying, ‘Go and prepare the Passover for us, so that we may eat it’”.

These examples show that πορεύομαι and ἀπέρχομαι are to some extent in a complementary distribution in the Gospels of Mark and Luke. Mark uses a PCC with πορεύομαι only once (Mk 16.10) and MVCs with ἀπέρχομαι seven times (5 PACs and 2 PCCs); Luke, conversely, uses πορεύομαι eleven times (9 PACs and 2 PCCs) and ἀπέρχομαι three times in PCCs. 3.4. Among the motion verbs that occur as V1 in MVCs, we also found ἐξέρχομαι “I go out”, προσέρχομαι “I go to, I approach”, and καταβαίνω “I go down”, which do not denote basic motion meanings but function as modifier verbs. The verb προσέρχομαι is particularly interesting in this respect, since it usually governs the dative of the person who is approached, e.g. (45); conversely, when the verb is part of an MVC, it may share the dative complement of V2, as in (46). In the latter case, the complement occurs, without exceptions, after the second verb. (45) τότε προσελθόντες οἱ μαθηταὶ τῷ Ἰησοῦ κατ’ ἰδίαν εἶπον (Mt 17.19). “Then the disciples came to Jesus privately and said”. (46) καὶ ἰδοὺ λεπρὸς προσελθὼν προσεκύνει αὐτῷ λέγων (Mt 8.2). “And a leper came and bowed down before Him, and said”.

The verbs ἐξέρχομαι and προσέρχομαι occur in both PACs and PCCs, with a clear predominance of PACs. We counted 13 PACs vs 3 PCCs with V1 ἐξέρχομαι and 9 PACs vs 1 PCC with V1 προσέρχομαι. The only PCC with προσέρχομαι shows the same argumental structure as (46). (47) καὶ ἰδοὺ ἄγγελοι προσῆλθον καὶ διηκόνουν αὐτῷ (Mt 4.11). “And behold, angels came and began to minister to Him”.

When it occurs as the V1 of PACs, the participle of ἐξέρχομαι is usually inflected in the aorist, as expected. One only exception was found in the Gospel of Luke, who is particularly careful to code aspectual differences in his reports.

  Felicia Logozzo and Liana Tronci (48) ἦν δὲ τὰς ἡμέρας ἐν τῷ ἱερῷ διδάσκων, τὰς δὲ νύκτας ἐξερχόμενος ηὐλίζετο εἰς τὸ ὄρος τὸ καλούμενον ἐλαιῶν (Lk 21.37). “Now during the day He was teaching in the temple, but at evening He would go out and spend the night on the mount that is called Olivet”.

The present participle in (48) codes imperfective (repetitive) aspectual meanings, as is also marked by the imperfective V2 ηὐλίζετο and the complement of time τὰς δὲ νύκτας. Two of the three PCCs with V1 ἐξέρχομαι exhibit a mismatch in the tenseaspectual inflection of V1 and V2. In both cases, the modifier verb ἐξέρχομαι occurs in the aorist and V2 in the present. (49) ἐν αὐτῇ τῇ ὥρᾳ προσῆλθάν τινες Φαρισαῖοι λέγοντες αὐτῷ· ἔξελθε καὶ πορεύου ἐντεῦθεν, ὅτι Ἡρώδης θέλει σε ἀποκτεῖναι (Lk 13.31). “Just at that time some Pharisees approached, saying to Him, ‘Go away, leave here, for Herod wants to kill You’”. (50) Ἰησοῦς οὖν εἰδὼς πάντα τὰ ἐρχόμενα ἐπ’ αὐτὸν ἐξῆλθεν καὶ λέγει αὐτοῖς (Jh 18.4). “So Jesus, knowing all the things that were coming upon Him, went forth and said to them”.

As we have already discussed, it is not surprising to find tense-aspectual mismatches in imperatival PCCs. In (49), for instance, the two imperatives point to an aoristic action (“go away”) followed by an imperfective one (“leave”). Luke is stressing here that Jesus must leave immediately (punctual action) and go away (imperfective action). Conversely, the tense-aspectual mismatch in (50) is more difficult to explain, since the clause is a narrative report and not an injunction. The opposition between aorist and present verbal stems codes the category of tense besides that of aspect. According to this, the two verbs ἐξῆλθεν and λέγει appear to point to two different moments along the timeline, i.e., the first one to the past, and the second one to the present. This apparent mismatch has been remarked on by scholars, who coined the label of “historic” present for the present tense that refers to the past. Robertson (1919: 134) explains this use by invoking the “vivid use of the tenses” that both John and Mark make. Porter (1989: 189–198) provides a detailed discussion of the so-called “historic” present, arguing that it is used “at those places where the author feels that he wishes to draw attention to an event or series of events. This includes the beginning of

Motion and Posture Verbs in Multiverb Constructions  

units of discourse, and thus it is used to highlight possibly the discourse unit itself but certainly the transition to the new unit”.10 This is just the case of our occurrence in (50), with respect to which we may add that the mismatch is only apparent, if we consider that the two verbs form an MVC and some inflectional features can be marked on only one verb and shared by the other one. 3.5. Let us now turn to MVCs with a verb of posture as V1. The most frequent verbs are ἀνίστημι and ἐγείρω, which are both translated by Latin surgo “I get up” (with few exceptions), even though they point to two different ways of getting up: ἀνίστημι usually relates to people who get up from a sitting position while ἐγείρω relates to people who get up from a lying position. The difference is evident in the following couple of clauses: (51) καὶ παράγων ὁ Ἰησοῦς ἐκεῖθεν εἶδεν ἄνθρωπον καθήμενον ἐπὶ τὸ τελώνιον, Μαθθαῖον λεγόμενον, καὶ λέγει αὐτῷ, ἀκολούθει μοι. καὶ ἀναστὰς ἠκολούθησεν αὐτῷ (Mt 9.9). “As Jesus went on from there, He saw a man called Matthew, sitting in the tax collector’s booth; and He said to him, ‘Follow Me!’ And he got up and followed Him”. (52) ἀναχωρησάντων δὲ αὐτῶν ἰδοὺ ἄγγελος κυρίου φαίνεται κατ’ ὄναρ τῷ Ἰωσὴφ λέγων, ἐγερθεὶς παράλαβε τὸ παιδίον καὶ τὴν μητέρα αὐτοῦ καὶ φεῦγε εἰς Αἴγυπτον (Mt 2.13). “Now when they had gone, behold, an angel of the Lord appeared to Joseph in a dream and said, ‘Get up! Take the Child and His mother and flee to Egypt’”.

It must be said, however, that the semantic difference between the two V1s turns out to be not relevant in certain cases, as is shown by (53) and (54). Here, two different V1s occur in Matthew’s and Luke’s Gospels in relationship to the same event. Note also that the PCC of Matthew corresponds to the PAC of Luke: (53) καὶ ἥψατο τῆς χειρὸς αὐτῆς, καὶ ἀφῆκεν αὐτὴν ὁ πυρετός· καὶ ἠγέρθη καὶ διηκόνει αὐτῷ (Mt 8.15). “He touched her hand, and the fever left her; and she got up and waited on Him”. (54) καὶ ἐπιστὰς ἐπάνω αὐτῆς ἐπετίμησεν τῷ πυρετῷ, καὶ ἀφῆκεν αὐτήν· παραχρῆμα δὲ ἀναστᾶσα διηκόνει αὐτοῖς (Lk 4.39). “And standing over her, He rebuked the fever, and it left her; and she immediately got up and waited on them”.

 10 Porter 1989: 196.

  Felicia Logozzo and Liana Tronci Moreover, the two participles ἀναστάς and ἐγερθείς alternate in PACs that host the same main verb in two successive clauses of Luke’s Gospel: (55) ἤδη ἡ θύρα κέκλεισται, καὶ τὰ παιδία μου μετ’ ἐμοῦ εἰς τὴν κοίτην εἰσίν· οὐ δύναμαι ἀναστὰς δοῦναί σοι (Lk 11.7). “The door has already been shut and my children and I are in bed; I cannot get up and give you anything”. (56) λέγω ὑμῖν, εἰ καὶ οὐ δώσει αὐτῷ ἀναστὰς διὰ τὸ εἶναι φίλον αὐτοῦ, διά γε τὴν ἀναιδίαν αὐτοῦ ἐγερθεὶς δώσει αὐτῷ ὅσων χρῄζει (Lk 11.8). “I tell you, even though he will not get up and give him anything because he is his friend, yet because of his persistence he will get up and give him as much as he needs”.

Concerning the distribution in the Gospels, we found a large number of PACs with ἀνίστημι in Mark’s and Luke’s Gospels (5 occurrences in Mark and 10 occurrences in Luke) and a preference for ἐγείρω in Matthew’s Gospel (7 PACs and 3 PCCs). As regards the relationship between the V1s and the types of MVCs, we remark that ἀνίστημι nearly always occurs in PACs (excepting the only PCC in Jh 11.31) and ἐγείρω is equally distributed in PACs (8 occurrences) and PCCs (8 occurrences). A detailed description of PCCs with ἐγείρω is provided in Logozzo/Tronci 2022, which we refer to. We would like to point out here only one particular aspect of these constructions, already mentioned in Section 3.3 on the subject of ὑπάγω. We observed that imperatival PCCs with ὑπάγω usually exhibit a tense-aspectual mismatch, since the V1 ὑπάγω is exclusively inflected in the present tense, whatever the inflection for tense and aspect of the V2 is. Likewise, ἐγείρω always occurs in the present tense in imperatival PCCs, even when the V2 is an aorist (e.g. Mk 2.9, 2.11, Lk 6.8, Jh 5.8). In addition to this, ἐγείρω shapes another inflectional peculiarity in imperatival PCCs. We would expect the middlepassive voice of ἐγείρω, as is the case when it occurs as a simple verb in intransitive constructions,11 e.g. (57), but it always occurs in the active voice in PCCs, e.g. (58), despite its intransitive meaning. (57) αὐτὸς δὲ κρατήσας τῆς χειρὸς αὐτῆς ἐφώνησεν λέγων· ἡ παῖς, ἐγείρου (Lk 8.54). “He, however, took her by the hand and called, saying, ‘Child, arise!’”  11 In three occurrences of Mark’s Gospel (Mk 3.3, 5.41, 10.49), the verb ἐγείρω is inflected in the active voice even outside PCCs. These forms can be analysed as a spread of the form of PCC to simple forms.

Motion and Posture Verbs in Multiverb Constructions  

(58) τί ἐστιν εὐκοπώτερον, εἰπεῖν τῷ παραλυτικῷ· ἀφίενταί σου αἱ ἁμαρτίαι, ἢ εἰπεῖν· ἔγειρε καὶ ἆρον τὸν κράβαττόν σου καὶ ὕπαγε; (Mk 2.9). “Which is easier, to say to the paralytic, ‘Your sins are forgiven’; or to say, ‘Get up, and pick up your pallet and walk’?”.

That this is a peculiarity of imperatival PCCs also emerges from the comparison with (a) non-imperatival PCCs, e.g. (59) and (b) PACs reporting the same scene, e.g. (60). In both of them, the verb ἐγείρω is inflected in the passive aorist, which is the expected inflection when the verb codes an intransitive meaning. (59) καὶ ἥψατο τῆς χειρὸς αὐτῆς, καὶ ἀφῆκεν αὐτὴν ὁ πυρετός· καὶ ἠγέρθη καὶ διηκόνει αὐτῷ (Mt 8.15). “He touched her hand, and the fever left her; and she got up and waited on Him”. (60) ἵνα δὲ εἰδῆτε ὅτι ἐξουσίαν ἔχει ὁ υἱὸς τοῦ ἀνθρώπου ἐπὶ τῆς γῆς ἀφιέναι ἁμαρτίας, τότε λέγει τῷ παραλυτικῷ, ἐγερθεὶς ἆρόν σου τὴν κλίνην καὶ ὕπαγε εἰς τὸν οἶκόν σου (Mt 9.6). “But so that you may know that the Son of Man has authority on earth to forgive sins, then He said to the paralytic, ‘Get up, pick up your bed and go home’”.

In our opinion, the general tendency of Hellenistic Greek towards intransitivization (Moulton/Turner 1963: 52) cannot account for the distribution of voice inflection in the verb ἐγείρω. We observed that, outside PCCs, the verb exhibits the expected active vs middle-passive markers for coding transitive vs intransitive constructions. The occurrence of the active form with an intransitive meaning is peculiar to PCCs. This lack of voice inflection matches the lack of tenseaspectual inflection, which we already pointed out in the verb ὑπάγω. The collapsing of both voice and tense-aspectual inflections into the unmarked active present can be related to the function played by ἐγείρω in PCCs. We refer to the discussion in Section 3.3. Attrition in inflectional features is systematically evident in only two verbs of our corpus, namely ἐγείρω and ὑπάγω. Robertson (1919: 855) already remarked that ὑπάγω and ἐγείρω are inflected in the present in e.g. ὕπαγε νίψαι (Jh 9.7) and ἔγειρε ἆρον (Jh 5.8) and considered this surprising because it goes against the trend of the NT and Hellenistic Greek in general to use the aorist imperative at the expense of the present.

  Felicia Logozzo and Liana Tronci

 Concluding remarks In this paper, we have investigated the distribution of PACs and PPCs in the Gospels. PACs and PPCs are multiverb constructions in which two verbs occur, namely a motion or a posture verb as V1 and the main verb of the clause, generally a verb of activity, as V2. The full verb of the clause is the second verb, since it governs the arguments. The first verb does not govern any complement and it functions as a modifier of the second verb. It shares the subject with the second verb and, if a complement of place occurs in the clause, it does not depend on the motion verb; rather it relates to the multiverb construction in its entirety, e.g., (5) ὕπαγε νίψαι εἰς τὴν κολυμβήθραν τοῦ Σιλωάμ (Jh 9.7) “go wash in the pool of Siloam”. These syntactic constraints make it clear that the two verbs point to single eventhood from a semantic point of view. The main difference between the two constructions concerns the syntactic arrangement of the verbs: the first verb is subordinate to the second one in PACs, whilst the two verbs are coordinated in PCCs. The coordination that occurs in PCCs is a particular type: the order of the two verbs is fixed, as in asymmetric coordination, and the first verb is not a full verb, but a modifier of the second one. Compared to the participial type, where the functional difference between the two verbs is overtly coded by the different mood, the coordinated type does not manifest any difference, in that the two verbs are in the same mood, person and number. The two types display a different distribution in many respects. Firstly, the coordinated type occurs in imperatival MVCs more than the participial type, which is preferred, conversely, in narrative contexts. Secondly, PACs and PCCs do not occur with the same verbal lexemes as V1: some V1s are typical of PCCs while others tend to occur in PACs. Thirdly, the distribution of PACs and PCCs in the four Gospels is not similar, as the following table summarizes.

Motion and Posture Verbs in Multiverb Constructions   Tab. 6: Distribution of PACs and PCCs in the four Gospels. V

PACs

PCCs

Total

Mt

Mk

Lk

Jh

Mt

Mk

Lk

Jh

ἀνίστημι



















ἀπέρχομαι



















ἐγείρω



















ἐξέρχομαι



















ἔρχομαι



















ἵστημι



















καταβαίνω



















πορεύομαι



















προσέρχομαι



















ὑπάγω





































Total

Matthew makes widespread use of both PACs and PCCs, but his preference is for PACs, except for the construction with V1 ὑπάγω, which only occurs in PCCs. Contrary to Matthew, John uses almost exclusively PCCs. This is not surprising, given the “Semitic” influence on John’s Gospel and the preference for coordination in both Hebrew and Aramaic. According to Turner (1963: 71), “Biblical Greek will often disguise the parataxis by making one of the verbs a participle, e.g., answering said, but John prefers the co-ordination (answered and said), avoiding some of the redundant participles appearing in Biblical Greek (e.g., coming, rising) and preferring they came and saw, he rose and went out. Turning to the Gospel of Mark, we may claim that our data confirm the shared opinion that parataxis is also a typical feature of Mark’s Gospel (Turner 1963: 34). The number of PCCs in Mark’s Gospel is actually more than half that of PACs (13 vs 21 occurrences); therefore, comparing the data from Mark with the synoptic Gospels of Matthew and Luke, it can be claimed that the latter two decidedly prefer PACs, with a ratio of 77% in Matthew and 82% in Luke against 61% in Mark. Before concluding this study, we would like to recall one aspect that also provides some insights for further steps in the research. As mentioned, both PCCs and PACs encode some deictic information at the verb phrase level. In both constructions, V1 is associated with some information about spatial deixis. The motion verbs provide information about the “place” where the main action of the clause is performed. In fact, information about spatial deixis is given with

  Felicia Logozzo and Liana Tronci respect to a reference point, which is the speaker when the MVC is used in direct speech (imperatival MVCs) or another reference point of the narrator himself, e.g., the place of the previous scene or the place of a relevant character, for instance Jesus. It is not surprising, then, that a great number of the MVCs at issue here occur in direct speech and in directive speech acts, i.e., in dialogical situations, with one speaker and one interlocutor. In his well-known paper on tense relationships in French verbs, Benveniste (1966 [1959]) distinguished two different and complementary tense systems. These two systems manifest two different levels of enunciation, i.e., the narrative story (histoire) and the discourse (discours). The story is the report of the events as they happened at a moment in time, without any involvement of the speaker in the report. In order to be reported as events that occurred, they must be placed in the past. Conversely, the discourse is every enunciation that involves a speaker and an interlocutor and the intention of the speaker to influence his interlocutor in some way. Because of this, the speaker organizes his speech according to the category of person.12 As suggested above, the opposition between story and discourse accounts for several aspects of the MVCs investigated here. First of all, it explains why PCCs are usually in the imperative and PACs occur in both types of construction. The opposition between story and discourse is not relevant for PACs, but it is relevant for PCCs, which tend to occur in discourse enunciation. In fact, it is in discourse enunciation that PCCs may occur in the syndetic and the asyndetic configurations. Asyndetic PCCs never occur in the few instances of narrative PCCs. A second and final aspect we would like to stress concerns the relationship between the category of person and the other two deictic categories of space and time. It is unquestionable that the motion verbs in imperatival MVCs point to spatial deixis. By giving his commands, the speaker also gives information on the “place” where the interlocutor must perform the ordered action. The interlocutor is requested to leave, to approach, to go out etc. What is interesting is the relationship between space deixis and time deixis, which emerges in some imperatival PCCs with V1 ὑπάγω. This verb, which is specialized in the Gospels as a marker of PCC, basically conveys spatial deixis (departure from the place where the speaker is). However, in some occurrences, temporal nuances of im 12 We translate roughly from Benveniste 1966 [1959]: 239 for story: “la présentation des faits survenus à un certain moment du temps, sans aucune intervention du locuteur dans le récit. Pour qu’ils puissent être enregistrés comme s’étant produits, ces faits doivent appartenir au passé” and from Benveniste 1966 [1959]: 242 for discourse: “toute énonciation supposant un locuteur et un auditeur, et chez le premier l’intention d’influencer l’autre en quelque manière”.

Motion and Posture Verbs in Multiverb Constructions  

mediacy and urgency combine with spatial deixis, as can be seen by comparing PCCs and simple imperatives with the same verb. This is a sporadic occurrence, as we have remarked, and seems to be a semantic effect of the verb combination.13 However, we leave it as an insight for further research, to investigate whether this is only a semantic effect without any consequence on the language system or, conversely, whether this heralds the onset of later changes in Greek language.

Bibliography Aikhenvald, A. (2010), Imperatives and Commands, Oxford. Aikhenvald, A. (2018), Serial Verbs, Oxford. Andrason, A. (2019), “Categorial gradience and fuzziness—The QWM Gram (serial verb construction) in Biblical Hebrew”, in: G. Kotzé/C.S. Locatell/J.A. Messarra (eds.), Ancient Texts and Modern Readers. Studies in Ancient Hebrew Linguistics and Bible Translation, Leiden/Boston, 100‒126. Benveniste, É. (1966 [1959]), “Les relations de temps dans le verbe français”, in: Problèmes de linguistique générale, 1, Paris, 237‒250. Biraud, M. (2010), Les interjections du théâtre grec antique: étude sémantique et pragmatique, Louvain-La-Neuve. Blass, F./Debrunner, A./Funk, R.W. (1961 [1896]), A Greek Grammar of the New Testament and Other Early Christian Literature, Chicago. Burney, C.F. (1922), The Aramaic Origin of the Fourth Gospel, Oxford. Burrows, M. (1930), “The original language of the Gospel of John”, Journal of Biblical Literature 49(2), 95‒139. Denizot, C. (2011), Donner des ordres en grec ancien, Mont-Saint-Aignan. Hock, H.H. (2002), “Vedic éta...stávāma: subordinate, coordinate, or what?”, in: M. Southern (ed.), Indo-European Perspectives, Journal of Indo-European Studies Monograph 43, 89–102. Hock, H.H. (2014), “Come and get it: the Indo-european background of the Vedic éta … stávāma construction”, in: S.W. Jamison/H. Craig Melchert/B. Vine (eds.), Proceedings of the 24th Annual UCLA Indo-European Conference, Bremen, 47–66. Kölligan, D. (2021), “Multiverb constructions in classical Armenian”, in: A. Bartolotta (ed.), Perspectives on Language and Linguistics. Essays in honour of Lucio Melazzo, Palermo, 235–260. Kühner, R./Gerth, B. (1976 [1898–1904]), Ausführliche Grammatik der griechischen Sprache. Zweiter Teil: Satzlehre, Hannover.

