Variation and Change in Ancient Greek Tense, Aspect and Modality, (Amsterdam Studies in Classical Philology, 23) (English and Ancient Greek Edition) [Bilingual ed.] 9004311645, 9789004311640

Citation preview

Variation and Change in Ancient Greek Tense, Aspect and Modality

Amsterdam Studies in Classical Philology Series Editors Irene J.F. de Jong Caroline H.M. Kroon

Editorial Board Rutger J. Allan Mark A.J. Heerink

volume 23

The titles published in this series are listed at brill.com/ascp

Variation and Change in Ancient Greek Tense, Aspect and Modality Edited by

Klaas Bentein Mark Janse Jorie Soltic

leiden | boston

Library of Congress Cataloging-in-Publication Data Names: International Colloquium on Ancient Greek Linguistics (7th : 2011 : Rijksuniversiteit te Gent) | Bentein, Klaas, editor. | Janse, Mark, 1959- editor. | Soltic, Jorie, 1988- editor. Title: Variation and change in Ancient Greek tense, aspect and modality / edited by Klaas Bentein, Mark Janse, Jorie Soltic. Description: Leiden ; Boston : Brill, 2017. | Series: Amsterdam studies in classical philology, issn 1380-6068 ; volume 23 | Based on the Seventh International Colloquium on Ancient Greek Linguistics, which was held at Ghent University, 21–23 September, 2011. | Includes bibliographical references and index. Identifiers: lccn 2017011864 (print) | lccn 2017018281 (ebook) | isbn 9789004315358 (E-book) | isbn 9789004311640 (hardback : alk. paper) Subjects: lcsh: Greek language–Temporal constructions–Congresses. | Modality (Linguistics)–Congresses. | Linguistic change–Congresses. Classification: lcc pa379 (ebook) | lcc pa379 .i58 2017 (print) | ddc 485–dc23 lc record available at https://lccn.loc.gov/2017011864

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface. issn 1380-6068 isbn 978-90-04-31164-0 (hardback) isbn 978-90-04-31535-8 (e-book) Copyright 2017 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill nv incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change. This book is printed on acid-free paper and produced in a sustainable manner.

Contents Preface vii List of Figures viii List of Tables ix List of Abbreviations x List of Contributors xi 1

Introduction 1 Klaas Bentein, Mark Janse and Jorie Soltic

2

Subjunctive and Optative in Herodotus’ Purpose Clauses as Relative Tense Markers 9 Antonio Lillo

3

Variation in Expressing Temporal and Aspectual Distinctions in Complement Clauses: A Study of the Greek Non-Literary Papyri of the Roman Period 22 Jerneja Kavčič

4

Syntactic Variation with Verbs of Perception and the ‘Oblique Imperfect’: Once Again on Aspect, Relative Time Reference and Purported Tense-Backshifting in Ancient Greek 56 Julián Méndez Dosuna

5

The Gnomic Aorist in Hesiod 84 Gerry C. Wakker

6

The Imperfect Unbound. A Cognitive Linguistic Approach to Greek Aspect 100 Rutger J. Allan

7

Aktionsart, Aspect and Category Change in the History of Greek Amalia Moser

8

Ὤφελ(λ)ον in Ancient Greek Counterfactual Desiderative Sentences: From Verb to Modal Particle 158 Antonio R. Revuelta Puigdollers

131

vi

contents

9

Modalized Future and Scheduled Present in Coan Inscriptions 189 Marina Veksina

10

‘High’ and ‘Low’ in Medieval Greek 219 Geoffrey Horrocks

11

Confusion of Moods in Greek Private Letters of Roman Egypt? 242 Martti Leiwo

12

Register Variation and Tense/Aspect/Mood Categories in Ancient Greek: Problems and Perspectives 261 Andreas Willi Index locorum 287 Index nominum et rerum 297

Preface This volume has its origins in the Seventh International Colloquium on Ancient Greek Linguistics, which was held at Ghent University, 21–23 September, 2011. The topic of the conference was ‘Variation in Grammar and Discourse’, with a focus on word order and tense-aspect phenomena. Since most participants concentrated on tense and aspect, we decided to publish a unified volume on linguistic variation and change in tense, aspect and modality. The International Colloquium on Ancient Greek Linguistics has a rich tradition: it was first organized by the Department of Classics at the University of Amsterdam in 1986, and then every fourth or fifth year thereafter. A variety of topics included the language of Homer (Grenoble), particles (Amsterdam), complement clauses (Saint-Étienne), word classes (Madrid), and discourse cohesion (Groningen). Up until 2011, the International Colloquium on Ancient Greek Linguistics ran parallel with another international series, the Incontro Internazionale di Linguistica Greca. On the initiative of the Ghent organizers, it was agreed that it would be in everyone’s interests to join forces, and to run a new unified series in parallel with the International Colloquium on Latin Linguistics. The first of these was held in Rome, 23–27 March 2015. It is our pleasure to thank a number of organizations for financial support: the Commission for Scientific Research of the Faculty of Arts & Philosophy at Ghent University, Brill, and the Embassy of Greece in Brussels. Various colleagues of the former Department of Latin and Greek at Ghent University provided invaluable practical help during the conference: many thanks to Lieven Danckaert, Sylvie Geerts, Delphine Nachtergaele, Maarten Taveirne, and Berenice Verhelst. Cleo Janse and Margaret Whibley provided assistance with the copy-editing of this volume, for which we are grateful. Any inaccuracies remain our own responsibility. Klaas Bentein, Mark Janse and Jorie Soltic Ghent, December 2016

List of Figures 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8.1 9.1 9.2

State (− telic, + durative, − dynamic) 113 Activity (− telic, + durative, + dynamic) 113 Accomplishment (+ telic, + durative, + dynamic) 114 Achievement (+ telic, − durative, + dynamic) 114 Climax 115 Elbow and hand: scope, base and profile 117 Present vs. aorist: (un)bounded within temporal scope 118 Perfect stem: resultant state profiled 119 Accomplishment in progress 121 Profiled ‘preface’ of achievement 122 Initial boundary profiled: ingressive aorist 123 Atelic situation viewed in its totality: complexive aorist 124 Accomplishments unbounded: continuing relevance 125 Comrie’s (1976:25) classification of aspectual opposition 135 ‘Unexpected’ imperfectives in Herodotus 149 ‘Unexpected’ imperfectives in Thucydides 149 ‘Unexpected’ imperfectives in Xenophon 149 ‘Unexpected’ imperfectives in Polybius 151 ‘Unexpected’ imperfectives in Diodorus 151 ‘Unexpected’ imperfectives in Arrian 151 ‘Unexpected’ imperfectives in Procopius 152 Uses of the imperfective 153 Grammaticalization process of ὤφελον 186 Modalization of future 204 From generics to deontic modality 216

List of Tables 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.1 4.2 6.1 6.2 7.1 7.2 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 8.10 8.11 8.12 9.1 12.1

Frequency of the verb εἰμί 33 Frequency of the verb ἔχω 33 Frequencies of present, perfect, aorist, and future infinitives used in DInf Aspect frequencies in private letters according to Porter & O’Donnell (2010) 40 Tense frequencies in non-literary papyri (Mandilaras 1973: 59–60) 40 Tense frequencies in ὅτι clauses (non-literary papyri) 44 Tense frequencies in ὅτι clauses (New Testament) 45 Maximum/minimum frequencies of aorist/future infinitives 47 DInf in private letters 47 Aspect in the Ancient Greek verb 57 Absolute time reference in the indicative 60 πλέω and χωρέω: imperfect vs. aorist indicative (3sg/pl) 102 Aktionsarten: semantic features 113 Aktionsart categories, based on Vendler (1957) 133 The indicative of the verb in Classical and Modern Greek 136 Diachronic evolution of ὀφείλω’s stems and tenses 162 Conjugation of ὤφελ(λ)ον 162 Possible and counterfactual desiderative constructions 165 Counterfactual desiderative constructions with ἐβουλόμην 167 Ὤφελον: construction a 168 Ὤφελον: construction b 168 Ὤφελον: construction c 170 Synoptic view of ὤφελον’s constructions 172 Sentence type and illocutionary force 178 Sentence type and negation 180 Analogical change of ὤφελον 184 Diachronic evolution of ὤφελον 185 Verbal moods and nominal style in Coan calenders 206 A sample study on the use of active perfects in Classical Greek 282

35

List of Abbreviations acc AcI agr aor aux dinf fut fut.pf imp ind inf ipf neg nom nt opt part pers pf plpf pres ptcl rtr RTTh SoA subj tam tlg v vp

Accusative Accusativus cum infinitivo Agreement Aorist Auxiliary Declarative infinitive clause Future Future perfect Imperative Indicative Infinitive Imperfect Negation Nominative New Testament Optative Participle Personal form Perfect Pluperfect Present Particle Relative time reference Relative tense theory State of affairs Subjunctive Tense-aspect-mood/modality Thesaurus Linguae Graecae Verb Verb phrase

List of Contributors Rutger J. Allan Rutger Allan (PhD University of Amsterdam, 2002) is a lecturer in Ancient Greek Language and Literature at the Free University Amsterdam and a researcher at the University of Amsterdam. His main research interests are Ancient Greek verbal semantics and discourse linguistics. Klaas Bentein Klaas Bentein (PhD 2012) is a post-doctoral research fellow at Ghent University. He wrote a dissertation on the history of verbal periphrasis, which has recently appeared as Verbal periphrasis in Ancient Greek: Have- and be- constructions (oup 2016). In his most recent research, he further explores the relationship between language and society, focusing on morpho-syntax in Greek documentary writing. Geoffrey Horrocks Geoffrey Horrocks is Professor of Comparative Philology at the University of Cambridge and a fellow of St John’s College. He has wide-ranging interests in theoretical and historical linguistics, but his principal focus has always been on the history and structure of Greek: Ancient, Medieval and Modern. Mark Janse Mark Janse is bof-zap Research Professor in Ancient and Asia Minor Greek at Ghent University. His research interests include Ancient and Modern Greek linguistics and dialectology; Homeric and Byzantine metre and versification; Ancient Greek profane and obscene language; language variation and change; and language contact and language death, with particular attention to GreekSemitic and Greek-Turkish language contacts. He is well-known for his research on Cappadocian (Asia Minor) Greek on which topic he has published extensively, including a grammar. His personal involvement with speakers of the language has earned him the honorary title ‘Ambassador of the Cappadocians’ and is the topic of a documentary film, Last Words (seriousFilm, 2014). Jerneja Kavčič Jerneja Kavčič is an associate professor of Greek at the University of Ljubljana, Slovenia. Her research interests concern Greek in all of its historical stages (Ancient, Byzantine, and Modern Greek) as well as linguistic theory (mostly syntax). She authored a volume on the syntax of the infinitive and participle in

xii

list of contributors

Early Byzantine Greek (2005); compiled the first Modern Greek–Slovenian and Slovenian–Modern Greek dictionary (2006); and wrote the first Modern Greek grammar in Slovenian (2011). Martti Leiwo Martti Leiwo is a senior lecturer at the University of Helsinki, Department of World Cultures, and an adjunct professor at the University of Turku, Department of Languages. He is also principal investigator of the project ‘Act of the Scribe: transmitting linguistic knowledge and scribal practices in GraecoRoman Antiquity’, funded by the Finnish Academy (2015–2019). He was professor of Greek Philology, University of Helsinki (2005–2006), leader of ‘Variation and change in Greek and Latin’, a subproject of Ancient Greek written sources, Centre of Excellence of the Academy of Finland (2006–2011), and director of the Finnish Institute at Athens, Greece (2008–2013). Antonio Lillo Antonio Lillo is Professor of Classics at the University of Murcia. His research interests focus on Ancient Greek dialectology and linguistics. He has also published on the Greek of the 19th century. Julián Méndez Dosuna Julián Méndez Dosuna (PhD 1983) is Professor of Greek Linguistics at the University of Salamanca. He has published a substantial number of articles on various issues in Ancient and Modern Greek linguistics, always endeavoring to bridge traditional philology and modern linguistic theory. His two current research interests are the phonology of the ancient dialects, especially those of the oracular lamellae of Dodona, and double entendres in Aristophanes. He is editor of the journal Minos. Amalia Moser Amalia Moser is Professor of Linguistics at the University of Athens. Her main research interests are the semantics and pragmatics of aspect and tense, and historical linguistics with emphasis on semantic change and grammaticalization. Antonio R. Revuelta Puigdollers Antonio R. Revuelta Puigdollers (Universidad Autónoma de Madrid) is an assistant professor of Ancient and Modern Greek at the Universidad Autónoma de Madrid and an official translator of Modern Greek (Spanish Ministry of Foreign Affairs and Cooperation). He is the author of several works on Greek

list of contributors

xiii

and Latin linguistics. His research interests include the syntax, semantics and pragmatics of the verb (valency, preverbation, reciprocity/symmetry, semantic functions, alternations), discourse particles, information structure (focus and topic) and subordination. Jorie Soltic Jorie Soltic studied classical languages at Ghent University and holds a PhD in Greek linguistics with a dissertation on ‘The Late Medieval Greek πολιτικὸς στίχος poetry: Language, Metre and Discourse’. Her research is focused on discourse linguistics, including issues such as information structure, intonation units and discourse markers, with particular attention to Late Medieval Greek. Marina Veksina Marina Vladislavovna Veksina is a research associate at the Humboldt University of Berlin (Excellence Cluster Topoi) and a post-doctoral fellow at the University of Basel. She works in the fields of Greek epigraphy and the history of the Greek language. Her research is focused on diachronic, regional, social and other variations of Greek, as well as on the development of epigraphic formulas. As part of the Topoi-team she is developing a digital edition of the early Christian inscriptions. Since 2013 she has been writing The grammar of Coan inscriptions. Gerry Wakker Gerry Wakker is Professor of Ancient Greek Linguistics at the University of Groningen (Department of Classics). Her main interests are the pragmatics and semantics of Ancient Greek. She graduated at the University of Amsterdam (PhD thesis: ‘Conditions and conditionals. An investigation of Ancient Greek’. Amsterdam 1994). She now mainly studies issues related to aspect, tense and discourse cohesion, e.g., the use of particles. Andreas Willi Andreas Willi is Diebold Professor of Comparative Philology at the University of Oxford. He is the author of The languages of Aristophanes: Aspects of linguistic variation in Classical Attic Greek (2003) and Sikelismos: Sprache, Literatur und Gesellschaft im griechischen Sizilien (2008) and has edited or co-edited, among other things, The language of Greek comedy (2002) and Laws and rules in IndoEuropean (2012). His research interests include ancient sociolinguistics, Greek dialectology, the interaction of language and literature in Ancient Greece, and Greek, Latin, and Indo-European comparative grammar.

chapter 1

Introduction Klaas Bentein, Mark Janse and Jorie Soltic

If structure is at the heart of language, then variation defines its soul. wolfram 2006:333

∵ 1.1

Setting the Scene

Variation was long considered a problem in linguistic theory, one from which attention needed to be deflected.1 Yet nowadays variation has become a central concept in variationist (socio)linguistics,2 a discipline which draws attention to its synchronic and diachronic importance, for variation is ubiquitous in language, and can be considered ‘the inevitable synchronic face of long-term change’ (Guy 2003:370). This variationist approach has mainly been adopted in relation to the spoken language, which can be most accurately observed, described and analysed. Written texts have received relatively little attention, with Labov (1994:11) describing historical linguistics as ‘the art of making the best use of bad data’. Scholars have come to realize, however, that written texts also form a valuable object of investigation and that research into spoken language faces its own methodological difficulties. As a result, ‘socio-historical’ linguistics has, in recent years, become established as a discipline.3 The main purpose of this volume is to explore linguistic variation in Ancient Greek texts, concentrating on one of the core areas of grammar: tense, aspect

1 See e.g., Dufter-Stark (2008:81): ‘La variation à l’ intérieur d’une langue historique fut considérée comme noise dont il fallait absolument minimiser l’influence’. 2 The term ‘variationist linguistics’ is broader than ‘sociolinguistics’, with the latter focusing specifically on socially related variation. The two terms are often combined as ‘variationist sociolinguistics’. 3 Romaine (1982) is a pioneering work.

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_002

2

bentein, janse and soltic

and modality (commonly abbreviated as tam).4 The different chapters explore a number of central research questions, such as: – Which domains of variation (e.g., syntax, semantics, morphology) can be related to the categories of tense-aspect-modality? – Which dimensions of variation (e.g., diachrony, register/style, dialect) are relevant? – Which forces stimulate linguistic variation? What role does the existing linguistic system play? – What does variation look like from a synchronic point of view? – How is the synchronic competition between variant expressions resolved from a diachronic point of view? – What can modern linguistic theory contribute to the study of variation? – How might variation be approached? What is the value of quantitative vs. qualitative approaches? The present volume does not limit itself to one specific research question: the contributors were asked to focus on tense, aspect and modality, but were stimulated to discuss these phenomena from various time periods, and to take different approaches. We thus hope that this volume will stimulate the further investigation of (tam-related) linguistic variation in Ancient Greek.5

1.2

Some Theoretical Background

The space available for this introduction is too limited for us to provide a full background to the variationist literature and its application to Ancient Greek. It is worth giving some historical background, however, as well as introducing some central concepts. Variationist (socio)linguistics can be traced back to the 1960s, and in particular to the work of William Labov. Hymes (1974) observes that the study of linguistic variation followed that of phonology, morphology, syntax and semantics:

4 A terminological note is in order here. The tam-triad is often referred to as tense, aspect and mood. Since we understand these categories in a functional (rather than a formal) sense, we prefer ‘modality’ over ‘mood’. 5 For another recent volume on linguistic variation, see Leiwo et al. (2012). This volume takes an even broader approach as it does not focus on one specific linguistic area, and deals with both Latin and Greek.

introduction

3

To a great extent it was the conquest of speech sounds as an area of pattern belonging to linguistics that gave structural linguistics its impetus […]. Morphology came to be intensively cultivated in the late 1930s and the 1940s. Syntax came more to attention in the 1950s […]. Semantics has become a major concern in the 1960s […]. Very recently the notion of sociolinguistic description has been advanced. hymes 1974:89–90

In the variationist literature, it is common to distinguish between a number of dimensions of variation (see e.g., Seidl & Wirth 2008). According to the model first introduced by Coseriu (1969), four such dimensions can be distinguished: (a) the ‘diachronic’ dimension (variation in time); (b) the ‘diatopic’ dimension (variation in space); (c) the ‘diastratic’ dimension (variation according to the speaker’s social status); and (d) the ‘diaphasic’ dimension (variation in communicative settings).6 The diastratic/diaphasic dimension lies at the heart of modern-day sociolinguistics: various models have been introduced to connect language to its social context, centering around notions such as ‘politeness’, ‘register’, and ‘accommodation’.7 Some of these have been applied to Ancient Greek (for further details see Bentein 2016). Next to dimensions of variation, we can also recognize domains of variation (see e.g., Croft 2006:98–103). Three main types of variation are distinguished in this regard, corresponding to the three main areas in grammar: (a) phonetic variation (e.g., pronunciation vs. non-pronunciation of final t/d in words such as ‘soft’ and ‘missed’); (b) lexical variation (e.g., ‘trousers’ vs. ‘pants’); and (c) morpho-syntactic variation (e.g., ‘there is a lot of people’ vs. ‘there are a lot of people’). A key construct in the study of these different types of variation is that of the ‘variant field’ (see e.g., Rydén 1987), denoting a pool of linguistic features that are considered roughly equivalent.8

6 There is some discussion with regard to these dimensions. Considerable overlap exists between the diaphasic and diastratic dimensions in particular, to the extent that one could doubt whether any real distinction can be made. (Coseriu, for example, originally considered factors such as age, occupation and sex to belong to the diaphasic dimension, see Dufter-Stark 2008:87.) Other scholars argue that yet another dimension should be introduced, that is, communicative immediacy vs. communicative distance (Dufter-Stark 2008:90–94). 7 See Eckert (2012) for a historic overview. Eckert (2012:87) argues that ‘the treatment of social meaning in sociolinguistic variation has come in three waves of analytic practice’. 8 This ‘equivalency’, especially concerning lexical and syntactic variation, has been subject to considerable debate.

4

bentein, janse and soltic

Slightly less common is the distinction made by Croft (2006:98–103) between levels of variation:9 (a) first-order variation (individual utterances, e.g., a student who always writes ‘chose’ for ‘choose’); (b) second-order variation (sociallyvalued variants, e.g., copula deletion in Black English Vernacular, as in ‘you from outta town’); and (c) third-order variation (linguistic differences across dialects and languages, e.g., the pronunciation of ‘tomato’ in American vs. British English [təˈmeɪtoʊ/təˈmɑːtəʊ]). Generally speaking, interest in language variation focuses on second- and third-order variation. As Wolfram (2006:333) notes, however, ‘socially meaningful aspects of individual speaker performance are of interest to those interested in language variation’.10 Within variationist linguistics, two main approaches can be identified: the ‘qualitative’ and the ‘quantitative’ approaches. In the former type, exhaustive and detailed analysis is central, attention being paid to both rare and more frequent phenomena; in quantitative analysis, features are classified and counted on the basis of which generalizations are made (McEnery & Wilson 2001:76). In variationist linguistics, quantitative analysis has dominated. As Leiwo (2012:2) notes, however, this quantitative methodology has not been easy for languages such as Ancient Greek, ‘because of the lack of adequate linguistically-annotated corpora’.11 Moreover, when dealing with restricted data (e.g., ‘first-order’ variation), qualitative analysis forms the starting point. Often the two approaches complement each other. When it comes to Ancient Greek, many papers and books deal with linguistic variation. When we apply the above-discussed notions, however, it turns out that they have a rather specific profile: they typically discuss diachronic, and especially diatopic variation in the Classical period,12 focusing on phonology and morphology.13 Much rarer are works dealing with diastratic/diaphasic

9 10 11

12 13

For an application to Ancient Greek, see Bentein (2012). Evans (2010), for example, draws attention to the interest of individual linguistic preferences in documentary papyri. Several such projects are underway, however. Most of these focus on the Post-classical period: see e.g., Dag Haug’s proiel-project, Marja Vierros’ sematia, and PorterO’Donnell’s OpenText.org. At Ghent University, Klaas Bentein is compiling a corpus of linguistically and socio-pragmatically annotated Roman and Byzantine papyri. Mention should also be made of the Perseus database. For the latest outline of (the history of) Greek dialectology, see e.g., Consani (2014). In the recently published Encyclopedia of Ancient Greek language and linguistics (Giannakis 2014), for example, there is a separate article for ‘Linguistic variation in Classical Attic’ (Poccetti 2014), but none for the other historical periods (see more generally, Consani 2014).

introduction

5

variation in the Post-classical and Byzantine periods, focusing on lexical and syntactic variation. Moreover, a common characteristic of several studies and grammars is that they are ‘taxonomic’ in nature: they list variation, but do not explain the reasons or chronology behind it (Leiwo 2012:5). The present volume attempts to offer a more balanced picture by discussing variation in a functional area over a broad period of time.

1.3

Outline of the Volume

That the notions of tense, aspect, and modality are interconnected is a wellknown fact (see, for example, the discussion about whether the future should be considered in terms of tense or modality, and similarly, whether the perfect should be considered in terms of tense or aspect; as well as the observation that perfective aspect typically combines with past tense in the indicative). The volume is organized as follows: the first three papers are concerned with tense (ch. 2–4), the next three papers with aspect (ch. 5–7), and the following four papers with modality (ch. 8–11). The volume ends with a paper that contains some more general, methodological considerations, with illustrations from all three functional areas (ch. 12). In the area of tense, the main topic concerns the expression of relative tense (or the absence thereof), through various modes of realization. Antonio Lillo (ch. 2) discusses the use of subjunctive and optative in purpose clauses introduced by ἵνα and ὅπως, focusing on Herodotus’ Histories. He argues that the choice for one or the other is a matter of relative tense, rather than primary versus secondary purpose (as is traditionally assumed for Homer): ‘The optative in purpose sentences reports an action in the present with respect to the action of the main clause, while the subjunctive reports an action in the future with respect to the action of the main clause’. In both cases, the main clause can contain a primary or secondary tense. Jerneja Kavčič (ch. 3) analyzes the use of the infinitive in complement clauses following verbs of saying/thinking, focusing on a corpus of documentary papyri from the Roman period. She observes that the aorist infinitive tends to be excluded from this type of complementation pattern, and that there is a tendency towards the use of the present infinitive for simultaneity, the perfect infinitive for anteriority and the future infinitive for posteriority. Interestingly, the aorist does appear more often in complement clauses introduced by ὅτι or ὡς. The spirit of the final tense-related paper (ch. 4) is somewhat different from that of these first two papers: Julián Méndez Dosuna opposes the view that relative tense was an operative category in the grammar of Ancient (Classical) Greek. He draws attention to the

6

bentein, janse and soltic

fact that verbs of immediate perception, which necessarily take a temporally simultaneous complement, can be followed by a present or aorist participle. The following three papers concern aspect and are more varied in nature. Gerry Wakker (ch. 5) discusses expressions of omnitemporality, concentrating on the use of the so-called ‘gnomic’ aorist. After studying the possible origin of the gnomic aorist, she discusses its synchronic value in Hesiod’s Theogony and Works and Days. Another question she addresses is how the gnomic aorist relates to the omnitemporal present tense. Rutger Allan (ch. 6) treats aspectual variation with verbs of motion and speech in Thucydides. In a number of cases where the aorist might be expected, the imperfect tense is used for completed states of affairs. Allan focuses on one specific use, called the ‘imperfect of continuing relevance’, which he distinguishes from ‘the imperfect creating a temporal framework’, and ‘the discourse-structuring imperfect’. In the second part of his paper, Allan argues for the benefits of modern linguistic theory, Cognitive Linguistics in particular, for our understanding of this particular use of the imperfect. Amalia Moser (ch. 7) offers the most general discussion: she discusses the relationship between the categories of aspect and Aktionsart, covering a very broad time period (from Archaic to Standard Modern Greek). Moser argues that while there has been little change morphologically, the system itself has been altered in a fundamental way: the three stems have shifted from the expression of Aktionsart to the expression of aspect. To demonstrate this, she examines a corpus of historiographical texts. The authors treat various topics in the functional area of modality also, using different methodologies and corpora. Antonio Revuelta Puigdollers (ch. 8) discusses the grammaticalization of (counterfactual) desiderative constructions introduced by ὤφελ(λ)ον, from Homer up to the 2nd c. ce. He describes in detail how ὤφελ(λ)ον developed from an auxiliary verb into an illocutionary particle (semantically and morpho-syntactically), and also outlines several other constructions competing for the expression of desire (such as ἴθε/αἴθε, εἰ/αἰ γάρ, ὡς with the optative and ἐβουλόμην with the infinitive). Marina Veksina (ch. 9) analyzes mood fluctuations in Coan inscriptions, showing that whereas in decrees on sacral matters the substance is normally expressed in the imperative, the present indicative and future indicative can also express permission/obligation. Similarly to Antonio Revuelta Puigdollers’ contribution, Marina Veksina’s analysis is well embedded in current linguistic theory. Geoffrey Horrocks (ch. 10) concentrates on futurity-related expressions in a period and corpus that has received relatively little attention so far, that is, later Medieval Greek. He argues for the need to investigate both lower and higher registers. The language of high-register texts is often considered identical to that of the Classical period, but this, Horrocks shows, is not the case. By way of illustration,

introduction

7

he analyses the use of the indicative future, subjunctive and optative in Anna Comnene’s Alexiad. Finally, Martti Leiwo’s paper (ch. 11) concentrates on confusion of moods in documentary texts from Mons Claudianus. He notes that in many documents there is variation between the infinitive, participle and imperative, arguing that this variation may be both language-internal and contact-induced, drawing attention to the possible influence of Coptic. Leiwo’s discussion concentrates on the idiolectal level: he argues that a small set of data can show ongoing trends in language use, trends which are relevant at a higher level as well. In the article that concludes the volume, Andreas Willi (ch. 12) offers some more general considerations centered around the notion of ‘register’. He investigates the extent to which it is possible to determine an interrelation between linguistic registers and certain distributional patterns in the use of various grammatical phenomena relating to tense/aspect/mood, and stresses the importance of paying attention to both synchronic and diachronic aspects of register variation, and of being aware of the dialectal diversity of Ancient Greek when investigating registers. Another important point that is made concerns the need to account for certain register-specific grammatical preferences. By way of illustration, Willi draws attention to the frequency of the so-called ‘resultative’ perfect in oratory.

References Bentein, K. 2012. ‘Perfect periphrases in Post-classical and Early Byzantine Greek: an ecological-evolutionary account’. jgl 12:205–275. Bentein, K. 2016. ‘Towards a socio-historical analysis of Ancient Greek? Some problems and prospects’. Towards a sociolinguistic approach to textual interpretation, ed. by E. Trapp & A. Cuomo. Turnhout. In print. Consani, C. 2014. ‘Ancient Greek sociolinguistics and dialectology’. In: Giannakis 2014:117–124. Coseriu, E. 1969. Einführung in die strukturelle Linguistik. Tübingen. Croft, W. 2006. ‘The relevance of an evolutionary model to historical linguistics’. Different models of linguistic change, ed. by O. Nedergård Thomsen, 91–132. Amsterdam. Dufter, A. & E. Stark. 2002. ‘La variété des variétés: combien de dimensions pour la description?’ RJb 53:81–108. Eckert, P. 2012. ‘Three waves of variation study. The emergence of meaning in the study of sociolinguistic variation’. ara 41:87–100. Evans, T.V. 2010. ‘Identifying the language of the individual in the Zenon Archive’. The language of the papyri, ed. by T.V. Evans & D. Obbink, 51–70. Oxford.

8

bentein, janse and soltic

Giannakis, G. (ed.). 2014. Encyclopedia of Ancient Greek language and linguistics. Leiden. Guy, G. 2003. ‘Variationist approaches to phonological change’. The handbook of historical linguistics, ed. by B. Joseph & R. Janda, 369–400. Oxford. Hymes, D. 1974. Foundations in sociolinguistics: an ethnographic approach. Philadelphia. Labov, W. 1994. Principles of linguistic change. i: Internal factors. Oxford. Leiwo, M., Halla-aho, H. & M. Vierros (eds.). 2012. Variation and change in Greek and Latin. Helsinki. Leiwo, M. 2012. ‘Introduction: variation with multiple faces’. In: Leiwo et al. 2012:1–12. McEnery, T. & A. Wilson. 2001. Corpus linguistics: an introduction. Edinburgh. Poccetti, P. 2014. ‘Linguistic variation in Classical Attic’. In: Giannakis 2014:364–367. Romaine, S. 1982. Socio-historical linguistics: its status and methodology. Cambridge. Rydén, M. 1987. ‘Syntactic variation and paradigmatic typology’. sl 41:48–58. Seidl, C. & T. Wirth. 2008. ‘Eine Sprache ist kein Monolith’. bsav 72:5–18. Wolfram, W. 20062. ‘Variation and language: an overview’. Encyclopedia of language and linguistics, ed. by K. Brown, 333–341. Oxford.

chapter 2

Subjunctive and Optative in Herodotus’ Purpose Clauses as Relative Tense Markers Antonio Lillo

Final clauses are usually constructed with an eventual subjunctive—without ἄν when the clause marker is ἵνα—but may use the modal particle when the markers are ὅπως (Hdt. ὅκως) and ὡς, and when the verb of the main clause is in the primary tense. When the verb of the main clause is in the secondary tense, the final clause is constructed with an optative without ἄν when the clause marker is ἵνα, but it may use the particle when the markers are ὅπως (Hdt. ὅκως) and ὡς. This optative without a modal particle dependent on main clauses in secondary tense verbs is considered an oblique optative or optative of subordination.1 Yet these rules are not always followed mechanically. The occurrence of a subjunctive rather than an optative after the main clause in the historical tense, which is an even more common procedure than the use of the optative (Goodwin 1889:114–115), has been explained as an attempt to reproduce the original form of the final clause, making the expression more ‘vivid’. Thus we have ἦλθεν ἵνα ἴδῃ instead of ἤλθεν ἵνα ἴδοι, because it tries to reproduce the original construction, ἔρχομαι ἵνα ἴδω. Kühner-Gerth (1904:387) point out that in texts in which two final clauses depending on a main clause in the historical tense appear, the subjunctive refers to an immediate purpose, while the optative to a mere assumption or consequence, uses which they define as Hauptzweck and Nebenzweck respectively. Similarly, Vandaele (1897:119) thinks that the subjunctive expresses the intention with some force, while the optative does so in a more attenuated fashion, and adds that, were these purpose clauses to be considered independent, the subjunctive would be equivalent to an indicative future and the optative to a potential.2 Goodwin (1889:115) also notes that the subjunctive marks an immediate consequence of the action expressed in the main clause and the optative a remote consequence, and adds that the construction with optative would mark a consequence of the 1 Goodwin (1889:112); Kühner-Gerth (1904:378); Rijksbaron (2006:61). 2 See also Humbert (1960:234): ‘En réalité, ἵνα exprime à l’ optatif une possibilité, beaucoup plus voisine de la consécution que la finalité, tandis que la finalité est directement rendue par le subjonctif’.

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_003

10

lillo

action expressed in the clause with subjunctive. Like the previous authors, Chantraine (1963:269) speaks of a more ‘vigoreux’ subjunctive than the optative, and Rijksbaron (2006:61) also explains the appearance of the subjunctive in purpose clauses depending on main clauses with a verb in the secondary tense as the result of maintaining ‘direct speech’. Similarly, he considers that when subjunctive and optative occur in a couple of purpose clauses dependent on the secondary tense in the main clause, the subjunctive represents the immediate purpose and the optative a more remote purpose. It is more difficult to explain the occurrence of purpose clauses with an optative dependent on main clauses with a primary tense—passages which mostly have transmission problems or need specific explanations (Kühner-Gerth 1904:383)—and which Goodwin (1889:115) considers ‘a mere irregularity of construction’. These distributions of modes, already appearing in Homer, are explained by Willmott (2007:155–174) within the system, as a result of their own modal meanings. Owing to the meaning of the subjunctive as a future reference, she considers that ‘many of the examples of the subjunctive in secondary sequence may be explained in exactly the same way as those in primary sequence, that is, they refer to an event that will take place in an actual future moment’ (Willmott 2007:167), while the optative, which does not indicate time, ‘could well express purpose in any time, though its “unreal” nature suggests that the purposes will be more loosely connected to the main clause’. Hence her conclusion that ‘the optative after primary as well as secondary main verbs confirm that its use is semantically rather than grammatically governed’ (Willmott 2007:174). This distribution of ‘anomalous’ modal forms, in which a final clause with optative depends on a verb in the primary tense, is also found in Herodotus, for example when purpose clauses with subjunctive and optative both depend on the same verb in the primary tense, a construction that also appears in Thucydides. These couples of purpose clauses with modal variation in Homer can, as seen above, be explained as Hauptzweck and Nebenzweck by referring to the final clauses with subjunctive and optative respectively. An example from Homer is: (1) Ἕκτορι γάρ οἱ θυμὸς ἐβούλετο κῦδος ὀρέξαι Πριαμίδῃ, ἵνα νηυσὶ κορωνίσι θεσπιδαὲς πῦρ ἐμβάλῃ3 ἀκάματον, Θέτιδος δ’ ἐξαίσιον ἀρὴν πᾶσαν ἐπικρήνειε·

3 The optative form ἐμβάλοι is Hermann’s conjecture.

subjunctive and optative in herodotus’ purpose clauses

11

‘For his heart was set on giving glory to Hector, son of Priam, so that he might cast upon the beaked ships unwearied, wondrous-untiring fire, and so fulfill to the uttermost the portentous prayer of Thetis’. hom. Il. 15.596–15.599

We can see that in (1) both final clauses indicate actions which refer to the same purpose, so that the final clause with subjunctive indicates the specific purpose, while the optative relates to the overall objective, the result of a fulfilment of the actions in the main clause and the subordinate clause with subjunctive. The reverse order may also occur: the first final clause having an optative and the second a subjunctive.4 In any case, the construction with the subjunctive indicates Hauptzweck, while the one with the optative indicates a contingency also to be taken into account. Yet these constructions in Herodotus with two final clauses, in which subjunctive and optative alternate, differ from the Homeric constructions, since the actions to which they refer are, or may be, of a different nature, with little indication which expresses the Hauptzweck and which the Nebenzweck. Hence we will analyze these passages from Herodotus with two final clauses in which subjunctive and optative alternate, since the analysis of the state of affairs in which the actions of both subordinates take place will allow us to establish more clearly whether the optative is being used to express a remote or secondary purpose with respect to the final clause with subjunctive, as is normally proposed, or whether other usage rules apply. (2) οὗτοι ὦν ἐπείτε τὴν Ἱστιαίου αἱρέοντο γνώμην, ἔδοξέ σφι πρὸς ταύτῃ τάδε ἔργα τε καὶ ἔπεα προσθεῖναι· τῆς μὲν γεφύρης λύειν τὰ κατὰ τοὺς Σκύθας ἐόντα, λύειν δὲ ὅσον τόξευμα ἐξικνέεται, ἵνα καὶ ποιέειν τι δοκέωσι ποιεῦντες μηδὲν καὶ οἱ Σκύθαι μὴ πειρῴατο βιώμενοι [καὶ βουλόμενοι] διαβῆναι τὸν Ἴστρον κατὰ τὴν γέφυραν ‘When these accepted then Histiaeus’ counsel, they [the Ionians] resolved to follow it by act and word: to break as much of the bridge on the Scythian side as a bowshot from there reached, so that they seem to be doing something when in fact they were doing nothing, and that the Scythians not try to force a passage across the bridge over the Ister’. hdt. 4.139.1

4 See also Hom. Il. 24.582–586; Hom. Od. 3.76–78; 12.156–157.

12

lillo

With respect to (2), since the clause with subjunctive precedes that with optative, both Kühner-Gerth (1904:387) and Vandaele (1897:116) point to Herodotus’ insistence on what the final clause with subjunctive indicates. After further analysis of the state of affairs in the text, however, it is possible to see that there are two purposes of a different nature, unrelated to one another. What is indicated in the subordinate clause with optative (the second one) is the immediate and specific purpose, which is directly related to what is described in the main clause within the same state of affairs. The bridge is destroyed because the Ionian tyrants of the Hellespont, whose tyranny was provoked by Darius, were not interested in the Scythians—enemies of the Persians—crossing the river and entering their domains. Moreover, the very fact that the bridge was destroyed implies that the Scythians did not think of crossing it. By contrast, what the final clause with subjunctive indicates (the first one) is a purpose which obeys another strategy. It gives the appearance that they were working without actually doing so, with two future goals. Firstly, the Scythians might not think of crossing the bridge, because they thought they could not do so if it had been destroyed. Secondly, once the danger of the Scythians actually using it had passed, they could use it themselves, if this suited, since it had not really been destroyed. Later indeed the bridge was repaired by the same Ionians (Hdt. 2.141), so the Persians, after outwitting the Scythians, could cross it and escape. It is not appropriate to speak, therefore, of a Hauptzweck and of a Nebenzweck, since both purpose clauses express two different and equally relevant actions within an overall strategy. (3) τῶνδε δὲ εἵνεκα ἀνῆγον τὰς νέας, ἵνα δὴ τοῖσι Ἕλλησι μηδὲ φυγεῖν ἐξῇ, ἀλλ’ ἀπολαμφθέντες ἐν τῇ Σαλαμῖνι δοῖεν τίσιν τῶν ἐπ’ Ἀρτεμισίῳ ἀγωνισμάτων. ‘They launched their ships in this way, so that the Hellenes could not escape, but, caught in Salamis, pay the penalty for the battles at Artemisium’. hdt. 8.76.2

With reference to (3), Kühner-Gerth (1904:387),5 Vandaele (1897:116),6 Humbert (1960:234) and Rijksbaron (2006:62) consider that the punishment is a consequence of preventing escape, but the situation described in the text is not

5 ‘Konj. von dem direkt erstrebten Ziele, Opt. von der möglichen Folge’. 6 ‘Le but direct, important, qui aura pour conséquence le châtiment entrevu en idée, δοῖεν τίσιν’.

subjunctive and optative in herodotus’ purpose clauses

13

precisely that. Earlier, in Herodotus 8.75, it is said that a servant of Themistocles informs the Persians of the previous decision of the Greek fleet to flee out of fear of them, and also because they thought it more suitable to go and defend the Peloponnese than stay where they were (Hdt. 8.74.2). The ships of the Persians thus had two different missions. On the one hand they had to prevent the Greeks from sailing to the Peloponnese, which would make their strategy more difficult, and on the other, once they had caught the Greeks, they had to take their revenge for the events in Artemisium. This is not a global purpose to be broken down into two specific actions, as if it were a single mission, that is, preventing the Greeks from escaping in order to take revenge. Accordingly, the construction with optative refers to the immediate purpose, which is a direct consequence of the arrival of the ships and their subsequent seizure, revenge for the events in Artemisium. The final clause with subjunctive indicates the future plan of the Greek fleet, the fulfilment of which the Persians intend to prevent, i.e., the escape of the fleet to the Peloponnese as a defense prior to the arrival of the Persians. The interpretation that the Greeks would be punished as a result of their failure to escape—with no indication that the plan preceding such an escape has any other purpose than to evade the punishment for the events at Artemisium—would, we believe, require an optative ἐξείη instead of a subjunctive ἐξῇ. (4) ἐς τοῦτον δὴ τὸν χῶρον ἐβουλεύσαντο μεταναστῆναι, ἵνα καὶ ὕδατι ἔχωσι χρᾶσθαι ἀφθόνῳ καὶ οἱ ἱππέες σφέας μὴ σινοίατο ὥσπερ κατιθὺ ἐόντων· ‘To that place then they planned to go so that they might have plenty of water for their use and not be harmed by the horsemen, as now when they were face to face with them’. hdt. 9.51.3

In (4), as in the previous texts, we find two distinct actions: the harassment of the Persian cavalry and the shortage of water. The final clause with optative does not point to the more remote purpose, but to the more immediate one, which will happen as soon as the action indicated in the main clause is fulfilled. The Greek riders would not be injured or harassed by the Persians after their change of location. A new state of affairs is indicated with the subjunctive construction, which it is intended will take place in the future, as a consequence of what is indicated in the main and final clauses: that the Greeks might get water. We do not think this interpretation is appropriate in the case of (5).

14

lillo

(5) Αἰγύπτιοι δὲ ταῦτα ἐπιμεμφόμενοι ἀπέστησαν ἀπ’ αὐτοῦ, δοκέοντες τὸν Ἀπρίην ἐκ προνοίης αὐτοὺς ἀποπέμψαι ἐς φαινόμενον κακόν, ἵνα δὴ σφέων φθορὴ γένηται, αὐτὸς δὲ τῶν λοιπῶν Αἰγυπτίων ἀσφαλέστερον ἄρχοι. ‘The Egyptians blamed him for this and rebelled against him; for they thought that Apries had knowingly sent his men to their doom, so that their extermination occurred and he reigned safer over the rest of the Egyptians’. hdt. 2.161.4

In (5) there is a transmission problem, because ἄρχῃ appears as a variant of ἄρχοι in the second final clause. In addition, this is the only place in Herodotus where a substantial difference between the two final clauses is difficult to see, so we believe it is not sound evidence, since the variant ἄρχῃ is possible here. However, if the ἄρχοι variant is accepted, we would have a text in which, as in (1), the subjunctive construction indicates the immediate result, the Hauptzweck, while the one with optative, the result stated in the main clause and the former purpose clause, the Nebenzweck. To see to what extent the uses of subjunctive and optative in (2) to (4) can be applied to any final construction with any of these modes in Herodotus, we will now analyze some minimal pairs with alternation between purpose clauses of subjunctive and optative dependent on the same verb in the main clause. (6) οἱ δὲ τοκέες παραλαβόντες τὸ οὔνομα τοῦτο, ἵνα θειοτέρως δοκέῃ τοῖσι Πέρσῃσι περιεῖναί σφι ὁ παῖς, κατέβαλον φάτιν ὡς ἐκκείμενον Κῦρον κύων ἐξέθρεψε. ‘His parents, after hearing this name, in order to make the story of his salvation seem more marvellous to the Persians, circulated a story that Cyrus was suckled by a dog when exposed’. hdt. 1.122.3

(7) τὰς δὲ οἰκίας οὐ κατέβαλλε ὁ Λυδὸς τῶνδε εἵνεκα, ὅκως ἔχοιεν ἐνθεῦτεν ὁρμώμενοι τὴν γῆν σπείρειν τε καὶ ἐργάζεσθαι οἱ Μιλήσιοι, αὐτὸς δὲ ἐκείνων ἐργαζομένων ἔχοι τι καὶ σίνεσθαι ἐσβάλλων. ‘The Lydian did not destroy the houses for this reason, so that the Milesians might have homes from which to plant and cultivate their land, and that he could plunder their fruits with his invading army if they did so’. hdt. 1.17.3

subjunctive and optative in herodotus’ purpose clauses

15

We can see in (6) that the final clause with subjunctive refers to the future, and that Cyrus was to gain the aura of a person protected by the deity after his story was spread about. By contrast, in (7) the final clause indicates an action that is occurring while the Lydians take the decision not to destroy the houses of the Milesians. In other words, the Milesians can continue planting and working the land because the Lydians have not destroyed their homes. From what has been said, it is reasonable to assume that a form of the subjunctive in (7), i.e. ἔχωσιν instead of ἔχοιεν, presupposes that the Milesians did not cultivate the land at that time, and that the non-destruction of the houses presupposes they would farm the land after finding their homes thus preserved by the Lydians. (8) τοῦτο δὲ τὰς διόδους τὰς ἐς τὰς κορυφὰς τῆς Μυκάλης φερούσας προστάσσουσι τοῖσι Μιλησίοισι φυλάσσειν ὡς ἐπισταμένοισι δῆθεν μάλιστα τὴν χώρην· ἐποίευν δὲ τούτου εἵνεκεν, ἵνα ἐκτὸς τοῦ στρατοπέδου ἔωσι. ‘The Persians appointed the Milesians to guard the passes leading to the tops of Mycale with the idea that they knew the place perfectly, and for them to stay away from the camp’. hdt. 9.99.3

(9) τοῦδε δὲ εἵνεκεν ταῦτα ἐποίεε, ἵνα μὴ συντυχίης δεινῆς τε καὶ μεγάλης Πολυκράτεα καταλαβούσης αὐτὸς ἀλγήσειε τὴν ψυχὴν ὡς περὶ ξείνου ἀνδρός. ‘[Amasis] did this [to cancel the bond of hospitality] because of the following, so that he himself might not have to sadden his heart for a friend when some great and terrible mischance overtook Polycrates’. hdt. 3.43.2

In (8), the order of the Persians that the Milesians check the passes will result in a new state of affairs, the removal of the Milesians from the camp. By contrast, in (9) the final clause refers to an action that already occurs from the moment the action of the main clause takes place, as well as being its necessary consequence.7 The cancellation of the bond of hospitality implies that, from that very moment, Amasis does not sadden his heart for what might happen to Polycrates. Let us analyze the three cases in Herodotus where the final clause with optative precedes the one with subjunctive. 7 Similar constructions with optative in Hdt. 5.37.2; 5.98.1; 8.76.3. There is another minimal pair in Hdt. 7.8.1.

16

lillo

(10) ἐποίεε δὲ ἀμφότερα ταῦτα, τόν τε ποταμὸν σκολιὸν καὶ τὸ ὄρυγμα πᾶν ἕλος, ὡς ὅ τε ποταμὸς βραδύτερος εἴη περὶ καμπὰς ἀγνύμενος, καὶ οἱ πλόοι ἔωσι σκολιοὶ ἐς τὴν Βαβυλῶνα, ἔκ τε τῶν πλόων ἐκδέκηται περίοδος τῆς λίμνης μακρή. ‘He made these two works, the course of the river tortuous and that the excavation remain flooded, so that the current of the river might be slower because of the many windings and that the trips by boat to Babylon might be crooked and that right after this trip, a long detour to the lake had to be made’. hdt. 1.185.6

In (10) the construction with optative indicates the purpose and immediate consequence of the action referred to in the main clause: the work on the riverbed implies that the current runs slower. By contrast, the final clause with subjunctive indicates what will be achieved as a result of what is previously stated in both the main clause and the subordinate clause with optative: that the boat will make more turns and take a long detour. (11) ὁ μέν νυν κάλλιστος νόμος οὗτός σφι ἦν, οὐ μέντοι νῦν γε διατελέει ἐών, ἄλλο δέ τι ἐξευρήκασι νεωστὶ γενέσθαι ἵνα μὴ ἀδικοῖεν αὐτὰς μηδ’ [ἐς ἑτέραν πόλιν] ἄγωνται·8 ‘This, then, was their best custom; nevertheless, it does not continue at this time; but they have invented lately a new one, so that the women not be wronged or taken [to another city]’. hdt. 1.196.5

Just as in (10), the purpose or the immediate consequence of the action indicated in the main clause is expressed with an optative in (11): the change of law implies that no injustice will be done to the women if they are purchased for slavery, given the poverty of which the text speaks. Moreover, the final construction with the subjunctive indicates that the purpose to be subsequently fulfilled is a result of the change to the law: that the Babylonian women will not be taken away. (12) ἐκ μὲν δὴ τῆς ἀντίης προσπλέειν οὔ κώ σφι ἐδόκεε τῶνδε εἵνεκα, μή κως ἰδόντες οἱ Ἕλληνες προσπλέοντας ἐς φυγὴν ὁρμήσειαν φεύγοντάς τε εὐφρόνη καταλαμβάνῃ· 8 I follow Rosen’s edition.

subjunctive and optative in herodotus’ purpose clauses

17

‘Then it seemed good [to the Persians] not to make a head-on attack [against the ships] because of this, so that the Greeks, seeing them attack, would not be put to flee and the night would close upon them as they fled’. hdt. 8.6.2

In (12) the construction with optative indicates the purpose or immediate consequence of the action referred to in the main clause (that the Greeks might not be put to flight if they see they are being attacked), while the clause with subjunctive indicates the situation will arise later, if the attack and the ensuing night-time flight go ahead, whereupon the situation would be more favourable to the Greeks. Unlike examples (2), (3) and (4), where both final clauses describe actions that are not consequences of each other, so that it is not entirely appropriate to speak of a Hauptzweck for the final clause with subjunctive and of a Nebenzweck for the other with optative, because these are of a different nature, in (10), (11) and (12) the action described in the final clause with subjunctive is the ultimate consequence of fulfilment of the action indicated in the main clause and in the final clause with optative. This does not allow us to say that the final clause with optative indicates a secondary action, more loosely connected to the main clause, however, since the action expressed by this final clause is crucial in fulfilling what is indicated in the following clause with subjunctive. On the contrary, the final clause with optative indicates an action prior to that with the subjunctive, because the clause with optative indicates the natural result of the fulfilment of the action expressed in the main clause, which occurs while producing that which is indicated in this main clause. It is thus the order of the final clauses, the first with optative and the second with subjunctive, with a previous action (that of the optative) and the subsequent action (that of the subjunctive) that makes it possible to speak of a Hauptzweck expressed by a subjunctive and a Nebenzweck expressed by an optative. This is due to the nature of the specific actions, the fact that one is necessarily prior to the other, and not because one is a secondary or probable result and the other a main or certain result. In short, these two modal uses refer to the moment each of the actions take place, in relation to the moment the action of the main clause is completed. In short, we believe it is necessary to change the approach to the use of the optative in these constructions from Herodotus. Based on its use as a means for expressing the remote possibility and a lesser degree of vividness with which the speaker presents the proposition in a series of texts and authors, mainly in Homer, it has come to be proposed for Herodotus, too, that the final clauses with optative express Nebenzweck, a probable result, when these are compared

18

lillo

with those constructed with subjunctive, which express a Hauptzweck, a certain result. The final clause with optative, which is an oblique optative, indicates that it depends on a main clause in the secondary tense, though not only this. In all the previous evidence from Herodotus, the final clause with optative makes reference to an action that would take place precisely because the action of the main clause also takes place, and at the same time as the fulfillment of what is indicated in the main clause. Thus, in (2) the Scythians would not try to cross the bridge because it is destroyed; in (3) the mere presence of the Persian ships would make the Greeks think the Persians would take revenge for the defeat at Artemisium; in (4) the Greek cavalry would not be harassed by the Persian army because of its change of location; in (7) the Milesians could continue planting and working their land because the Lydians had not destroyed their homes; in (9) Amasis will not be hurt by what happens to Polycrates because he has canceled the bond of hospitality with him; in (10) the current of the river flows more slowly because of the works on the river; in (11) the change to the law makes it impossible to commit an injustice with the Babylonian women, and in (12) the Greeks would not flee because they would not be attacked. By contrast, the final clause with subjunctive refers to an action that would take place after the action of the main action occurred. Thus, in (2) the final clause with subjunctive reports the strategy to be followed by the Ionian tyrants, which consists in appearing to destroy the bridge without actually destroying it; in (3) the final clause with subjunctive indicates the future plan of the Greek fleet, which the Persians intend to quash with their arrival, by preventing its departure for the Peloponnese; in (4) the possibility of access to water by the Greeks after eluding the influence of the Persians is reported; in (6) the divine aura of Cyrus spreads after the report of his history; in (8) the order of the Persians that the Milesians control a given area will result in their removal from the camp; in (10) the boat trip to Babylon will make a long detour after the work on the riverbed is completed; in (11) the failure to take the women out of Babylon occurs after a change of law, and in (12) the night which will close in following an attack by the Persians, with the consequent inability to capture the bearer of fire, advises against such an attack. From the above it follows that, in Herodotus, the optative in purpose clauses reports an action in the present with regard to the action of the main clause, while the subjunctive reports an action in the future with regard to the action of the main clause. We are thus dealing with modal uses which imply a time concerning the action indicated in the main clause. Moreover, in these constructions, in which the final clause with subjunctive precedes that with optative, what is indicated in each is usually of a different nature, so that, in contrast to the Homeric evidence, what is indicated in the latter of these, with optative,

subjunctive and optative in herodotus’ purpose clauses

19

is not the consequence of what is generally stated in the main and in the final clause with subjunctive, as Goodwin (1889:115) suggests. There is an odd use of a main clause with the verb in the primary tense and an optative in the final clause which also appears in Homer, which is not easy to explain: (13) ἐχόμενοι τῆς γῆς ἐπ’ ἀριστερὰ καταπλώουσι ἐς θάλασσαν, καὶ ἀναπλώοντες ὀπίσω τῆς αὐτῆς ἀντέχονται, ἐγχριμπτόμενοι καὶ ψαύοντες ὡς μάλιστα, ἵνα δὴ μὴ ἁμάρτοιεν τῆς ὁδοῦ διὰ τὸν ῥόον. ‘[The fishes] swim seawards close to the left bank and keep to the same bank also on their return, coming up close and touching it as much as they can, lest the current make them miss their way’. hdt. 2.93.4

In the explanation of this text Kühner-Gerth (1904:383) follow Stein, who thinks that Herodotus is speaking ‘ironically’, and does not accept the purpose of the fish as a real fact. Chantraine (1963:271) gives these optatives of the Homeric texts ‘la valeur de souhait’. Goodwin (1889:115) considers them ‘a mere irregularity of construction’, while Willmott (2007:165) explains the Homeric passages of this construction as ‘a possible and conditional consequence of the main clause’, resorting to Jebb’s explanation of a similar construction in his edition of Oedipus Coloneus, v. 11, where he assumes in Hom. Od. 17.250, ἄξω τῆλ’ Ἰθάκης, ἵνα μοι βίοτον πολὺν ἄλφοι that the final clause with optative involves a prior conditional construction εἰ ἄγοιμι, ἄλφοι ἄν. Jebb’s explanation might be applicable here, though we consider it unnecessary, since from the evidence it appears that the optative in such purpose clauses refers to a present time in relation to the state of affairs of the verb in the main clause. Consequently, in spite of being a unique construction, the action indicated by the optative—not losing the path because of the current—occurs at once and as a consequence of the fact that the fish continuously swim close to the shore. We would thus be dealing with a special development deriving from the use of the optative in the above examples by Herodotus, because there would be absolutely no case here for a form of the subjunctive as a function of the meaning of this mode in these constructions.9

9 It is not necessary to consider here the text of Hdt. 1.110.3, κελεύει σε Ἀστυάγης τὸ παιδίον τοῦτο λαβόντα θεῖναι ἐς τὸ ἐρημότατον τῶν ὀρέων, ὅκως ἂν τάχιστα διαφθαρείη, because the optative with modal particle gives the action a potential sense.

20

lillo

These double final constructs, one with subjunctive and the other with optative, appear again five times in Thucydides,10 which may well be considered an influence of Herodotus. The detection of differences between both modal uses is more problematic, however, and thus in (14) the final clause with subjunctive appears before the one with optative: (14) παρανῖσχον δὲ καὶ οἱ ἐκ τῆς πόλεως Πλαταιῆς ἀπὸ τοῦ τείχους φρυκτοὺς πολλοὺς πρότερον παρεσκευασμένους ἐς αὐτὸ τοῦτο, ὅπως ἀσαφῆ τὰ σημεῖα τῆς φρυκτωρίας τοῖς πολεμίοις ᾖ καὶ μὴ βοηθοῖεν ‘But the Plataeans in the town at once displayed from the wall many fires, which had been prepared beforehand for this purpose, so that the enemy’s signals might be rendered unintelligible and they [the Thebans] might defer bringing aid’. thuc. 3.22.8

Kühner-Gerth (1904:387)11 consider that what is stated in (14) by the optative is a consequence of what is indicated by the subjunctive, and Amigues (1977:129) thinks that ‘le but visé par le sujet (subjonctif) n’est qu’ un aspect particulier d’un objectif plus large dont l’historien a une vue d’ ensemble rétrospective (optatif oblique)’.12 The reverse order appears in (15): (15) φυλάξαντες ἔτι νύκτα καὶ αὐτὸ τὸ περίορθρον ἐχώρουν ἐκ τῶν οἰκιῶν ἐπ’ αὐτούς, ὅπως μὴ κατὰ φῶς θαρσαλεωτέροις οὖσι προσφέροιντο καὶ σφίσιν ἐκ τοῦ ἴσου γίγνωνται, ἀλλ’ ἐν νυκτὶ φοβερώτεροι ὄντες ἥσσους ὦσι τῆς σφετέρας ἐμπειρίας τῆς κατὰ τὴν πόλιν. ‘While they were watching their opportunity for the time of night just before dawn, they went out of their houses against the enemy, so that they would not attack them by day, meeting men full of courage, and would be on equal terms with them, but at night, so that they would be more frightened and at a disadvantage from their enemy’s knowledge of the locality’. thuc. 2.3.4 10 11 12

Thuc. 2.3.4; 3.33.8; 6.96.3; 7.17.4; 8.87.3. This double construction also appears in Lysias 32.22. Kühner-Gerth (1904:387), as in Thuc. 7.17.4, with a reverse order in the appearance of the modes. Amigues’ point of view can also be applied to Thuc. 7.17.4. Kühner-Gerth (1904:387) consider that in Thuc. 7.17.4 the subjunctive indicates the ‘nächster Zweck’ and the optative the ‘mögliche Folge’.

subjunctive and optative in herodotus’ purpose clauses

21

Amigues (1977:244) considers that the optative form of (15) indicates ‘le but visé dans le passé’, while the subjunctive indicates the ‘visée prospective’.13 By contrast, Duhoux (2000:188, 232) considers these uses of the oblique optative as mere doublets of the subjunctive form.14 The appearance of two purpose subordinates dependent on the same verb, one with subjunctive and the other with optative, would be a device used by Thucydides under the probable influence of Herodotus, though, in our opinion, it is difficult to find differences in the use of each of these modes. Rather, it would appear that the variation is due to Thucydides’ ‘obsession’ with the μεταβολή, not because it is the expression of two distinct purposes, or a temporal gradation, or a function of the importance of both. The very fact that this variation is no longer found in Xenophon suggests that this variation was a purely formal procedure in Attic, without the meaning noted for the Ionian dialect.

References Amigues, S. 1977. Les subordonées finales par ὅπως en attique classique. Paris. Chantraine, P. 1963. Grammaire homérique. ii: Syntaxe. Paris. Duhoux, Y. 20002. Le verbe grec ancien: éléments de morphologie et de syntaxe historiques. Louvain-la-Neuve. Goodwin, W.W. 1889. Syntax of the moods and tenses of the Greek verb. London. Humbert, J. 1960. Syntaxe grecque. Paris. Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Chicago. Vandaele, H. 1897. L’optatif grec. Paris. Willmott, J. 2007. The moods of Homeric Greek. Cambridge. 13 14

The same explanation is valid for Thuc. 6.96.3. In his work he does not study the preceding texts of Herodotus.

chapter 3

Variation in Expressing Temporal and Aspectual Distinctions in Complement Clauses: A Study of the Greek Non-Literary Papyri of the Roman Period* Jerneja Kavčič

This paper draws attention to phenomena concerning the expression of aspectual and temporal distinctions in declarative infinitive clauses dependent on verbs of saying and thinking (DInf) in Hellenistic and Roman Greek, as well as in their syntactic variant, finite complement clauses, introduced with ὅτι. According to Rijksbaron (2006:97), the declarative infinitive ‘represents a statement or thought of the subject of the main verb concerning some state of affairs in the “real” world. This means that […] an aorist infinitive is anterior to, a present infinitive simultaneous with […], and a future infinitive posterior to the main verb’. The declarative infinitive is distinguished from the so-called dynamic infinitive, in which case ‘the state of affairs exists δυνάμει “potentially”’ (ibid.). The distinction between declarative and dynamic infinitive clauses is particularly evident in the period investigated because a DInf can be substituted by a ὅτι clause after governing verbs of saying and thinking (only some of the former allowed for such a substitution in Classical Greek), whereas ἵνα rather than ὅτι introduces finite complement clauses substituting for dynamic infinitive clauses. Finite complement clauses introduced with ἵνα are also attested in cases in which the use of an infinitive clause was obligatory in earlier periods.1 In other words, the distinction between dynamic and declarative infinitive clauses in the period investigated is particularly evident from * I would like to express my sincere thanks to the Alexander S. Onassis Public Benefit Foundation for awarding me a research scholarship for the academic year 2008–2009. The papyrological material which this paper is based upon was gathered during my stay in Athens in the spring of 2009. 1 For example, after λέγω ‘say, order’ (Mt 5.29, Lk 10.40), θέλω ‘want’ (Mk 6.25), and ἱκανός εἰμι ‘be able’ (Mt 8.8; see Blass et al. 2001:392–393; Mandilaras 1973:329–331; Joseph 1983:51–53; Horrocks 2010:173). In addition, Hellenistic and Roman Greek display the use of ὅτι clauses after verbs of thinking, which governed infinitive clauses in earlier periods (see Blass et al. 2001:327–328).