 13 Cf. Andrason 2019 for a similar occurrence in Biblical Hebrew and Aikhenvald 2010: 128–129 for a typological overview of the relationship between imperative and time reference.

  Felicia Logozzo and Liana Tronci Ledgeway, A. (2016), “From coordination to subordination: the grammaticalization of progressive and andative aspect in the dialects of Salento”, in: A. Cardoso/A.M. Martins/S. Pereira/C. Pinto/F. Pratas, Coordination and Subordination, Oxford, 171–199. Logozzo, F./Tronci, L. (2019), “Subordination vs coordination: la traduction latine des participes du verbe ἔρχοµαι dans les Évangiles”, Revue de linguistique latine du Centre Ernout – De Lingua Latina 18, 1‒35. Logozzo, F./Tronci, L. (2019 [2020]), “Nota sulle costruzioni “a participio pleonastico” in greco antico: il tipo ἐλθὼν κατῴκησεν”, Atti del Sodalizio Glottologico Milanese 14, 85‒108. Logozzo, F./Tronci, L. (2022), “Pseudo-coordination and serial verb constructions in Hellenistic Greek? Some insights from the New Testament and the Septuagint”, Journal of Greek Linguistics 22(1), 72–144. LSJ = Liddell, H.G./Scott, R./Jones, H.S. (1996 [1843]), A Greek-English Lexicon, Oxford. Muraoka, T. (2009), A Greek-English Lexicon of the Septuagint, Louvain/Paris/Walpole, MA. Orlandini, A./Poccetti, P. (2008), “Liens de coordination: une approche sémantique à travers les langues anciennes”, Revue de sémantique et pragmatique 24, 93–113. Pompei, A. (2006), “Participles as a non-prototypical word class”, in: E. Crespo/J. de la Villa/ A. Revuelta (eds.), Words Classes and Related Topics in Ancient Greek, Louvain-la-Neuve, 361‒388. Pompei, A. (2012), “Participio greco e converbi”, Archivio Glottologico Italiano 97(2), 160‒204. Pompei, A. (2015), “Participle constructions in Ancient Greek, cosubordination, and symmetry”, Inverbis 5(1), 143‒157. Porter, S.E. (1989), Verbal Aspect in the Greek of the New Testament, with Reference to Tense and Mood, New York. Robertson, A.T. (1919), A Grammar of the Greek New Testament in the Light of Historical Research, New York. Ross, D. (2016), “Between coordination and subordination: typological, structural and diachronic perspectives on pseudocoordination”, in: F. Pratas/S. Pereira/C. Pinto (eds.), Coordination and Subordination: Form and Meaning, Newcastle upon Tyne, 209–243. Turner, N. (1963), A Grammar of New Testament Greek - J.H. Moulton. Vol III. Syntax, London/ New York. van Den Hout, T. (2003), “Studies in the Hittite phraseological construction I: its syntactic and semantic properties”, in: G. Beckman/R. Beal/G. McMahon (eds.), Hittite Studies in Honor of Harry A. Hoffner Jr. on the Occasion of his 65th Birthday, Winona Lake, 177–203. van den Hout, T. (2010) “Studies in the Hittite phraseological construction II: its origin”, in: R. Lebrun/J. De Vos (eds.), Hethitica XVI: Studia Anatolica in memoriam Erich Neu dicata, Louvain-la-Neuve, 191–204. van Emde Boas, E./Rijksbaron, A./Huitink, L./de Bakker, M. (2019), The Cambridge Grammar of Classical Greek, Cambridge. Yates, Anthony (2014), “On the PIE ‘quasi-serial verb’ construction: origin and development”, in: S.W. Jamison/H. Craig Melchert/B. Vine (eds.), Proceedings of the 25th Annual UCLA Indo-European Conference, Bremen, 237–255.

Mario Lombardo and Giovanni Boffa

Contact and Interaction between Greeks and Messapians Abstract: After a brief update on the large Messapian epigraphic corpus, the first part of this paper offers a broad outline of the main phases in the history of the Messapians’ relations with the Greeks (primarily the Tarentines), from ca. 700 BCE onwards, an indispensable premise to a study of the reception and uses of alphabetic writing in the Messapic world. The second part deals with the presentation and discussion of a fascinating epigraphic document that has recently shed new light on this question. It is a partially preserved column drum, found in the area of the ancient Messapian site of Veretum (in the municipality of Patù, Lecce), in whose grooves an interesting archaic Greek abecedarium is engraved, together with some Messapian inscriptions and various boat profiles.

 The historical and cultural framework (Mario Lombardo) In a volume devoted to multilingualism and multiculturalism, it seems appropriate to offer a broad outline of the history of the relations between Greeks and Messapians, focusing mainly on the Messapic epigraphic documents and the alphabet they are written in; for these are important pieces of evidence for understanding the complex processes of cultural interaction that took place between the local population and the Greek colonists in southern Apulia, from the foundation of Tarentum in the late 8th cent. BCE to the period following the Roman conquest in the 3rd cent. BCE (Fig. 9). The Messapic epigraphic corpus, already quite substantial, was recently enhanced by a welcome development. We previously had more than 600 inscriptions, mostly of a funerary or votive nature1, found in almost all the Messapic settlements, albeit with considerable and interesting variations.2 The chronological distribution of these inscriptions covers the period from the sixth to the

 1 See MLM; Marchesini 1999; 2020; cf. Lombardo 2013b; 2016. 2 On this point, see especially Lombardo 2013b. https://doi.org/10.1515/9783110779684-010

  Mario Lombardo and Giovanni Boffa 2nd cent. BC, with a large number of them being datable to the 4th and especially the 3rd centuries BCE (Fig. 10).3 This already remarkable epigraphical corpus has recently been increased by the several hundred inscriptions, mostly of a votive nature, and some of them written in Greek and Latin, engraved on the walls of the so-called Grotta Poesia at Roca Vecchia (Figs. 11, 12, 13). These inscriptions were discovered by Cosimo Pagliara in the 1980s,4 but only a few of them have since been published. Recently they have become the subject of a large research project made possible by an agreement between the Department of Cultural Heritage of the University of Salento, the Department of Scienze dell’Antichità of the University “La Sapienza” at Rome, and several other partners, among which the local Archaeological Superintendency, the Alteritas Association, and the Municipality of Melendugno. Lastly, this epigraphic corpus was even further enhanced by the discovery of Messapic texts by Francesco D’Andria in the sanctuary area of Castro,5 as well as by a problematic, although very interesting fragment of a lithic column, which was uncovered in Patù (at the southernmost tip of the Salento peninsula). As Giovanni Boffa will explain in greater detail below, this latter fragment presents numerous grooves in which there are engraved images of boats, along with some archaic texts probably in the Messapic language, and a “complete” archaic alphabet, in all likelihood a Greek one (infra, Fig. 19). Let us therefore proceed to illustrate the most significant stages of interaction between Greeks and Messapians, with a focus on those aspects especially relevant to the present discussion. In a first stage the Greeks (probably Euboeans and Corinthians for the most part) began, from about 800 BCE, to frequent the Adriatic coast of Messapia and, towards the end of the 8th cent. BCE, the Spartans established the colony of Tarentum. Unlike what occurred in Etruscan territory, however, it is impossible to detect any signs of the transmission of alphabetic writing from the Greeks to the Messapic world in the 8th as well as in the 7th cent. BCE.6 By contrast, the period from the late 7th to the early 5th cent. BCE is characterized by lively and significant, albeit diversified, “Hellenizing” phenomena in the context of local communities. These Hellenizing phenomena are especially noticeable in the northern, “isthmic” part of Messapia, which was the “frontier

 3 See De Simone 1984; 2018; Marchesini 1999; 2020; Lombardo 2013b; 2016. 4 See Pagliara 1987. See also the paper by Marchesini in this volume. 5 See D’Andria-Lombardo 2009 and now D’Andria 2020. 6 See D’Andria 1988; 1991; Lombardo 1991; 2002.

Contact and Interaction between Greeks and Messapians  

area” in most direct contact with Tarentum, though they would later spread to other sub-regional areas.7 In addition to the adoption of alphabetic writing, to which we will return, these phenomena concerned building and “urban planning” practices,8 the production of ceramics, especially figured pottery,9 as well as funeral and cultic practices, which are particularly interesting here, because they constitute the main functional fields in which the epigraphic use of writing by the Messapians can be found (Figs. 10, 14, and 16).10 If we take a look at funeral practices, we can clearly see the “adaptive” nature of the Messapic borrowings: between the 7th and the 6th cent. BCE, and starting with the “isthmic” settlements (Oria, Mesagne, Muro Tenente, Valesio), the Messapians adopted from the Greeks the practice of inhuming the dead in earth-cut tombs along with vases and other offerings; but they adapted in various ways the Greek practice to the peculiarities of their own society and culture, for instance by locating the tombs within their settlements, by disposing of the dead in a crouched position, and by reusing the same tomb, evidently conceived as a family grave, for subsequent interments.11 It is also noteworthy that, while the Greeks of Tarentum seem to have used very little writing in connection with funerary practices, on the contrary the Messapians made the greatest use of it precisely in funerary contexts.12 Furthermore, especially from the 4th cent. BCE onwards (the period for which we know of the greatest number of tombs), the Messapians often placed funerary inscriptions inside the tombs themselves, a practice almost unattested in the Greek world (Fig. 14).13 The same phenomena of “adaptive borrowing” can be seen in the adoption of Greek alphabetic writing by the Messapians. This started probably around ca. 600 BCE and was a relatively rapid and straightforward process, resulting in several clever adaptations of the original Greek alphabet to the requirements of the Messapic language, such as the elimination of some unneeded signs (like

 7 Since I myself have dealt directly and repeatedly with these issues, I limit myself here to referring to other works of mine in which extensive additional bibliographical references can be found: Lombardo 1991; 2002; 2013b; 2014; 2016. See also D’Andria 1988. 8 See especially D’Andria 1991; 2005. 9 See D’Andria 1988. 10 See Marchesini 1999; 2020; Lombardo 2013b; 2014; 2016. 11 Lombardo 1994; 2014; 2016. See also Iacono 2007–2008; Giannotta 2016. 12 Lombardo 2013b; 2016. 13 On this point see especially Frisone 2013.

  Mario Lombardo and Giovanni Boffa hypsilon), and the addition of new graphemes required by Messapic phonology (like the trident-shaped grapheme seen in Fig. 15).14 It is worth noting that in the cultic sphere too (see Figs. 16 and 17), the Messapians appear to have borrowed Greek cults and Greek theonyms, such as Damatra and Aprodita. These phenomena are again first attested in the Isthmus area (in Oria, Mesagne, Valesio, Ceglie).15 But the Messapians developed also very peculiar usages, such as the adoption of hieronyms, especially feminine ones such as tabara, -as, attested in funeral inscriptions (Fig. 17),16 or those linked to the cult of Thaotor, which are documented especially in the Grotta Poesia at Roca Vecchia. The next phase in the relations between the Greeks and the Messapians, from the early 5th cent. BCE onwards, looks very different from the earlier ones. Whereas previously there had been largely peaceful interactions, this period saw a serious and longlasting conflict — once labeled barbaros polemos by Giuseppe Nenci17 — between the Tarentines and the Messapians, accompanied by attitudes, on the part of the latter, of seeming closure towards, and even rejection of, the Hellenizing elements coming from Tarentum. However, in the second half of the 5th century we also have notice of interesting phenomena of political contact between the Messapians and other Greeks, such as the Athenians.18 Finally, there is the Hellenistic phase (late 4th and 3rd cent. BCE), when, after a further period of conflict in the central decades of the 4th cent. BCE, we can see not only the development of strong interactions between the Messapians and Tarentum in the cultural sphere,19 but also their sharing of its anti-Roman political and strategic choices when the Spartan prince Cleonymos (303–302 BCE) and the Epirot king Pyrrhus (281–275 BCE) conducted their expeditions in Italy.20  14 See Lombardo 2014 and now especially De Simone 2018 and Marchesini 2020. 15 See Lombardo 2013b; 2016; Mastronuzzi 2018; Marchesini 2020. 16 See De Simone 1984; Lombardo 2013b; Marchesini 2020. 17 Nenci 1978. 18 See Lombardo 1991; 1992; 2002; 2014. See also D’Andria, Dell’Aglio, Klaohi Zis. 19 Among the many aspects of intense cultural interaction and among the various and rich testimonies that the archaeological documentation offers — handicraft productions (Apulian vases and Gnathia vases); extensive necropoleis, such as that at Manduria, along with richly decorated underground chamber tombs, such as the Hypogeum of the Caryatids at Vaste; city walls (at Manduria; Vaste; Muro Leccese etc.), large houses and sizeable residential buildings (as at Vaste) —, I limit myself here to referring to the basileion of Oria, which Strabo (6.3.1) mentions and of which a beautiful pebble mosaic fragment, dating to the late 4th cent. BCE, has recently come to light (Fig. 18), whose style and technique recall the mosaics of Pella: see Masiello et al. 2013; Giannotta 2015; D’Andria 2017. 20 See Lombardo 1991; 2014; 2016.

Contact and Interaction between Greeks and Messapians  

It is difficult to gauge how significantly the aforementioned “scenario” was altered by the Roman conquest of southern Apulia during the bellum Sallentinum of 267–266 BCE, which culminated in the conquest of Brindisi with its splendid port, bound to become, with the foundation of the “Latin colony” of Brundisium in 243 BCE, the main bridgehead for the Roman expansion towards Greece and the East. In any case, during the 3rd cent. BCE the Messapians not only continued, but even intensified their practice of producing epigraphic documents, mainly funeral, in their epichoric language — a practice that continued even in the 2nd cent. BCE, albeit in decreasing numbers, and only came to a close at the beginning of the 1st cent. BCE, not surprisingly at the time of the bellum sociale, which was to trigger the subsequent process of municipalization and Romanization.21

 The column from Patù: some provisional notes (Giovanni Boffa) The aim of this section of the work is to present a new, complex and fascinating epigraphic document. It poses many historical and cultural problems, which are only beginning to be understood properly, and therefore the present contribution must be considered the preliminary stage of a work in progress. The item in question consists of a small grooved column (Fig. 19), now kept in the “Sigismondo Castromediano” Museum of Lecce (Italy). It is made of local stone and was found in the dry stone wall of a farmhouse located in the municipality of Patù — which is to say, precisely in the area of the ancient Messapic site of Veretum, inland of southern Salento and near the sea.22 The column is slightly tapered; it is about 75 cm high and its bottom diameter is about 55 cm. It is not fully intact, showing instead many breaks and abrasions, some of them quite large (Fig. 20). Fourteen grooves (between 2.5 and 3 cm wide) are clearly identifiable, and we can assume the existence of five others, considering the width of the damaged areas. Many of them bear various engraved graphic elements, including a remarkable graffito representing two ships next to each other, whose sterns are almost in contact (Fig. 21a–b). In both of these ships we can recognize an attempt to reproduce the planking by means of a grid of lines, thicker in the ship on the left (whose design is partly ruined by a large abrasion of the surface) than in the one to the right; the stern of the left 21 See Lombardo 1991; 2016. 22 See Sammarco 2012.

  Mario Lombardo and Giovanni Boffa hand ship, moreover, seems to have a different and more curved shape than that of the vessel on the right, whose stern ends with a circular decorative element. In the drawing of the ship on the right we can also observe the presence of a rudder and a pointed bow. It is difficult to determine whether both vessels were carved by the same hand.23 Two letters, heta and epsilon, drawn side by side, occupy the space between the ships; the heta cuts into the stern of the vessel on the left. If we observe the object in its natural position, i.e., standing, we find a similar letter (inscription a) in the groove immediately on the left. This appears to be a “closed” heta, with three horizontal strokes (Fig. 22). Each of the four grooves found to the right of the one with the ships (still from the viewpoint of someone looking at the object in upright position) bears an inscription (Fig. 23a–b). All these inscriptions seem to be carved by different hands, judging by the shape and size of the letters and the type of engraving, which always proceeds from left to right. The first of these inscriptions is certainly an abecedarium, partly traced under the ship on the right (Fig. 24). We can recognize the twenty-six letters listed below, which vary in height from 0.8 to 1.5 cm; they are engraved with fairly sure strokes, regular spacing, and fairly good horizontal alignment, perhaps obtained with the help of one or two reference lines. In fact, it is possible to notice a faint line placed immediately below the letters, extending precisely along the bottom of the first eight letters, up to the heta. After this latter sign, the sequence continues slightly higher and it seems to be possible to perceive a second reference line. To be more precise, we find: 1) alpha, with the internal stroke tilted to the right and with a rounded right vertical stroke; 2) rounded beta, with the upper eyelet smaller than the lower one; 3) hooked gamma; 4) two-stroked curvilinear delta; 5) epsilon with the three oblique strokes and the vertical one extended downwards; 6) digamma with two oblique strokes; 7) zeta, with small horizontal strokes; 8) “closed” heta, with an oblique internal stroke; 9) theta with the internal “upright cross”; 10) single-stroke iota; 11) kappa, hardly distinguishable, with the oblique strokes attached to the vertical one in a single point;  23 The ships were studied by M. Di Bartolo (2003–2004: 109–117).

Contact and Interaction between Greeks and Messapians  

12) 13) 14) 15) 16) 17) 18) 19) 20) 21) 22) 23) 24) 25) 26)

lambda, with a hook at the top; four-stroked mu, scratched upside-down; retrograde nu; koppa; omikron; cross-shaped xi; hooked pi; rho with very narrow eyelet; “snake-shaped” sigma, with six short strokes; tau with a very short horizontal stroke; hypsilon, with a small oblique stroke; phi, with the vertical stroke extended downwards; phi, with the vertical stroke not extended downwards; “hourglass-shaped” psi; “trident-shaped” khi.

This sequence of signs is remarkable. The absence of omega suggests that there is no distinction between long and short vowels, so we can suppose that omikron indicates both the short and long variants of the vowel /o/ just as epsilon represents both the short and long variants of the vowel /e/; heta stands for the aspiration /h/; the cross-shaped letter is used to indicate the consonant cluster /ks/, as testified by its position in the seventeenth place; sigma renders the sibilant sound /s/; the last, trident-shaped letter is likely to stand for the aspirated guttural consonant /kh/, since our sequence, considering the place of discovery,24 should be Laconic-Tarentine,25 and as a result it should belong to the “red” type; finally, the hourglass-shaped sign should render the consonant cluster /ps/,26 which means that it is placed before chi, not after, as in the “standard” sequence of the Greek alphabet.27 (The sequence koppa, omikron, xi, pi is also wrong — xi, omikron, pi, koppa being the standard one.)

 24 On the relationship between Tarentum and Messapia with respect to writing, see MLM: 6; Marchesini 1999: 201–202; Marchesini 2009: 140–141; Lombardo 2014: 46–51; Matzinger 2019: 13–14. 25 See the alphabetic sequence defined by M. Guarducci (1967: 278), within which the koppa, not attested at the time, is missing. 26 This sign was attested in a Laconian (metropolitan or colonial) context only by IG V, 1, 828, an archaic and higly problematic inscription from Sparta (see Guarducci 1967: 279). 27 The fact that the hourglass sign occupies the penultimate place suggests that it was obtained by modifying the cross sign with the addition of two horizontal strokes.

  Mario Lombardo and Giovanni Boffa The internal stroke of alpha, the parallel strokes in epsilon and digamma, the six strokes of sigma, and the reduced size of the horizontal strokes in zeta and tau all deserve some attention. As a matter of fact, these features support a date for the abecedarium in the 6th cent. BCE. This chronology is also supported by a comparison with archaic epigraphic documents from the colonial Tarentine context — such as the inscription of Dexylos, on a stone funerary cippus, and the more famous stone slab from Torricella28 (see Figs. 25 and 26). The inscriptions on the second, third, and fourth groove after the one with the ships are very difficult to classify, in part because the reading of some letters is in doubt (the last sign in inscription c in particular), as is indicated in the transcription. Linguistically, they do not seem Greek, so they are probably Messapic, even if the use of the “hourglass-shaped” sign presents some difficulty, since it is not otherwise attested in Messapic epigraphy.29 I will return to this issue further on. The first inscription (b) is especially difficult to read because of the considerable abrasion suffered by the written surface and because of the presence of many unintentional scratches (Fig. 27). We can read: [- - -]𐅡𐅡a𐅡𐅡he𐌗𐌗a aiheFa 𐅡𐅡e ẹhi[- - -]. The letters, drawn with shallow lines and with a certain tendency to tilt to the left, occupy the entire writing surface and have a height of about 2 cm. Moreover, they are fairly regularly spaced. A space of about 3.4 cm separates the last two signs from the previous ones. One should note the alpha with a curved stroke; the internal oblique stroke of heta; epsilon with its three parallel oblique strokes and a vertical one protruding both above and below; and finally, the “hourglass-shaped” sign. In the third groove (Fig. 28) we can read the inscription [- - ]ọ𐅡𐅡amanaṣzaFoio󰃹󰃹 (inscription c). This inscription is engraved more deeply than the previous one, but in common with the previous one the letters show a certain tilt to the left. The signs have a height of between 0.7 cm (omikron) and 1.7 cm (iota). The beginning of this text is not preserved (this is also the case for the previous and the following texts, where the end too is missing); it stops shortly before the midpoint of the writing surface, at the extreme right of which it is possible to see three more letters, slightly larger and drawn by a different hand: digamma with oblique parallel strokes (h. 1.8 cm), iota (h. cm 2) and alpha (h. cm 2). Notable features include the alpha with curved vertical stroke and internal oblique stroke; the four-stroked mu; and (once again) the “hourglassshaped” sign. It is difficult to say if the last letter is the “simple” trident-shaped  28 Arena 1998: 24, 39–42; Ferrandini Troisi 2015: no. 120. 29 MLM: 7–18.