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_004

variation in expressing temporal and aspectual distinctions

23

the perspective of their syntactic variants (finite complement clauses), which eventually fully replaced the latter. As is well known, Modern Greek lacks the infinitive (see Joseph 1983:74–80). Thus, when speaking of infinitive clauses in Hellenistic and Roman Greek, Mandilaras (1973), refers to ἵνα and ὅτι infinitive clauses instead of dynamic and declarative infinitive clauses.2 The phenomena examined in this paper seem to indicate that between Classical Greek and the spoken Greek of the period investigated, significant changes occurred in terms of expressing temporal and aspectual distinctions in DInf. Some texts dating back to Hellenistic and Roman Greek seem to be characterized by the avoidance of the aorist infinitive in DInf. This infinitive was quite commonly used in Classical Greek DInf, the aorist infinitive referring to anteriority, as already mentioned. An example is Pl. Resp. 377e: ἅ φησὶ δρᾶσαι αὐτὸν Ἡσίοδος ‘what Hesiod says that he did’.3 The observation that this clause type is avoided in the New Testament goes back to Burton (1898:53) and is repeated in several modern studies (see Thorley 1989:295–296; Porter 1989:389; Fanning 1990:401; Kavčič 2009). Undoubtedly the aorist infinitive is also frequently attested in Hellenistic and Roman Greek; for example, Ἀριστέαν μὲν γὰρ ἔν τινι κναφείῳ τελευτῆσαί φασι ‘they say that Aristeus died in a fuller’s shop’ (Plut. Rom. 28.4–5).4 However, because the aorist infinitive was quite commonly used in Classical Greek DInf, DInf containing aorist infinitives in authors such as Plutarchus could perhaps reflect Atticistic influence. Because the phenomena examined in this paper were first observed in New Testament Greek, which is believed to offer—at least to a degree higher than in the case of literary texts—an insight into the ‘natural development’ of the 2 The two terms go back to Kurzová (1968; see Rijksbaron 2006:98 on other terms). Another option would be to use the (syntactic) term AcI clauses (dependent on verbs of saying and thinking) instead of the term declarative infinitive clauses. As is well known, the subject is omitted from the infinitive clause in Classical Greek if it is co-referential with the subject of the governing verb. In this case, the infinitive clause is not an AcI clause at all, which is why the term DInf could hardly substitute for AcI in the case of Classical Greek. In the case of co-referential subjects, the subject of the infinitive clause occurs in the accusative only when emphasized in this period (see Schwyzer 1959:376). In Hellenistic and Roman Greek, on the other hand, it seems that the subject of the infinitive can occur in the accusative even if no particular emphasis is placed on it (see Mayser 1926:336; Jannaris 1968:481). An example is Lk. 23.2: λέγοντα ἑαυτὸν Χριστὸν βασιλέα εἶναι ‘saying that he is Christ, a king’. Nevertheless, I avoid the use of the term AcI because the aforementioned Classical Greek construction with subject omission also occurs in the period investigated (see P.Oxy 1672, P.Oxy 2352, P.Ryl 2, p. 381). 3 See Pl., Ap. 33b, Symp. 178b, Resp. 377e, 586c; Hdt. 1.2, 1.129, 4.13, 4.151, 6.68; Thuc. 1.67, 4.83. 4 See Plut., Cam. 14.3, Mor. 5.4, Them. 18.6, 20.1; Luc., Nigr. 14.1, Symp. 20.4.

24

kavčič

language,5 this paper investigates whether the aforementioned avoidance of DInf containing aorist infinitives could be considered a property of the spoken language of the period. As a consequence, the aim is to examine whether the same phenomenon occurs in the non-literary papyri contemporary with the New Testament. The corpus examined consists of around 150 non-literary papyri dating to the 1st and early 2nd c. ce (around 30,000 words). Because the paper is primarily concerned with what could be considered the spoken language of the period, it distinguishes between private and official documents. The latter are in general believed to be more archaic in style than the former, although private letters are not stylistically uniform and do not lack traces of the written tradition.6 It has also been suggested, most explicitly by Thorley (1989:296), that DInf were mostly used to express states in the spoken language of the period. This claim, as pointed out below, raises several issues concerning other processes in Hellenistic and Roman Greek. Nevertheless, similar observations, although without explicitly drawing the same conclusion as Thorley (1989), have been made elsewhere (see Fanning 1990:401; Kavčič 2009). As a consequence, this paper studies the use of the present, the perfect, and the future infinitive in DInf, in addition to the aorist infinitive, examining whether the aforementioned assumption concerning the stativity of DInf applies in the case of the corpus of non-literary papyri. In order to examine whether or not this assumption applies in the case of DInf rather than in the case of their syntactic variants (finite complement clauses introduced by ὅτι), the paper also studies the use of corresponding finite verbal forms in such clauses. After presenting previous findings on DInf in the New Testament, the paper examines the use of the present, the aorist, the perfect, and the future infinitive in DInf in the corpus under investigation, as well as the use of corresponding finite forms in complement clauses introduced by ὅτι. The last part of the paper is dedicated to interpreting these data, and pointing out the complexity of their relation to other developments in Hellenistic and Roman Greek, primarily the development of the Ancient Greek synthetic perfect and its relation to the aorist.

5 See Horrocks (2010:149). 6 See Luiselli (2010) as well as the editors’ note concerning the style of one of the letters (P.Oxy 2190) from the corpus investigated in this paper: ‘The letter is vigorous and idiomatic in expression and may occasionally show the influence of the writer’s studies’. (Lobel et al. 1941:147).

variation in expressing temporal and aspectual distinctions

3.1

25

DInf in the New Testament

Based on Burton (1898), Fanning (1990), Porter (1989), Thorley (1989), and Kavčič (2009), this section summarizes previous findings on the use of present, perfect, aorist, and future infinitives in New Testament Greek. New Testament Greek frequently employs the present infinitive in DInf. However, there is a very strong tendency towards using stative present infinitives. The high frequency of DInf containing the infinitive εἶναι ‘be’ is particularly striking. Although the frequency of this verb is approximately the same in all Greek texts as in the New Testament (1.5–2.2 percent), its frequency is much higher in the construction discussed: 44 out of 396 words or 11.11 percent of New Testament DInf are the infinitive εἶναι. This verb is therefore about five to eight times more frequent in DInf than in Greek texts in general. In addition, the frequency of another stative infinitive, ἔχειν ‘have’, is also very high. The infinitive ἔχειν is the second most frequent present infinitive in DInf and occurs six times,7 thus more common in DInf than the average in Greek texts and in the New Testament. Although its frequency is approximately half a percent in all Greek texts, its frequency in DInf is approximately 1.5 percent.8 Several other present infinitives also occur in New Testament DInf.9 Although non-stative present infinitives seem to occur, certain cases of such clauses are rare, reaching at most 7 percent of all present infinitives used in New Testament DInf.10 As already stated, an observation that goes back to Burton (1898) claims that the aorist infinitive is not used in New Testament DInf. This claim has been repeated and partially supplemented in several modern studies.11 Nevertheless, Porter (1989:389) finds four counterexamples. One of these passages is Acts 15.11: διὰ τῆς χάριτος τοῦ κυρίου πιστεύομεν σωθῆναι ‘we believe that we are saved through the grace of the Lord’.12

7 8 9

10 11 12

Acts 12.15, 24.9; Lk 8.18; Jn 5.39; 1 Cor 7.40; and Jas 2.14. See Kavčič (2009:156). Other stative present infinitives in the New Testament include: ὑπάρχειν ‘exist’ (1Cor 7.26), ζῆν ‘be alive, live’ (Lk 24.23, Acts 25.19), μέλλειν ‘be about to’ (Acts 11.28, 25.4), δικαιοῦσθαι ‘be deemed righteous’ (Rom 3.28), λανθάνειν ‘be hidden, unknown’ (Acts 26.26), συνιέναι ‘understand’ (Acts 7.25), possibly ἄρχειν ‘rule, be a ruler’ (Mk. 10.42), and in at least one case λέγειν ‘say’ (Mt 3.9, Rom 3.8, Heb 11.24), and also the passive infinitive λέγεσθαι ‘be named’. See Kavčič (2009) for a more detailed account. See Porter (1989:389); Thorley (1989:295–296); and Fanning (1990:401). Other cases include Lk 2.26; Acts 10.43, 23.12. The infinitive in Acts 23.12 (ἀνεθεμάτισαν

26

kavčič

One of the most striking features of DInf in the period investigated concerns the frequencies of perfect and aorist infinitives in DInf. The perfect is believed to be the least common tense in Ancient Greek, as shown in the following scale (Porter 1989:181, 246): Perfect < present/imperfect (10,602 words) < aorist (11,653 words)13 The situation in New Testament DInf is quite different: perfect infinitives represent 21 percent of all infinitives, thus being less frequent than the present infinitive (more than 70 percent), but clearly more common than the aorist infinitives (at most 6 percent, if all dubious cases are counted).14 Thus the New Testament DInf display the reverse ratio between the aorist and the perfect compared to what is believed to be the situation in language in general.15 In addition, three of the perfect infinitives (ὑστερηκέναι in Hebrews 4.1 and in 2Corinthians 11.5, εὐαρεστηκέναι in Hebrews 11.5, and ἐκπεφευγέναι in Acts 16.27) used in DInf are the only perfect forms of their respective verbs found in the New Testament. Future infinitives are rather rare in the New Testament, which presents only one certain case of a DInf containing a future infinitive: ἔσεσθαι (Acts 23.30).16

3.2

Non-Literary Papyri of the Roman Period

3.2.1 The Present Infinitives Used in DInf The corpus investigated contains the following present infinitives used in DInf: ἀγνοεῖν ‘not know’, ἀναγράφ]εσθαι ‘be registered’, ἀμελεῖν ‘be neglectful’, ἀνήκειν (?) ‘belong’, ἀπέχειν ‘have (in full)’ (eight cases), ἀπετίστε ‘demand’ (?), γρά]φειν ‘write’, διατρίβειν ‘reside’, διδόναι (διδόν[αι, one or two cases) ‘give’, δύνασθαι ‘can’ (three cases), ἐγκαλεῖν ([ἐγ]καλεῖν, two cases) ‘charge’, εἰδέναι (εἰδέν⟨αι⟩, [εἰδ]έν̣ ̣αι, six cases) ‘know’,17 εἶναι (ε]ἶναι, ε[ἶν]αι, εἶν[α]ι, sixteen cases)

13 14 15 16 17

ἑαυτοὺς λέγοντες μήτε πιεῖν μήτε φαγεῖν: ‘they bound themselves with an oath not to eat or drink anything’) seems to be dynamic rather than declarative. In addition, it contains the infinitives πιεῖν and φαγεῖν, on which see Dolger (1953) and Joseph (1983:60). See below 3.2.3, Table 3.4. An example is Jn 12.29: ὁ οὖν ὄχλος ἔλεγεν βροντὴν γεγονέναι ‘the crowd said that it had thundered’ (see Kavčič 2009:159). See Porter (1989:392) on high frequencies of perfect infinitives in the New Testament. Another potential case is Jn 21.25 (see Kavčič 2009:163, and below, 3.2.6). It may seem unusual that this infinitive is treated as a present rather than as a perfect

variation in expressing temporal and aspectual distinctions

27

‘be’, ἐνέχεσθαι ‘be held in, be subject to’, ἐπικρατεῖν (two cases)/ἐπικρατεῖσθαι (ἐπικ̣ [ρατ]εῖσθαι, two cases) ‘have power over’, ἐπ ̣ ̣ ιστασθαι ̣́ ‘understand’, ἔχειν (ἔχ]ειν, ἔχιν, eighteen cases) ‘have’, εὐδοκεῖν ‘be satisfied’, γεωργεῖν ‘cultivate’/ γεωργεῖσθαι ‘be cultivated’, ισφερι ̣ν ‘contribute’, κρατεῖν ‘rule over, take control’, ποιεῖν ‘do’ (one or two cases), συνεῖναι ‘be together’, ὑγιαίνειν ‘be healthy’, ὑπάρχειν (ὑπάρχιν) ‘exist’, and ἐκτίνειν ‘pay off’. This list of present infinitives is rather extensive, although their total number depends on the interpretation of some less clear passages, which are discussed below. The infinitives εἶναι, (ἀπ)έχειν, and εἰδέναι, as well as some other present infinitives, occur more than once in the DInf investigated, some of them in fixed expressions such as καλῶς ἔχειν ὑπέλαβον: ‘I thought it was right’ and διὰ τὸ φάσκειν (τινὰ) μὴ εἰδέναι/ἐπίστασθαι γράμματα: ‘because of saying that (someone) is illiterate’. Other examples commonly occurring in various periods include bgu 1776, sb 7341, sb 7463, P.Lond 1168, bgu 2246, and P.Ryl 173 (see Mandilaras 1973:331). These DInf are clearly stative. Nevertheless, the present infinitive of nonstative verbs is used as well. An example is this petition from the 1st c. ce: (1) τοῦ δὲ κωμογραμματέ]ω̣ ς ἀντιφωνήσαντός σοι εὑρηκέναι μὲν ἐκ [τινων βιβλίων σ]ωζομένω[ν ἐν τῷ ἱερῷ] δι’ ἡμῶν τῶν ἱερέων ἀναγ[ρ]αφόμενος τὰς ἀρούρας διὰ σπόρου [γραφῆς τοῦ […] (ἔτους) καὶ?] νενεμῆσθ[αι] ταύτας [κατὰ] διαδοχὴν τῶν γονέων καὶ γεωργεῖν ἀπὸ τῶν Πετρωνίου [χρόνων ἐπὶ […] οι]ς ἐκφορίοις [τ]οσούτων ἐτῶν ἱερέων ἡμῶν ὄντων νομῆι διὰ τὸ μὴ ἐκ τοῦ [δημοσίου σύντα]ξιν ἡμᾶς [λ]αμβάνειν, ἄλλως δὲ καὶ δι’ ὧν καταχωρίζομεν λόγων καθ’ ἔτος [τῷ στρατηγῷ ὑπὸ τ]ῶν ἱερέων γεωργεῖσθαι τὰ ἐδάφη ταῦτα. […] ἐδηλώσαμεν τὴν τῶν προ[κειμένων ἐδαφῶν γεωργίαν ἡμῖν ἀναγράφ]εσθαι ἀπὸ τῶν προγεγραμμένων χρόνων, ὃς λόγος [σώζεται ἐν τοῖς […]

infinitive. It is clear that, historically speaking, οἶδα is a perfect of εἶδον. As a consequence, it is usually referred to as a perfect, although Smyth (1963:217) treats it apart from the perfect among ‘irregular mi-verbs’ because of its irregular conjugation. Moreover, it has to be stressed that, according to Chantraine (1927:217), at least some of the ag stative perfects tended in Hellenistic and Roman Greek towards being separated from their original conjugations and can be considered irregular presents. This is highly significant in terms of this paper. Namely, it is claimed below (see 3.2.2) that perfect infinitives are more common in Hellenistic and Roman Greek DInf than aorist infinitives. Because it is not entirely clear whether (particularly in Hellenistic and Roman Greek) εἰδέναι can still be treated as a perfect infinitive, the latter claim is more convincing if εἰδέναι is treated as a present infinitive rather than as a perfect infinitive; see footnote 32.

28

kavčič

‘The comogrammateus answered you, saying that he had discovered firstly from certain documents preserved at the temple that the land was registered in the name of us the priests in a list of crops of the … year, and had been possessed by us by inheritance from our ancestors and cultivated from the time of Petronius at the (fixed?) rents, the possession of it by us the priests for so many years being due to the circumstance that we received no subvention from the State, and secondly that according to the yearly reports which we make to the strategus this land was cultivated by the priests. […] we showed that the cultivation of the aforesaid land was registered in our names since the above-mentioned period, which list is preserved […]’.18 P.Tebt 302, ce 71–72

In the passage quoted, the participle ἀντιφωνήσαντος governs the perfect infinitives εὑρηκέναι and (probably) νενεμῆσθαι, in addition to the present infinitives γεωργεῖν and γεωργεῖσθαι. According to the modern edition, another present infinitive used in the same passage could be the passive infinitive ἀναγράφεσθαι, dependent of ἐδηλώσαμεν, but partially restored. The verb γεωργῶ ‘cultivate’ is not stative, nor is ἀναγράφω ‘register’. However, it can be argued that γεωργεῖν does not refer to the activity of cultivating the land, but to the state or habit of having and cultivating it. Landowners in the 1st c. ce (in this case a group of priests) were presumably not reduced to tilling the land themselves. Although this is not the only option, this clause can be interpreted as habitual and therefore stative.19 The passage contains another infinitive of a non-stative verb, which is passive (γεωργεῖσθαι),20 the same as the partially restored ἀναγράφεσθαι (see below 3.5, footnote 57, on the voice of DInf). In addition, the corpus investigated contains the following instances (some of them uncertain), of non-stative present infinitives occurring in DInf. i. If the reading of the modern edition is accepted, a potential instance of such a clause is found in cpr 15.2qtpl, which contains the infinitive ποιεῖν and perhaps γράφειν.

18 19 20

Translation Grenfel-Hunt-Goodspeed (1907:91). See Smith (1997:34). The infinitive λαμβάνειν ‘get, receive’, on the other hand, does not seem to occur in a DInf, as is also suggested by the English translation of the passage.

variation in expressing temporal and aspectual distinctions

29

(2) ἐπικ[ελεύει γυνὴ ἱέρεια Σούχο]υ Τομσά̣εις ἐμνίθου Χα[ι]ρήμονος μητρὸς Ταμεσ̣ θασ̣ ύτμιος ἡ τούτου γυνὴι λέγουσα [γράφειν] καὶ ποιεῖν κατὰ π[άντα τὰ] π̣ ρογεγραμμένα ‘Consente la moglie, sacerdotessa di Suchos, Tomsais, figlia dell’ emnithes Chairemon e di Tamesthasytmis sua moglie, dicendo: “Scrivi e agisci secondo tutto ciò è sopra scritto” […]’.21 cpr 15.2qtpl = sb 5231, ce 11

According to the Italian translation, the infinitive clause is dynamic rather than declarative. ii. κρατεῖσθαι in P.Congr xv.15, and perhaps ἀπετιστε in sb 8070, representing, as it seems, ἀπαιτεῖσθαι.22 The reading of the latter passage (sb 8070), which seems to contain another instance of a passive non-stative present infinitive, is far from certain. Because the governing verb is unclear, it is not certain whether it is a DInf or not. (3) ἐδαφῶν δὲ ἐκφορίοις λογιζο(μένων) καὶ ἄλλων ἐν ἰδιωτικοῖς οἱ κωμογραμματεῖς τὰ ἐν ἀπαιτήσει ὄντα ἐκφόρια καὶ καθήκο(ντα) τῷ β (ἔτει) ἠλασώθη(σαν) δηλοῦντες τὴν γῆν μὴ δύνασθαι διὰ γεωρ(ῶν) σπείρεσθαι μηδὲ ἄλλως προσοδεύειν τούς τε γενομένο(υς) αὐτῶν δεσπότας οὓς μὲν τετελευτηκέναι οὓς δὲ ἀνακεχω(ρηκέναι) καὶ ἐκ τῶν ἐδαφῶν τούτων κρατεῖσθαι διὰ τὸ ὀφείλειν αὐτοὺς τὰ πλεῖστα κεφάλ(αια) ὑπὲρ ὧν ἐξέβαλον μισθώσεων. ‘Of lands reckoned in the class paying rent in kind and others in the class of private land, the village clerks were in the second year relieved, as to the payments in kind and the fixed dues which were collectible, on their showing that the land cannot be sown by the public farmers nor otherwise yield revenue, and that of the former owners, some have died, and others have run away, and that, out of the lands (paying rents in kind) they

21

22

Translation Messeri Savorelli (1990:20). I cite modern translations if available, which are, as also indicated by this very passage, particularly significant in ambiguous cases (see also below 3.2.3 on P.Mich 2.121). If the source is not stated, the translation is mine. English translation of the passage above: ‘The woman, the Suchos priest Tomsais, daughter of emnithes Chairemon and his wife Tamesthasytmis, agrees saying to write and act in accordance with everything written above’. The infinitives σπείρεσθαι ‘sow’ and προσοδεύειν ‘receive revenue’ (P.Congr xv.15) are governed by δύνασθαι ‘can’.

30

kavčič

(the clerks) have taken control of these (ownerless or unleased lands) on account of the fact that they (the lessees) owe very large sums of the leases they had assumed’.23 P.Congr xv.15, ce 80

(4) ω̣ σεν ὀφιλ(όμενα?) ἀπετιστε διὰ τῶν ἀπὸ τῆς κώμης ‘[…] that demands (?) what they owe (?) through those from the village’. sb 8070 = P.Fouad 14 = P.Bacch 24, ce 125–175

iii. The infinitive διδόναι in P.Dura 18, probably dependent on ἐξωμολογήσατο, in addition to ἐκγαλεῖν and ἐκτίνειν, dependent of ὁμολογεῖ in P.Mich 194. The infinitive διδόναι is restored. However, the same sentence occurs twice in this (double) document, and the other reading is more certain. (5) […]ἐξω[μ]ο̣λ̣ογ̣ ̣ ή̣ σ̣ α̣[το] Ν̣ [ι]κ̣ [αν]ω̣ [ρ] ὁ̣ α̣[ὐ]τ̣ο̣[̀ ς] [διδόναι ̣ ̣ ̣]ι ̣σε̣ι ̣ν̣ κ̣ α̣τ[̣ ὰ τὰ π]ε̣ρὶ̣ τ[ῶ]ν δό̣σ̣εων δεδ[ο]γ̣ματ̣ι ̣σ̣ μέ[̣ ν]α [Τι]μων̣α̣σ́ σηι τ̣[ῆι αὐτῆι, οὔ]ση̣ ι ̣ αὐτοῦ [πεν]θ̣ερ̣ᾶι, ἀνθ’ ὧν ἔφη [εὐχρ]ηστηθῆναι αὐ[τ]ὸ̣ν̣ ὑπ’ α[ὐτῆς ἐν] πλ̣ είοσιν̣ ‘This Nicanor has acknowledged that he gives under the provisions of the ordinance on gifts to this Timonassa, who is his mother-in-law, in consideration of favors which he has stated that he has received from her on numerous occasions’.24 P.Dura 18, 87 ce

(6) ὁμολογεῖ Θερμούθιον Ἀχιλλέως τῶν ἀπ’ Ὀξυρύγχων πόλεως μετὰ κυρίου [τ]οῦ τῆς μητρὸς αὐτῆς Πλουτάρχης ἀνδρὸς Δ[ι]ογένους τοῦ Πασίωνος τῷ γενομένῳ αὐτῆς [ἀ]νδρὶ Διοσκοῦτι Παυσίριο(ς) τῶν ἀπὸ τῆς αὐτῆς πόλεως, ἐν ἀγυιᾷ, ἀπέχειν παρ’ αὐτοῦ ἀργυρίου Σεβαστοῦ καὶ Πτολεμαικοῦ νομίσματος δραχμὰς διακοσίας κεφαλαίου […] καὶ μηθὲν [ἐγ]καλεῖν μηδ’ ἐγκαλέσειν μηδ’ ἐπελεύσεσασθα[ι Θερμούθ[ι]ον μηδ’ ἄλλον ὑπὲρ αὐτῆς τῷ Δι[ο]σκοῦτι μηδὲ τοὺς παρ’ αὐτοῦ μήτ(ε) περὶ τούτων μηδὲ περὶ μηδενὸς ἁπλῶς μέχρι τῆς ἐνεστώσης ἡμέρας ‘Thermouthion, daughter of Achilleus, a resident of the city of Oxyrhynchus, acting under the guardianship of Diogenes, son of Pasion, husband

23 24

Translation Pearl (1978:84). Translation Perkins (1959:102).

variation in expressing temporal and aspectual distinctions

31

of her mother Ploutarche, acknowledges in the street to her former husband Dioskous, son of Pausiris, a resident of the same city, that she has received from him a capital sum of two hundred drachmai […] and neither Thermouthion nor another on her behalf makes or shall make any accusation or shall proceed against Dioskous or his heirs about these things or about anything at all up to the present day’.25 P.Mich 194, 61 ce

Because these infinitive clauses seem to be posterior to rather than simultaneous with the main clause, they can be interpreted as dynamic. iv. Additional cases of non-stative present infinitives are perhaps (if the readings below are accepted) ἐλπίδα ποιεῖν and εἰσφέρειν in P.Bad 35. The former could perhaps be interpreted as stative (ἐλπίδα ποιεῖν ‘hope’). (7) γίνoσκε ἐλπίδα ἀνευρε[τικὸν ποιεῖν]. […] γινοσκε […] ισφερι ̣ν (δραχμὴν) α καὶ Ἄνθουσαν (ὀβολοὺς) […] ὡς καὶ] Ταῆσιν· γίγνοσκε μὴ αἰληφέναι με παρὰ Προσδοκίμου τὸν χαλκόν. ‘Know that I hope desperately(?) […] Know […] that contributes a drachma and Anthousa obolous […] as well as Taesis. (?) Know that I have not received bronze from Prosdokimos’. P.Bad 35, 87 ce

Although at least some of these passages may not be stative, there are many others that do permit a stative interpretation. The most certain cases of DInf containing non-stative present infinitives seem to occur in P.Tebt 302 (see above). Examples of DInf containing a stative present infinitive include the fixed expressions that I mentioned earlier, in addition to passages such as P.Lips 105 and sb 5233: (8) ἐχθὲς κατέσχον τὸν φύλακα νομίζων σοι δύνασθαι πέμψαι ὃν ἐπεζήτησας λόγον ‘Gestern wurde ich des Wächters habhaft, in der Meinung dir die Berechnung, die du verlangtest, schinken zu können’.26 P.Lips 105 = Chr.Wilck 237, 1st–2nd c. ce 25 26

Translation Winter (1936). Translation Mitteis (1906:307). English translation: ‘I held back the guard yesterday, thinking that I can send you the account you asked me to’.

32

kavčič

(9) ὁ αὐτὸς ἐδήλωσεν ἔχιν ἐν τῇ κώμ(ῃ) ἐλαιουργ(εῖον) ‘He declared that he had an oil press in the village’. sb 5233, 14 ce

In general, stative infinitive clauses seem to be significantly more frequent in the corpus investigated than non-stative ones. The frequency of non-stative present infinitives in DInf ranges between 8 percent and 14 percent, depending on which of the potential cases of non-stative present infinitive are counted. If some of the less probable cases of non-stative DInf are not counted, then the ratio between stative and non-stative present infinitives in DInf is 92:8 percent.27 However, this does not necessarily prove that DInf tended to be stative in the corpus investigated. It seems reasonable to assume that stative verbs are generally more frequent than non-stative verbs (the most frequent verb, εἰμί ‘be’, is stative, as is another very frequent verb, ἔχω ‘have’).28 It has been claimed above that in the New Testament the infinitives ‘be’ and ‘have’ are significantly more frequent in DInf than in the language in general. However, the infinitive ‘be’ is not as frequent in the contemporary non-literary papyri, its frequency in DInf being only around 1 percent of words (whereas it reaches approximately 11 percent of words in the New Testament).29 This frequency is slightly higher when only private letters are taken into account (reaching approximately 3 percent of words in DInf). In the New Testament, 47 percent of all present infinitives used in Dinf are the infinitive εἶναι.30 Again, its frequency is lower in the non-literary papyri, where approximately 22 percent (sixteen out of seventy-three cases) of all present infinitives are the infinitive εἶναι. These data are presented in Table 3.1 (the data concerning the frequencies

27

28

29

30

This number does not take into account the restored infinitives (e.g., ἀναγράφεσθαι in P.Tebt 302, γράφειν in cpr 15.2qtpl, ποιεῖν in P.Bad 35), the dynamic infinitive ποιεῖν (cpr 15.2 qtpl), and the infinitives ἀπετιστε (sb 8070) and ισφερι ̣ν (P.Bad 35). In the latter two cases, the entire context is unclear. Compare maximum and minimum frequencies of some commonly used (stative and non-stative) Greek verbs: εἰμί ‘be’ (1.5–2.26), ἔχω ‘have’ (0.35–0.54), οἶδα ‘know’ (0.11–0.15), λέγω ‘say’ (0.34–0.45), γίγνομαι ‘become’ (0.51–0.5), φέρω ‘bring’ (0.091–0.093), and ἔρχομαι ‘come’ (0.12–0.13); data according to the Perseus Digital Library. This number refers to the frequency of the verb ‘be’ in DInf in the corpus investigated. The number of words for all DInf is approx. 1,825, and DInf occurring in private letters contain approximately 200 words. See Kavčič (2009:155–156).

variation in expressing temporal and aspectual distinctions table 3.1

33

Frequency of the verb εἰμί

Frequency of the verb εἰμί (in %)

Greek texts in general New Testament DInf (New Testament) DInf (non-literary papyri) DInf (private letters)

table 3.2

Min.

Max.

1.5 1.41

2.26 1.79 approx. 11.11 approx. 1 approx. 3

Frequency of the verb ἔχω

Frequency of the verb ἔχω (in %)

Greek texts in general New Testament DInf (New Testament) DInf (non-literary papyri)

Min.

Max.

0.35 0.28

0.54 0.51 1.5 1

of the verb εἰμί in Greek texts in general and in the New Testament were collected from the Perseus Digital Library). Nevertheless, the frequency of the second most common stative verb (ἔχω) in DInf contained in the corpus investigated is remarkably high, its frequency being slightly greater than that of εἰμί (the latter is represented with sixteen instances, whereas ἔχω occurs eighteen times) and only slightly lower than the frequency of the same verb in New Testament DInf. This is shown in Table 3.2. (As in the case of the verb εἰμί, the data concerning frequencies in Greek texts in general and in the New Testament were collected from the Perseus Digital Library.) This frequency could be an indication of DInf tending to be stative, although the fact that the infinitive ἔχειν is more frequent than εἶναι at the same time suggests that the corpus of non-literary papyri investigated does not entirely represent the situation in the language as a whole.31 31

Many ἔχειν infinitives go back to one document (P.Mich 2.121).

34

kavčič

3.2.2 The Perfect Infinitives Used in DInf Passages already quoted (P.Tebt 302, P.Bad 35, P.Congr xv.15) contain cases of perfect infinitives used in DInf; for example, νενεμῆσθαι, αἰληφέναι (= εἰληφέναι), and τετελευτηκέναι (see above, 3.2.1). An additional example is: (10) Πομπηιανός μοι ἔγραψε τὸν ἵππον τοῦ ἀδελφοῦ σου κατεληλ[υ]θέναι εἰς Ἀλεξάνδριαν καὶ τετειμ̣ ογ̣ ραφῆσθαι ὑπὸ τοῦ ἡγεμόνος δραχμῶν̣ διακοσίων· ‘Pompeianos wrote to me that your brother’s horse has come down to Alexandria and has been evaluated by the prefect at two hundred drachmae’. sb 18.13303, 1st c. ce

The corpus examined contains the following cases of perfect infinitives used in DInf: ἀνακεχω(ρηκέναι) ‘have run away’, ἀνῆχθαι ‘have been brought up’, ἀπενηνοχ(έναι) ‘have carried away’, ἀπεσχηκέναι (ἀπησ̣ χηκέναι, three cases) ‘have had (in full)’, ἀποδεδωκέναι ‘have given back’, ἀπολελοιπέναι ‘have left behind’, ἀπομεμερι(κέναι) ‘have divided’, γεγε[ννῆ]σθαι ‘have been born’, γεγονέναι (γενονέν̣ αι, two cases) ‘have become’, δεδανεικέναι (δ̣εδ̣ ̣α̣ν̣εικέν̣ ̣α̣ι ̣, one or two cases) ‘have borrowed’, δεδωκέναι ‘have given’, διηρπακέναι ‘have torn’, εἰληφέναι (αἰληφέναι, two cases) ‘have got’, ἐληλ̣υ̣θ̣[έ]ναι ‘have come’, ἐνενηνέχθαι ‘have been brought’, ἐνπεριει[λη]φέναι ‘have included’, ἐπι[δ]ε̣δ̣η̣[μ]ηκέναι ‘have come home’, εὑρηκέναι (εὑρηκέ[̣ ναι, two or three cases) ‘have found’, ἠγορακέν(αι) (ἠγ]ο̣ρακέναι, three cases) ‘have bought’, κατεληλ[υ]θέναι ‘have come (down)’, κατηντη[κέναι ‘have arrived’, λελα]χέναι ‘have obtained (by lot)’, μεταδεδό[σθαι] ‘have been transmitted’, μεταδεδωκέν̣ ̣[α]ι ̣ ‘have transmitted’, μετεστάσ̣ θαι ‘have changed (position)’, νενεμῆσθ[αι] ‘have been possessed’, παραπεπτωκέναι (παραπεπτωκ[έναι, two cases) ‘have fallen (away)’, παρεστακέναι ‘have placed (beside)’, πεπομφέναι ‘have sent’, πεπρακέναι (πεπ[ρακέ]ναι, π]επρακέναι, four cases) ‘have done’, περιλελύσθαι ‘have loosened (around)’, πεφροντικέν̣ αι ‘have taken care’, προσπεφ[ωνηκέναι ‘have addressed’, [προσανειλη]φέν[αι] ‘have taken in (beside)’, συμ[πε]φωνηκέναι ‘have agreed’, συνκεχωρηκέναι ([σ]υνκεχωρηκέναι, two cases) ‘have come together’, τετελειῶσθαι ‘have been completed’, τετελευτηκέναι ([τετελευτηκέν]α̣ι ̣, four cases) ‘have passed away’, and τετιμογραφῆσθαι ‘have been evaluated’. Some of the readings are uncertain (see 3.2.4), which is why the number of perfect infinitives may vary according to which readings are accepted. Nevertheless, this rather extensive list in itself suggests that the perfect infinitive is very commonly used in DInf. It occurs in private letters (nine cases) and in official documents (forty-four cases). As a consequence, it seems that DInf con-

variation in expressing temporal and aspectual distinctions table 3.3

35

Frequencies of present, perfect, aorist, and future infinitives used in DInf

pres.inf

pf.inf

aor.inf

fut.inf

73 47 %

53 34.5%

7 4.5%

21 14%

taining a perfect infinitive were not marked stylistically. Mandilaras (1973:329) observes that the perfect is the most common tense in what he calls ‘ὅτι infinitive clauses’ (this refers to non-literary documents in general): ‘Certain verbs such as γράφω, δοκιμάζω, ἐλπίζω, ἡγοῦμαι, λέγω, μανθάνω, οἶδα, φημί are combined with an infinitive corresponding either to a ἵνα or to a ὅτι clause. In the latter case the most common tense of the verb of the ὅτι-infinitive is the perfect’. Even so, the present infinitive is slightly more frequent in the corpus investigated than the perfect infinitive, as shown in Table 3.3.32 The present infinitives counted are all listed in 3.2.1 except for the restored ones (e.g., ἀναγράφεσθαι in P.Tebt 302, ἀνήκειν in P.Vind.Bosw 1, ποιεῖν in P.Bad 35) and the infinitive απετίστε in sb 8070 (on which see above, 3.2.1). In the case of the perfect infinitives, all instances listed are counted except for the (partially) restored ones (e.g., [προσανειλη]φέν[αι] in sb 5236, [τετελευτηκέν]α̣ι ̣ in psi 1109). In the case of the aorist infinitives, all potential cases are counted except for the restored ones (ἐλθεῖν in P.Vind.Bosw 1 and one case of ἐκστῆναι in P.Dura 18). Particularly in the case of aorist infinitives, the number of convincing cases is even smaller (three; namely, the aorist infinitives occurring in P.Dura 18 and in P.Oxy 985; see below 3.2.3). It has been claimed above (in 3.1) that in the New Testament three of the perfect infinitives used in the construction are the only perfect forms of their respective verbs found in the New Testament. Non-literary papyri seem to indicate that this phenomenon of New Testament Greek is a reflection of the use of the perfect infinitive form in Hellenistic and Roman Greek rather than a mere coincidence. Thus Mihevc notes ‘clichés’; that is, infinitive clauses containing perfect infinitives. These were mainly dependent on ὁμολογῶ ‘admit’ and δηλόω ‘clarify, declare’, and were used well into the Byzantine period, although the

32

However, the frequency of the perfect infinitive would be even higher if I counted the infinitive εἰδέναι (which is a perfect infinitive from a historical perspective) in this group. I include this verb among the present infinitives (see footnote 17).

36

kavčič

corresponding perfect indicative forms seem to have already disappeared by then.33 An example is bgu 371 (7th c. ce): ὁμολογῶ ἐ[σ]χη̣ κ[έναι] καὶ δεδέχθαι καὶ πεπληρῶσθαι παρὰ σοῦ ‘I admit that I have got, that I have received, and that I have been paid by you’. 3.2.3 The Aorist Infinitives Used in DInf Aorist infinitives also occur in the construction investigated. The reading of the next passage is uncertain: if ποῆ̣ σ̣ ε is interpreted as an incorrectly spelled ποιῆσαι, then it is a case of an aorist infinitive. However, it refers to a posterior action, as do some cases of non-stative present infinitives (see 3.2.1). For this reason, such clauses should probably be interpreted as dynamic. (11) ἐπορεύθην πρὸς Ἑρμογένην τὸν κωμογρ[α]μματέαν, καὶ ὁμολόγησέ μοι ποῆ̣ σ̣ ε τὴν ἀναβολήν. ‘I went to Hermogenes the comogrammateus and he consented to make the delay’.34 P.Oxy 1480, 32 ce

Although the reading of the passage is not entirely clear, it seems to contain a weak aorist infinitive ποιῆσαι, formed from the s-stem. According to Gignac (1981:296), these verbal forms were frequently confused with others in the period investigated: ‘This mutual influence of future and aorist led to the frequent interchange of endings between these tenses and to the occasional substitution of one tense for the other (if not simply misspellings in the endings)’.35 As a consequence, the passage could also be read as containing a future infinitive rather than an aorist one (see below 3.2.6).36 The corpus investigated contains another potential case of an s-stem aorist infinitive, namely ἐπιτελέσ̣ α̣ι in P.Mich 2.121. However, the reading is uncertain. In addition, it does not seem to refer to anteriority, as also suggested by the English translation of the passage, which is ‘shall cede’ (Boak 1933:43).

33 34 35 36

Mihevc (1959:9). Translation Grenfell-Hunt (1916:239). See also Gignac (1981:33); Markopoulos (2009:54–55); and Leiwo (2010:116) on the confusion between the future and the aorist infinitive in the μέλλω + infinitive construction. Leiwo (2010:102) also draws attention to the confusion between aorist infinitive ending in -σαι and the aorist imperatives ending in -σε (Classical Greek -σον) in clauses of the type καλῶς/εὖ ποιήσεις. However, the infinitive clause is not a DInf in this case, which is why I do not consider the option of cases like ποῆ̣ σ̣ ε and ἐπιτελέσ̣ α̣ι representing an imperative.

variation in expressing temporal and aspectual distinctions

37

Another instance of a potential aorist infinitive is P.Oxy 3274 with the infinitive ἐπιδοῦναι. However, the aorist infinitive should be interpreted as dynamic, both according to the context and to the English translation (‘should submit a memorandum’): (12) [ἐ]δ̣ήλωσά̣ σ̣ ο̣ι ̣ τῷ εὐεργέτῃ \τὴν αὐτὴν/ \ἡμίσιαν/ ἀποδεδωκέναι με ὑπὲρ αὐτοῦ χρέα δανισταῖς καὶ ἀπε̣φ̣ή̣νω ἐπιδοῦναί [μ]έ σοι ἀναφόρ[ιον] περὶ τούτων, [ὦν] ἐστιν τ[ὸ] κα[θ’ ἕν.] τῷ μὲν τρίτῳ [ἔτει] Τραιανοῦ Κ̣ [αίσαρος το]ῦ̣ κυρίου [ἀπέ]δ̣ω̣κ̣α̣. ‘I informed you, my benefactor, that I had paid back the said half as debts to creditors on his behalf and you delivered a judgment that I should submit a memorandum to you about these matters, of which the detailed story is as follows. In the third year of Trajan Caesar the lord I paid back’.37 P.Oxy 3274, 54 ce

Another possible case of an aorist infinitive is συνπαθεῖν in the following passage: (13) μεταδόν[τος δ’ ἐμοῦ Φι]λ̣ οξένωι τὴν σὴν γνώμην, τὰ αὐτὰ μὲν [ἐπῄνει…α]ύτην μόνην τὴν τῶν σοφιστῶν ἀπ̣ ορ̣ [̣ ία]ν̣ σ̣ υ̣ν̣π̣α̣θ̣εῖ̣ ν τῆι π[ό]λ̣ ει φάσ̣ κων ‘When I informed Philoxenus of your opinion he agreed, saying that just in this shortage of professors […] was in the same condition as the city’.38 P.Oxy 2190, ca. 100 ce

In this case, too, the reading is uncertain. If the proposed reading is accepted, it could be an example of an aorist infinitive used in a DInf, but only if συνπαθεῖν is interpreted as an aorist infinitive of the verb συμπάσχω. Another option is to interpret it as a present infinitive of the verb συμπαθέω, which seems to have been commonly used in Hellenistic and Roman Greek.39 As indicated by the editors, the passage is rather unclear.40 Nevertheless, the English translation supports a stative interpretation: ‘was in the same condition’ (Lobel et al. 1941:147). 37 38 39 40

Translation Rea (1978:16). Translation Lobel et al. (1941:147). See lsj, s.v. συμπαθέω. ‘σ̣ υ̣ν̣π̣α̣θ̣εῖ̣ ν: the π and the α are both very probable, the σ̣ υ̣ν̣ is very far from certain’ (Lobel et al. 1941:148).

38

kavčič

According to the modern edition, another possible case of an aorist infinitive is ἐλθεῖν in P.Vind.Bosw 1 and (occurring twice in the document, in a repeated sentence) ἐκστῆναι in P.Dura 18; the document also contains a potential case of DInf containing a non-stative present infinitive (see 3.2.1). Although one of the infinitives ἐκστῆναι is partially restored, the reading of the other is more certain. Thus particularly the infinitive ἐλθεῖν in P.Vind.Bosw 1 is far from certain. If these readings are correct, these are DInf, referring to anteriority as in Classical Greek. (14) [Ν]ι ̣κ̣ [άνωρ Ξε]ν̣ο̣κ[ρά]του[ς] τ̣ο̣ῦ Ἀδδαίου, Εὐρω̣ [παῖος, φάμενος Τι]μ̣ ώ̣ ν̣[ασσ]α̣ν̣ [Σ]ε̣[λεύκ]ο̣υ̣ τ̣ο̣[ῦ] Λυσίου, Εὐρ̣ω̣[παίαν ἐ]κ̣ σ̣τη̣ ̣͂ν̣αι ε̣ις̣̓ ̣ αὐ̣τὸν κατὰ χρημα̣̣ τι [τ]ισ[μὸν γεν]ό̣μ̣εν̣ ̣ο̣ν̣ δ̣ια̣ ̣̀ τ̣ο̣ῦ̣ α̣ὐ̣το̣ ̣θ́ ̣ι ̣ β̣α̣σ̣ιλ̣ ̣ ικ̣ ̣ οῦ δι ̣κ̣ αστηρίου̣ [ἐν] τ̣[ῶι ἐ]ν̣εσ̣ ̣ τῶ ἔτει, μην̣ὸς Δίου δεκά[τηι, ἀντὶ] δ̣α̣ν̣ει̣ ό̣̣ υ̣ [οὗ] δ̣εδ̣ ̣α̣ν̣εικέν̣ ̣α̣ι ̣ α̣ὐ̣το̣ ̣ν̀ ̣ Διοκλεῖ Δ̣ αν̣ύ̣[μου] τοῦ Ἀπ̣ ολλ̣ωνίο̣ ̣υ, Εὐρω̣ π̣α̣ίωι ̣, τ̣ῶ̣ ι ̣ γενομ[ένωι τῆ]ς Τιμ̣ ων̣ά̣[σση]ς ̣ ἀ̣ν̣δ̣ρι̣ ,̣́ κ̣ [α]ὶ̣ α̣ὐ̣τῆι τῆι Τιμων̣α̣σ́ ̣ [σ]η̣ [ι ‘Nicanor son of Xenocrates, grandson of Addaeus, of Europos, stating that Timonassa daughter of Seleucus Lysias’ son, had ceded to him by an act accomplished through the local court in the present year, the 10th of the month Dius, in place of a loan which he had lent to Diocles son of Danymus, grandson of Apollonius, of Europus, then husband of Timonassa, and to Timonassa herself […] (had ceded to him)’.41 P.Dura 18, 87 ce

(15) ἐδήλωσε Σοχώτη[ν] [τοῦ δεῖνος καὶ] Πανομιέ[α τοῦ δεῖνος καὶ τὸν δεῖνα τοῦ Στοτ]οήτιος καὶ Σαταβοῦν ἕτ[ερ]ο̣ν Στοτοήτιος καὶ τοὺς σὺν αὐτοῖς ἐκ̣ πληρώσιν [ ̣ ̣ καὶ ἐλθεῖν? ἡ]γουμένου̣[ς καὶ πρεσβυτέρους τ]ῆς προκειμένης κώμ[ης] ‘[…] teilte […] mit, dass nämlich Sochotes, Sohn von […] und Panomieus, Sohn […] und […] Sohn von Stotoëtis, und Satabus, der zweite Sohn vοn Stotoëtis und diejenigen, die mit ihnen waren, bezahlen würden […] und das leitende und ältere Priester des obengenannten Dorfes […] gekommen wären […]’.42 P.Vind.Bosw 1, post 87 ce

41 42

Translation Perkins (1959:102). Translation Boswinkel (1941:5). English translation: ‘He declared that Sochotes, the son of […], and of Panomieus, the son of […], and the son of Stotoetis as well as Satabus, the other son of Stotoetis, and those who came along would pay […] and that (?) the leading and elderly (priests) came (?) from the village which lies in front […]’.

variation in expressing temporal and aspectual distinctions

39

A rather convincing case of an aorist infinitive is [εὐχρ]ηστηθῆναι in P.Dura 18. The passage is cited above (see 3.2.1). The ending -ηστηθῆναι (this reading seems to be quite certain) makes this interpretation plausible.43 There cannot be much doubt that this is a DInf referring to anteriority. Like some nonstative present infinitives, this infinitive is passive (see footnote 57). Another potential case of a passive aorist infinitive is ἀνηλω(θῆναι) in P.Oxy 985 (1st c. ce): σημα(ίνει) ὁ Σαραπίων ἀνηλω(θῆναι) ἐργά(ταις) δυσὶ σκάπτοντ(ι) χοῦν […] ‘Serapion notes that he spent for two workers who were digging the land […]’. The temporal augment (-η-) in the form ἀνηλω(θῆναι) could indicate that this is a perfect rather than an aorist infinitive, or, if the aoristic interpretation is accepted, could be interpreted as a sign of the confusion between the aorist and the perfect or even as a case of a perfect rather than of an aorist infinitive (see 3.2.4). However, this particular verb also occurs with the temporal augment in the present stem, and this phenomenon seems to have been rather common in Hellenistic and Roman Greek (e.g., ἀνηλουμένων in cpr 7.1 and ἀνηλοῦντι in P.Oxy 1143), resulting in the restructured present ἀνηλίσκω.44 As a consequence, the passage is hardly an indication of confusion between the aorist and the perfect and is treated here as a (potential) case of a DInf containing a passive aorist infinitive. Of the seven potential instances of DInf containing an aorist infinitive, three seem to be more convincing than the rest (ἐκστῆναι/εὐχρηστηθῆναι in P.Dura 18, and ἀνηλωθῆναι in P.Oxy 985). Even if all the cases above are accepted as DInf, the aorist infinitives are still significantly less frequent in DInf than the perfect infinitives, the same as in the New Testament (see above, Table 3.3, and below, 3.2.4). As stated, this is contrary to the situation in Ancient Greek in general, where the aorist is, as a rule, more frequent than the perfect (see 3.1). In addition, Porter-O’Donnell (2010) suggest that the perfect is in general less frequent than the aorist and the present in private letters in Hellenistic and Roman Greek, as can be deduced from Table 3.4. Nevertheless, to a certain extent the non-literary papyri of the 1st c. ce show a different picture concerning frequencies of the aorist and perfect forms in general. This is shown in Table 3.5. The relation between the aorist and the perfect is also discussed in the subsequent section (see 3.2.4).

43

44

Because the infinitive is partially restored, it is interesting to observe that Gignac (1981:305) finds cases of a presumed confusion between the middle-passive perfect and the passive aorist. The two cases he quotes go back to the end of the 2nd c. ce; see Joseph (1983:56). See Gignac (1981:251); lsj, s.v. ἀναλίσκω.

40

kavčič

table 3.4

+ = −

Aspect frequencies in private letters according to Porter & O’Donnell (2010)45

Perfective

Imperfective

Stative

94 (41%) 68 (40%) 58 (40%)

93 (41%) 72 (42%) 61 (42%)

41 (18%) 32 (18%) 24 (18%)

table 3.5

Tense frequencies in non-literary papyri (Mandilaras 1973: 59–60)

century 3rd c. bce 2nd c. bce 1st c. bce 1st c. ce 2nd c. ce 3rd c. ce 4th c. ce 5th c. ce 6th c. ce 7th–8th c. ce total

texts

pres

ipf

aor

pf

fut

plpf

fut.pf

43 46 20 58 133 70 54 8 15 9 456

66 95 36 144 378 210 167 19 50 32 1,197

20 28 2 27 45 19 16 0 0 4 161

66 93 16 93 390 222 209 27 50 20 1,186

45 56 27 90 130 90 65 9 25 6 543

17 34 6 53 114 74 64 5 13 4 384

6 21 1 1 11 3 2 0 0 0 45

1 5 1 0 0 0 0 0 0 2 9

According to these data, in all centuries the perfect is significantly less frequent than the aorist, except for the 1st c. bce and the 1st c. ce. Although the perfect is only slightly less frequent than the aorist in the 1st c. ce, it is even more frequent than the aorist in the 1st c. bce (although the number of occurrences for the latter century is rather low). These data are discussed in detail by Mandilaras (1973:59–60), who observes that ‘the perfect holds its frequency quite well throughout the period, although it never (except in the 1st c. bce) surpasses the aorist’, which allows him to attribute this relatively strong position of the

45

The stative = the perfect according to Porter (1989); + refers to letters written from a person of higher social status to a person of lower social status, whereas = and − signify letters addressed to persons of equal social status and higher social status, respectively.

variation in expressing temporal and aspectual distinctions

41

perfect in the non-literary papyri to the ‘aoristic sense’ of the perfect (loc. cit.). This possibility, namely, that the aorist and the perfect had merged by the period investigated in this paper, is also significant in terms of interpreting the construction examined and is discussed below (see 3.5). If the aorist and the perfect had about the same frequency in the period investigated, or if the perfect was even more frequent than the aorist, then the high frequencies of the perfect infinitive in DInf are less striking. Nevertheless, the assumption that the perfect was in general more frequent than the aorist or that the two tenses had about the same frequency in the period investigated could hardly account for what seems to be avoidance of the aorist infinitive in DInf. 3.2.4

On the Loss of Morphological Distinctions between the Aorist and the Perfect The spoken language of the period investigated (and of later periods) was also characterized by a gradual loss of morphological distinctions between the perfect and aorist, a process that is believed to be linked with the functional merger between the two tenses. Thus the perfect stem sometimes occurs with an augment instead of being reduplicated, and the reduplication could be transferred to other stems.46 Examples from the period include ἔμελκε in P.Oxy 1155 (104 ce) and διαγεγραφεῖσαι in O.Tait 1950 (112 ce?; see Gignac 1981:242–248; Horrocks 2010:177). It seems that an additional sign of the confusion between the aorist and the perfect in Hellenistic and Roman Greek can be found in the rather frequent use of perfects such as ἔγνωκα, εἴληφα, and εὕρηκα, which were originally formed without the reduplication and survived well into the subsequent periods (see Chantraine 1927:241–243). The development finally resulted in abandoning the reduplication and the reduplicated perfects. In addition, the functional merger between the aorist and the perfect led to the spread of the suffix -κ, originally used both in the active perfect stem (e.g., εὕρηκα and πεποίηκα) and in some of the active aorist stems (ἔθηκα and ἔδωκα), whereas some of the original perfect forms (with an augment having replaced the reduplication) substituted for the (mostly irregular) aorists. Examples of such perfects include ἕστηκα, ἔβηκα (Classical Greek βέβηκα), and εὕρηκα, resulting in Modern Greek forms such as βρήκα, ανέβηκα, and κατέβηκα (see Chantraine 1927:242; Jannaris 1969:440; Horrocks 2010:302). This raises the question of whether all the perfect (and aorist) infinitives above should be considered as such. An assumption that aorist infinitives could be reduplicated in the period investigated (as in διαγεγραφεῖσαι; see above) 46

See also 3.2.3 on the infinitive ἀνηλωθῆναι and the present ἀνηλίσκω, and Evans (2010:64, fn. 17) on ἀφέσταλκα.

42

kavčič

could of course yield additional potential cases of DInf containing an aorist infinitive and could impact the data above concerning the frequencies of the perfect and the aorist infinitives in DInf. However, in DInf occurring in the corpus investigated, there are no certain cases of this phenomenon. Cases that could potentially allow such an interpretation include the partially restored infinitives κατηντη[κέναι, πεπ[ρακέ]ναι, and προσπεφ[ωνηκέναι. In other passages containing a reduplicated infinitive, the reduplication is clearly added to a perfect rather than to an aorist stem. As stated above, the period investigated also yields examples of perfects such as ἔμελκε, with the augment having substituted for the reduplication. This phenomenon seems to be linked to the functional merger between the aorist and the perfect and could, if attested in DInf, support an aoristic interpretation of perfect infinitives. The list of perfect infinitives above does not contain certain cases of a perfect stem with an augment having substituted for the reduplication. Nonetheless, the question can be raised whether modern restorations such as συμ[πε]φωνηκέναι, λελα]χέναι, and π]επρακέναι are correct. An augmented perfect stem cannot be excluded in these cases, at least as a theoretical option. The substitution of some of the aorist forms by corresponding perfect forms formed with the suffix -κ also raises the question of whether some of the perfect infinitives listed above could be interpreted at least as ambiguous, if not aoristic. As suggested above, this development mostly concerns aorists that can be characterized as irregular (e.g., ἔστην, ἔβην, and εὗρον). As a consequence, the corresponding infinitives such as εὑρηκέναι are not necessarily interpreted as perfect infinitives. Although the final stages of this development seem to primarily concern later periods of Greek,47 Chantraine, as mentioned above, also draws attention to a significantly frequent use of perfect forms that were originally formed without reduplication. This phenomenon seems to concern Hellenistic and Roman Greek and could be related to the functional merger between the aorist and the perfect. Note that Apoc 5.7, which contains the perfect εἴληφεν, is considered in Blass et al. (2001:281) to be a certain case of an

47

See Horrocks (2010:302), who places the final stages of the aforementioned development concerning the morphological confusion between the aorist and the perfect in the Byzantine period. In the case of the infinitive εὑρηκέναι, non-literary papyri from the period investigated (1st–2nd c. ce) as well as from subsequent centuries clearly still use the corresponding aorist form (εὑρεῖν); for example, P.Oxy 2198, P.Muench 3.1.118, P.Oxy 393. In the non-literary papyri dating back to the period investigated (1st–2nd centuries ce), I found three cases of the infinitive εὑρηκέναι and five cases of the infinitive εὑρεῖν. The New Testament contains five instances of εὑρεῖν and one instance of εὑρηκέναι.

variation in expressing temporal and aspectual distinctions

43

aoristic use of the perfect.48 As a consequence, all perfect infinitives that were formed without reduplication (e.g., εὑρηκέναι and εἰληφέναι) from the list above (see 3.2.2) can be considered ambiguous. In addition, it is even more likely that the infinitive εἰληφέναι was aoristic, given that the period investigated displays aoristic use of its indicative form. When all of these are taken into consideration, then the list above contains forty-four clear cases of what were at least originally perfect infinitives (this number does not include the restored cases like συμ[πε]φωνηκέναι). Among these cases, twenty-nine are reduplicated perfects (e.g., γεγονέναι and παραπεπτωκέναι), and the rest are formed without reduplication (e.g., εὑρηκέναι, εἰληφέναι, διηρπακέναι, ἠγορακέναι, and ἀπεσχηκέναι). As stated, the latter particularly could be considered ambiguous from the perspective of the historical development of the Ancient Greek perfect. Even in this case, the perfect infinitives occurring in DInf surpass in number the aorist infinitives in DInf because there are no more than seven of the latter (as stated, three of them seem more certain than the rest, and two of them occur in the same document; see 3.2.3). This applies even if orthographic issues concerning the aorist and the future infinitive are taken into account—in which case, however, the predominance of DInf containing a perfect infinitive over those containing an aorist infinitive is weaker (see 3.2.6). 3.2.5 The Use of Finite Forms in Complement Clauses Introduced with ὅτι The lack of convincing cases of DInf containing an aorist infinitive is not a mere coincidence caused by an incidental uncommon use of the aorist in complement clauses in general.49 The corpus investigated provides clear examples of the aorist (as well as non-stative present) indicative in ὅτι clauses. The following passages contain both stative and non-stative verbs used in the aorist indicative: (16) περὶ οὗ καλῶς ποιήσεις ἀντιφωνήσασά μοι ὅτι ἐκομίσου. ‘Concerning it, please send me an answer that you received it’. P.Oxy 300, 75–99 ce

48 49

The passage is quoted below in 3.5. In this paper, this term refers to finite complement clauses after verbs of saying and thinking introduced with ὅτι. With rare potential exceptions (see sb 11644, 100–199, ὁ θεὸς οἶδεν πῶς σε κατὰ ψυχὴν φιλῶ ‘God knows that I love you in my soul / how I love you in my soul’), I found no convincing cases of finite complement clauses introduced with ὡς (or πῶς) in the corpus investigated.

44

kavčič

table 3.6

Tense frequencies in ὅτι clauses (non-literary papyri)

pres/ipf.ind

aor.ind

pf.ind

fut.ind

58 39%

49 33%

26 17%

17 11 %

(17) ἐδήλωσάς μοι ̣ διὰ τῆς πεμφθείσης μ̣ ο̣ι ἐπιστολῆς ὑπὸ σοῦ ὅτι κατεφρονήσατέ̣ μου. ‘You revealed to me in the letter you sent me that you despised me’. sb 11644, 1st/2nd c. ce

Although the perfect is used in ὅτι clauses, it is significantly less frequent in these clauses than the aorist indicative. An example of a ὅτι clause containing a perfect indicative is P.Oxy 295 (35ce): γίνωσκε ὅτι Σέλευκος ἐλθὼν ὧδε πέφευγε ‘I must tell you that Seleucus came here and has fled’.50 However, the aorist is about twice as frequent as the perfect and about as frequent as the present. This is presented in Table 3.6. The ratio between the perfect and the aorist in ὅτι clauses thus corresponds to the frequency scale which I referred to earlier, applying to tense frequencies in the language as a whole and indicating that the perfect was less frequent in Greek than the aorist (see 3.1 and 3.2.3, Table 3.4). The data in Table 3.6 indicate that, whereas the aorist tended to be avoided in DInf, its use in finite complement clauses (introduced with ὅτι) was quite common. As already stated, the corpus investigated also contains cases of non-stative present indicatives used in ὅτι clauses (whereas, as claimed above, non-stative present infinitives seem to be avoided in DInf). An example is P.Fay 109 (10 bce– 34ce): τοὺς τρεῖς στατῆρες οὓς εἴρηκέ σοι Σέλευκός σοι δοῦναί μοι ἤδη δὸς Κλέωνι, νομίσας ὅτι κιχρᾷς μοι αὐτούς ‘the three staters that Seleucus told you to give me and consider that you are lending them to me’.51 50 51

Translation Grenfell-Hunt (1899:296). Transl. Grenfell et al. (1900:261). Other indicatives occurring in ὅτι clauses include: ἀγωνιῶ, ἀμελῶ, ἀσθενεῖ, βούλομαι, δεῖ, δείρομεν, δοκεῖ, δύναμαι, εἶ, εἰμι (2), εἰσιν, εἶχο̣[ν], ἐνῆν, ἐστιν (8), ἔλεγας, ἐναποκίχραμαι, ἐνισχύω, ἔχει, ἔχω (2), [ἤλπ]ι ̣ζον, ζῆι, ἤθε[λες, ἥκει, ἤμην, θέλει, θέλεις, θέλω (2), θερίζουσιν, κιχρᾷς, λαμβάνει, λέγουσιν, λέγω, κατέβαλλεν, μανθάνει, μέλει (2), μέλλει, μέλλεις, ὀφε[ίλ]ει, ὀφείλω, παραλαμβάνει, παρέχει, πέμπω, ποιεῖς, πωλεῖται, τηρεῖται, φέρει, φοβούμεθα (pres/ipf); ἀνεχώρησαν, ἀνηρχόμην, ἀπέδρα, ἀπέλθε, ἀποκατεστάθησαν,

variation in expressing temporal and aspectual distinctions table 3.7

45

Tense frequencies in ὅτι clauses (New Testament)

pres/ipf.ind

aor.ind

pf.ind

fut.ind

321 58 %

110 20%

43 8%

75 14%

The same—namely, that the lack of aorist infinitives used in DInf could hardly be considered a coincidence—applies to New Testament Greek, which yields many examples of ὅτι clauses containing the aorist indicative. An example is Matthew 5.17: Μὴ νομίσητε ὅτι ἦλθον καταλῦσαι τὸν νόμον ἢ τοῦς προφήτας ‘Do not think that I have come to abolish the law’ (cf. Mk 9.26, Mt 6.29, 13.17, 16.11, Lk 12.51, Jn 17.21). Tense frequencies in ὅτι clauses occurring in the New Testament are shown in Table 3.7. The table shows that in ὅτι clauses in the New Testament, the aorist is—as in the case of non-literary papyri—more frequent than the perfect. However, the same table also shows that the aorist is—unlike the case of non-literary papyri and in the language in general—considerably less frequent than the present (although still more common than the perfect). This difference between the New Testament and the non-literary papyri is perhaps related to a fairly frequent use of direct speech in the New Testament. The latter often contains the aorist indicative. An example is Matthew 19.20: Λέγει αὐτῷ νεανίσκος, Πάντα ταῦτα ἐφύλαξα ‘The young man said to him: “I have obeyed all these”’ (cf. Mt 9.18; Lk 19.18; Jn 4.17, 20.2). 3.2.6 The Future Infinitives Used in DInf As shown in Table 3.3, the corpus examined contains several cases of future infinitives used in DInf. These infinitives include: ἀδικήσειν ‘will do wrong’, ἀπο-

ἐβλάβης, ἔβ̣ ̣λαψ̣ε, ἐγ̣ ̣ ενό̣μ̣ην, ἐγγενήθη, ἔγνωκας, ἔγραψε, ἔδωκα, ἔδωκας, ἔδωκε, εἶπε, ἐκομίσου, ἔλαβες, ἔλαβον (2), ἔμέλησε, ἐμέτ̣ρησαν, [ἐξ]ήντλησα, ἔπεμψα (2), ἔπεμψας (2), ἐξῆ̣λθε, ἐσώθημεν, [ἐφρόν]τισα, ἐπληροφόρησα, ᾐτήσω, ἠδυνήθην, ἤκουσα, εὗρεν, ἔφυ]γεν, ἐχειρογράφησεν, ἦλθε, ἦλθες, ἦλθον, ἤχθην, ἱκόμην, κατέπλευσα, κατεφρονήσατε̣, μετέδωκα, με[τέ]πεμψε, παρήνγειλας, [πα]ρεγενόμεθα, παρεγράφης, προέγραψα (aor); ἀναγορευθήσονται, ἄξει, ἀσπασ̣ ο̣μ́ εθα̣, ἀφήσ[εις, διαμισθώσει, δυνηθήσο̣μαι, εἰσπεσεῖται, ἐργο[λ]αβήσομα[ι], ἕξω, θεραπε[ύ]σω, κριθήσονται, λυπήσεις, μελήσει, μετρήσει, ποιήσει, ποιήσεις, ποιήσω (fut); ἀπέσπακα, ἀπέσταλκεν, βέβληται, γέγονα (2), γέγονεν, δέδωκα, δεδ[ώ]κατε, ἐπιβέβηκα, ἔμελκε, ἕστακα, ἔσχηκας, ἐφθάκαμεν, εὕρηκα, ἤλλαγμαι, ἡμάρτηκα, ἦρκα, ἦρκας, ἠσθένηκας, [π]ροέσχη[κας], συνέστακας, παραπεποίηται, πέφευγε, συνέσχηκε, τεθήκαμ[εν], and ὑπόκειται (pf).

46

kavčič

δώσειν (two cases) ‘will give back’, βεβαιώσειν (βεβαιωσιν, two cases) ‘will confirm’, δεήσειν ‘will have to’, ἐγκαλέσειν ‘will charge’, ἐκτείσειν ‘will pay off’, ἐκπληρώσειν (ἐκπληρώσιν) ‘will pay off’, ἐνοικήσειν (ἐνοικίσειν, two cases) ‘will lodge’, ἐπελεύσεσθαι (ἐπελε]ύσεσθαι, ἐπελεύσασθα[ι], four cases) ‘will come upon’, ἐπιμελήσεσθαι ‘will take care of’, ἐφάψεσ[θ]αι ‘will touch’, παρέξεσθαι (three cases) ‘will supply’, and μεταδώσειν ‘will give a share’. DInf containing these infinitives refer to (posterior) states or activities. Most non-stative present infinitive such as ἐγκαλέσειν, βεβαιώσειν, and ἐπελεύσεσθαι occur in official documents after ὁμολογῶ ‘admit’ or βεβαιόω ‘confirm’. A private letter containing a future infinitive used in DInf is P.Oxy 2190 with the infinitive ἐπιμελήσεσθαι; another potential case is P.Oxy 1480 with the infinitive ποῆ̣ σ̣ ε, which was discussed above in 3.2.3. Unlike in the case of other infinitives, none of the future infinitives occurring in DInf is passive. As already mentioned, the s-stem future and the s-stem aorist forms were frequently confused with one another in the period investigated (see 3.2.3). Thus in fourteen of the cases above an orthographic change of the active ending -ειν into -αι and a change of the thematic vowel -ε- into -α- changes the presumed future infinitive into an s-stem aorist infinitive.52 Examples of such infinitives from the list above include ἀδικήσειν, βεβαιώσειν, and ἐφάψεσθαι. Taking this into consideration could, of course, yield additional cases of aorist infinitives in the corpus investigated. This option has an impact on the statistical data concerning the aorist infinitives in the corpus. DInf containing aorist infinitives remain less common than those containing perfect (and present) infinitives, although the predominance of perfect infinitives over aorist infinitives is weaker in this case. This is presented in Table 3.8. As already discussed, these data are unexpected from the perspective of what is believed to be the situation in the language in general, where the perfect was less frequent than the aorist. In addition, the available translations do not support an aoristic interpretation of the passages: none of the DInf containing one of the infinitives above seems to refer to anteriority, which is a

52

I do not count among these ambiguous cases verbs that do not seem to form an s-stem aorist (e.g., ἐπέρχομαι) in the period investigated. According to Gignac (1981:386), the sstem aorist of the verb δίδωμι is found in the non-literary papyri. However, the examples he cites are quite late (mostly from the 6th and 7th c. ce), which is why the infinitives ἀποδώσειν and μεταδώσειν are not counted as ambiguous; the aorist infinitive δοῦναι seems to have been the normal form in the period investigated (Gignac 1981:393). On the other hand, an s-stem aorist is not excluded in the case of the verb παρέχομαι, which is why this verb is treated as ambiguous in Table 3.8 (see Gignac 1981:295).

variation in expressing temporal and aspectual distinctions table 3.8

Maximum/minimum frequencies of aorist/future infinitives

pres.inf

pf.inf

aor.inf

fut.inf

73 47 %

53 34.5%

7 (+14) 4.5% (14%)

21 (−14) 14% (4.5 %)

table 3.9

47

DInf in private letters

pres.inf

pf.inf

aor.inf

fut.inf

11

9

2?

1

property of DInf containing an aorist infinitive. As a consequence, the statistics presented in Table 3.3 (see 3.2.2) seem to be more probable than those from Table 3.8.

3.3

Distinctions between DInf Occurring in Private Letters and in Official Documents

As stated, private letters differ from official documents in terms of the frequency of the future infinitive in Dinf. Although such clauses are rather common in official documents, only one instance occurs in private letters. Because private letters are generally believed to offer an insight into the spoken language (at least to a higher extent than official documents), the data concerning the use of the DInf occurring in private letters seem particularly significant. They are presented in Table 3.9. The number of DInf occurring in private letters in itself is rather low. Nevertheless, these data seem to support the tendencies that apply to the corpus as a whole: DInf containing an aorist infinitive seem to be avoided, whereas perfect infinitives are used in DInf more often. Three of these are formed without reduplication (μετεστάσθαι in P.Sarap 90, παρεστακέναι in P.Oxy 3057, and αἰληφέναι in P.Bad 35), and the rest are reduplicated (see 3.2.4). Present infinitives are also common, and stative present infinitives greatly predominate over non-stative ones: nine present infinitives are stative, whereas only two cases of DInf containing a present infinitive could be interpreted as non-stative (ἀπετίστε in sb 8070 and ισφερι ̣ν in P.Bad 35). However, these

48

kavčič

two cases are far from certain (see 3.2.1). The two (potential) cases of a DInf occurring in a private letter and containing an aorist infinitive (ποῆ̣ σ̣ ε in P.Oxy 1480 and σ̣ υ̣ν̣π̣α̣θ̣εῖ̣ ν in P.Oxy 2190) are also uncertain (see 3.2.3).

3.4

Tendencies Common to the New Testament and Non-Literary Papyri: Summary

Owing to editorial problems and potential alternative readings of some passages, the interpretation of DInf occurring in the investigated corpus of nonliterary papyri is a more intricate task than seems to be the case with the New Testament. However, there are evident parallels between the language of the non-literary papyri and New Testament Greek. First, DInf in both the New Testament and the contemporary non-literary papyri are characterized by an avoidance of aorist infinitives. This avoidance is quite surprising, particularly in light of the following observation: ‘The infinitive is, in fact, a pre-eminent element of the language of the papyri, and is abundantly used in every kind of document. In particular, official writings follow the Attic norm’ (Mandilaras 1973:329). The corpus does not seem to be under any significant influence of the Attic norm in terms of Dinf, however. DInf containing the aorist infinitive seem to be highly uncommon even in the official documents and even if potential alternative readings are taken into account. Although most potential instances of such clauses occur in official documents, they seem to be too rare to claim such influence. On the other hand, finite complement clauses found in the corpus do not reveal the same tendency towards avoiding the aorist, whereas the perfect is significantly less frequent in these clauses than the aorist. This distinction between infinitive and finite complement clauses indicates that in the (spoken) language of the period investigated a distinction occurred between the two types of complement clauses, which was uncharacteristic of Classical Greek.53 Other properties common to DInf in both New Testament Greek and the non-literary papyri include relatively high frequencies of stative present infinitives and perfect infinitives. Stative present infinitives prevail over non-stative ones in both New Testament Greek and the contemporary non-literary papyri, although the most common stative verb ‘be’ is more common in the former than in the latter. Perfect infinitives are more common than aorist infinitives in

53

In Classical Greek, the distinction between finite and non-finite complement clauses was pragmatic, according to Cristofaro (2008).

variation in expressing temporal and aspectual distinctions

49

both New Testament Greek and in the contemporary non-literary papyri (and even if the aforementioned editorial issues are taken into account). On the other hand, non-literary papyri also seem to contain higher frequencies of future infinitives used in DInf than the New Testament. Most of them are used in official documents and can refer to states or activities.

3.5

The Relation of the Construction to Other Developments in Hellenistic and Roman Greek

The aforementioned tendencies seem to be in a particularly complex relation to other developments in Hellenistic and Roman Greek. These developments primarily include the development of the Ancient Greek perfect and its functional merger with the aorist, as well as the emergence of periphrastic future forms and the retreat of the infinitive (as suggested below). As already mentioned, it has been argued (most explicitly in Thorley 1989) that DInf are used to express states in New Testament Greek. On the one hand, Thorley’s solution seems to be supported by the fact that the perfect infinitive tends to be associated in DInf with the present infinitive of stative verbs, whereas the aorist infinitive is avoided in the construction. Note that even the aorist of a stative verb usually has a non-stative, inchoative meaning: ‘The aorist aspect with states denotes most frequently the entrance of the subject into the condition denoted by the verb. Thus, it makes a shift in sense and in effect becomes a kind of active verb’ (Fanning 1990:137). This can account for the fact that aorist infinitives of both stative and non-stative verbs tend to be avoided in DInf. The tendency towards DInf associating with stative present infinitives is particularly strong in New Testament Greek, where the infinitive ‘be’ is by far the most frequent infinitive used in DInf. In addition, a great majority of DInf found in my corpus of non-literary papyri are stative as well, although the predominance of the infinitive ‘be’ is not as strong as in the New Testament. If Thorley’s solution is accepted, the tendency of Hellenistic and Roman Greek DInf could have parallels in English AcI clauses as well as in some other languages.54 In Greek, however, this explanation faces the problem of the functional merger between the aorist and the perfect. If, by the early centuries ce, the

54

English infinitive clauses of this type can contain the verb ‘be’ or ‘have’ or (rarely) some third stative verb (see Stockwell et al. 1973:569–571; for other languages: Pollock 1980:410; Skytte 1983:292 ff.; Grevisse 1993:1278).

50

kavčič

perfect had developed its aoristic use or was undergoing this process, the construction investigated could hardly be accounted for in terms of stativity. The issue of the merger between the aorist and the perfect has been discussed rather extensively in recent years. Although some scholars (mostly McKay and Porter) place this development in late antiquity (the 5th c. ce), others (e.g., Wackernagel, Chantraine and Horrocks)55 find at least sporadic occurrences of the phenomenon considerably earlier. Horrocks (loc. cit.) cites cases of the aoristic use of the perfect dating to early 2nd c. ce, and instances of morphological confusion between the two tenses also occur in the same period. Nevertheless, Blass et al. (2001:281) cite only one certain case of the aoristic use of the perfect in the New Testament. On the one hand, the construction investigated in this paper is very unlike passages in which the perfect associates in the same sentence with the aorist and that are sometimes cited in support of the functional merger between the two tenses. An example is Revelation 5.7: καὶ ἦλθεν καὶ εἴληφεν ἐκ τῆς δεξιᾶς τοῦ καθημένου ἐπὶ τοῦ θρόνου ‘he came and took [the scroll] from the right hand of the one who was seated on the throne’.56 In DInf, perfect infinitives avoid associating with aorist infinitives. The avoidance of aorist infinitives and the high frequency of perfect infinitives (both occur in New Testament Greek as well as in the contemporary non-literary papyri) could be used against the assumption that the aorist and the perfect had merged (completely) by the 1st or early 2nd c. ce. If this was the case, it seems that either avoidance of both aorist and perfect infinitives or more frequent use of aorist infinitives should be expected in DInf.57 55 56 57

See Chantraine (1927:441–445); Jannaris (1969:439–440, 485); Mandilaras (1973:225); McKay (1980); Porter (1989:273); and Horrocks (2010:176–178); see also 3.2.4 above. See Moulton (1911:226); Chantraine (1927:241–245); Jannaris (1969:439–450); and Blass et al. (2001:281). The paper also drew attention to a few DInf containing an aorist infinitive or a non-stative present infinitive. At least some of them were passive, examples including: [εὐχρ]ηστηθῆναι (P.Dura 18), ἀνηλω(θῆναι) (P.Oxy 985), and γεωργεῖσθαι (P.Tebt 302). The list can be expanded with some of the New Testament DInf containing an aorist infinitive or a nonstative present infinitive (σωθῆναι in Acts 15.11, τηρεῖσθαι in Acts 25.4, and λέγεσθαι in Hebrews 11.24). The tendency towards the passive infinitives of non-stative verbs associating with the active infinitives of stative verbs could be accounted for in terms of the observation that the passive encodes a situation ‘where the referent of the passive voice subject has come to be in some relatively stable end state’ (Gries-Stefanowitsch 2004:110). Thus the aforementioned passive infinitives could be an indication that DInf tended to express states rather than activities in the period investigated. Of course, the question of the semantics of the perfect remains unanswered.

variation in expressing temporal and aspectual distinctions

51

On the other hand, an explanation in terms of stativity is a less plausible one if it is assumed that the aorist and the perfect had (at least to a certain extent) functionally merged by the 1st c. ce. The high frequencies of DInf containing perfect infinitives are not accounted for in this case. As argued by Mandilaras (see 3.2.3, Table 3.5), the frequencies concerning the aorist and the perfect in Hellenistic and Roman Greek non-literary papyri in general can be used in support of the merger between the aorist and the perfect. In addition, it could be argued that some of the data concerning frequencies of the aorist and the perfect infinitives in DInf are not incongruous with this view. Thus the predominance of perfect infinitives over aorist infinitives in DInf seems to be less evident when only private letters are taken into account (because private letters contain a relatively low number of relevant passages), and when editorial and orthographic issues are considered (see above, 3.2.4, 3.2.6, and 3.3, Table 3.9). The corpus investigated in this paper seems to suggest another explanation for the phenomena. A difference between the New Testament and the contemporary non-literary papyri concerns the use of future infinitives in DInf. New Testament Greek yields only one clear instance of a DInf containing a future infinitive (ἔσεσθαι), but the contemporary non-literary papyri reveal a more common use of both stative and non-stative future infinitives. This could indicate that, in the period investigated, present, future, and perfect infinitives were used in DInf in the function of expressing temporal relations. As a consequence of the emergence of the fully transitive resultative perfect58 and, later, of its functional merger with the aorist, the system of expressing temporal relations in DInf was restructured. In the period investigated, present infinitives are thus simultaneous with, and future infinitives posterior to, the main verb, whereas perfect infinitives are anterior to the main verb as a consequence of having developed the new function.59 In terms of the corpus investigated, this solution also seems to be supported by a somewhat less convincing predominance of the infinitive ‘be’ in DInf (than what seems to be the case in the New Testament) and by a relatively frequent use of future infinitives, in addition to the aforementioned data concerning the frequency of the aorist and the perfect in the non-literary papyri in general and in private letters. The following observation seems to corroborate this as well: many of the perfect infinitives occurring in DInf (δεδανεικέναι ‘have borrowed’, δεδωκέναι ‘have

58 59

As is well known, the Ancient Greek perfect was originally stative and predominantly passive/intransitive (e.g., Chantraine 1927). In Classical Greek, anteriority was conveyed by the aorist infinitive.

52

kavčič

given’, πεπρακέναι ‘have done’, etc.) can be interpreted as resultative; that is, as referring to a completed action in the past (with potential significance for the present). The use of future infinitives in DInf perhaps concerns another development that took place in Hellenistic and Roman Greek, which is the emergence of periphrastic future forms. According to Markopoulos (2009), several periphrastic future forms occur in this period; namely, μέλλω + infinitive and ἔχω + infinitive in addition to θέλω + infinitive.60 However, the Ancient Greek future infinitive was disappearing in this period without being replaced by the future periphrasis μέλλω + infinitive, which was the most commonly acceptable future periphrasis of the period (Markopoulos 2009:49, 85). Thus, on the one hand, there seems to have been no significant tendency, from the historical perspective, towards maintaining the system of expressing temporal relations in these clauses, as was suggested above, which is why this interpretation seems less plausible. On the other hand, the aforementioned absence of periphrastic future infinitives can also be related to the retreat of the infinitive in general.

3.6

Conclusion

In Greek texts dating back to early centuries ce that are believed to provide an insight into the developments of the spoken language of Hellenistic and Roman Greek, it seems that DInf tended to avoid aorist infinitives, instead mostly associating with stative present infinitives, as well as with perfect infinitives. This phenomenon is not limited to the New Testament; it also occurs in the corpus of non-literary papyri that were examined in this paper. Because at least some of these tendencies were uncharacteristic of Classical Greek, it seems that between Classical Greek and the period investigated, DInf underwent significant changes in the expression of aspectual and/or temporal distinctions.

60

It is beyond the scope of this paper to examine the use of potential future periphrastics in the corpus investigated. In any case, the non-literary papyri investigated in this paper do not contain phrases of the type μέλλειν + infinitive in DInf. In the New Testament, on the other hand, DInf contain two instances of the phrase μέλλω + infinitive: ὁ μὲν οὖν Φῆστος ἀπεκρίθη τηρεῖσθαι τὸν Παῦλον εἰς Καισάρειαν, ἑαυτὸν δὲ μέλλειν ἐν τάχει ἐκπορεύεσθαι ‘Then Festus replied that Paul was being kept in Caesarea, and he himself intended to go there shortly’ (Acts 25.4) and ἐσήμανεν διὰ τοῦ πνεύματος λιμὸν μεγάλον μέλλειν ἔσεσθαι ἐφ ὅλην τὴν οἰκουμένην (Acts 11.28) ‘[he] predicted by the spirit that a severe famine was about to come over the whole inhabited world’.

variation in expressing temporal and aspectual distinctions

53

The phenomena observed in the paper also raise questions concerning their relation to other processes that took place in Hellenistic and Roman Greek, primarily to the development of the Ancient Greek perfect and its relation to the aorist.

The Corpus Investigated Contains: bgu 214, 248, 451, 530, 531, 584, 594, 595, 665, 830, 843, 846, 923, 1078, 1095, 1097, 1563, 2246, 2493; cpr 1.103, 7.52, 15.2 (= sb 5275), 15.8; P.Alex.Giss 37; P.Amh 68, 130; P.Ashm 24; P.Athen 29, 62; P.Bad 35; P.Berl.Moeller 4, 7, 9; P.Brem 6; P.Col 215; P.Congr xv.15; P.Corn 49; P.Dura 18, 19; P.Fay 109, 111, 123; P.Flor 61 (= Chr.Mitt 80); P.Fouad 21; P.Fuad.Univ 31; P.Giss 20, 21, 77, 78; P.Hamb 4, 29, 30; P.Haun 28; P.Iand 104; P.Koeln 56; P.Oslo 50; P.Oxy 37, 44, 73, 286, 295, 298, 300, 325 (= 10240), 745, 985, 1027, 1155, 1480, 1672, 2190, 2339, 2353, 2725, 2756, 2757, 3057, 3070, 3264, 3274, 3357, 3503; P.Lips 105 (= Chr.Wilck 237); P.Lond 897, 1168; P.Mert 12; P.Mich 121 (= sb 6705), 194, 203 (= sb 7356), 236, 263, 344, 635 (= 10779), 656; P.Mich.Mchl 25 (= sb 11127); P.Michael 15; P.Princ 188; P.Ryl 2, pg. 381, 173a; P.Sarap 90 (= P.Bad 90); psi 317, 690, 777, 967, 1109; P.Stras 274, 334; P.Tebt 286, 302, 409; P.Turner 18; P.Vind.Bosw 1; P.Vind.Tand 10; P.Yale 77, sb 3924, 5230, 5233, 5236, 5761, 7258, 7268, 7461, 7463, 7600, 7737, 8070 (= P.Fouad 14 = P.Bacch 24), 8947, 9120, 9271, 10967, 11017, 11585, 11625, 11644, 12172, 13222, 13303, 13590; and StudPal 4, pg. 58– 78, 22.33.

References Blass, F., A. Debrunner, & F. Rehkopf. 200118. Grammatik des neutestamentlichen Griechisch. Göttingen. Boak, A.E.R. (ed.). 1933. Papyri from Tebtunis. i. Ann Arbor. Boswinkel, E. (ed.). 1941. Einige wiener papyri. Leiden. Burton, E. de Witt. 1898. Syntax of moods and tenses in New Testament Greek. Edinburgh. Chantraine, P. 1927. Histoire du parfait grec. Paris. Cristofaro, S. 2008. ‘A constructionist approach to complementation: evidence from Ancient Greek’. Linguistics 46, 571–606. Dolger, F. 1953. ‘Φαγεῖν καὶ πιεῖν’. Ἐπετηρὶς Ἑταιρείας Βυζαντινῶν Σπουδῶν 23, 57–64. Evans, T.V. & D.D. Obbink (eds.). 2010. The language of the papyri. Oxford. Evans, T.V. 2010. ‘Identifying the language of the individual in the Zenon archive’. In: Evans & Obbink 2010:51–70. Fanning, B.M. 1990. Verbal aspect in New Testament Greek. Oxford.

54

kavčič

Gignac, F.T. 1981. A grammar of the Greek papyri of the Roman and Byzantine periods. ii: Morphology. Milan. Grenfell, B.P. & A.S. Hunt (eds.). 1899. The Oxyrhynchus papyri. ii. London. Grenfell, B.P., A.S. Hunt, & D.G. Hogarth (eds.). 1900. Fayûm towns and their papyri. London. Grenfell, B.P., A.S. Hunt, & E.J. Goodspeed. 1907. The Tebtunis papyri. London. Grenfell, B.P. & A.S. Hunt (eds.). 1916. The Oxyrhynchus papyri. 12. London. Grevisse, M. 1993. Le bon usage. Louvain-la-Neuve. Gries, S.T., & A. Stefanowitsch. 2004. ‘Extending collostructional analysis. A corpusbased perspective on “alternations”’. ijcl 9, 97–129. Horrocks, G.C. 20102. Greek: a history of the language and its speakers. Chichester. Jannaris, A.N. 1968. An historical Greek grammar chiefly of the Attic dialect: as written and spoken from classical antiquity down to the present time. Hildesheim. Joseph, B.D. 1983. The synchrony and diachrony of the Balkan infinitive. Cambridge. Kavčič, J. 2009. ‘A feature of Greek infinitive clauses dependent on verbs of saying and thinking’. Early European languages in the eyes of modern linguistics, ed. by K. Loudová & M. Žáková, 151–168. Brno. Kurzová, H. 1968. Zur Syntaktischen Struktur des Griechischen Infinitiv und Nebensatz. Amsterdam. Leiwo, M. 2010. ‘Imperatives and other directives in the Greek letters from Mons Claudianus’. In: Evans & Obbink 2010:97–119. Lobel, E., C.H. Roberts, & E.P. Wegener (eds.). 1941. The Oxyrhynchus papyri. vxii. London. lsj = Liddell, H.G., R. Scott, & H.S. Jones. 1996. A Greek-English lexicon. Oxford. Luiselli, R. 2010. ‘Authorial revision of linguistic style in Greek papyrus letters and petitions (1st–4th c. ce)’. In: Evans & Obbink 2010:71–96. Mandilaras, B.G. 1973. The verb in the Greek non-literary papyri. Athens. Markopoulos, T. 2009. The future in Greek: from Ancient to Medieval. Oxford. Mayser, E. 1926. Grammatik der griechischen Papyri aus der Ptolemäerzeit: mit Einschluss der gleichzeitigen Ostraka und der in Agypten verfassten Inschriften. ii.1: Satzlehre. Berlin. McKay, K.L. 1980. ‘On the perfect and other aspects in the Greek non-literary papyri’. bics 27:23–50. Messeri Savorelli, G. (ed.). 1990. Corpus Papyrorum Raineri. 15. Vienna. Mihevc, E. 1959. La disparition du parfait dans le grec de la basse époque. Ljubljana. Mitteis, L. (ed.). 1906. Griechische Urkunden der Papyrussammlung zu Leipzig. i. Leipzig. Moulton, J.H. 1911. Einleitung in die Sprache des Neuen Testaments. Heidelberg. Pearl., O.M. 1978. ‘Document on the Episkepsis at Karanis’. Actes du xve Congrès International de Papyrologie (Seconde partie), ed. by J. Bingen & G. Nachtergael, 75–85. Brussels.

variation in expressing temporal and aspectual distinctions

55

Perkins, A. (ed.). 1959. The excavations at Dura-Europos. Final report 5, part 1. New Haven. Pollock, J.-Y. 1980. ‘Verb movement, universal grammar, and the structure of ip’. li 20:367–424. Porter, S.E. 1989. Verbal aspect in the Greek of the New Testament. New York. Porter, S.E., & M.B. O’Donnell. 2010. ‘Building and examining linguistic phenomena in a corpus of representative papyri’. In: Evans & Obbink 2010:287–311. Rea, J.R. (ed.). 1978. The Oxyrhynchus papyri. xlvi. London. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Chicago. Schwyzer, E. & A. Debrunner. 1950. Griechische Grammatik. ii: Syntax und syntaktische Stilistik. München. Skytte, G. 1983. La sintassi dell’ infinito in italiano modern. Copenhagen. Smith, C. 19972. The parameter of aspect. Dordrecht. Smyth, H.W. 1963. Greek grammar. Cambridge. Stockwell, R.P., Schachter, P. & B.H. Partee. 1973. The major syntactic structures of English. New York. Thorley, J. 1989. ‘Aktionsart in New Testament Greek: infinitive and imperative’. nt 31:290–315. Wackernagel, J. 1904. Studien zum griechischen Perfectum. Göttingen. Winter, J.G. et al. 1936. Papyri in the University of Michigan Collection. Ann Arbor.

chapter 4

Syntactic Variation with Verbs of Perception and the ‘Oblique Imperfect’: Once Again on Aspect, Relative Time Reference and Purported Tense-Backshifting in Ancient Greek* Julián Méndez Dosuna

4.1

Introduction

According to a view that seems to be increasing in popularity, relative tense was an operative category in the grammar of Ancient Greek. In this paper I join the opponents of this idea by focusing on the evidence provided by predicates of immediate perception. In the second part of the paper I concentrate on the so-called ‘oblique imperfect’, which allegedly serves the function of tense backshifting (consecutio temporum).

4.2

Aspect in Ancient Greek (Plus a Clarification on Terminology)

It is generally acknowledged that aspect was a dominant category in the Ancient Greek verb.1 It was relevant to all moods, to the infinitives and to the participles. The present stem denoted a durative (aka imperfective) aspect; the aorist stem denoted a punctual (aka perfective) aspect; the perfect stem denoted a resultative aspect. The future stem was the odd one out in the system so to speak, inasmuch as, disregarding a few desultory attempts at cre-

* This paper was part of the project ffi2009-13402-c04-04 funded by the Spanish Ministry of Science and Innovation (micinn). I am grateful to Alcorac Alonso Déniz, Klaas Bentein, Stephen Colvin, Emilio Crespo and Alberto Pardal Padín for their helpful comments on earlier versions of the paper. As regards responsibility for the views expressed here and for any errors or shortcomings, the usual disclaimers apply. 1 For a general framework, I refer the reader to Ruipérez (1954), a landmark in the study of time and aspect in Greek. Except for some matters of terminology and minor details, Ruipérez’s views remain largely valid; see also Crespo et al. (2003:255–281) and, in a different theoretical framework, Amalia Moser’s contribution to this volume (chapter 7).

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_005

syntactic variation with verbs of perception table 4.1

57

Aspect in the Ancient Greek verb

Stem

Aspect

pres aor fut pf

+durative –durative ±durative

–resultative –resultative –resultative +resultative

ating aspect contrasts which ultimately proved fruitless, it remained aspectneutral.2 Before proceeding further, a brief clarification concerning terminology is in order. Some authors have objected to the label punctual (alternatively [– durative]) on the grounds that the aorist may be combined with adverbs and prepositional phrases that denote duration in well-known examples such as the following: (1) a. Ψαμμήτιχος δὲ ἐβασίλευσε [aor.ind] Αἰγύπτου τέσσερα καὶ πεντήκοντα ἔτεα ‘Psammetichus ruled Egypt for fifty-four years’. hdt. 2.157.1

b. χρόνον μὲν οὖν τινα ὀλίγον οὕτω πρὸς ἀλλήλους ἠκροβολίσαντο [aor.ind]· ‘Thus they skirmished for a certain time with each other in that way’. thuc. 4.34.1

In order to overcome the apparent contradiction, they assume that the aorist is [+perfective], sc. denotes an event temporally bounded and complete, as against the imperfect, which is [–perfective], sc. denotes an event temporally unbounded and incomplete. 2 See e.g., διδώσομεν ‘we will give [repeatedly?]’ (Hom. Od. 13.358) built on the present stem (contrasting with aoristic δώσομεν?) and σχήσουσιν ‘they will hold [once?]’ (Hom. Od. 11.70) built on the aorist stem (contrasting with aoristic ἕξουσιν?). Eventually, a systematic aspect contrast was established with the emergence of the periphrastic θα-future in the Medieval period: see Modern Greek θα δίνω ‘I’ll be giving’ vs. θα δώσω ‘I’ll give’.

58

méndez dosuna

However, the feature [±perfective] is not without its own shortcomings. The signatures of Greek artists are a case in point. When signing their works, potters, vase painters or sculptors used indifferently the aorist (2a) or, less often, the imperfect (2b): (2) a. Πῦθις ἐποίε̄σεν: ‘Pythis sculpted (this statue)’ (ig i3 680.I1, Athens, ca. 500bce?) b. Πῦθις τόδε ἐποίε̄: ‘Pythis sculpted this’ (ig i3 1118(3), Prasiai, ca. 500bce?) Of course, one should not infer from the imperfect in (2b) that Pythis had left his work unfinished.3 Priding oneself on one’s own incompetence or negligence is certainly not the most effective way to build up a reputation. Thus, in this example the feature [± perfective] fails to capture the essence of the Ancient Greek aorist and imperfect (or, for that matter, of its Modern Greek descendants).4 Tellingly, in the absence of an adequate context, the imperfect of (2b) does not easily translate into the imperfect of the modern Romance languages: It. Putis scolpiva (!?) questo, Fr. Pythis sculptait (!?) ceci, Sp. Pitis esculpía (!?) esto. With regard to the untranslatability of (2b), the important thing to bear in mind is that the imperfect of Greek and the imperfect of Romance languages are equipollent categories inasmuch as they partake of the general idea of duration. Still, when it comes to more specific values, they are not always interchangeable. As in (2b), the imperfect of Ancient Greek admits of a progressive reading, which is incompatible with the imperfect of the Romance languages.5 These have developed special periphrases for the progressive aspect. The nuance conveyed by the Greek imperfect in (2b) can be reflected in Spanish by combining the perfectiveness of the pretérito indefinido with the progressiveness of the periphrasis estar ‘to be [temporarily]’ + gerund: Pitis estuvo esculpiendo esto ‘Pythis spent some time sculpting this’. In cases like this, the

3 Stephen Colvin called my attention to a passage of Pliny the Elder’s Natural History (praef. 26), in which the Greek imperfect is misinterpreted through the distorting lens of Latin. Pliny believed that a dedication like Apelles faciebat (Gr. Ἀπελλῆς ἐποίει) implied that the work was still incomplete, as against ille fecit (Gr. ὁ δεῖνα ἐποίησε), which indicated that the work was finished and its author was proud of it. 4 A similar point is made by Crespo (2014) on the basis of cases like e.g. Δευκαλίων δ’ ἐμὲ τίκτε πολέσσ’ ἄνδρεσσιν ἄνακτα ‘Deucalion sired me, a lord over many men’ (Hom. Il. 13.452). 5 For similar facts in Modern Greek, see Moser (this volume, chapter 7).

syntactic variation with verbs of perception

59

Greek speaker was confronted with the dilemma of whether to highlight progressiveness with the imperfect (ἐποίει) at the expense of perfectiveness or accord precedence to perfectiveness with the aorist (ἐποίησεν) at the expense of duration. The issue is more terminological than conceptual. Duration must be conceived of, in the spirit of the grams of Bybee et al. (1994), as an abstract coverterm capable of encompassing a wide and complex variety of more concrete interrelated readings: imperfective, continuous, habitual, iterative, distributional, ingressive, progressive, continuative, etc. (these subcategories vary from language to language). The context will usually clarify which of these shades of meaning is (ir)relevant. Punctuality must be understood as lack of duration, real or conceived. In (1a–b) the aorist telescopes the events into a compact whole intentionally disregarding its internal time structure. Conversely, the imperfect of (2b) stresses the amount of time required by the task of sculpting the statue.6 For convenience, throughout this paper I will stick to the traditional labels with a certain preference for durative / punctual over imperfective / perfective, even though I admit that something like diffuse vs. compact, a contrast that can boast a long tradition in studies of phonology, would probably best describe the facts.

4.3

Absolute and Relative Time Reference in Ancient Greek

Absolute time reference in the tenses of the indicative is fairly straightforward and needs no extensive comment.7 An overview is given in Table 4.2. Conversely, in the moods other than the indicative, the contrast between present and past does not hold anymore, only aspect being relevant (see Kühner-Gerth 1904:§389). This amounts to saying that the imperative, the subjunctive, the optative, the infinitive and the participle were time-neutral or, put another way, that they failed to express time reference. The future stem is again deviant since the future optative, the future infinitive and the future participle either encode future time reference or convey a volitive meaning. Two examples may help to clarify this point. An imperative can hardly refer to the past or to the present. By definition, orders are future-oriented. Present 6 It is no coincidence that dedicators invariably use formulas in the aorist (ὁ δεῖνα τόδε ἀνέθηκεν ‘so-and-so dedicated this’), since a dedication is an intrinsically brief event. 7 Needless to say, this is a simplification of a more complex issue that extends well beyond the scope of this paper.

60

méndez dosuna

table 4.2

Absolute time reference in the indicative

Tenses of the indicative

Absolute time reference

ipf/aor/plpf pres/pf fut/fut.pf

past present future

t−1 t0 t+1

imperatives envisage actions conceived of as extended in time while aorist imperatives envisage actions conceived of as compact events: see e.g., λέγε ‘tell me the whole matter, speak at leisure’ vs. εἰπέ ‘tell me one thing, go to the point’. By the same token, the subjunctive of final clauses refers to an event that by definition is contingent on the action of the main verb, and, accordingly, cannot precede it. In (3a) the present subjunctive denotes a mental process involving a certain amount of duration while the aorist subjunctive (3b) denotes a sudden intuition. Neither simultaneity nor precedence is here at stake: (3) a. ἵνα γιγνώσκῃς [pres.subj] τὸν τιθασευτήν ‘So that you may come to know [gradually] your tamer’. aristoph. Vesp. 704

b. τῶνδε τοίνυν τῶν θεωμένων σκόπει τὰ πρόσωφ’ ἵνα γνῷς [aor. subj] τὰς τέχνας. ‘So look at the faces of these spectators so that you may identify [instantly] their occupations’. aristoph. Pax 543

The third tenet of the traditional doctrine holds that relative time reference (hereafter, rtr) was a notion entirely alien to the grammar of Ancient Greek.8 Demonstrably, the Greek ‘perfect’ tenses were resultatives denoting a state persisting into the past, the present or the future as a consequence of a previous

8 Most concisely phrased by Crespo et al. (2003:255): ‘En griego clásico no hay expresión de tiempo relativo’. See also Humbert (1960:169–170).

syntactic variation with verbs of perception

61

action: ἐτεθνήκει ‘he was dead’, τέθνηκε ‘he is dead’, τεθνήξει ‘he will be dead’.9 Neither the pluperfect nor the future perfect were anteriors denoting an action completed prior to a primary point of time made explicit or remaining implicit in the context: English he had died (before a given event), he will have died (before a given event). For this reason, the Greek ‘perfect’ tenses could occur in abstraction from any time context, as the following examples demonstrate. Unlike its homonyms in English, Latin or the Romance languages, which are past anteriors, the Greek pluperfect in (4a–b) is a past resultative: (4) a. εἰ δέ τις καὶ ἀντείποι, εὐθὺς ἐκ τρόπου τινὸς ἐπιτηδείου ἐτεθνήκει [plpf.ind] ‘Whenever someone disapproved, he was immediately dead in some convenient way’. thuc. 8.66.1

b. οἱ δ’ ἐγκατακείμενοι παρ’ αὐτῷ […] τὸν Πλοῦτον ἠσπάζοντο καὶ τὴν νύχθ’ ὅλην ἐγρηγόρεσαν [plpf.ind], ἕως διέλαμψεν ἡμέρα. ‘Those who were lying near him kept on congratulating Wealth and stayed awake all night long till day broke’. aristoph. Plut. 742–744

In (4a) the death of the opponent ensues from his disapproval. In (4b) the accusative of duration of time τὴν νύχθ’ ὅλην indicates that ἐγρηγόρεσαν expresses a state. In either case, a past anterior (‘He had died’, ‘They had awoken’) would make no sense.

9 For a good overview of the issue see Bentein (2014). Bentein sympathizes with the view defended by e.g., Haspelmath (1992) that by 450 bce the old stative/resultative perfect (e.g., τέθνηκε ‘(he) is dead’, τόδε γέγραφε ‘(he) has this written’) often allowed for anterior readings (‘(he) has died’, ‘he has written this’). Admittedly, the grammaticalization path from resultative to anterior is very common in the languages of the world, but to my mind, the Greek evidence produced by the adherents of this view is anything but compelling. In any case, this is not the place to address a contentious issue that remains largely marginal to my argument. An anterior (continuing) perfect (≈ immediate past) is anterior in that it expresses a past action which is relevant or continues into the present, but crucially, unlike a past or future anterior, it does not presuppose a secondary reference time different from the moment of speech.

62

méndez dosuna

Similarly, the future perfect expressed a future resultative (‘I’ll remain in a given state’), not a future anterior (‘I’ll have done something’): (5) Ἑλένη τὸ λοιπὸν ἐν βροτοῖς κεκλήσεται [fut.pf.ind] ‘[The island] shall (always) be called Helen among men from now on’. eur. Hel. 1674

The future perfect does not simply mean that the island will be given the name Helen, but makes clear that it will keep that name forever. A future anterior interpretation (‘It will have been called’) is completely impossible. As happens in numerous languages, e.g., Russian, Ancient Greek simply lacked a grammatical category serving the specific function of expressing temporal depth (usually precedence) in a sequence of events. If ambiguity had to be dispelled, Greeks had recourse to temporal adverbs like πρίν ‘before’, ἔπειτα ‘after’, πρῶτον ‘first’, etc. Otherwise, the ordering of events was a matter of discourse pragmatics depending on the context and commonsensical implicatures, as in the following example: (6) οἱ δὲ […] ὀλίγον χρόνον ὑπέμειναν [aor.ind] [t-2], ἔπειτα δὲ ἐτράποντο [aor.ind] [t-1] ἐς τὸν Πάνορμον, ὅθενπερ ἀνήγαγον [aor.ind] [t-3]. ‘The others resisted a little time and then headed for Panormus, from where they had set out’. thuc. 2.92.1

In the absence of the relevant information provided by Thucydides in 2.86 that the Peloponnesians had coasted along to Panormus in Achaea, it would be almost impossible to decide whether the aorist ἀνήγαγον refers to an event prior (‘from where they had set out before’) or subsequent (‘from where they set out afterwards’) to ὑπέμειναν and ἐτράποντο. To sum up, according to the traditional view, aspect was an all-pervasive category in the Ancient Greek verb. Absolute time was secondary to aspect, the contrast between past and present being neutralized outside the indicative. No tense was capable of expressing rtr.

syntactic variation with verbs of perception

4.4

63

rtr and Participles: Anything Goes

In spite of overwhelming evidence in its support, the traditional doctrine has been repeatedly challenged by a number of scholars who contend that rtr can be denoted by some tenses, especially by participles. The Relative Tense Theory (henceforth, RTTh) is clearly stated by Rijksbaron (2006:117, original italics):10 In general, the participles of the various tense stems express the following relationships with respect to the main verb. The present participle expresses that the state of affairs of the participle is simultaneous with that of the main verb, the aorist participle that it is anterior, the future participle that it is posterior with respect to the main verb. The perfect participle expresses that the state of affairs of the participle is simultaneous with that of the main state of affairs. The rationale behind this approach is roughly the following. The present stem expresses an imperfective event. An unbounded event with no defined endpoint naturally tends to be interpreted as simultaneous with the action of the main predicate. Conversely, the perfective aspect of the aorist participle, which denotes an action with a defined endpoint, is more likely to be interpreted as preceding the action of the main predicate. The perfect participle expresses the state resulting from a previous event and is consequently liable to be interpreted as expressing a state simultaneous with the main predicate. Being fully aware that the principle confronts a host of exceptions, Rijksbaron is cautious enough to introduce the paragraph with the phrase ‘in general’ as a sort of disclaimer. I belong with those who do not believe the participles are capable of expressing rtr, or that their aspect is instrumental in the temporal ordering of the events.11 Consider the following examples with different participles of the verb ἀποθνῄσκω:

10

11

Rijksbaron adopts the views of Ruijgh (1991:199–200). See also Hettrich (1976). The idea goes as far back as Stahl (1907). The most explicit and radical attempts to vindicate rtr in Greek are probably those put forward by Martínez Vázquez (1995) and de la Villa (2010; 2014), but in my opinion, their claims remain unsubstantiated. See e.g., Kühner-Gerth (1904:§389.3, 590c, original emphasis): ‘Der Infinitiv und das Partizip […] bezeichnen ebenfalls nicht die Zeitstufe, sondern nur die Beschaffenheit des durch sie ausgedrückten Verbalbegriffes’. Crespo et al. (2003: 266): ‘Los participios, excepto el de futuro […], no expresan tiempo, ni absoluto ni relativo, y, por tanto, pueden referirse a situaciones pasadas, presentes, futuras o atemporales’.

64

méndez dosuna

(7) a. […] τὰ χρήματα ἃ Κίρων ὁ πάππος ἀποθνῄσκων [pres.part] κατέλιπεν ‘The fortune that our grandfather Ciron had left when he was dying’. is. 8.3

b. ἀποθανὼν [aor.part] δὲ καὶ ταριχευθεὶς ἐτάφη ‘When [Amasis] died and was embalmed, he was buried’. hdt. 3.16.2

c. εἰ […] Ἐπίλυκος […] τεθνεὼς [pf.part] πολλὰ κατέλιπε χρήματα ‘If Epylicus had left a large fortune when he was dead […]’. and. 1.119.3

To judge by appearances, the present and perfect participles in (7a) and (7c) denote an event simultaneous with the main verb κατέλιπεν while the aorist participle in (7b) denotes an event anterior to ἐτάφη. Actually, anteriority and simultaneity are not expressed by the different verb stems of the participles, but are mere inferences based on common sense. In (7a) the grandfather must have drawn up his will before his death since we know that the dead cannot speak or write. In (7b) we interpret that Amasis died before he was buried simply because people are rarely buried before or at the very moment of dying. It is also self-evident that the simultaneity of actions in (7a) differs from that in (7c).12 In (7a) the grandfather dictated his last will while he was dying (ἀποθνῄσκων). In (7c) the temporal relation of τεθνεώς to κατέλιπε is doublefaced: the will was executed after Epilykos’ death, but is simultaneous with the ensuing state of being dead.13 In short, in these examples, aspect is crucial while (relative) tense is epiphenomenal at best. It has been observed more than once that, if the RTTh works in a considerable number of cases, it is due to the fact that, as indicated above, durative and stative events are much more likely to coincide or overlap in time with an event other than punctual events. This notwithstanding, a legion of counterexamples demonstrate that the rule at issue is spurious: 12

13

The difference is clearly stated by Ruijgh (1991:200): ‘On voit que la valeur du thème de parfait est complexe: l’ état résultatif est simultané, mais l’action achevée est antérieure à l’ action du verbe principal’. For the interplay of the actionality and the aspect of the different stems of ἀποθνῄσκω, see Méndez Dosuna (2008).

syntactic variation with verbs of perception

65

(8) a. θάπτουσι δὲ τοὺς ἀποθνῄσκοντας [pres.part] οἱ νομάδες κατά περ οἱ Ἕλληνες ‘The nomads bury the dead much like the Greeks do’. hdt. 4.190.1

b. ἤδειν γὰρ ὅτι ἀποθνῄσκοντος [pres.part] Λέοντος τοῦ Σαλαμινίου […] οἱ ὁμοῖοι τούτῳ φοβήσοιντο ‘For I knew that, once Leon the Salaminian had been sentenced to death, those who were like him would be scared’. xen. Hell. 2.3.39

In (8a) ἀποθνῄσκοντας cannot imply that the Greeks or the nomads buried people while they were dying. The present stem is best interpreted as expressing an iterative (distributive) aspect: the nomads buried the dead as they died successively, one after the other. In (8b) ἀποθνῄσκοντος expresses the period during which Leon of Salamis was, so to speak, on death row, i.e., the period extending from the moment he was arrested and sentenced to death until he was finally executed. At least the beginning of this period must precede the event indicated by φοβήσοιντο. Likewise, it has long been noted that the present participle frequently cooccurs with adverbs meaning ‘before’, which, needless to say, are incompatible with the idea of simultaneity:14 (9) οἱ †Κύρειοι† πρόσθεν [+past] σὺν ἡμῖν ταττόμενοι [pres.part] νῦν ἀφεστήκασιν. ‘The [followers of Arieus] who were formerly formed in order of battle with us, are now deserters’. xen. An. 3.2.17

In his defence of rtr in Greek, Martínez Vázquez (1995:186) explains away ‘exceptions’ of this type as ‘special cases’ and endeavours to accommodate them in the Procrustean bed of the RTTh under various pretexts. He is bound to assume that the participles may or may not express relative time depending on

14

For other examples, see Crespo et al. (2003:266–267).

66

méndez dosuna

the context, a claim that at the very least is illogical since one and the same grammatical category cannot convey contradictory meanings. Grammatical meaning can be modulated by context, but cannot entirely depend on it. Either it is part of the grammar or it is not. The RTTh also comes across difficulties from the viewpoint of diachrony and typology. On the one hand, to judge from the evidence of other languages, participles are not good initiators of the grammaticalization of rtr.15 On the other hand, the hypothetical difference left no trace whatsoever in Medieval and Modern Greek. As is known, the rich array of participles of Classical Greek has undergone a severe reduction down the centuries. In the middle-passive voice only the perfect participle has survived. More to the point, in the active voice, the present and aorist participles gradually evolved into indeclinable gerunds in Medieval Greek. Contrary to what the RTTh would have us expect, the contrast of the present vs. the aorist gerund was not one of simultaneity vs. precedence, but, once more, one of durative vs. punctual aspect. As expected, their temporal construals depended on the semantics of the verb and on the context.16 (10) a. κ’ ὑπάγαινεν [ipf.ind] ἐβλέποντα17 [pres.gerund] τὰ κάστρη καὶ τὲς χῶρες ‘And he drove inspecting [one by one] the castles and the villages’. Chron. Mor. 329

b. ὁ πρίγκιπας γὰρ βλέποντας [pres.gerund] τὴν τόση ἀλαζονείαν ἀπὸ χολῆς του καὶ θυμοῦ ὤμοσε [aor.ind] εἰς τὸ σπαθί του ‘For, when the prince [gradually] perceived that much arrogance, he swore upon his sword out of wrath and irritation’. Chron. Mor. 240–241

15

16 17

Resultatives are the most common source for anteriors in the languages of Europe. In more exotic languages anteriors may arise from verbs meaning ‘finish’, ‘throw away’ or ‘pass by’ (Bybee et al. 1994:68–74). Mirambel (1961); Tsimpli (2000); Moser (2002; 2006); Manolessou (2005); and Horrocks (2010: 181–183, 298). The hypercorrect ἐ- results from the analogy of alternations in augmented forms like ἐβλέπαμεν ~ βλέπαμεν, where the loss of unstressed initial /e/ was regular.

syntactic variation with verbs of perception

67

c. ἰδόντας [aor.gerund] τοῦτο ὁ ἅγιος ἐκεῖνος ὁ ἐρημίτης, μεγάλως ἐβαρέθηκεν [aor.ind] ‘When that holy hermit [instantly] saw this, he felt very upset’. Chron. Mor. 18–19

In time the aorist gerund was displaced by the present gerund, which remains time-neutral in Modern Greek, its temporal relation to the main verb being inferred from the pragmatic context (11a–b). Ambiguity can be avoided by having recourse to the more recent and less frequent compound gerund (11c– d): (11) a. φεύγοντας [gerund] η δασκάλα, γέλασαν [aor.ind] τα παιδιά. Default construal (anteriority): ‘Once the teacher had left, the children laughed.’ Non-default construal (simultaneity): ‘As the teacher was leaving, the children laughed’. b. φεύγοντας [gerund] η δασκάλα, γελούσαν [ipf.ind] τα παιδιά. Default construal (simultaneity): ‘As the teacher was leaving, the children were laughing’. Non-default construal (anteriority): ‘When the teacher had left, the children continued laughing’. c. έχοντας φύγει [ant.gerund] η δασκάλα, γέλασαν [aor.ind] τα παιδιά. Unambiguous construal (anteriority): ‘Once the teacher had left, the children laughed’. d. έχοντας φύγει [ant.gerund] η δασκάλα, γελούσαν [ipf.ind] τα παιδιά. Unambiguous construal (anteriority): ‘When the teacher had left, the children continued laughing’.

4.5

Further Evidence against RTTh: Participles in Predicates of Immediate Perception

Complements to immediate perception predicates have Determined Time Reference with respect to the main clause since the two events are necessarily simultaneous, as illustrated by the following English examples:

68

méndez dosuna

(12) a. John saw [t−1] with his own eyes Mary eat / eating [t−1] an egg.18 (≈ John saw (t−1) with his own eyes that Mary was eating (t−1) an egg) b. † John saw [t−1] with his own eyes Mary have eaten / having eaten [t−2] an egg.19 (≠ John saw [t−1] with his own eyes that Mary had eaten / had been eating [t−2] an egg) In Ancient Greek immediate perception verbs can take a participle. RTTh predicts that the present and perfect participles should be used to the exclusion of the aorist participle, which allegedly expresses an event prior to the event of the matrix (see 4.4). This prediction, however, is not borne out by the data, since verbs of perception combine indifferently with present and aorist participles, as illustrated by the following sentences with ἀκούω and ὁράω: (13) a. σοῦ δ’ ἐγὼ λόγους λέγοντος [pres.part] οὐκ ἀκούσομαι μακρούς ‘I won’t listen to you delivering long speeches’. aristoph. Ach. 303

b. ἀκούσας οὖν μου ὁ Κριτίας ταῦτ’ εἰπόντος [aor.part] ‘When Critias heard me say that’. pl. Chrm. 157c

c. ὅστις ἀγχόθεν χιόνα ἁδρὴν πίπτουσαν [pres.part] εἶδε, οἶδε τὸ λέγω· ‘Whoever has seen snow falling thick at a short distance, knows what I say’. hdt. 4.31.2

d. οὔτε γὰρ πεσόντα [aor.part] μιν εἶδον ἀπὸ τοῦ ἵππου οὔτε ἀποθνῃσκόντα [pres.part] ‘For they neither saw him falling from his horse nor dying’. hdt. 9.22.3

18 19

For the different shade of meaning of the two constructions, see Quirk et al. (1985:204). These sentences are grammatical if saw is taken to mean ‘witnessed’: e.g. John saw the remains of the egg and inferred that Mary had eaten it.

syntactic variation with verbs of perception

69

e. ὡς δὲ εἴδομεν πίνοντά [pres.part] τε καὶ πεποκῶτα [pf.part] ‘When we saw [him] sipping [the hemlock] and having drunk it’. pl. Phdr. 117c

In (13a) the present participle is congruous with the duration of the long speeches delivered. Conversely, in (13b) duration is backgrounded as irrelevant information. A snowfall (13c) and the agony of a wounded warrior (13d) typically last for some time, while a fall from a horse (13d) is an event of very short duration. In (13e) the act of drinking is successively conceptualized as a durative action (‘to sip’) and as a resultative state.20 The natural conclusion of all this is that the participles in (13a–e) do not encode rtr, but merely convey aspect.21

4.6

The Oblique Optative and Purported Tense Backshifting

We are moving now to a different subject matter, which also bears on the verbs of immediate perception. Sequence of tenses or consecutio temporum as attested in Latin or, currently, in English and in most Western European languages, involves a shift of temporal deixis in reported speech. With a verb in the past in the matrix clause, a present tense in direct speech is changed to a past in reported speech (14a) and a past tense becomes a past anterior (14b):22 (14) a. Direct: John said [t−1]: ‘My father is [t0]: a doctor’. reported: John said [t−1] that his father was [t−1] a doctor.

20

21

22

Note that the case of εἴδομεν … πεποκῶτα is not comparable to the acceptable reading of (12b) (see previous footnote). Socrates’ disciples did not infer that he had drunk the cup of hemlock (cognition). They actually saw Socrates with the hemlock inside him so to speak (immediate perception). The same rationale explains why the verbs of phase (‘begin’, ‘end’, ‘continue’), which presuppose duration, can only be constructed with the present participle (or infinitive): e.g., ἐκλείπειν μοι φαίνεται ἡ ψυχὴ ὅθενπερ […] πᾶσιν ἄρχεται ἀπολείπουσα [pres.part] ‘My soul seems to me to be leaving from those parts of the body, from which it begins to leave everyone’ (Xen. Cyr. 8.7.26), οὔποτε ἐπαυόμην ἡμᾶς μὲν οἰκτίρων [pres.part], βασιλέα δὲ καὶ τοὺς σὺν αὐτῷ μακαρίζων [pres.part] ‘I never ceased commiserating ourselves and congratulating the King and his followers’ (Xen. An. 3.1.19). Once again, for the sake of brevity, the facts concerning sequence of tenses have been simplified.

70

méndez dosuna

b. Direct: John said [t−1]: ‘My father was [t−1] a doctor’. reported: John said [t−1] that his father had been [t−2] a doctor. Demonstrably, Ancient Greek lacked such a rule of sequence of tenses so that the tenses of direct speech underwent no changes in reported speech (15a–b). Optionally they could be replaced by the oblique optative (15c–e). (15) a. Ἐργοκλῆς ἔλεγεν ὡς ἤδη συκοφαντεῖτε [pres.ind] ‘Ergocles said that there you were at your slander-mongering’. lys. 28.5

direct speech: ἤδη συκοφαντεῖτε [pres.ind] b. περὶ Ἀγοράτου τουτουὶ ἔλεγεν ὅτι ⟨οἱ⟩ αἴτιος ἦν [ipf.ind] τοῦ θανάτου ‘As regards Agoratus, that man there, [Dionysodorus] said repeatedly that he had been responsible for his death’. lys. 13.41

direct speech: ἐμοὶ αἴτιος ἦν [ipf.ind] τοῦ θανάτου c. ἄλλα τε πολλὰ εἶπε καὶ ὅτι παρασπόνδους ὑμᾶς ἔχοι [pres.opt] ‘He said many things and specifically that he held you to be trucebreakers’. lys. 12.74

default construal: παρασπόνδους αὐτοὺς ἔχω [pres.ind] non-default construal: παρασπόνδους αὐτοὺς εἶχον [ipf.ind] d. ἐρομένου δέ μου τί αἱ θύραι νύκτωρ ψοφοῖεν ‘When I asked why the doors had been banging in the night’. lys. 1.14

default construal: τί αἱ θύραι νύκτωρ ψοφοῦσιν; [pres.ind] non-default construal: τί αἱ θύραι νύκτωρ ἐψόφουν; [ipf.ind]23 23

The context makes clear that this is the right reading: ἀναμιμνῃσκόμενος δὲ ὅτι ἐκείνῃ τῇ

syntactic variation with verbs of perception

71

e. ἔλεγεν ὅτι Ἐρατοσθένης αὐτὸν […] εἰς τὸ δεσμωτήριον ἀπαγάγοι [aor.opt]. ‘He told me that Eratosthenes had taken him [sc. my brother] off to prison’. lys. 12.16

default construal: Ἐρατοσθένης αὐτὸν εἰς τὸ δεσμωτήριον ἀπήγαγεν [aor.ind] non-default construal: Ἐρατοσθένης αὐτὸν εἰς τὸ δεσμωτήριον ἀπάγει [pres.ind] In Méndez Dosuna (1999), I argued that the oblique optative had nothing to do with a temporal backshift, but was a marker of reported speech. With regard to tense, the oblique optative retains the aspect of direct speech, but the time reference becomes opaque. The temporal construal of the oblique optative depends once again on discourse pragmatics. An optative present (15c) usually stands for a present in direct speech, but it can also stand for an imperfect as in (15d); an optative aorist (15e) usually corresponds to an aorist in direct speech for the simple reason that instantaneous presents are uncommon. The conception of the oblique optative as a backshifting device is no less problematic from the viewpoint of diachrony. The construction was still in its embryonic stages in Homer and consolidated itself at a relatively late date. By the time it became a full-fledged category, the optative mood could denote a wish (deontic modality), a contingency (epistemic modality) or a habitual aspect in some subordinate clauses, but was no longer capable of encoding time, if it ever had been.24 A development potential > (relative) tense makes little sense.25 I posited a more natural grammaticalization path starting from

24

25

νυκτὶ ἐψόφει ἡ μέταυλος θύρα καὶ ἡ αὔλειος: ‘Remembering that on that night the inner and outer doors had been banging’. (Lys. 1.17) Like the oblique optative, the so-called iterative optative occurs in past-oriented contexts (see (4a) for an example), but it does not denote past temporal reference by itself. Interestingly, the polysemy habitual-irrealis is not uncommon (see Horrocks 1985 for Greek): see the double reading of e.g., Classical Greek ἔλεγον ἄν (imperfective habitual, imperfective counterfactual), εἶπον ἄν (perfective habitual past, perfective counterfactual), λέγοιμι (ἄν) (imperfective potential, imperfective habitual), εἴποιμι (ἄν) (perfective habitual past, perfective counterfactual), Modern Greek θα έλεγα (past future, inferential past imperfective, counterfactual, imperfective habitual), θα είπα (inferential past perfective), or English I would say (counterfactual, habitual past). The pluperfect of Modern Greek (e.g., είχα πει < Classical Greek εἶχον εἰπεῖν) resulted from

72

méndez dosuna

the idea of a potential (epistemic modality), which in contexts of implicit or explicit reported speech was reanalyzed as an evidential (epistemological modality). This in turn underwent semantic bleaching to become a mere marker of reported speech (quotative).26 Crucially for the point discussed in this paper, the oblique optative in final clauses provides incontestable evidence against the RTTh, since, as indicated above, simultaneity, and especially anteriority with respect to the verb of the main clause, are excluded in final clauses: (16) a. ἐγὼ πρότερον μὲν ἐξέπλευσα [aor.ind] ἐκ τῆς πόλεως ἔχων τὸ μειράκιον, ἵνα μὴ τούτῳ μαχοίμην [pres.opt] ‘First I left the city on a voyage keeping the boy with me to avoid fighting with this man [so I said]’. lys. 3.32

b. οἱ τριάκοντα […] προσέταξαν [aor.ind] ἀγαγεῖν ἐκ Σαλαμῖνος Λέοντα τὸν Σαλαμίνιον ἵνα ἀποθάνοι [aor.opt] ‘The Thirty Tyrants ordered [us] to bring Leon the Salaminian from Salamis to be put to death [so they said]’. pl. Ap. 32c

4.7

The ‘Oblique Imperfect’

In Méndez Dosuna (1999:340–341), I touched on another phenomenon, which on the surface looks like the perfect match of sequence of tenses proper in Greek: the so-called ‘oblique imperfect’.27 Consider the variation of the verb of

26

27

a past future reanalyzed as a past counterfactual (Horrocks 2010:300). Paradoxically, the Latin pluperfect has become a past subjunctive in Spanish (Lat. dixeram ‘I had said’ > Sp. dijera ‘I would say’). In both cases the pivotal category was past counterfactual in conditional contexts. Recently, de la Villa (1999:2014) and Faure (2009:2014) have insisted in interpreting the oblique optative in terms of relative tense (backshift and a mystifying narrative tense respectively), but their interpretation of the data is often arbitrary and sometimes incorrect; their arguments are not cogent and the diachronic evolution they implicitly take for granted involves an inexplicable somersault from mood to relative tense. As it seems, the term ‘oblique imperfect’ was first coined by Hermann (1915); see also Fraenkel (1949:132–136) and Ruipérez (1972).

syntactic variation with verbs of perception

73

the subordinate clauses governed by the verbs ἐμφανίζω and δηλόω ‘show, make see’ in the following examples: (17) a. ἐνεφανίσσοεν [ipf.ind] αὐτοῦ πόκκι ἁ ἀμμέουν πόλις διὲ τὸς πολέμος ποτεδέετο [ipf.ind] πλειόνουν τοῦν κατοικεισούντουν· ‘[The envoys] made clear to him that on account of the wars our city was in need of a larger contingent of those willing to inhabit it’. ig ix 2 517, 12–13; Larisa, 214 bce

b. ἐνεφάνιζον [ipf.ind] μοι ὅτι ἡ ὑμετέρα πόλις διὰ τοὺς πολέμους προσδεῖται [pres.ind] πλεόνων οἰκητῶν· ‘[The envoys] made clear to me that on account of the wars your city is in need of more inhabitants’. ibid. l. 5

c. καὶ ἂν ἐδήλου [ipf.ind] πᾶσι ὡς ἐπὶ Σκύθας ἐλαύνει [pres.ind] ‘He would keep showing to everyone that he was marching against the Scythians’. hdt. 4.118.4

d. ἐδήλουν [ipf.ind] δὲ οὐδὲν ὅτι ἴσασι [pres.ind] τὰ πρασσόμενα ‘They did not disclose that they knew the scheme’. thuc. 4.68.6

e. ἑωυτοῖσι τε ἐδήλου [ipf.ind] λόγῳ ὡς εἴη [pres.opt] καὶ πόλις καὶ γῆ μέζων ἤ περ ἐκείνοισι ‘[Themistocles] declared by word of mouth that both their city and their land [sc. of the Athenians] were greater than theirs [sc. of the Corinthians]’. hdt. 8.61.2

f. πανταχοῦ ἐδήλου [ipf.ind] ὡς ἐλευθερώσων τὴν Ἑλλάδα ἐκπεμφθείη [aor.opt].