Contact and Interaction between Greeks and Messapians  

sign (which indicates a velar fricative sound) or the Messapic “trident-shaped” sign (which indicates a dental fricative sound), with a horizontal stroke.30 In both cases the position of the letter, at the end of the word, appears rather strange. Finally, in the fourth groove (Fig. 29) we find the sequence [- - -] e𐅡𐅡eḥenaFo𐌗𐌗aihe𐌗𐌗oi[- - -] (inscription d). As compared with the previous two texts, in this case the letters, between 1 cm (omikron) and 2 cm (the second heta) high, are much more deeply engraved and do not show the slight inclination to the left that is typical of the letters in the previous two inscriptions; however, alpha, heta, epsilon and digamma have the same shape. The simultaneous presence of the hourglass sign and the cross sign is remarkable. With the sole exception of the hourglass-shaped sign, the palaeographic character of these three inscriptions seems compatible with the first phase of development of the Messapic alphabet, which runs from the first half of the 6th cent. BCE to the first half of the following century.31 Moving to the right, in the fifth groove after the one with the ships, we find a thick grid that appears to have been drawn intentionally (Fig. 30). In the following (sixth) groove, it is possible to recognize an alpha (inscription e), with a curved stroke and an internal oblique stroke (Fig. 31). Another grid occupies the entire writing surface in the next (seventh) groove, in which, on the right, we can read the sequence he𐅡𐅡aaị (inscription f), whose letters are between 1.5 and 2.5 cm high. The first alpha has a “potbellied” shape and the second one is written upside down (Fig. 32). Next, in the eighth groove, there is a more faintly engraved grid which, on the right side, shows a tendency to curve in its long horizontal strokes; it might be an additional, and more elementary, drawing of a ship (Fig. 33). Other alphabetic signs (inscription g), faintly traced, are found in the ninth groove (Fig. 34), which also features horizontal lines intersected, on the left side, by occasional vertical lines. We can identify kappa, alpha, nu, perhaps followed by iota and, certainly, by phi; the letters are between 1 and 1.5 cm high. On the right-hand side, in a scratched and abraded area, alpha and kappa (between 2 and 1.7 cm high) are also legible. The following groove, the tenth one, has a sketchy grid, of a type also found further on (fourteenth and fifteenth grooves). The abecedarium is of particular interest, not only as far as the Messapic and Tarantine environments are concerned, but also with reference to the La 30 MLM: 27. 31 MLM: 7.

  Mario Lombardo and Giovanni Boffa conic-Spartan one. As a matter of fact, in relation to all three cultural contexts, our document presents certain significant aspects that render it a rare and precious specimen. As for Laconia, the only other evidence from Archaic times attesting an abecedarium (and a partial one at that) are the letters used by the authors of the bronze Vix Crater (third quarter of the 6th cent. BCE) as a reference point for mounting the decorative elements of the frieze on the neck of that monumental artifact. As is well known, these letters, located both on the neck of the vase and on the appliqués, run from alpha to kappa (drawn in retrograde direction but with left-to-right orientation); then they are interrupted by a series of numeral signs (vertical strokes ranging from one to seven); and finally they continue with a trident-shaped sign, a cross-shaped sign, tau, sigma (with five strokes), and koppa. After long and intense discussion about the manufacturing place of the crater, and, in parallel, about the cultural origins of the aforementioned letters, S. Verger, following Rolley, identified among the team that created the artifact the presence of two “model makers”, respectively of Laconic and of Corinthian origin, and of three different “assemblers”. In order to correctly mount the appliqués, these three “assemblers” employed letters from different epichoric Greek alphabets, among them possibly, but not certainly, the Laconic alphabet. Therefore, the evidence provided by the crater for the Laconic alphabet is remarkably problematic. To date, no abecedaria from ancient Taranto are known to us, and the archaic Tarantine alphabet is attested only by a small number of epigraphic documents. With regard to the Messapic area, a fragment of an imported lekane (now lost) from the Masseria Fani (Salve, Lecce), not far from Patù, bears five letters of an abecedarium, scratched on the object in left-to-right direction (Fig. 35). It is possible to identify beta with triangular eyelets, hooked gamma with a second very long stroke, rounded delta, and epsilon with tilted parallel strokes, exactly as in the digamma. This object can be dated to the second half of the 6 th cent. BCE.32 We should also mention the Vaste abecedarium (Fig. 36), dated to the 4th cent. BCE and unfortunately known to us only via a drawing made by L. Cepolla in 1805. The letters drawn by Cepolla are in many cases not easily recognizable, and they are separated by punctuation marks that sometimes isolate single elements, but at other times define groups of three or four signs:  32 MLM: 7.

Contact and Interaction between Greeks and Messapians  

α̣ .β.γδ.FFι.h.ι.κλ.μ / νο𐌗𐌗 .ϙ.ρhσ.τρ󰃤󰃤󰃤󰃤 is the sequence which can to be read as Greek, while a: b: gd: ẹvẓ: h: i: kl: mnoš: q: rs: tr .. is its Messapic transcription, in the modern extended Latin alphabet, as suggested by C. De Simone and S. Marchesini. Many letters are of uncertain reading (in particular the first, the fifth, the seventh, the eighteenth, and the last two signs); there are letters missing (theta, pi, hypsilon); and some signs (digamma, rho, heta, the two tridentshaped signs) are repeated, while others are displaced (in the sequence from the thirteenth sign to the eighteenth). These oddities are truly remarkable, and they have been explained in various ways that unfortunately cannot be discussed in the present contribution. Vis-à-vis Laconia, and especially Sparta, the presence of the hourglassshaped sign (see fn. 5) in the abecedarium of our new document is particularly remarkable. This, however, constitutes an intricate problem to which a specific study will be devoted on another occasion. For the moment, I would like to point out that this new document represents the first concrete evidence for the circulation of the Laconic-Tarantine alphabet in a Messapic environment — a fact which directly concerns the problem of the origin of the Messapic alphabet. In this regard, as is well known, two different hypotheses have been proposed: one sees the Tarentine alphabet as the unique model script, whereas the other assumes that additional Greek contributions also played a role — Euboean and Corinthian in particular. This issue will be addressed in a future study. Here I limit myself to suggesting that our alphabetical series is located “at the border”, so to speak, between the Greek and Messapic worlds; in other words, one might say that, in this new document, we find a not-yet-modified model series, of Laconic-Tarantine origin, being used to write Messapic.33 This seems particularly probable if one considers the actual texts that are written on the column, which attest to the use of the hourglass sign, which, as noted before, is not otherwise found in Messapic epigraphy (presumably because it was soon abandoned). The graphic and iconographic elements of the column of Patù pose numerous difficult problems, which concern the classification of the monument both within the specific Veretine context and within the more general Messapic environment. Indeed, the object represents a veritable “palimpsest” of graphic expressions — different in shape, size, and semantic content — which is almost unparalleled: in Greek epigraphy34 only a few columns with inscriptions in the

 33 For an overview of the Messapic language, see Matzinger 2019; Marchesini 2020. 34 See Guarducci 1967: 451.

  Mario Lombardo and Giovanni Boffa grooves are known to us, and very few are extant in the Messapic epigraphic corpus (pace Pagliara 1983). Generally speaking, cippi — sometimes inscribed and similar in size to our monument — are a recurrent feature in Messapic centres. In the case of Cavallino,35 one is found within a religious context, and at Leuca and Porto Cesareo they are found inside maritime sacred areas linked to mooring points.36 Significant finds have also been made at Muro Leccese and Nardò.37 The place in which the greatest number of quadrangular cippi have been discovered is probably Vaste. There, cippi have been found with and without inscriptions, as in the case of the agricultural complex of Fondo Melliche, where votive deposits containing libation vessels and animal remains have been found placed at the base of numerous cippi featuring onomastic formulas. Additionally, anepigraphic cippi have been found in the areas near the walls, put in place starting from the second half of the 4th cent. BCE.38 Three similar objects come from the sites Fondo Padulella, Fondo Giuliano and Fondo Lucernara. In the latter area, a limestone slab with the theonyms Venas and Artamis has also been found.39 A closer typological comparandum for our document comes from Muro Leccese, where archaeologists have found (Figs. 37–38) a small fluted pillar with 16 grooves, slightly tapered (h. 45 cm, diam. 22 cm), with seven inscriptions, engraved by different hands in different times, from the 6th to the 4th cent. BCE, in in both left-to-right and right-to-left arrangements: a) klaias [.] or?roisdazima [.?]; b) θθlla.sbalaidratis; c) 󰃹󰃹aa (4th cent. BCE?); d) hnthaša (mid-5th cent. BCE? On two contiguous grooves); e) bo (6th cent. BCE.?); f) bro.o.[? (4th cent. BCE?); g) pa.40 But the most interesting pieces of comparanda, for our purposes, are to be found in two similar inscribed columns, both from the archaic period, which also both come from Vereto, i.e., the same ancient Messapic area that is the origin of the new column we are currently examining. These are:

 35 D’Andria/Mastronuzzi, 2008: 229. It is a cippus (h. 135 cm), dated to the second half of the 5th cent. BCE, bearing an anthroponym in the genitive case. 36 D’Andria/Mastronuzzi, 2008: 231. 37 D’Andria/Mastronuzzi, 2008: 234. 38 D’Andria/Mastronuzzi, 2008: 231–233. For an interpretative overview about the cippi at Fondo Melliche (and Fondo Padulella) of Vaste see Lombardo 1994; 2013a. See also, about the archaeological features of these areas, Mastronuzzi/Melissano 2015. 39 Pagliara 1981: 207–215. 40 MLM 3 Mu, 337–339.

Contact and Interaction between Greeks and Messapians  

1. MLM 2 Ve 2 (Ta MARTA, IG 6339). Fragmentary ten-sided column (h 142 cm, maximum diam. 16 cm), found together with other fragments of non-inscribed columns41, bearing an archaic (6th–5th cent. BCE) retrograde inscription: ]rzaideias haiθraaθi vareti (Fig. 39). 2. MLM 19 Ve (= Santoro IM 27,121, pp. 417–421. Private collection, from the so-called Centopietre of Patù). Column with cylindrical base and 11 grooves, tapered (total h. 90 cm), bearing on two sides an archaic (6th–5th cent. BCE) retrograde inscription: a) ].ddaskalan; b) zaka (Fig. 40). However, even when compared with these (admittedly few) similar items, the column from Patù appears quite exceptional in the plurality and the variety of graphic expressions it exhibits. In order to understand the column and its writing, we must start by formulating some specific questions: 1) Are these graphic expressions to be read in a substantially synchronic or diachronic order (and if so, in what order?) 2) Are they to be read as a whole (i.e., as a set of coexisting and/or overlapping graphic interventions that are in some way related to each other), or instead as the result of distinct graphic interventions that differ from each other in terms of chronology, nature, authors, recipients, aims and meanings? 3) In the first case, how ought we to interpret the “document/palimpsest” as part of an original inscribed monument? And what kind of monument? 4) In the second case, how ought we to reconstruct and interpret the sequence of the graphic interventions, and how should we understand the use (or the different/subsequent uses) that was made of the column? Was it perhaps used as a functional support for graphic teaching and/or learning experiences? 5) Does the coexistence of drawings, inscriptions and single letters carved by different hands, possibly at different times, suggest an open and uncontrolled access to the column? And what connection, if any, will have existed between the alphabetic expressions and the “figurative” drawings? These new, (probably) Messapic texts invite, indeed they require, extensive further research. Under the auspices of the Convention for the Study, Valorisation and Promotion of the Grotta Poesia of Roca Vecchia, Melendugno (LE), this research is currently underway, and hopefully it will lead to decisive steps towards a better understanding of this fascinating monument.

 41 See Pagliara 1979: 77.

  Mario Lombardo and Giovanni Boffa

Figures Strabo, VI 3, 1 (C 277) Συνεχὴς δ’ ἐστὶν ἡ Ἰαπυγία· ταύτην δὲ καὶ Μεσσαπίαν καλοῦσιν οἱ Ἑλληνες, οἱ δ’ ἐπιχώριοι κατὰ μέρη τὸ μέν τι Σαλεντίνους καλοῦσι τὸ περὶ τὴν ἄκραν τὴν Ἰαπυγίαν, τὸ δὲ Καλαβρούς

Fig. 9: Strabo’s Messapia.

Contact and Interaction between Greeks and Messapians  

Fig. 10: Chronological and typological distribution of Messapian inscriptions (from Lombardo 2013b).

  Mario Lombardo and Giovanni Boffa

Fig. 11: Grotta Poesia di Rocavecchia (Melendugno, Lecce): view of the interior.

Contact and Interaction between Greeks and Messapians  

Fig. 12: Grotta Poesia di Rocavecchia (Melendugno, Lecce): detail of a portion of the wall with a Messapian inscription (MLM 23 Ro).

Fig. 13: Grotta Poesia di Rocavecchia (Melendugno, Lecce): cast of a portion of a wall with ‘palimpsest’ of inscriptions: Messapian (red boxes), Greek (blue box), Latin (yellow boxes).

  Mario Lombardo and Giovanni Boffa

Chronological distribution of the funeral inscriptions

Inscriptions inside the tombs: circa 

period

period

VI-V sec. a.C.

percentage %

number

percentage

VI-V sec. a.C.



%

IV sec. a.C.

 %

IV sec. a.C.



%

III (-II) sec. a.C.

 %

III (-II) sec. a.C.



%

Fig. 14: Percentage of funeral inscriptions on the total known inscriptions: about 230 out of about 640 (= about 38%) and their chronological distribution (from Lombardo 2013b).

Contact and Interaction between Greeks and Messapians  

Fig. 15: Tarentine and Messapian alphabets with values of graphemes (from Lombardo 2014).

  Mario Lombardo and Giovanni Boffa

Fig. 16: Messapia: cults and cult sites (from Lombardo 2013b).

Contact and Interaction between Greeks and Messapians  

about  religious references in Messapian inscriptions of which about  in funeral inscriptions (= %) Catergories

Number of references

Percentage in funeral inscriptions

Theonyms:

ca 

ca %

Divine epiclesis or appellatives:

ca 

%

Priestly figures:

More than 

%

) Main theonyms

) Priestly figures

Theonym

Number of references in inscriptions

Number of references in funeral inscriptions

Name

Zis





Tabaras, -a 

% funeral

Thaotor

+

 (?)

Prespolis

% funeral

Aprodita





Damatra



+

Number of references in inscriptions



Percentage of references in funeral inscriptions

Fig. 17: Names and terms referable to the religious-cultic sphere present in the Messapian inscriptions and their percentages in the funeral inscriptions (from Lombardo 2013b).

Fig. 18: Oria (Brindisi), Episcopio. Pebble mosaic (ca mq. 2.50). Last quarter of the fourth century B.C. (from Giannotta 2015).

of

  Mario Lombardo and Giovanni Boffa

Fig. 19: The column from Patù: grooves and inscriptions.

Contact and Interaction between Greeks and Messapians  

Fig. 20: The column from Patù: damaged sides.

  Mario Lombardo and Giovanni Boffa

Fig. 21: The column from Patù: the drawing of the ships.

Fig. 22: The column from Patù: detail of the letters near the ships.

Contact and Interaction between Greeks and Messapians  

Fig. 23a-b: The column from Patù: inscriptions within grooves. a) photo of the object; b) epoxy resin cast.

  Mario Lombardo and Giovanni Boffa

Fig. 24: The column from Patù: the greek abecedarium.

Fig. 25: Dexylos’ inscription from Lizzano (Taranto). Mid-sixth century B.C. (from Ferrandini Trosi 2015).

Contact and Interaction between Greeks and Messapians  

Fig. 26: Archaic inscription from Torricella (Taranto). First half of the sixth century B.C. (from Ferrandini Troisi 2015).

Fig. 27: The column from Patù: Messapic inscription.

Fig. 28: The column from Patù: Messapic inscription.

  Mario Lombardo and Giovanni Boffa

Fig. 29: The column from Patù: Messapic inscription.

Fig. 30: Thick grid intentionally drawn.

Fig. 31: The column from Patù: isolated letters.

Fig. 32: The column from Patù: grid with Messapic inscription.

Fig. 33: The column from Patù: drawing of a thick grid.

Fig. 34: The column from Patù: alphabetic signs.

Contact and Interaction between Greeks and Messapians  

Fig. 35: Salve (LE), Masseria Fani: partial abecedarium scratched on an imported lekane. Second half of the 6th cent. BCΕ.

Fig. 36: The Vaste abecedarium.

  Mario Lombardo and Giovanni Boffa

Fig. 37: Muro Leccese (Lecce): fluted pillar with Messapic inscriptions.

Contact and Interaction between Greeks and Messapians  

Fig. 38: Muro Leccese (Lecce): fluted pillar with Messapic inscriptions.

  Mario Lombardo and Giovanni Boffa

Fig. 39: Ancient Vereto (Lecce): Fragmentary ten-sided column with an archaic Messapic inscription.

Contact and Interaction between Greeks and Messapians  

Fig. 40: Ancient Vereto (Lecce): fluted column with two archaic Messapic inscriptions.

  Mario Lombardo and Giovanni Boffa

Bibliography Andreassi, G./Cocchiaro, A./Dell’Aglio, A. (eds.) (2013), Vetustis novitatem dare. Temi di antichità e archeologia in ricordo di Grazia Angela Maruggi, Taranto. Arena, R. (1998), Iscrizioni greche arcaiche di Sicilia e Magna Grecia, V. Iscrizioni di Taranto, Locri Epizefiri, Velia e Siracusa, Alessandria. Bloch, R./Joffroy, E. (1953), “L’Alphabet du cratère de Vix”, RPh 27, 175–191. D’Andria, F. (1988), “Messapi e Peuceti”, in: AA.VV., Italia omnium terrarum alumna. La civiltà dei Veneti, Reti, Liguri, Celti, Piceni, Umbri, Latini, Campani e Iapigi, Milano, 653–715. D’Andria, F. (1991), “Insediamenti e territorio: l’età storica”, in: I Messapi. Atti del trentesimo Convegno di Studi sulla Magna Grecia, Taranto-Lecce, 4-9 ottobre 1990, Taranto, 393–478. D’Andria, F./Dell’Aglio, A. (eds.) (2002), Klaohi Zis. Il culto di Zeus a Ugento, Cavallino. D’Andria, F. (eds.) (2005), Cavallino: pietre, case e città della Messapia, Ceglie Messapica. D’Andria, F. (2017), “Ipotesi sul basileion di Oria”, in: L. Cicala/B. Ferrara (eds.), Kithon Lydios. Studi di storia e archeologia con Giovanna Greco, Napoli, 743–755. D’Andria, F. (2020), “L’Athenaion di Castro in Messapia”, MDAIR 126, 79–140. D’Andria, F./Lombardo, M. (2009), “Due nuove iscrizioni messapiche da Castro”, in: F. D’Andria, (ed.), Castrum Minervae, Galatina, 67–78. D’Andria, F./Mastronuzzi, G. (2008), “Cippi e stele nei santuari magno-greci”, in: G. Greco/ B. Ferrara (eds.), Doni agli dei. Il sistema dei doni votivi nei santuari. Atti del Seminario di Studi, Napoli, 21 aprile 2006, Napoli, 223–240. De Simone, C. (1964), “Die messapischen Inschriften und ihre Chronologie”, in: H. Krahe (ed.), Die Sprache der Illyrier, II, Wiesbaden, 7–151, 214–361. De Simone, C. (1982), “Su Tabaras (femm. -A) e la diffusione di culti misteriosofici nella Messapia”, SE 50, 177–197. De Simone, C. (2018), “Messapic”, in: J. Klein/B. Joseph/M. Fritz (eds.), Handbook of Comparative and Historical Indo-European Linguistics, Berlin/Boston, 1839–1850. Di Bartolo, M., La costruzione navale in età arcaica, Graduation Thesis, Università del Salento, A.A. 2003–2004. Ferrandini Troisi, F. (2015), Iscrizioni greche d’Italia. Puglia, Roma. Frisone, Flavia (2013), “Le iscrizioni all’interno di tombe messapiche: una riflessione”, in: G. Andreassi/A. Cocchiaro/A. Dell’Aglio (eds.), Vetustis novitatem dare. Temi di antichità e archeologia in ricordo di Grazia Angela Maruggi, Taranto, 279–286. Giannotta, M.T. (2015), “Le βασίλειον d’Oria? Découvertes récentes sur l’acropole du centre messapien”, in: E. Lévy (ed.), La Royauté dans la Grèce Antique (Ktema, 40), Strasbourg, 211–218. Giannotta, M.T. (2016), “Aspetti del rituale funerario in Messapia”, in: M.T. Giannotta/F. Gabellone/F.M. Stifani/L. Donateo, Soleto ritrovata. Ricerche archeologiche e linguaggi digitali per la fruizione, Lecce, 27–40. Guarducci, M. (1967), Epigrafia Greca, I, Roma. Iacono, F. (2010), “Burial & society in non-Greek Salento (Southeast Italy) 600–250 BC”, ARP 11, 95–11. Jeffery, L. (19902), The Local Scripts of Archaic Greece. A Study of the Origin of the Greek Alphabet and Its Development from the Eighth to the Fifth Centuries B.C., with a supplement by A.W. Johnston, Oxford.