74

méndez dosuna

‘He repeatedly made plain everywhere that he had been sent out to free Greece’. thuc. 4.108.2

(17a) and (17b) occur in an inscription which records a letter written in Koine that Philip v of Macedon sent to the citizens of Larisa (Thessaly). Apart from some minor changes in the wording, in the clause governed by ἐνεφανίσσοεν / ἐνεφάνιζον, the present προσδεῖται in the original Koine version (17b) has become an imperfect ποτεδέετο in the translation into the local vernacular (17a).28 Example (17c) illustrates the use of the present in the embedded clause in Ionic and Attic. Examples (17d–e) illustrate a third option: the use of the oblique optative. A recently published inscription from Larisa documents new occurrences of the ‘oblique imperfect’ in Thessalian. The relevant passage of the text in the local vernacular in (18a) is followed by a tentative translation into Attic (18b) and by the French translation (18c) provided by the first editors (Tziafalias et al. 2006:457): (18) a. ὀπειδεὶ Νούιος Λατῖνος Οὐίοι Μαμερτῖνος παργενόμενος καὶ ἐπελθοὺν ἐτ τὰν παρελθόνσαν ἀγορὰν ἐνεφά̣[νισ]σε [ipf.ind] πόκ̣ [κ]ι, Πατροκλέα τοῖ Ἀστοκλέα παιδὸς ποππετόντος ἑυτοῦ ἐν Σικελία καὶ ἐμφανίσσοντος [pres.part] διέκι ἀνδραποδισθὲς ὕκ κινουν στρατιουτᾶν ἐδούλευε [ipf.ind], ἐπιγνὸς [aor.part] διέκι εἶς [ipf.ind] ἐλεύθερος καὶ Λαρισαῖος ἐξαγγρε(ι)μένος εἶς [plpf.ind], κα[ὶ] κατεσταούκουν ἐλεύθερον καὶ ἑυτοῖ κύριον29 seg lvi 636.41–47, Larisa, late 2nd c. bce

b. ἐπειδὴ Νούιος Λατῖνος Οὐίου Μαμερτῖνος παραγενόμενος καὶ ἐπελθὼν ἐπὶ τὴν παρελθοῦσαν ἀγορὰν ἐνεφάνισε ὅτι Πατροκλέου τοῦ Ἀριστοκλέου παι-

28

29

It is important to stress that the difference between the present in the Koine and the imperfect in Thessalian does not replicate the contrast in English between He explained that the city needed [past] more inhabitants vs. He explained that the city needs [present] more inhabitants, where the lack of backshift implies that the need continues to exist at the moment of speech. In this intricate sentence, the clauses of the ‘oblique imperfects’ ἐδούλευε and εἶς [ipf.ind] are governed by the participles ἐμφανίσσοντος and ἐπιγνούς respectively. The relevant time reference is provided by the imperfect ἐνεφάνισσε, which in turn governs πόκ̣ [κ]ι … ἐξαγγρε(ι)μένος εἶς κα[ὶ] κατεσταούκουν ἐλεύθερον καὶ εὑτοῖ κύριον.

syntactic variation with verbs of perception

75

δὸς προσπεσόντος ἑαυτοῦ ἐν Σικελίᾳ καὶ ἐμφανίζοντος ὅτι ἀνδραποδισθεὶς ὑπό τινων στρατιωτᾶν ἐδούλευε, ἐπιγνοὺς ὅτι ἦν ἐλεύθερος καὶ Λαρισαῖος, ἐξῃρημένος ἦν, καὶ καθεστηκὼς ἐλεύθερον καὶ ἑαυτοῦ κύριον. c. Attendu que Novius Latinus fils d’Ovius, Mamertin étant venu et s’étant présenté à l’assemblée lors de sa session passée, a fait savoir que, après que Patrokléas fils d’Astokléas s’ était présenté devant lui en Sicile comme suppliant et lui avait fait savoir que, ayant été capturé et réduit en esclavage par des soldats, il demeurait en esclavage, il avait eu confirmation que Patrokléas était un homme libre et qu’ il était effectivement originaire de Larissa, et qu’ alors il l’ avait libéré par rachat, l’ayant (ainsi) rétabli dans sa condition d’ homme libre et lui ayant rendu la maîtrise de lui-même. For the ‘oblique imperfect’ in Homer, Ruipérez adduces the following example: (19) οὐ γάρ οἵ τις ἐτήτυμος ἄγγελος ἐλθὼν ἤγγειλ’ [aor.ind] ὅττι ῥά οἱ πόσις ἔκτοθι μίμνε [ipf.ind] πυλάων ‘For no true messenger had come and announced to her that her husband remained outside the gates’. hom. Il. 22.438–439

The contrast between the present in (17b–c) and the imperfect in (17a, 18a, 19) has been interpreted in terms of a dialectal variation (see Bechtel 1921:202; Ruipérez 1972 and García Ramón 2011): allegedly, the ‘oblique imperfect’ was characteristic of some dialects, notably Thessalian and Homer (an Aeolicism of the epos?). The imperfects and pluperfects of (17a, 18a, 19) purportedly serve the syntactic function of adapting a present of direct speech to the past of the reporting verb.30 For those who believe that the oblique optative is also a backshifting device, the ‘oblique imperfect’ and the oblique optative in (17d–e) are just alternative manifestations of one and the same syntactic phenomenon. One of the most characteristic features in the verbal morphology of the Thessalian dialect is the ending -(ι)εν in the 3rd pl. of the imperfects and indicative aorists: e.g., imperfect ἐνεφανίσσοεν (Att. ἐνεφάνιζον), εἴχοιεν (Att.

30

Hermann (1915) uses the term Tempusassimilation.

76

méndez dosuna

εἶχον), aorist ἐκρίνναεν (Att. ἔκριναν), ἐστάσαιεν (Att. ἔστασαν). Building on a suggestion put forward by Hoffmann (1893:319), Morpurgo Davies (1965) came to the conclusion that Thessalian -εν and -ιεν were taken over from 3pl optative -ιεν. García Ramón (2011) carries the argument one step further and sees in the use of the ‘oblique imperfect’ and the oblique optative the connection that induced this improbable analogy. All these views are mistaken. To begin with, Bechtel (1921:202) and Fraenkel (1949:133–134) did not fail to note that the ‘oblique imperfect’ is far from alien to Attic,31 as the following examples illustrate: (20) a. ἀνέδησε [aor.ind] τὸν ἡνίοχον βουλόμενος δηλῶσαι ὅτι ἑαυτοῦ ἦν [ipf.ind] τὸ ἅρμα· ‘He crowned the charioteer with the intention of making plain that the chariot was his’. thuc. 5.50.4

b. ὁ δ’ ἀποθανὼν ἐδήλωσεν [aor.ind] ὅτι οὐκ ἀληθῆ ταῦτα ἦν [ipf.ind]· ‘And, when he died, he made plain that this was not true’. lys. 19.52

c. ἐμφανίζων [pres.part] δὲ καὶ τοῦτο ὅτι περὶ πολλοῦ ἐποιεῖτο [ipf.ind] μηδένα μήτε φίλον ἀδικεῖν μήτε σύμμαχον […] μᾶλλον καὶ τοὺς ἄλλους ᾤετ’ [ipf.ind] ἂν τῶν μὲν αἰσχρῶν κερδῶν ἀπέχεσθαι ‘He thought that, by making plain also the following, that he attached a great importance to never being unfair to a friend or an ally, others also would be more likely to abstain from improper gains’. xen. Cyr. 8.1.26

Nor did it escape Ruipérez that the use of the ‘oblique imperfect’ in embedded clauses governed by verbs of perception and cognition (including causative verbs of perception and cognition) in the past is attested both in Homer and in Attic: (21) a. γνῶ [aor.ind] δ’ Ὀδυσεὺς ὅ οἱ οὔ τι τέλος κατακαίριον ἦλθεν [aor.ind!] 31

Fraenkel refers to Kühner-Gerth (1904:§ 550.3, 590c); see also Rijksbaron (2006:50–51).

syntactic variation with verbs of perception

77

‘Odysseus realized that his fatal hour had not yet come’. hom. Od. 11.439

b. ἐπεὶ ᾔσθετο [aor.ind] ὅτι τὸ Μένωνος στράτευμα ἤδη ἐν Κιλικίᾳ ἦν [ipf.ind] εἴσω τῶν ὀρέων ‘Because [Cyrus] had learned that Menon’s army was already in Cilicia, inside of the mountains’. xen. An. 1.2.21

c. ἐδήλωσαν [aor.ind] ὅτι οὐ πλαστὴν τὴν φιλίαν παρείχοντο [ipf.ind]. ‘They showed that their affection was unfeigned’. xen. Ages. 1.38

d. τί γὰρ ᾔδειν [ipf.ind] εἴ τι κἀκεῖνος εἶχε [ipf.ind] σιδήριον; ‘For how did I know whether he too had some weapon?’ lys. 1.42

e. ὑμῖν ἐδήλωσαν [aor.ind] ὅτι οὐδὲ τὸν Πειραιᾶ Λακεδαιμονίων προσταττόντων περιεῖλον [aor.ind!] ‘They revealed to you that they did not dismantle the Peiraeus because the Lacedaemonians imposed it either’. lys. 12.40

f. ὅτι δ’ ἀληθῆ ἔλεγεν [ipf.ind], ὧδ’ ἐδίδασκεν [ipf.ind]· ‘This is how he demonstrated that he was telling the truth’. xen. Mem. 7.2.1

In view of this evidence, the label ‘oblique imperfect’ is doubly inadequate. Firstly, as the aorists of (21a) and (21e) and the pluperfect of (19) demonstrate, the phenomenon, whatever its nature may be, concerns past tenses in general rather than the imperfect alone. Imperfects happen to be more frequent for the trivial reason that perceived events are often durative. Secondly, the examples (21a–e) are not instances of reported speech (oratio obliqua). Similarly, the verb ἐπιγιγνώσκω, which takes an ‘oblique imperfect’ in (18b), is a verb of cognition, not a reporting verb. Unlike the oblique optative, the ‘oblique imperfect’

78

méndez dosuna

is never attested in combination with prototypical reporting verbs like λέγω or ἀγορεύω.32 Consequently, despite superficial similarities, the ‘oblique imperfect’ differs substantially from sequence of tenses. Verbs like δηλόω and ἐμφανίζω were originally causative verbs of perception (‘show’ = ‘make see’) (see δηλόω in 20d) that secondarily evolved into reporting verbs through a common semantic metamorphosis (note the pleonasm ἐδήλου λόγῳ in 16e). As a consequence, these verbs partake of a dual nature. On the one hand, the construction with a verb in the past (17a, 18a, 19, 20a–c, 21a–f) is a syntactic relict which conforms to their etymological meaning. Conversely, the present (17b–c) and the oblique optative (17d–e) are so to speak syntactic updatings more consonant with their new role as reporting verbs. Ἀγγέλλω has evolved in the reverse direction. Etymologically, it was a reporting verb (‘to act as a messenger’ (ἄγγελος) > ‘to deliver a message’) which took an object clause with ὅτι/ὡς with the tense of direct speech (22a) or the oblique optative (22b), or, alternatively, a declarative infinitive (22b): (22) a. ἠγγέλλετο [ipf.ind] ὅτι ἐν Ῥηγίῳ αἱ νῆές εἰσι [pres.ind] ‘It was repeatedly announced that the ships were at Rhegium’. thuc. 6.45.1

b. ὄντι δ’ αὐτῷ ἐν Ἀμφιπόλει ἀγγέλλει [pres.ind] Δερκυλίδας ὅτι νικῷέν [pres.opt] τε αὖ Λακεδαιμόνιοι, καὶ αὐτῶν μὲν τεθνάναι [pf.inf] ὀκτώ ‘When [Hagesilaus] was at Amphipolis, Dercylidas brings him word that the Lacedaemonians are victorious, and that eight of them are dead’. xen. Hell. 4.3.1

32

Pace de la Villa (1999:360) and García Ramón (2011), ἦν in (15b) and similar instances are not ‘oblique imperfects’ (sc. backshifted presents). It is worth noting in this regard that, while the oblique optative governed by a historical present is commonplace (23b is a case in point), the use of the ‘oblique imperfect’ in such a context is unheard of. Of course, in an example like δηλοῖ [pres.ind] δὲ μάλιστα Ὅμηρος ὅτι τοιαῦτα ἦν [ipf.ind] ἐν τοῖς ἔπεσι τοῖσδε ‘Homer makes clear that this was so in the following verses’ (Thuc. 3.104.4), in which a present in the main clause is combined with a past in the subordinate clause, the present cannot be interpreted as ‘historical’ since Homer’s works still remained a speaking testimony at the time when Thucydides was writing his history.

syntactic variation with verbs of perception

79

However, on the analogy of its quasi-synonyms δηλόω, ἐμφανίζω, it developed the ability to take the ‘oblique imperfect’ in (19). The same ambiguity explains why ἀγγέλλω can also be constructed with participles as a causative verb of perception: (23) a. ἔλεγεν [ipf.ind] ὡς ἀγγέλλοιτο [pres.opt] ὁ μὲν Πείσανδρος τετελευτηκώς [pf.part] ‘He kept saying that Peisander was reported as being dead’. xen. hell. 4.3.13

b. αὐτῷ Κῦρόν τε ἐπιστρατεύοντα [pres.part] πρῶτος ἤγγειλα [aor.ind] ‘I was the first to inform him [sc. the King of Persia] that Cyrus was marching against him’. xen. An. 2.3.19

It becomes clear from the above evidence that the ‘oblique imperfect’ was not exclusive to Thessalian and that it was essentially different from the oblique optative. The overlapping area of both constructions is restricted to a very limited number of verbs. Consequently, pace García Ramon (2011), the interplay of the ‘oblique imperfect’ and the oblique optative is unlikely to be the fulcrum of the spread of the ending -(ι)εν to the 3rd pl. of the past indicative.33 As regards the ‘oblique imperfect’, the crucial point here is that, as explained above, the two events involved in predicates of immediate perception are necessarily simultaneous. When the perception predicate belongs in the past, the perceived situation belongs also in the past. There is no backshift as the following English examples demonstrate: (24) a. John saw Mary eating an egg ≈ Mary was eating an egg and John saw her. b. John said that Mary was eating an egg ≠ Mary was eating an egg and John said so (?).

33

In Méndez Dosuna (forthcoming) I try to demonstrate that Thessalian 3pl -εν was borrowed from 3pl imperfect εἶεν [eːen] (< ἦεν) (Att. ἦσαν) as Hoffmann (1893:319) correctly guessed, while -ιεν was taken over from 3pl imperfect *εἶεν [eien] (< *ἦιεν) (Att. ᾖσαν).

80

méndez dosuna

In (24a) the perceived event belongs in the past. It is factive and independent of the act of perception (see Rijksbaron 2006:50–51 on Greek). Conversely, in a declarative sentence like (24b), the state of affairs of the embedded sentence is not factive and entirely dependent of the reporting verb. The periphrasis does not necessarily correspond to the declarative sentence since John may have told a lie. Notice that a negative version like Mary was not eating an egg, but John said so sounds much more natural than Mary was not eating an egg, but John saw her: lies are much more frequent than hallucinations. These differences explain why declarative predicates and perception predicates may behave quite differently in the languages of the world. The constructions in Russian (Comrie 1986:294; Noonan 2007:98–99) provide a close parallel with the Greek facts. (25) a. Petr ponjal [past], čto on [pres.] bolen. b. Petr ponjal [past], čto byl [past] bolen. ‘Peter realized that he was ill’. c. Petr skazala [past], čto on [pres] bolen. d. † Petr skazala [past], čto byl [past] bolen. ‘Peter said that he was ill’. The verb of a clause governed by a perception verb can be put interchangeably in the present or in the past with a fake sequence of tenses (25a–b). Conversely, subordinate clauses governed by a declarative verb are refractory to backshift (sequence of tenses proper) (25c–d). The morphology of the Russian verb crucially shares another feature with Ancient Greek: it has no tenses expressing rtr.

4.8

Conclusion

In the preceding sections I have argued against the proponents of the RTTh that Ancient Greek had no tenses capable of denoting rtr. To all appearances, this absence had important consequences for other realms of grammar. It precluded, for instance, the possibility of a formal differentiation of

syntactic variation with verbs of perception

81

present vs. past counterfactual sentences comparable to the following English examples:34 (26) a. If John had money, he would buy a car. b. If John had had money, he would have bought a car. The odds are that the existence of tenses denoting rtr is a precondition for the existence of rules of backshift (consecutio temporum) in reported speech in a given language. The typological coincidences between Greek and Russian can hardly be attributed to sheer chance, since Russian also lacks relative tenses. The implicative generalization that the existence of tenses denoting rtr is a prerequisite both for a differentiation of past and present counterfactual conditionals and for a rule of tense backshift may have a cross-linguistic validity, but I leave to typologists the task of confirming or disconfirming it with the evidence gathered from a critical mass of languages. When viewed at a closer distance, the ostensible tense backshift with Greek causative verbs of immediate perception, which secondarily turned into reporting verbs, is a mirage. The label ‘oblique imperfect’ is uncalled for, since this construction was only marginally related to reported speech (oratio obliqua).

References Adrados, F.R. 1992. Nueva sintaxis del griego antiguo. Madrid. Bartolotta, A. (ed.). 2014. The Greek verb: morphology, syntax, and semantics. Louvainla-Neuve. Bechtel, F. 1921. Die griechischen Dialekte. i. Berlin. Bentein, K. 2014. ‘Perfect’. In: Giannakis 2014:3, 46–49. Bybee, J., Perkins, R. & W. Pagliuca. 1994. The evolution of grammar: tense, aspect, and modality in the languages of the world. Chicago. Comrie, B. 1986. ‘Tense in indirect speech’. fl 20:265–296. Crespo, E., Conti, L. & H. Maquieira. 2003. Sintaxis del griego clásico. Madrid. Crespo, E. 2014. ‘A rule for the choice of aorist and imperfect’. In: Bartolotta 2014:71–82. Louvain-la-Neuve. 34

The rule mentioned in most handbooks that the imperfect is used in present counterfactual periods and the aorist indicative in past counterfactual periods is specious (Wakker 1994:146–150).

82

méndez dosuna

Faure, R. 2009. ‘L’optatif oblique serait-il un temps?’ lalies 30:281–293. Faure, R. 2014. ‘The oblique optative: a case of narrative tense. The example of the future optative’. In: Bartolotta 2014:131–148. Fraenkel, E. 1949. ‘Griechisches und Italisches’. if 60:131–155. García Ramón, J.L. 2011. ‘The Thessalian secondary ending 3rd pl. -(ι)εν and the optative’. Book of abstracts of the 6th international colloquium on Ancient Greek dialectology, Nicosia, University of Cyprus, September 26–29, 2011. Giannakis, G. (ed.). 2014. Encyclopedia of Ancient Greek language and linguistics. Leiden. Haspelmath, M. 1992. ‘From resultative to perfect in Ancient Greek’. Nuevos estudios sobre construcciones resultativas, ed. by J.L. Iturrioz Leza, 187–224. Guadalajara. Hettrich, H. 1976. Kontext und Aspekt in den altgriechischen Prosa Herodots. Göttingen. Hermann, E. 1915. ‘Graeca 5: Thessalisches Imperfektum obliquum’. if 35:166. Hoffmann, O. 1893. Die griechischen Dialekte in ihrem historischen Zusammenhange. ii. Göttingen. Horrocks, G.C. 1985. ‘On condition: aspect and modality in the history of Greek’. PCPhS 41:153–173. Horrocks, G.C. 20102. Greek: a history of the language and its speakers. Chichester. Humbert, J. 1960. Syntaxe grecque. Paris. Jacquinod, B. (ed.). 1999. Les complétives en grec ancien: actes du colloque international de Saint-Étienne (3–5 septembre 1998). Saint-Étienne. Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. Manolessou, I. 2005. ‘From participles to gerunds’. Advances in Greek generative syntax in honor of Dimitra Theophanopoulou-Kontou, ed. by M. Stavrou & A. Terzi, 241–283. Amsterdam. Martínez Vázquez, R. 1995. ‘Tiempo relativo en griego antiguo’. Philologia Hispalensis 1995:175–197. Méndez Dosuna, J. 1999. ‘La valeur de l’optatif oblique grec: un regard fonctionneltypologique’. In: Jacquinod 1999:331–352. Méndez Dosuna, J. 2008. ‘To die in Ancient Greek: on the meaning of ἀπο- in ἀποθνῄσκειν’. Θερμή και φως. Licht und Wärme. Αφιερωματικός τόμος στη μνήμη του A.-Φ. Χρηστίδη / In memory of A.-F. Christidis, ed. by M. Theodoropoulou, 245–255. Thessaloniki. Méndez Dosuna, J. Forthcoming. ‘Thessalian secondary 3rd. pl. –(ι)εν and the optative: dangerous liaisons’. Studies in ancient Greek dialects from Central Greece to the Black Sea, ed. by G. Giannakis, E. Crespo & P. Filos. Berlin. Mirambel, A. 1961. ‘Participe et gérondif en grec médiéval et moderne’. bsl 56:46–79. Morpurgo Davies, A. 1965. ‘A note on Thessalian’. Glotta 43:235–251. Moser, A. 2002. ‘Μετοχές, γερούνδια, -οντας: χρόνος και άποψη’. Recherches en linguistique

syntactic variation with verbs of perception

83

grecque: Procédés du 5ème congres international de linguistique grecque, Sorbonne, 13–15 septembre 2001, ed. by Ch. Clairis, ii:111–114. Paris. Moser, A. 2006. ‘The Greek forms in -ontas: a study in “converbiality”, temporality, aspectuality and finiteness’. Glossologia 17:43–67. Noonan, M. 19852. ‘Complementation’. Typology and syntactic description, ed. by T. Shopen, ii:52–150. Cambridge. Quirk, R. et al. 1985. A comprehensive grammar of the English language. London. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Chicago. Stahl, J.M. 1907. Kritisch-historische Syntax des historischen Verbums der klassischen Zeit. Heidelberg. Ruijgh, C.J. 1991. ‘Les valeurs temporelles des formes verbales en grec ancien’. The function of tense in texts, ed. by J. Gvozdanovic & T.A.J.M. Janssen, 197–217. Amsterdam. Ruipérez, M.S. 1954. Estructura de tiempos y aspectos del verbo griego antiguo: análisis funcional sincrónico. Salamanca. Ruipérez, M.S. 1972. ‘Imperfecto oblicuo y pretérito de futuro en griego antiguo’. Homenaje a Antonio Tovar, ed. by J.L. García Ramón, 425–435. Madrid. Tsimpli, I.-M. 2000. ‘Gerunds in Greek’. jgl 1:133–169. Tziafalias, A., García Ramón, J.L. & B. Helly. 2006. ‘Décrets inédits de Larissa (2)’. bch 130:435–483. de la Villa, J. 1999. ‘L’indicatif du passé dans les propositions complétives du grec ancien’. In: Jacquinod 1999:353–365. de la Villa, J. 2010. ‘Estrategias lingüísticas. La expresión del tiempo relativo en griego antiguo’. Plenary presentation at the xxxix simposio de la Sociedad Española de Lingüística (Santiago de Compostela, 1–4 de febrero, 2010). de la Villa, J. 2014. ‘Relative tense’. In: Giannakis 2014:3, 231–232. Wakker, G.C. 1994. Conditions and conditionals: an investigation of Ancient Greek. Amsterdam.

chapter 5

The Gnomic Aorist in Hesiod* Gerry C. Wakker

5.1

Introduction

One of the most vexed questions in Ancient Greek linguistics concerns the so-called gnomic aorist and especially how this use of the (past) indicative aorist stem fits into the system of Greek aspect and time reference.1 It is widely accepted that the indicative ii of the aorist stem expresses a past and completed state of affairs. In general truths and descriptions of habits, the gnomic aorist therefore seems an anomaly. Note, for example, the proverbial: (1) Ὦ Πέρση, σὺ δ’ ἄκουε δίκης μηδ’ ὕβριν ὄφελλε· ὕβρις γάρ .... τε κακὴ δειλῷ βροτῷ, οὐδὲ μὲν ἐσθλὸς ῥηιδίως φερέμεν ............ δύναται, ............ βαρύθει δέ ... θ’ ὑπ’ αὐτῆς ἐγκύρσας ἄτῃσιν· ὁδὸς δ’ ἑτέρηφι παρελθεῖν κρείσσων ἐς τὰ δίκαια· δίκη δ’ ὑπὲρ ὕβριος ἴσχει ....... ἐς τέλος ἐξελθοῦσα· παθὼν δέ .... τε νήπιος ἔγνω. ‘But you, Perses, listen to right and do not foster violence; for violence is bad for a poor man. Even the prosperous cannot easily bear its burden, but is weighed down under it when he has fallen into delusion. The better path is to go by on the other side towards justice; for Justice beats Outrage when she comes at length to the end of the race. But once he has suffered the fool realizes this’. hes. Op. 213–2182

* My thanks are due to the participants of the 7th International Colloquium on Ancient Greek Linguistics (icagl7) in Ghent (2011) for their valuable comments and especially to Sander Orriens for critically reading an earlier draft of this paper. 1 For earlier descriptions, see e.g., Goodwin (1889:53–56); Kühner-Gerth (1904:158–163); Stahl (1907:131–134); Humbert (1960:145–146); Péristérakis (1962); Chantraine (1963:185–187); Ruijgh (1971:255–265); Monro (1986:67); McKay (1988); Bakker (1999; 2001; 2005); and Rijksbaron (2006:31–33; 2009:241–247). 2 The translations are those of H.G. Evelyn-White (1914), unless otherwise mentioned. When necessary I adapt my translation to my interpretation (as explained in the text). In the

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_006

the gnomic aorist in hesiod

85

The statement with the gnomic aorist in v. 218 is part of a passage which illustrates the advice given to Perses to follow the path of justice (Op. 213).3 This passage (Op. 214–218) gives a general or omnitemporal description (with present indicatives) of the general nature of violence (ὕβρις) and its counterpart justice (δίκη).4 This omnitemporal nature is also marked by the use of the epic particle τε (214, 215, 218), which indicates that permanent characteristics are described (Ruijgh 1971). Given this context, the indicative ii aorist ἔγνω clearly does not refer to the past, but is omnitemporal. A question arises about how to explain this use of the indicative ii aorist and how to link it to other uses of the indicative ii aorist: – What is the origin of this use of the indicative ii aorist? – What is the synchronic value of this use of the indicative ii aorist in Ancient Greek? – How do we explain that both the indicative i present and indicative ii aorist are used in omnitemporal statements? Which factors influence the choice between the two options and what is the (semantic) difference between them? Consider, for instance, the following example: (2) νικήσας δὲ βίῃ καὶ κάρτει καλὸν ἄεθλον ῥεῖα ........ φέρει χαίρων τε, τοκεῦσι δὲ κῦδος .......... ὀπάζει. […] καὶ τοῖς οἳ γλαυκὴν δυσπέμφελον ἐργάζονται, εὔχονται δ’ Ἑκάτῃ καὶ ἐρικτύπῳ Ἐννοσιγαίῳ, ............. ῥηιδίως ἄγρην κυδρὴ θεὸς ὤπασε πολλήν, ῥεῖα δ’ ἀφείλετο φαινομένην, ἐθέλουσά γε θυμῷ. ‘He who by might and strength gets the victory wins the rich prize easily with joy, and brings glory to his parents. […]

Greek examples, ............................. dotted underlining marks indications of the omnitemporal character of the passage (particles, indicative i present, distributive iterative subjunctives etc.), whereas boldface marks (gnomic) aorist forms. 3 Closely parallel is Il. 20.198 ῥεχθὲν δέ τε νήπιος ἔγνω. 4 See Rijksbaron (2009:243–244) for the preference for ‘omnitemporal’, rather than ‘timeless’. Whereas timeless refers to what is outside time altogether (e.g., mathematical truths), an omnitemporal statement is ‘time-bound, but temporally unrestricted’, i.e., ‘something has been, is and always will be so’ in a finite or infinite set of time-points or time intervals, see Lyons (1977:680).

86

wakker

And to those whose business is in the grey discomfortable sea, and who pray to Hecate and the loud-crashing Earth-Shaker, easily the glorious goddess [= Hecate] gives great catch, and easily she takes it away as soon as seen, if so she will’. hes. Theog. 437–443

Here we see that the same verb in the same omnitemporal passage is found in both the indicative i present (ὀπάζει) and the indicative ii aorist (ὤπασε). For what reason? I will study the above questions in relation to one corpus, Hesiod’s Theogony and Works and Days, in which the gnomic aorist is less studied than in Homer. Moreover, the number of examples is limited,5 so that it is possible to study all relevant examples exhaustively in the framework of one article. After examining the possible origin of the gnomic aorist (5.2), I discuss the synchronic value of the gnomic aorist (5.3), evaluate the dubious examples (5.4) and present my conclusions (5.5).

5.2

Origin of the Gnomic Aorist

Despite many hypotheses, it seems impossible to prove how the gnomic aorist came into being. At most one can point to three factors that perhaps influenced the interpretation of an aorist as gnomic: i. Often the state of affairs in the gnomic aorist is anterior to another state of affairs in the context, as is often the case when a (non-indicative or indicative) aorist form is used in other (i.e., not omnitemporal) contexts, cf. the use of the aorist participle and the aorist in subordinate clauses. The idea of anteriority may be exemplified by (3): (3) [Happy is he whom the Muses love: from his mouth the stream flows sweet. If a man has sorrow, a singer will chant the deeds of men of old and of the gods.] αἶψ’ ὅ γε δυσφροσυνέων ................ ἐπιλήθεται οὐδέ τι κηδέων μέμνηται· ταχέως δὲ παρέτραπε δῶρα θεάων. .............

5 In total the number of examples in Hesiod is about forty (depending on the interpretation of the dubious or ambiguous examples, for which see 5.4, and because of textual uncertainties).

the gnomic aorist in hesiod

87

‘At once [the man with sorrows] forgets his heaviness and remembers not his sorrows at all; but the gifts of the goddesses soon turn him away from these’. hes. Theog. 102–103

Here one could say that παρέτραπε expresses an (omnitemporal) activity that is terminative and completed, and hence anterior to ἐπιλήθεται and μέμνηται, verb forms that are also omnitemporal and that describe situations.6 However, one could ask whether this explains the origin of the gnomic aorist. Rather, this is a characteristic and frequent usage linked to the perfective value attributed to the aorist stem in general. It does not explain why the aorist indicative came to be used in non-past contexts. ii. The use of the gnomic aorist is influenced by the use of the empiric aorist (also called aorist of experience) with πολλάκις/ἀεί etc., see (4), which describes the punishment inflicted by Zeus on wrongdoers: τε μέμηλε κακὴ καὶ σχέτλια ἔργα, (4) οἷς δ’ ὕβρις ............... τοῖς δὲ δίκην Κρονίδης .................. τεκμαίρεται εὐρύοπα Ζεύς. πολλάκι καὶ ξύμπασα πόλις κακοῦ ἀνδρὸς ἀπηύρα, ............ ὅστις ἀλιτραίνῃ καὶ ἀτάσθαλα μηχανάαται. ....................... ................. τοῖσιν δ’ οὐρανόθεν μέγ’ ἐπήγαγε πῆμα Κρονίων, λιμὸν ὁμοῦ καὶ λοιμόν, ἀποφθινύθουσι ..................... δὲ λαοί· ‘But for those who practise violence and cruel deeds far-seeing Zeus, the son of Cronos, ordains a punishment. Often even a whole city suffers for a bad man who sins and devises presumptuous deeds, and upon them, from heaven, the son of Cronos lays great trouble, famine and plague together, and the men perish away’. hes. Op. 238–243

Wackernagel, Humbert and Stahl interpret the aorists ἀπηύρα (240) and ἐπήγαγε (242) as empiric. West and Verdenius, however, point to the subordinate clause ὅστις ἀλιτραίνῃ (with subjunctive) and to the fact that the πολλάκι clause explains the general δίκην […] τεκμαίρεται before (as marked by the explicative asyndeton at the start of the πολλάκι clause). They argue that this proves that the clause focuses on the ‘general’ (omnitemporal) situation. 6 See Rijksbaron (2009:257–259) for the syntax and interpretation of vv. 98–103.

88

wakker

In my opinion it is indeed difficult to prove whether the aorists are gnomic or empiric in this case. It is even more difficult to prove that the gnomic aorist developed out of the empiric aorist. They seem to be simply akin. The gnomic aorist, see (1)–(3), primarily marks the omnitemporality of the statement, whereas the empiric aorist (with πολλάκι etc.) focuses on the fact that the state of affairs in question often happened in the past (fact of experience). By conversational implicature it is usually suggested that this observation is not restricted to the past.7 This may also be explicitly indicated by adverbs as ἀεί and/or by the use of epic τε in the context, just as in (4) above. Of course, the empiric aorist may have facilitated the interpretation of the gnomic aorist as referring to the non-past, but in my opinion there are no indications that the gnomic aorist developed out of the empiric aorist. iii. The origin and assumed original value of the augment is often put forward as an argument in the debate about the origin of the gnomic aorist. However, this supposedly original value is also used as an argument for opposite claims. Everyone agrees that in Homer and Hesiod the augment is still optional—a fact interpreted as an archaic feature of the Homeric language,8 but that in similes and gnomic passages the augment tends to be used. In the traditional view, the augment marks ‘pastness’ and this is seen as evidence for the claim that the gnomic aorist came into being relatively late in the development of the language, i.e., at a moment when the use of the augment was already obligatory in the spoken language (see, for instance, Ruijgh 1971:164–165). Bakker and other scholars,9 however, give another, narratologically and semantically motivated interpretation: they claim that the augment does not indicate pastness, but rather nearness, ‘with respect to the speaker’s present and immediate situation’ (Bakker 2001:15). This would explain why the augment prefers discourse pertaining to the speaker’s now, such as speech introductions, similes, proverbs and general statements. I do not consider Bakker’s theory very convincing. First, he pays no attention to the peculiarities of the Homeric text, which is the result of a long oral and finally a written tradition, for which reason it is difficult to make purely synchronic analyses.10 Second, he fails to explain how a marker of ‘nearness’ 7 8 9 10

See Grice (1975:50) and Wakker (1994:30) for the term conversational implicature. See, for instance, Bakker (2001:2). Bakker (1999; 2001; 2005:117–135); McKay (1988:193); Basset (1989); and Pagniello (2002; 2007). As argued by Orriens (2008). For the nature of the Homeric corpus see, for instance, Shipp (1972) and Wilmott (2007:5–9).

the gnomic aorist in hesiod

89

developed into an obligatory marker of pastness, not only for aorist forms (the only forms dealt with by Bakker), but also for the imperfect (and pluperfect). Third, he does not discuss the facultative element ἐ- in non-proximal deictics like (ἐ)κεῖνος, (ἐ)κεῖ, (ἐ)κεῖθεν, (ἐ)κεῖσε, or words like (ἐ)χθές, that seems to point to a spatial or temporal remoteness and that would perfectly well comply with the augment indicating pastness, and hence, remoteness from the present.11 This would also explain why augmented forms can be used to express counterfactuality, the counterfactual being removed from the present reality as well. These phenomena are not touched upon by Bakker. Finally, his explanation is not in harmony with the omnitemporal character of the passages in which the gnomic aorist is found. How do we explain the use of permanent digressive epic τε or of general or distributive subordinate clauses with subjunctive in the same context as the aorist if this aorist expresses—as Bakker argues—a sense of ‘nearness’? Bakker’s explanation is that the aorist in these cases expresses one instantiation of the general rule, but one may wonder whether this is a convincing argument. Why does this instantiation hold only for the aorist forms in such omnitemporal contexts, whereas the other characteristic elements point instead precisely to the omnitemporality? Moreover, Bakker does not explain the reason for using an aorist instead of a present tense form. An alternative view is presented by Willi (2007:45–48), who argues that the augment originated as the reduplication syllable of verbs with initial first laryngeal (duplication being argued to be a perfectivity marker): *h1e-h1. By analogy this *h1e-syllable, i.e., the later augment, was then transferred to roots with other initial consonants. Next, the aspect marker converted into an allpurpose past tense marker, as Willi argues. While this hypothesis may indeed explain the frequency of the augment with gnomic aorists (which often have a perfective aspect, see 5.3 below), it does not explain why this marker only occurred with aorist indicatives, and not with, for instance, aorist subjunctives or optatives. In my opinion, this makes the hypothesis less likely. For all these questions and observations I prefer the traditional view: the augment marks pastness, first facultatively (in contexts where this marker helped in interpreting the text), later obligatorily. The gnomic aorist belongs to the younger text layers in Homer, and is, therefore, nearly always augmented.

11

Of course, from an etymological point of view it is highly disputable whether the augment ἐ- and the ἐ- in non-proximal deictics have the same origin, but one may wonder whether the language user of the time of Hesiod or Homer was aware of this fact.

90

wakker

The latter remark also holds for Hesiod, where of the thirty-eight indisputable gnomic aorists there are only five12 without augment (= 13 percent).13

5.3

Synchronic Value of the Gnomic Aorist in Hesiod

Let us now turn to the Hesiodic corpus and the question why an aorist indicative is used in these general contexts. We start by observing that all passages concerned: – contain general and omnitemporal statements, see (1) and (3). Usually these descriptions are easily recognizable by the addressee on the basis of his experience and/or general knowledge. The description can be applied also to the specific situation involved. Its value is: it always happens like this, so it can possibly happen now, or, – are so-called attributive passages in which characteristic features and courses of action of someone are described, see (2), where Hecate’s characteristic actions towards fishermen are described. The character of ‘omnitemporality’ is never conveyed by the aorist alone. Instead, the aorist functions in a context where the omnitemporal character is already conveyed by other indications that also help to interpret the aorist as a gnomic one. The following indications are found: – use of the indicative i present with general or omnitemporal value, see (2), (3), (4), (6); – use of particles like τοι, τε, καὶ γάρ (τε). For τε see (1) and for more particles in one context, see, for instance, (5), a somewhat uncertain example, since the interpretation of ἔμμορε as an aorist is disputed:14

12 13

14

Theog. 447 (θῆκεν); Op. 20 (ἔγειρεν), 345 (ζώσαντο), 705 (δῶκεν), 741 (δῶκαν). In past, non-gnomic contexts we see that of the 356 aorists (in the text of the Budé series) there are 190 augmented and 166 not augmented (53 vs. 47 percent). For the imperfect the percentages are comparable: 55 vs. 45 percent. For the interpretation as aorist, see the lexicon of lsj s.v. Most scholars interpret these verb forms as perfects, however. Péristérakis (1962:204) argues that the perfects are used here not for semantic reasons, but because of the lack of an aorist form, which in my opinion is a rather weak reason. Whether an aorist or a perfect verb form, the verb form occurs in any case in an omnitemporal context in which the omnitemporal character is underlined by the use of the particles.

the gnomic aorist in hesiod

91

(5) ἔμμορέ ..... τοι τιμῆς ὅς ... τ’ ἔμμορε γείτονος ἐσθλοῦ. ‘He who enjoys a good neighbor has—I tell you—a precious possession’. hes. Op. 347

The statement gives a further explanation for the observation that a bad neighbour is a disaster, whereas a good one is a great blessing; the fact that it is a general observation and that we deal with gnomic aorists is underlined by both τοι and epic τε. Τοι indicates that the speaker wants the addressee to take notice of this statement as it is important for him (‘I tell you’), whereas τε marks the permanent character of the statement. The statement holds for each individual who has a good neighbour. This individualization explains the use of the indicative (rather than the subjunctive) in the subordinate clause. The presence of relative or conditional subordinate clauses with subjunctive (sometimes with κε/ἄν) shows that reference is to the present and future, rather than to the past, see (3)15 and (6). In (3) the passage starts with a conditional clause with subjunctive, εἰ γάρ τις […] | .......... ἄζηται κραδίην ἀκαχήμενος, αὐτὰρ ἀοιδὸς | ὑμνήσῃ (vv. 98–101). (6) .... ὃς ποταμὸν ......... διαβῇ κακότητ’ ἰδὲ χεῖρας ἄνιπτος, τῷ δὲ θεοὶ .............. νεμεσῶσι καὶ ἄλγεα δῶκαν ὀπίσσω. ‘Whoever crosses a river with hands unwashed of wickedness, the gods are angry with him and bring trouble upon him afterwards’. hes. Op. 740–741

Note the coordination of the present and aorist indicatives in the main clause. There are no examples where it is only the aorist that points to the omnitemporality or habituality of the passage or statement. Whereas the present indicates here that the gods are permanently angry, the aorist indicates that once, afterwards, they give trouble. To cite Péristérakis (1962:58), the gods’ ‘courroux n’est pas passager et voilà qu’il aboutit à ἄλγεα δῶκαν’. What seems clear is that most aorist verb forms16 indicate terminative actions referring to:

15 16

See e.g., Theog. 607, 783, 973; Op. 330–332, 344. Problematic is perhaps προσελέξατο in Op. 499, for which see (14) below.

92

wakker

– crucial actions, which are turning points in the situation in question, see (2), – actions that precede other actions or states in the context, see (3),17 – completed actions, see (4) and (6); often these actions are suddenly completed, which may be explicitly marked by the use of adverbs as ῥεῖα, ταχέως, etc., as in (2), (3),18 or, in the case of stative verbs they refer to: – ingressive actions (see ἔγνω (1), ἐπάχνωσεν ‘is turned to hoar frost’19 in Hes. Op. 360). These observations confirm that I agree with McKay (1988:194–208); Rijksbaron (2006:31–33); and Bary (2009:20–21, 131–132, 175), who state that the aorist is used for aspectual reasons but that due to the lack of the optimal verb form, i.e., an indicative i aorist, a second best option is chosen, the indicative ii aorist, expressing the desired aspect. In that case, the (past) time reference has to be taken for granted. Choosing the other suboptimal form, the indicative i present, would imply that the time reference would be fine, but not the aspect. We may conclude that expressing the adequate aspect was sometimes considered more important than expressing the adequate time reference. In these contexts, however, the time reference cannot be troublesome, because of all other indications marking that an omnitemporal or habitual passage is concerned. Also, the repetition is left out of consideration and is not marked as such, but is implied within the context (the subject being not an individual, but a group or a class, or, in the case of an attributive passage, a description concerning habitual activities). The omnitemporality or habituality of the passage as a whole is marked by other means, as indicated above. Most cases in Hesiod are easily explicable along these lines. They also help to explain why in omnitemporal passages the same verb is sometimes used in the present, sometimes in the aorist. Thus, in (2), the present ὀπάζει seems to focus on the fact that κῦδος is a permanent gift20 to the parents, whereas the aorist

17 18

19 20

See also Op. 289, 508, 512, 580. See also e.g., Theog. 87, 103, 418, 442, 443; note that these adverbs are in no way exclusively used with the aorist; they are found combined with the indicative i present (Theog. 102, 254, 438; Op. 5, 6, 7, 215, 288, 292, 325, 762), potential optative (Op. 45, 362, 379, 721) and indicative future (Op. 185, 312, 401). When combined with the indicative ii aorist it is always a gnomic aorist. See Verdenius (1985 ad loc.). The same explanation holds for ὀπάζει in Theog. 420.

the gnomic aorist in hesiod

93

ὤπασε (and its opposite ἀφείλετο)21 refers to the (sudden) moment of the giving and taking away, states of affairs that are easily (ῥηιδίως […] ῥεῖα) completed by performing them.

5.4

Disputable Examples

If the above semantic value of the gnomic aorist is accepted, one might ask whether this description can also help us interpret cases of which the interpretation, at least at first sight, seems disputable. In total I identified eight possible cases. First, an example where, after some hesitation, I interpret the aorist as gnomic: (7) [But the other Eris is the elder daughter of dark Night, and she is far kinder to men] ἥ .... τε καὶ ἀπάλαμόν περ ὁμῶς ἐπὶ ἔργον ἔγειρεν· εἰς ἕτερον γάρ τίς τε ἰδὼν ἔργοιο χατίζει .......... πλούσιον, ὃς σπεύδει μὲν ἀρόμεναι ἠδὲ φυτεύειν ζηλοῖ δέ .... τε γείτονα γείτων οἶκόν τ’ εὖ θέσθαι· ........ εἰς ἄφενος σπεύδοντ’· ‘She stirs up even the shiftless to toil; for a man grows eager to work when he considers his neighbour, a rich man who hastens to plough and plant and put his house in good order; and neighbour vies with his neighbour as he hurries after wealth’. hes. Op. 20–24

Here τε in ἥ τε (20) marks the permanent character of this specification of the previous statement ‘she is far kinder to men’. I therefore argue that ἔγειρεν is a gnomic aorist (without augment, see note 13) in a clause attributing a characteristic action to Eris. In the following explanation (note γάρ) this gnomic aorist alternates with the indicative i present (χατίζει in 21, ζηλοῖ in 23). Uncharacteristically, the indicative i present follows the indicative ii aorist, but since the section starts with ἥ τε its omnitemporal character is clear. I note that ἔγειρεν is the reading of manuscript c, whereas the scholia plead for ἐγείρει,

21

See in a gnomic context Il. 16.689–690 = 17.177–178 (ἀφείλετο νίκην | ῥηιδίως).

94

wakker

which, however, seems to be the lectio facilior, and has to be rejected for that reason. The aorist ἔγειρεν is a completed action anterior to stative χατίζει and ζηλοῖ. Next, there are four examples where one may hesitate between an interpretation of a normal or a gnomic aorist: (8) ἣ δ’ ἕπεται κλαίουσα πόλιν καὶ ἤθεα λαῶν, ἠέρα ἑσσαμένη, κακὸν ἀνθρώποισι φέρουσα, οἵ .... τέ μιν ἐξελάσωσι ............... καὶ οὐκ ἰθεῖαν ἔνειμαν. ‘And she [= Justice], wrapped in mist, follows to the city and haunts of the people, weeping, and bringing mischief to men, who have driven her forth in that they did not deal straightly with her’. hes. Op. 222–22422 What in 224 begins as an omnitemporal description (cf. τε and the use of the subjunctive in a relative clause) gains definition so that the types become like individuals (see West 1978 ad loc.), hence the subsequent use of the indicative ii aorist. The more literal translation may be: ‘who have driven her forth and have dealt not straightly with her’, where οὐκ belongs to ἰθεῖαν23 and the second verb ἔνειμαν is a specification of ἐξελάσωσι before. We may interpret ἔνειμαν as a gnomic aorist (see also (5) above), but in view of the coordination with the subjunctive aorist also as a normal indicative ii aorist anterior to the state of affairs of the main clause. Examples (9)–(11) are likewise ambiguous examples that allow for a double interpretation: (9) ἄλλος δ’ εἰαρινὸς πέλεται πλόος ἀνθρώποισιν· ἦμος δὴ τὸ πρῶτον, ὅσον ... τ’ ἐπιβᾶσα κορώνη ἴχνος ἐποίησεν, τόσσον πέταλ’ ἀνδρὶ φανήῃ ἐν κράδῃ ἀκροτάτῃ, τότε δ’ ἄμβατός ἐστι θάλασσα· ‘Another time for men to go sailing is in spring when a man first sees leaves on the topmost shoot of a fig-tree as large as the foot-print that a crow makes; then the sea is passable’. hes. Op. 678–681

22 23

Cf. Op. 289–290, 680; Theog. 7–8. See Chantraine (1963:354–356).

the gnomic aorist in hesiod

95

Just as in (8), τε (in ὅσον τ’) refers to permanent facts (Ruijgh 1971:89) and a generic description starts, but the description gains definition so that the type ‘crow’ (κορώνη) becomes like an individual crow, which explains the use of ἐποίησεν. This ἐποίησεν may for that reason be explained either as a gnomic aorist or as a normal one anterior to φανήῃ, as if a past picture is described.24 (10) Badness (κακότητα, 287) can be got easily τῆς δ’ ἀρετῆς ἱδρῶτα θεοὶ προπάροιθεν ἔθηκαν ἀθάνατοι· μακρὸς δὲ καὶ ὄρθιος οἶμος ἐς αὐτὴν ‘But between us and Goodness the immortal gods have placed the sweat of our brows: long and steep is the path that leads to her’. hes. Op. 289–290

This is an omnitemporal passage about κακοτής and ἀρετή. Here ἔθηκαν may be interpreted either as a gnomic aorist or as a normal one, referring to the gods’ activity in the past (probably after the deceit committed by Prometheus). At the same time, however, it is an omnitemporal observation holding for everyone at all times (since Prometheus’ deceit). Whereas in (8)–(10) two interpretations seem equally plausible, in (11) there seems to be a gradual shift of interpretation: (11) καί .... τε […] ἀκροτάτῳ Ἑλικῶνι χοροὺς ἐνεποιήσαντο, καλοὺς ἱμερόεντας, ἐπερρώσαντο δὲ ποσσίν. ἔνθεν ἀπορνύμεναι κεκαλυμμέναι ἠέρι πολλῷ ἐννύχιαι στεῖχον […] ‘and they [i.e., the Muses] make their fair, lovely dances upon highest Helicon and move with vigorous feet. Thence they arose and went abroad by night, veiled in thick mist’. hes. Theog. 7–8

For the interpretation of these verses I refer to Rijksbaron (2009:241–247), who convincingly shows that these indicative ii aorists are multi-interpretable. The interpetation changes following the course of the passage. Within the 24

See also Hom. Od. 4.356 τόσσον ἄνευθ’ ὅσσον τε νηῦς ἤνυσεν, where Chantraine (1963:185) speaks of a ‘cas typique’.

96

wakker

omnitemporal context (cf. καί τε) these aorists are first interpreted as gnomic ones. Since, however, in principle, an indicative ii aorist may also express a past fact, a gradual transition is made to the imperfect στεῖχον (9), which clearly refers to a past fact. The aorists are then reinterpreted as normal aorists referring to the past, i.e., to the time the poet met the Muses. Since in Greek the same verb form performs both functions, this double interpretability is not a problem. It is only in translating that one has to choose. Evelyn-White opt for a gnomic translation, while Mazon’s Budé makes another choice: ‘ont formé des choeurs […] ont voltigé leurs pas’. Finally, there are three examples where one may hesitate between an empiric and gnomic interpretation. (12) καὶ δόμεν ὅς κεν δῷ καὶ μὴ δόμεν ὅς κεν μὴ δῷ· δώτῃ μέν τις .... ἔδωκεν, ἀδώτῃ δ’ οὔ τις ἔδωκεν· δὼς ἀγαθή, ἅρπαξ δὲ κακή ‘Give to one who gives, but do not give to one who does not give: a man gives to the free-handed, but no one gives to the close-fisted. Give is a good girl, but Take is bad’. hes. Op. 354–356

Just before, by means of infinitives with imperative value, the following advice is given: ‘Do not get base gain: base gain is as bad as ruin. Be friends with the friendly, and visit him who visits you’. This last advice is motivated (note the asyndeton with explicative value) by the general observation in v. 355. Here τις indicates the general character, but it is not clear whether an empiric or a gnomic aorist is involved. A gnomic aorist usually refers to activities, but this does not hold for οὐ … ἔδωκεν. This rare use of οὐ may, however, be explained here by the opposition ἔδωκεν vs. οὐ … ἔδωκεν. An interpretation both as empiric and as gnomic aorist therefore seems possible. See Mazon’s Budé translation: ‘on donne à un donneur: à qui n’est pas donneur, nul jamais ne donna’, where both interpretations are combined in the translation. (13) [… || … ..................... ἄλλοτε δ’ αὖτε] ἢ τῶν γε στρατὸν εὐρὺν ἀπώλεσεν ἢ ὅ γε τεῖχος ἢ νέας ἐν πόντῳ Κρονίδης .................. ἀποτείνυται αὐτῶν. ‘And again, at another time, the son of Cronos either destroys their wide army, or their walls, or else makes an end of their ships on the sea’. hes. Op. 245–247

the gnomic aorist in hesiod

97

For the context of (13) see (4) above, which is perhaps directly followed by vv. 246–247, since vv. 243–245 are lacking in some manuscripts.25 In any case, vv. 246–247 describe further punishments inflicted by Zeus on wrongdoers. If vv. 243–245 do belong in the text, an interpretation as empiric aorist seems likely given ἄλλοτε δ’ αὖτε (245). However, in view of the omnitemporal character of the passage as a whole (239–247) and in view of the uncertain status of the text in 244–245, an interpretation as gnomic aorist is equally likely (comparable to ἐπήγαγε in 242). (14) ......... πολλὰ δ’ ἀεργὸς ἀνήρ, κενεὴν ἐπὶ ἐλπίδα μίμνων, χρηίζων βιότοιο, κακὰ προσελέξατο θυμῷ. ἐλπὶς δ’ οὐκ ἀγαθὴ κεχρημένον ἄνδρα ........... κομίζει ‘The idle man who waits on empty hope, lacking a livelihood, addresses his heart with many reproaches;26 it is not a wholesome hope that accompanies a need man’. hes. Op. 498–500

With πολλὰ (κακὰ) an interpretation as empiric aorist is possible, but also as a gnomic one (cf. the coordination with κομίζει), though the terminative character of this state of affairs is less obvious than with verbs like ἔδωκε, ἔθηκε, see note 16. All in all, then, if one uses the description of the gnomic aorist given in par. 3, the disputable examples in Hesiod seem either indisputable, see (7), ambiguous and allowing for a double interpretation, see (8)–(10) and (12)–(14), or allowing for a gradual shift of interpretation, see (11).

5.5

Concluding Remarks

By way of conclusion, I would like to argue that the reason for using the indicative ii aorist in omnitemporal sections seems to be its (perfective) aspectual value. Due to the non-existence of the optimal form to express both omnitemporality and perfectivity (indicative i aorist), the suboptimal form (indicative ii aorist) is used in Greek. We are dealing with a crucial, anterior, com25 26

As indicated in the critical apparatus of most text editions, it is uncertain whether these lines, including ἄλλοτε δ’ αὖτε, belong here. My translation. Evelyn-White (‘lays to heart mischief-making’) seems to take προσελέξατο as form of *προσλέχομαι rather than of προσλέγομαι.

98

wakker

pleted, or ingressive state of affairs. Its time reference is taken for granted. The choice for the indicative ii aorist is optional and depends on the choice of the speaker, depending on the question whether or not he wants to emphasize the aspect of the state of affairs concerned. Of course there are tendencies: some verbs in omnitemporal sections are preferably used in the gnomic aorist: ἔδωκε, ἔθηκε, ὤλεσε. In other cases the choice is free, see ὤπασε—ὀπάζει in (2). The aorist is never the first or only signal of the omnitemporality of the section. This observation helps in interpreting dubious or borderline cases. The existence of borderline cases is understandable if we take the so-called Prototype theory27 into account. According to this theory, linguistic elements all have their own semantic value. In actual usage some semantic values are more central, other more peripheral. Thus, as for the indicative ii aorist, one can argue that its central meaning is perfective aspect and past reference. More peripheral are the empiric aorist (perfective and repetition, hence not only past but also—by conversational implicature—non-past reference) and gnomic aorist (perfective, non past). It is quite logical that there are no sharp borders between these interpretations and that borderline cases exist. After all, Greek does not have to choose. It has one form to express these different nuances. It is only in translating and interpreting that we encounter problems. These conclusions are based on a study of Hesiod and apply to Hesiod (at least). On the basis of the Hesiodic results, I would hypothesize that most conclusions also hold for Homer and later examples of the gnomic aorist. Of course, this remains to be proven in other studies.

References Bakker, E.J. 1988. Linguistics and formulas in Homer: scalarity and the description of the particle per. Amsterdam. Bakker, E.J. 1999. ‘Pointing to the past: verbal augment and temporal deixis in Homer’. Euphrosyne: studies in ancient epic and its legacy in honor of Dimitris N. Maronitis, ed. by J. Kazazis & A. Rengakos, 50–65. Stuttgart. Bakker, E.J. 2001. ‘Similes, augment and the language of immediacy’. Orality and literacy in the Greek and Roman world, ed. by J. Watson, 1–23. Leiden. Bakker, E.J. 2005. Pointing at the past: from formula to performance in Homeric poetics. Cambridge, ma.

27

For a brief and clear description of Prototype theory, see for instance Bakker (1988:14–18).

the gnomic aorist in hesiod

99

Basset, L. 1989. ‘L’augment et la distinction discours/récit dans l’Iliade et l’Odyssée’. Études homériques, ed. by M. Casevitz, 9–16. Lyon. Bary, C. 2009. Aspect in Ancient Greek: a semantic analysis of the aorist and imperfective. Unpublished PhD thesis. Nijmegen. Chantraine, P. 1963. Grammaire homérique. ii: Syntaxe. Paris. Goodwin, W.W. 1889. Syntax of the moods and tenses of the Greek verb. London. Grice, H.P. 1975. ‘Logic and conversation’. Syntax & semantics, vol. iii: Speech acts, ed. by P. Cole & J. Morgan, 41–58. New York. Humbert, J. 1960. Syntaxe grecque. Paris. Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. lsj = Liddell, H.G., Scott, R. & H. Stuart Jones. 1996. A Greek-English lexicon. Oxford. Lyons, J. 1977. Semantics. Cambridge. McKay, K.L. 1988. ‘Aspectual usage in timeless contexts in Ancient Greek’. In the footsteps of Raphael Kühner, ed. by A. Rijksbaron et al., 193–208. Amsterdam. Mazon, P. 1986. Hésiode. Théogonie—Les travaux et les jours—Le bouclier. Paris. Monro, D.B. 1891. A grammar of the Homeric dialect. Oxford. Orriens, S.A.J. 2008. The augment and the imperfect in Homer. Unpublished paper. Cambridge. Pagniello iii, F.J. 2002. The Homeric augment. Unpublished PhD thesis. Athens, ga. Pagniello iii, F.J. 2007. ‘The past iterative and the augment in Homer’. if 112:105–123. Péristérakis, A.E. 1962. Essai sur l’aoriste intemporel en grec. Athens. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Chicago. Rijksbaron, A. 2009. ‘Discourse cohesion in the proem of Hesiod’s Theogony’. Discourse cohesion in Ancient Greek, ed. by S.J. Bakker & G.C. Wakker, 241–265. Leiden. Ruijgh, C.J. 1971. Autour de τε épique. Amsterdam. Shipp, G.P. 19722. Studies in the language of Homer. Cambridge. Stahl, J.M. 1907. Kritisch-historische Syntax des griechischen Verbums der klassischen Zeit. Heidelberg. Verdenius, W.J. 1985. A commentary on Hesiod: works and days, vv. 1–382. Leiden. West, M.L. 1966. Hesiod: Theogony. Oxford. West, M.L. 1978. Hesiod: Works and days. Oxford. Willi, A. 2007. ‘Of aspects, augments, aorists—or how to say to have killed a dragon’. Greek and Latin from an Indo-European perspective, ed. by C. George et al., 34–48. Cambridge. Willmott, J. 2007. The moods of Homeric Greek. Cambridge.

chapter 6

The Imperfect Unbound. A Cognitive Linguistic Approach to Greek Aspect Rutger J. Allan

6.1

Introduction

A widely-accepted view of the semantic contrast between the aorist and the imperfect in narrative is formulated by Rijksbaron as follows: Since the imperfect characterizes the state of affairs as ‘not-completed’ it creates a framework within which other states of affairs may occur, while the aorist indicative characterizes the state of affairs as ‘completed’, as a mere event. rijksbaron 2006:11

While this explanation of the function of the imperfect and the aorist in narrative works very well and therefore meets with broad approval, there are still many elusive instances of the imperfect which seem to evade an explanation of markers of non-completedness. Rijksbaron himself mentions one class of verbs which appears to behave differently: verbs of speech. The imperfect of verbs of speech, Rijksbaron observes, can be used in contexts where it is clear that the act of speaking is completed: contexts where at first glance one would expect an aorist. In those cases, the imperfect ‘serves to direct the attention to the consequences of the completion of the state of affairs. […] In general the use of the imperfect suggests that the state of affairs and what follows, i.e. the reaction of the person addressed, are closely connected: the message is delivered, the request is made, the command given, but what does the addressee do as a result?’ (Rijksbaron 2006:18–19). An example given by Rijksbaron (loc. cit.) is: (1) ταῦτα λέγων ὁ Μεγάβαζος εὐπετέως ἔπειθε Δαρεῖον […] μετὰ δὲ πέμψας ἄγγελον […] ‘With this speech Megabazus readily persuaded Darius […] presently he sent a messenger’. hdt. 5.24.1 © koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_007

the imperfect unbound

101

The adverb εὐπετέως shows that the imperfect should not be interpreted as conative: Megabazos successfully persuades Dareios. What the imperfect does, instead, is draw attention to the consequences of Dareios being persuaded.1 In this paper, I describe this type of imperfect as a marker of the continuing relevance of the event. The imperfect of continuing relevance is not limited to verbs of speech, but also occurs with other semantic classes such as verbs of motion. I will argue here that a number of puzzling imperfects of the motion verbs πλέω and χωρέω can be explained as marking continuing relevance. This is the topic of 6.2. A second objective of the paper is to show that aspectual semantics in general, and this special use of the imperfect in particular, can be insightfully described by means of a number of theoretical concepts derived from Cognitive Linguistics (e.g., Langacker 2008 & 2009; Croft 2012). These are construal, profile vs. base, and (temporal) scope of predication. These Cognitive Linguistic notions and the way they relate to the more generally accepted aspectual notions of telicity and boundedness are discussed in 6.3. Finally, in 6.4, I briefly turn to the verb πέμπω and use it as a test case of whether its notably varied use of the imperfect can also be elucidated in terms of the theoretical apparatus introduced in 6.3.

6.2

The Imperfect of Verbs of Motion: πλέω and χωρέω in Thucydides

Verbs of motion are an interesting subject for examination, not only because they occur with high frequency, but also because they show a remarkable flexibility in their use of tense and aspect. In this section, the use of the imperfect and the aorist indicative of two verbs of motion, πλέω and χωρέω, will be the focus of interest. As a starting point for our discussion, let us examine the following statistics relating to the distribution of aspect forms in Thucydides.2 The figures in Table 6.1 show an unexpected relationship between preverbs and aspect form. It has been noted in the literature on aspect that prepositions can influence the Aktionsart of the verb by providing a telos to the verb meaning.3 In the case of verbs of motion specifically, the telos may involve the 1 The use of the imperfect of verbs of speech to express a lasting effect (‘dauernde Wirkung’, ‘bleibende Gültigkeit’) had already been observed in the older literature, e.g., Blass (1889); Svensson (1930); and Schwyzer-Debrunner (1950:277–278). For important recent accounts of the imperfect of completed actions, see Buijs (2007) and Basset (2009). 2 I left out compound forms of these two verbs which occur less than five times in Thucydides. 3 See e.g., Fanning (1990:32, 151). The term telos will be a topic of 6.3.

102 table 6.1

πλέω ἀποπλέω ἐπιπλέω καταπλέω παραπλέω περιπλέω

allan πλέω and χωρέω: imperfect vs. aorist indicative (3sg/pl)

Imperfect

Aorist

40 9 16 3 14 10

18 30 2 11 11 5

χωρέω ἀναχωρέω ἀποχωρέω προσχωρέω προχωρέω ὑποχωρέω

Imperfect

Aorist

62 11 7 6 12 6

12 67 29 11 6 5

addition of a source or a destination. In other words, the verb is made telic (more specifically, an accomplishment) by means of the addition of a preverb.4 This well-known phenomenon cannot explain the distributional pattern in the tables above, however, since many compound verbs show a preference for the imperfect. Instead, there appears to be a split within the compounded verbs. Upon closer inspection, this split seems to be linked to the particular direction expressed by the preverb. Compound verbs meaning ‘sailing towards’ (ἐπιπλέω) and ‘advancing towards’ (προχωρέω)5 tend to be used in the imperfect, while compounds meaning ‘sailing away from’ (ἀποπλέω) and ‘withdrawing from’ (ἀναχωρέω and ἀποχωρέω) tend to be used in the aorist form. The simple verbs πλέω and χωρέω are in alignment with this pattern. They tend to be accompanied by a goal-expression and they show a preference for the imperfect form. How can these correlations be explained? I shall argue here that the use of the imperfect and the aorist form of these verbs is closely linked to the function of these verbs within the structure of the narrative. Three functions of the imperfect appear to be chiefly responsible for the distribution in the table: i. the imperfect creating a temporal framework, ii. the discourse-structuring imperfect, iii. the imperfect of continuing relevance.

4 In 6.3, I return to the issue of Aktionsarten and their relation to grammatical aspect. 5 The verb is also used with the meaning ‘succeed’, which accounts for many of the imperfect and all aorist occurrences.

the imperfect unbound

103

In order to demarcate the imperfect of continuing relevance clearly from the other two uses, I first discuss the imperfect which creates a temporal framework (6.2.1), and the discourse-structuring imperfect (6.2.2). Most attention, however, is devoted to the third discourse-function of the imperfect (6.2.3) since it is probably the least recognized in the literature on Greek aspect.6 6.2.1 The Imperfect Creating a Temporal Framework Undoubtedly the most frequent function of the imperfect in narrative is setting up a temporal framework within which other events may occur, as in the following example: (2) καὶ περιέπλεον τὸν Ἰχθῦν καλούμενον τὴν ἄκραν ἐς τὸν ἐν τῇ Φειᾷ λιμένα, οἱ δὲ Μεσσήνιοι ............. ἐν τούτῳ καὶ ἄλλοι τινὲς οἱ οὐ δυνάμενοι ἐπιβῆναι κατὰ γῆν χωρήσαντες τὴν Φειὰν αἱροῦσιν. ‘And they sailed around Point Ichthys into the port of Pheia. In the meantime the Messenians, and some others who could not get on board, marched over by land and took Pheia’.7 thuc. 2.25.4

In the following example, the discourse is structured somewhat differently: (3) καὶ οἱ μὲν ἄραντες τῇ παρασκευῇ ταύτῃ περιέπλεον, οἱ δὲ Πελοποννήσιοι […] ἀνεχώρησαν διὰ Βοιωτῶν ‘This armament weighed anchor and started on its cruise, and the Peloponnesians […] retired through Boeotia’. thuc. 2.23.3

After mentioning the start of the Athenians’ expedition around the Peloponnesus, the narrative focuses on the Peloponnesian army returning home from Attica. With the imperfect περιέπλεον, the narrator signals to the reader that 6 Other functions of the imperfect are, the conative imperfect and the imperfect used to describe a scene of simultaneous occurring events (Rijksbaron 2006:12–13; see also Bakker’s (1997) mimetic mode and Allan’s (2009:2013) descriptive mode). Less typical uses are the imperfect of likelihood (Rijksbaron 2006:17–18) and the immediative (also called inceptive) imperfect (Rijksbaron 2006:16–17). These uses will be addressed in 6.3. 7 Translations of Thucydides are Crawley’s (Strassler 1998), sometimes with modifications.

104

allan

this story-line is to be continued. And, indeed, in 2.25.1 we turn back to the Athenians who were meanwhile still sailing around the Peloponnesus (oἱ δ’ ἐν ταῖς ἑκατὸν ναυσὶ περὶ Πελοπόννησον Ἀθηναῖοι). In filmic terms, the imperfect is employed as a device enabling the narrator to ‘cross-cut’ from one scene to another which is taking place simultaneously in another space. Once we are following the Athenian fleet around the Peloponnesus, the stages of the continuing operation are again marked by imperfects παρέπλεον (2.25.3) and περιέπλεον (2.25.4).8 The Aktionsart of πλέω construed with a goal expression is an accomplishment since it designates a gradual development towards a telos. The state of affairs is presented as ongoing (telos not yet reached) so that it can function as a temporal framework for other events.9 Such examples of the imperfect in my view demonstrate that the imperfect is not necessarily used to refer to background events, in the sense that the events at issue are less significant to the story-line. The fact that the Corinthians set out to sail for Epidamnus is a major turning point in the course of events and is therefore a part of the backbone of the narration. Likewise, the Athenians’ expedition around the Peloponnesus is a significant operation in the development of the war. What the imperfect form indicates is not the subsidiary discourse status, but that the event is construed as unbounded, that is, as continuing while other events are taking place. By way of contrast, consider the following aorists (the Athenians became masters of Sestos held by the Persians): (4) a. καὶ μετὰ τοῦτο ἀπέπλευσαν ἐξ Ἑλλησπόντου ὡς ἕκαστοι κατὰ πόλεις. ‘And after this they sailed away from the Hellespont to their respective cities’. thuc. 1.89.2

b. καὶ οἱ περὶ Πελοπόννησον Ἀθηναῖοι ἐν ταῖς ἑκατὸν ναυσίν […] ὡς ᾔσθοντο τοὺς ἐκ τῆς πόλεως πανστρατιᾷ ἐν Μεγάροις ὄντας, ἔπλευσαν παρ’ αὐτοὺς καὶ ξυνεμείχθησαν. στρατόπεδόν τε μέγιστον δὴ τοῦτο ἁθρόον Ἀθηναίων ἐγένετο

8 Other examples are: 1.46.1; 2.33.2 (ἀπέπλεον). 9 This is represented in Fig. 6.9 in 6.3 below. Issues of Aktionsart and telicity are discussed in more detail in section 6.3.

the imperfect unbound

105

‘The Athenians in the hundred ships around the Peloponnese […], hearing that the citizens at home were in full force at Megara, now sailed over and joined them. This was without doubt the largest army of Athenians ever assembled’. thuc. 2.31.1

In 1.89.2, the Athenians and their allies sail back to their cities after the siege of Sestos. The return to their cities is only briefly mentioned by the narrator. There are no more intervening incidents. In 2.31.1, the aorist ἔπλευσαν marks the final return home of the ships around the Peloponnesus. This concluding event of the narrative of the operation is followed by a closing narratorial intervention about the unprecedented strength of the Athenian army. The contrast between the aforementioned examples makes it clear why the imperfect tends to be used to refer to the journey towards a certain destination. This type of imperfect serves to prepare the way for subsequent events through which the tension of the conflict gradually builds up. On the other hand, in the aftermath of the narrative episode, the aorist forms tend to appear. In the closure of an episode, the protagonist’s arrival at the final destination (typically a return to the original location) is only briefly mentioned, and without potential interference of further complicating events. In Rijksbaron’s words (2006:11): ‘The imperfect creates an expectation on the part of the reader/hearer: what else happened?; the aorist on the other hand, does not have this effect: the state of affairs has simply occurred’. 6.2.2 The Imperfect as a Discourse-Structuring Device There are cases in which the imperfect cannot be interpreted as a temporal framework to subsequent events in the narrative. For example: (5) αὐτὸς δὲ ἀπολεξάμενος ἐκ πάντων ἑξήκοντα ὁπλίτας καὶ τοξότας ὀλίγους ἐχώρει ἔξω τοῦ τείχους ἐπὶ τὴν θάλασσαν ᾗ μάλιστα ἐκείνους προσεδέχετο πειράσειν ἀποβαίνειν, ἐς χωρία μὲν χαλεπὰ καὶ πετρώδη πρὸς τὸ πέλαγος τετραμμένα, σφίσι δὲ τοῦ τείχους ταύτῃ ἀσθενεστάτου ὄντος ἐσβιάσασθαι αὐτοὺς ἡγεῖτο προθυμήσεσθαι· οὔτε γὰρ αὐτοὶ ἐλπίζοντές ποτε ναυσὶ κρατήσεσθαι οὐκ ἰσχυρὸν ἐτείχιζον, ἐκείνοις τε βιαζομένοις τὴν ἀπόβασιν ἁλώσιμον τὸ χωρίον γίγνεσθαι. κατὰ τοῦτο οὖν ..... πρὸς αὐτὴν τὴν θάλασσαν χωρήσας ἔταξε τοὺς ὁπλίτας. ‘He himself picked sixty heavy infantry and a few archers from his whole force, and with these went outside the wall down to the sea, where he thought that the enemy would most likely attempt to land. Although the

106

allan

ground was difficult and rocky, looking towards the open sea, the fact that this was the weakest part of the wall would, he thought, encourage their ardor, as the Athenians, confident in their naval superiority, had here paid little attention to their defences, and the enemy if he could force a landing might feel secure of taking the place. At this point, accordingly, going down to the water’s edge, he posted his heavy infantry’. thuc. 4.9.2–4

The sentences following ἐχώρει do not advance narrated time. Instead, we are informed about the motives of Demosthenes for going to that particular place down by the sea. We are told that it was Demosthenes’ expectation that the Spartans would attack there. Demosthenes’ ideas about the locations did not just occur to him while he was descending to the sea: instead, they are Demosthenes’ reasons for descending to the sea. This means that the imperfect ἐχώρει does not signal a ‘real world’ temporal relationship with the following imperfect ἡγεῖτο. What the imperfect ἐχώρει does instead, is to organize the structure of the discourse. The imperfect is used by the narrator to suspend narrated time in order to dwell briefly on Demosthenes’ thoughts. The imperfect gives a signal to the reader that the story-line will be resumed after the digression about the position of the Spartans. To put it in more technical terms, the imperfect signals that the discourse will be resumed after the embedded discourse segment (push).10 In (5), the resumption of the main story-line (pop) is marked explicitly by the particle οὖν (κατὰ τοῦτο οὖν). Another example of this use of the imperfect is the following (the Athenians have landed on Sphacteria): (6) καὶ ἐχώρουν δρόμῳ ἐπὶ τὸ πρῶτον φυλακτήριον τῆς νήσου. ὧδε γὰρ διετετάχατο· […] οὕτω μὲν τεταγμένοι ἦσαν. oἱ δὲ Ἀθηναῖοι τοὺς μὲν πρώτους φύλακας, οἷς ἐπέδραμον, εὐθὺς διαφθείρουσιν ‘And they advanced with a run against the first post in the island. The enemy had distributed his force as follows: […] Such was their position. The advanced post thus attacked by the Athenians was at once put to the sword’. thuc. 4.31.1–32.1 10

This special discourse-cohesive function of the imperfect is also treated by Rijksbaron (1988). For the terms push and pop applied to discourse-structural particles in Greek, see Slings (1997).

the imperfect unbound

107

After landing on Sphacteria, the Athenians run up against the first post of the Spartans. The imperfect ἐχώρουν is, again, not used to create a temporal framework in which subsequent events take place. Instead, it suspends narrated time in order to insert a narratorial digression on the division of the Spartan troops. The shifts into (push) and out off (pop) the digression are marked with a number of lexical and grammatical means. The particle γάρ introduces the inserted digression, while the return to the main story-line is marked by the combination μέν … δέ. The opening and closing of the digression is also signalled lexically by the forward- and backward-referring clauses ὧδε γὰρ διετετάχατο and οὕτω μὲν τεταγμένοι ἦσαν. There is, furthermore, a ring composition encompassing the digression: (a) ἐχώρουν δρόμῳ ἐπὶ τὸ πρῶτον φυλακτήριον τῆς νήσου → (b) οἱ δὲ Ἀθηναῖοι τοὺς μὲν πρώτους φύλακας, οἷς ἐπέδραμον […]. The ring is opened by an imperfect ἐχώρουν δρόμῳ and closed off with an aorist ἐπέδραμον.11 6.2.3 The Imperfect as a Marker of Continuing Relevance The third use of the imperfect explaining the distribution as shown in Table 6.1 involves imperfects which, unlike the two previous types, refer to completed states of affairs. Let us first examine a number of examples: (7) a. καὶ παραλαβόντες τῶν Ἀργείων ὁπλίτας ἐπὶ τὰς ναῦς ἔπλεον ἐς τὴν Λακωνικήν· καὶ πρῶτον μὲν τῆς Ἐπιδαύρου τι τῆς Λιμηρᾶς ἐδῄωσαν, ἔπειτα σχόντες ἐς τὰ καταντικρὺ Κυθήρων τῆς Λακωνικῆς. ‘Taking on board the heavy infantry from Argos [the Athenians] sailed to Laconia, and after first plundering part of Epidarus Limera, landed on the coast of Laconia, opposite Cythera’. thuc. 7.26.1–2

b. οἱ δ Ἀθηναῖοι […] ὑπὸ νύκτα ἔπλεον ἐπὶ τὰς Συρακούσας. καὶ οἵ τε Ἀθηναῖοι ἅμα ἕῳ ἐξέβαινον ἐς τὸ κατὰ τὸ Ὀλυμπιεῖον ὡς τὸ στρατόπεδον καταληψόμενοι, καὶ οἱ ἱππῆς οἱ Συρακοσίων πρῶτοι προσελάσαντες ἐς τὴν Κατάνην καὶ αἰσθόμενοι ὅτι τὸ στράτευμα ἅπαν ἀνῆκται, ἀποστρέψαντες ἀγγέλλουσι τοῖς πεζοῖς. ‘Meanwhile the Athenians […] sailed by night to Syracuse. Thus, when morning broke the Athenians were landing opposite the Olympieum

11

Another example of this use of the imperfect is 8.17.1 (ἔπλεον; cf. subsequent γάρ [push] + οὖν [pop] in 8.17.2 and 3) and 8.98.1 (ἐχώρει, cf. οὖν [pop] in 8.98.3).

108

allan

ready to seize their camping ground, and the Syracusan horse having ridden up first to Catana and found that all the armament had put to sea, turned back and told the infantry’. thuc. 6.65.2–3

c. καὶ καταστήσαντες ἐν τῷ Λαβδάλῳ φυλακὴν ἐχώρουν πρὸς τὴν Συκῆν οἱ Ἀθηναῖοι, ἵναπερ καθεζόμενοι ἐτείχισαν τὸν κύκλον διὰ τάχους. καὶ ἔκπληξιν τοῖς Συρακοσίοις παρέσχον τῷ τάχει τῆς οἰκοδομίας· ‘After posting a garrison in Labdalum, they advanced to Syca, where they sat down and quickly built the Circle or centre of their wall of circumvallation. The Syracusans were appalled at the rapidity with which the work advanced’. thuc. 6.98.2

d. καὶ δόξαν αὐτοῖς εὐθὺς ἐχώρουν ἐς τὸ ἄστυ καὶ ἔθεντο αὖ ἐν τῷ Ἀνακείῳ τὰ ὅπλα. ‘And they decided to march into the city, and setting forth accordingly halted in the Anaceum’. thuc. 8.93.1

In each of these examples, the context shows that the imperfect is not used to create a temporal framework within which the ensuing actions occur. The context shows that the spatial goal denoted by the verb of motion has been reached. In other words, the events are accomplishments of which the telos has been attained before the subsequent event takes place. In (7a), the imperfect does not indicate that the Athenians were still sailing towards Laconia: in order to be able to plunder Eupidaurus Limera (which is in Laconia), they must have reached Laconia first. In (7b), the Athenians arrive in Syracusan territory and land at the Olympieum. Note that the following imperfect ἐξέβαινον also refers to a completed event. There is no mention of events taking place while the Athenians were sailing to Syracuse and landing. Thus, the two imperfects appear to refer to subsequent events. In (7c), the Athenians go to Syca and build a Circle fort. In (7d), hoplites go to the city and lay their weapons in the Anaceum (the temple of the Dioscuri on the slope of the Acropolis). The imperfect referring to a completed state of affairs is a marked use of the imperfect form.12 What the imperfect signals in these cases is that the event 12

The issue of markedness is addressed in 6.3 below.

the imperfect unbound

109

expressed by the imperfect should be viewed not as a self-contained event, but as preparatory step towards some larger aim pursued by the protagonist. The imperfect indicates that the event at issue should be construed as a prelude to, and form a continuum with, the following event; in other words, this type of imperfect has a cataphoric (forward-cohesive) pragmatic function: it signals that the event, even though physically completed, is of continuing relevance to the following discourse. The thematic continuity is demonstrated by the fact that in each of the examples in (7) and (8), the grammatical subject of the following clause is the same. In other words, there is Topic continuity. How this works in a larger stretch of narrative, in which the aorist also occurs, can be seen in the following example (Alcibiades intends to sail by a number of coastal cities to offer an alliance): (8) μετὰ δὲ τοῦτο Ἀλκιβιάδης τῇ αὑτοῦ νηὶ διαπλεύσας ἐς Μεσσήνην καὶ λόγους ποιησάμενος περὶ ξυμμαχίας πρὸς αὐτούς, ὡς οὐκ ἔπειθεν, ἀλλ’ ἀπεκρίναντο πόλει μὲν ἂν οὐ δέξασθαι, ἀγορὰν δ’ ἔξω παρέξειν, ἀπέπλει ἐς τὸ Ῥήγιον. καὶ εὐθὺς ξυμπληρώσαντες ἑξήκοντα ναῦς ἐκ πασῶν οἱ στρατηγοὶ καὶ τὰ ἐπιτήδεια λαβόντες παρέπλεον ἐς Νάξον, τὴν ἄλλην στρατιὰν ἐν Ῥηγίῳ καταλιπόντες καὶ ἕνα σφῶν αὐτῶν. Ναξίων δὲ δεξαμένων τῇ πόλει παρέπλεον ἐς Κατάνην. καὶ ὡς αὐτοὺς οἱ Καταναῖοι οὐκ ἐδέχοντο (ἐνῆσαν γὰρ αὐτόθι ἄνδρες τὰ Συρακοσίων βουλόμενοι), ἐκομίσθησαν ἐπὶ τὸν Τηρίαν ποταμόν, καὶ αὐλισάμενοι τῇ ὑστεραίᾳ ἐπὶ Συρακούσας ἔπλεον ἐπὶ κέρως ἔχοντες τὰς ἄλλας ναῦς· δέκα δὲ τῶν νεῶν προύπεμψαν ἐς τὸν μέγαν λιμένα πλεῦσαί τε καὶ κατασκέψασθαι εἴ τι ναυτικόν ἐστι καθειλκυσμένον, καὶ κηρῦξαι ἀπὸ τῶν νεῶν προσπλεύσαντας ὅτι Ἀθηναῖοι ἥκουσι Λεοντίνους ἐς τὴν ἑαυτῶν κατοικιοῦντες κατὰ ξυμμαχίαν καὶ ξυγγένειαν· τοὺς οὖν ὄντας ἐν Συρακούσαις Λεοντίνων ὡς παρὰ φίλους καὶ εὐεργέτας Ἀθηναίους ἀδεῶς ἀπιέναι. ἐπεὶ δ’ ἐκηρύχθη καὶ κατεσκέψαντο τήν τε πόλιν καὶ τοὺς λιμένας καὶ τὰ περὶ τὴν χώραν ἐξ ἧς αὐτοῖς ὁρμωμένοις πολεμητέα ἦν, ἀπέπλευσαν πάλιν ἐς Κατάνην. ‘After this Alcibiades sailed in his own vessel across to Messina with proposals of alliance, but met with no success, the inhabitants answering that they could not receive him within their walls, though they would provide him with a market outside. Upon this he sailed back to Rhegium. Immediately upon his return the generals manned and victualled sixty ships out of the whole fleet and coasted along to Naxos, leaving the rest of the armament behind them at Rhegium with one of their number. Received by the Naxians, they then coasted on to Catana, and being refused admittance by the inhabitants, there being a Syracusan party in the town, went on to the river Terias. Here they bivouacked, and the next

110

allan

day sailed in single file to Syracuse with all their ships except ten which they sent on in front to sail into the great harbour and see if there was any fleet launched, and to proclaim by herald from shipboard that the Athenians were come to restore the Leontines to their country, as being their allies and kinsmen, and that such of them, therefore, as were in Syracuse should leave it without fear and join their friends and benefactors the Athenians. After making this proclamation and reconnoitring the city and the harbours, and the features of the country which they would have to make their base of operations in the war, they sailed back to Catana’. thuc. 6.50

Alcibiades’ plan is to make a tour of a number of coastal cities in order to find out which of them are inclined to become allies of the Athenians. Alcibiades starts off from Rhegium, proceeds to Messina, returns to Rhegium in order to increase his fleet substantially, then sails to Naxos, Catania and Syracuse. The expedition is concluded with a return to Catania. The main stages of the expedition are marked by imperfects of the verb πλέω (in bold): ἀπέπλει ἐς τὸ Ῥήγιον,13 παρέπλεον ἐς Νάξον, παρέπλεον ἐς Κατάνην, ἐπὶ Συρακούσας ἔπλεον. Only the movement closing off the series is referred to by an aorist (in italics): ἀπέπλευσαν πάλιν ἐς Κατάνην.14 The imperfects are not used to indicate that the sailing is still ongoing when the subsequent events take place. It is clear from the context that the spatial goal is attained before any other events take place. The imperfect, therefore, is not used to create a temporal framework within which other events occur. What the imperfects do, instead, is to mark that the completion of each of the stages is somehow relevant to the following discourse: they constitute an element of a continuous series of movements connected with one another as part of some larger aim. The imperfects, in other words, expressly signal to the reader that the sequence of actions is not completed yet—there is more to come. The aorist ἀπέπλευσαν is then used to close off the tour around the coastal cities. 13 14

Note that an imperfect ἀπέπλει is used here to signal the continuation of the sequence, in spite of the general preference of ἀποπλέω for the aorist. The visit to Messina is not reported in a main clause, but in a number of participial and finite subordinate clauses, thus marking it as of secondary importance (a false start) within the expedition. The use of aspect in participial and temporal subordinate clauses is not relevant here since this is determined by the relative time (anterior or simultaneous) with respect to the main clause (see Rijksbaron 2006:76, 117).

the imperfect unbound

111

The chain of imperfects in the main clauses is interrupted once by an aorist ἐκομίσθησαν. The aorist presents the event in a neutral way: it is viewed as completed and self-contained. Contrary to the imperfects of πλέω, the fact that the Athenians go to the river Terias after their visit to Catania is not particularly relevant to the overall cohesion of the larger discourse, but only to the particular stage of their voyage. There is another aorist in the passage which follows the chain of imperfects: προύπεμψαν. It has been observed that verbs of sending (e.g., πέμπω, στέλλω) also display a special use of the imperfective aspect, a use which is very similar to the continuing relevance imperfect of πλέω and χωρέω. This marked use of ἔπεμπον is characterized by Steup (ad Thuc. 1.26.1) as: ‘Das Imperfektum ἔπεμπον von der Handlung selbst und ihren fortwirkenden Folgen’. According to Steup, the imperfect of πέμπω can be employed to designate the action itself and its continuing consequences. The narrator uses the imperfect in a marked way to signal that the action of sending remains relevant to the following discourse because the reaction of the person or persons to whom someone (or a message) is sent will subsequently be attended to in the narrative. The aorist προύπεμψαν in 6.50, however, presents the act of sending in an unmarked way: there is no forward-reference to a possible reaction on the part of the Leontines or Syracusans. The Athenians simply convey their message and leave without expecting a response. I will return to the aspectual characteristics of πέμπω below (6.4).

6.3

A Cognitive Linguistic Approach to Greek Aspect

6.3.1 Aspect and Aktionsart In the unmarked case, completed events, that is, events which have reached their telos, are referred to by an aorist verb which construes the state of affairs as bounded. The imperfect of continuing relevance is a deviation from this canonical pattern: a completed event is conceptualized as unbounded. In this section, I will argue that a number of theoretical concepts which have developed within the field of Cognitive Linguistics may be helpful in describing the semantics of aspect in Greek, and more specifically, the semantics of the imperfect of continuing relevance. Cognitive Linguistic concepts that are relevant to aspectual semantics are construal, temporal scope, profile and base. But before I turn to these issues, it is necessary first to address another fundamental issue of aspectual semantics: the relationships between Aktionsart and grammatical aspect and between telicity and boundedness.

112

allan

Aktionsart is often described as lexical aspect in contrast to grammatical aspect, which is aspect in the narrow sense. However, the term lexical aspect, which is used to define Aktionsart, is misleading in that it suggests that the Aktionsart of a clause is only determined by the lexical semantics of the verb stem. It is clear, however, that other elements of the clause are also relevant to its Aktionsart, especially the clausal subject and object, and adverbs.15 Furthermore, contextual factors and world knowledge also play a role in determining the Aktionsart of the clause. A well-known classification of Aktionsarten is the one which has been proposed by Vendler (1957). Vendler distinguishes four classes of verbs: (i) states, e.g., have, desire, want, rule, know; (ii) activities, e.g., walk, swim, push, pull; (iii) accomplishments, e.g., build a house, write a novel, grow up, get ready; and (iv) achievements, e.g., recognize, lose/find object, reach summit, start/stop, be born, die. There are a number of semantic features that crucially distinguish these four classes from one another.16 Probably the most relevant to aspect is the distinction pertaining to telicity. Clauses referring to a telic state of affairs crucially involve a change of state. Telic states of affairs have an initial state and a natural endpoint (a telos), leading to a resultant state. Accomplishments and achievements have a natural endpoint and are therefore telic. States and activities can also have endpoints. These endpoints are not natural, however, that is, they are not inherent in the semantics of the predicate, but are pragmatically implicated by our knowledge of the world. Even though it is not indicated explicitly by linguistic means, we know that loving, sleeping and swimming do not continue indefinitely. Furthermore, states and activities may come to an end, but they do not lead to a resultant state. There is, in other words, no reference to a change of state. The state following a state or activity is typically presented as identical to the state preceding the activity. Initial boundaries of atelic states of affairs are arbitrary rather than natural. Telic states of affairs can be further divided with regard to durativity. Accomplishments are durative since they consist of successive stages leading up to the final point, whereas achievements are non-durative in that they only designate a change in state of which the initial and final boundary coincide. The final semantic feature relevant to the classification of Aktionsarten is the feature

15 16

Take, for example, the distinction between to write (atelic) and to write a letter (telic). See also Verkuyl (1972); for Ancient Greek see Fanning (1990:126–127) and Napoli (2006). The following discussion of the Aktionsarten is based on Fanning (1990); Boogaart (1999); and Napoli (2006).

113

the imperfect unbound table 6.2

Aktionsarten: semantic features

Telic Durative Dynamic State Activity Accomplishment Achievement

– – + +

+ + + –

– + + +

of dynamicity. Unlike states, activities involve internal change (dynamicity). Subintervals of the interval in which an activity takes place cannot count as instances of the whole activity. Lifting a foot from the floor does not count as walking. Activities are internally heterogeneous. States, on the other hand, are homogenous. Even the smallest subintervals can count as instances of the state as a whole. The three distinctive features, telicity, durativity and dynamicity and their relation to the four Aktionsarten, are shown in Table 6.2.17 The semantics of the four Aktionsarten can be depicted graphically in the following way. An obvious way to represent a state is by means of a straight line (see Smith 1991):

figure 6.1 State (− telic, + durative, − dynamic)

Examples of Greek verbs expressing states are: βούλομαι, δύναμαι, εἰμί, ἐλπίζω, ἔχω, βασιλεύω, κεῖμαι, νομίζω (see also Fanning 1990:135–136). The internal dynamicity of activities can be indicated by means of a wavy line:

figure 6.2 Activity (− telic, + durative, + dynamic)

Examples of verbs denoting activities are: ἄγω, ἀγωνίζομαι, αὐξάνω, γελάω, γράφω, δακρύω, διώκω, ἔρχομαι, ζητέω, πράττω, τρέχω, φέρω (see Fanning 1990: 144–145). It has to be stressed, however, that many of these verbs can be turned

17

This table is taken from Boogaart (1999:89); see also Napoli (2006:34).

114

allan

into accomplishments by adding a reference to a telos, for example by adding a goal (in the case of verbs of motion) or a countable effected object (in the case of verbs of making).18 Accomplishments are dynamic and durative States of Affairs in which an initial state gradually develops towards the final boundary (telos) at which the resultant state commences.

figure 6.3 Accomplishment (+ telic, + durative, + dynamic)

Examples are: ἄγω (+ source/goal), γράφω (+ countable effected object), θάπτω, ποιέω (+ countable effected object), πείθω, πληρόω, φέρω (+ source/goal).

figure 6.4 Achievement (+ telic, − durative, + dynamic)

Achievements relate to punctual events that bring about a transition from an initial to a resultant state. There is a subclass of achievements which have a preparatory phase (a ‘preface’) preceding the event. These are sometimes called climaxes (see e.g., Fanning 1990, who refers to the un-prefaced achievements as punctuals). Climaxes are momentaneous events taking place as the culmination of a separate preparatory phase. For example, arriving implies a preceding phase of going towards, approaching. The event of finding is often preceded by a searching process. Climaxes can be depicted thus:

18

See Napoli (2006).

the imperfect unbound

115

figure 6.5 Climax

Examples of achievements given by Fanning (1990:156–157) are: ἀποθνῄσκω, ἀπόλλυμι, ἀφικνέομαι, δίδωμι, εὑρίσκω, λαμβάνω (climaxes); βάλλω, πίπτω (punctuals). 6.3.2 Aspect: Boundedness and Construal Against the background of the Aktionsarten as presented above, we can now state the contribution of the aspects (present vs. aorist) more clearly. A common way of explaining the function of aspectual distinctions is by invoking the idea of viewpoint. Aspect is a means to code a particular way of viewing or conceptualizing a state of affairs. This brings us to a fundamental notion in Cognitive Linguistics: construal. A central tenet in Cognitive Linguistic theorizing is that semantics is not so much concerned with the relationship between language and a real or possible world, but more with the way in which a subject conceptualizes (construes) the world. According to Cognitive Linguistics, meaning is not only a matter of objective properties of some referent in the world, it is very often substantially dependent on the—subjective—way some conceptualizer ‘views’ a situation in the world. Situations can, of course, be conceptualized in different ways. A classic example is that it is possible to describe the same glass as the glass is half-full or the glass is half-empty, depending on the particular conception of the situation one wishes to highlight.19 That a situation can be conceptualized in alternative ways is referred to by the term construal. Langacker distinguishes four broad classes of construal phenomena. These are: (i) specificity (one may describe a situation in more or less detail); (ii) focusing/prominence (some aspect of a situation can be foregrounded (‘profiled’) while another is backgrounded or ignored altogether); (iii) perspective (a situation can be viewed from different vantage points); and (iv) dynamicity (the temporal order in which components of a conceived situation are pro-

19

This example is discussed by Langacker (2008:43).

116

allan

cessed may vary) (Langacker 2007:435–438; 2008:3–89). As I will argue later in this section, two construal types are crucial to aspectual semantics: focusing/prominence and perspective. Now let us start with the notion of ‘viewpoint’. What exactly is meant by the term ‘viewpoint’, appearing so often in discussions of aspect?20 A set of theoretical notions developed in Cognitive Grammar may enable us to pinpoint the idea of viewpoint. The relationship between aspect and viewing is analyzed in Cognitive Grammar by means of the notions of temporal scope, profile and base. Scope is defined by Langacker as ‘those aspects of a scene that are specifically included in a particular predication’ (Langacker 1987:493). The scope is that portion of the world that is mentally encompassed at a certain moment. It is the ‘viewing frame’ or ‘onstage region’ delimiting the world around us. A scope is always limited because there is only so much that we can direct our attention to at a certain moment. Our visual field, for example, contains only a limited portion of our spatial surroundings. Every linguistic expression, be it lexical or grammatical, has a scope (called scope of predication by Langacker), i.e., that portion of the scene it specifically includes in its meaning. For example, within the immediate scope of predication of the word elbow, the concept of an arm is included because it is essential to understanding the meaning of the word elbow (Langacker 1987:119). So we can say that the conception of an arm is within the scope of predication of the word elbow. Similarly, the meaning of the word hypotenuse cannot be captured without reference to a complete right-angled triangle (Langacker 1991:64). The scope of the word uncle not only encompasses the concept of the person designated by uncle, but also encompasses a network of kinship relations including, for example, the ego and his or her parents (Langacker 1987:184–187). To sum up, the immediate scope of a lexical concept is the ‘on stage’ region to which viewing attention is directed and which is directly relevant for the understanding of its semantic value. Within the scope of predication, Langacker makes a distinction between the profile and the base. The base is a body of conceptual content that serves as a basis for the meaning of an expression. The profile of an expression stands out as the specific focus of attention within the scope. The profile can be

20

An example is Fanning’s definition of the present stem: ‘The present reflects an internal viewpoint concerning the occurrence which focuses on its development or progress and sees the occurrence in regard to its internal make-up, without beginning or end in view’. (Fanning 1990:103). See also Smith’s notion of viewpoint aspect vs. situation aspect (Smith 1997).

the imperfect unbound

117

characterized as what the expression designates. To illustrate what is meant by the notions of scope, base and profile, consider the following figures (from Langacker 2008:64).

figure 6.6 Elbow and hand: scope, base and profile

What is relevant for a complete understanding of what the expression elbow conveys is the combination of a base (in this case the concept of arm) and its profile elbow (indicated by the thick line segment) which stands out as the designated substructure within the base. In the case of the lexical item hypotenuse, for example, the profile (i.e., what is designated by the expression) is the longest side in a right-angled triangle. The figure also shows a distinction between the immediate scope, which comprises the conceptual content that is directly relevant to the expression, and the maximal scope. The maximal scope is the full array of conceptual content evoked by the expression. The lexical item elbow, for example, evokes a conception of the entire human body as background knowledge. Thus the immediate scope (foreground) is embedded within the maximal scope (background). A comparison between the two pictures also makes clear what the crucial difference between the semantic value of elbow and hand is. The distinction between the two concepts is a matter of profile selection. The two concepts invoke exactly the same base (the conception of arm and body), but crucially differ in the specific profile singled out within the base.21 In the dimension of time, we are dealing with a temporal scope, which can be defined as the relevant span of time which is selected (‘put on stage’) for focused viewing: a temporal window of attention. What I would like to propose here is that the semantics of Greek aspect revolve around the relationship between the boundaries of the event and the temporal scope. Fundamental to

21

For a more extended discussion of scope, base and profile, I refer to Langacker (2008:62– 70).

118

allan

Greek aspect is construal: grammatical aspect is a device to focus on (‘profile’) a particular portion (‘slice of time’) of the event while backgrounding others. The semantic content of the present and aorist stems in Greek can be defined as follows. – Present (imperfective) stem: unbounded within temporal scope The present stem profiles a portion of the event which lacks initial and final boundaries, i.e., the boundaries are outside the immediate temporal scope. – Aorist (perfective) stem: bounded within temporal scope The aorist stem profiles one or both boundaries of the event, i.e., one or both boundaries are included within the immediate temporal scope. The abstract conceptual content invoked by the two aspect forms can be represented graphically in the following figures:

figure 6.7 Present vs. aorist: (un)bounded within temporal scope

The straight horizontal lines are meant as abstract representations of the events. They do not indicate that the event is necessarily static (as in Fig. 6.1). The temporal scope (‘viewing frame’) is represented by the boxes.22 The arrow labeled ‘t’ stands for time. The portion of the total event that is profiled is indicated by a thick line. Dashed parts of a line indicate that the boundaries of the situation are not specified. They are outside the scope of view. The event may extend indefinitely (see Langacker 1987:258–262; 1991:88; 2000:224). As for the aorist, only the variant with two boundaries included within the temporal scope is represented. The aorist, however, may also profile only one boundary; see the ingressive aorist below (Fig. 6.11). These abstract schematic meanings are the invariable semantic core of the present and aorist stems. In this view, the various discourse functions of aspect mentioned in the literature (such as the marking of foreground vs. background

22

The notion of temporal scope and its application to aspectual semantics show similarities with the notion of Topic Time in formal semantics (Klein 1994; Bary 2009).

the imperfect unbound

119

in the discourse23 or of temporal ordering/relative tense24) should be seen as pragmatic side-effects rather than inherent to the semantics of the aspects. A definition of aspectual meaning in terms of boundedness is preferable to definitions based on the term completeness or completedness.25 Characterizations of the semantics of the aorist in terms of complete(d)ness are especially valid for telic verbs, verbs that have an inherent goal (telos) that has to be reached in order to be able to say that the event is complete(d). For atelic verbs complete(d)ness is a less fitting term. The aorist of atelic verbs typically has either an ingressive or a complexive meaning. In these cases, the aorist does not convey a sense of completion or completeness, but merely indicates that the state or activity has either started or come to an arbitrary end (see also Fanning 1990:91). In the same vein, it makes no sense to characterize ongoing states or activities (e.g. ἐβασίλευε, ἐγέλα) as not yet complete(d) since these types of events do not possess any natural point of completion. I would like to conclude this section by noting that the prototypical meaning of the perfect stem can also be described elegantly by means of this theoretical apparatus. The semantic import of the perfect stem might be captured thus: the perfect stem profiles a portion of the resultant state of a previous (telic) event. The previous event is not put ‘on stage’ (immediate scope) but is nevertheless still part of the conceptual background (maximal scope) of the meaning of the perfect.

figure 6.8 Perfect stem: resultant state profiled

23 24 25

Sicking is the best-known proponent of a pragmatic approach to Greek aspect (e.g., Sicking 1991; 1996). Ruijgh (1971; 1985; 1991) is perhaps the most prominent advocate of the relative tense approach. For boundedness as a defining property of Ancient Greek aspect, see also Napoli (2006) and Bary (2009).

120

allan

6.3.3

Marked and Unmarked Relationships between Aktionsart and Aspect Form For a better understanding of the interplay between grammatical aspect and Aktionsart, the notion of markedness relation is relevant. There has been some debate about the question of markedness with regard to aspect. Perhaps the best-known scholar who has applied the structuralist concept of markedness to the Greek aspect system is Ruipérez (1954). He treats the Greek present and aorist as a privative opposition, that is, as an opposition between one marked and one unmarked member. In his view, the present is marked for duration, whereas the aorist is not. Other scholars have argued that the aorist is the marked member.26 As has become clear from the way I see the abstract semantic core of the present and aorist stem (disregard of vs. focus on boundaries), I see no semantic, morphological or frequency-based reason to describe the opposition as privative. A better case can be made for the view that the present and the aorist constitute an equipollent opposition, that is, an opposition of which the two members both contribute a positive value.27 There is, however, another manner in which the notion of markedness might be relevant to aspectual distinctions. The notion of markedness as discussed above relates to the general opposition between present and aorist. It is, however, also useful to take a more local perspective, that is, to see whether or not an aspect form is marked with respect to a particular Aktionsart. The distinction between telic and atelic Aktionsarten especially is relevant here. Telic and atelic states of affairs are associated with different modes of viewing. Since telic states of affairs inherently have a salient terminal point (telos), these have to be observed in their totality to be fully identified and to be counted as an instance of the telic type (see Langacker 2009:191, 208). This preference of telic states of affairs to be viewed in their totality makes it more natural to employ an aorist (i.e., perfective) form since it portrays the event as bounded. Since atelic states of affairs inherently lack a culmination point and consist of identical subparts, it is sufficient to view only a limited portion of the full event to identify what kind of event it is. This lack of inherent boundaries makes it more natural to mark them with a present (i.e., imperfective) form since it describes the event without regard to its boundaries. These unmarked relationships (telic-perfective vs. atelic-imperfective) are cross-linguistically very common. Bybee (1985), for example, points out that the various indicators of general markedness are often in conflict when applied

26 27

For a discussion, see Fanning (1990:64–65). See Porter (1989) and Fanning (1990).

the imperfect unbound

121

to aspect and she explains this situation by the close connection of aspect with the inherent lexical meaning (Aktionsart) of the verb. For activity and stative (i.e., atelic) verbs, the imperfective aspect is the conceptually unmarked member, while for telic verbs the perfective aspect is the unmarked member. This type of markedness relation is therefore best described as a form of local (as against general) markedness (Bybee 1985:147).28 6.3.4 Marked Relationships between Aktionsart and Aspect In the preceding section, I described the abstract meaning of the present and aorist aspect as a matter of inclusion or exclusion of boundaries within the temporal scope. In this section, I will demonstrate how this abstract meaning interacts with Aktionsart. I will argue that aspect morphology is employed to impose a particular construal on a situation. This becomes apparent especially in marked relationships between Aktionsart and grammatical aspect. Imperfective morphology is used to defocus from inherent initial and final boundaries, while perfective morphology can be employed to impose boundaries on an atelic state of affairs lacking inherent boundaries. The semantic content provided by the Aktionsart of the state of affairs serves as the conceptual background (base) from which a ‘slice’ is selected and focused upon. The most typical effect of imperfective aspect on a telic state of affairs is that the temporal phase anterior to the telos is profiled (see Fig. 6.9).

figure 6.9 Accomplishment in progress

The event is envisaged as going on before and after the temporal scope (viewing frame). With climaxes, i.e., achievements preceded by a preparatory phase (‘preface’), the present stem effects a focus on the preface of the event (see Fig. 6.10). Typical discourse-pragmatic motivations for viewing the event in this way are: (i) to present the event as (a) a temporal framework within which

28

For this view of aspectual markedness in Ancient Greek see also Napoli (2006:41–44, 60) and Allan (2013b).

122

allan

another event occurs (see 6.2.1.), or (b) as simultaneous to other ongoing events (descriptive function); (ii) to imply pragmatically that the event does not reach its telos and thus remains incomplete (conative implicature of the present stem).29

figure 6.10

Profiled ‘preface’ of achievement

An example of a present stem form of an achievement focusing on the preface is: (9) ἐν τούτῳ δὲ ὁ Βρασίδας […] ἠπείγετο προκατασχεῖν, εἰ δύναιτο, τὴν πόλιν, μὴ ἀφικνουμένου αὐτοῦ τὸ πλῆθος τῶν Ἀμφιπολιτῶν […] οὐκέτι προσχωροίη. ‘Meanwhile Brasidas […] hastened to gain the town, if possible, before the people of Amphipolis should be encouraged by his approach […] and refuse to surrender’. thuc. 4.105.1

Thucydides has not yet arrived at Amphipolos. The present stem ἀφικνουμένου profiles the preface immediately anterior to the actual arrival. Focus on the preface of the event is also involved in the conative reading and the imperfect of likelihood. An example of the latter, taken from Rijksbaron (2006:16–17) is: (10) μεταρσία ληφθεῖσ’ ἐκαινόμην ξίφει. ἀλλ’ ἐξέκλεψεν ‘Held up high I was about to die by the sword but [Artemis] stole me from them’. eur. it 27–28 29

Conative readings are possible with telic events and therefore also occur with achievements (see below).

the imperfect unbound

123

The verb καίνω is an achievement (climax). The imperfect form profiles the ‘preface’ to the imminent event of killing, which was very likely but did not actually occur.30 Hitherto, we have seen a number of cases in which an unbounded construal is imposed on telic states of affairs by the use of present stem morphology. Now I will discuss two cases in which atelic states of affairs are construed as bounded through aorist morphology. A typical way of imposing a boundary on an atelic state of affairs is by using an ingressive aorist. The aorist aspect of atelic verbs typically designates the subject’s entrance into a state or activity. This aorist form focuses on the initial boundary of the state or activity while the state or activity itself and its potential terminal boundary remain backgrounded. This can be represented as:

figure 6.11

Initial boundary profiled: ingressive aorist

Typical examples of ingressive aorists are ἐβασίλευσα ‘I became king’, ἐνόμισα ‘I began to think, got the idea’, ἡσύχησα ‘I came to rest’ (entrance into state); ἐδάκρυσα ‘I started to cry’, ἐγέλασα ‘I started to laugh’ (start activity).31 Another usual way of imposing boundaries on an atelic situation is the use of a complexive aorist.32 In this case, both boundaries are included within the temporal scope. In the case of atelic states of affairs the boundaries are arbitrary rather than inherent in the event type. The state or activity just stops without reaching a natural point of completion (telos). The situation—which

30 31 32

In my view, ‘imperfect of imminence’ would therefore be a more adequate term for this use of the imperfect. The mirror-image of the ingressive aorist, is the effective aorist which only profiles the final boundary (telos) of a telic state of affairs. See Smyth (1956:430–431); Fanning (1990:92–94); and Rijksbaron (2006:12). In the literature, complexive meaning (i.e., as a semantic notion) and summarizing function (discourse pragmatic) are often used interchangeably. However, complexive aorists need not function as a summary of a discourse segment, nor does an aorist with a summarizing function always have to be complexive.

124

allan

may have lasted for a short or long period of time—is thus viewed in its totality (holistically), as represented in the following figure:

figure 6.12

Atelic situation viewed in its totality: complexive aorist

An example is ἡμέρας γὰρ πέντε καὶ δέκα ἔμειναν ἐν τῇ Ἀττικῇ (Thuc. 4.6.2) (‘Because they stayed in Attica for fifteen days’). The Aktionsart of the verb μένω is a state. Through the aorist inflection the state is viewed in its totality and thus construed as bounded.33

6.4

The Imperfect of Continuing Relevance

It is time to return to the imperfect of continuing relevance and see how this particular use of the imperfect might be described in terms of construal, temporal scope, profiling and boundedness. What the imperfect in cases such as ἔπλεον ἐς τὴν Λακωνικήν (ex. 7a) does, in my view, is to mark that the inherent final boundary (telos) should be disregarded as a salient endpoint. The event is construed as if lacking a final boundary in order to highlight that the event also continues to be somehow relevant to the subsequent discourse and that the result of the event should be attended to by the addressee. This semantic structure can be represented in two ways:

33

In a comparable way, the aorist can also appear in clauses with iterative Aktionsart, to mark that the series of iterated events is construed as bounded (completed), cf. the empiric aorist of the type πόλλακις ἔπαθον (Rijksbaron 2006:33). For the bounded construal of iterative events, see also Langacker (2000:248–249). The aorist also occurs in iterative constructions such as the iterative optative, the generic subjunctive plus ἄν, and the iterative use of the secondary indicative plus ἄν. This shows that iterativity is not necessarily associated with the present stem. Once we realize this, the gnomic use of the aorist also becomes less problematic. In its gnomic use (which is only permitted with inherently telic events), the aorist is used to profile one arbitrary bounded instance (a subevent) which is seen as representative of a whole type of event. This type of generic expression is analyzed insightfully by Langacker (2000:276; 2008:526–527).

the imperfect unbound

figure 6.13

125

Accomplishments unbounded: continuing relevance

In representation (13a), both the initial boundary (moment of departure in the case of verbs of motion such as ἐχώρει and ἔπλει) and the telos (the arrival at the spatial goal) are disregarded, resulting in an unbounded construal of the event. In the alternative representation (13b), the initial boundary is included within the viewing scope. However, even though the initial boundary is within the scope of view, imperfective morphology is used to de-emphasize the final boundary and to highlight the continuing relevance of the event.34 As a conclusion to this paper, I would like to turn to the verb πέμπω and examine how the semantics of this verb might be described by means of the theoretical concepts introduced in 6.3. Πέμπω is known to have a remarkably flexible use of the imperfect (see, for example, Kühner-Gerth 1904:143; Svensson 1930:69). A point that has often been overlooked—probably under the influence of its usual translation ‘to send’—is that πέμπω does not exclusively refer to the punctual moment of sending away from the location of departure.35 The 34 35

This alternative representation seems more attractive in cases such as ἀπέπλει ἐς τὸ Ῥήγιον (ex. 8), in which the preposition ἀπο- explicitly refers to the moment of departure. For πέμπω as an achievement, see e.g., Fanning (1990:157).

126

allan

lexical semantics of πέμπω also encompass the movement towards, and the arrival at, the spatial goal of the mission. In other words, the lexical meaning of the verb is not an achievement (denoting only the moment of departure) but an accomplishment (Fig. 6.3). The role of aspect morphology is to focus on a particular stage of the process of sending. Aorist morphology construes the event as bounded, that is, either the event is viewed as in its totality including the moment of departure (initial boundary) and the moment of arrival (terminal boundary), or—less frequently—it will focus on the moment of departure only, disregarding the following stages. The imperfective form construes the event as unbounded and this may manifest itself in several ways. The first way is to focus on the intermediary stage of motion while disregarding departure and arrival. This is the case in examples such as: (11) πέμποντο δὲ οὗτοι παρὰ Ἀμύντην αἰτήσοντες γῆν τε καὶ ὕδωρ Δαρείῳ βασιλέϊ. ‘These were being sent to Amyntas to demand earth and water for Darius the king’. hdt. 5.17.1

The imperfect ‘zooms in’ on the circumstances accompanying their voyage (‘they were underway on their mission’): in 5.17.2 the route taken by the envoys is elaborated upon. The imperfect disregards the moments of departure and arrival, as they are irrelevant to the current purposes of the narrator. Departure and arrival are, instead, mentioned elsewhere: the moment of their being sent off has already been referred to in 5.17.1 (πέμπει ἀγγέλους) and their arrival is mentioned later in 5.18.1 (οἱ ὦν Πέρσαι οἱ πεμφθέντες οὗτοι παρὰ τὸν Ἀμύντην ὡς ἀπίκοντο).36 Finally, the present stem may also be used to emphasize the continuing relevance of the act of sending, even after the person sent has reached the spatial goal. (12) ἀγνοεῦντες δὲ τὰς συμφορὰς οἱ Θηραῖοι ἔπεμπον ἐς Δελφοὺς περὶ τῶν παρεόντων κακῶν· ἡ δὲ Πυθίη σφι ἔχρησε […]

36

More examples in Svensson (1930:71). Note that the special meanings ‘accompany, escort’ and ‘conduct/take part in a procession’ tend to attract imperfective aspect (e.g., Hdt. 5.56). Nonetheless, if the event is completed, aoristic aspect is used (Hom. Od. 11.626).

the imperfect unbound

127

‘The Theraeans, not understanding what happened to them, sent to Delphi to ask about their present problems. The Pythia replied to them […]’. hdt. 4.156.1

It is clear from the fact that the Pythia gives an answer that the envoys of the Theraeans have been sent on their way and have indeed arrived in Delphi. A translation ‘they were sending’ would clearly not be fitting. Herodotus often uses imperfects to refer to missions of inquiry to the oracle. Svensson explains these imperfects as highlighting the fact that one expects an answer to the inquiry. In other words, the mission is not over when the envoys arrive at their goal and ask their question—the continuation (in this case, the oracle’s response) is also seen as relevant. In this regard, the imperfect of πέμπω behaves in a way very much like the imperfect of verbs of saying, asking and commanding.37 This use also explains the following remarkable example which combines two different aspect forms referring to the same event: (13) βραδεῖαν ἡμᾶς ἆρ’ ὁ τήνδε τὴν ὁδὸν πέμπων ἔπεμψεν, ἢ ’φάνην ἐγὼ βραδύς ‘The one who is sending me on this journey apparently sent me too late, or I turned out to be slow’. soph. Aj. 738–739

The aspect forms are associated with different perspectives on (construals of) the event. The present participle focuses on the current relevance of the sender, Teucer, and the mission, while the aorist profiles the moment of sending off (initial boundary), which may have been too late. I conclude my discussion of πέμπω by noting that the feature of continuing relevance is possibly not restricted to the imperfect but may also be present in other tenses and moods of the present (imperfective) stem. For example, there may be a connection with the use of the present tense (primary indicative) of πέμπω referring to the temporal stage following the arrival at the destination. An example in case is: (14) δὸς δὴ τὴν ἐπιστολὴν ἥν […] πέμπει πρὸς τοὺς ἐν Πελοποννήσῳ συμμάχους ὁ Φίλιππος

37

See Svensson (1930:72) for more examples.

128

allan

‘Now hand me the letter which […] Philip sent to his allies in the Peloponnesus’. dem. 18.156

The dispatch of the letter is physically completed at the moment of arrival at its destination, Philippus’ Peloponnesian allies. In other words, the telos has been accomplished. The present tense, however, emphasizes that the letter sent by Philippus is currently still relevant at the moment of speech.38

6.5

Conclusion

Imperfects referring to completed states of affairs are undoubtedly among the most intriguing phenomena in the aspectual system of Ancient Greek. These imperfects are atypical since they involve a marked relationship between telicity and boundedness: telic states of affairs are construed as unbounded for discourse-pragmatic purposes. The inherent telos of the event is de-emphasized while its continuing relevance is profiled. Cognitive Linguistics concepts such as construal, (temporal) scope of view, profile and base appear to be highly useful theoretical concepts to describe the intricate interplay between Aktionsart and aspect, and to analyze the semantics of this and other uses of the imperfect in an insightful way. It will be interesting to see whether a Cognitive Linguistics approach is able to shed new light on other old aspectual conundrums, such as the ‘praesens pro perfecto’, the ‘tragic’ aorist and the gnomic aorist,39 to name but a few.

38

39

The present tense used to designate past states of affairs of which the effects or resultant states still continue at the present moment of speech (typically verbs of perception, saying, motion and begetting/dying) is sometimes called praesens pro perfecto. See also Kühner-Gerth (1904:1.135–137); Schwyzer-Debrunner (1950:2.274–275). In cases where there is a choice between a perfect and this type of present, I suspect that what the perfect expresses is that the subject’s state is somehow relevant to the present (subjectresultativity), whereas the present is used in case the object is more relevant (see the letter in (14)). On the gnomic aorist see Wakker (this volume, chapter 5).

the imperfect unbound

129

References Allan, R.J. 2009. ‘Towards a typology of the narrative modes in Ancient Greek: text types and narrative structure in Euripidean messenger speeches’. In: Bakker & Wakker 2009:171–203. Allan, R.J. 2013a. ‘History as presence: time, tense and narrative modes in Thucydides’. Thucydides between history and literature, ed. by A. Tsakmakis & M. Tamiolaki, 371– 389. Berlin. Allan, R.J. 2013b. ‘Exploring modality’s semantic space: grammaticalisation, subjectification and the case of ὀφείλω’. Glotta 89:1–46. Bakker, E.J. 1997. ‘Verbal aspect and mimetic description in Thucydides’. Grammar as interpretation: Greek literature in its linguistic contexts, ed. by E.J. Bakker, 7–54. Leiden. Bakker, S.J. & G.C. Wakker (eds.). 2009. Discourse cohesion in Ancient Greek. Leiden. Bary, C. 2009. Aspect in Ancient Greek: a semantic analysis of the aorist and imperfective. Unpublished PhD thesis. Nijmegen. Basset, L. 2009. ‘The use of the imperfect to express completed states of affairs: The imperfect as a marker of narrative cohesion’. In: Bakker & Wakker 2009:205–220. Blass, F. 1889. ‘Demosthenische Studien, iii: Aorist und Imperfekt’. RhM 44:406–430. Boogaart, R. 1999. Aspect and temporal ordering: a contrastive analysis of Dutch and English. Den Haag. Buijs, M., ‘Aspectual differences and narrative technique: Xenophon’s Hellenica & Agesilaos’. The language of literature: linguistic aspects of Classical texts, ed. by R.J. Allan & M. Buijs, 122–153. Leiden & Boston. Bybee, J.L. 1985. Morphology: a study of the relation between meaning and form. Amsterdam. Croft, W. 2012. Verbs: aspect and causal structure. Oxford. Fanning, B.M. 1990. Verbal aspect in New Testament Greek. Oxford. Klein, W. 1994. Time in language. London. Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. Langacker, R.W. 1987. Foundations of cognitive grammar. i. Stanford. Langacker, R.W. 1990. ‘Subjectification’. CognL 1:5–38. Langacker, R.W. 1991. Concept, image, and symbol. Berlin. Langacker, R.W. 2000. Grammar and conceptualization. Berlin. Langacker, R.W. 2007. ‘Cognitive grammar’. The Oxford handbook of Cognitive Linguistics, ed. by D. Geeraerts & H. Cuyckens, 421–508. Oxford. Langacker, R.W. 2008. Cognitive grammar. Oxford. Langacker, R.W. 2009. Investigations in Cognitive grammar. Berlin. Napoli, M. 2006. Aspect and actionality in Homeric Greek: a contrastive analysis. Milan.

130

allan

Porter, S.E. 1989. Verbal aspect in the Greek of the New Testament. New York. Rijksbaron, A. 1988. ‘The discourse function of the imperfect’. In the footsteps of Raphael Kühner, ed. by A. Rijksbaron, H.A. Mulder & G.C. Wakker, 237–254. Amsterdam. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Chicago. Ruijgh, C.J. 1971. Autour de τε épique. Amsterdam. Ruijgh, C.J. 1985. ‘L’emploi “inceptif” du thème du présent du verbe grec: esquisse d’une théorie de valeurs temporelles des thèmes temporels’. Mnemosyne 38:1–61. Ruijgh, C.J. 1991. ‘Les valeurs temporelles des formes verbales en grec ancien’. The function of tense in texts, ed. by J. Gvozdanovic & T.A.J.M. Janssen, 197–217. Amsterdam. Ruipérez, M.S. 1954. Estructura del sistema de aspectos y tiempos del verbo greco antiguo: análisis funcional sincrónico. Salamanca. Schwyzer, E. & A. Debrunner. 1950. Griechische Grammatik. ii: Syntax und syntaktische Stilisitik. München. Sicking, C.M.J. 1991. ‘The distribution of aorist and present tense stem forms in Greek, especially in the imperative’. Glotta 69:14–43; 154–170. Sicking, C.M.J. 1996. ‘Aspect choice: time reference and discourse function’. Two studies in the semantics of the verb in Classical Greek, ed. by C.M.J. Sicking & P. Stork, 3–118. Leiden. Slings, S.R. 1997. ‘Adversative relators between push and pop’. New approaches to Greek particles, ed. by A. Rijksbaron, 101–129. Amsterdam. Smith, C.S. 19972. The parameter of aspect. Dordrecht. Strassler, R. (ed.). 1998. The landmark Thucydides: a comprehensive study. New York. Svensson, A. 1930. Zum Gebrauch der erzählenden Tempora im Griechischen. Lund. Vendler, Z. 1957. ‘Verbs and times’. The Philosophical Review 66:143–160. Verkuyl, H. 1972. On the compositional nature of the aspects. Dordrecht.

chapter 7

Aktionsart, Aspect and Category Change in the History of Greek Amalia Moser

7.1

Introduction

Aspect is more or less universally recognized as the defining tam category of the Greek verbal system. There is a consensus on this in both the philological tradition and the linguistic literature on aspect. It is, moreover, almost universally accepted that diachronically the Greek verbal system has been extremely conservative, remaining practically unchanged down to the present day. This paper largely accepts the former assumption, but not the latter, which has been increasingly challenged in the last few years. It is argued here that, while there has been little change morphologically, the system has altered in a fundamental way. More specifically, it takes up the analysis proposed in Moser (2008; 2009), according to which the history of the Greek verb indicates that the system of three (and later two) stems moves from the expression of Aktionsart (actionality/lexical aspect) to the expression of (grammatical) aspect,1 and develops it further, taking a fresh look at the data from a different point of view and in light of recent studies. The claim of a transition from Aktionsart to aspect is controversial in several respects. Apart from challenging the assumption of the conservatism of the Greek verb, it assumes that there is a meaningful distinction between the two categories, Aktionsart and aspect, a theoretical question that remains unresolved in the literature on aspect. This assumption can easily be—and has been—challenged, although in the last decade or so it seems to be gaining ground; and its theoretical merits as well as the main objections against are discussed in 7.2. The way in which this study diverges from the usual path is in its approach to the data. An attempt is made to counteract the obvious difficulties analyzing past stages of the language by adopting a retrospective point of view: a thor1 Moser (2008; 2009) explores the relationship between the categories of aspect and tense, which are claimed to have developed together from the initial Aktionsart-based system. Tense will not be dealt with in this paper.

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_008

132

moser

ough knowledge of the contemporary situation of the verbal system in terms of aspect makes it easy to pinpoint the differences in earlier stages. Logically, these differences, i.e., characteristics which have not survived into Modern Greek, should represent (the traces of) earlier stages of development. The paper is organized as follows: after a brief outline in 7.2 of the theoretical framework adopted here with respect to aspect and Aktionsart, 7.3 provides an overview of the morphological changes to the verbal system in the history of Greek. Section 7.4 analyzes the Modern Greek situation and 7.5 discusses the historical data from a specific genre—historiography—selected on the grounds that it is both prose, and therefore closer to the spoken language, and narrative, a well-studied genre. Since no prose texts are available from before the Classical period, it was necessary to resort to Homer, taking into account the characteristics of the epic, in particular its formulaic character and the dialectal and temporal layering. The conclusions are presented in 7.6.