Contact and Interaction between Greeks and Messapians  

Lejeune, M. (1983), “Rencontres de l’alphabet grec avec les langues barbares au cours du Ier millénaire av. J.C.”, in: Modes de contacts et processus de transformation dans les sociétés anciennes. Actes du colloque de Cortone, Rome, 24-30 mai 1981, Roma, 731–752. Lombardo, M. (1991), “I Messapi: aspetti della problematica storica”, in: I Messapi. Atti del trentesimo Convegno di Studi sulla Magna Grecia, Taranto-Lecce, 4-9 ottobre 1990, Taranto, 35–109. Lombardo, M. (1992), “Greci e Messapi nel V sec. a.C.: fonti, eventi e problemi storici”, in: Aspetti della storia del Salento nell’antichità. Atti del Convegno Nazionale AICC, Lecce, novembre 1989, Cavallino, 76–109. Lombardo, M. (1994), “Tombe, necropoli e riti funerari in ‘Messapia’: evidenze e problemi”, StAnt 7, 25–45. Lombardo, M. (2002), “ΠΗΜΑ ΙΑΠΥΓΕΣΣΙ: rapporti con gli Iapigi e aspetti dell’identità di Taranto”, in: Taranto e il Mediterraneo. Atti del quarantunesimo Convegno di Studi sulla Magna Grecia, Taranto, 12-16 ottobre 2001, Taranto, 254–290. Lombardo, M. (2013a), “Cippi, iscrizioni e contesti: i rinvenimenti di Fondo Padulella nel quadro della documentazione di Vaste”, in: G. Andreassi/A. Cocchiaro/A. Dell’Aglio (eds.), Vetustis novitatem dare. Temi di antichità e archeologia in ricordo di Grazia Angela Maruggi, Taranto, 225–235. Lombardo, M. (2013b), “Tombe, iscrizioni, sacerdoti e culti nei centri messapici: aspetti peculiari tra sincronia e diacronia”, in: L. Giardino/G. Tagliamonte (eds.), Archeologia dei luoghi e delle pratiche di culto. Atti del Convegno, Cavallino 26-27 gennaio 2012, Bari, 155–164. Lombardo, M. (2014), “Iapygians: the indigenous population of ancient Apulia in the fifth and fourth centuries B.C.E.”, in: T.H. Carpenter/K.M. Lynch/E.G.D. Robinson (eds.), The Italic People of Ancient Apulia. New Evidence from Pottery for Workshops, Markets, and Customs, Cambridge, 36–68. Lombardo, M. (2016), “I Messapi: origini, vicende e civiltà”, in: M.T. Giannotta/F. Gabellone/ F.M. Stifani/L. Donateo, Soleto ritrovata. Ricerche archeologiche e linguaggi digitali per la fruizione, Lecce, 9–26. Marchesini, S. (1999), “Confini e frontiera nella grecità d’Occidente: la situazione alfabetica”, in: Confini e frontiera nella grecità d’occidente. Atti del trentasettesimo Convegno di Studi sulla Magna Grecia, Taranto, 3-6 ottobre 1997, Taranto, 173–212. Marchesini, S. (2009), Lingue frammentarie dell’Italia antica. Manuale per lo studio delle lingue preromane, Milano. Marchesini, S. (2020), “Messapico / Messapian”, Palaeohispanica, 20, 495–530. Masiello, L./Giannotta, M.T./Cocchiaro, A./Lombardo, M./Quarta, G. (2013), “Recenti scoperte sull’acropoli di Oria: un mosaico policromo a ciottoli”, in: G. Andreassi/A. Cocchiaro/ A. Dell’Aglio (eds.), Vetustis novitatem dare. Temi di antichità e archeologia in ricordo di Grazia Angela Maruggi, Taranto, 355–364. Mastronuzzi, G. (2017), “Lo spazio del sacro nella Messapia (Puglia meridionale, Italia)”, in: S. Bourdin/A. Pagliara (eds.), Magno e Latio totaque Ausonia: etnografia virgiliana e Italia augustea, Mélanges de l’École française de Rome - Antiquité [Online], 129/1 (Online since 26 September 2017, connection on 31 March 2021. URL: http://journals.openedition.org/ mefra/4236; DOI: https://doi.org/10.4000/mefra.4236). Mastronuzzi, G./Melissano, V. (2015), Museo Archeologico di Vaste. Catalogo, Poggiardo. Matzinger, J. (2019), Messapisch. Kurzgrammatiken Indogermanischer Sprachen und Sprachstufen, Band 2, Wiesbaden. MLM = De Simone, C./Marchesini, S. (2002), Monumenta Linguae Messapicae, I, Wiesbaden.

  Mario Lombardo and Giovanni Boffa Nenci, G. (1976), “Il ΒΑΡΒΑΡΟΣ ΠΟΛΕΜΟΣ fra Taranto e gli Iapigi e gli ΑΝΑΘΗΜΑΤΑ tarentini a Delfi”, ANSP S. III, 6, 3, 719–738. Pagliara, C. (1979), “Materiali iscritti arcaici del Salento”, in: Salento arcaico. Atti del Colloquio internazionale, Lecce, 5-8 aprile 1979, Galatina, 57–91. Pagliara, C. (1981), “Materiali epigrafici da Vaste e Muro”, StAnt 2, 207–220. Pagliara, C. (1983), “Materiali iscritti arcaici del Salento, II”, ASNP S. III, 13, 1, 21–89. Pagliara, C. (1987), “La Grotta Poesia di Roca (Melendugno – Lecce). Note preliminari”, ANSP S. III, 17, 2, 267–328. Prosdocimi, A. (1989), La trasmissione della scrittura in Etruria e nell’Italia antica: insegnamento e oralità tra maestri e allievi, in: G. Maetzeke (ed.), Atti del secondo Congresso Internazionale etrusco, Firenze, 30 maggio – 2 giugno 1985, III, Roma. Rolley, C. (1982), Les vases de bronze de l’archaïsme récent en Grande Grèce, Naples. Rolley, C. (1998), “Ateliers sans vestiges: l’atelier du cratère de Vix”, Topoi 8, 2, 711–718. Rolley, C. (2003), éd. La tombe princière de Vix, Paris. Rumpf, A. (1957), “Krater Lakonikos”, in: K. Schauenburg/E. Langlotz (eds.), Charites. Studien zur Altertumswissenschaft: Festschrift für Ernst Langlotz zum 60. Geburtstag, 1955, Bonn, 127–135. Sammarco, M. (2012), “Vereto”, in: M.I. Gulletta/C. Cassanelli (eds.), Bibliografia Topografica della colonizzazione greca in Italia e nelle isole tirreniche, XXI, Pisa/Roma/Napoli, 817–825. Santoro, C. (1984), Nuovi studi messapici. Primo supplemento, Galatina. Verger, S. (2020), “L’équipe qui fabriqua le cratère de Vix: habitudes scripturales, traditions stylistiques, organisation et localisation de l’atelier”, RA 70, 2, 263–324. Vokotopoulou, I. (1972), Χαλκάϊ κορινθιουργείς πρόχοι, Athens. Whatsmough, J. (1925), “The alphabet of Vaste”, CQ 19, 2, 68–70. Woodhead, A.G. (1955), “Recensione di R. Bloch / R. Joffroy, L’Alphabet du cratère de Vix. (Reprinted from Revue de Philologie, xxvii, 1953) Pp. 17; 2 plates, 3 figs. Paris: Klincksieck, 1953. Paper, 200 fr.”, CR 5, 2.

Marco Mancini

The Etymology and Semantics of Oscan pukam Abstract: In an Oscan fragmentary dedication (ca. 2nd cent. BCE) […] hanuseís ・ pukam ・ prúffed, found in the area of the domus publica of Pietrabbondante, the hapax pukam is a remarkable addition to the Oscan lexicon. According to La Regina, Osc. pukā- means a kind of visible monument and stems from the Indo-European root *kwVḱ- “to see” but also “to appear”. This new word, which is thoroughly investigated in this paper from an etymological viewpoint, significantly increases the attested outcomes of a root that until now seemed productive only in Greek (τέκμωρ/τέκμαρ) and Indo-Iranian (Old Ind. cakṣ- “shine; appear; see”, cakṣate “appears”, cakṣas- “eye, gaze”, Old Pers. cašma, Avestan čašman- “eye”, Man. Parth. āgas “visible”, Middle-Pers. ākāh “informed, aware” etc.). On the basis of the productive Proto-Iranian morpheme *-kāśa(> Old Persian *-kāθa- “be visible”), there is no doubt that Oscan pukam corresponds to a /ˈpukam/ with /u/ as an outcome of inherited *-ō-. Besides, the Oscan new entry allows us, albeit indirectly, to reconsider the Latin semantic field of terms such as statua/signum/effigies/simulacrum. Above all, it opens up the possibility of defining an analogous semantic subfield in Oscan, which was shared by both the well-attested segúnúm and the hapax pukā-.

 Introductory remarks In the “Rivista di Epigrafia Italica” contained in the seventy-fifth issue of “Studi Etruschi” Adriano La Regina commented upon an unpublished Oscan inscription carved on a stone found in the area of the domus publica of Pietrabbondante. This inscription consists of two fragments, which originally formed part a donarium; the donarium itself was later reused as building material in a wall during the Imperial era (La Regina 2012: 322–325).

 I would like to thank my colleagues Maria Carmela Benvenuto and Federico Cinquepalmi for supporting me in the translation and Luca Lorenzetti and Paolo Milizia for their fruitful discussions. Above all, I am indebted to my friend Albio Cesare Cassio for his invaluable linguistic and scientific review of this article. Of course, the responsibility for any errors that remain is mine alone. https://doi.org/10.1515/9783110779684-011

  Marco Mancini All the details of the discovery can be found in the editor’s meticulous epigraphic description. In this paper, I will focus on one of the two fragments, in which the dedication to a deity similar to Latin Honos, i.e., Oscan *hanús, is clearly attested according to La Regina’s interpretation.1 Oscan *hanús is immediately comparable to Paelignian hanustu, a fem. sg. form attested in the socalled inscription of Herentas (Pg9 Rix = Corfinium 6, Imagines Italicae, Crawford). This latter form, as stated by Vittore Pisani (Pisani 1959: 141) among others, can be regarded as morphologically parallel to Lat. honesta, even though the vocalism is not fully clear and, above all, there is no evidence of any etymological connection outside the Italic area.2

 1 The Italic inscriptions are quoted according to the two Corpora of Rix 2002 and Crawford 2011; the Latin ones according to CIL and ILLRP. All translations of Homer are quoted from the Loeb edition by A.T. Murray; Pindar’s Odes are cited according to Diane Arnson Svarlien’s translation. 2 Regarding the etymology of Lat. honos, DELL (299b) is tersely pessimistic, as is usual for Ernout and Meillet in uncertain cases like this: “pas d’étymologie”; for LEW (656) any etymological connection remains uncertain; De Vaan correctly points out that “Latin hon- can hardly reflect anything but *ǵhon- or *ghon-, but no further etymology is known” (EDLIL 288). The parallelism between Paelignian hanustu ( puer “child”. This series also includes, of course, South Picene puqloh “son”, where the meaning is unambguous (Marinetti 1985: 151–152). It is clear from the above that none of these words has any connection with the new Oscan word pukam “representation”.

Bibliography Abaev, V.I. (1958), Istoriko-étimologičeskij slovar’ osetinskogo jazyka, tom I, Moskva/Leningrad. Adams, D.Q. (2013), A Dictionary of Tocharian B, rev. and greatly enlarged, Amsterdam/New York. Adams, J.N. (2003), Bilingualism and the Latin Language, Cambridge. AE = Année Épigraphique, 1889–. AiWb = Bartholomae, C., Altiranisches Wörterbuch, zusammen mit den Nacharbeiten und Vorarbeiten [1904–1906], Berlin/New York, 1979. Antonini, R. (1989), “Materiali per un’indagine preliminare sull’alfabeto osco. Documenti, problemi, prospetttive concernenti la riforma”, Quaderni dell’Istituto di Glottologia dell’Univ. di Urbino 6, 33–91. Back, M. (1978), Die sassanidischen Staatsinschriften. Studien zur Orthographie und Phonologie des Mittelpersischen der Inschriften zusammen mit einem etymologischen Index des mittelpersischen Wortgutes und einem Textcorpus der behandelten Inschriften, Leiden/ Téhéran/Liège.

  Marco Mancini Bakkum, G.C.L.M. (2009), The Latin Dialect of the Ager Faliscus: 150 Years of Scholarship, Part 1, Amsterdam. Bartholomae, C. (1901), “Arica XIV”, IF 12, 92–150. Bechtel, F. (1914), Lexilogus zu Homer. Etymologie und Stammbildung homerischer Wörter, Niemeyer, Halle an der Saale. Belardi, W. (1961), “Axš-aina, axša-ina o a-xšaina-?”, AION-L 3, 1–39. Belardi, W. (1999), “Forma, semantica ed etimo dei termini greci per “segno”, “indizio”, “simbolo” e “sintomo””, in: M.L. Bianchi (ed.), Signum – IX Colloquio Internazionale del Lessico Intellettuale Europeo, Firenze, 1–22. Benedetti, M. (2003), “L’apparizione dell’Aurora e l’occhio del linguista: a proposito della radice *derḱ-”, SSL 40–41, 31–44. Benzing, J. (1982), Chwaresmischer Wortindex. Mit einer Einleitung von Helmut Humbach, Wiesbaden. Bombi, R. (2009), “Neologia e formazioni produttive. Tra lingue speciali e lingua comune”, in: R. Bombi, La linguistica del contatto. Tipologie di anglicismi nell’italiano contemporaneo e riflessi metalinguistici, II ediz. riveduta e aggiornata, Roma, 378–387. Boyce, M. (1977), A Word-List of Manichaean Middle Persian and Parthian, Téhéran/Liège. Brachet, J.-P. (1994), “Esquisse d’une histoire de lat. signum”, Revue de Littérature et d’Histoire Anciennes 68, 33–50. Brachet, J.-P. (1999), “Réflexions sur l’évolution sémantique de significare”, in: Conceptions latines du sens et de la signification, Recherches linguistiques du Centre Alfred Ernout, Paris, 29–39. Brandenstein, W./Mayrhofer, M. (1964), Handbuch des Altpersischen, Wiesbaden. Brust, M. (2018) Historische Laut- und Formenlehre des Altpersischen. Mit einem etymologischen Glossar, Innsbruck. Bugge, S. (1878), Altitalische Studien, Christiania. Buonocore, M./de Simone, C. (2014), “Ex Samio efuit in una nuova dedica ad Ercole da Alba Fucens”, in: E. Benelli (ed.), Per Mariastella Pandolfini cên zic ziχuχe, Pisa/Roma, 79–98. Camodeca, G. (2013), “Osco e latino nelle istituzioni pubbliche di Cuma nel II-I secolo a.C.”, in: C. Cascione/C.M. Doria/G.D. Merola (eds.), Modelli di un multiculturalismo giuridico. Il bilinguismo nel mondo antico. Diritto, prassi, insegnamento, I, 1–26. Campanile, E. (1985), Questioni metodologiche nell’analisi dei testi oschi, in: E. Campanile (ed.), Lingua e cultura degli Oschi, Pisa, Giardini, 11–20. Campanile, E. (2008)[1995], “Per l’interpretazione di osco prúffed”, ora in Id., Latina & Italica. Scritti minori sulle lingue dell’Italia antica, a cura di Paolo Poccetti, Pisa/Roma, vol. II, 793– 797. Cannizzaro, F.A. (1990), Vendidad. La Legge di abiura dei demoni dell’AVESTA zoroastriano [1916], con un saggio sulla Storia dello Zoroastrismo di Antonio Panaino, Milano. Cappelletti, L. (2011), “Le magistrature italiche. Problemi e prospettive”, Index. Quaderni camerti di studi romanistici 39, 323–338. Ciancaglini, C. (2008), Iranian Loanwords in Syriac, Wiesbaden. Ciancaglini, C. (2015), “Allomorphic variability in the Middle Persian continuants of the Old Iranian suffix -ka-”, in: U. Bläsing/V. Arakelov/U. Weinreich (eds.), Studies on Iran and the Caucasus, Festschrift presented to Prof. Dr. Garnik S. Asatrian on the Occasion of his 60th Birthday, Leiden/Boston, 291–308.

The Etymology and Semantics of Oscan pukam  

Ciancaglini, C. (2020), Middle Persian bāzūg and bāzāy, in: M. Ashtiany/M. Maggi (eds.), A Turquoise Coronet. Studies in Persian Language and Literature in Honour of P. Orsatti, Wiesbaden, 97–109. CIL = Corpus Inscriptionum Latinarum, 1863–, Berolini, apud G. Reimerum. Crawford, M.H. (2011), Imagines Italicae. A Corpus of Italic Inscriptions, 3 vols., London. Dedè, F. (2013), I nomi greci in -αρ e -τωρ. Eteroclisi e classi nominali, Roma. Degrassi, A. (1965), Inscriptiones Latinae liberae rei publicae, editio altera aucta et emendata, Firenze. De Sanctis, G. (2015), La logica del confine. Per un’antropologia dello spazio nel mondo romano, Roma. DELG = Boisacq, É., Dictionnaire étymologique de la langue grecque, étudiée dans ses rapports avec les autres langues indo-européennes, Heidelberg/Paris 1916. DELL = Ernout, A./Meillet, A., Dictionnaire étymologique de la langue latine. Histoire des mots, 4ème éd., Paris 1959. Dettori, E. (1997), Osservazioni su lat. signum, AIΩN 19, 213–227. Di Giovine, P. (1990), Studio sul perfetto indoeuropeo, I, Roma. Dionisio, A. (2011), “Alcune osservazioni a proposito dell’iscrizione del Guerriero di Capestrano”, in: G. Firpo (ed.), Fides amicorum. Studi in onore di Carla Fayer, Pescara, 169–190. DKS = Bailey, H.W., Dictionary of Khotan Saka, Cambridge/London/New York/Melbourne 1979. Dorothée, S. (2002), “Aux origines de la notion de signe: les emplois de signum chez Plaute”, Revue de philologie, de littérature et d’histoire anciennes 76, 33–48. Dorothée, S. (2006), À l’origine du signe: le latin signum, Paris. Durante, M. (1958), Prosa ritmica, allitterazione e accento nelle lingue dell’Italia antica, RL 4, 61–98. Durante, M. (1978), “I dialetti medio-italici”, in: A.L. Prosdocimi (ed.), Popoli e civiltà dell’Italia antica, VI, Lingue e dialetti, Roma, 789–824. Durkin-Meisterernst, D. (2004), Dictionary of Manichaean Texts, III, Texts from Central Asia and China, Part I, Dictionary of Manichaean Middle Persian and Parthian, Turnhout. EDG = Beekes, R., Etymological Dictionary of Greek, Leiden/Boston, 2010. EDIL II = Rastorgueva, V.S./Èdel'man, Dž. I., Etymological Dictionary of the Iranian Languages / Ètymologičeskij slovar’ iranskix jazykov, Vol. 2, Moscow, “Vostochnaya Literatura” Publishers, 2003. EDIL IV = Èdel'man, Dž. I., Etymological Dictionary of the Iranian Languages / Ètymologičeskij slovar’ iranskix jazykov, Vol. 4, Moscow, “Vostochnaya Literatura” Publishers, 2011. EDIW = Cheung, J., Etymological Dictionary of the Iranian Verb, Leiden, 2007. EDLIL = de Vaan, M., Etymological Dictionary of Latin and the Other Italic Languages, Leiden/ Boston, 2008. ÈSRJA II = Fasmer [Vasmer], M., Ètimologičeskij slovar’ russkogo jazyka, tom II, Moskva, 1986. EWA I = Mayrhofer, M., Etymologisches Wörterbuch des Altindoarischen, 1. Band, Heidelberg, 1992. EWG = Hofmann, J.B., Etymologisches Wörterbuch des Griechischen, München, Oldenbourg, 1949. Franchi De Bellis, A. (1981), Le iovile capuane, Firenze. Friedman, J, (1999), Analysis of Simple *-r/n- Heteroclisis in Proto-Indoeuropean, in: V.V. Ivanov/ B. Vine (eds.), UCLA Indo-European Studies, I, Los Angeles, 31–69. Froehde, Friedrich (1891), “Griechische und lateinische etymologien [sic]”, in: Beitr. zur Kunde der indogerm. Sprachen 17, 303–319.