7.2

Aktionsart and Aspect

The controversy over whether the two categories, Aktionsart/actionality and aspect can, or more importantly, should be distinguished, has been latent since the introduction of the concept in the mid-19th century, although the debate did not become prominent in the theoretical literature until the 1980s.2 The picture is confusing, to a great extent because of the different terminology used by different scholars or, even more misleadingly, because of the same terminology used with different meaning. Generally, the proponents of the distinction base their view, explicitly or implicitly, on subjectivity: Aktionsart is supposed to be objective, since it expresses an inherent property of the situation/event /eventuality (duration, punctuality, iteration, habituality, completion etc.). Of the various classifications available in the literature, Vendler’s (1957) is the most widespread. His four categories are distinguished on the basis of two features, which will be referred to here as terminativity and incrementality, and can be presented as in Table 7.1. Aspect, by contrast, is subjective in the sense that it expresses the point of view of the speaker; more specifically the speaker’s choice of presenting the 2 The Stoics categorize verb forms on the basis of aspectual criteria, but they use the term χρόνος indiscriminately for both aspect- and tense-based distinctions. Curtius (1846) introduced the notion and the term Zeitart (in contrast to Zeitstufe for tense), later replaced by Aktionsart (Brugmann 1855) and Aspektus (Meltzer 1899, translated from the Russian vid, first used by Smetritsky in the 17th c.); see Binnick (1991) on these and all other questions on the history of tense and aspect theory and terminology.

aktionsart, aspect and category change in the history of greek 133 table 7.1

Aktionsart categories, based on Vendler (1957)

Homogeneous

Incremental

atelic

states know believe have love

activities run walk paint read

telic

achievements recognize find be born die

accomplishments paint a portrait build a house design/make a dress read a book

situation either as a whole, without reference to its internal temporal constituency (perfective), or as a developing process (imperfective). In other words, as Comrie (1976:3–4) puts it: The perfective looks at the situation from outside, without necessarily distinguishing any of the internal structure of the situation, whereas the imperfective looks at the situation from inside, and as such is crucially concerned with the internal structure of the situation. It follows that, in theory at least, any situation, regardless of its Aktionsart (duration, continuity, etc.), can be presented in one or the other way, as is the case in the following examples from Modern Greek: (1) Ο Άρης χτες όλη μέρα διάβασε ένα βιβλίο. The Ares yesterday all day read[perfective] a book (2) Ο Άρης χτες όλη μέρα διάβαζε ένα βιβλίο. The Ares yesterday all day read[imperfective] a book ‘All day yesterday Ares read/was reading a book’.

134

moser

The strongest objections raised against the distinction stem from two opposing theoretical paradigms: formal semantics and Cognitive Linguistics. The former is represented by Verkuyl (1993), who does not object to categorizations such as Vendler’s, but considers them purely ontological and irrelevant for linguistic analysis. According to him, aspectuality is a property of the entire sentence and is determined by two features, [±sqa] (Specified Quantity of a), which refers to the opposition between types of arguments, and [±add to], which refers to the incrementality mentioned above. Apart from the fact that [±add to] would be an Aktionsart feature according to the definition adopted here, there are indications that Aktionsart interacts with aspect crosslinguistically in several cases, including Greek (see 7.4 below). However, both the formalization of the analysis and the recognition of the significance of all the arguments of the verb (and not just the object) for the determination of aspectuality are major contributions to the study of aspect. Most Cognitive linguists reject the distinction based on the fundamental tenet of the theory that every statement involves choice on the part of the speaker, since it expresses one of the possible construals (by definition subjective) of each situation (see e.g., Taylor 2002:389–409).3 Even so, it is clear that some circumstances leave a much larger margin of choice than others, and, more importantly, that not all languages incorporate the same choices into their grammar. Languages that have morphological aspect, for instance, afford a set of options for expressing differences such as the one in examples (1) and (2) above grammatically. De Smet-Verstraete (2006) capture this in their proposal of a gradient of subjectivity, distinguishing pragmatic and semantic subjectivity, the latter subdivided into ideational and interpersonal. Even supporters of the Aktionsart-aspect distinction raise objections to subjectivity as its defining feature. Bache (1982) was the first to point out that subjectivity only works for the contrast between Aktionsart and [±perfective], but not for the progressive and the habitual, generally considered as subdivisions of the imperfective (Fig. 7.1). Bache offered the solution of a different definition. In Moser (2008) the solution opted for is for the progressive to be considered a non-prototypical (or not fully grammaticalized) aspectual category, since it is constrained by objective situational features, and for the habitual to remain outside the category

3 Chilton (2007) is a notable exception within the Cognitive paradigm, in that he claims the necessity of the distinction and also proposes a geometrically-based formal model for its analysis.

aktionsart, aspect and category change in the history of greek 135

figure 7.1 Comrie’s (1976:25) classification of aspectual opposition

altogether, since it is entirely determined by the objective feature of regular repetition.4 Moser (2013) proposes an Aktionsart-to-aspect continuum based on subjectivity as understood by De Smet-Verstraete (2006) with Aktionsart at the [-subjectivity] and aspect at the [+subjectivity] end. In agreement with Aarts’ (2007:201) claim that ‘the well-motivated setting up of discrete categories of form classes is logically prior to claiming that gradience obtains between them’, and in the belief that definitions of these categories should be abstract enough to transcend individual languages, (prototypical) aspect is defined as entirely free of any restrictions imposed by the actual temporal structure of situations. It is further claimed that Modern Greek aspect is very close to this abstract notion, a point which will be taken up again in 7.4.

4 In support of this view, it should be noted that several languages with an overt [±perfective] distinction possess extra forms for expressing the progressive and/or habitual (Dahl 1985:96– 102; Bybee et al. 1994:151–160). The habitual depends entirely on the nature of the situation (regular repetition) and under the definition adopted here is therefore closer to Aktionsart than to aspect. The fact that cross-linguistically it is widely expressed through the imperfective can be explained by the possibility of interpreting the regularly repeated occurrences of the situation as phases of a single event. A counterargument to this would be the claim that repetition is not a necessary condition for the habitual (Comrie 1976:27), based on the growing use of the English habitual (used to) as a continuous; rather than proving that the habitual does not require repetition. This seems to indicate a tendency of the habitual to assume the function of a continuous in absence of the basic [±perfective] distinction in the language, however.

136

moser

table 7.2

The indicative of the verb in Classical and Modern Greek

Classical Greek

Modern Greek

pres ipf (imperfective past) aor (perfective past) fut

πράττω ἔπραττον ἔπραξα πράξω

pf plpf fut.pf

πέπραχα ἐπεπράχειν πεπραχὼς ἔσομαι

πράττω έπραττα έπραξα θα πράττω (imperfective) θα πράξω (perfective) έχω πράξει είχα πράξει θα έχω πράξει

Subjectivity is the feature that differentiates the two ends of the continuum and the intermediate points. A gradient, however, presupposes the existence of a uniting feature. In the case of Aktionsart and aspect, this feature is [±telicity], which is understood concretely, i.e., as the existence of an end point, in the case of Aktionsart (terminativity) and abstractly in the case of aspect (perfectivity), in the sense that the perception of a situation as a whole entails the perception of at least some kind of notional boundary.5

7.3

The Diachrony of Greek Aspect Morphology

Table 7.2 juxtaposes the Classical and the Modern Greek verb paradigm in order to highlight the similarities and differences, which are few, but crucial to the argument put forward in this paper. The table omits changes outside the indicative paradigm, namely the loss of the optative and the infinitive, as well as of most of the participles. It shows, however, all the changes that are most pertinent to the current investigation. The obvious ones, i.e., the replacement of the synthetic future and perfect by periphrases, are much-studied phenomena, but their significance for the aspectual system has gone largely unnoticed. The future periphrases introduce the aspectual distinction into the future, a rare occurrence cross-linguistically

5 A similar though not identical understanding of telicity can be found in Horrocks-Stavrou (2007), where terminativity is used in the same way as here, while the aspectual property is labeled ‘delimitedness’, and ‘telicity’ refers to the situation types denoted by whole vps.

aktionsart, aspect and category change in the history of greek 137

(Dahl 1985:79–84; Bybee et al. 1994:275–278), while the replacement of the synthetic perfect by periphrases means the disappearance of the perfect stem, one of the three stems on which the entire system is built. This morphological loss in an otherwise remarkably conservative morphological paradigm must be of significance;6 nevertheless, many analyses of Modern Greek insist on assuming a tripartite aspectual distinction. In 7.4 I will try to dispel this misconception.

7.4

Modern Greek Aspect

The Modern Greek binary aspectual distinction is very close to the prototypical distinction described in 7.2: practically every verb in the language forms both perfective and imperfective forms with the aorist and the present stem respectively, in the indicative (past and future), the subjunctive and the imperative. The imperfective can be used with a continuous or a habitual meaning (the [±progressive] opposition is for all practical purposes irrelevant to Greek). The habitual, which, as predicted by the theoretical discussion in 7.2, can be used with every verb, is in fact compulsory when the pragmatic conditions of habituality hold. In all other cases choice between the perfective and the imperfective is free. Adverbials do not restrict choice, as is often implied, but are themselves part of the speaker’s choice of aspect. Significantly, choice is not even constrained by the type of the verb’s arguments, which undoubtedly affect Aktionsart and the aspectuality of the sentence. A comparison with Verkuyl’s (1993:6) English examples shows the remarkable degree of freedom in Modern Greek (habitual readings of the imperfective are not taken into consideration either here or in Verkuyl 1993): (3) a. Judith ate a sandwich *for an hour/in an hour. Judith ate sandwiches for an hour/?in an hour. Judith ate bread for an hour/?in an hour. Judith ate no sandwich for an hour/?in an hour. b. Η Ρέα έφαγε[perfective] ένα κουλούρι επί μία ώρα/?σε μια ώρα. Η Ρέα έφαγε κουλούρια επί μία ώρα/?σε μια ώρα. 6 It should be noted that the phonological change, which is usually blamed for the loss, did not influence the aorist-perfect distinction. Apart from the fact that the perfect stem was very often based on a different root or a markedly different Ablaut formation, even in ‘regular’ verbs like πράττω in the example, in which the stems are marked by affixes, the difference, if anything, becomes more pronounced after the phonological change: [ks] vs. [kh] > [ks] vs. [x].

138

moser

Η Ρέα έφαγε ψωμί επί μία ώρα/?σε μια ώρα. Η Ρέα δεν έφαγε κανένα κουλούρι επί μία ώρα/?σε μια ώρα. c. Η Ρέα έτρωγε[imperfective] ένα κουλούρι επί μία ώρα/*σε μια ώρα. Η Ρέα έτρωγε κουλούρια επί μία ώρα/*σε μια ώρα. Η Ρέα έτρωγε ψωμί επί μία ώρα/*σε μια ώρα. Η Ρέα δεν έτρωγε κανένα κουλούρι επί μία ώρα/*σε μια ώρα. The point of this juxtaposition is to show that in Modern Greek no restriction is imposed by the type of argument. The time span (one hour) remains unchanged; the adverbial in an hour simply adopts a perfective point of view and does not allow, as for an hour does, for the imperfective. The latter adverbial combines perhaps more naturally with an imperfective, but allows the perfective with a simple shift of emphasis (e.g., in a context where the speaker wants to stress the waste of time on the part of the eater). The point therefore is that neither transitivity nor the existence of count vs. mass arguments influences aspectual choice. The only factor that plays some minor role is definiteness: an object with a definite article (e.g., τα μπισκότα) makes it slightly more difficult—but not impossible—to use the perfective with the adverbial for an hour. The behaviour of the same verb as intransitive (and consequently as activity rather than accomplishment) is exactly the same. The only context in which aspect choice seems to be constrained by external factors is να-complements. Historically, these complements (να + finite forms, usually though not always analyzed as subjunctives) have replaced infinitives, which, however, were also aspectually marked. In theory, there should be free choice. In practice, every imperfective complement is interpreted as habitual, which essentially means that there is no real choice. (4) Προσπάθησα[perfective] να ξεκουραστώ[perfective] περισσότερο, αλλά μάταια. Προσπαθούσα[imperfective] να ξεκουραστώ[perfective] περισσότερο, αλλά μάταια. Προσπάθησα[perfective]/προσπαθούσα[imperfective] να ξεκουράζομαι[imperfective] περισσότερο, αλλά μάταια. ‘I tried to rest longer but in vain’. Paradoxically, however, this is not an argument against the view that Modern Greek aspect is very close to the maximum subjectivity attributed to the prototypical category in 7.2. It is simply a result of an aspectual restriction concerning these non-factive complements: something that is envisaged in the

aktionsart, aspect and category change in the history of greek 139

future has to be envisaged in its entirety. ‘Has to’ is a relative term here. It is merely a rule imposed by the particular language. The point of significance for the purpose of this investigation is that it is not subject to any Aktionsart restrictions, but applies equally to all verbs and all situations in the language. A more obvious argument in this vein can be made about the two groups of verbs that accept only imperfective complements: ‘aspectual’ verbs (‘begin’, ‘continue’, ‘cease’ etc.) and verbs of perception, sensory or mental (‘see’, ‘hear’ etc. and ‘know’, ‘learn’ etc.): the situation needs to be perceived in its development. Again this is not a necessary restriction, but it is one that is imposed on all the verbs of the language, without Aktionsart considerations. The regularity of the system and the freedom of choice (outside να-complements) is broken by a few significant exceptions: i. A number of state verbs do not have perfective forms; ii. The perfective of some state verbs can have an inchoative meaning along with the perfective one (thus turning the verb into an achievement): (5) Κοιμήθηκα εφτά ώρες. I-slept[perfective] seven hours ‘I slept for seven hours’. (6) Κοιμήθηκα στις εφτά. I-slept[perfective] at-the seven ‘I slept (= went to sleep) at seven’. iii. The imperfective of achievements can only be used to present a situation as background to another (7) or when the meaning is extended to include the process leading to the achievement, effectively turning the verb into an accomplishment (8): (7) Τη στιγμή ακριβώς που πατούσα το διακόπτη κόπηκε το ρεύμα. The moment exactly that I-pressed[imperfective] the switch was-cut the power ‘Just as I was pressing the switch, there was a power cut’.

140

moser

(8) Η αδερφή τους πέθαινε από φυματίωση. The sister their died[imperfective] of tuberculosis ‘Their sister was dying of tuberculosis’. All three exceptions are Aktionsart-related. It is argued in this paper that they are relics of older stages in the development of the language. The Modern Greek perfect is usually described as denoting an action completed in the past whose results last into the present, or as denoting current relevance. It is claimed here that, as argued in Moser (2003) and Moser-Bella (2003), the perfect paradigm expresses anteriority in the three time spheres (past, present and future). Aspectually it belongs to the perfective, as would be expected for an anterior and as is evinced through the fact that it is formed by the auxiliary έχω ‘have’ and what is a fossilized relic of the aorist infinitive.7

7.5

The Historical Data

This study is based, as stated in the introduction, exclusively on historiographical texts in order to achieve as great a consistency of style and dialect as possible, given the dialectal peculiarities of the various genres in Greek Antiquity. Historiography is moreover a genre which is well represented in all periods from the 5th c. bce onwards. It is also a narrative, and any investigation can thus rely on a considerable body of research on its use of tense and aspect. For data from earlier periods, Homer had to be resorted to, with all the problems of non-uniformity that the epic presents. Contrary to previous work (Moser 2008; 2009) in support of the claim that the system of the Greek verb moves gradually from the expression of Aktionsart to that of aspect and tense, in which the central argument is the disappearance of the perfect stem, this article focuses on the changing use of the two stems

7 This is not the prevailing view. It appears as one of the alternatives in only one grammar (Clairis-Babiniotis 2005), but otherwise the perfect is most frequently analyzed as a third aspect (e.g. Seiler 1952; Mirambel 1959; and Holton et al. 1997). The present perfect has also been analyzed as a category whose main function is not temporal or aspectual but marking situations as ‘given’ in conversation (Veloudis 2003) and even as an imperfective form (Yu 1983; Tsangalidis 1999) because of its auxiliary and of its current relevance. The choice, however, of the perfective formant when an imperfective alternative was also available, seems to me conclusive.

aktionsart, aspect and category change in the history of greek

141

that carry the [±perfective] opposition: aorist and present. It takes into account forms in the indicative (imperfects and aorists) and infinitival complements. It leaves out the imperative because the use of its different aspectual alternatives display a marked divergence from the usual picture that aspect presents even in Modern Greek. It also leaves out the optative and the subjunctive on the grounds that the former has disappeared, with its functions divided between the descendant of the latter and other forms of the verb in a way that makes any comparison too complicated and too lengthy for the limited space of this paper. Given the retrospective point of view adopted, the decision was also made because the finer points of the change are most evident in the way in which exactly those elements of the system that were preserved almost intact differ from period to period as to their use in context. It will perhaps come as a surprise that there is no mention of transitivity. There is no doubt that all issues relevant to transitivity are of paramount importance for Greek throughout its history. They are also inextricably tied up with Aktionsart and thus, at least indirectly, with aspect.8 This investigation, however, concerns what it hopes to prove is a diminishing degree of influence of Aktionsart on aspectual choice. The factors that determine Aktionsart are not of central importance for this purpose. 7.5.1 The Three Stems in Homer The greatest divergence from the modern use of the three forms of the verb is found, as expected, in Homer. Indeed, their use in Homer diverges from that in Classical Greek much more significantly than is usually acknowledged. Napoli (2006) provides a detailed and insightful analysis of the use of aorist and present stem-based forms. Her main conclusion is that the explanation lies in the prominence that Homeric Greek gives to actionality and aspect (thus resembling the Russian verb system), rather than to tense and aspect, as is the case in English, the Romance languages and Modern Greek. Moser (2008; 2009), while reaching similar conclusions, claims that the aspectual system is still underdeveloped in Homer and that Aktionsart predominates. Apart from the correlation of atelic and telic Aktionsart with the use of the present and aorist stem respectively, the claim is based on: i. Perfect indicatives appearing where Modern Greek and all contemporary Indo-European languages would have presents, almost exclusively with state

8 E.g., De Boel (1987); Lavidas (2006; 2012); Tsimpli-Papadopoulou (2009); and HorrocksStavrou (2010) for the connections between transitivity, aspect, null and cognate objects.

142

moser

verbs as well as with the unexpected group of animal cries. Sometimes they co-occur with presents: (9) αἰεὶ παρμέμβλωκε[pf] καὶ αὐτοῦ κῆρας ἀμύνει[pres]· ‘She is always by his side and protects him from death’.9 hom. Il. 4.11

It is a well-documented fact that the perfect in Homer often denotes simple states rather than states resulting from a prior action, and this is generally considered its original use (see e.g., Monro 1891:31–32; Schwyzer-Debrunner 1950:263–264; Chantraine 1953:197; Duhoux 1992:138–142; and Sihler 1995:564– 568). Resultative uses, already present in Homer, are thought to have arisen as implicatures (because states are often the result of some past action) and according to Chantraine (1927) to have been the cause of the loss of the perfect. ιι. Aorist indicatives appearing where Modern Greek and even Classical Greek would have presents, exclusively with achievements and accomplishments. This use is entirely regular in similes and gnomic uses,10 where an aorist of an achievement or accomplishment sometimes coincides with the present of an activity: (10) ὡς δ’ ὑπὸ φρικὸς Βορέω ἀναπάλλεται[pres] ἰχθὺς θῖν’ ἐν φυκιόεντι, μέλαν δέ ἑ κῦμ’ ἐκάλυψεν[aor] ‘And as beneath the ripple of the North Wind a fish leaps up on the seaweed-strewn sand of a shallow, and then the black wave hides it’. hom. Il. 23.692–693

(11) ἥ τ’ ὀλίγη μὲν κορύσσεται[pres], αὐτὰρ ἔπειτα οὐρανῷ ἐστήριξε[aor] κάρη καὶ ἐπὶ χθονὶ βαίνει[pres]· ‘She first rears her crest only a little, but then her head is fixed in the heavens while her feet tread on the earth’. hom. Il. 4.442–443

9 10

All translations come from the Loeb editions. On the gnomic aorist see Wakker (this volume, chapter 5).

aktionsart, aspect and category change in the history of greek 143

ιιι. Suppletion: usually assumed to be a mechanism for filling morphological gaps, in a language as morphologically resourceful as Greek and in light of later developments (when sigmatic aorists replace some of the suppletive forms) it is more satisfactorily explained as a relic of a stage in which different Aktionsarten were expressed by different roots, because they were understood as different verbs. The same applies to ablaut formations, some of which were later also replaced by ‘regular’ forms (cf. λείπ-ω / ἔ-λιπ-ον > λείπ-ω / ἔ-λειπ-σ-α = ἔλειψα). The frequent absence of the augment, which is later indispensable as a past marker, indicates that not only aspect, but also tense is originally of secondary importance for the verbal system. Nevertheless, the aorist is used overwhelmingly for the past, and the imperfect exclusively so. They often appear together, with the imperfect sometimes used for an eventuality which serves as background to another, but sometimes for one that seems to be equally in the foreground. (12) […] ὅτε Ἴλιον εἰσανέβαινον[ipf] Ἀργεῖοι, μετὰ δέ σφιν ἔβη[aor] πολύμητις Ὀδυσσεὺς. ‘When the Argives embarked for Ilios and with them went Odysseus of many wiles’. hom. Od. 2.172–173

Interestingly, habitual imperfects are often formed with the iterative affix -σκ-: (13) μνηστῆρες δ’ ὁμάδησαν[aor] ἀνὰ μέγαρα σκιόεντα, ὧδε δέ τις εἴπεσκεν[ipf] ἰδὼν ἐς πλησίον ἄλλον ‘Then all the wooers broke into uproar throughout the shadowy halls, and thus would one man speak with a glance at his neighbour’. hom. Od. 18.399–400

This verse is a formula (see e.g., Od. 18.72) and thus an instance of the preservation of older and possibly already obsolete elements. It should be made clear that this is not a uniform picture. The epic, originating as it does in oral poetry and not having been written down for many centuries, is a many-layered text both temporally and dialectally. It seems reasonable to assume, however, that phenomena absent in subsequent periods are traces of older phases of the history of the language. The conclusion reached here is supported by several theories about Indo-European. Unfortu-

144

moser

nately, evidence from early phases of other Indo-European languages is not conclusive on this point. 7.5.2 The Three Stems in Greek Historiography Historiography was chosen as the main source of data because it is a prose genre, and hence lacks the metric restrictions imposed by poetry (which can paradoxically lead to a greater freedom of choice in terms of grammar), and a narrative. Narrative is a well-studied text type, whose default (unmarked) tense is known to be the past and default aspect (for languages that possess the grammatical category) the perfective (see e.g., Fleischman 1990, esp. 52–63). The imperfective is normally reserved for background events and permanent states.11 The historians selected, in chronological order, are: i. Classical: Herodotus, Thucydides, Xenophon ii. Hellenistic: Polybius, Diodorus of Sicily, Arrian iii. Byzantine: Eusebius, Theodoret, Procopius The discussion concerns the entire production of these authors, but the examples and the quantitative data come from randomly chosen 5,000 word sections of the works of each. The differences from Modern Greek, apart from the obvious case of the perfect, are limited to a much more frequent use of the imperfective, both in the indicative (imperfect where a modern speaker would expect an aorist) and in other forms, particularly infinitives as verb complements. Not surprisingly, given the versatility of the aorist and the freedom of Greek with respect to consecutio temporum, there are no uses of the aorist which would not be acceptable to Modern Greek speakers, with the sole exception of its use in the protasis and apodosis of counterfactual conditionals (a context in which Modern Greek only accepts the imperfect and the past perfect): (14) εἰ δὴ καὶ ἠπιστήθη[aor] μὴ πολυπρηγμονέειν, ἀπέλαβε[aor] ἂν Αἴγυπτον ὥστε ἐπιτροπεύειν αὐτῆς ‘And had he but been wise enough to mind his own business, he would have so far won back Egypt as to be its governor’. hdt. 3.15.2 11

Campbell (2007) bases his analysis of verbal aspect in the New Testament entirely on its narrative function, believing that each genre correlates with specific rules on the use of aspect.

aktionsart, aspect and category change in the history of greek 145

7.5.2.1 The Classical Period Compared to the epic, the picture in the Classical texts is much closer to Modern Greek and much further from the Homeric texts than might be expected. In the historians of this period, the ‘unexpected’ forms (from the point of view of Modern Greek speakers) occur as a rule in three types of contexts: i. With verbs that are inherently durative, i.e., activities and accomplishments. (States also appear predominantly in the imperfective, but their Modern Greek counterparts behave in almost exactly the same way). The significant role that Aktionsart plays in this choice is shown by the many passages in which aorists and imperfects coexist, the only difference between them being Aktionsart: (15) εἰ δὲ μή, Μυτιληναίοις εἰπεῖν[perfective] ναῦς τε παραδοῦναι[perfective] καὶ τείχη καθελεῖν[perfective], μὴ πειθομένων δὲ πoλεμεῖν[imperfective]. ‘If not, the generals were to order the Mytilenaeans to deliver up their ships and pull down their walls, and if they disobeyed, to go to war’. thuc. 3.3.3

ii. When an iterative action is described. Interestingly and contrary to Modern Greek, these are cases where there is no regular repetition, i.e., habituality, only simple iteration, hence the unexpectedness. The iterativity can concern actions repeated either by the same or by different agent(s): (16) κρητῆρα ἐν μέσῳ ἔστησαν[perfective] ἀμφοτέρων τῶν στρατοπέδων, μετὰ δὲ ἀγινέοντες κατὰ ἕνα ἕκαστον τῶν παίδων ἔσφαζον[imperfective] ἐς τὸν κρητῆρα· ‘They set a great bowl between the two armies; then they brought the sons one by one and cut their throats over the bowl’. hdt. 3.11.2

(17) κατὰ οὖν μεταπύργιον προσέμισγον[imperfective] πρὸς τὰς ἐπάλξεις, εἰδότες ὅτι ἐρῆμοί εἰσι, πρῶτον μὲν οἱ τὰς κλίμακας φέροντες, καὶ προσέθεσαν[perfective]· ἔπειτα ψιλοὶ δώδεκα ξὺν ξιφιδίῳ καὶ θώρακι ἀνέβαινον[imperfective], ὧν ἡγεῖτο[imperfective] Ἀμμέας ὁ Κοροίβου καὶ πρῶτος ἀνέβη[perfective]· μετὰ δὲ αὐτὸν οἱ ἑπόμενοι, ἓξ ἐφ’ ἑκάτερον τῶν πύργων, ἀνέβαινον[imperfective]. ‘So they came up to the battlements at a space between two towers, knowing that the battlements were deserted. First came the men with the ladders, who set them against the wall; next came twelve light-armed

146

moser

men, with dagger and corslet only, who mounted the ladders. These were led by Ammeas son of Coroebus, who was the first to ascend, and after him his followers ascended, six men going against each of the adjoining towers’. thuc. 3.22.3

This is a particularly interesting example, because the same verb appears twice as an imperfective and once as a perfective describing the same action. The only difference is the single subject of the intermediate verb. It is perhaps again quantity, this time of the object, that explains the imperfective in seemingly inexplicable juxtapositions of the two aspects, as in the following passage: (18) παρέδοσαν[imperfective] σφέας αὐτοὺς ἀμαχητὶ καὶ φόρον τε ἐτάξαντο[perfective] καὶ δῶρα ἔπεμπον[imperfective]. ‘They surrendered unresisting, laying tribute on themselves and sending gifts’. hdt. 3.13.3

Here, the verb in the imperfective (‘send’) has a plural object. The multiplicity of gifts creates some sort of iteration which could induce the use of the imperfective, aided by the fact that a period of time necessarily elapses between sending and receiving the gifts. iii. When an ingressive meaning is conveyed: (19) καὶ αὐτοὶ πρῶτοι ἀφίκοντο[perfective], καὶ ὁλκοὺς παρεσκεύαζον[imperfective] τῶν νεῶν ἐν τῷ ἰσθμῷ ὡς ὑπεροίσοντες ‘And the Lacedaemonians themselves arrived first and proceeded to construct on the Isthmus hauling-machines with which to transfer the ships’. thuc. 3.15.1

One set of verbs that seems to defy any attempt to explain the authors’ predilection for the imperfective on the basis of either Aktionsart or aspectual considerations is the set with meanings such as ‘say’, ‘announce’, ‘order’, ‘ask for’, ‘persuade’ (verba dicendi). Most grammars and studies (see e.g., SchwyzerDebrunner 1950:77–78) offer the explanation that these are verbs whose meaning becomes complete, in the sense that their goal is reached only when

aktionsart, aspect and category change in the history of greek 147

another action takes place. It is not an entirely satisfactory solution, given that this argument would hold for any verb that accepts a complement, and not all of these verbs show the same predilection for the imperfect, but it is accepted by all preceding and subsequent scholars and no other seems to be forthcoming. It is certainly a use that survives well into the Hellenistic era. The uses described in the preceding paragraphs may be very frequent, but this is not to say that aspect is completely tied to Aktionsart during this period; on the contrary, the authors can and often do choose to use aspect in order to convey a certain point of view of the situation: (20) οὓς ἐγὼ λαβὼν οὐκ εἰς τὸ ἴδιον κατεθέμην[perfective] ἐμοὶ οὐδὲ καθηδυπάθησα[perfective], ἀλλ’ εἰς ὑμᾶς ἐδαπάνων[imperfective]. καὶ πρῶτον μὲν πρὸς τοὺς Θρᾷκας ἐπολέμησα[perfective], καὶ ὑπὲρ τῆς Ἑλλάδος ἐτιμωρούμην[imperfective] μεθ’ ὑμῶν ‘And I, receiving this money, did not lay it up for my own personal use or squander it in pleasure, but I proceeded to expend it on you. First I went to war with the Thracians, and for the sake of Greece I inflicted punishment upon them with your aid’. xen. An. 1.3.3

In the following excerpt the imperfective (‘descend’), in opposition to the perfective (‘ascend’), is followed by a detailed description of what Cyrus saw during his descent. It is possible that the imperfective is used to convey the duration of the event, in which case we have an example of a more subjective use of the aspect: (21) Κῦρος δ’ οὖν ἀνέβη[perfective] ἐπὶ τὰ ὄρη οὐδενὸς κωλύοντος, καὶ εἶδε[perfective] τὰς σκηνὰς […]. ἐντεῦθεν δὲ κατέβαινεν[imperfective] εἰς πεδίον μέγα καὶ καλόν ‘At any rate Cyrus climbed the mountains without meeting any opposition, and saw the camp where the Cilicians had been keeping guard. From there he descended to a large and beautiful plain’. xen. An. 1.2.22

There is, however, a strong tendency to use the imperfective in the type of contexts described above. The essential point of difference from Modern Greek is that Aktionsart plays a significant role in the choice of aspect. It seems, in fact, that aspect tends to follow Aktionsart, unless there is a specific reason for which the author does not want to comply with this rule.

148

moser

The graphs (7.2–7.4) show the distribution of ‘unexpected’ imperfectives (by Modern Greek standards) in the three authors on the basis of the texts studied. The virtual absence of unexpected uses of the imperfective of state verbs is due to the fact that choice of aspect in their case has remained more or less unchanged in Modern Greek, especially in narrative, where, as pointed out above, permanent states are typically expressed through the imperfective cross-linguistically. 7.5.2.2 The Hellenistic Period Traditional grammars and more recent studies alike stress the changes in the use of various forms of the verb and devote relatively long discussions to the imperfect, whose uses seem to be rather baffling. From the point of view of Modern Greek, they are much less so than the Classical uses. MoultonTurner (1963:64) mention that ‘the imperfects are retreating before aorists in the Koine’. Fanning (1990:191) is right to point out the different narrative function of the two and their importance in New Testament Greek. The division of functions (the aorist and the imperfect having a foregrounding and backgrounding effect respectively, mentioned in the introduction to 7.5.2 above) seems to be gaining ground in all Koine texts. It is indicative of the degree of change that only a very small percentage of the ‘unexpected’ uses are imperfects: out of thirty-four instances in Polybius, all but two (5.55 percent) were infinitival complements. The general picture of the use of aspect is not as consistent as that in Classical Greek (this being in itself a sign of ongoing change). Moulton-Turner (1963:64) note the co-occurrence of the imperfect with the aorist as an inexplicable phenomenon. There are instances of aorist-imperfect pairs with achievements and accomplishments/activities/states respectively, but there are indeed others where both verbs belong to the same category: (22) οἱ μὲν Αἰγύπτιοι […] ἀπέστησαν[perfective] καὶ προῆγον[imperfective] ἐπ’ Αἰθιοπίας ‘The Egyptians revolted […] and set out for Ethiopia’. diod. sic. 1.67.3

(23) τούτους Ἀντίγονος πεῖσαι[perfective] μὲν ἢ δωροδοκεῖν[imperfective] ἀγεννὲς ἡγεῖτο ‘Antigonus considered it fruitless to persuade or bribe them’. diod. sic. 19.19.4

aktionsart, aspect and category change in the history of greek 149

figure 7.2 ‘Unexpected’ imperfectives in Herodotus

figure 7.3 ‘Unexpected’ imperfectives in Thucydides

figure 7.4 ‘Unexpected’ imperfectives in Xenophon

150

moser

This passage is particularly indicative of both the change and the variation of the time, as the author reserves the imperfective for a verb that could easily be interpreted as an achievement (and which might well appear in imperfective form in the earlier classical authors because of the plural object) and uses the imperfective for a verb that traditionally (and surprisingly from the point of view of the modern speaker) appears almost exclusively in the imperfective. The continued frequent use with verbs such as κελεύω, παραγγέλλω, προστάττω, λέγω, πυνθάνομαι is often mentioned and explained mostly as a result of their inherently incomplete action, following Blass et al. (2001:146), to whom Schwyzer-Debrunner (1950:77–78) also refer in this context. It is likely that these particular verbs are more resistant to change because of their very high frequency in narrative, which gives them a near-formulaic status within the genre. The perfect occupies a very controversial position in this period. There is wide disagreement as to its status. Fanning (1990:290–305) analyzes it as combining stativity, anteriority and perfectivity, thus producing ‘a condition resulting from an anterior occurrence’. Porter (1989) considers it a third aspect (stative); and Evans (2001:26–32) classifies it as an imperfective, because of its original (and still extant) stative meaning. For this paper, what is of significance is that, whatever its central meaning (if it is at all possible to define it in these terms, since in this period it has all the meanings and uses attributed to it in all the studies mentioned), the perfect is about to disappear as a category. It is very widely used, but more and more in ways similar to that of the aorist; indeed, it tends to appear much more frequently alongside aorists, especially in the New Testament, though there was no such occurrence in the texts analyzed here. 7.5.2.3 The Byzantine Period Two of the three historians of this period, Eusebius and Theodoretus, use the imperfect and the aorist in a way identical to that of Modern Greek. The only dissimilarity from Modern Greek is the excessive use of the synthetic perfect (present and past) functioning either as a simple preterite or, more frequently, as an anterior. The significance of this occurrence for the question addressed in this paper is firstly that the perfect is evidently felt as perfective, since perfectivity is a prerequisite for both these functions, and secondly that it tends to take over one of the functions delegated so far to the versatile aorist, that of the anterior. It should be noted, however, that after the final disappearance of the synthetic perfect, the speakers of the language did not seem to feel the need for creating a form to express anteriority until more than ten centuries later, if the analysis in Moser (2008), which attributes this meaning to the modern periphrastic perfect, is correct.

aktionsart, aspect and category change in the history of greek

figure 7.5 ‘Unexpected’ imperfectives in Polybius

figure 7.6 ‘Unexpected’ imperfectives in Diodorus

figure 7.7 ‘Unexpected’ imperfectives in Arrian

151

152

moser

figure 7.8 ‘Unexpected’ imperfectives in Procopius

The Atticist Procopius, by contrast, uses the imperfect, the aorist and the perfective and imperfective infinitives in complements in a thoroughly Classical manner. His text is much further removed from Modern Greek in this sense, not only than those of his near contemporaries, but even than those of the historians of the Hellenistic period, including the Atticist Arrian. It is very likely that he was able to realize the importance of the use of the imperfective for creating a convincingly Attic text exactly because it was so different from the use in his own time.12 Figure 7.8 shows the distribution of ‘unexpected’ imperfective pasts and infinitives in Procopius. In his case also, achievements in the imperfective are almost entirely limited to the same type of verbs that were discussed with regard to the classical authors. 7.5.2.4 The Overall Picture The following graph shows the number of those uses of the imperfective past and infinitive in verb complements in the historians studied that are unexpected for native speakers of Modern Greek. As expected, the number of these uses diminishes with time to reach zero in the early Byzantine period. There is only one slight increase which interrupts the downward flow: Arrian’s text presents marginally more such uses than that of his immediate predecessor, Diodorus. Considering that Arrian is an Atticist and thus consciously trying to imitate the classical historians, it is

12

It is also interesting that the perfects in his text are almost half the number of present and past perfects in Eusebius and that they are neither substitutes for aorists nor anteriors. Again, they are very much in the style of Thucydides.

aktionsart, aspect and category change in the history of greek 153

figure 7.9 Uses of the imperfective

interesting that the increase is so slight. It seems that he has missed this particular characteristic. By contrast, five centuries later, when such uses seem to have completely disappeared from the language, Procopius is much more efficient in his imitation of the classics, the number of his ‘unexpected’ uses bringing him right up to the level of Thucydides.

7.6

Conclusions

The retrospective analysis of the data in the preceding sections leads to the following conclusions. Differences in the use of aspectual forms in earlier phases of the language and Modern Greek, as well as differences among the former stages, can best be explained through a shift from the morphological expression of Aktionsart to that of aspect. The original tripartite distinction of states, punctual/telic actions and durative/atelic actions, expressed by the three stems of the verb (perfect-aorist-present respectively), seems to have gradually given place to a system based on the binary opposition [±perfective], expressed by two of the original stems, aorist and present. This explanation accounts for a number of phenomena, including the crucial, though usually disregarded, loss of one of the three stems of the verb (the perfect): the binary aspectual opposition has no need for three separate stems. The original system has not been preserved intact in any of the surviving texts, but strong traces can be found in Homer, particularly in the frequent use of perfects and aorists as ‘presents’, in the sense that they refer to current states or ‘eternal truths’. This use is limited to states and achievements respectively. It is true that the evidence of the epic is not consistent, as different phases of the history of the language coexist under the metric restrictions of the dactylic

154

moser

hexameter. What is but a conjecture, however, is reinforced not only by the evidence from other ancient Indo-European languages, itself inconclusive and subject to different interpretations, but also by the evolution of Greek itself up to the present day. The retrospective analysis of the historical data opted for in this paper presents a clear picture of the gradually diminishing importance of Aktionsart in the choice of aspect. Modern Greek has a system which allows almost entirely free choice on purely aspectual terms; even in cases where the choice is not free (e.g., in verb complements) the restrictions apply to all verbs regardless of their Aktionsart. The only interferences of Aktionsart that Modern Greek aspect is subject to are the following: the perfective of states sometimes has an ingressive meaning alongside the perfective one, essentially effecting a change in Aktionsart (state > achievement), while the imperfective of achievements can sometimes refer to a process leading up to the punctual change of state, thus also changing their Aktionsart (achievement > accomplishment). Aspectual choice is available to even the earliest historians, who do take advantage of it, but still have a strong predilection for the imperfective when they refer to states and activities and even to accomplishments, especially if they express a collective or iterative (not necessarily habitual) action. The result from the point of view of the Modern Greek speaker is that their texts are full of ‘inappropriate’ imperfectives. These are mostly activities and achievements, as it is still natural for states to be viewed imperfectively in narrative, serving as they normally do as background to action. Aktionsart-determined choices become markedly fewer in the Hellenistic period, where Aktionsart seems to influence rather than impose choice, hence the much smaller number of imperfectives that sound odd for the modern language speaker. The interference of Aktionsart seems to reach the minimal level of Modern Greek already in early Byzantine times, with only the most dedicated Atticists employing imperfectives in a manner that is alien to speakers of the modern language. This account of the history of the verb system in Greek has some additional theoretical implications. It lends support to the proponents of the existence of two separate categories (Aktionsart/actionality and aspect) while at the same time highlighting their affinity: the possibility of a gradual transition would not exist if there was no common ground between the two categories. Their affinity rests on their shared feature of telicity (terminativity in the case of Aktionsart and perfectivity in the case of aspect). The aorist stem seems to have been more versatile, but in the earlier texts the absence of terminativity, which can be the result either of the lack of a terminal boundary or of the multiple repetition of an action, is favourable to the use of the present stem. Later texts show

aktionsart, aspect and category change in the history of greek 155

that the number of such uses diminishes significantly, until the present stem is unconnected to the presence or absence of terminativity and is only used when it is required by the context to present a state or action as developing or simply, in accordance with the rules of narrative, as background to another event; in other words, the present stem becomes a fully-fledged imperfective in binary opposition to the perfective aorist stem.

References Aarts, B. 2007. Syntactic gradience: the nature of grammatical indeterminacy. Oxford. Bache, C. 1982. ‘Aspect and Aktionsart: towards a semantic distinction’. jl 18:57– 82. Binnick, R.I. 1991. Time and the verb: a guide to tense and aspect. Oxford. Blass, F., Debrunner, A. & F. Rehkopf. 200118. Grammatik des neutestamentlichen Griechisch. Göttingen Bybee, J., Perkins, R. & W. Pagliuca. 1994. The evolution of grammar: tense, aspect, and modality in the languages of the world. Chicago. Campbell, C.R. 2007. Verbal aspect, the indicative mood, and narrative: soundings in the Greek of the New Testament. New York. Chantraine, P. 1927. L’histoire du parfait grec. Paris. Chantraine, P. 1953. Grammaire homérique. Paris Chilton, P. 2007. ‘Geometrical concepts at the interface of formal and cognitive models: Aktionsart, aspect, and the English progressive’. Pragmatics and Cognition 15:91– 114. Clairis, Ch. & G. Babiniotis. 2005. Γραμματική της Νέας Ελληνικής. Athens. Comrie, B. 1976. Aspect. Cambridge. Dahl, Ö. 1985. Tense and aspect systems. Oxford. De Smet, H. & J.-C. Verstraete. 2006. ‘Coming to terms with subjectivity’. CognL 17:365– 392. De Boel, G. 1987. ‘Aspekt, Aktionsart und Transitivität’. if 92:35–57. Duhoux, Y. 1992a. Le verbe grec ancien: éléments de morphologie et de syntaxe historiques. Louvain-la-Neuve. Evans, T.V. 2001. Verbal syntax in the Greek Pentateuch: natural Greek usage and Hebrew interference. Oxford. Fanning, B.M. 1990. Verbal aspect in New Testament Greek. Oxford. Fleischman, S. 1990. Tense and narrativity: from Medieval performance to modern fiction. London. Holton, D., Mackridge, P. & I. Philippaki-Warburton. 1997. Greek: a comprehensive grammar of the modern language. London.

156

moser

Horrocks, G.C. 2007. ‘The language of Homer’. A history of Ancient Greek: from the beginnings to Late Antiquity, ed. by A.-F. Christidis, 475–481. Cambridge. Horrocks, G.C. & M. Stavrou. 2007. ‘Grammaticalized aspect and spatio-temporal culmination’. Lingua 117:605–644. Horrocks, G.C. & M. Stavrou. 2010. ‘Morphological aspect and the distribution of cognate objects across languages’. Lexical semantics, syntax, and event structure, ed. by M. Rappaport-Hovav et al., 284–308. Oxford. Lavidas, N. 2006. Μεταβολές στη μεταβατικότητα του ρήματος της Ελληνικής. Unpublished PhD thesis. Athens. Lavidas, N. 2012. ‘Null vs. cognate objects and language contact: evidence from Hellenistic Greek’. Acta Linguistica Hafniensia 44.2:142–168. Monro, D.B. 1891. A grammar of the Homeric dialect. Oxford. Mirambel, A. 1959. La langue grecque moderne: description et analyse. Paris. Moser, A. 2003. ‘Tense, Aspect and the Greek Perfect’. Perfect explorations, ed. by A. Alexiadou et al., 235–252. Berlin. Moser, A. 2008. ‘The changing relationship of tense and aspect in the history of Greek’. stuf 61: 5–18. Moser, A. 2009. Άποψη και χρόνος στην ιστορία της Ελληνικής. Athens. Moser, A. 2013. ‘Aspect and Aktionsart: a study on the nature of grammatical categories’. Major trends in theoretical and applied linguistics, ed. by N. Lavidas et al., 99–121. Berlin. Moser, A. & S. Bella. 2003. ‘Παρελθόν, παρόν, οριστικότητα και παρακείμενος’. Proceedings of the 6th international conference on Greek Linguistics, 18–21 September 2003, ed. by G. Katsimali et al., Rethymnon. Moulton, J.H. 1908. A grammar of New Testament Greek. i: Prolegomena. Edinburgh. Moulton, J.H. & N. Turner. 1963. A grammar of New Testament Greek. iii: Syntax. Edinburgh. Napoli, M. 2006. Aspect and actionality in Homeric Greek. Milano. Porter, S.E. 1989. Verbal aspect in the Greek of the New Testament. New York. Schwyzer, E. & A. Debrunner. 1950. Griechische Grammatik. ii: Syntax und syntaktische Stilistik. München. Seiler, H. 1952. L’aspect et le temps dans le verbe néo-grec. Paris. Sihler, A.L. 1995. New comparative grammar of Greek and Latin. Oxford. Taylor, J.R. 2002. Cognitive grammar. Oxford. Tsangalidis, A. 1999. Will and tha. Thessaloniki. Tsimpli, I.-M. & D. Papadopoulou. 2009. ‘Aspect and argument realization: a study on antecedentless null objects in Greek’. Lingua 116:1595–1615. Veloudis, I. 2003. ‘Possession and conversation: the case of the category perfect’. Perfect explorations, ed. by A. Alexiadou et al., 381–399. Berlin. Vendler, Z. 1957. ‘Verbs and times’. The Philosophical Review 66:143–160.

aktionsart, aspect and category change in the history of greek 157 Verkuyl, H.J. 1993. A theory of aspectuality: the interaction between temporal and atemporal structure. Cambridge. Yu, J. 1983. Η ακολουθία των χρόνων στην Νέα Ελληνική. Unpublished PhD thesis. Athens.

chapter 8

Ὤφελ(λ)ον in Ancient Greek Counterfactual Desiderative Sentences: From Verb to Modal Particle* Antonio R. Revuelta Puigdollers

8.1

Ancient Grammarians

From the 2nd c. ce onwards, ancient grammarians insist in their treatises on the differences between ὤφελον and εἴθε/αἴθε, as in the following passage taken from Herennius’ work (1st–2nd c. ce): (1) αἴθε καὶ ὤφελον διαφέρει. τὸ μὲν γάρ ἐστιν ἀπαρέμφατον πρόσωπον· τὸ δὲ ὤφελον ἐμφαίνει πρόσωπα, οἷον· ὤφελον ἐγώ, ὤφελον σύ, ὤφελον ἐκεῖνος. ἄλλως τε τὸ μὲν αἴθε ἐστιν ἐπίρρημα, τὸ δὲ ὤφελον ῥῆμα καὶ ἐπίρρημα. ‘αἴθε and ὤφελον are different, because the first one is an impersonal form, whereas ὤφελον expresses person, for example, “I should”, “you should”, “he should”’. Besides, αἴθε is an adverb, whereas ὤφελον is a verb and an adverb.’ herennius philo, a.335

Grammarians usually define εἴθε/αἴθε, giving ὤφελον as a synonym, and establishing the differences between the two. For some εἴθε/αἴθε is an adverb and ὤφελον only a verb (indicative aorist of the verb ὀφείλω); for others ὤφελον can be both a verb and an adverb.1 The fact that they define εἴθε/αἴθε in terms of

* This paper was written within the framework of the research projects ‘Funciones y marcas del griego moderno’ (hum2007–61974) and ‘Corpus Morfológico y Formación de Palabras en Griego Moderno’ (ffi2012–31567) financed by the Spanish Ministry of Science and Innovation. I want to express my gratitude to Hartley Ferguson for making my English more understandable. 1 See also Ptolemeus (Diff. 391.8–9), Apollonius Sophista (ὄϕελον and the like are used to express a desire, 125.8, 1st–2nd c. ce), Ammonius (only adverb, Diff. 19.1–4; αἴθε adverb and ὤφελον verb, 19.3, 1st–2nd c. ce), Apollonius Dyscolus (about person and mood, Adv. 142.9, 144.1; Synt. 353.6–354.9, 2nd c. ce), Herodianus Grammaticus (about the writing with omega

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_009

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

159

ὤφελον should be taken as a reflection that the first terms (εἴθε and αἴθε) were already out of use or at least foreign to most speakers of their time, whereas the second (ὤφελον, the defining term) was still in use. On the other hand, the fact that ὤφελον can be treated as both a verb and adverb poses problems. When reading these grammarians, one perceives that they were trying to correct their contemporaries by suggesting the ‘correct’ usage; they were trying to impose on them the grammar that was in use some centuries before (most of the time a reference to Attic classical grammar), as occurs in the following example from Phrynichus (later 2nd c. ce, see Horrocks 2010:138–139): (2) ἀκμὴν ἀντὶ τοῦ ἔτι· Ξενοϕῶντα μέν ϕασιν ἅπαξ αὐτῷ κεχρῆσθαι, σὺ δὲ ϕυλαττόμενος ἔτι λέγε. ‘Ἀκμὴν for ἔτι. They say Xenophon used it once, but you should avoid using it and say ἔτι’. phryn. 93.193

Phrynichus tries to dissuade readers from using ἀκμήν instead of ἔτι, although even Xenophon once made use of it. The adverb ακόμα/ακόμη derives from ἀκμήν and is the common word in Modern Greek for ‘still’ nowadays, whereas ἔτι has disappeared from use. From this it is clear that the change from ἔτι to ἀκμήν/ακόμη/ακόμα was a fact in Phrynichus’ times, or at least this change was underway. In a similar way, when advising readers to consider and use ὤφελον as an inflected verb, or at least as a verb and an adverb, grammarians make clear that in their time this word was already working as an adverb or was at the least changing from verb into adverb. The question is how the past tense (aorist and imperfect) of a verb like ὀφείλω (but no other forms) can become an adverb and express mainly (but not exclusively) counterfactual desires. That is to say, how can tense develop into mood? The second question is how a verb and an adverb can be confused. The remainder of this paper is devoted to these two issues.

or omicron according to the verbal mood, Epim. 270.12–16; those who use ὤϕελον with aorist make a mistake (ἁμαρτάνουσιν), Philet. 255.1–3, 2nd c. ce), Moeris Grammaticus (205.16–18, 2nd c. ce).

160 8.2

revuelta puigdollers

The Verb ὀφείλω

Although it is common knowledge, it is convenient to present some information about the verb ὀφείλω that can help to clarify the general picture.2 This verb has two basic meanings. The first is to ‘owe’, ‘have to pay’, or ‘account for’, as exemplified in (3); the second meaning is ‘to be bound/to be obliged to do’, as exemplified in (4): (3) κἀργύριον ὠφείλησας […] ‘You owed money’. aristoph. Av. 115

(4) ὀφείλεις γὰρ ἐμέο προποιήσαντος χρηστὰ ἐς σὲ χρηστοῖσί με ἀμείβεσθαι ‘You are obliged to reward me for my previous favours’. hdt. 1.41.2

The second meaning derives from the first and has both an epistemic and a deontic interpretation. From this deontic meaning both the counterfactual assertive and the counterfactual desiderative meanings arose (Allan 2013:11– 29), and later the verb became a particle. Example (4) is an instance of deontic ‘must’, (5) of epistemic ‘must’, (6) of counterfactual ‘should (have)’ and (7) of counterfactual desiderative ‘if only/I wish’: (5) εἰ δὲ ἐρίζων πρὸς πᾶν τὸ λεγόμενον μὴ τὸ βέβαιον ἀποδέξεις, σϕάλλεσθαι ὀϕείλεις ἐν αὐτοῖσι ὁμοίως καὶ ὁ ὑπεναντία τούτοισι λέξας· ‘But if you quarrel with all that is said and cannot put forth a secure position, you must be mistaken just as he who holds the contrary opinion’. hdt. 7.50.2

(6) μάλιστα μέν νυν ὠϕείλετε ἐπ’ ὑμέων αὐτῶν βαλλόμενοι ποιέειν ταῦτα· ‘You ought to have done it by yourselves’. hdt. 3.71.5

2 For other aspects of this verb see Allan (2013). I want to thank Professor Allan for providing me with a copy of his article long before it was available in print.

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

161

(7) οὐλόμεν’ αἴθ’ ὤφελλες ἀεικελίου στρατοῦ ἄλλου σημαίνειν, μὴ δ’ ἄμμιν ἀνασσέμεν ‘Ruinous! I wish you directed some other, unworthy army, and were not lord over us’. hom. Il. 14.84–85

These values of the verb ὀφείλω seem to have a diachronic evolution, as Table 8.1 shows (adapted from Allan 2013:30). Apart from the development from a lexical verb into a modal auxiliary (with many different meanings), the table reflects the connection between the several stems and some of these evolutions.3 In particular, the development of counterfactual values is restricted to some past forms (ὤφειλον, and especially ὤφελλον and ὤφελον).4 This paper will focus on the evolution of ὤφελ(λ)ον from a verb into a desiderative illocutionary particle. As for the conjugation of ὤφελ(λ)ον, we find forms with and without augment (in Homer and writers in his tradition).5 Table 8.2 gives a general view of the forms taken into consideration in this paper. 3 Cf. ὀφείλω (pres), stage i: τῷ Ἀσκληπιῷ ὀϕείλομεν ἀλεκτρυόνα ‘we owe a cock to Aesculapius’ (Pl. Phd. 118a); stage iia: ὀφείλεις γὰρ ἐμέο προποιήσαντος χρηστὰ ἐς σὲ χρηστοῖσί με ἀμείβεσθαι ‘you are obliged to reward me for my previous favours’ (Hdt.1.41.2); stage iib: εἰ δὲ ἐρίζων πρὸς πᾶν τὸ λεγόμενον μὴ τὸ βέβαιον ἀποδέξεις, σϕάλλεσθαι ὀϕείλεις ἐν αὐτοῖσι ‘but if you quarrel with all that is said and cannot put forth a secure position, you must be mistaken’ (Hdt. 7.50.2). 4 Cf. ὤφειλον (ipf.i), stage i: εἴπερ μὴ ὤϕειλεν τοῦτο τὸ ἀργύριον ‘if he did not really owe the money’ (Dem. 49.47); stage iia: εἰ δὲ πρὸς ἀλήθειαν διὰ τὸ ἀδύνατον μὴ κατισχύσαι βοηθῆσαι, μηνῦσαί γε πάντως ὤϕειλε τοὺς λῃστὰς ‘and if he was truly prevented from aiding the person because of inability, he was in any case required to lodge information against the bandits’ (Diod. Sic. 1.77.3); stage iii: μάλιστα μέν νυν ὠϕείλετε ἐπ’ ὑμέων αὐτῶν βαλλόμενοι ποιέειν ταῦτα ‘you ought to have done it by yourselves’ (Hdt. 3.71.5); stage iv: οὐ προσοιστέος / ἄλλος πρὸς ἄλλωι· μηδὲ τόνδ’ ὠϕείλομεν ‘one must not be heaped on another; would that we required not even this one’ (Eur. Hec. 394–395); ὤφελον (aor.ii), stage iii: οὐκ ὄϕελον Τρώεσσι κοτεσσαμένη κακὰ ῥάψαι ‘how was I not in my wrath against the Trojans to devise against them evil?’ (Hom. Il. 18.367); stage iv: ὡς ὄϕελέν τις / ἀνδρῶν ἄλλος ἔχειν ‘if only some other of the fighters had your age!’ (Hom. Il. 4.315–316); ὤφελλον (ipf.ii), stage i: ἦλθε μετὰ χρεῖος, τό ῥά οἱ πᾶς δῆμος ὄϕελλε ‘he came to recover a debt that was owing from the whole demos’ (Hom. Od. 21.117); stage iii: ἐπεί μ’ ἔτεκές γε μινυνθάδιόν περ ἐόντα / τιμήν πέρ μοι ὄϕελλεν Ὀλύμπιος ἐγγυαλίξαι ‘since you bore me to be a man with a short life, therefore the Olympian should grant me honour at least’ (Hom. Il. 1.352–354); stage iv: οὐλόμεν’ αἴθ’ ὤϕελλες ἀεικελίου στρατοῦ ἄλλου / σημαίνειν ‘ruinous! I wish you directed some other, unworthy army’ (Hom. Il. 14.84–85). 5 Dictionaries and grammars write ὄφελον when referring to the adverb and record ὤφελον as a bad spelling (see lsj ‘sic’ remark) or a varia lectio for ὄφελον (Blass et al. 2001:544, 290).

162

revuelta puigdollers

table 8.1

Diachronic evolution of ὀφείλω’s stems and tenses6

Lexical Stage i

ὀφείλω (pres) ὠφείληκα (pf) ὀφειλήσω (fut) ὠφείλησα (aor.i) ὤφειλον (ipf.i) ὤφελον (aor.ii) ὤφελλον (ipf.ii)

table 8.2

Modal Stage iia

Stage iib

Stage iii

Stage iv

deontic

epistemic

counterfactual assertion / interrogation

counterfactual wish

+

+

± + +

– + +

+ + + + +

+

–

Conjugation of ὤφελ(λ)ον

ind.aor./ipf sg

aor.part ptcl pl

1 ὤφελ(λ)ον ὄφελ(λ)ον ὠφέλ(λ)ομεν ὀφέλ(λ)ομεν ὀφέλων 2 ὤφελ(λ)ες ὄφελ(λ)ες ὠφέλ(λ)ετε ὀφέλ(λ)ετε ὀφέλουσα ὄφελ(λ)ον ὄφελον 3 ὤφελ(λ)ε ὄφελ(λ)ε ὤφελ(λ)ον

(ὤφελον) ὄφελον

There is homography (graphical identity) between (i) the non-augmented forms in the first singular and third plural (aorist), (ii) the neuter participle and (iii) the adverb (see the forms in bold). But from the 3rd c. bce onwards begins the homophony (phonetic, but not graphical identity) between all these forms and the augmented forms of the same persons (see all the forms in bold and/or

6 In Tables 8.1 and 8.10, ‘-’ signals ‘least frequent, but possible’, ‘±’ ‘a possible, but unusual combination’, ‘+’ ‘a usual combination’, and ‘void’ ‘not attested in the corpus/grammars’.

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

163

italics), since, as stated in grammars (e.g., Browning 1969:26), quantitative differences between omicron and omega were disappearing: ‘But by the end of the 3rd or early 2nd c. bce, confusion between ο and ω begins to occur in letters, and a little later we find grammarians giving elaborate rules for the length of vowels which suggests that their pupils were no longer observing distinctions of length in practice’.7 These similitudes (graphical and/or phonetical) between the adverb ὄφελον (and its variant or ‘misspelling’ ὤφελον) and so many forms of the verbal paradigm, impose some caution when deciding on the adverb’s origin.

8.3

Desiderative Sentences

Before analysing desiderative ὤφελον constructions, it is useful to offer a full view of desiderative sentences in Ancient Greek. Greek has several different ways of formulating desiderative sentences and in the following subsections I will describe three of them.8 In all cases these constructions express two separate States of Affairs (hence SoAs): (i) the speaker’s (1st person) desire and (ii) the SoA that refers to the content of his/her wish. As we shall see, each construction encodes these two SoAs in different ways. 8.3.1 Morpho-Syntactic Structures: Possible and Counterfactual Desires The first group of constructions could be called morpho-syntactic, since they are largely based on verbal moods: the lexeme of the main verb refers to the wish’s content, whereas the verbal mood and the particles convey the illocutionary force (desiderative) and the modality (possible or counterfactual).9 Possible desideratives are marked by the optative mood and their desiderative illocutionary force may optionally be underlined by several particles and particle combinations (see εἴθε in (8) and ὡς in (9)), but their presence is not compulsory (see (10)), since the optative mood conveys by itself both possibility (modality) and desire (illocutionary force):

7 See also Horrocks (2010:163–165). 8 For a description of desiderative sentences in Ancient Greek, see Rijksbaron (2006:27, 28, 41). Caragounis (2006:160–163) points out some other later constructions and offers a comparison with Modern Greek. The whole evolution from Ancient to Modern Greek will be undertaken in a forthcoming article. For lack of space I will restrict my paper to the most representative constructions in Ancient Greek. 9 For this type of desiderative sentence and its relationship with conditionals, see Wakker (1994:384–402).

164

revuelta puigdollers

(8) εἴθε με πυρφόρος αἰθέρος ἀστὴρ τὸν δύσμορον ἐξολέσειεν. ‘I wish that the blazing shaft of the lightning would destroy this barbarian for me!’ aristoph. Thesm. 1050–1051

(9) […] ὡς ὁ τάδε πορὼν ὄλοιτ’, εἴ μοι θέμις τάδ’ αὐδᾶν. ‘May the one who did that perish, if I may speak such a curse without breaking the gods’ laws’. soph. El. 126–127

(10) εἰ δ’ οἵδ’ ἁμαρτάνουσι, μὴ πλείω κακὰ πάθοιεν ἢ καὶ δρῶσιν ἐκδίκως ἐμέ ‘But if the guilt lies with my judges, I could wish for them no greater evils than they inflict unjustly on me’. soph. Ant. 927–928

On the other hand, counterfactual desiderative sentences choose past indicative as mood, but demand the compulsory presence of the same particles and particle combinations, since past indicative can be used by itself to express counterfactuality (modality), but not desire (illocutionary force).10 (11) εἴθε σοι, ὦ Περίκλεις, τότε συνεγενόμην ὅτε δεινότατος ἑαυτοῦ ἦσθα. ‘Ah, Pericles, if only I had known you intimately when you were at your cleverest in these things’. xen. Mem. 1.2.46

Table 8.3 summarizes the characteristics of these constructions (brackets refer to the optionality of elements):

10

Without the presence of the particles, these past indicative forms would express declarative or interrogative factual sentences. Therefore the particles here are directly responsible for the desiderative illocutionary force.

῎ωφελ(λ)ον in greek counterfactual desiderative sentences table 8.3

165

Possible and counterfactual desiderative constructions

Type

Particle

Negation

Mood

Possible Counterfactual

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς] εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς

[μή] [μή]

opt ind.past

8.3.2 Auxiliaries i: Performative ἐβουλόμην A second way of expressing desires is the use of an explicit performative auxiliary verb,11 as for example βούλομαι:12 the verb itself lexically introduces the notion of desire and for this reason it is restricted to the 1st person, since desires refer to the speakers: ‘If you only could […]’ = ‘I (the speaker) wish you could’. The verb βούλομαι in counterfactual contexts (ἐβουλόμην normally with ἄν, but also without) can be used for these purposes in several stages of Greek: it appears in Classical Greek,13 in New Testament Greek (MoultonTurner 1963:91) and, according to Mayser (1970:227), this verb in the imperfect 1st person without the particle ἄν (ἐβουλόμην) is the only attested way to express desires in the Egyptian papyri. The following examples illustrate its use in different periods (with and without ἄν): (12) ὅμως δὲ καὶ δέσποιναν ἐς τέκν’ εὐτυχεῖν ἐβουλόμην ἂν τούς τ’ Ἐρεχθέως δόμους.

11

12

13

As an example of auxiliaries in Greek, see μέλλω’s use as a semi-auxiliary for expressing future (Wakker 2006). For the notion of performatives and their use for expressing wishes in modern colloquial English (I wish he would come) see Palmer (2001:167–171). Other performative verbs like εὔχομαι (ηὐχόμην, εὐξαίμην ἄν) and θέλω (ἤθελον) with the same meaning can be used, although they are less frequent and tend to appear later (Moulton-Turner 1963:91). In Modern Greek, apart from the structures ας/(μακάρι) να + verb (dependent or past indicative) (Holton et al. 2012:517–518), the performative expression θα ήθελα να + verb (dependent or past indicative; lit. ‘I wish that’) is a very frequent means for formulating wishes: θα ήθελα να είχα γνωρίσει τον Θεμιστοκλή (‘I wish I had known Themistocles’, hnc 890831); θα ήθελα να συνεργαστώ μαζί σας (‘I wish I could collaborate with you’, hnc 173904). Its use in Classical Greek does not seem to be recognized in grammars. Perhaps this is due to the overwhelming number of other markers during this period of the language, or to its unfinished process of grammaticalization. Grammars only explain the counterfactual meaning of the imperfect ἐβουλόμην without the particle ἄν (e.g. Kühner-Gerth 1904:ii.1 214).

166

revuelta puigdollers

‘Yet I wish that my mistress too, and the race of Erechtheus, were happy in children’. eur. Ion 567–568

(13) ἐβουλόμην μὲν, ὦ ἄνδρες, τὴν δύναμιν τοῦ λέγειν καὶ τὴν ἐμπειρίαν τῶν πραγμάτων ἐξ ἴσου μοι καθεστάναι τῇ τε συμϕορᾷ καὶ τοῖς κακοῖς τοῖς γεγενημένοις· νῦν δὲ τοῦ μὲν πεπείραμαι πέρᾳ τοῦ προσήκοντος, τοῦ δὲ ἐνδεής εἰμι μᾶλλον τοῦ συμϕέροντος. ‘I could have wished, gentlemen, that my powers of speech and my experience of the world were as great as the misfortune and the severities which have been visited upon me. Instead, I know more of the last two than I should, and am more wanting in the first than is good for me’. antiph. 5.1

(14) ἐβουλόμην δ’ ἂν μὴ ἀπολαχεῖν αὐτὸν κριτὴν Διονυσίοις, ἵν’ ὑμῖν ϕανερὸς ἐγένετο ἐμοὶ διηλλαγμένος, κρίνας τὴν ἐμὴν ϕυλὴν νικᾶν· νῦν δὲ ἔγραψε μὲν ταῦτα εἰς τὸ γραμματεῖον ἀπέλαχε δέ. ‘I wish that he had not been excluded by lot from the judges at the Dionysia, so that you might have seen clearly that he had been reconciled to me, from his decision that my tribe was the winner. In fact he recorded it thus on his tablet, but he was excluded by lot’. lys. 4.3

(15) ἠβουλόμην δὲ καὶ σὲ παραγεγονέναι εἰς τὴν πόλιν ‘Ich hätte gewünscht, auch du wärest in die Stadt gekommen’/ ‘I wish that you too had come to the city’. upz 1.60, 2nd c. bce; mayser 1970:227

An important peculiarity of this construction is that the wish is decomposed into its two components: (i) the desire of the 1st person (therefore the person restriction) marked by the performative verb in 1st person (ἐβουλόμην (ἄν)), and (ii) the object of the desire, represented by the accusativus cum infinitivo (complement sentence). The subjects of both the main and subordinate verb need not be co-referential, although they can be. Table 8.4 summarizes all the features of this construction.

῎ωφελ(λ)ον in greek counterfactual desiderative sentences table 8.4

167

Counterfactual desiderative constructions with ἐβουλόμην

Type

Auxiliary

Negation

Mood

Person

Counterfactual

ἐβουλόμην (ἄν)

[μή]

inf

1

8.3.3 Auxiliaries ii: ὤφελ(λ)ον The third way to express desire is the use of ὤφελ(λ)ον.14 The use of this verb has been very complex throughout the history of the Greek language. I have tried to reduce this large variety of uses to three main constructions.15 Table 8.5 presents the first of them, which I refer to as construction a.16 From a quantitative and diachronic point of view, the first and most extended is construction a, where the verb ὤφελ(λ)ον/-ες/-ε etc. (with or without augment) functions as an auxiliary and is inflected for person and number. The subject appears in the nominative and the verb ὤφελ(λ)ον takes an infinitive as complement: main and subordinate verbs have co-referential subjects. The modal value is counterfactual desiderative. The presence of particles and particle combinations is optional. Some of the possible combinations appear in examples (16) to (19): (16) ἀλλ’ ὤφελε μὲν Κῦρος ζῆν· ἐπεὶ δὲ τετελεύτηκεν […] ‘I wish that Cyrus were still alive! But since he is dead […]’. xen. An. 2.1.4

(17) […] αἴθ’ ἅμα πάντες Ἕκτορος ὠϕέλετ’ ἀντὶ θοῇς ἐπὶ νηυσὶ πεϕάσθαι. ‘I wish that all of you together, instead of Hector, had been slain at the swift ships!’ hom. Il. 24.253–254 14 15

16

For a general study of ὀφείλω and other semantically related auxiliaries, see Ruiz Yamuza (2008). There are other constructions as exemplified in the following example with both forms (ὄϕελες and παρείης) inflected: αἴθ’ ὄϕελες καὶ ἄμικτος ἀνύμϕευτός τε παρείης, ἄϕθορος ὡς κούρη πρὸς πόσιν ἐρχομένη ‘I wish that I had known you when you were single, like a virgin coming to her husband’ (Ant. Gr. 9.229). For the distribution of the forms ὤφελον and ὤφελλον for expressing counterfactuality in the past or the present and their differences in meaning, see Allan (2013:22).

168

revuelta puigdollers

table 8.5

Ὤφελον: construction a

Particle

Lexical (aux)

Subject Negation Mood

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς] ὤφελ(λ)ον/-ες … in agr nom

table 8.6

Ὤφελον: construction b

Particle

Lexical

Subject

Negation

Mood

part ὄφελον

acc

[μή]

inf

[μή]

inf

(18) οὐλόμεν’ αἴθ’ ὤϕελλες ἀεικελίου στρατοῦ ἄλλου σημαίνειν […] ‘You are a doomed man, and I wish that you were in command of some other disgraced army’. hom. Il. 14. 84–85

(19) εἰ γὰρ ὤϕελον, ὦ Κρίτων, οἷοί τ’ εἶναι οἱ πολλοὶ τὰ μέγιστα κακὰ ἐργάζεσθαι, ἵνα οἷοί τ’ ἦσαν καὶ ἀγαθὰ τὰ μέγιστα, καὶ καλῶς ἂν εἶχεν. ‘I only wish, Crito, that the people could accomplish the greatest evil, that they might also be able to accomplish the greatest good’. pl. Crit. 44d

The second construction, construction b (Table 8.6), is based on the participle ὄφελον, which has as subject an accusativus cum infinitivo-construction.17 Using the latest edition of the tlg, the four examples provided below are the only ones available. The last two are the same and they can be understood differently from the first two (as an example of the third construction, construction c, discussed below).18 (20) ὄϕελον καὶ Θηβαίους ἔχειν Δημάδην· ἔτι γὰρ ἂν ἦσαν αἱ Θῆβαι πόλις· 17 18

Moulton-Turner (1963:191, 252) understand ὄφελον as a case of ‘dropped augment’. Examples (22) and (23) can be understood as a combination of the particle ὄφελον and the imperfect ἦν of εἰμί with the meaning ‘to be possible’.

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

169

‘If only the Thebans had possessed a Demades; then Thebes would still be a city’. demad. Fr. 78.16, c. 380–318 bce

(21) ὤ μοι ἐγὼν, ὄϕελόν με διαρραισθεῖσαν ὀλέσθαι Κυανέαις πέτραισιν ἐν ἀξείνῳ τε κλύδωνι· ‘Ah! Woe! If only I had been destroyed by the Cyanae on the Inhospitable Sea’. orph. A. 1159, 4th–6th c. ce

(22) ἐπὶ τῆς ἔναγχος γενομένης συνόδου ἐν τῇ Ἐϕεσίων μητροπόλει, ἣν ὄϕελόν γε ἦν μὴ γενέσθαι ‘In the recent Council of Chalcedon in the metropoly of Ephesus, which I wish had not taken place’. Concilium Chalcedonense 2.1.1.66, Post 451 ce

(23) ἐπὶ τῆς ἔναγχος γενομένης συνόδου ἐν τῇ Ἐϕεσίων μητροπόλει—ἣν ὄϕελόν γε ἦν μὴ γενέσθαι ‘In the recent Council of Chalcedon in the metropoly of Ephesus, which I wish had not taken place’. evagrius, he 68.25, 6th c. ce

Except for Demades’ quotation (4th c. bce), the other three are from Late Antiquity. The scarcity of this construction (only four examples in the corpus) is of importance, as I will try to show later. In the third construction ὄ/ὤφελον does not agree in person and number and behaves rather like an adverb or particle that just expresses illocutionary force. The main SoA is expressed through a verb in personal form (not in infinitive) and is responsible for the final modality: possibility (optative, future and subjunctive) and counterfactuality (past indicative). The exact form of the particle/adverb varies: (i) the form ὄφελον with omicron is the most frequently attested, appearing 147 times; (ii) ὤφελον with omega appears in nineteen cases and is considered a ‘misspelling’ or a varia lectio for ὄφελον; (iii) ὤφελε is attested in two cases.19 Unlike the previous cases, none of these forms should 19

Actually there is only one token: the original text by Callimachus; the other is a quote of the same passage by a later author.

170

revuelta puigdollers

table 8.7

Particle

Ὤφελον: construction c

Lexical

Subject Negation Mood

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς] ὄφελον (147) nom ὤφελον (19) ὤφελε (2) not in agr

[μή]

ind.past opt (-) fut (-) subj (-)

be understood as verbs, since there is another main verb and there is no person agreement with the subject (see Table 8.7). This construction appears in use from the 3rd c. bce onwards (examples are sorted by date). (24) ὤϕελε μηδ’ ἐγένοντο θοαὶ νέες· ‘I wish that swift ships had never even existed!’ callim. Epigr. 17.1, 4th–3rd c. bce

(25) ὄϕελον ἀπεθάνομεν πληγέντες ὑπὸ κυρίου ἐν γῇ Αἰγύπτῳ ‘Through God’s grace, I wish we had died by the hand of the Lord in the land of Egypt’. Ex 16.3, 3rd–2nd c. bce

(26) ὄϕελον κατευθυνθείησαν αἱ ὁδοί μου τοῦ ϕυλάξασθαι τὰ δικαιώματά σου ‘Oh I wish that my ways were steadfast to obey your statutes!’ Ps 118.5, 3rd–2nd c. bce

(27) ὤϕελον γάρ, ἔϕησεν, ὥσπερ τῆς γλώττης εἰμὶ κύριος, οὕτως ὑπῆρχον καὶ τοῦ σώματος. ‘“I wish”—he said—“that I were as much the master of my body as I am of my tongue!”’ diod. sic. 10.18.2, 1st c. bce

(28) καὶ ὄϕελόν γε ἐβασιλεύσατε, ἵνα καὶ ἡμεῖς ὑμῖν συμβασιλεύσωμεν. ‘And I wish you were the sovereign, that we also might reign with you’. 1 Cor 4.8, 1st c. ce

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

171

(29) ὄϕελον ἔμεινας ἐν Βυζαντίῳ· ‘I wish you had remained in Byzantium’. ach. tat. 2.24.3, 2nd c. ce

(30) ὤφελον γὰρ ἐν ἐκείνῳ τῷ τόπῳ ἦμεν. ‘If only we were there, at that place’. aesop. 4.5

(31) ἐπεὶ ὄϕελον καὶ νῦν ἀκολουθῆσαι δυνήσῃ.20 ‘I wish you could follow me now too’. luc. Sol. 15.1, 2nd c. ce

(32) ὄϕελον πάντας τοὺς λόγους μου γράϕωσιν ἐπ’ ἀληθείας ἐπὶ τὰ ὀνόματα αὐτῶν ‘I wish that they would write all my words in their languages truthfully’. Apocalypsis Enochi 97.11, 3rd c. bce–1st c. ce

Table 8.8 gives a synoptic view of the above mentioned constructions of ὤφελον. Although Wackernagel (2009:292) in his Vorlesungen über Syntax proposes the participle (construction b) as the origin for construction c, where the verb works as an adverb or particle, most probably it derives from construction a.21 There are several reasons for this alternative account. i.

ii.

20 21

The participial construction (construction b) is irrelevant from a quantitative point of view and most of the very few attested examples are too recent and allow a different interpretation, whereas the personal construction a is very frequent and is already attested in Homer. The phonetic confusion between ὤφελον (with omega) and ὄφελον (with omicron) from the end of the 3rd or beginning of the 2nd c. bce (Browning 1969:26) should prevent us from mechanically deriving the adverbial ὄφελον with omicron from the homographic participial form. Besides, ὄφελον can be a Homeric form of ὤφελον without augment.

Lucian presents this passage as a clear example of solecism. See the remark in Blass et al. (2001): ‘ὄφελον is nicht ein unaugmentiertes ὤφελον, sondern Ptz. [sc. Partizip] mit ursprünglich zu ergänzendem ἐστίν’.

172

revuelta puigdollers

table 8.8

Synoptic view of ὤφελον’s constructions

Particle

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς]

Lexical

Subject

Negation

Mood

ὤφελ(λ)ον/–ες … in agr nom

inf

part ὄφελον (-)

acc

inf

ὤφελον, ὄφελον, ὤφελε not in agr

nom

[μή]

ind.past opt (-) fut (-)22 subj (-)

iii. The omicron of the adverbial form can be understood as an erosion suffered by forms when grammaticalized and here we are dealing with a grammaticalization process from a verb into an adverb.23 iv. Apart from ὄφελον, there are other adverbial forms derived from the personal verbal paradigm (ὤφελον and ὤφελε), which, although scarce, are more frequent than the participle. v. The evolution from a personal construction to an impersonal one and later back again to a personal one is not very plausible: in construction a, the subject in the nominative is the logical subject of both the auxiliary and the infinitive; in construction b the subject in the accusative depends on the infinitive and the whole AcI is the logical subject of the participle; in construction c the nominative is the subject of the personal verb. The evolution from a to c is more probable than the evolution from b to c. vi. There are parallel cases (within and outside the Ancient Greek tradition) where an inflected verbal form is used as a metalinguistic expression that develops into an adverb or particle. At the end stage only one of the inflected forms is chosen for all cases irrespective of the person, or sometimes, all of them are used irrespective of their original person. For example, the imperative of the verbs ἄγω, φέρω and εἶμι are used as a reinforcement of the directive illocutionary force, as in examples (33) and (34).24 At the beginning, it agrees with the main verb in person

22 23 24

See Moulton-Turner (1963:300). For the concept of erosion see Hopper-Traugott (20032:101). See Goodwin (1889:86–87).

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

173

and number (see example (33)), but later the second person singular is extended to all cases both plural and singular (see (34)).25 (33) ἄγετε, ὦ ἄνδρες, δειπνήσατε ‘Come, my men, get dinner’. xen. Hel. 5.1.18

(34) […] ἄγε δή, Ξανθία καὶ Μανόδωρε, λαμβάνετε τὰ στρώματα. ‘Come on then, Xanthias and Manodorus, pick up our baggage’. aristoph. Av. 656

Likewise, ὤφελ(λ)ον/-ες/-ε/-ομεν/-ετε/-ον in construction a is inflected and changes according to person and number, but since its subordinate verb changes from impersonal infinitive to personal moods, its own person and mood become irrelevant and only one form of ὤφελον’s paradigm is chosen for all cases and this form is eroded (and written with omicron). The phonetic confusion between long and short o vowels in the oral language further contributes to the erosion.26

8.4

Diachronic Development

So far I have tried to answer the question of how the same form can be considered a verb and an adverb: initially we have an auxiliary verb that by a process of grammaticalization has become a particle expressing desiderative illocutionary force. In this section I will elaborate on some aspects of the diachronic evolution of ὤφελον/-ες/-ε and I will try to explain how the grammaticalization process has taken place and a modal value has arisen from a temporal value. This explanation is necessary since only the aorist indicative (ὤφελον/-ες/-ε) and the

25

26

In Spanish the form venga is used in a similar way: it is a 3rd person singular subjunctive/imperative from the verb venir (‘to come’), but it is used as an illocutionary reinforcement with directives. For the parallel between semantic change and phonetic erosion see Bybee-Pagliuca (1985: 59–60) (cited in Hopper-Traugott 20032:101, 157).

174

revuelta puigdollers

imperfect (ὤφελλον/-ες/-ε and less frequently ὤφειλον/-ες/-ε) of ὀφείλω display this desiderative meaning. 8.4.1

Counterfactuality: From Conversational Implicature to Meaning Postulate The first aspect that has to be accounted for is the counterfactual use of the past form ὤφελον/-ες/-ε, since the present, perfect or future forms of the same verb do not develop this additional meaning.27 In many languages, modal expressions of possibility and necessity that refer to the past develop a conversational implicature in the sense that their complement sentence has not taken place, although from a logical point of view these expressions are compatible with both a positive or negative interpretation (see b and c in example (35)), as in the Spanish examples taken from Silva-Corvalán (1995): (35) a. Pudiste llamarme (factual: ‘you could call me’). b. Pudiste llamarme y lo hiciste (‘You could call me, and you did so’). c. Pudiste llamarme, pero no lo hiciste (‘You could call me, but you didn’t’ = ‘you could have called me’). d. Conversational implicature: no me llamaste (‘you did not call me’). Although (35a) is compatible both with (35b) and (35c), the conversational implicature in (35d) is the most frequent inference and makes (35c) (counterfactual) the most probable interpretation. This conversational implicature in some cases is very strong and develops into part of the verbal form’s meaning (conventionalization), that is to say, it is (partially or totally) grammaticalized. This fact is acknowledged in the grammars of languages such as Ancient Greek and Latin, where modal verbs referring to the past exhibit a very frequent counterfactual interpretation, as grammars usually point out. i. Ancient Greek’s ‘irrealis’ meaning of ἐχρῆν/χρῆν, ἔδει, εἰκὸς ἦν, προσῆκεν without ἄν:28

27 28

See Allan (2013:16–19). See Goodwin (1889:151 ff.) and Kühner-Gerth (1904:204ff.).

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

175

(36) χρῆν γάρ σε μήτ’ αὐτόν ποτ’ ἐς Τροίαν μολεῖν ‘You shouldn’t have gone in person to Troy’. soph. Phil. 1363

(37) ὁ δὲ σιγήσας ἡνίκ’ ἔδει λέγειν, ἄν τι δύσκολον συμβῇ, τοῦτο βασκαίνει. ‘The charlatan holds his peace when he ought to speak, and then complains about any unwanted result’. dem. 18.189–190

(38) καίτοι ἄνδρα γε τύραννον ἄϕθονον ἔδει εἶναι, ἔχοντά γε πάντα τὰ ἀγαθά· τὸ δὲ ὑπεναντίον τούτου ἐς τοὺς πολιήτας πέϕυκε· ϕθονέει γὰρ τοῖσι ἀρίστοισι περιεοῦσί τε καὶ ζώουσι ‘And yet an absolute ruler ought to be free of envy, having all good things; but he becomes the opposite of this towards his citizens; he envies the best who thrive and live’. hdt. 3.80.4

(39) ἀλλά, ϕαίητε ἄν, ἔδει τὰ ἐνέχυρα τότε λαβεῖν, ὡς μηδ’ εἰ ἐβούλετο ἐδύνατο ἐξαπατᾶν. ‘But, you might say, sureties should have been taken [but they were not] at the time, so that he could not have deceived us even if he had wanted to’. xen. An. 7.6.23

ii. Latin: counterfactual modal verbs in the indicative:29 (40) Bonus vates poteras esse, nam quae sunt futura dicis. ‘You might have made a good prophet; for you tell what is to be’. pl. Mil. 911

29

For the Latin facts see Leumann et al. (1963–1965:317–328); Ernout-François (1964:247– 248); and Touratier (1994:131–132). Some of the verbs affected are possum, debeo, oportet, decet, necesse est.

176

revuelta puigdollers

(41) Cato, qui Siciliam tenere nullo negotio potuit (et, si tenuisset, omnes boni ad eum se contulissent), Syracusis profectus est. ‘Cato, who could have held Sicily without any trouble—and, if he had held it, all the loyalists would have flocked to him—sailed from Syracuse’. cic. Att. 10.16.3

But this is just a conversational implicature and can be cancelled, as in the following passage taken from Demosthenes, where the modal verb does not imply the negation of the SoA referred to by the infinitive:30 (42) ἔδει δὲ μένειν. πῶς οὖν; ἀρρωστεῖν προϕασίζεται, καὶ λαβὼν Ἐξήκεστον τὸν ἰατρὸν ἁδελϕὸς αὐτοῦ καὶ προσελθὼν τῇ βουλῇ ἐξώμοσεν ἀρρωστεῖν τουτονὶ καὶ αὐτὸς ἐχειροτονήθη. ‘However, he had to remain. What was to be done? He pleaded ill-health; and his brother, taking Execestus the physician with him, repaired to the council-house, made affidavit of the illness, and received the appointment himself’. dem. 19.124–125

The same happened to ὤφελον. Initially it only expressed an obligation about the past (as happens in other forms in Ancient Greek and still in Modern Greek), as some examples from Homer demonstrate. (43) οὐκ ὄϕελον Τρώεσσι κοτεσσαμένη κακὰ ῥάψαι; ‘How was I not in my wrath against the Trojans to devise against them evil?’ hom. Il. 18.367

From the obligation in the past arose the implicature that the SoA has not taken place (counterfactuality). That implicature must have arisen before Homer, because most examples taken from this author already present this counterfactual meaning, as in the following example:

30

See also Hdt. 1.8.6. For the cancellability of this implicature and its conversational/contextual value see Rijksbaron (2006:25–27).

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

177

(44) αἴθ’ ὄϕελες παρὰ νηυσὶν ἀδάκρυτος καὶ ἀπήμων ἧσθαι […] ‘I wish you had remained by your ships without tears and without grief’. hom. Il. 1.415–416

a) b)

You had to remain by your ships without tears and without grief (obligation in the past). Implicature: you did not remain by your ships without tears and without grief.

From Homer onwards this implicature has been incorporated into the meaning of ὤφελ(λ)ον: the conversational implicature has been conventionalized. The change from conversational implicatures (cancellable implicatures) into conventional implications (not cancellable and part of a construction’s semantic meaning) is a frequent step in semantic change.31 8.4.2 From Assertion to Desire The second change is that from assertion to desire: unlike the other forms of ὀφείλω, ὤφελ(λ)ον/-ες/-ε expresses desires.32 This section discusses this change.33 i. Primary and Secondary Illocutionary Force: Indirect Speech Acts34 Languages have grammaticalized sentence types that are prototypically used for expressing certain classes of speech acts. Declaratives are prototypically used to formulate statements, interrogatives to express questions, and so on, as represented in the first two columns of Table 8.9. However, sometimes these sentence types are used with a different second illocutionary force (indirect speech acts) and they formulate different speech acts (see the third column of Table 8.9): questions (45) can be used to formulate polite requests, future declaratives to formulate strong commands (46) and so on, as in the following English examples:35 31 32 33 34 35

See Hopper-Traugott (2003:78 ff.). Except for the very few cases in Homer and very late authors mentioned in the previous section. For this point see also Allan (2013:19 ff.). For these concepts see Dik (1997:299–308). For this phenomenon in general linguistics see Levinson (19874:226–283) and for their application to Ancient Greek and Latin see Revuelta (2005) and Risselada (1993) respectively.

178

revuelta puigdollers

table 8.9

Sentence type and illocutionary force

Sentence type

Primary illocutionary force

Secondary illocutionary force

Declarative Interrogative Imperative Exclamatory Desiderative

Statements Questions Directives Exclamations (surprise, etc.) Desires

Directives ( futurus pro imperativo) Directives Directives

(45) Could you close the window? (Interrogative ⇒ mild directive ‘please close the window’). (46) Tomorrow you will come here at 8 (declarative ⇒ strong directive ‘Come tomorrow at 8, it is an order’). This secondary force can be just a contextual conversational implicature and it can be cancelled as in (47): (47) a: Can you close the window? b: I’ll do it immediately. a: No, I just wanted to know whether you can close the window. I see you can. Or this second illocutionary force can be very frequent and end up grammatized.36 In this case it is recognized by grammars, as happens with the futurus pro imperativo in many languages. In a third case, the secondary illocutionary force may replace the first. The use of ὤφελ(λ)ον in counterfactual wishes reaches this third phase. (48) αἴθ’ ὄϕελες παρὰ νηυσὶν ἀδάκρυτος καὶ ἀπήμων ἧσθαι […] ‘I wish you had remained by your ships without tears and without grief’. hom. Il. 1.415–416

36

There are different degrees of grammaticalization (Hopper-Traugott 20032:113).

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

a) b)

179

(Deontic declarative) You had to remain by your ships without tears and without grief (but you did not) (initial meaning). ⇒ (Counterfactual desiderative) I wish you had remained by your ships without tears and without grief (but you did not) (grammaticalized conversational implicature).

ii. Formal Reflection: Negation μή The degree of grammaticalization of ὤφελον is complete as the negation type shows. Table 8.10 roughly summarizes the main uses of negations in independent sentences.37 iia. Futurus pro imperativo in Ancient Greek: οὐ/μή: Semi-Grammaticalized The futurus pro imperativo alternates between both negations because it is a declarative sentence (for that reason it sometimes takes οὐ, see example 49), but it is used with a directive illocutionary force (for that reason it sometimes takes μή, see 50). The process is not complete and declarative sentences using the future tense (i) do not always express strong commands (the conversational implicature has not been conventionalized in all cases) and (ii) commands in the future can take both negations, as the following examples show: (49) καὶ οὐ τοῦτο λογιεῖσθε ‘You will not think’ => ‘Don’t think’. eur. Med. 1320

(50) μὴ βουλήσεσθ’ εἰδέναι ‘You will not want to know’ => ‘Do not wish to know […]’. dem. 23.117

iib. Ὤφελον (Desiderative) Unlike the futurus pro imperativo, ὤφελον constructions take μή in a very consistent way.38 If we find the other negation (οὐ), the interpretation is not 37 38

For this see Revuelta (2005:1994). Wackernagel (2009:754) wonders that a declarative sentence takes μή instead of οὐ. His explanation is that although ‘the form would have required negative οὐ, the sense led

180

revuelta puigdollers

table 8.10 Sentence type and negation

Sentence type Declarative Interrogative Negation οὐ(κ) μή

+

Imperative Exclamatory Desiderative

Epistemic Epistemic/Deontic

+

±

+

desiderative (see the few early cases in Homer). Therefore the process of grammaticalization is complete: the conversational implicature has become part of the semantic meaning of this construction and this is reflected in the choice of the negation: μή is selected when it conveys wish (in the majority of cases, see (51)), whereas οὐ is the choice in the (few and old) remaining cases (see (52)): (51) ὤφελε γὰρ μηδεὶς ἄλλος Ἀριστογείτονι χαίρειν ‘If only nobody had greeted Aristogeiton’. dem. 25.44

(52) οὐκ ὄϕελον Τρώεσσι κοτεσσαμένη κακὰ ῥάψαι; ‘How was I not in my wrath against the Trojans to devise against them evil?’ hom. Il. 18.367

8.4.3 Adoption of Illocutionary Modality Particles From the beginning of the Greek textual transmission it is possible to use ὤφελον with no desiderative particle as an illocutionary marker or reinforcement. The conclusion is, therefore, that the origin of this construction did not include any particle and its derivation from a declarative sentence has a good foundation. (53) μὴ ὄϕελες λίσσεσθαι ἀμύμονα Πηλεΐωνα, μυρία δῶρα διδούς·

inescapably to μή’. See also Schwyzer-Debrunner’s parallelism between εἴθε μὴ γένοιτο/ἐγενόμην and εἴθε ὤφελον γενέσθαι (1950:346) and Cirac Estopañán (1957:151–152).

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

181

‘I wish you had never petitioned the peerless son of Peleus, nor offered him countless gifts’. hom. Il. 9.698–699

(54) ἀλλ’ ὤφελε μὲν Κῦρος ζῆν· ἐπεὶ δὲ τετελεύτηκεν ‘I wish that Cyrus were still alive! But since he is dead […]’. xen. An. 2.1.4

However, from Homer onwards αἴθε/εἴθε and ὡς may optionally appear, reinforcing the illocutionary force; in later authors the combination εἰ γάρ/αἰ γάρ is added as well: (55) εἴθ’ ὤφελέν μοι κηδεμὼν ἢ ξυγγενὴς εἶναί τις ὅστις τοιαῦτ’ ἐνουθέτει. ‘If only I had a relative, an ally to speak to me like this!’ aristoph. Vesp. 731–732

(56) σπεύσατέ μοι κακὰ τέκνα κατηφόνες· αἴθ’ ἅμα πάντες Ἕκτορος ὠφέλετ’ ἀντὶ θοῇς ἐπὶ νηυσὶ πεφάσθαι. ‘Come to me at once; worthless sons who do me shame; I wish that you had all been killed at the ships rather than Hector’. hom. Il. 24.253–254

(57) ὡς ὄϕελον θανέειν καὶ πότμον ἐπισπεῖν αὐτοῦ ἐν Αἰγύπτῳ· ‘I wish that I had rather died and met my fate there in Egypt’. hom. Od. 14.274–275

(58) εἰ γὰρ ὤϕελον, ὦ Κρίτων, οἷοί τ’ εἶναι οἱ πολλοὶ τὰ μέγιστα κακὰ ἐργάζεσθαι, ἵνα οἷοί τ’ ἦσαν καὶ ἀγαθὰ τὰ μέγιστα, καὶ καλῶς ἂν εἶχεν. ‘I only wish, Crito, the people could accomplish the greatest evils, that they might be able to accomplish also the greatest good things’. pl. Crit. 44d

182

revuelta puigdollers

This reinforcement is probably due to the semantic equivalence of this ὤφελον construction to the more prototypical, my so-called morpho-syntactic desiderative patterns (see 8.3.1). The presence of these particles with ὤφελον constructions is a case of re-characterization motivated by the following analogy: since the ὤφελον constructions express wishes and those particles appear in wishes, they are incorporated to the ὤφελον constructions in a redundant way. It is just a case of hyper-characterization and highlights the fact that this construction expresses wishes in a fully grammaticalized way. 8.4.4 Analogy: Adoption of the Indicative Past Forms and Other Moods In a fourth stage the ὤφελον constructions replace the infinitive (see example (59)) by a personal form in the past indicative (see examples (60) to (62)), while ὤφελον becomes just an illocutionary marker and takes the form of an adverbial or particle (ὄϕελον, ὤϕελον, ὤϕελε). Despite the change in mood the whole construction still expresses counterfactual wishes. (59) ἀλλ’ ὤφελεaux μὲν Κῦρος ζῆνinf· ‘I wish that Cyrus were still alive!’ xen. An. 2.1.4

(60) ὤϕελεptcl μηδ’ ἐγένοντοpers θοαὶ νέες· ‘I wish that swift ships had never even existed!’ callim. Epigr. 17.1, 4th–3rd c. bce

(61) ὄϕελονptcl ἀπεθάνομενpers ‘I wish we had died’. Ex. 16.3, 3rd c. bce

(62) ὤϕελονptcl γάρ, ἔϕησεν, ὥσπερ τῆς γλώττης εἰμὶ κύριος, οὕτως ὑπῆρχονpers καὶ τοῦ σώματος. ‘I wish that I were as much the master of my body as I am of my tongue!’ diod. sic. 10.18.2, 1st c. bce

In later stages it extends its meaning beyond counterfactuality and includes other moods: optative (fewer cases, example (63)) or future (even fewer cases, example (64)) or subjunctive (one case, example (65)):

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

183

(63) εἰ γὰρ ὄϕελονptcl δυναίμηνpers ἐμαυτὸν χειρώσασθαι, ἢ δεηθείς γε ἑτέρου, καὶ ποιήσει μοι τοῦτο. ‘Oh I wish I could lay hands upon myself, or ask another to do it for me’. Job 30.24, 2nd–3rd c. bce

(64) ἐγὼ μὲν λέγω καὶ σολοικίζω, σὺ δ’ οὐχ ἕπῃ τοῦτο δρῶντι· ἐπεὶ ὄϕελονptcl καὶ νῦν ἀκολουθῆσαι δυνήσῃpers. ‘I am saying things and making mistakes, but you don’t keep up with me as I do it. I hope39 you can follow me this time’. luc. Sol. 15.1.28–15.2.2, 2nd c. ce

(65) καὶ ὄϕελονptcl πάντας τοὺς λόγους μου γράϕωσινpers ἐπ’ ἀληθείας ἐπὶ τὰ ὀνόματα αὐτῶν καὶ μήτε ἀϕέλωσινpers μήτε ἀλλοιώσωσινpers τῶν λόγων τούτων, ἀλλὰ πάντα ἐπ’ ἀληθείας γράϕωσινpers ἃ ἐγὼ διαμαρτυροῦμαι αὐτοῖς. ‘I wish that they would truthfully write all my words into their languages, and neither alter nor omit anything from my words but write them all down truthfully—all that I first testified against them’. Apocalypsis Enochi 97.11, 3rd–1st c. bce

With this change the ὤφελον/-ες/-ε constructions modify themselves in two dimensions. First, they extend their range from counterfactuality to possibility. At the same time the auxiliary verb becomes fossilized as a particle for many reasons. i. The features the auxiliary ὤφελον/-ες/-ε contributed to the sentence (person, number, mood) are now marked by the independent personal verb (past indicative, optative, subjunctive or future). ii. The form of the adverb ὄφελον/ὤφελον probably derives from the 1st singular or 3rd plural of ὤφελον/-ες/-ε (the two most frequent in texts), but it has been eroded by the process of grammaticalization and by the phonological changes described by Browning (confusion between long and short /o/ vowel). Other

39

See the translator’s note: ‘A gross mistake involving (a) the late usage of ὄφελον for ὤφελες, and (b) its combination with a future indicative’.

184

revuelta puigdollers

table 8.11 Analogical change of ὤφελον

Type

Particle

Independent clause

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς]

ὤφελον a

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς]

ὤφελον b ὤφελον c

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς] ὤφελον, ὄφελον, ὤφελε

Auxiliary

Subject

Neg.

Mood

nom

[μή]

Opt ιnd.past

ὤφελον/-ες/-ε

nom

[μή]

inf

part ὄφελον (-)

acc nom

inf [μή]

ind.past opt (-) fut (-) subj (-)

forms could be possible in previous stages as ὤφελε in Callimachus’ example shows, before ὄφελον/ὤφελον became the rule. iii. The adverb/particle fixes its position to first sentence position, as suits its function, since the particle ὄφελον establishes the illocutionary force of the whole speech act. This change of word order is already visible in ὤφελον/ες/ε, which changes its word order when passing from declarative to desiderative (Allan 2013:20–21). As we have seen before (8.3.3), there are parallels of such processes of conversion and attrition within Ancient Greek (e.g., ἄγετε, ἄγε), and in other languages like Spanish (e.g., venga) or English (e.g., come on). Table 8.11 shows how the ὤφελον/-ες/-ε a constructions have developed into the c constructions under the pressure and in analogy with the morpho-syntactical pattern (εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς + optative/indicative past). The analogy of the morpho-syntactical pattern is clearly reflected in the transition from expressing counterfactual wishes to expressing any kind of wishes. In fact at the end of its evolution the ὄ/ὤφελον constructions convey any kind of wishes. 8.4.5 Decategorialization: From Verb to Particle Table 8.12 gives a possible chronology in the evolution of desiderative constructions headed by ὤφελον and the evolution from a fully inflected auxiliary verb into an illocutionary particle following the pattern established by the morphosyntactic constructions (εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς + past indicative/optative).

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

185

table 8.12 Diachronic evolution of ὤφελον

Construction

ὤφελον desiderative

Morphosyntactic desiderative

Phases

Particle

Main v

Subordinate v

PreHomeric

Ø

ὤφελ(λ)ον/-ες/-ε … (aux, inflected, in agr)

inf

Homer

[εἴθε/αἴθε, ὡς]

ὤφελ(λ)ον/-ες/-ε … (aux, inflected, in agr)

inf

3rd c. bce

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς] ὤφελε/ὤφελον/ὄφελον (no agr)

ind.past

1st–2nd c. ce?40

[εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς] ὤφελον/ὄφελον (no agr)

ind.past/opt/fut/subj

εἴθε/αἴθε, εἰ/αἰ γάρ, ὡς

ind.past/opt

Originally ὤφελον/-ες/-ε was an auxiliary verb that took an infinitive (co-referential subject) and simply expressed an obligation in the past, as some examples in Homer still attest (see 8.4.1). Already in Homer it was used for conveying counterfactual desires and ‘copied’ some of the features exhibited by other equivalent constructions (mainly the presence of particles, see εἴθε/αἴθε). From the 3rd c. bce onwards, due to phonological changes and the analogous pressure of other constructions, the auxiliary turned into a particle expressing just illocutionary force (desire), whereas the SoA in the infinitive occupied the position of main verb and exhibited the features of mood (counterfactual or possible), person and tense. At a later phase the particle accepted other modalities apart from counterfactuality (see optative, future and subjunctive).

40

This dating is just approximate and requires further research. There are examples since the 3rd c. bce, but they come from translation works (Septuaginta). Lucianus (Sol. 15.1.28– 15.2.2) presents these cases as mistakes, which should be taken as evidence that the change was under way in his time.

186

revuelta puigdollers

figure 8.1 Grammaticalization process of ὤφελον

The verb experienced a process of decategorialization (Hopper-Traugott 20032: 106ff.). Probably both the former use (as an auxiliary) and the new one (as a particle) coexisted for a long time, since written language is conservative. Furthermore, Atticism brought the revival of Attic grammar into Roman times. Passages like the one quoted at the beginning of this article are just a reflection of this coexistence. According to the grammarian’s linguistic ideology (Atticism or not), the status of ὄ/ὤφελον as an adverb (the new grammar) is rejected or not, whereas all the grammarians seem to accept its status as a verb (the old grammar). As the following example taken from Hopper-Traugott (20032:208) shows, there are similar processes in other languages. (66) a. I thinkmain verb Commander Dalgleish writes poetry. b. Commander Dalgleish writes poetry, I thinkparenthetical/adverb. After these explanations we can now better understand the information available in the works of the ancient grammarians. For some of them ὄ/ὤφελον was an inflected verb, because they were looking at Classical Attic Grammar, and they disregarded and despised the uses presented by their contemporaries, for whom it had become an adverb. Just a few grammarians were ready to accept that language had changed or was changing.

῎ωφελ(λ)ον in greek counterfactual desiderative sentences

187

Figure 8.1 (slightly adapted from Allan 2013:30), represents the evolution of ὀφείλω from a lexical verb to an auxiliary and from an auxiliary to a particle. The evolution is restricted to certain stems and to certain tenses. The present stem is compatible with both epistemic and deontic modality, whereas counterfactuality seems to be restricted to past tenses (see 8.4.1) and particularly to the aorist ὤφελον and the imperfect ὤφελλον (ὤφειλον is rather infrequent). All forms seem to retain the initial lexical meaning ‘to owe’ (ὤφελλον is very restricted in this sense), except ὤφελον, which is the most grammaticalized and the only one ending up as an adverb. The cline, that is to say, the path of grammaticalization, is not abrupt, but follows a series of small transitions, as is usual in these kinds of processes, before the lexical verb ends up as a desiderative illocutionary particle equivalent to εἴθε/αἴθε.41

References Allan, R.J. 2013. ‘Exploring modality’s semantic space: grammaticalization, subjectification and the case of ὀφείλω’. Glotta 89:1–46. Blass, F., Debrunner, A. & F. Rehkopf. 200118. Grammatik des neutestamentlichen Griechisch. Göttingen. Bybee, J. & S. Fleischman (eds.). 1995. Modality in grammar and discourse. Amsterdam. Caragounis, C.C. 2006. The development of Greek and the New Testament: morphology, syntax, phonology, and textual transmission. Tübingen. Cirac Estopañán, S. 1957. Manual de gramática histórica griega. iv: Lecciones de sintaxis del verbo y de las oraciones. Barcelona. Dik, S.C. 1997. The theory of functional grammar. i: The structure of the clause. Berlin. Ernout, A. & T. François. 19643. Syntaxe latine. Paris. Goodwin, W.W. 1889. Syntax of the moods and tenses of the Greek verb. London. hnc = Hellenic National Corpus http://hnc.ilsp.gr/en/default.asp Holton, D., Mackridge, P. & I. Philippaki-Warburton. 2012. Greek: a comprehensive grammar of the modern language. London. Hopper, P.J. & E.C. Traugott. 20032. Grammaticalization. Cambridge. Horrocks, G.C. 20102. Greek: a history of the language and its speakers. Chichester. Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. lsj = Liddell, H.G., R. Scott & H.S. Jones. 1996. A Greek-English lexicon. Oxford. Leumann, M., Hofmann, J.B. & A. Szantyr. 1963–1965. Lateinische Grammatik. ii: Lateinische Syntax und Stilistik. München. 41

For the concept of a cline see Hopper-Traugott (20032:6ff.).

188

revuelta puigdollers

Levinson, S.C. 19874. Pragmatics. Cambridge. Mayser, E. 1970. Grammatik der griechischen Papyri aus der Ptolemäerzeit. ii.1: Satzlehre. Leipzig. Moulton, J.H. & N. Turner. 1963. A grammar of New Testament Greek. iii: Syntax. Edinburgh. Palmer, F.R. 20012. Mood and modality. Cambridge. Revuelta Puigdollers, A.R. 1994. ‘El significado del subjuntivo deliberativo en griego antiguo’. Quid ultra faciam? Trabajos de griego, latín e indoeuropeo en conmemoración de los 25 años de la Universidad Autónoma, ed. by L.M. Aparicio et al., 75–86. Madrid. Revuelta Puigdollers, A.R. 2005. ‘Modo y modalidad en griego antiguo. La negación’. Sintaxis griega, ed. by D. Jiménez López. Downloaded at: http://www.liceus.com/bonos/compra1.asp?idproducto=517&modo-y-modalidad-en -griego-antiguo-la-negacion. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Amsterdam. Risselada, R. 1993. Imperatives and other directive expressions in Latin: a study in the pragmatics of a dead language. Amsterdam. Ruiz Yamuza, E. 2008. Tres verbos que significan ‘deber’ en griego antiguo. Zaragoza. Silva-Corvalán, C. 1995. ‘Contextual conditions for the interpretation of poder and deber in Spanish’. Modality in grammar and discourse, ed. by J. Bybee & S. Fleischman, 67–106. Amsterdam. Schwyzer, E. & A. Debrunner. 1950. Griechische Grammatik. ii. München. Touratier, Ch. 1994. Syntaxe latine. Louvain-la-Neuve. Wackernagel, J. 2009. Lectures on syntax with special reference to Greek, Latin, and Germanic. Translated by D. Langslow. Oxford. Wakker, G.C. 1994. Conditions and conditionals: an investigation of ancient Greek. Amsterdam. Wakker, G.C. 2006. ‘Future auxiliaries or not?’ Word classes and related topics in Ancient Greek, ed. by E. Crespo et al., 237–255. Louvain-la-Neuve.

chapter 9

Modalized Future and Scheduled Present in Coan Inscriptions* Marina Veksina

9.1

Preliminaries

The present paper adopts the framework for mood and modality set forth by Van der Auwera-Plungian (1998), which is based to a great extent on Bybee et al. (1994) and Bybee-Fleischman (1995). Within this approach, the term ‘mood’ is reserved for purely inflectional categories, while the term ‘modality’ refers to ‘those semantic domains that involve possibility and necessity as paradigmatic variants’ (Van der Auwera-Plungian 1998:80).1 This two-member paradigm applies to the ‘epistemic’ and ‘non-epistemic’ domains: epistemic modality has scope over the whole proposition, whereas non-epistemic modality contributes to the proposition’s meaning. Within non-epistemic modality the ‘participant-internal’ and ‘participant-external’ realms are usually distinguished according to (the degree of) the involvement of a participant in the state of affairs. Legislative texts naturally constitute a welcome field for the use of the participant-external modality, and in particular its subdomain of ‘deontic modality’ as defined by Van der Auwera-Plungian (1998:81): ‘Deontic modality identifies the enabling or compelling circumstances external to the participant as some person(s), often the speaker, and/or as some social or ethical norm(s) permitting or obliging the participant to engage in the state of affairs […] Permission is thus a matter of deontic possibility’. The ‘participant’ does not necessarily refer to an animated subject, but can also refer to an inanimate one, and with several participants involved, the ascription of modality can become problematic. This ‘modality orientation’ problem (actual in a language with so rich an inflectional system as Greek) cannot be investigated in any detail in this study, nor can notions such as evidentiality and volition, commonly associated with modality. Looking from the diachronic perspective, modality markers can

* This study was supported by the Fritz Thyssen Foundation as part of my work on the ‘Grammar of Coan inscriptions’ project. 1 For various uses of the term see e.g., Portner (2011).

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_010

190

veksina

develop from different expressions, and in their turn can become sources of other grammatical markers. Drawing on cross-linguistic data, Van der AuweraPlungian (1998) present a range of notions that are attested as sources for modal meanings (both epistemic and non-epistemic). Thus, there is a path of development from ‘be strong’, ‘know’, ‘arrive at’, ‘suffice’ towards participant-internal possibility; such notions as ‘owe’, ‘belong’, ‘be good/proper’ become sources of deontic necessity; ‘dare’ can evolve into deontic possibility, and so on. There are paths of semantic development leading from epistemic necessity and possibility towards condition, concession and complementation. Future is found both at pre-modal and post-modal stages. For the modalization of future, Van der Auwera-Plungian (1998:98) only present a path towards epistemic necessity (cf. 9.2.3. below). This article provides evidence for the path from future to deontic modality. In the strict sense outlined above, modality pertains to a level of analysis other than illocution and should not be confused with it.2 Modality markers can be used performatively, but not necessarily so.3 Legal deeds constitute a dimension of their own in the theory of speech acts, and the opposition between ‘primary or natural speech acts’ and ‘secondary or institutional speech acts’ is to be evoked here.4 The principal illocutionary types of these texts are: a) communicating legal actions (declarations) and b) imposing legal norms (directives). As far as Greek epigraphic discourse is concerned, two textual realms should be distinguished with different illocutionary purposes and different conventionalized means of expression.5 First is the introductory part of a legal deed defining its character and purpose, its issuing authority and the circumstances of its issuing, the common elements of this part being invocation, heading, prescript, motion and enactment formulas. Second is the ‘substance’ of a legal deed, where norms are imposed or declarations are made. These two realms are differently organized and behave differently in relation to grammatical categories. It is to be expected that grammatical variations of mood and modality have different characters and grounds in the introductory parts on the one side, and in the substance of decrees on the other. In fact, in the Coan corpus, all the indicatives expressing deontic modality are found only in the 2 This is not the general approach. Bybee-Fleischman (1995:6), for instance, subsume ‘markers of directives, such as imperatives, optatives or permissives, which represent speech acts through which a speaker attempts to move an addressee to action’ under the category of ‘speaker-oriented modalities’. 3 Zanutti et al. (2012:1262). 4 Wunderlich (1980:296). 5 See Rhodes-Lewis (1997:4 ff.).

modalized future and scheduled present in coan inscriptions

191

substantial part of the decrees. These indicatives are restricted to independent clauses, where they can be and are used performatively. The occurrences of these indicatives are further confined to the acts on religious matters, where the normal mood is imperative both in and outside Cos.6 Concerning epigraphical corpora, the problem of discerning whether observable regularities in the use of the grammatical forms reflect language properties or fossilized formulaic patterns exploited in the epigraphic, and in particular, bureaucratic discourse, is not to be underestimated. The formulaic nature of the epigraphic texts becomes prominent in a case study of a corpus, in particular a corpus as extensive as that of the Coan. A linguistically operational model of epigraphic formulas remains a task for future research, though some significant foundations have already been laid by Morpurgo-Davies (1999) and Wachter (1999). In this study I have addressed the problem on a case-by-case basis and have tried to account for possible meaningful differences as well as phraseological boundedness.

9.2

Indicative Future

9.2.1 Coan Data Surprisingly enough, almost all the instances of the finite future forms in an independent clause are restricted to the expression of deontic modality. The future forms are found to have commissive force in two oath specimens: first, in a vow text introduced in the homopoliteia decree between Cos and Calymna 15214– 31,7 and second, in a vow text for some religious officials in 3351– 7. Though commissives are often grouped with imperatives and exhortatives,8 this relates essentially to their performativity, and commissives do not necessarily express modality in the sense outlined above.9 All the other instances are deontic futures. Notably enough, they are only found in sacral laws.10 The vast majority of the examples are restricted to one particular kind of article in the provi-

6 7 8 9 10

Rhodes-Lewis (1997:555 f., 561). Throughout this paper I will refer to the Coan inscriptions by the numbers of the recently published corpus ig xii 4, with subscript numbers indicating lines. E.g., Zanutti et al. (2012). See Van der Auwera-Plungian (1998:83 and 118, note 2). Including 100. In fasc. 2 of the Coan corpus, there is one future form which could build an independent clause; this is a restored future in a presumably honorary inscription: [ἱεράσε]ται διὰ β[ίου] 8879 (early 1st c. ce). Due to the fragmentary state of the text the modal reading cannot be ascertained.

192

veksina

sions on the sale of priesthoods. This is the article that determines the payment modalities for the priesthood purchasers.11 These passages exhibit great uniformity in vocabulary and form so that they can be recognized as formulaic. The following formulaic pattern has been established: Pattern i: [verbal expression for payment operation] [object] [number of instalments] [target] [term fixed] The subject of the verb is always ‘the one who has purchased the priesthood’ (either explicitly stated or easily inferable from the context as in 306, 326, 319, and 304). Some optional elements of the phrase, such as an adverbial expression ποτὶ μέρος, need not be discussed here. The linear order of the lexical units representing these formulaic items varies and will not be discussed here either. The commonest realization of this formula (ia) is as follows: Pattern ia: [verbal expression for payment operation] [object] [target] [number of instalments] [term fixed] The formula for the payment operation is καταβολὰς ποιήσει/ποιησεῖται; the object is a genitive dependent on καταβολάς, e.g. τοῦ εὑρέματος/τοῦ εὑρόντος ἀργυρίου etc.; the target is optional, e.g., ἐπὶ τὰς ταμίας; each installment is expressed by a separate syntagm, e.g.: (1) κατα/βολὰς δὲ ποιήσει (sc. ἁ πριαμένα τὰν ἱερωσύναν) τὰς τοῦ εὑρόντος ἀργυρίου ποτὶ μέ/ρος τέσσαρας, τὰν μὲν πράταν ἐμ μηνὶ Ἀλσείωι […] τὰν δὲ δευτέραν ἐμ μηνὶ Γερα⟨σ⟩τίωι […] ‘She (i.e., the one who has bought the priesthood) shall pay the sum realized (by the sale) part by part in four instalments: the first in the month Alseios […] the second in the month Gerastios […]’. ig xii 4 3028– 11

This structure is further found in 3048 sq., 31526 sqq., 32610 sqq., 31322 sqq., 31816 sqq., 31940 sqq. and 32417 sqq.. The formula is to a great extent restored in 32417 sqq. with the internal object καταβολάς and ‘number of instalments’ item omitted. Such omissions are not found in the extant texts, but the extension of the lacuna does not allow for the whole formula in any case. The other realization (ib) splits the formula into two parts (clauses):

11

On historical details see Wiemer (2003:281 f.).

modalized future and scheduled present in coan inscriptions

193

Pattern ib: [verbal expression for payment operation] [object] [target] [number of instalments] [term fixed] The expression for the payment operation contains the finite verb καταβαλεῖ followed by ποεύμενος καταβολάς or καταβολὰς ποιησεῖται. This pattern is found in two inscriptions: (2) ὁ δὲ πρι/άμενος καταβαλεῖ τὸ ἀργύριον ἐπὶ τὸς ταμίας ποεύμε/νος καταβολὰς τρεῖς, τὰν μὲν πράταν ἐμ μηνὶ Γεραστίωι,/ τὰν δὲ δευτέραν ἐν Ὑακινθίωι, τὰν δὲ τρίταν ἐν τῶι Ἀλσείωι […] ‘The buyer shall pay the money to the treasurers making three instalments, the first in the month Gerastios, the second in Hyakinthios, the third in Alseios […]’. ig xii 4 298136 sqq.

(3) [τὰ]ν τιμὰν καταβαλεῖ ἐπὶ τὸς ταμ[ί]ας· καταβολὰς ποιησ̣ [εῖ/τ]α̣ι ̣ τρεῖς, τὰν μὲν πράταν ἐμ μηνὶ Καφισίωι […] ‘(The buyer) shall pay the price to the treasurers: he shall make three instalments, the first in the month Kaphisios […]’. ig xii 4 30615 sq.

The ‘number of instalments’ items are always rhematic, as are the date specifications within the ‘term fixed’ items. There are two further patterns describing singular payment acts that should take place simultaneously with some other instalment(s), which represent compressed variants of the formulaic pattern established above: Pattern ii: [verbal expression for payment operation] [object] [term fixed] The first variant (iia) has καταβαλεῖ followed by an accusative object (τὸ γενόμενον ἀνάλωμα etc.) and the term fixed (ἅμα τᾶι πράται καταβολᾶι), and is found in 31559– 60, 31948– 50, 32879 (preceded by the full formula in 31526 sqq. and 31940 sqq. respectively). The second variant (iib) has ἀποδώσει followed by an accusative object (τὸ δαπάναμα/τὸ ποτικατάβλη̣ μα) and the term fixed (ἅμα τᾶς ἄλλας τιμᾶς) and is found in 30443– 46 and 32527– 29. Beyond the priesthood sale regulations there are two possible occurrences of the future forms of the same verbs: [– – – κατα/β]α̣λεῖ ὁ φάνας τὸ πέμπτον μέρος [ – – – ] 10033– 3412 12

‘The denouncer shall pay the fifth part […]’.

194

veksina

and [– –]ν ποιήσε̣[ι – – – ] 34114. But the fragmentary state does not allow any certain interpretation. All the instantiations of the described payment formula (hereafter referred to as ‘καταβαλεῖ formula’) occur in the instruction lists on equal terms with the other provisions, which are normally held in imperatives, sporadically interrupted by infinitives. The placement of the formula in the text can vary. Thus in 298 and 306 the καταβαλεῖ formula immediately follows the sale formula (ἀποδόσθων τοὶ πωληταὶ τὰν ἱερωσύναν), and both are found at the end of the whole text with most of the other provisions preceding. In other cases, as for example in 302 and 304, the junction of both formulas appears before the other provisions. But whatever its context, the καταβαλεῖ formula is never a kind of digression, explication or addition to the previously stated action; in none of the occurrences has the action of payment already been indicated by any other expression in the preceding text.13 Thus the discourse status of the καταβαλεῖ formula cannot be described as supplementary or secondary; on the contrary, it does not differ from the general status of most other prescriptions. Therefore the discourse conditions cannot be referred to as a possible factor in the specific morphology of this formula. The described use of the future is subject to strict thematic limitation. The same verb is in the imperative when it is used in contexts that differ thematically from the καταβαλεῖ formula, e.g., ὁ ἱερεὺς […] [λόγον ἐ]ς κιβωτὸν καταβαλ[λέτω] 34313– 14.14 Even in the inscriptions with the καταβαλεῖ formula the imperatives of the same verb are used to refer to situations involving different agents (ἱερεύς, ἐπιστάτας, στρατευόμενοι), objects (λόγον, δραχμάς), and targets (ἐς κιβωτόν, ἐς τὰ δαμόσια γράμματα, ἐς τὸν θησαυρόν): (4) (sc. ὁ ἱερεὺς) [λ]όγον δὲ ἐς κι/βωτὸν καταβαλλέτω τᾶ̣ς γ̣ε̣ν̣ομ̣ [έν]ας δ̣α̣[πά]ν̣ας ἐν̣ αἷς κα̣ ἁμέ/ραις θυσίαι συντελῶ̣ ν̣ται κατὰ τάνδε τὰν διαγραφάν· ‘The priest should put the expenditure record into the archive for all the days when the sacrifices are performed in accordance with this diagrapha’. ig xii 4 298129– 131

13 14

The fragmentary state of some of the inscriptions precludes any secure generalizations, but it can be reasonably assumed from the extant material that the general situation is so. ‘The priest should put the account […] into the archive’.

modalized future and scheduled present in coan inscriptions

(5)

195

(sc. ὁ ἐπιστάτας) […] λόγον καταβαλέσθω ἐς τὰ δαμόσια γράμματα […] καὶ ὁ ἱερεὺς λόγον καταβαλέσθω περὶ τᾶς παραλάμψιος αὐτῶν̣ […] ‘(The overseer) should put the record […] into the public archive […] and the priest should put the record of receiving those […]’. ig xii 4 31551– 52

(6) τοὶ στρατευόμενοι ἐν ταῖς μακραῖς ναυσὶν […] καταβαλλόντω τῶν γερῶν καθ’ ἑκάσταν σκα/νὰν τᾶι ἱερείαι δραχμὰς δεκαπέντε καὶ ἐς τὸν θησαυρὸν δραχμάν· ‘Those serving in warships […] shall pay 15 drachmas as the substitution for the perquisites to the priestess for each tent (-unit), and a drachma into the treasury box’. ig xii 4 3198– 915 There is also ‘the one who has purchased the priesthood’ functioning as agent, but the situation is again different and the imperative καταβαλλέτω is used: (7) ὅπως δὲ τὸ ἱερὸν τοῦ Ἀντιγόνου ἐπισκευᾶς / [τ]ε̣ καὶ ἐπιμε̣[λε]ίας τυγχάνηι, ἁ πριαμένα τὰν ἱερωσύ/[ν]αν […] καταβαλλέτω ἐπ[ὶ] / τὸν λόγον ἐπιστάτ̣[α]ι ̣ς ̣ [– – – – – – καθ’] ἕκαστον ἐνιαυ/τὸν εἰς ἐπισσκε[υὰν καὶ εἰς ἐπιμέλειαν δρα]χ̣μὰς ἑκατὸν – – – ‘In order for the sanctuary of Antigonos to be restored and maintained, the one who has bought the priesthood […] should add into her account […] 100 drachmas every year for the epistatai for restoration and maintenance’. ig xii 4 30448– 52, cf. 32668– 70

On the other hand, the following passage in the decree on the subscription for the completion of the temple of Apollo at Halasana contains all the necessary items of the established formulaic pattern ia, but has its predicate expressed in the infinitive, as have the other provisions in this decree: (8) τὰν δὲ καταβολὰν τοῦ ἀργυρίου πο/εῖσθαι (sc. τὸς δηλομένος ἐπαγγείλασσθαι εἰς ταῦτα) ἐπὶ τὸς ἄνδρας τὸς αἱρεθέντας ἐν τρισὶν / καταβολαῖς, διδόντες

15

See Parker-Obbink (2000:418, 434–438).

196

veksina

τὰ⟨ν⟩ μὲμ πράταν μετὰ τὸ / [τὰ]ν ἐπαγγελίαν γενέσθαι παραχρῆμα, τὰν δὲ ἐ/[χομ]έναν ἐν ἑξαμήνω[ι] κ̣ αὶ τὰν ἄλλαν ὡσαύτως. ‘Those willing to make pledges for this purpose should make payment of the money to the men elected in three instalments, paying the first straightaway after the making of the pledge, the next within six months and the remaining one in the same way’.16 ig xii 4 94a23– 27 (iii2)

The formal properties of this predicate differ slightly from those of the καταβαλεῖ formula (there is the double nominal argument καταβολὰν […] ἐν τρισὶν καταβολαῖς instead of καταβολὰς), but the difference in the situation and participants seems to be more essential. The situation is similar with ἀποδώσει, cf. imperatives for other agents and other situations than those of the formulaic pattern iib: (9) κατὰ ταὐτὰ τοὶ ἐπιστά/ται τοῖς ταμίαις τὸ ἀργύριον ἀποδόντω […] ‘In accordance with this the epistatai should deliver the money to the treasurers […]’. ig xii 4 29892– 93

(10) χρήματα ἀποδόντω τοὶ ἄν/δρες […] ‘The men […] should hand over the money […]’. ig xii 4 94b3– 4, cf. 27826, 2959, 33733

The καταβαλεῖ formula expresses deontic necessity (‘is obliged to pay’), though it could also be taken as expressing deontic possibility (‘is allowed to pay’).17 The construction in ia and ib is not trivial from the point of view of the aspectual composition.18 καταβαλεῖ is an achievement, whereas καταβολὰς ποι-

16 17

18

Translation of Parker-Obbink (2001:254 f.). The sums the purchasers of priesthoods had to pay were very large, and paying by several instalments could be seen as a kind of concession. Nevertheless, precisely because the prices were high it could be emphasized that the money was to be paid in due time. I differentiate between lexical aspect (Aktionsart) and aspect proper, assuming that ‘the former has to do with various properties of event predicates (e.g., telicity), whereas the latter determines how the running time of the event is related to the reference time’

modalized future and scheduled present in coan inscriptions

197

ήσει/ποιησεῖται + ‘number of instalments’ is an accomplishment.19 The complex predicate καταβολὰς ποιήσει/ποιησεῖται + ‘number of instalments’ is an example of a quantized telic predicate in the sense of Krifka (1998), with several telic unique subevents designated by the nominal arguments (καταβολάς) of the semantically deficient verb ποιέω, whereas καταβαλεῖ is telic, but not quantized. Neither action is specified grammatically as durative or punctual. Unlike the present and the past tense systems, the Greek future is deficient in the formal means available for the indication of aspect oppositions.20 Note that the almost identical complex predicate in 94a23– 27 takes the present, and not the aorist infinitive. We can conclude that in the juridical language of Cos, the modal use of the future forms ποιήσει/ποιησεῖται/καταβαλεῖ/ἀποδώσει is restricted both syntagmatically and paradigmatically to one formulaic pattern employed only on one particular occasion, viz. to set the conditions on how the payment for a purchased priesthood should be delivered. As is obvious from the comparison with 94a23– 27, a close structural similarity and lexical crossovers do not presuppose the use of the same grammatical form. For the καταβαλεῖ formula, no internal structural motivation for the choice of the future indicative over the imperative or the provisional infinitive has been established. It can be speculated, whether, since the future indicative is the form of the commissives, this meaning is not subsumed in the καταβαλεῖ formula: the persons in question are not just ordered to pay according to the schema, they are invited to assume the obligation to pay in accordance with the schema. Does that mean that the 3rd person future indicative represents a transformed formula of the act of commitment? Remaining examples of the modal future indicative in the Coan corpus are scarce: i.