  Marco Mancini Gershevitch, I. (1945), “Sogdian compounds”, TPhS, 137–149. Gershevitch, I. (1954), A Grammar of Manichean Sogdian, Oxford. GEW = Frisk, H., Griechisches etymologisches Wörterbuch, 2. Band, Heidelberg, 1973. Ghilain, A. (1939), Essai sur la langue parthe. Son système verbal d’après les textes manichéens du Turkestan Oriental, Louvain. Gignoux, P. (1972), Glossaire des Inscription Pehlevies et Parthes, London/Lund. Ginevra, R. (2020), “Gods who shine through the millennia: Old Norse Baldr, Celtic Belinos, Old Irish Balar, and PIE *bhelH‑ ‘be white, shine’”, in: L. Repanšek/H. Bichlmeier/ S. Sadovski (eds.), Vácāṁsi miśrā kr̥ ṇavāmahai. Akten der Arbeitstagung der IndogermanischenGesellschaft und des IWoBA XII vom 4. bis 7. Juni 2019 in Ljubljana aus Anlass der Hundertjahrfeier der dortigen Indogermanistik (= Studien zur historischvergleichenden Sprachwissenschaft 17). Hamburg, 187–206. Girardi, C. (2018), “Le divinità plurali dei confini nelle iscrizioni in lingue epicorie e in latino”, Anuari de filología. Antiqua et mediaeualia 8, 393–410. Grienberger, Th. (1926), “Italica”, KZ 54, 57–76. Gusmani, R. (1986), Saggi sull’intereferenza linguistica, II ediz. accresciuta, Bagno a Ripoli. Ḥasandust, M. (2014), Farhang-e rišešenåxti-ye zabån-e fårsi, ǰeld avval, Tehrān 1393 [2014]. Henning, W.B. (1932), “Ein manichäischer kosmogonischer Hymnus”, Nachrichten der Gesellschaft der Wissenschaften zu Göttingen, 214–228. Henning, W.B. (1971), A Fragment of a Khwarezmian Dictionary, ed. by D.N. MacKenzie, London/ Lund. Hinz, W. (1975), Altiranisches Sprachgut der Nebenüberlieferungen, Wiesbaden. Hoffmann, K. (1992), Aufsätze zur Indoiranistik, Band 3., herausgeg. v. S. Glauch-Robert Plath Sabine Ziegler, Wiesbaden. Horn, P. (1893), Grundriss der neupersischen Etymologie, Strassburg. Horn, P. (1898–1901), Neupersische Schriftsprache, in: W. Geiger/E. Kuhn (eds.), Grundriss der iranischen Philologie, 1, 2, Strassburg, 1–200. Hübschmann, H. (1895), Persische Studien, Strassburg. Hübschmann, H. (1897), Armenische Grammatik, 1. Teil. Armenische Etymologie, Leipzig. IEW = Pokorny, J., Indogermanisches etymologisches Wörterbuch, 1. Band, Bern/Stuttgart 1959. ILLRP = Degrassi, A. (1965), Inscriptiones Latinae liberae rei publicae, editio altera aucta et emendata, Firenze. IosPE I2 = Inscriptiones antiquae orae septentrionalis Ponti Euxini Graecae et Latinae, vol. I2, Petropoli 1916. Jiménez Zamudio, R. (1986), Estudio del dialecto peligno y su entorno lingüístico, Salamanca. Kellens, J. (1995), Liste du verbe avestique, Wiesbaden. Kent, R.G. (1953), Old Persian. Grammar. Texts. Lexicon, 2nd ed., New Haven. KEWA I = Mayrhofer, M., Kurzgefasstes etymologisches Wörterbuch des Altindischen, Band I, Heidelberg, 1956. Kümmel, M.J. (2000), Das Perfekt im Indoiranischen. Eine Untersuchung der Form und Funktion eine ererbten Kategorie des Verbums und ihrer Weiterentwicklung in den altindoiranischen Sprachen, Wiesbaden. La Regina, A. (1966), “Le iscrizioni osche di Pietrabbondante e la questione di Bovianum Vetus”, Rheinisches Museum für Philologie, N.F., 109, 260–286. La Regina, A. (2011), “Il Guerriero di Capestrano e le iscrizioni paleosabelliche”, in: L. Franchi dell’Orto (ed.), Pinna Vestinorum e il popolo dei Vestini, Roma, 230–273.

The Etymology and Semantics of Oscan pukam  

La Regina, A. (2012), “Sannio. Pietrabbondante”, SE 75, 2009, 315–336. Lazzeroni, R. (1978), “Contatti di lingue e di culture nell’Italia antica: i dati delle iscrizioni posteriori alla silloge del Vetter”, in: La cultura italica, Atti S.I.G., Pisa, 91–102. Lazzeroni, R. (1983), “Contatti di lingue e di culture nell’Italia antica. Modelli egemoni e modelli subordinati nelle iscrizioni osche in grafia greca”, AIΩN 5, 171–182. Lazzeroni, R. (1997), “Il dativo “sabellico” in -a. Contributo alla conoscenza della latinizzazione dei Peligni” [1965], ora in: T. Bolelli/S. Sani (eds.), Scritti scelti, Pisa, 300–318. Lazzeroni, R. (2019)[recte 2020], “Etimologie creative: ritorno a Varrone?”, AGI 104, 2, 129–143. Lejeune, M. (1964), “Venus romaine et Venus osque”, in: M. Renard/R. Schilling (eds.), Hommages à Jean Bayet, Bruxelles/Berchem, 383–400. Lejeune, M. (1970), “Phonologie osque et graphie grecque”, RÉA 72, 271–316. Lejeune, M. (1975), “Réflexions sur la phonologie du vocalisme osque”, BSL 70, 233–251. Leumann, M. (1952), Morphologische Neuerungen im altindischen Verbalsystem, Amsterdam. LEW = Walde, A./Hofmann, J.B., Lateinisches etymologisches Wörterbuch, 1. Band, 4. Aufl., Heidelberg, 1965. LIV = Rix, H./Kümmel, W. (eds.), Lexicon der indogermanischen Verben, 2. erweit. und verbess. Aufl., Wiesbaden, 2001. MacKenzie, D.N. (1967), “Notes on the transcriptions of Pahlavi”, BSOAS 30, 17–29. MacKenzie, D.N. (1971), A Concise Pahlavi Dictionary, London. MacKenzie, D.N. (1978), “Shapur’s shooting”, BSOAS 41, 499–511. MacKenzie, D.N. (1980), “Mani’s Šābuhragān – II”, BSOAS 43, 288–310. Mancini, M. (2004), “Uno scioglilingua da Falerii Veteres e l’etimologia di fal. umom", AGI 89, 200–211. Mancini, M. (2018), “Sul metodo linguistico-epigrafico: l’iscrizione latina di Alba Fucens tra norma e uso”, in: R. Bombi/F. Costantini (eds.), Percorsi linguistici e interlinguistici. Studi in onore di Vincenzo Orioles, Udine, 371–395. Mancini, M. (2021), “Does Prenestinian fhe⋮faked Actually Exist?”, JLS 20, 75–108. Marchese, M.P. (1974), “Sulle nuove (e antiche) iscrizioni di Pietrabbondante”, SE 42, 371–378. Marchese, M.P. (1997), “Osco prupukid sverruneí”, Quaderni Dipart. di Linguistica - Univ. di Firenze 8, 1–6. Marinetti, A. (1985), Le iscrizioni sudpicene, I, Testi, Firenze. Marinetti, A. (2000), La romanizzazione linguistica della Penisola, in: J. Herman/A. Marinetti (eds.), La preistoria dell’italiano, Tübingen, 61–79. Marinetti, A./Cresci Marrone, G. (2011), “Ideologia della delimitazione spaziale in area veneta nei documenti epigrafici”, in: G. Cantino Wataghin (ed.), Finem dare. Il confine tra sacro, profano e immaginario, Vercelli, 287–311. Meiser, G. (1986), Lautgeschichte der umbrischen Sprache, Innsbruck. Montagnaro, L. (2011), “Venetico termon. Lessico e istituzionalità nella terminologia della confinazione”, Alessandria 5, 419–427. Morgenstierne, G. (1927), A New Etymological Vocabulary of Pashto, Wiesbaden. Narten, J. (1968), “Zum “proterodynamischen” Wurzelpräsens”, in: J.C. Heesterman/G.H. Schokker/V.I. Subramoniam, V.I. (eds.), Pratidānam: Indian, Iranian and Indo-European Studies Presented to Franciscus Bernardus Jacobus Kuiper on his 60th Birthday, Mouton, 9–19. Nussbaum, Alan J. (2014), “Greek τέκμαρ ‘sign’ and τέκμωρ ‘sign’: Why both?”, in: N. Oettinger/ Steer, T. (eds.), Das Nomen im Indogermanischen, Wiesbaden, 215–260. Nyberg, H.S. (1931), Hilfsbuch des Pehlevi, 2. Glossar, Uppsala. Nyberg, H.S. (1974), Manual of Pahlavi, Part II: Glossary, Wiesbaden, Harrassowitz.

  Marco Mancini Orioles, V. (2002), “Il termine compagno nella storia del movimento operaio. Gli anni della Prima Internazionale (1864–1876) [1979]”, ora in: V. Orioles, Percorsi di parole, Roma, 69–85. Pagliaro, A. (1955), “Il nome della turchese”, AGI 39, 142–165. Peruzzi, E. (1978), Aspetti culturali del Lazio primitivo, Firenze. Piccaluga, G. (1974), Terminus. I segni di confine nella religione romana, Roma. Pinault, G.-J. (2017), “In the heat of the day”, in: I. Hajnal/D. Kölligan/K. Zipser (eds.), Miscellanea Indogermanica. Festschrift für José Luis García Ramón zum 65. Geburtstag, Innsbruck, Institut f. Sprachen u. Literaturen der Universität Innsbruck, 641–659. Pisani, V. (1959), “Über eine pälignische Inschrift (Co. 208 bis, Pl. 246 d) und die Herkunft des Oskisch-Umbrischen [1952]”, in: V. Pisani, Saggi di linguistica storica, Torino, 136–159. Pisani, V. (1964), Le lingue dell’Italia antica oltre il latino, seconda ediz., Torino. Piwowarczyk, D.R. (2011), “Formations of the perfect in the Sabellic languages with the Italic and Indo-European background”, Studia linguistica Universitatis Jagellonicae Cracoviensis 128, 103–126. Poccetti, P. (1979), Nuovi documenti italici a complemento del Manuale di E. Vetter, Pisa. Poccetti, P. (1981), “Elementi culturali negli epitafi poetici peligni. II. Modelli formulari”, AIΩN 3, 259–270. Poccetti, P. (1994), “Per un dossier documentario dei riflessi di dottrine misteriche e sapienziali nelle culture indigene dell’Italia antica. Note sulle iscrizioni osche Ve 161,185; Po 103”, in: A.C. Cassio/P. Poccetti (eds.), Forme di religiosità e tradizioni sapienziali in Magna Grecia, Atti del Convegno, AION-fil.let. 16, 109–136. Poccetti, P. (2000), “La diffusione di dottrine misteriche e sapienziali nelle culture indigene dell’Italia antica: appunti per un dossier”, in: AA.VV., Tra Orfeo e Pitagora. Origini e incontri di culture nell’antichità, Napoli, 93–126. Poccetti, P. (2009), “Paradigmi formulari votivi nelle tradizioni epicoriche dell’Italia antica”, in: J. Bodel/M. Kajava (eds.), Religious Dedications in the Greco-Roman World. Distribution, Typology, Use, Rome, 43–93. Poccetti, P. (2016), “Note sulla nuova iscrizione osca da Cuma. Il nuovo teonimo nel contesto dei culti della città in fase di romanizzazione”, in: A. Ancillotti/A. Calderini/R. Massarelli (eds.), Forme e strutture della religione nell’Italia mediana antica - Forms and Structures of Religion in Ancient Central Italy, III Convegno Internazionale IRDAU, Roma, 573–592. Pucci, G. (1996), “Terminus. Per una semiotica dei confini nel mondo romano”, in: G. Manetti, (ed.), Knowledge through Signs. Ancient Semiotic Theories and Practices, Turnhout, 295–307. Rix, H. (1994), “Südpikenisch kdúiú”, HS 107, 105–122. Rix, H. (1996), “Variazioni locali in osco”, in: L. Del Tutto Palma (ed.), La Tavola di Agnone nel contesto italico, Firenze, 243–261. Rix, H. (2002), Sabellische Texte. Die Texten des Oskischen, Umbrischen und Südpikenisch, Heidelberg. Salemann, C. (1895–1901), “Mittelpersisch”, in: W. Geiger/E. Kuhn (eds.), Grundriss der iranischen Philologie, 1, 1, Strassburg, 249–342. Salemann, C. (1912), “Manichaica IV”, Izvěstija Imperatorskoj Akademii Naukŭ - Bulletin de l’Académie Impérial des Sciences de St.-Pétersbourg 1912, 33–50. Schmitt, R. (2014), Wörterbuch der altpersischen Königsinschriften, Wiesbaden. Schwyzer, Eduard (1977). Griechische Grammatik. 1. Bd.: Allgemeiner Teil. Lautlehre. Wortbildung. Flexion, 5. unver. Aufl., München.

The Etymology and Semantics of Oscan pukam  

Skjærvø, P.O. (2007), Zoroastrian Texts, by the author. Triantafillis, E. (2011), “Ancóra sull’iscrizione ‘ernica’ sic Rix He 2”, Alessandria 5, 311–331. Van Windekens, A.J. (1976), Le tokharien confronté avec les autres langues indo-européennes, Vol. I: La phonétique et le vocabulaire, Louvain. Ve = Vetter, E., Handbuch der italischen Dialekte, 1. Band, Heidelberg 1953. Wolff, F. (1910), Avesta. Die heiligen Bücher der Parsen, übersetzt auf der Grundlage v. Chr. Bartholomae altiranischem Wörterbuch, Strassburg. WOU = Untermann, J., Wörterbuch des Oskisch-Umbrischen, Heidelberg 2000. Zair, N. (2016a), Oscan in the Greek Alphabet, Cambridge. Zair, N. (2016b), “Vowel Weakening in the Sabellic Languages as Language Contact”, IF 121, 295–315. Zavaroni, A. (2016), “L’iscrizione falisca su oinochoe EF 4 e i termini oschi con base *puk-”, AION, n.s. 5, 223–258. Zuin, F. (2020), Umbro fakust, Osco fefacust, Presannita fεfικεδ. La morfologia del Perfettivo nelle varietà sabelliche e nelle lingue frammentarie dell’Italia antica, Alessandria. Zvetaieff, I. (1884), Inscriptiones Italiae mediae dialecticae ad archetyporum et librorum fidem, Lipsiae.

Simona Marchesini

The Messapic Inscription from Grotta Poesia MLM 3 Ro: Analysis with Frame Semantics Abstract: The exploration of the methodological frameworks called “frame semantics” and “ethnographic semantics” led me to reconsider the epigraphical environment of Grotta Poesia, Roca Vecchia, a maritime cave sanctuary with walls covered with hundreds of Messapic, a few Greek, and almost two dozen Latin inscriptions. A a new survey campaign shed new light onto the text of a Messapic inscription (MLM 3). Also, in Frame Semantics (cf. Fillmore 1982), “words represent categorisations of experience, and each of these categories is underlain by a motivating situation occurring against a background of knowledge and experience”. Fillmore defines the “frame” as a system of categories structured following some motivating context, assuming that “The background context is absolutely essential to understanding the category”. In other words, in the process of using a language, a speaker “applies” a frame to a situation and shows that he/she intends this frame to be applied by using words recognised as grounded in such a frame. This theoretical scenario has been applied in interpreting the inscription MLM 3, the meaning of which was deemed certain by past contributions, despite some hermeneutical issues. Special attention is paid to the verb eipeigrave/ ipigrave, for which I propose here a new etymology.

 The epigraphical environment of Grotta Poesia as medium The context of Grotta Poesia, characterised by the co-existence of both prehistoric engravings and historic-era inscriptions, can be labelled as a multifunctional medium. Analysis of the epigraphic evidence has shown that at least two different

 A research-stay at the Trinity College Dublin Long Room Hub (from March 13th to April 13th 2019), conducted within the framework of the Research Theme “Manuscript, Books, and Printed Culture”, led me to reconsider the relationship between medium and message within the cave of Grotta Poesia, Roca, and to approach this special epigraphic environment from a new perspective. I am thankful to the MuMil project and its organizers, Prof. Albio Cesare Cassio and Dr. Sara Kaczko, for offering me the opportunity to present the results of my research during the project’s final meeting. https://doi.org/10.1515/9783110779684-012

  Simona Marchesini kinds of texts were written on the surface of the walls in the historical period: on the one hand, there are votive texts, and on the other, a network of overlapping inscriptions that resemble so-called tags or (better) “ego-inscriptions”. This latter term should not be confused with the widespread Etrusco-Italic or Greek class of “speaking inscriptions”, in which the object speaks in the first person and expresses its belonging to a (dead) person. Instead, the term “egoinscriptions” refers to the tendency in humans to leave a sign of their individual presence in a particular place. The votive inscriptions of Grotta Poesia are generally quite legible even to modern readers, since they are fairly uniform and do not display different, overlapping writing acts; the other group, however, consists of a multi-layered network of inscriptions, which can be difficult to decipher even for experienced scholars. Ego-inscriptions reflect the tendency to leave a sign (compare the Latin tituli memoriales) — a signature, a sign of presence, as has been observed in many special places, in the past and the present: the September 11th Memorial in New York City; the ladies bathroom of the Elephant House restaurant in Edinburgh, where J.K. Rowling began to write Harry Potter; the Door of Reconciliation at St. Patrick Cathedral in Dublin, which commemorates the reconciliation between the O’Connor and Fitzgerald families; Juliet’s balcony in via Cappello in Verona, as inspired by Shakespeare’s Romeo and Juliet — these are all examples of this kind of writing. Each of these sites, along with the writing contained in it, fulfills the need to mark someone’s presence in a place. The message does not need to be read, and the medium does not communicate.1 Writing has, in these cases, a performative function: written texts substantiate the action they express while they are written.2 Votive inscriptions, on the other hand, are related to a different human tendency: they express a more profound need to get in touch with the gods that are worshipped in a given place. In terms of their writing function, votive inscriptions communicate with a target outside the subject; tags, however, are addressed to the person who writes.

 1 On the question of texts that are not read, see the essays on hidden texts in Campus, Marchesini/Poccetti 2020. 2 Marchesini 2016: 190–191; Cardona 1981; 1986; Fraenkel 2006; 2007.

The Messapic Inscription MLM 3 Ro and Frame Semantics  

 The inscription MLM 3 Ro Among the votive inscriptions found in Grotta Poesia, one of them (MLM 3 Ro) deserves particular attention for several different reasons: firstly, for the kind of worship described in the text; secondly, for the role played by the woman named in the first line; and thirdly, for the question of the morphological analysis and meaning of the verb eipeigrave found in the first line. Initially, this verb seemed to be crucial for interpreting the relationship between medium and message: the verb was thought by scholars to express the activity of writing itself, and was linked (as a loanword? as a commonly inherited Indo-European verb?) to Greek ἐπιγράφω. However, after undertaking a new analysis of the text, and after reconsidering it from the perspectives of the phenomenology of votive actions and of frame semantics I was prompted to explore other possible explanations. As is well known, all the texts from Grotta Poesia were discovered during explorations carried out in the late 1980s and 1990s by Cosimo Pagliara, who in 1987 presented a report on the discovery of the Grotta, with its engravings and its epigraphic remains, in the Annali della Scuola Normale Superiore di Pisa.3 Pagliara’s provisional transcriptions of the inscriptional texts were partially published in the Monumenta Linguae Messapicae in 2002.4 Thanks to the acquisition of resin casts produced by the Consortium for the study, consolidation, and valorisation of Grotta Poesia (led by The University of Salento with seven public and private partners), established in 2017, and to the photogrammetric images produced by the team at La Sapienza under the auspices of the HERA project, I and my colleagues were able to recognise some of the inscriptions previously published in MLM. Unfortunately, we have not yet been able to identify the inscription MLM 3 Ro on these resin casts. However, Cosimo Pagliara and Riccardo Guglielmino presented images of the inscription MLM 3 Ro during a conference in Kalamata, Greece, in 2003. I owe to Guglielmino the availability of the photos taken from the presentation made at that time. Even though it is of rather low resolution, the photo of MLM 3 Ro allowed me to identify most of the letters, and was sufficient to produce a first transcription of the text (Fig. 41).

 3 Pagliara 1987. 4 Under the label MLM 3 Ro.

  Simona Marchesini

Fig. 41: Photo of MLM 3 Ro. Courtesy of Riccardo Guglielmino.

Another photo, however, was recently brought to my attention by Mario Lombardo. This photo seems to have been taken after the wall was polished, with the result that part of the text is more legible, even if the resulting text now contains some differences from the text published in MLM. The last survey carried out by Teodoro Scarano in the Grotta in June 2019 made it possible to take some more detailed pictures, which confirmed the readings of the new text (Fig. 42).

Fig. 42: Detail of the inscription MLM 3 Ro. Courtesy of Teodoro Scarano.

The Messapic Inscription MLM 3 Ro and Frame Semantics  

The text is written from left to right in scriptio continua in four lines. The letterforms present the typical features of the sixth phase of Messapic epigraphy as established in the Monumenta, such as the prevalence of A with angled medial stroke, isosceles delta, and three-stroke sigma; this phase probably belongs in the second half of the 3rd cent. BCE (Fig. 43).

Fig. 43: 6th phase of Messapic epigraphy. Drawing S. Marchesini.

In the second line, instead of θaotori ipigrave aton, as published in MLM, we can now read θaotori ipigraveston. After considering all possible word divisions, I have opted for ipigraves ton. To sum up, the new reading is (Fig. 44): dazomadazinnahaakreθθihi|2θaotoriipigraveston|3θaotor akreθθesθaotori|4taosiiassetiandahezzandaosapistaθi

With word-divisions: dazoma dazinnaha akreθθihi θaotori ipigraves ton θaotor akreθθes θaotori taosi iasseti anda hezzan daos apistaθi.

  Simona Marchesini

Fig. 44: Drawing of MLM 3 Ro from a photo taken after polishing (drawing: S. Marchesini).

The inscription is skillfully composed: stylistic expedients such as alliteration (dazoma dazinnaha) and anaphora (θaotor on the second and third line), or the position of the personal/divine name Θaotor expressed three times (in a kind of epanalepsis in the third line), reveal that the text was written by a skilled individual. I attempted to explore possible prosodic content, but I was unfortunately not able to reach any conclusive results on that score. As regards the analysis of the text itself: at the beginning of the text (first line) we find a feminine name composed of three elements in the nominative case and in topic position. The first element, Dazoma, is the feminine version of the masculine praenomen Dazomas, which is common in Messapic epigraphy, especially in the stem Dazimo-.5 In the following word, Dazinnaha, the reduplication of — due to the meeting of /n/ and the /j/ of the suffix -ja (*dazin-yā) — tells us that this component of the name is the feminine appositive, which in Messapic usage is generally a patronymic, or else (but less commonly) the nomen gentilicium, i.e., family name, according to Untermann’s analysis.6 The third element of the name is probably the gamonymic, i.e., the husband’s family name in the genitive: “wife of Akreθθes” (*Akreθ-yos). In the second line, the name of the god Θaotor is written in the dative case: Θaotori. Immediately after this, we find a verb, initially read as ipigrave, but now corrected to ipigraves; this is the 3rd person singular of the perfect, and must have as its subject the individual announced in the nominative case in the first line of the text, namely the woman (I will come back to the meaning of the verb below). We do not have any exact Messapic comparanda for the word ton that follows. A form atan occurs in another inscription of the Grotta, MLM 13 Ro,

 5 de Simone 1988: 339–341; 1991: 316; Däuber 1991: 326. 6 Untermann 1964.

The Messapic Inscription MLM 3 Ro and Frame Semantics  

and, at least to judge by the ending, it seems probable that it is an accusative. In the Greek dialects the deictic το- is already used in Homer (PIE *to-, *teh2- “that” as a demonstrative and anaphoric pronoun, and later as an article);7 the stem is spread throughout the entire Indo-European linguistic area.8 The third line begins with an onomastic formula, Θaotor Akreθθes, which presents a praenomen formally identical to that of the god’s, followed by a family name which is the same as that of the woman’s gamonymic: Akreθθes (: *akreθ-yos). This is Dazima’s husband. The name of the god Θaotor then follows in the dative. The following sequence, iasseti anda hezzan, might contain the direct object to which the verb refers (thus de Simone 1988). The sequence iassetianda presents some difficulties in the divisio verborum: on the one hand, the word iassetian (which should be considered an -ia stem) occurs in another inscription of the Grotta (MLM 13 Ro, see below); on the other hand, anda (< *n̥ do), the preposition from an Indo-European stem (*h1ndo(m) “inside”) occurs frequently in Messapic epigraphy9 and can be compared with Gallic ando-,10 Hethitic anda (“inwards”, “into”), Greek ἔνδον (“inside”, “at home”), and perhaps Old Latin endo of the Duenos inscription (ted endo: “in/onto you”) and the XII Tables, from an Indo-European stem (*h1ndo(m) “inside”). Now, given that iasseti seems to be morphologically incomplete (it is attested as [i]assetian in MLM 13 Ro, from the same Grotta Poesia), we might be dealing with a sandhi phenomenon, whereby the morpheme -an, phonetically the same at word-end and at the beginning of the following preposition (iassetian + anda), has been reduced by a kind of haplography. The word hezzan seems to have a convincing etymology: *sedyā̆m might be the abstract verbal noun formed from *sodéyō- (where /*s-/ > /h-/ and /dy-/ > /z-/, as in Zis < *dyeus), which is causative formation of *sed-/sod- “to sit” (compare Italian seggio).11 The verb is known from other Messapic evidence as well, in the form of the perfect hadive ( /b/ — as it appears in the IndoEuropean *to-bhero- attested in Messapic tabara —, we do not have evidence for medial /-bh-/ > /v/, since medial /-v-/ in Messapic is elsewhere attested only in the perfect, which is problematic for other reasons (see below). – The proposal (de Simone 1988) that the Messapic form might be a loanword from the Greek, adapted in Messapian — thanks to the morpheme ave — suffers from the drawback of merging the Greek stem in -/ph/ with the -v- suffix of the perfect. – Moreover, if we accept the meaning “write” for eipeigrave, we encounter semantic and historical inconsistencies, which I will try to outline below. Another problematic feature is the use of two different tenses for the two verbs: the first line refers to the woman in the perfect tense, whereas the final verb  14 See Kajanto 1982: 178. 15 Starting from Vittore Pisani 1964: 244, no. 85; see also Ribezzo 1914/1915: 12; Krahe 1959: 254; Parlangèli 1960: 306; de Simone 1988: 389. 16 Chantraine 1984: s.v. 17 Beekes 2010: s.v. 18 Pisani 1964: 244, no. 85; see also de Simone 1988: 389–391. 19 Ribezzo 1914/1915: 12. 20 Krahe 1959: 254. 21 Parlangèli 1959: 254.

  Simona Marchesini apistaθi (“he sets up”, “donates”) is in the present tense.22 The man and the woman are linked by marriage. They might have presented the dedication to the god Taotor together, but it seems that only the man makes the dedication. Why should the verbs be in different tenses? The only reason one can think of is that the two actions take place at different moments of the story. The woman has dedicated something to Taotor in the past, and therefore her husband donates or sets up the throne/the seat to the god. As for the socio-historical aspects: should we really suppose that the writing was carried out by women? Following Untermann (1964), we know that women did not have a nomen gentilicium in most cases and were instead designated by the patronymic and the gamonymic. In the two inscriptions in which the verb eipeigrave is associated with a feminine name formula, however, we should probably assume that the second name is a nomen gentilicium (see above). Furthermore, what would the woman have written to the god Taotor, before the man later dedicated a seat to him? What was the reason for the ex-voto? Let us therefore abandon the root for “to write” in search of other possible PIE roots. A first possible candidate might be the root *ĝher-/ĝhr-,23 which has several meanings. One meaning is: “gefallen finden, begehren” (“to enjoy”, “to reap benefits”: cf. Paleo-Umbrian heriseí “dem, der gewünscht hat”, Vedic háryati “freut sich an etwas, hat gern”; Umbrian pis-her “wer auch immer”; probably also the Venetic dual participle *horeionte “freudig”). Another, more fitting meaning is “nehmen”, “holen” (“to take”, “to get”: Vedic ahr̥ thās “hast genommen”). The sense that this root takes on in Sanskrit haráti means “bears, offers”, may shed light also on the Messapic verb. If we assume this meaning, we can better explain the accusative ton and the dative of the god Taotor; the epi- preverb might have restricted the meaning of the action to a sense “set upon, offer on”. In both cases, the development -gra- might come from a syllabic r̥ . If we suppose that epi-gra-v-es(t) is a perfect tense with the meaning “(she) offered, brought (something) upon”, then this meaning would also be suitable for the other occurrences of the verb in Messapic epigraphy: i.e., the long inscription from Carovigno with the initial invocation to Zis, and the text from Diso with the woman’s dedication. Let us now consider the question of the -v- perfect. As is well known, the -vinfix is one of the possible ways to build the Latin perfect, together with stemreduplication (mĕ-mĭni), the lengthening of the stem vowel (vĕnio : vēni), the -s 22 de Simone 1988: 384–389. 23 LIV: 176–177 = Pokorny 442.

The Messapic Inscription MLM 3 Ro and Frame Semantics  

perfect (dico: dixi), and the use of the unmodified stem (as in pando, pandi). The -v- perfect is unknown to the neighbouring Italic languages.24 Meiser quotes two different potential explanations of the ŭ suffix: on the one hand, the perfect of the stem *bhuh- meaning “to become”, Lat. “fui”; on the other hand, the suffix *-u̯ os of the perfect active participle. With a few modifications, Meiser’s analysis was accepted by H. Rix25 and Meier-Brügger.26 Recently A. Willi27 took up the question again, presenting three possible solutions: (1) the comparison with Vedic Sanskrit forms in -au such as dadáu (“gave”), dadháu (“put”); (2) the reanalysis of the root *bhuH-, reinterpreted as a tense-morpheme and incorporated in the verbal paradigm bhu-bhuH-h2ai, and 3) a synthesis of the two foregoing explanations. None of the abovementioned scholars considered the Messapic evidence, where the -v- perfect is well documented. The occurrences are: 1) dupave: 3rd person singular (MLM 12 Ruv, from Ruvo di Puglia, 3rd cent. BCE): to dupa-. The perfect appears at the end of a text which mentioned a gift from Artos Atotios (to a woman?) and is written on a bronze tablet in five lines. The inscription — artos atotios tai θoitai /taiθoitai gunakēai pensklen dupave — belongs to a northern variant of Messapic epigraphy, called Daunian, as is revealed by the use of the grapheme for [ē]. 2) hadive: again, 3rd person singular (MLM 23 Cae, from Caeliae, 3rd cent. BCE). The text is written on a stone block, and many letters are missing at the beginning of the first line as well as at the end of each of the five lines: ]......[--]aθida[---] plator[---] lahon[---]hadive[. Besides the perfect ]hadive[, only the personal name Plator is readily recognisable in this text. As in the previous case, the verb is placed at the end of the text, as often happens in formulaic expressions related to gifts and dedications. 3) niligaves: all occurrences are from Roca (MLM 6, 9, 10, 18 Ro): this is a 3rd person (*ni-legh-a-v-e(s), (*legh- ?) which may perhaps be comparable to Gallic legasit, 3rd person -s perfect from a root *legh- “coucher, allonger”, Greek λέχεται, Gotic ligan etc.;28 or else it may perhaps come from a root *leg- “(faire) couler” (compare Ancient Irish legaid “il fond”, avec un sens ± de “faire une libation”).29 The term in any case could be explained as form-

 24 Meiser 1998: 204; 2003: 220–222. 25 Rix 1992: 221. 26 Meier-Brügger 2010: 296. 27 Willi 2009: 229. 28 Delamarre 2003: s.v. 29 Lambert 2002: L-79.

  Simona Marchesini ing part of the semantic field “to lay down” within the context of worship. The fact that the same verb is documented without initial ni- in the Grotta (MLM 14–16 Ro), speaks in favour of a separation of ligaves as a verb from ni as a preposition. It is now clear that discussions of the Latin -v- perfect and its Indo-European roots should include the Messapian evidence, which significantly enlarges our frame of reference.

 Conclusions Reconsidering the semantic frame of the inscription MLM 3 Ro has led to a new explanation of the text’s meaning. In this new interpretation, the verb in the first line referring to the woman does not mean “write”, but rather “offer”, “take on (as an offering)”. The woman’s nuncupatio voti, expressed in the past, should correspond to the ex-voto effectively donated by the man in the last line. Additionally, the donation of a seat or a throne as part of the votive action might be understood as belonging to the context of a Hellenistic devotional practice that finds its closest comparisons in representations of gods and goddesses seated on a throne. As for the relationship between medium and message, it is worth emphasizing that even a difficult epigraphic setting such as the projecting walls of the maritime cave of Grotta Poesia may host refined inscriptions, which express the act of votive dedication in stylized language. Once again, the text’s function and its message are more important than the medium itself.

Bibliography Beekes, R.S.P. (2010), Etymological Dictionary of Greek, Volume I, Leiden/Boston. Bentini et al. (2018): Bentini, L./von Eles, P./Esposito, A./Mazzoli, A./Rodriguez, E. “Wooden thrones: ritual and function in Italian Iron Age”, Arimnestos. Ricerche di Protostoria Mediterranea 1, 171–186. Cardona, G.R. (1981), Antropologia della scrittura, Torino. Cardona, G.R. (1986), Storia universale della scrittura, Milano. Chantraine, P. (1984), Dictionnaire Étymologique de la langue grecque, I-II, Paris 1984–1990. Däuber, U. (1991), “L’onomastica messapica: continuità e rinnovamento ermeneutico alla luce dei nuovi dati”, in: I Messapi, Atti del Trentesimo Convegno di Studi sulla Magna. Grecia, Taranto-Lecce, 4-9 ottobre 1990, Taranto, 323–343.

The Messapic Inscription MLM 3 Ro and Frame Semantics  

de Simone, C. (1988), “Iscrizioni messapiche della Grotta della Poesia (Melendugno, Lecce)”, ASNP S. III, 18.2, 325–415. de Simone, C. (1991), “La lingua messapica oggi: un bilancio critico”, in: I Messapi, Atti del trentesimo convegno di studi sulla Magna Grecia, Taranto-Lecce, 4-9 ottobre 1990, Taranto, 297–322. de Vaan, M. (2008), Etymological Dictionary of Latin and the other Italic Languages, Leiden/ Boston. Delamarre, X. (2003), Dictionnaire de la langue gauloise: une approche linguistique du vieux celtique continental, 2. éd. revue et augmentée, Paris. Fillmore, C.J. (1982), “Frame semantics”, in: Linguistics in the Morning Calm, Seoul, 111–137. Fraenkel, B. (2006), “Actes d’écriture: quand écrire c’est faire”, Language et société, Éditions de la Maison des sciences de l’homme, 3, nos. 121–122, 101–112. Fraenkel, B. (2007), “Actes écrits, actes oraux: la performativité à l’épreuve de l’écriture. Written Acts and Speech Acts: Performativity and Writing Practices”, Études de Communication 29, 69–93. Krahe, H. (1959), “Beiträge zur Illyrische Wort- und Namenforschungen”, IF 64, 26–33. Lambert, P. (2002), Recueil des Inscriptions Grecques 2.2, Textes gallo-latins sur “instrumentum”, Paris. LIV: Rix, H., Lexikon der indogermanischen Verben (LIV): die Wurzeln und ihre Primärstammbildungen, bearbeitet von Martin Kümmel et al., Wiesbaden. Marchesini, S. (2016), “Die Rolle der Schrift in Selbstwahrnehmung und Identitätskonstitution bei antiken Völkern. Das Beispiel des rätischen Gebietes (Jüngere Alpine Eisenzeit)”, Archaeologia Austriaca 100, 189–198. Matasović, R. (2009), Etymological Dictionary of Proto-Celtic, Leiden/Boston. Meier-Brügger, M. (2010), Indogermanische Sprachwissenschaft, 9. Auflage, Berlin/New York. Meiser, G. (1998), Historische Laut- und Formenlehre der lateinischen Sprache, Darmstadt. Meiser, G. (2003), Veni vidi vici: Die Vorgeschichte des lateinischen Perfektsystems, München. MLM = de Simone, C./Marchesini, S., Monumenta Linguae Messapicae, 2 volumes, Wiesbaden. Pagliara, C. (1987), “La grotta Poesia di Roca (Melendugno, Lecce). Note preliminari”, ANSP S. III, 17/2, 267–328. Parlangèli, O. (1960), Studi Messapici, Milano. Pisani, V. (1964²), Le lingue dell’Italia antica oltre il Latino, Torino. Pokorny = Pokorny, J., Indogermanisches etymologisches Wörterbuch, Bern/München 19591969. Ribezzo, F. (1914/1915), “Diso”, Neapolis 2, 293–303. Rix, H. (1992), “Zur Entstehung des lateinischen Perfektparadigmas”, in: O. Panagl/ T. Krisch (eds.), Latein und Indogermanisch, Innsbruck, 221–240. Strøm, I. (2000), “A Fragment of an Early Etruscan Bronze Throne in Olympia?”, in: S. Isager/ I. Nilsen (eds.), Proceedings of the Danish Institute at Athens 2, 67–95. Untermann, J. (1964), “Die messapische Personennamen”, in: H. Krahe, Die Sprache der Illyrier, II, Wiesbaden, 161–213. Vassileva, M. (2009), “Phrygian rock-cut thrones, ‘idols’ and Phrygian royal symbolism”, Thracia 18, 111–124. Willi, A. (2009), “To be or not to be: the Latin perfect in -v-”, HS 122, 228–247.

Michael Weiss

Latin uncia à la Heron Abstract: Most modern scholars agree that Lat. uncia “ounce” should be derived from *oinikia and connected to the Italic numeral *oinos “one”. In fact, as I show, there are insuperable phonological difficulties to this account. Instead we should prefer an idea first put forth by Heron of Alexandria who derived uncia from Greek and connected it with Gk. ὄγκος “mass, volume”. The creation of the common Italic system of measurement was a joint production of Italic and Greek speakers.

 Introduction The major Latin etymological dictionaries from Vaniček to de Vaan are united in their belief that Lat. uncia “ounce” is a derivative of Lat. ūnus “one”.1 In fact, this idea has ancient roots. Varro (Ling. 5.171) explained uncia de uno. But this is not the only ancient view of the matter. The Alexandrian engineer Heron (1st cent. CE) suggested that Latin uncia (ὀγκία as he spelled it) had its name from Gk. ὄγκος “mass, volume” (De mensuris 60.20.4: ὄνομα ἐξ Ἑλληνίδος ἀπὸ

 I would like to thank Alexander Nikolaev and Alan Nussbaum for advice and discussion. All views, correct or incorrect, are my own.  1 de Vaan 2008: 642 lists uncia, unciātim, and deunx under ūnus with no comment; Walde and Hofmann 1954: 816 “als *oin(i)cia “Einheit” zu ūnus”; Ernout and Meillet 1985: 746 “le nom de l’unité est évidemment dérivé de ūnus; et tout les autres s’y rattachent”. Vaniček 1874: 25 sub ūnus; Curtius 1862: 286 “uncia das doch wohl mit unus, unicus zusammenhäng”. Curtius 1879: 320 “unc-ia (alt oncia) will Corssen II 187 nicht hierher sondern zu ὄγκος Masse stellen. Das alte o hindert wie cōraverunt neben coirare, cūrare zeigt, die Herleitung von oinu-s nicht”. Curtius was reacting here to Corssen’s claim (1870: 187) that the epigraphical form SESCONCIA precluded a connection with ūnus, but (a) Curtius was correct that such a spelling would not preclude a connection with ūnus and (b) the forms SESCONCIA does not exist. The correct reading on the 1st cent. BCE inscription from Cremona (ILLRP 954) is SESCUNCIA. Two exceptions to this general unanimity: Bréal/Bailly 1885: 411: “uncia est une mot emprunté du grec, comme libra dont il marque une subdivision. C’est le sicilien οὐγκία.” Tucker 1931: 261: “little lump”; ὄγκος (mass, lump) < *(e)neq- “gather together, make close, compact” (v. nox) [Not < *oinicia (with ūnus). The phonetics would be difficult & the sense of “unit” does not suit the fraction]”. https://doi.org/10.1515/9783110779684-013

  Michael Weiss τοῦ ὄγκου). In this brief paper I will show that the engineer got the better of the philologian and that future etymological dictionaries of Latin should prefer Heron’s explanations to Varro’s.

 Uncia and its derivatives The Latin word uncia is first attested in Plautus in the meaning “small unit (of weight)”: (1) Men. 524

ad aurificem deferas, Atque huc ut addas auri pondo unciam [Erotium asks] you to bring [this bracelet] to the goldsmith and to add an ounce of gold to it.

(2) Rud. 912 Neque piscium ullam unciam pondo hodie cepi nisi hoc quod fero hic in rete. I haven’t caught any ounce of fish today except what I carry here in my net.

The word makes several derivatives, most quite unremarkable (unciālis “weighing one twelfth of a lībra” (Plin.); unciatim “ounce by ounce” (Ter. +); unciola “a mere twelfth” (Juv. 1.40). One derivative, unciārius, is used as an attribute of fēnus in the phrase unciarium fenus which, according to Tacitus was mentioned in the Twelve Tables: (3) Ann. 6.16

nam primo duodecim tabulis sanctum, ne quis unciario faenore amplius exerceret, cum antea ex libidine locupletium agitaretur; For it was prohibited for the first time in the Twelve Tables that anyone charge more than the unciarial interest, when previously the rate was set at the will of the rich men.

There are also a number of compounds. These fall into two groups. The fractionals .5 (sēmuncia)2 and 1.5 (sescuncia) are made with expected second

 2 As the name of a coin Varro Ling. 5.171, as the name of a measure of weight Cic. Verr. 4.57, as the name of a mysterious piece of farm equipment already in Cat. Agr. 10.3.

Latin uncia à la Heron  

member -uncia,3 but whole number multiples of uncia are formed with the athematic second member -unx (quincunx,4 septunx,5 and deunx6). This mathematical distribution is peculiar. It is likely that the athematic whole number compounds should be explained as secondary. For what could be the semantic reason to limit the root noun to whole number multiples? In fact the other members of the whole ounce series are all athematic: sextans (2 oz.), quadrans (3 oz.),7 triens (4 oz.), semis (6 oz.), bes (8 oz.), dodrans (9 oz. < *dequadrans), dextans (10 oz. < *de-sextans also called decunx), as (12 oz.). This pattern suggests that quincunx etc. owe their athematic inflection to the other members of the series, a nice instance of series contamination on a fairly abstract level. In addition the existence of pairs like līcium “cord” beside trilix “having a triple thread” and commercium “commerce” beside commers, i.e., commerx “friendly commerce” (Plaut. Stich. 519) could have helped provide more immediate models for the creation of the athematic compounds.8 Thus, as far as Latin goes, the only form that needs to be considered further is uncia itself.

 3 Pompon. fr. 111 Pappus hic in medio habitat, senica non sescunciae “Here lives a beggar Pappus, an oldster not worth a penny”. The athematic from sescunx is not found before Priscian (De figuris nummorum p. 408, l. 23). 4 First attested as the name for the arrangement of five items in the fashion that they are displayed on a die in Cic. Cat. Mai. 59 cum autem admiraretur Lysander et proceritates arborum et derectos in quincuncem ordines “But when Lysander admired both the height of the trees and their arrangement into quincunx patterns” and Caes. BGall. 7.5; this extension of meaning apparently arises from the marking of the coin with five dots sometimes, but not always, arranged in the quincunx pattern. As the name of a fraction first in Hor. Ars 326–8 dicat /filius Albini: si de quincunce remota est /uncia, quid superat? “Let the son of Albinus answer “what remains if you take away an ounce from a quincunx?”... 5 Varro Ling. 5.171 septunx a septem et uncia conclusum “a septunx is contracted from septem “seven” and uncia “ounce”. 6 Varro Ling. 5.172 reliqua obscuriora quod ab deminutione, et ea quae deminuuntur ita sunt, ut extremas syllabas habeant, ut de una dempta uncia deunx “The remaining words are less clear, because they are expressed by subtraction, and those elements from which the subtraction is made are such that they keep their last syllables: as that from which dempta uncia “one ounce is taken” is a deunx”. [Translation R. Kent, Loeb edition]. 7 Interestingly, the older term for quadrans was the thematic terruncius. According to Plin. NH 33.45 quadrans antea teruncius uocatus a tribus unciis. The word terruncius, first in Plautus, predates quadrans, first in Lucilius, and the preservation of the geminate rr (Plaut., Varr., CIL 8.17408 terr(uncio)) from rs shows that the compound must be quite old. So we see here another related but different effect of series homogenization. H. Chantraine 1962: 61 argues that the Romans originally used multiples of the uncia to name the subparts of the as and that the fractional method originated in Sicily. 8 See Garnier 2016: 310.

  Michael Weiss

 Uncia and ūnus? If we exclude any other evidence, there is no serious phonological problem in deriving Lat. uncia from *oinVkia. True, a medial vowel is not normally syncopated in such an environment (cf. tribūnicius “tribunician”), but, given that the word is part of a counting series, we could well imagine that it was subject to changes resulting from hypoarticulation. Changes of this sort are well paralleled in counting series (undecim < *oi̯nodekem) and even among the other names for fractions of an as (dodrans etc.). A somewhat more serious problem is the use of the suffix -cius, -cia. If uncia is a substantivized adjective from *oi̯nVkii̯o- this would presumably be a derivative of the adjective *oinoko- (Lat. ūnicus).9 But the semantics of ūnicus “one and only”, “unique” are far removed from uncia.10

 Uncia and ὀγκία But the real fly in the ointment is the presence of an obviously related ὀγκία which is attributed to the Sicilian writers Epicharmus and Sophron in the 2nd cent. CE work of the Antiatticist: (4) ὀγκίαν · τὸν σταθμόν. Σώφρων (fr. 148) καὶ Ἐπίχαρμος (fr. 138)11

This attribution, if taken seriously, would mean that “ounce” was attested in Sicily in the form ὀγκία in the 6th–5th cent. BCE, because Epicharmus lived between 530 and 460 and Sophron’s floruit was ca. 430. But at that date it is entirely excluded that the putative Latin source12 could have been anything other than *oi̯n(V)kia. The monophthongization of the diphthong *oi̯ did not take place in Latin until the 2nd cent. BCE at the earliest.13 Nor is it likely that the  9 An alternative derivation with the conglomerate suffix -icius is less likely since that suffix has a notable semantic specialization, making genitival adjectives from the names of officeholders (aedilicius) or public games (ludi compitalicii). See Leumann 1959: 4–12. 10 And this meaning and form are apparently quite old since they match PGmc. *ainah/ga- ~ *ainah/gan- “only” (Goth. ainaha, OHG einag). 11 See Valenti 2015: 223 for other loci and commentary. 12 Chantraine 2009: 806: “Emprunt ancien par des Grecs de Sicile du lat. uncia”. 13 See Weiss 2020: 111. The reading COEṆALIA on the Praenestine mirror CIL 12.560b dated to 260–240 BCE, if an inverse spelling for cēṇalia “dinner things” would only back date the monophthongization by a half century. In fact this reading is quite uncertain. Klug 2017 offers good arguments in favor of CREVI ALIA “I’ve decided on something else”.

Latin uncia à la Heron  

monophthongization took place in some potential intermediate language. Oscan, of course, retained the diphthong *oi̯ for its entire lifespan and we know, by a rare bit of good luck, that Sicel apparently did too on the basis of Sicilian Greek proverb (Sophron fr. 168)14 μοῖτον ἀντὶ μοίτου “like for like”.15 The only recourse for this problem would be to simply deny the attribution of these forms to Epicharmus and Sophron. But this escape hatch is not viable either because the form ὀγκία is attested epigraphically on two 5th or 4th cent. weights from Sicily: δ(αμοσία) ὀγκία and ὀγκί(α) (SEG 45.1423.1–2).16 Thus there is no doubt that the form ὀγκία is the genuine Sicilian form in the 5th and 4th centuries and that the Antiatticist’s attribution of this form to Epicharmus and Sophron is correct.17 It follows from this fact that the word cannot be derived from a Latin, Sabellic or Sicel *oi̯nVkia. I should note that these objections are hardly new. In 1964 Karl M.D. Rosen made more or less the same points in an article published in Language, apparently with little echo in subsequent scholarship.18 This may be because Rosen had no solution to offer to the question of the ultimate etymology, merely suggesting that the Latin and Greek word were two independent loans from a third unnamed language.

 14 The phrase is given in Hesychius in the corrupt form μοιτοὶ ἄντιμοι as a saying of the Sicilians (παροιμία Σικελοῖς). The current reading is due to Kaibel. Its attribution to Sophron depends on Huschke’s plausible emendation of Varro L. 5.179 itaque scribit Sophron moiton antimo et. 15 Willi 2008: 32 suggests that “*oinikiā (vel sim.) im Sikulischen zu *oinkia > *onkia geworden sein kann”. This suggestion is similar to the idea of Schmoll 1958: 95. Of course, we cannot exclude that a diphthong *oi̯ was monophthongized only in a closed syllable in Sicel and that a short medial vowel was on occasion syncopated, but we have, to my knowledge, no independent evidence for either of these developments. We have positive evidence that *oi̯ was not monopthongized in an open syllable (μοῖτον) in Sicel and *-ōi̯ was retained or shortened in the dat. sg. (αδαιοι Agostiniani 1992: 150). The word κυβιτίζειν “to strike with the elbow” attributed to Epicharmus (fr. 220) and related in some fashion to Lat. cubitus, perhaps via Sicel, shows no medial syncope. 16 ογκι is also attested on a 4th cent. BCE copper coin from Syracuse. See MacDonald, 1895: 233, n. 105. 17 The Sicilian associations of the word ὀγκία are confirmed by the quotation from Aristotle preserved in Pollux (4.174, 9.80) ἐν δὲ Ἱμεραίων πολιτείᾳ φησὶν ὡς οἱ Σικελιῶται τοὺς μὲν δύο χαλκοῦς ἑξᾶντα καλοῦσι· τὸν δὲ ἕνα οὐγκίαν. And πεντόγκιον (frr. 9, 10) and ἡμιο(ύ)γκιον (fr. 8), compounds of ὀγκία occur in quotes from Epicharmus. 18 Rosen 1964: 22: “*oiniciā would not come out onkíā in Greek. For this reason we must reject the explanation of onkíā as a loan from Latin”. The only reference to this article I have found in the subsequent literature is in Coleman’s discussion (1992: 411) of the Latin numerals, but Coleman misidentified the author as H.B. Rosén.

  Michael Weiss

 The semantic of ὄγκος What then of the other ancient explanation that derives uncia and Sicilian Greek ὀγκία from Gk. ὄγκος? Our understanding of the semantics of this word has been greatly advanced by Jacques Jouanna’s 1985 article but, though this article has made a crucial contribution, I cannot follow its ultimate etymological conclusions. On the basis of a minute examination of the usage of the Hippocratic corpus Jouanna 1985 argues that ὄγκος does not refer to weight but volume and especially swelling in the medical authors and that “swelling” is the original meaning. Such a meaning, according to Jouanna, precludes a connection with the root *h1nek̑- (Gk. ἔνεγκεῖν) as is usually assumed.19 Instead Jouanna believes that what are normally regarded as two distinct words ὄγκος1 “barb”20 and ὄγκος2 “swelling”, “volume” are two different semantic specialization of an original *h2ónkos “bending”. This account, though accepted enthusiastically in the CEG by Lamberterie and mentioned as a possibility by Beekes 2010, does not seem completely satisfactory to me. The word ὄγκος is first attested in the 5th cent. BCE in a fragment of Parmenides describing the mass or volume of the universal whole. (5) Parm. fr. 8.43 αὐτὰρ ἐπεὶ πεῖρας πύματον, τετελεσμένον ἐστί πάντοθεν, εὐκύκλου σφαίρης ἐναλίγκιον ὄγκωι But since its limit is distant, it is completed on all sides, like to the volume of a well-rounded sphere.

But the slightly younger Attic tragedians Euripides and Sophocles give us examples of less philosophical meanings. ὄγκος is used for the little bit of matter constituted by the ashes of a human body (Electra address what she thinks are Orestes ashes): (6) Soph. El. 1142 σμικρὸς προσήκεις ὄγκος ἐν σμικρῷ κύτει You have arrived a little volume in a little jar.

Or for the burden of child:

 19 See Chantraine 2009: 806; Frisk 1960: s.v. 20 ὄγκος “barb” of an arrow is attested already in Homer.

Latin uncia à la Heron  

(7) Eur. Ion 15–16 γαστρὸς διήνεγκ’ ὄγκον. ὡς δ’ ἦλθεν χρόνος, τεκοῦσ’ ἐν οἴκοις παῖδ’ ἀπήνεγκεν βρέφος (Secretly from her father, Creusa) carried the burden of her womb and when the time came gave birth to the child in the house and carried away the baby (to the cave).

Here the emphasis is not so much on volume but on the load of carrying the child. Despite Jouanna’s learned commentary, one does not carry a volume but a burden. Herodotus uses the word just twice to describe the great heap of brushwood piled up by the Scythians to serve as an altar for sacrifices to their Ares: (8) Herodot. 4.62 ἐπὶ τούτου δὴ τοῦ ὄγκου ἀκινάκης σιδήρεος ἵδρυται ἀρχαῖος ἑκάστοισι, καὶ τοῦτ’ ἐστὶ τοῦ Ἄρεος τὸ ἄγαλμα. On this heap an ancient iron sword stands for each and this is their icon of Ares. (9) Herodot. 4.62 καὶ ἔπειτα ἀνενείκαντες ἄνω ἐπὶ τὸν ὄγκον τῶν φρυγάνων καταχέουσι τὸ αἷμα τοῦ ἀκινάκεω. And then carrying up the vessel onto the heap of brushwood, they pour the blood on the sword.

From this concrete meaning “mass”, “volume” various metaphorical extensions have developed including in a negative sense “arrogance”21 or in a positive sense “pride”.22 The word is attested for Sicilian Greek by Sophron: (10) Sophr. fr. 165 κύμβου ὄγκον ἔχων having the volume of a drinking cup

The noun ὄγκος with a meaning much like those attested in the 5th cent. BCE must have existed already in Homeric times because it is the evident derivational base of the hapax ὄγκιον or ὀγκίον  21 Soph. Aj. 128 μηδ’ ὄγκον ἄρῃ μηδέν’, εἴ τινος πλέον /ἢ χειρὶ βρίθεις ἢ μακροῦ πλούτου βάθει “Don’t take up any arrogance, if you outweigh somebody in strength of hand or in depth of great riches” where the connection between the concrete sense “bulk” and the abstract “selfimportance” pride is stressed by the verbs ἄρῃ and βρίθεις. 22 Soph. Trach. 817–818 (Hyllos speaking) ὄγκον γὰρ ἄλλως ὀνόματος τί δεῖ τρέφειν/ μητρῷον, ἥτις μηδὲν ὡς τεκοῦσα δρᾷ; “Why should she nourish in vain the dignity/weight of the maternal name who acts nothing like a mother?”.

  Michael Weiss (11) Hom. Od. 21.61 ὄγκιον, ἔνθα σίδηρος/ κεῖτο πολὺς καὶ χαλκός. a chest in which there was much iron and bronze

This describes the box that the servants carry for the ax competion. It doesn’t contain the arrows which are obviously in the φαρέτρη carried by Penelope herself. Thus it is a container of weighty things.23 No doubt related to this form is the Hesychian gloss ὀγκίαι ·θημῶνες χώματα σιδηροθήκη “heaps, mounds, iron chest”. Jouanna is quite correct that in the technical language of the Hippocratic corpus ὄγκοs clearly means specifically “volume” or “swelling” and not βάρος “weight” with which it is contrasted. But non-scientifically minded speakers must constantly have conflated volume and weight in ancient Greece as we do today and the word “ounce” and its ancient ancestors are perfect examples of exactly this lack of precision. An ounce is a measure of weight and a measure of volume. This very point was made by Irigoin (Jouanna 1985: 61–62) in a comment made immediately after Jouanna’s oral presentation: “Or toute charge comporte deux éléments: son poids et son encombrement.” Thus the original meaning of ὄγκος need not have been “swelling”, but “delimited amount of matter taking up space and possessing mass”.

 The root *h2enk- “bend” Let us next address the question of etymology. Jouanna rejects the traditional connection of ὄγκοs with ἐνεγκεῖν and prefers to see ὄγκοs as another semantic specialization of the noun that also resulted in Hom. ὄγκοs “barb”. He reconstructs a noun *h2ónkos from the root *h2enk- “bend”. He proposes that from “bend” a verbal abstract *h2ónkos could have developed to both “barb” and “swelling” and from “swelling” all the other meanings of ὄγκοs could have arisen. But the semantics of the root *h2enk- are much more in the area of “bend”, “hook” than “be convex”, and, though in the abstract one can image that “bend” or “curve” could develop the meaning “swelling”, the fact is that the root *h2enk- does not ever develop such a meaning. Ved. añc-, the sole certain verbal form of this root, refers to the bending of limbs at joints (10.15.6 ā́cyā (← ā́ ácyā) jā́nu) “having bent your knee”) the bending of tree (5.78.6), the bend 23 See Mader s.v. LfgrE iii: 486–487.

Latin uncia à la Heron  

ing of a chariot basket by an artisan (10.119.5) and later develops the meaning “draw water” via the idea of bending of the arms in drawing water.24 The nominal derivatives refer either to a sharp or gentle change in direction and fall into a number of basic categories:25 (1) “bent thing”: Ved. aṅká- “hook”, YAves. aka- “hook”, Gk. ἄγκιστρον “fish hook”, ἄγκυρα “anchor”, Lat. uncus “hook”, OIr. écath “fish hook”, OHG ango “sting, door hinge”, OE onga “sting”, ON angi “prickle”, Old and dial. Lith. ánka “loop in a rope”, Serbian CS ǫkotь “hook”, ORu. ukotь “claw, anchor”, ON áll, óll “sprout” < *anhula-, Skt. aṅkura- “sprout”,26 TB āṅkär “tusk”, TA āṅkari “tusks”,27 TA añcäl “bow”28 (2) “bendable thing”: YAves. ąxnah- “reins”, Gk. ἀγκύλη “strap”,29 ON ǫ́ l, ál f. “belt, strip”, OE ōl-(þwong) “halter, snare, strap” < *anhulō- ”strap” (3) “bendable body part”, i.e., “elbow” “bent arm”, “hand”, or “knee”: ἀγκάλη “bent arm”, ἀγκών “elbow”, Lat. ancus “with bent arm”, MW angad “hand”, craf-anc “paw”, OCS paǫkŭ “spider” < “with bending limbs” (4) bending terrain: Ved. áṅkas- “bend of a path” (RV 4.40.4), áṅkasá- “race course” (RV 4.40.3),30 ἄγκος “valley”, ON angr “bay”

It is clear that the root *h2enk- does not mean “be convex” and does not develop such a meaning in Greek or elsewhere and therefore is not a suitable starting point for ὄγκος, even on the assumption that the original meaning of the word was “swelling”.

 24 Hoffmann 1975: 165. 25 Obviously, the lines between categories can be drawn differently. 26 Referring probably to the hooked shape of a germinating seed before the true leaves escape the cotyledon. Jouanna cited the meaning “tumor” for aṅkura- but what is actually attested is the compound māṃsāṅkura- “flesh sprout” first in Suśruta, which shows that the meaning “tumor” is a metaphor. 27 On these see Pinault 2011: 162–163. 28 Cf. ἀγκύλος “bent” specifically of τόξα in Homer. 29 This word also means “hook for a door” and “bending of a limb” and so could go in either (1) or (3) as well. 30 See Jamison’s commentary ad loc. http://rigvedacommentary.alc.ucla.edu/?pageid=19, accessed March 31st, 2023.

  Michael Weiss

 *h2ónk̑os “what is taken”, “capacity” There is a better way to account for the meanings of ὄγκος while maintaining connection with *h1nek̑-. García Ramón 1999 has argued convincingly that the original meaning of this root was “take”. This meaning can be seen best in Tocharian where the verbal root eṅk-, which has not entered into a suppletive relationship with the stem *bher-, preserves precisely this meaning.31 The Schwebeablauting barytone thematic noun meaning “share” is preserved in Indo-Iranian (Ved. áṃśa- “portion, share”, OAves. ąsa- “party”).32 These forms are derived by Mayrhofer (1992: 36) from the root *h2nek̑- “reach” “attain”, so the share would be “what is attained” and that is possible, but it is more straightforward to derive the meaning from *h1nek̑- in its original meaning “take”. Cf. PDE take “that which is taken or received in payment, or as the proceeds from a transaction or piece of business; (U.S. colloquial) one’s personal income or earnings.” (OED s.v. take no. 4a). The semantics of Greek ὄγκος, on the other hand, have traditionally been hard to square with the Indo-Iranian evidence. The standard view connects the meaning of ὄγκος directly with the synchronic meaning of the Greek verb ἐνεγκεῖν “to carry, bear” on the assumption of a development “what is carried” > “burden” > “weight”. But this account is problematic for two reasons. First, as Jouanna correctly points out, there is good evidence that the Greek noun did not mean “what is carried” or “weight”. Therefore it would be preferable to find an original meaning that might have been specialized to “delimited amount of matter taking up space and possessing mass”. Second, it is attractive to derive the Greek form and the Indo-Iranian forms from a common and inherited protoform *h1ónk̑os precisely because they share the feature of Schwebeablaut. It is unlikely that the conditions for creating the Schwebeablauting forms would have existed so late in the history of the two branches. The heart of this root *h1nek̑-’s averbo is the aorist — it doesn’t make an old present — and the most archaic aorist form is the root aorist continued in the TB subjunctive eṅk-tär “will take, seize” < *h1n̥ k̑-. The Schwebeablauting forms are the result of an erroneous full-grade insertion based on the zero-grade allomorphs. If the root aorist was middle only, as are its descendants in TB, there would have been no clear evidence for where the full-

 31 See Malzahn 2010: 538–539. 32 The Proto-Indo-Iranian word was borrowed into some stage of Uralic to eventually give Finn. and Eston. osa “part” etc. See Holopainen 2019: 170–171 for discussion. I thank Martin Kümmel and Mikhail Zhivlov for the reference and advice about the interpretation of these forms.

Latin uncia à la Heron  

grade was to be located.33 Therefore it is not attractive to have to derive the meaning of the Greek noun from a secondary sense of the root *h1nek̑- viz. “carry”, which probably developed from the semantic influence of the suppletive partner *bher- (φέρειν). One could assume that the original meaning of the verbal abstract *h1ónk̑os “what is taken” shifted in Greek as the semantics of the verb shifted but that is less than ideal.34 We can solve this problem if we assume that Gk. ὄγκος is indeed inherited from PIE *h1ónk̑os. In Indo-Iranian the passive reading “what is taken” was specialized as “one’s take” but in Greek “what is taken” came to refer to the capacity (from Lat. capācitās ← capere “to take”)35 or volume of a container and from here the various attested meanings derive. Sophron’s κύμβου ὄγκον ἔχων would then be quite close to the original meaning. Returning to ὀγκία we may propose that it is a substantivization of the adjectival *onkii̯o- “pertaining to mass or volume”.36 That a word for mass or volume should be specialized in the meaning of a specific measure is easy to parallel. Compare PDE pound, originally from lībra pondō “a pound by weight”, or Welsh pwys “pound” from Lat. pē(n)sum. In Sicilian Greek the word became part of the duodecimal Italic system and made its way quite early to Latin where it became, with the raising of o to u before a velar nasal, uncia.37 The creation of the common Italic system of measurement was a joint production of Italic and Greek speakers.38 The unit lībra ~ λίτρα was supplied by Italic, but the name for the twelfth was purely Greek.  33 This account is not completely satisfying because the correct locus full-grade did apparently survive in Greek at least (Hesych. κατήνοκα and Myc. e-ne-ka). 34 This is the line of explanation explored by García Ramón 2020: 91–95, who kindly sent me his paper. 35 An interesting parallel is Lith. síekas, siẽkas, saĩkas Latv. sieks “measure of grain” derived from the verb síekti “to take” “to encompass”. These nominal forms are cognate with PGmc. *saiha“measure of quantity” > ON sár “cask”, OE sā “ladle”. Cf. also the Slavic words for “volume” Russ. объем, Cz. objem etc. from jęti “to take” though these may be calques on NHG Umfang. 36 Note that uncia is not just a unit of weight but also of volume and even of area. Cf. Plautus’s use of the specification pondo. 37 See Weiss 2020: 151. The short vowel is confirmed by the Romance reflexes: Ital. oncia Fr. once, Prov. onsa, Sp. onza, Po. onça. 38 H. Chantraine 1962: 56 argued that since the duodecimal system is not found elsewhere in the Greek world and since it is unlikely that the Siciliot Greeks would have taken terms for their system of measurement from distant Rome the system must have originated in Sicily among the Sicels. I doubt we can insist on the specifically Sicilian origin of the system as firmly as Chantraine wished. The duodecimal metrological system may have been another aspect of the Italic cultural koine. The Greeks need not have adopted a native word to describe the native unit. Cf. Lith. svaras or Finn. naula for “pound”. It is likely that the fractional names are Greek in origin. See Cassio in this volume.

  Michael Weiss

Bibliography Agostiniani, L. (1992), “Les parlers indigènes de la Sicile pré-grecque”, Lalies 11, 125–157. Beekes, R.S.P. (2010), Etymological Dictionary of Greek, Leiden. Bréal, M./Bailly, A. (1885), Dictionnaire étymologique latin, Paris. Chantraine, H. (1962), “Bemerkungen zum ältesten sizilischen und römischen Münzwesen”, Jahrbücher für Numismatik und Geldgeschichte 12, 51–64. Chantraine, P. (2009), Dictionnaire étymologique de la langue grecque, 2nd ed., Paris. Coleman, R. (1992), “Italic”, in: J. Gvozdanović (ed.), Indo-European Numerals, Berlin, 389–446. Corssen, W. (1870), Über Aussprache, Vokalismus und Betonung der lateinischen Sprache, vol. 2, 2nd ed., Leipzig. Curtius, G. (1862), Grundzüge der griechischen Etymologie. Zweiter Theil, Leipzig. Curtius, G. (1879), Grundzüge der griechischen Etymologie, 5th ed., Leipzig. Ernout, A./Meillet, A. (1985), Dictionnaire étymologique de la langue latine: histoire des mots, revised 4th ed., Paris. Frisk, H. (1960), Griechisches etymologisches Wörterbuch, Heidelberg. García Ramón, J.L. (1999), “Zur Bedeutung indogermanischer Verbalwurzeln: *h2nek-̑ ‘erreichen, reichen bis’, *h1nek-̑ “erhalten, (weg)nehmen”, in: J. Habisreitinger/R. Plath/S. Ziegler (eds.), “Gering und doch von Herzen”. 25 indogermanische Beiträge Bernhard Forssman zum 65. Geburtstag, Wiesbaden, 47–80. García Ramón, J.L. (2020), “Proto-indo-european continuity in Anatolian after the split: when Hittite and Luwian forms require a proto-indo-european source”, in: M. Serangeli/ T. Olander (eds.), Dispersal and Diversification. Linguistic and Archaeological Perspectives on the Early Stages of Indo-European, Leiden, 77–106. Garnier, R. (2016), La derivation inverse en Latin, Innsbruck. Hoffmann, K. (1975), “Materialien zum altindischen Verbum, 1–10”, in: J. Narten (ed.), Aufsätze zur Indoiranistik, vol. 1. Wiesbaden, 162–182. Originally published in ZvS 79, 171–191. Holopainen, S. (2019), Indo-Iranian Borrowings in Uralic. Critical Overview of the Sound Substitutions and Distribution Criterion, Ph.D. diss. University of Helsinki. Klug, F. (2017), Der altlateinische Dialekt von Praeneste, Ph.D. diss. Ernst-Moritz-ArndtUniversität, Greifswald. Jouanna, J. (1985), “Le mot grec ὄγκος ou de l’utilité d’Hippocrate pour comprendre les textes poétiques”, Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres 129, 31–62. Leumann, M. (1959), “Die Adjektiva auf -īc̆ ius”, in: Kleine Schriften, Zürich, 3–35. [Originally published in Glotta 9, 1918, 129–168.] MacDonald, G. (1895), Catalogue of Greek Coins in the Hunterian Collection, vol. 1, Glasgow. Malzahn, M. (2010), The Tocharian Verbal System, Leiden. Mayrhofer, M. (1992), Etymologisches Wörterbuch des Altindoarischen, vol. 1, Heidelberg. Pinault, G.-J. (2011), “Let us now praise famous gems”, Tocharian and Indo-European Studies 12, 239–242. Rosen, K.M.D. (1964), “Latin uncia”, Language 40, 21–22. Schmoll, U. (1958), Die vorgriechischen Sprachen Siziliens, Wiesbaden. Tucker, T.G. (1931), A Concise Etymological Dictionary of Latin, Halle. de Vaan, M. (2008), Etymological Dictionary of Latin and the Other Italic Languages, Leiden. Valente, S. (2015), The Antiatticist: Introduction and Critical Edition, Berlin.

Latin uncia à la Heron  

Vaniček, A. (1874), Etymologisches Wörterbuch der lateinischen Sprache, Leipzig. Walde, A./Hofmann, J.B. (1954), Lateinisches etymologisches Wörterbuch, vol. 2, 3rd ed., Heidelberg. Willi, A. (2008), Sikelismos. Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.-5. Jh. v. Chr.), Basel. Weiss, M. (2020), Outline of the Historical and Comparative Grammar of Latin, 2nd ed., Ann Arbor.

List of Contributors Luca Alfieri is Associate Professor in Linguistics and Sociolinguistics in the Department of Human Sciences of the University Guglielmo Marconi, Rome. He is author of several papers in historical linguistics, with special focus on history of linguistic terminology, ancient grammatical theories, language change, historical sociolinguistics, and parts of speech typology. He wrote a monograph on the history of word-formation (La morfologia derivazionale e il problema del tempo, Mimesis, 2023) and is currently principal investigator in the Prin project “The lexicalisation of the adjective class in Indo-European and Semitic” (2023–2025). Giovanni Boffa is Laboratory Manager and Contract Professor in History and Economy of the Ancient World at the Department of Cultural Heritage of the University of Salento, Lecce. His works mainly focus on Greek archaic History and Epigraphy, with special attention to the Greek colonial environments and to the social implications of the adoption and the use of writing within the Greek world and the Italic cultures of Southern Italy. Carla Bruno is Associate Professor at the University for Foreigners of Siena, where she teaches General and Historical Linguistics. Her research focuses on morpho-syntactic aspects of ancient Indo-European languages (especially ancient Greek) from both a synchronic and a diachronic perspective, with special attention to linguistic variation and change. Albio Cesare Cassio is Emeritus Professor of Classical Philology at the University of Rome “La Sapienza”. After a period of work on Aristophanes, culminating in a book on the Peace (Commedia e Partecipazione, Naples 1985), he has concentrated on the study of the Greek language and its dialects in an Indo-European perspective, publishing a long series of articles in Italian and foreign journals, and organizing conferences on these subjects (Katà Diálekton. III Colloquio Internazionale di Dialettologia Greca, Naples 1997). He has especially studied the most recent phases of the Homeric language and the dialects of the Greek colonies of Southern Italy and Sicily (The Language of Doric Comedy in A. Willi [ed.], The Language of Greek Comedy, Oxford 2002). He is the editor (and in part the author) of a book on the history of the Greek literary languages (Storia delle lingue letterarie greche, 2nd edition, Milano 2016). Emilio Crespo is Professor Emeritus of Greek Philology at the Universidad Autónoma de Madrid and Chairman of the Steering Committee of the Pastor Foundation for Classical Studies (Madrid). He was vice rector of the Universidad Autónoma de Madrid and Head of its Department of Classical Philology. Doctor honoris causa from the Aristotle University of Thessaloniki, he is, among other publications, co-author of Sintaxis del griego clásico (2003), associate editor of Encyclopedia of Ancient Greek Language and Linguistics (2014) and author of a well-known Spanish translation of Homer’s Iliad (1991, revised 2019). Panagiotis Filos is Associate Professor of Historical and Indo-European Linguistics at the University of Ioannina, while he has also taught Linguistics and Classics at various other universities in Greece and abroad. He has received BA degrees in Archaeology and Classics from the University of Ioannina, while he also holds an MPhil in General Linguistics and Comparative Philology as well as a DPhil in Comparative Philology from the University of Oxford. He is,

  List of Contributors inter alia, Associate Editor of the Encyclopedia of Greek Language and Linguistics (forthcoming) and co- author of the annual review of Modern Greek linguistics in This Year’s Work in Modern Language Studies. He has published particularly on Graeco-Latin bilingualism as well as on Ancient Greek dialectology, onomastics and etymology. Paloma Guijarro Ruano is assistant professor at the Universidad Complutense of Madrid. Her general academic interests focus on the history of Ancient Greek, from Mycenaean to Modern Greek, and on linguistic and cross-dialectal contacts. Since her doctorate, she has concentrated on the epigraphical poetry of the Archaic period. She has also extended her research to the analysis of language contact in the northern borders of Greece and the study of the lexicon of slavery thanks to subsequent European post-doctoral fellowships. Mark Janse is BOF-ZAP Research Professor in Ancient & Asia Minor Greek at Ghent University and a Member of the Academia Europaea. He is a former Visiting Fellow of All Souls College (Oxford), the Onassis Foundation (Greece) and Harvard’s Center for Hellenic Studies (DC), and a former Onassis Senior Visiting Scholar at Harvard, Princeton, Stanford and the University of Arizona. He gave various named lectures: Gennadeion (Athens, 2008), F.B.J. Kuiper (Leiden, 2008), Johannes Sundwall (Athens, 2011), Naylor Memorial (Ohio State, 2021), Gaisford (Oxford, 2022) and Jeremie Septuagint (Cambridge, 2024). His research covers the entire history of the Greek language from Homer via Classical, Biblical and Medieval Greek up to the Modern Greek dialects of Asia Minor. Sara Kaczko is Assistant Professor in Classical Philology at the Department of Classics of the University of Rome “La Sapienza”. She is the author of several papers on Greek literary languages and dialects; Greek epigraphy; the relationship between material object and text; classical philology. Her monograph Archaic and Classical Attic Dedicatory Epigrams. An Epigraphic, Literary, and Linguistic Commentary, De Gruyter, Berlin/Boston 2016 was awarded the “Prix d’épigraphie” by the AIEGL – Association International d’Épigraphie Grecque et Latine. Felicia Logozzo is Associate Professor of General and Historical Linguistics at the University for foreigners of Siena. Her favourite research subjects are: Greek and Latin morphosyntax; GreekRomance texts; the languages of the Vulgata and the Septuaginta. Graduated in 2008 in Linguistics at the University of Rome “Tor Vergata”, she received her PhD in ‘Linguistic history of Eurasia’ at the University of Macerata in 2012. After being involved in two research projects as Postdoctoral fellow at the University of Rome “Tor Vergata” (“SCRIBE - Short writings: Linguistic simplification, social inclusion: models and applications” (2013–2014) and “The authority of words. The forms of Roman preceptive discourse between preservation and change” (20152016), she worked on the project MuMiL-EU (Multilingualism and minority languages in Ancient Europe – HERA Uses of the past), as junior fixed-term Researcher, at University for foreigners of Siena. At the same university she was senior fixed-term Researcher until 2022. Mario Lombardo is Professor Emeritus in Ancient Greek history and epigraphy at the Salento University (Lecce), where he has been full Professor until 2018. He has been Assistant Professor of Greek history and then Associate Professor of Greek epigraphy at the Scuola Normale Superiore di Pisa (1977–1987) and Professor of Greek antiquities and Greek epigraphy at the Scuola Archeologica Italiana di Atene (2002–2016). He is the author of many papers on Greek

List of Contributors  

epigraphy and Greek history. He has investigated especially the relationships between the Greeks and the peoples with whom they came to interact in the “colonial” contexts of Southern Italy, Sicily, and the Adriatic. He is Corresponding Member of the German Archaeological Institute and President of the scientific committee organizing the annual Taranto Conferences on Magna Grecia. Marco Mancini (Rome, 1957) is full professor of Historical Linguistics at “Sapienza”, University of Rome. He is a national member of the prestigious Accademia dei Lincei and Ordinary Member of the Istituto Nazionale di Studi Etruschi e Italici. He is co-director of the “Archivio Glottologico Italiano”, the oldest Italian journal of historical linguistics, and member of the scientific boards of several linguistic journals. Mancini is a specialist in languages and philologies of the ancient world, in both Western (Latin, Romance, Italic, Germanic) and Oriental fields (Iranian, ancient and Medieval pre-Islamic, Hebrew and Middle Aramaic). He has written volumes and papers concerning Ancient Latin and Greek, Italian dialects, Judaeo-Italian varieties, Old and Middle Persian, Parthian, Oscan, Faliscan, Syriac and Judaeo-Aramaic; he is also a specialist of linguistic historiography and ethnography of writing. Simona Marchesini is the scientific coordinator of Alteritas, a research organisation based in Verona. A specialist in fragmentary languages of ancient Italy and archaeologist, she has lectured at the universities of Tübingen and Verona. She has been research fellow at the Scuola Normale Superiore in Pisa and at Trinity College in Dublin. She has published eight monographs (including Prosopografia Etrusca. Gentium Mobilitas, Rome 2007 and Le lingue frammentarie dell’Italia antica, Milan 2009) and almost one hundred articles in international scientific journals. She is currently scientific coordinator of two European projects and is head the editorial board of Alteritas Academy Press. Liana Tronci is Associate Professor of General and Historical Linguistics at the University for Foreigners of Siena. Her research focuses on the morphosyntax of Ancient Greek and Latin and its interactions with the lexicon, information structure, and pragmatics. She published a book on passive aorists in Ancient Greek (Guerra 2005) and is co-editor (with Camille Denizot) of the book “Building modality with syntax. Focus on Ancient Greek” (De Gruyter 2023). She has devoted many recent publications to the syntax of Biblical Greek and Biblical translations from Greek to Latin, with special attention to participial constructions and pseudo-coordination. Johan Vandewalle is Professor of Turkish and Head of the Turkish section at the Department of Translation, Interpreting and Communication at Ghent University and an acknowledged polyglot. He has degrees in Civil Engineering and Architecture, Slavic Languages, Oriental Languages and General Linguistics, and a PhD in Linguistics with a dissertation on motion verbs in Russian. He has published on Russian, Turkish, Uzbek and, together with Mark Janse, Cappadocian linguistics. Together with his wife he founded “Orientaal vzw”, a centre for Oriental languages and cultures, where he has taught Turkish and Arabic. They developed an original teaching method for Turkish, published as Haydi Türkçe Konuşalım 1 & 2. He has been awarded several prizes for Turkish Language Teaching and for his extensive knowledge of a variety of languages.

  List of Contributors Michael Weiss (Ph.D., Cornell, 1993) is a professor in the Department of Linguistics at Cornell University in Ithaca, New York. He is the author of Language and Ritual in Sabellic Italy (2009), Outline of the Historical and Comparative Grammar of Latin (2nd ed. 2020) and Kuśiññe Kantwo: Elementary Lessons in Tocharian B (2022) as well as articles on various Indo-European topics. He has taught in the Classics departments at Yale, and the University of North Carolina at Chapel Hill and has been a visiting lecturer at the Program in Indo-European Studies, UCLA, the Center for Hellenic Studies, The Leiden Summer School in Languages and Linguistics, the University of Vienna, and the University of Ljubljana.

General Index Accidentia nominum (πάθη τῆς λέξεως) 4, 5, 10, 22, 24 Adopted dialect in Macedonian 73–74 Aegean Thrace 115, 117, 120, 123–133 Affixation, suspended 148–155 Alphabetic writing 216–217 ἀνίστημι 207–208, 211 ἀπέρχομαι 192, 193, 197, 201, 204, 205, 211 Attic dialect 116, 120, 123–126 Compositio (σύνθεσις) 8, 9, 13, 16 Compound – dvandva 145–147 – phrasal 139–158 – verbal 139–158 Corruption of nouns (παραφθορά) 4–5 Declinatio (κλίσις, πτῶσις) 6, 8–9, 17 Decompositio (παρασύνθεσις) 13–14 Derivatio, sono or sensu (παραγωγή, κατὰ φωνήν or κατὰ σημασίαν) 13, 16, 18, 21, 24 Derveni papyrus 71, 72 Dexylos’ inscription 222 Dialect – contact 115, 116, 123–126 – vernacular in Macedeonia 73–75 Dialectal – convergence 61, 71, 74 – features common to Aigeai and Pella 66–72 – hybrid documents in Macedonia 72 Dialects in the inscriptions found at Aigeai and Pella 67–69 Discourse vs. story 191, 204, 207, 212 Doric and Northwest Doric varieties 70 E, H, I, EI variations in Macedonian 69–70 ἐγείρω 192, 207–209, 211 Epiphany vs. episode dream 33, 38 ἔρχομαι 185, 186, 191, 192, 194, 195, 196 Etymology, ancient theories of 3, 4, 6, 10, 21, 23 Figura or forma (σχῆμα) 8, 11–17 First-person dreams accounts 30, 35–37, 43, 45 Frame Semantics 283, 285 Geographical variations in Macedonian documents 61–66 Greek – -ᾶς, -ᾶντος suffixes 54–56 – Asia Minor 140, 157–158 – Cappadocian 139–158 – Medieval 140–144, 155–157 – Modern 139–147 https://doi.org/10.1515/9783110779684 015

  General Index – Pharasiot 141, 155–157 – Pontic 141 Greek-Roman bilingualism 2, 22 Grotta Poesia, Melendugno, Lecce 218, 227, 283–289, 291, 296 Impositor nominum (νομοθέτης) 3–5, 17 Ionic (Cycladic Ionic, Asia Minor Ionic) 116, 122 Interference of/from Italic languages with Greek 163, 165, 177, 179, 180, 181 Italic measurement system 53–54, 299, 309 καταβαίνω 192, 205, 211 Katochoi of the Serapeum 31 Koine 62, 117–118, 120, 125–123 Koineization 61, 117, 129–132 Latin adjectives inflected like participles 57 Latin borrowings from Greek 56–57 Latin triens 57 Lecce (Italy), “Sigismondo Castromediano” Museum 219 Lecce, Masseria Fani, Salve, abecedarium 224 Leveling 74–75 Linguistic variations in lower Macedonian 61–63 Macedonian 59–78 Messapians 164, 215–219 Messapic – eipeigrave/ipigrave 283, 285, 292–294 – hezzan 289 – ipi(-eipi) 293 – Kezareihei 292 – name system 294 Motion verbs 186–189, 193, 203–205, 210 Muro Leccese (inscribed pillar) 226 New Testament Greek 185–210 nuncupatio voti 296 Oscan – language 177–179, 251–281 – pukam = “visible monument” vel sim. 253, 272 – pukam, its etymology 251–281 – seg(ú)núm and Latin signum 269–274 Παιστανο coins 163, 165, 180, 181 Pattern replication 139–158 P(roto-)I(ndo-)E(uropean) *gerbh 293 PIE – *ĝher-/ĝhr 294 – *kwVḱ 262–263 – *kwVḱand its Greek outcomes τέκμωρ/τέκμαρ 256–259 – *kwVḱwith oor ōgrade, Iranian outcomes 263–269 – *sedyā̆m 289 – -vperfect 294–296 Ποσδαν 163, 165, 168, 169, 171, 172, 173, 175, 177, 179, 180, 181

General Index  

Poseidon, variants of the name in Greek 169–176 Poseidonia / Paestum 163, 164, 165, 170, 173, 179, 180 Positio, prima or thema (θέσις, πρώτη) 11, 12, 14, 15, 17, 18, 19, 23 Posture verbs 187, 189, 190, 192, 210 πορεύομαι 192, 193, 197, 201–205, 211 προσέρχομαι 192, 205, 211 Revelative evidentiality 30, 39 serialization 187–190, 197, 199 South Italian red-figure vases 163, 164, 165, 176 Species (εἶδη) 11, 12, 13, 14, 16, 22 Suffixes, Greek 54–56 Syncope 173, 174, 176, 177, 178, 179, 180, 181 Syracusan coinage 53–54 τέκμωρ/τέκμαρ and the Indo-Iranian forms for “eye”, “gaze” 259–262 Terminatio, extremitas, finis (παραγωγή, κατάληξις, τέλος) 16, 17, 19, 20, 23 Torricella inscription, Taranto 222 Turkish 139–140, 148–158 uncia, etymology 308–309 ὑπάγω 155, 187, 192, 193, 197, 198, 199, 200, 202, 203, 204, 208, 209, 211 Vaste, abecedarium 224, 243 Vaste, cippi 226 Vereto, inscribed columns 226–227 Vix Crater 224 Word-formations based on numbers 56