[… πᾶσα]ν πρόνοιαν ποιήσει […] 29029 ‘[…] shall take extreme care […]’.

19 20

(Ivanov-Tatevosov 2009:852). See Brown (2006:i.537 f.) and Tatevosov (2002). There is no uniform taxonomy for describing these phenomena, neither in general linguistics nor in particular for Greek. For an overview of terminologies on lexical aspect see Tatevosov (2002). As for Greek aspect, Duhoux (2000:138ff.) distinguishes between durative, punctual, and state aspects expressed morphologically. For other terminologies, see Tatevosov (2002:320 f.). Duhoux (2000:143, 145).

198

veksina

The fragmentary state of this citation leaves open whether it is an independent clause and the properties of the predicate remain unclear. ii.

[τελ]έσει δὲ τὸ[ν ἱερῆ ἁ πόλις·] 31160 ‘The priest is to be consecrated by the city’.

This is, according to the restoration, a complete sentence. The initiation was a ritual; as such it was not momentaneous and included several different actions.21 But as long as we are dealing with the linguistic conceptualization, nothing indicates that τελέσει should be taken to be anything other than a telic predicate, which is neither quantized nor cumulative.22 It is a complete action that is prescribed, and not just a process of its fulfillment. The same is indicated by the use of the aorist in the infinitive and imperative, cf. ἀπομισθωσάντω δὲ καὶ] τ̣ελέσαι τὰν [ἱέρ]ειαν […] 31947, [… μισθωσάντω τελέσ]α̣ι τὸν ἱερῆ […] 31810, and τὰν δὲ ἱερῆ τε̣[λε/σά]τω ἁ πόλις 30415– 16. iii. […] δὶς ὀφηλεῖ τῶι Ἀπόλλω[νι …] 3359– 10 ‘[…] shall owe twice as much to Apollo […]’. This is a state predicate with the lexical meaning of deontic necessity. The future apparently provides here for the ‘inception of state’ meaning. For the present imperative see also the following example: [αἰ δέ κα μὴ ἀποδῶντι, ὀφ]ειλόντω διπλάσια […] 35017 ‘If they do not hand over, they should owe twice as much […]’. iv. In terms of modality all the modal futures analyzed so far express deontic necessity. There is one instance of a future form expressing deontic possibility where a permission to economically use the sanctuary is given to a priest:

21 22

See Parker-Obbink (2000:446) and Wiemer (2003:280). From a more sophisticated perspective it can be described as consisting of two subevents, viz., ‘become subevent’ (relating to the patient) and ‘activity subevent’ (relating both to the agent and the patient), of which only the latter can be incremental, see Ivanov-Tatevosov (2009:106 ff.).

modalized future and scheduled present in coan inscriptions

199

καρπεύσεται δὲ / καὶ τὸ τέμενος καθότι καὶ ὁ πρότερον ἱερεὺς ἐ/καρπεύετο 32818– 20 ‘She is allowed to economically use the sanctuary the way the previous priest did’. This predicate corresponds to Vendler’s ‘activity’; it is atelic and not-quantized. The durative character of the activity is also indicated by the imperfect ἐ/καρπεύετο. In two other cases permissions of the same kind are granted in the present imperative (3109– 10 and 31943– 45). To conclude, no certain rules have been established dictating the choice of the future indicative as opposed to the imperative or the provisional infinitive. A great majority of cases are constituted by one bureaucratic cliché used in one particular type of text (priesthood sale regulations), but the modal futures are also found beyond this discourse and involve different predicates. The aspectual properties of the predicates vary. The modality is restricted to the deontic realm comprising both necessity and possibility; that the future forms are mainly used to express obligations and only exceptionally express permissions reflects the needs of the legislative texts. The jussive force of the provisions expressed with future forms is presumably neither weaker nor stronger than that of the other provisional expressions. There is the possibility that the formulation with the future indicative in the priesthood sale regulations is an echo of a commissive formula. 9.2.2 Parallels beyond the Coan Corpus Since syntax and grammatical semantics are not usually treated in studies of the Greek dialects and inscriptions, and instances of the future indicative used in independent clauses to express deontic modality, have not, to my knowledge, been gathered anywhere, what follows is only an illustrative collection of such examples. The situation with the Gortynian inscriptions of the first part of the 5th c. bce exhibits striking similarities with the Coan usage.23 Along with regular provisional infinitives and imperatives, deontic futures sometimes occur. Most of these instances are thematically and lexically homogeneous: the vast majority of the examples is confined to one particular verb, καταστασεῖ, stating that (in a certain case) someone will have to provide some financial compensation. There are fourteen instances of καταστασεῖ in the Gortynian Law

23

See Bile (1988:254).

200

veksina

(ic iv 72),24 to which two occurrences of ἐπικαταστασεῖ should be added (ic iv 72:i 48 and ix 15).25 Further attestations of καταστασεῖ in Gortys are the following: ic iv 73:a 5; ic iv 43:A.a. 6–7 and A.b. 6–7; ic iv 46:b 4 and 11–12; ic iv 47:25–26 and 30–31; ic iv 41:ii 2, iii 6–7, iv 5, v 3.26 Yet the use of the deontic future was not lexically restricted. In ic iv 41, in addition to the regular καταστασεῖ, we also find a synonymous τεισῆται (ic iv 41: i 7), also used in other parts of Creta. The modal use of the future was not thematically restricted either. The future forms state, for instance, a legal solution in case someone injures a free man, pawned for his debts. Thus in the same ic iv 41 the futures μο̄λησεῖ and πρακσῆται τὰς τιμάνς denote initiating a lawsuit and exacting compensation: αἰ δέ τις / [τ]ὸν κατακείμεν/ον ἀδικήσει, ὀ κατ/αθέμενος μο̄λησε/ῖ καὶ πρακσῆται τ/ὰς τιμὰνς ἆι ἐλευθέ/ρο̄ ic iv 41:vi 2–8. The clause states the only possible legal way of defence in such a case. The choice between deontic necessity and possibility is not obvious in ic iv 47:10–16: αἰ δέ κα τὸν κατακεί/μενον ἀδικήσει ἄλλος, αἰ μὲν / κ’ ἀνπότεροι μο̄λίοντες νικά/σοντι, τὰν ἠμίναν ϝεκάτερο/ς ἐκσίοντι· αἰ δέ κ’ ὀ ἄτερος μ/ὴ λῆι, ὀ ἄτερος μο̄λίο̄ν αἴ κα νι/κάσει αὐτὸς ἐκσεῖ. Here ἐκσίοντι and ἐκσεῖ, with their remarkable inceptive meaning ‘shall receive’ (state predicate with future markers) denote legal outcomes of a case.27 The subject of the predicate ἐκσίοντι / ἐκσεῖ is the recipient and is thus not agentive. Since the situation also involves an agent clearly inferable from the context (aggressor), the modal interpretation of the predicate becomes ambiguous between the deontic necessity (obligation) imposed on the agent, and the deontic possibility open to the patient.28 But this is an infrequent case. The main domain of the modal futures in Cretan was deontic necessity. It remains unclear what motivates the choice of the future forms as opposed to the imperatives and provisional infinitives.29

24 25

26 27 28 29

For the editions of the non-Coan inscriptions I follow the abbreviations of the seg. However, I use the more common ic for their I.Cret. and lss for lscg Suppl. Bile (1988:254) counted thirteen examples and did not mention ἐπικαταστασεῖ. In addition, there are two infinitives in this inscription, καταστᾶσαι vi 42, κατιστάμεν i 54, and one imperative κατισ/[τάτ]ο̄ in i 45–46. For more Cretan examples, see Villaro (1998:192 f.). See the above-mentioned example, δὶς ὀφηλεῖ τῶι Ἀπόλλω[νι] 335 9– 10. See Van der Auwera-Plungian (1998:99 ff.) for such ambiguous and/or transitional cases. The Cretan modal future has been variously approached. For καταστασεῖ Bile (1988:254) suggests the following explanation: ‘Le “il paiera” … ne laisse pas de doute sur la réalisation prochaine de l’ acte’. This suggestion remains ungrounded. An unquestioning character of an obligation imposed or the confidence in its fulfillment should not necessarily combine with the idea of the nearest future. Villaro (1998:189ff.) attempts to explain the future

modalized future and scheduled present in coan inscriptions

201

As far as the archaic corpora are concerned, there is one Elean inscription of the late 6th c. bce yielding two future indicatives: [– – – τοὶ μ]/ιαντε̑ρες ἐνοισέονται καὶ τούτοι ὐπασχε̄σέον[ται – – ] (Minon 2007:nr. 5, lines 2–3). Despite the fragmentary context, the deontic necessity reading is very probable, see Minon (2007:454). In the Heraclean Tablets (late 4th–early 3rd c. bce), the prevailing provisions for the rent contract of Dionysus’ sanctuary estate (Συνθήκα Διονύσω χώρων i 94– 18730) are made in the future, interchangably with provisional infinitives (e.g., in lines 152–158 as if this were a citation from some other document)31 and sporadic imperatives (lines 127, 171), which could be due to the Koine influence. Again, the main realm of this future is deontic necessity, but sometimes it also indicates deontic possibility, e.g., ἐς δὲ τὰ ἐποίκια χρησόνται ξύλοις ἐς τὰν οἰκοδομὰν ℎοῖς κα δήλωνται καὶ ἐς τὰς / ἀμπέλως· τῶν δὲ ξηρῶν κοψόντι, ℎόσσα αὐτοῖς ποτ’ οἰκίαν ἐς χρείαν· i 146f. In the Olbian decree, iospe i 24, = I.Kalchedon 16 = igdop 14 (early 4th c. bce), the future indicative is used in vindicatory parts, the subjects being both infringers (patientive), [φευ]ξεῖται, line 11, στερήσεται, line 17, and those acting in the interests of the law (agentive), πράξονται, line 19.32 The expression of deontic modality with future indicatives is very widespread in the funerary formulas of imperial times. Their usual topic is prohibiting violation of the tombs, in particular unauthorized burial, and imposing sanctions on violators.33 We often find futures used with a negation, and it is often οὐ, as the following examples from Ephesos demonstrate: οὐδεμιᾷ δὲ παρευρέσει πονηρῇ τε/θήσεται οὐδεὶς ἄλλος ἐν τῇ σορῷ I.Eph 2211b: 3–4; οὐδενὶ […] ἐξέσται et sim. I.Eph 2265, 2299b, 2304, 2311, 2435, 2446, 2476, 3216 etc. (but see μή in the above example from Pantikapaion). The inscriptions from Ephesos also exhibit indicative, subjunctive and imperative in such contexts.34 These cases are not trivial from the point of view of semantics: they negate deontic possibility by stating deontic necessity. For other regions, see I.Didyma 529, mama 8, nos. 544, 545, 552, 553 etc. (Aphrodisias), tam ii 55 (Telmessus, Lycia).35 For

30 31 32 33 34 35

as a form of an agreement or compromise, and not an obligation. However, except for the contracts where the source of authorization does not differ from the agent (Villaro 1998:195), I find this interpretation difficult to reconcile with the texts. The references are made according to the edition of Uguzzoni-Ghinatti (1968). Cf. also the negation μή with these infinitives and the negation οὐ with the future. In this inscription the future interchanges with imperatives and infinitives. For further references, see McLean (2002:271 ff.). Ortuño (1999:255–256). For regions outside Asia Minor, see Pantikapaion iospe iv 342 = cirb 731 (3rd c. ce?),

202

veksina

sanction, see εἰ δέ τις τολμήσι, δώσι εἰς τὸ ἱερώ/ατον ταμῆον […] I.Eph 2265. Funerary inscriptions do not confine their use of the deontic future to prohibitions and punishments for tomb violators, as the following examples illustrate: ποιή/{σου}σουσιν δὲ τὴν εὐωχίαν μη(νὸς) Ποσ(ιδεῶνος) ἡ(μέρᾳ) έ / ἀπι(όντος)· εἰ δὲ μὴ ποιήσουσι, ἀλὰ παραπένψονται, ἐκδικήσου/σιν καὶ ἐκπράξουσιν οἱ ἐν Ἐφέσῳ προμέτραι I.Eph 3216; see also I.Smyrna 223b: 8–9 with the future found in the provision to place the copy of the inscription into the archive. Modal futures are quite widespread in later times and their use is not confined to funerary discourse. For the expression of deontic necessity in an honorary decree, cf. τεθήσονται δὲ ἐν τῷ εὐθέτῳ τό/πῳ ἐν τῷ δημοσίῳ I.Eph 4337: 22–23 (after imperatival infinitive). For further prescriptions, cf. δαπανήσουσιν I.Eph 27: 496, δώσει I.Eph 2226: 3. The number of examples can easily be multiplied. The evidence provided allows the following conclusions. The non-epistemic modal future was in use in various dialectal sub-groups from archaic to later times. Its semantic scope comprised both deontic necessity and possibility. There is so far no evidence to demonstrate that this usage was not specific to juridical language. The imperatival future was denied by Magnien (1912:171f.) for Classical Greek. In fact, some literary passages mentioned in the grammars express strong assurance on the part of the speaker that something will (not) take place, rather than impose obligations, grant permissions, call to action or prohibit action.36 However, participant-external, probably deontic, possibility seems to be expressed by the future form in Soph. oc 956. Neither deontic necessity nor imperatival force can be denied for ὣς οὖν ποιήσετε καὶ πείθεσθέ μοι Pl. Prt. 338a, where, according to Dover (ad Aristoph. Nub. 1352), ‘ὥς points to an archaic formula’. This phrase can in fact reflect dialectal, in particular Elean, phrasing, since it is articulated by Hippias. In Classical Greek, futures could be used with jussive force in questions. This is a distinct construction where the imperatival force is due to the combination of the future form and the question status of the sentence. The modal semantics of these futures is secondary to their performative status: a speaker acting as an authority imposes a requirement on an addressee. There are several examples of this construction in Attic comedy, all of a rough speech style

36

ἐνθάδε κατοικῶν/τά με μηδείς μοι παρυβρίσει μη/δέ μοί τις σκυλεῖ τὰ ὀστέα παρυβρίσει 7– 8. See e.g., Aristoph. Nub. 1352, mentioned by Rijksbaron (2006:34); Xen. An. 1.3.5, mentioned by Schwyzer-Debrunner (1950:291); and Aristoph. Vesp. 671. No modal or imperatival meaning is perceptible in Lys. 29.13, see Schwyzer (ibid.).

modalized future and scheduled present in coan inscriptions

203

(Aristoph. Nub. 633, 1299; Av. 1572; Pax 259). A different case is the jussive force of the questions with negated future forms (Aristoph. Nub. 1296, 1298).37 9.2.3 Modalized Future That indicatives, and in particular future indicatives, can be used performatively to express deontic modality is an issue that has already received due attention both in theoretical and cross-linguistic studies. Forms lacking imperative marking can be and are used the way imperatives are. See the following example: ‘You’re gonna take off your shoes before you come in here!’ (Bybee et al. 1994:211). There are two dimensions to this problem: a) to what extent is the imperatival interpretation pragmatically inferred? b) to what extent is it grammatically grounded? Turning to the Coan material, the puzzles can be reformulated as follows: a) To what extent is the deontic interpretation of futures entailed by the discursive properties of the texts? Simply put, the readers of the inscriptions are presupposed to know that the state decisions are to be followed, even if the literal form of these decisions might perhaps suggest other possibilities. ‘In a situation in which the speaker has authority over the addressee, a prediction about the addressee can be interpreted as a command’, as Bybee et al. (1994:211) put it. b) What properties of a given syntactic construction actually allow such use? Or, in what way do the grammatical features contribute to the so-called ‘illocutionary potential’ of a sentence, to use Truckenbrodt’s term (2006:260)? After all it is not indicative imperfects that are used in the imperatival function, but indicative futures. While the problems of performativity go beyond the scope of this paper, the aspect of grammatical semantics should and will be addressed here. From the typological point of view, the Greek data collected in this paper exhibit nothing extraordinary. In their cross-linguistic study, Bybee et al. (1994: 210f., 273) report about thirteen languages that use future grammatical markers, or ‘primary futures’, to impose a requirement. It is a pity that they only take into consideration 2nd person forms as markers of deontic necessity used performatively, so that we lack data on the 3rd person forms used in this way, data on the expression of deontic possibility (permission) with primary futures, and data on the non-performative deontic use of futures. Nevertheless their results bear directly on this study since it is clear that deontic necessity is only one subcategory in the modality paradigm, and that in the discourse of the legal deeds the 3rd person expresses the addressee. Even more important is the

37

Herewith belong Aristoph. Vesp. 397 and Pl. Symp. 175a, quoted by Rijksbaron (2006:34). For further examples see Kühner-Gerth (1904:§ 387.7).

204

veksina

figure 9.1 Modalization of future

observation that the ‘imperative is the most commonly occurring other use for futures’ and ‘the futures that are used as imperatives in all other respects have the properties of primary futures’ (Bybee et al. 1994:273). Prototypical future meanings reside close to the domain of participant-external modality, and in fact overlap with it, as has often been pointed out in the literature on modals (see such notions as non-factuality, non-actuality, potentiality, nonveridicality).38 The semantic shift from future-projecting meaning to deontic modality is thus cross-linguistically grounded. As far as Ancient Greek is concerned, the data gathered in sections 9.2.1 and 9.2.2 demonstrate that the modalization of future leads to the unified domain of deontic modalities (both necessity and possibility). Presumably, both deontic and non-deontic participant-external values were accessible to the future forms (fig. 9.1). The issue of the cross-over between modality and future tense can find a broader perspective within the history of Greek. As is well known, the future markers in the languages of the world can also come to express epistemic modality.39 In Modern Greek the future forms can express epistemic necessity if the proposition has present or past reference.40 See examples from Giannakidou-Mari: I Ariadne tha ine arosti, ala dhen ime ke endelos sigouri, ‘Ariadne must be sick, but I am not entirely sure’, and I Ariadne tha milise xthes, ‘Ariadne must have spoken yesterday.’ This modern epistemic usage could develop directly from the pre-modal future, not through the modal

38

39 40

The paths of development between the future and non-epistemic modality are in fact bidirectional, cf. ‘ability and obligation as precursors or antecedents of action’ and ‘any non-epistemic modality will serve a secondary function of expressing future time’ (Brown 2006:viii.263); see Asher (1994:2540); Bybee et al. (1994:184); Van der Auwera-Plungian (1998:111); and Rijksbaron (2006:33ff.). See e.g., Van der Auwera-Plungian (1998:96 ff.). For Spanish, see Bybee et al. (1994, 224f.). As well as in Italian.

modalized future and scheduled present in coan inscriptions

205

use of the future forms described in this paper; both usages have always remained secondary.

9.3

Indicative Present

9.3.1 Coan Data The present indicative is the verbal form par excellence of the Coan sacral calendars. It is not the only mood used in these calendars; sometimes there also occur imperatives and infinitives, but their occurrences are sporadic. By contrast, the calendars consistently make use of what can be called ‘nominal style’: it includes non-verbal clauses, e.g.: (11) τούτων οὐκ ἐκφορὰ ἐκ τοῦ ναοῦ ‘Of these (offerings) no carrying out of the temple’. ig xii 4 27410

(12) [ἐ]πὶ̣ τούτων ἑκάστωι ἱερὰ οὐλομέτ/[ριο]ν̣, ἡμίεκτον ἑκατέρων, καὶ κύλικες καιναὶ τρεῖς ἑ/[κάσ]τ̣ωι καὶ πίναξ ἑκάστωι· ‘As additional sacrifice for each (heros): a portion of barley-meal, a halfhekteus of both (wheat and barley), as well as three new cups for each (heros) and a plate for each (heros)’. ig xii 4 2765– 7

(13) τὰ δὲ ἄλλα κρέα τᾶς πόλιο̣ς. ‘The rest of the meat (is to be the property) of the polis’. ig xii 4 27823

Nominal style also includes list structures, normally prescribing offerings in nominatives or accusatives without an overt verb like θύει or θύεται, e.g.: (14) δυωδεκ/άται· Ζηνὶ Μαχανῆι οἶες τρεῖς τέλεωι καὶ βοῦς ὁ κριθεὶς τὸ / ἅτερον ἔτος ἐφ’ οὗ κα ἔωντι Καρνεῖαι, τὸ δὲ ἅτερον ἔτος οἶες τ̣/ρεῖς τέλεωι· ‘On the twelfth: for Zeus Machaneus three full-grown male sheep and one choice ox every other year (in the year) when Karneiai are performed, and the other year three full-grown male sheep’. ig xii 4 27413– 16

206

veksina

table 9.1

Verbal moods and nominal style in Coan calenders

id

Date

274 state 275 state 276 state 277 state 278 state 279 demos 280 demos 281 gymnasium

mid iv bce mid iv bce mid iv bce mid iv bce mid iv bce iii² bce ii1 bce 158–138bce

ind.pres

Nominal

imp

inf

+ + +

+ + + + + + + +

+

+

+ + + +

+ +

Table 9.1 shows the distribution of the forms in the preserved Coan calendars.41 The indicative present is used unfailingly in verbs constituting the core set of the verbal expressions of the state sacral calendars: θύει, θύεται, παρέχει, δίδοται, λαμβάνει, ἐνδέρεται. These are of course not the only indicative verbs important in the composition of calendar provisions. I have excluded from this list indicatives that are (almost) singular and verbs recurring both in the indicative and imperative. The indicative present unquestionably dominates in the Coan calendars, involving numerous verbs not belonging to the established set of ‘core expressions’. Illustrative cases are two calendars that have a structure quite different from that of the state fasti, and that in particular lack the ‘core expressions’ established above. These are the fasti of Phyxiotai, 279, and the gymnasial fasti, 281, both written in an austere list form. In 279 the sacrificial victims are inconsistently listed in nominatives and accusatives, so that θύει or θύεται could easily be supplied but are never explicitly mentioned. Verbs are only used here when it is necessary to indicate any other action than ‘to sacrifice’, and are unexceptionally present indicatives: ἀγγέλεται 27914, πορεύονται 27939, 54, 56, and κρίνει 27940. The gymnasial fasti, 281, contain the list of activities expressed by nouns: θυσία 2812, 28130, πομπὴ 2813– 4, 28112, 28127, 28138, 28140, 28147, ἀγωνάριον 2817, 28125, 28131, 28135, 28142, δό[σις] 28116– 17, and ἀπόδειξις 28143. This preference for nominal constructions is striking, since all of these nouns have lexically

41

The fragmentary 282 does not bear on this investigation and is not taken into account here.

modalized future and scheduled present in coan inscriptions

207

equivalent verbs that could presumably stay in their place without any difference in meaning. There is no obvious grammatical motive for the occurrence of the single verbal form [οἱ παῖ]δ̣ες / [τρέ]χ̣ουσιν 2815– 6, but in the same month a [διαδρ]ομὰ / [παίδω]ν 28115– 16 takes place, and this is a different kind of contest. Thus the choice of the verbal form will have been motivated by a wish to differentiate between the two activities. The indicative present is not the only verbal mood used for prescriptions in the Coan calendars, although its dominance is challenged (but unsubstantially) by the imperative and the infinitive. Generally we can only speak about an occasional deviation. In the Coan calendars, the provisional infinitive occurs only once after a series of indicatives, non-verbal syntagmas and one imperative: (15) τοῖ[ς ̣ δὲ Καρνεί]οι ̣ς τὰμ πόλιν ὠνεῖσθαι δάμ̣ [α/λιν κριτὰν μὴ ἐλάσσονος πεντήκοντα] δραχ̣μᾶν· ‘At/for the Karneioi the polis should buy a choice heifer worth no less than 50 drachmas’. ig xii 4 27426– 27

The preceding sentence is in the indicative, whereas the following context is lost, since the inscription breaks off after line 27. A combination of the subject πόλις with the predicate of ὠνέομαι is not found elsewhere in the Coan calendars. The only imperatival form in the extant text of this inscription occurs in a provision concerning, again, the price of a victim: (16) Ἥραι Ἀργείαι Ἑλείαι Βασιλείαι δάμ/αλις κριτά· κρινέσθω δὲ μὴ ἐλάσσονος ⟦ον⟧ὠνημένα πεντ/ήκοντα δραχμᾶν· ‘For Hera Argeia Heleia Basileia a choice heifer: a heifer at a cost of no less than 50 drachmas should be chosen’. ig xii 4 2745– 7

Note the special status of this expression in the discourse: it is an extension of the previous provision (δάμαλις κριτά· κρινέσθω δὲ …). It is reasonable to suggest that both digressions from the default indicative indicate additions introduced into the original redaction. It is the indicative that is used in the provisions lacking any indication of a victim price in this calendar: (17) Ζηνὶ Μαχα/νῆι βοῦς κρίνεται τὸ ἅτερον ἔτος […]

208

veksina

‘For Zeus Machaneus an ox is (to be) chosen every other year’. ig xii 4 27410– 1142 The need to indicate the value of a sacrificial heifer could have arisen in later times. The choice between infinitive and imperative remains unexplained, in particular because the infinitive does not occur anywhere else in the calendars; it can be speculated that neither addition was contemporary. The imperative is somewhat more prominent. As we have already seen (exx. 16–17), there are verbs recurring both in the imperative and in the indicative within one and the same inscription. An extensive alternation of imperatives and indicatives is characteristic of 278. There are the following ‘doublets’: ἐλάντω 2788, 27812, 27814 vs. ἐλᾶντι 2786, 27828, ἐπελάντω 27811, 27815, 27816 vs. ἐπελᾶντ[ι] 27819, ἐπισπενδέτω 27836 vs. [ἐπι]/σπένδει 27829– 30, καρυσσέτω 27836 vs. καρύσσοντι 27832, cf. προκαρύσσεται 27413, ἐπιθυέτ̣[ω] 27837 vs. ἐπιθύει 27821. The context does not indicate any meaningful contrast between the two kinds of forms. They are found juxtaposed: ἐς δὲ / [τ]ὰν ἀγο̣ρὰν ἐλάντω Πάμφ̣υλοι πρᾶτοι, ἐν ἀγορᾶι δὲ συμμίσ̣ /[γ]ο̣ν̣ται 2787– 9, cf. ἐλά̣[σα]/ντες δὲ τούτους ̣ σ̣ υμμίσγο̣ν[ται το]ῖς̣ ἄλλοις 27817– 18. Despite the ample alternation of both forms in this inscription, the ‘core set’ of verbal expressions remains unchallenged. The same is true of the demos calendar 280 where imperatives are only used in other verbs: φερέτω 28027 (cf. indicative in 27847), [ἔ]στω 28044, but its fragmentary state does not allow any far-reaching conclusions. In accordance with what has already been observed in previous sections, deontic necessity is the main modality value in the calendars, whereas deontic possibility is a less usual case. Still, the priest/priestess’s privilege in collecting the hide and a leg from sacrificial victims relates to the ‘core set’ indicatives: (18) [γ]/έρη λαμβάνει δέρμα ‘As perquisites he receives the skin’. ig xii 4 2744– 5

(19) γέρη̣ λαμβά[νει] / δέρμα καὶ σκέλος· 42

Compare αἰ δέ κα τούτωγ κριθ̣[ῆι] / μ̣ ηδείς, ἐπικρίνοντα[ι] β̣οῦν ἐκ χι[λια]στύος ἑκάστας ‘if none of those has been chosen, they additionally pick out an ox from every chiliastys’ (27816– 17); ἐλά̣[σα]/ντες δὲ τούτους ̣ σ̣ υμμίσγο̣ν[ται το]ῖς̣ ἄλλοις καὶ εὐθὺ[ς κρίν]/οντι καὶ εὔχονται καὶ ἀποκαρύ[σσο]ντι ‘having driven these (oxen) (into the agora) they mix them with the other ones and right away they pick out (an ox) and pray and announce’ (27817– 19).

modalized future and scheduled present in coan inscriptions

209

‘As perquisites he receives the skin and a leg’. ig xii 4 2747– 8

This λαμβάνει is similar with δίδοται expressing the analogous privilege for other recipients: (20) γέρη̣ δ̣ὲ Φυλεομ/αχίδαις δίδοται τοῦ βοὸς ὁπλά, τ̣α̣[ρ]σ̣ ο̣ς́ , τῶν δὲ οἴων̣ τ̣ὸ ὠμ̣ όν̣ ‘As perquisites Phylomachidai receive a hoof and a sole from the ox, and from the male sheep the (piece of the) raw meat/shoulder’. ig xii 4 27418– 19

The only infinitive (27426– 27) expresses deontic necessity. Leaving aside the uncertain φερέτω (28027), which can express prerogative (the context is very fragmentary), there is no permissive imperative in the calendars, though this meaning was not alien to the imperative, cf. in the provisions for the cult of Hermes Enagonios: (21) γέρη δὲ λαμβανέτω τῶν θυομένων τῶι / Ἑρμᾶι τῶι Ἐναγωνίωι δέρμα καὶ σκέλος ‘Of the sacrificed for Hermes Enagonios he is to receive as perquisites the skin and a leg’. ig xii 4 29811– 12

Non-verbal constructions are characterized in terms of modality no less than verbal, a distinct case of deontic possibility being the permission to take away the sacrificial meat: ταύτας ἀποφορά 2748, cf. 2752. Negated possibility turns out to be necessity: τοῦ χοίρου οὐκ ἀποφορά 27846, cf. 27859, 27861, 27863, 27939, 27954, 27956. In terms of modality it is not possible to formulate any conditions of the interchange of indicative/infinitive/imperative/non-verbal constructions. The only discernible factor in the mood variation in the calendars that has been revealed so far is the pressure of the formulaic stereotypes. In its strongest form, this refers to a closed set of lexemes that do not occur in any form other than the indicative present. In a looser form it means the general preference for the indicative present in the calendar discourse along with the secondary character of the other moods.43 Of course even these generalizations are fragile and subject to the state of the evidence available to us.

43

There is also a closed set of recurring non-verbal syntagmas, for which no equivalent

210

veksina

There is evidence that the established formulaic stereotypes were not confined to the genre of the calendars. In the rest of the Coan corpus the occurrences of the indicative present are exceptional. Mostly they are found in the inscriptions containing various sacral regulations. In a great number of cases the ‘core expressions’ are reproduced: θύει 7234,44 3322, 6, 23, 28, 3585, [γέρη δὲ / λαμβάνε]ι δέρμα κ[αὶ σκέλος] 30714– 15, θύεται 33213, 26, 27, 53, 3591, παρέχει 3329, 25, 30, ἐνδέρεται 33227, cf. 2748, 27848, [δί]/δοται 33232– 33, cf. 27852, 27419, 27611. The sacral law 332 is further characterized by the extensive use of the indicative and adherence to the calendar formulas. Some indicatives falling outside the ‘core set’ of formulas nevertheless have parallels in the extant calendar texts, e.g., ἐπισπένδε̣[ι] 33229, cf. 27830; others do not, e.g., καθαίρεται 332122– 23, διαιρεῖται 35814. The periodic character of the prescribed actions is explicit in 358 with monthly date reference (ll. 6–8); no information is available for 359, which should be a supplement to some other document; 307 is likely to express periodicity, cf. [καθ’ / ἕκαστον ἐ]νιαυτόν 30716– 17; 332 contains a date specification in lines 61–62, but otherwise the context is too fragmentary to allow for any certainty. In 364 the non-calendar indicative present στεφανοῦντι is used for singular unique actions that are to be fulfilled on certain dates. One special case is 298 (iii2 bce), listing provisions concerning the priesthood of Hermes Enagonios. Here the strict formulaic prototypes are neglected, cf. the imperative in γέρη δὲ λαμβανέτω τῶν θυομένων τῶι / Ἑρμᾶι τῶι Ἐναγωνίωι δέρμα καὶ σκέλος 29811– 12. The imperative θυέτω/θυόντω is used here throughout for the prescription of periodic sacrifices, and in most sentences where it occurs there are explicit references to monthly dates. The single indicative present ἁ θυσία γί/νεται τοῦ Ὑακινθίου μ̣ η̣ν̣ὸς τρ̣ια̣ κ̣ α̣δ́ ̣ι 298120– 121 is therefore striking. Since the inscription violates the grammatical stereotype of the calendar discourse in the ‘core set’ of expressions, it would be strange if it followed it in an expression that does not have a direct parallel in the calendars (θυσία γίνεται). This sentence contains the time indication with reference to a month, but so do most other imperatival sentences in this inscription. It therefore seems altogether unlikely that this indicative is a kind of grammatical allusion to the calendar discourse. The context, however fragmentary, allows an interpretation of the sentence with θυσία γίνεται as an explication of the provision already given, that is, providing the date of the sacrifice: [θυσ]ά[τω ἀπὸ] τοῦ / διδομένου ἀργυρίου ὃ κατ̣α̣– – – – – – – –ασσ– – ἁ θυσία γί/νεται τοῦ Ὑακινθίου μ̣ η̣ν̣ὸς τρ̣ι ̣ακ̣ α̣δ́ ̣ι·

44

verbal syntagmas are ever found, and which are likely to pertain to the same stratum as the ‘core indicatives’. Or [ἐπι]θύει according to Herzog’s restoration.

modalized future and scheduled present in coan inscriptions

211

vac. 298119– 121. It can be tentatively assumed that here the shift from the default mood correlates with the complementary status of the sentence. Cf. ex. 16 for a choice of the imperative instead of the default indicative in similar discourse conditions. 9.3.2 Parallels beyond the Coan Corpus It is uncommon for the officially published Greek calendars to be formulated in indicatives. Where an austere list form is not upheld, we mostly find infinitives and/or imperatives. Nevertheless, the indicative is found in some calendars from different parts of Greece (in chronological order): i. lsam 41 = Milet vi 1, 3, 31a (Milet, end of the 6th c. bce) only uses indicatives: [πρ]οδόρπ̣ ια δίδοτ̣αι line 1, δωδεκάτηι ⋮ ἐς βασιλέως δίδοται line 3, [τρίτηι ἐπ]|ὶ δέκα […] χοῦς τῇ ἰερῆι δίδοται […] ἐς το̑ ἰερέως ⋮ δίδοται χοῦς ξύλα lines 6–8, [ἔκ]|τηι ἰσταμένου ⋮ ἐορτὴ κηρύσσεται line 12. ii. lss 10 (Athens, between 403 and 399bce) in the heading of the list: τάδε τὸ ἕτερον ἔτος θύεται line 30. iii. ig ii/iii2 1358 = lscg 20 (Marathon and Tricorynthos, 4th c. bce) in the headings of the lists: [τάδε ὁ ἄρχων θύ]ει i 11, [τάδε ὁ δήμαρχος/ ὁ Μα]ραθωνίων θύει b 1–2, τάδε ὁ δήμαρχος ὁ Μαραθωνίων θύει ii 23, τάδε τὸ ἕτερον ἔτος θύεται μετὰ Εὔβουλον ἄρχ[ο]/ντα Τετραπολεῦσι ii 39–40. These three inscriptions, the oldest in my list, adhere to the core set of calendar expressions as it was established for Cos. The indicative is the only verbal mood in these inscriptions, thus the same grammatical pattern is valid here too. By contrast, the later inscriptions only contain indicatives side by side with other moods: iv. ig ii/iii2 1363 = lscg 7 (demos Eleusis, late 4th c. bce): πένπτει ἱσταμένου […] τὴν ἑορτὴν / προαγορεύουσιν τῶν / Προηροσίων lines 2–6. For infinitives in this calendar, cf.: τράπεζαν κοσμῆσαι line 11, and παρέχειν σπονδ[άς] line 17. v. lscg 96 (Mykonos, 3rd c. bce): [τάδε ἔδοξεν […] θύειν] ὁ κριὸς εἰς πόλιν [ο]ὐ/κ εἰσάγεται· νῶτογ καὶ πλάτη κόπτεται· ἡ πλάτη σπένδε/ται lines 6–8, νῶτογ κόπτετα[ι] 12, τοῦτο ἐνα/τεύεται 23–24, νῶτον / τοῦ ταύρου κόπτεται· τῶι ἱερεῖ τοῦ ταύρου δίδοται γλῶσ/σα καὶ βραχίων 30–32, πρὸς τῶι β[ωμ]ῶι σ[φάτ]τετ[αι] 36. Here the actual prescriptions are given in the list form (nominatives) and imperatives, whereas the indicative is used in passages introduced to make

212

veksina

specifications about the already indicated victims; thus the special discourse status of these digressions correlates with the change from the default verbal mood, imperative, to the indicative. It can be concluded that the use of the present indicative in the calendar provisions is neither dialectally nor chronologically restricted to the Coan corpus. 9.3.3 Scheduled Present We have already seen that deontic modality and indicative can go together. As far as the future indicative is concerned, the non-factual character of any predication made of future events is a crucial factor for the semantic shift into the domain of modality. The conditions are more complicated with the present indicative. A sentence like [Γεραστ]ίου [date] […] [οἱ παῖ]δ̣ες / [τρέ]χ̣ουσιν 2811– 6 does not refer to a particular event; rather, it reports that on a particular day of the month Gerastios (due to the lacuna we do not know exactly which day) a running competition of the paides normally takes place. The sentence informs us that running on a certain day of Gerastios is one of the paides’ habits of conduct, but no concrete event of the paides’ running is either prescribed or referred to. As a logician would say, this sentence is intentional in its semantics. Linguists use such designations as ‘habitual’, ‘generic’, ‘non-episodic’, ‘customary’, ‘usitative’, ‘gnomic’, and ‘frequentative’, of which ‘generic’ and ‘habitual’ are the commonest. Generic sentences are a widely discussed topic in modern linguistics, and in particular in the domains of formal semantics and the study of aspect phenomena.45 To outline the fundamental properties of such sentences, I will adopt the approach developed by Carlson (2008). It has been assumed that sentences expressing ‘habits, broad generalizations which are regularly instantiated […] propensities, dispositions, abilities, laws, rules of games, customs, social norms, job descriptions, functions and other notions’ (Carlson 2008:20), are characterized by some common semantics and these semantics can be subsumed under the notion of pattern. In particular, ‘generic and habitual sentences express a pattern of occurrences of events’ (Carlson 2008:17). Such sentences are interpreted in terms of ‘restrictor’ and ‘matrix’ with restrictor ‘providing a context in which the matrix is interpreted’ (Carlson 2008:32). In other words, a sentence part qualifies as a ‘restrictor’ if it imposes some limitation on the whole set of

45

On Greek, see Rijksbaron (2006:10 f.).

modalized future and scheduled present in coan inscriptions

213

all the possible interpretations of a ‘matrix’. Our calendars’ articles illustrate the matter quite manifestly. A calendar date (we are mostly dealing with a certain day of a certain month) is a restrictor and events associated with this day are matrices. In the following example, for instance, [το]ῦ Βατ̣ρομίο̣̣ υ̣ […] δεκάται is the restrictor, the rest is the matrix: (22) [το]ῦ Βατ̣ρομίο̣ ̣υ̣ […] δεκάται Ἥραι Ἀργείαι Ἑλείαι Βασιλείαι δάμ/αλις κριτά· […] θύει ἱαρεὺς καὶ ἱερὰ παρέχει· γέρη̣ λαμβά[νει] δέρμα καὶ σκέλος· ταύτας ἀποφορά· ἔνδορα ἐνδέρεται καὶ θύε̣[ται] / ἐπὶ τᾶι ἱστίαι ἐν τῶι ναῶι τὰ ἔνδορα καὶ ἐλατὴρ ἐξ ἡμιέκτου [σπ]/υρῶν· τούτων οὐκ ἐκφορὰ ἐκ τοῦ ναοῦ. ‘On the tenth of [the month] Batromios for Hera Argeia Heleia Basileia a choice heifer […] the priest sacrifices and provides the offerings; as perquisites he receives the skin and a leg. Of (the meat of ) this (heifer) carrying away (is allowed). The offerings wrapped in skin are prepared and on the hearth in the temple the wrapped offerings and a cake from half a hekteus of wheat are sacrificed. Of these (offerings) no carrying out of the temple’. ig xii 4 2741– 10

The essential property of generic sentences is that they do not refer to any particular event and in this relation are opposed to so-called ‘episodic’ sentences. The semantic domains of genericity, non-factuality and modality overlap, and these semantic intersections are often noted in the modern works on generics.46 As far as modality in particular is concerned, the following observation is instructive: ‘Genericity […] resembles […] a deontic or normative modality, since facts about what ought to be the case similarly neither entail, nor are entailed by, facts about what is actually the case […] generics do not just resemble normative statements. They almost are normative statements’ (Asher-Morreau 1995:304f.). Thus the calendar generics are predisposed to prescribe some patterns of behavior, or even more strongly, to oblige the citizens to follow these patterns. It may be claimed that generics border on the domain of participant-external modality, if not fall within it. There are two refinements to be made at this point. First, not all present indicatives in the Coan corpus occur in generic sentences about periodic events. Thus 364 schedules the occurrence of individual events to take place on one occasion and for one time only:

46

Cf. Dahl (1995:419); Tatevosov (2005); and Carlson (2008:19).

214

veksina

(23) τᾶι v ι̅ζ ̅ v τοὶ ἱερατευ/κότες στεφανοῦντι / Πολύμναστον Ζωπύρου / στεφάνῳ χρυσῷ […] ‘On the seventeenth (of the month Dalios): the ex-priests crown Polymnastos, the son of Zopyros, with a golden crown […]’. ig xii 4 3642– 5

(24) ι̅η̅ τοὶ ἱερατευκότες στεφα/νοῦντι Ἁγησίαν Δαμοφῶν/τος στεφάνῳ χρυσέῳ […] ‘On the eighteenth: the ex-priests crown Hagesias, the son of Damophon, with a golden crown […]’. ig xii 4 36412– 14

While each of these sentences contains a proper name, and expresses not a custom, but a unique event, they still resemble calendars in terms of providing a schedule for the fulfillment of actions. One might imagine that these indicatives are a kind of secondary usage inspired by the calendar discourse. Yet the link seems to be deeper: though the events are neither periodic nor recurring, their sequence is constrained; they are thought of as an instantiation of a pattern. What distinguishes this case from the calendar schedules is that this pattern should only have one instantiation. Another case is θύει 7234. The ill-preserved context, according to Herzog (1928:21ff.), contains provisions on the initiation of a priestess. Here the restriction is situational (‘each time a priestess is to be initiated, such and such actions should be performed’), and the sentence carries a generic meaning.47 Second, not all generics are expressed by the present indicative in our corpus. Imperatives and infinitives (for references, see 9.3.1) are clearly not so well-grounded in the calendar discourse and can be seen as later intrusions into the domain of indicatives. As the parallels gathered in 9.3.2. suggest, the ‘scheduled present indicative’ was an older phenomenon, well preserved in Cos. In historical times, the position of the indicative present among the means of the expression of deontic modality was hardly anything more than peripheral. Rather, its deontic function was likely to be confined to some special kinds of contexts, of which the calendars and schedules constitute a clearly discernible group.

47

As Chung-Timberlake (1985:221) put it, actions can be repeated over time or ‘over possible occasions in possible worlds’; see Dahl (1995:423) and Carlson (2008). The same kind of analysis applies to the indicatives in 332 and probably in 7284.

modalized future and scheduled present in coan inscriptions

215

The observed interchange with the imperative and the infinitive is highly instructive for us, since it explicitly precludes a non-deontic interpretation of indicatives as well as of their intrinsic counterparts, non-verbal syntagms. The latter, unlike the imperative and the infinitive, are well-grounded in the calendar discourse. This is not incidental. Cross-linguistically, one of the usual means to express ‘habituality’ is a ‘maximally unmarked tensed’ form available within a language (Dahl 1995:416; Brown 2006:v.19), and ‘in some types of generics, an alternative nominalization construction can be used’ (Dahl 1995:422). Furthermore, an event expressed by a noun is not characterized in terms of aspect, whereas verbal generics are involved in the aspect-marking system, and constitute a category of their own within this system.48 Generic, or habitual aspect rarely possesses an overt morphological marker of its own though. If a language makes a distinction between perfective/imperfective aspect, and Greek does, it is usually the imperfective form that marks the generic aspect,49 and this is precisely the case with our scheduled present.50 Generic aspect is to be distinguished from the ‘lexical aspect’ characteristics of predicates. Among the scheduled presents in the Coan corpus we find diverse kinds of predicates except for state predicates:51 atelic processes, e.g., [οἱ παῖ]δ̣ες / [τρέ]χ̣ουσιν ‘the paides run’ (i.e., the running competition of paides takes place; 281 5– 6); πορεύοντα[ι] / τελέαν (sc. ἄρνα) ‘they lead the full-grown (ewe) in the procession’ (27954– 55); ἱαροποιοὶ δὲ ξενίζον̣τι τὸν̣ ἱερῆ καὶ τὸς κάρυκας τα[ύ]/τ̣αν τὰν νύκτα ‘hiaropoioi entertain the priest and the heralds this night’ (27840– 41); telic accomplishments, e.g., ἀγορεύει ‘announces’ (27825, the text of the announcement follows), ἔνδορα ἐνδέρεται (literally) ‘wrapped in hide offerings are wrapped in hide’ (2748); telic achievements, e.g., [γ]/έρη λαμβάνει δέρμα ‘as prerequisites he receives the skin’ (2744– 5), etc. The generic meaning is formalized within the present tense form, but the sentence structure is quite specific too, so that it is indispensable for the generic interpretation. Fig. 9.2 visualizes the transition from non-deontic to deontic generics, taking as a starting point a modalized reading of generics.

48 49 50

51

Chung-Timberlake (1985:220 ff.); Bybee et al. (1994); and Dahl (1995). Bybee et al. (1994:159 f.); Brown (2006:v.19). As Dahl (1995:424) and Carlson (2008:31 ff.) observe, non-past/past tense differentiation in principle is available for generics, but the issue of tense-marking in Greek generics goes beyond the scope of this paper. This would demand an investigation of present/aorist aspectual syncretism (to account for the gnomic aorist) and aorist/imperfect past tense syncretism (to account for generic propositions referred to the past); see Wackernagel (2009:229 ff.) and West (1989). Cf. the imperative καθήσθω in 2789.

216

veksina

figure 9.2 From generics to deontic modality

9.4

Conclusions

This paper has exemplified the present and future indicative forms as expressions of obligation and permission in the corpus of the Coan inscriptions and has investigated this phenomenon from the cross-dialectal as well as from the cross-linguistic perspective. The semantic and grammatical properties that enable the indicative forms to be used in directives have been examined.52 It is clear that future predictions and statements about scheduled events constitute semantic domains closely related to the domain of deontic modality, and the corresponding semantic paths have been visualized. The specific character of these paths can be subsumed under the following points: i) they reflect grammatical polysemy; ii) they are not productive, i.e., they reflect a development that did not lead diachronically to weakening or abandonment of the source meaning in favour of the secondary one; iii) this secondary deontic meaning becomes actualized in sentences which have directive force. As the last thesis suggests, the identified grammatical polysemy is closely related to the interactive character of the texts where it reveals itself. The officially published legal deeds are allocated in the context of social interaction where the distribution of power and authority is presupposed. The legislative discourse can be expected to exploit some secondary means of the expression of deontic modality in a consequent way. The pragmatic dimension of this usage remains a subject for further investigation, however.

52

For another ‘indirect’ use of the Greek indicative present to express suggestion, see Rijksbaron (2006:9).

modalized future and scheduled present in coan inscriptions

217

References Asher, R.E. (ed.). 1994. The encyclopedia of language and linguistics. Oxford. Asher, N., & M. Morreau. 1995. ‘What some generic sentences mean’. In: Carlson & Pelletier 1995:330–339. Bile, M. 1988. Le dialecte crétois ancien: étude de la langue des inscriptions, recueil des inscriptions postérieures aux ic. Paris. Bosnakis, D. & K. Hallof. 2010. ‘Alte und neue Inschriften aus Kos iv’. Chiron 40:323– 358. Brown, K. (ed.). 2006. The encyclopedia of language and linguistics. Oxford. Bybee, J., Perkins, R. & W. Pagliuca. 1994. The evolution of grammar: tense, aspect, and modality in the languages of the world. Chicago. Bybee, J. & S. Fleischman (eds.). 1995. Modality in grammar and discourse. Amsterdam—Philadelphia. Carlson, G. & F.J. Pelletier (eds.). 1995. The generic book. Chicago. Carlson, G. 2008. ‘Patterns in the semantics of generic sentences’. Time and modality, ed. by J. Guéron & J. Lecarme, 17–38. Dordrecht. Cassio, A.C. (ed.). 1999. kata dialekton. Napoli. Chung, S. & A. Timberlake. 1985. ‘Tense, aspect, and mood’. Grammatical categories and the lexicon, ed. by T. Schopen, 202–258. Cambridge. Dahl, Ö. 1995. ‘The marking of the episodic/generic distinction in tense-aspect systems’. In: Carlson & Pelletier 1995:412–427. Duhoux, Y. 20002. Le verbe grec ancien: éléments de morphologie et de syntaxe historiques. Louvain-la-Neuve. Herzog, R. 1928. ‘Heilige Gesetze von Kos’. Abhandlungen der Preußischen Akademie der Wissenschaften. Philosophisch-Historische Klasse. Nr. 6. Berlin. Ivanov, M. & S. Tatevosov. 2009. ‘Event structure of non-culminating accomplishments.’ Cross-linguistic semantics of tense, aspect and modality, ed. by L. Hogeweg et al., 83– 131. Amsterdam. Giannakidou, A. & A. Mari. In press. ‘Epistemic weakening with future and ‘must’: nonveridicality, evidentiality and partial knowledge’. Tense, mood, and modality: new perspectives on old questions, ed. by J. Blaszac et al. Chicago. Krifka, M. 1998. ‘The origins of telicity’. Events and grammar, ed. by S. Rothstein, 197– 235. Dordrecht. Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. Magnien, V. 1912. Le futur grec. ii. Paris. McLean, B.H. 2002. An introduction to Greek epigraphy of the Hellenistic and Roman periods. Ann Arbor. Minon, S. 2007. Les inscriptions éléennes dialectales (vie–iie siècle avant j.-c.). Genève.

218

veksina

Morpurgo Davies, A. 1999. ‘Contatti interdialettali: il formulario epigrafico’. In: Cassio 1999:7–33. Ortuño, F.J.M. 1999. Sintaxis de las inscripciones griegas de Efeso. Amsterdam. Parker, R. & D. Obbink. 2000. ‘Aus der Arbeit der Inscriptiones Graecae vi. Sales of priesthoods on Cos’. Chiron 30:415–449. Parker, R. & D. Obbink. 2001. ‘Aus der Arbeit der Inscriptiones Graecae viii. Three further inscriptions concerning Coan cults’. Chiron 31:253–276. Portner, P. 2011. ‘Verbal Mood’. Semantics, ed. by C. Maienborn et al., 1262–1291. Berlin. Rhodes, P.J. & D.M. Lewis. 1997. The decrees of the Greek states. Oxford. Rijksbaron, A. 20063. The syntax and semantics of the verb in Classical Greek: an introduction. Chicago. Schwyzer, E. & A. Debrunner. 1950. Griechische Grammatik. ii: Syntax und syntaktische Stilistik. München. Searle, J.R., Kiefer, F. & M. Bierwisch (eds.). 1980. Speech act theory and pragmatics. 10. Dordrecht. Tatevosov, S. 2002. ‘The parameter of actionality’. lt 6:317–401. Tatevosov, S. 2005. ‘From habituals to futures’. Perspectives on aspect, ed. by H.J. Verkuyl, 181–197. Dordrecht. Truckenbrodt, H. 2006. ‘On the semantic motivation of syntactic verb movement to c in German’. tl 32:257–306. Uguzzoni, A. & F. Ghinatti. 1968. Le tavole greche di Eraclea. Rome. Van der Auwera, J. & V.A. Plungian. 1998. ‘Modality’s semantic map’. lt 2:79–124. Villaro, B.O. 1998. ‘La distribución de imperativo, futuro e infinitivo-imperativo en las inscripciones dialectales cretenses’. cfc 8:185–208. Wachter, R. 1999. ‘Evidence for phrase structure analysis in some Archaic Greek inscriptions’. In: Cassio 1999:365–382. Wackernagel, J. 2009. Lectures on syntax with special reference to Greek, Latin, and Germanic. Translated by D. Langslow. Oxford. West, M.L. 1989. ‘An unrecognized injunctive usage in Greek’. Glotta 67:135–138. Wiemer, H.U. 2003. ‘Kaufliche Priestertumer im hellenistischen Kos’. Chiron 33:263–310. Wunderlich, D. 1980. ‘Methodological remarks on speech act theory’. In: Searle et al. 1980:291–312. Zanuttini, R., Pak, M. & P. Portner. 2012. ‘A syntactic analysis of interpretive restrictions on imperative, promissive, and exhortative subjects’. nllt 30:1231–1274.

chapter 10

‘High’ and ‘Low’ in Medieval Greek Geoffrey Horrocks

10.1

Introduction

The now standard prioritization of ‘naturally acquired’ spoken language over prescriptively defined written language has generally led historical linguists to focus on documents revealing of the vernacular rather than on those composed in more elaborated (‘artificial’) registers. This has tended to work against serious linguistic analysis of the ‘high’ varieties of languages characterized by long-term diglossia, though Greek is a partial exception in so far as the ancient language has enjoyed its own long and independent tradition of grammatical description based largely on texts written in just such registers. Greek of the Byzantine era is a different matter, however. The comparatively little grammatical work that has been carried out to date has centred on the usage of individual writers and on ‘low(er)’ varieties, while the language of belletristic texts has generally been characterized as faux ‘Classical Greek’, reproduced with varying degrees of skill and good taste, a variety without need of further study beyond the identification of its many ‘mistakes’ (see, for example, Browning 1983). The reasons are not hard to detect. Such an approach conforms not only with the expectations of theoretical linguists, but also with the views of the finally (more or less) triumphant demoticist movement, which advocated contemporary speech as the basis for a national written language in the period from Greek independence to the fall of the military dictatorship in 1974 (see Mackridge 1990 & 2008). More recently, however, it has come to be recognized that this narrowing of the domain of enquiry may ultimately prove misleading and even counterproductive (see, for example, Wahlgren 2002 & 2010, Hinterberger 2014). After all, a ‘dead’ written variety of a language used creatively beside spoken ones also changes with time, in part because those who use it, however well trained, subconsciously reconceptualize its older grammatical elements in contemporary terms. This is particularly likely to occur in the domain of syntax and semantics, where formal instruction is less precise and restrictive than in matters of orthography, morphology or lexicon, and where speakers inevitably fall back on their instincts and intuitions as native speakers of the contemporary language, thus opening the way for new norms to develop without obvious violation of ancient rules. Morphological resources that no longer have a role in

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_011

220

horrocks

the spoken language are particularly liable to such functional reanalysis, with the result that a ‘dead’ variety of a language is just as likely to take on underlying attributes of its ‘living’ counterpart1 as ‘natural’ speech is to be affected by the ‘artificial’ written language learned in the classroom (see Horrocks 2010:213– 214). The banal reality of interplay between high(er) and low(er) varieties of Medieval Greek should not be underestimated if the nature of the language is to be properly appreciated. In practice, Byzantine writers working in classical genres rarely (if ever) attempted virtuoso composition in the Attic Greek of the 5th and 4th c. bce, but instead maintained a weaker form of linguistic continuity through the genre-conditioned ‘upgrading’ of contemporary Greek by reference not simply to the classical language but to the entire tradition of classicizing prose up to their own time, a tradition reflecting successive reinterpretations, including their own, of the syntactico-semantic function of obsolete elements of ancient grammar. Seen in this light, high-register Byzantine writing is less an imitation of Ancient Greek—though it certainly owes a great deal to it—than a variety of Medieval Greek ‘antiqued’ according to convention. But while there were explicit paradigms of ancient morphology for the writer to follow, the actual understanding and use of obsolete forms at any given time was in part determined by the evolving grammar of contemporary Greek, with the resulting reanalyses being tacitly absorbed over time into the ‘normal’ practice of the written tradition.2 My purpose here, then, is to exemplify the process of reanalysis by identifying a Byzantine grammatical norm which, though not ancient, had become standard in its own time, and to demonstrate that it was shaped by the grammar of the contemporary vernacular. The focus will be on the language of the Alexiad of Anna Comnena (1083–c. 1153).3 In the prologue to her history, Anna draws special attention to her study of Greek language, literature and rhetoric,

1 The earliest example of this kind of development in the Greek world is provided by the orally transmitted Kunstsprache of epic poetry and its Hellenistic/Roman afterlife (see, for example, Horrocks 1997). 2 In this connection we should recall the description of the high style as ‘Atticized’ in Chapter 1 of Constantine Porphyrogenitus’ De Administrando Imperio: οὐ γὰρ ἐπίδειξιν καλλιγραφίας ἢ φράσεως ἠττικισμένης καὶ τὸ διηρμένον διογκούσης καὶ ὑψηλὸν ποιῆσαι ἐσπούδασα (Moravcsik 1967). 3 Anna composed her history of the reign of her father, Alexius i Comnenus, after retiring to a convent on the death of her husband, Nicephorus Bryennius the Younger, in 1137. Before his death, Bryennius had begun a history of his wife’s family. The fact that this remained incomplete (covering only 1070–1079) was perhaps an inspiration to Anna.

‘high’ and ‘low’ in medieval greek

221

together with history, philosophy, geography, military affairs, mathematics and science. Such rare erudition suggests that her literary Greek will be representative of the norms of the period and this expectation is readily confirmed by comparison with the language of other 12th century writers such as the statesman and historian Nicetas Choniates and the churchman and scholar Eustathius of Thessalonike (see also Stone 2009). The analysis below deals specifically with verb forms denoting futurity and future-referring modality. The classical and post-classical uses of the future indicative, subjunctive and optative are outlined in 10.2 and the corresponding future/modal system of the medieval vernacular is explained in 10.3. Section 10.4 then deals with the uses of the ancient future indicative, subjunctive and optative in books 1 (including prologue), 2 and 11 of the Alexiad (KambylisReinsch 2001). These are compared both with true classical practice and with the usage of the fragmentary, Palaeologan-era metaphrase that reproduces the content of 11.7–13.12 in a simpler written style (Hunger 1981; see Davis 2010 for the role of metaphrases). Many examples in the Alexiad indeed turn out to be unclassical, but far from constituting a mere collection of errors, the attested usage is shown to reflect directly the underlying future/modal system of normal contemporary Greek. I conclude in 10.5 that the traditional dichotomy between (pseudo-)ancient written Greek and the spoken vernacular may be more realistically conceptualized as a continuum of genre-determined variation in the realization of grammatical categories characterizing contemporary Medieval Greek in all its manifestations.

10.2

Futurity and Modality in Ancient Greek

10.2.1 Future Indicative The future indicative (with negative οὐ) is used in main and subordinate clauses to locate ‘actual’ events4 at potentially determinate times in the future relative to the time of utterance (see (1a)) or the time referred to by the main verb (see 1b). There is no formal distinction of aspect: (1) a. τοῦτο (οὐ) γενήσεται. ‘This will (not) happen’.

4 No future ‘event’ is certain to occur and futurity is therefore inextricably bound up with modality (see 10.2.4), but speakers may still refer to the future with strong conviction.

222

horrocks

b. εἶπεν ὅτι τοῦτo (οὐ) γενήσεται. ‘S/he said that this would (not) happen’. In the 2nd person the future may be taken to express a wish or command, or to make a concession, cf. will you do that? (I hope so), you will do that (now), you will do that (and there’s nothing I can do about it). In all these cases there is functional overlap with the optative, see 10.3. The future is also used with ὅπως ‘how’ (negative μή) to complement verbs of ‘taking care’, ‘seeing to it’ etc. This construction may be used elliptically with imperatival function: c. (ἐπιμελεῖται/ἐπεμελήθη) ὅπως (μὴ) τοῦτο γενήσεται. ‘S/he is taking/took care that this will/would (not) happen’. ‘Make sure this does (not) happen’. The future is also used in relative clauses to express purpose (negative μή): d. ἔρχεταί/ἦλθέ τις ὃς τοῦτο ποιήσει. ‘Someone is coming/came to do this.’ 10.2.2 Subjunctive Use of the subjunctive mood requires a choice between two sets of forms marking perfective (‘aorist’) and imperfective (‘present’) aspect respectively. It does not appear in declarative sentences, but is used to locate possible events in the future in deliberative questions (negative μή, see (2a)), exhortations/commands (mainly 1st person plural, often shading into a wish) and prohibitions (perfective only with negative μή, see (2b)), and also, with οὐ + μὴ combined, in strong denials (see (2c)): (2) a. (μὴ) ἔλθωμεν; ‘Shall we (not) go?/Are we (not) to go?’ b. (μὴ) ἔλθωμεν. ‘Let us (not) go’.

‘high’ and ‘low’ in medieval greek

223

c. οὐ μὴ ἔλθωμεν. ‘We will not go’. In conditional and temporal clauses the subjunctive regularly combines with ἄν in its indefinite function to locate a possible event at an indeterminate future time (see (3a)) or to locate an indefinite number of such events at a series of unspecified times (see (3b)). The appropriate reading is determined by the temporal/aspectual character of the main verb, future-referring versus atemporal/habitual:5 (3) a. ἐὰν/ὅταν τοῦτο γένηται, φευξόμεθα. ‘If/when this happens, we shall flee’. b. ἐὰν/ὅταν τοῦτο γένηται, φεύγομεν. ‘If/when this happens, we flee’. The subjunctive is also used in final clauses with ἵνα, ὡς and ὅπως (ἄν) (negative μή, see (3c)), and with μή after verbs of ‘fearing’ (negative οὐ, see (3d)) to denote potential future events relative to the time referred to by the main verb: c. τοῦτο ποιεῖ/ἐποίησε ἵνα διαφύγῃ. ‘S/he is doing/did this to escape’. d. φοβεῖται/ἐφοβήθη μὴ (οὐ) τοῦτο γένηται. ‘S/he is/was afraid that this will/would (not) happen’. 10.2.3 Optative Use of the optative mood also involves a compulsory choice between forms marking perfective (aorist) and imperfective (present) aspect. Used alone in main clauses it expresses a wish for the future (negative μή, see (4a)):

5 The present indicative, as in (3b), is an imperfective non-past tense that contrasts temporally with the imperfect indicative (see note 7); it therefore denotes continuousness or progressiveness in the present as well as temporally unspecified habituality.

224

horrocks

(4) a. τοῦτο (μὴ) γένοιτο. ‘May/let this (not) happen!’ In combination with ἄν in its consequential-inferential function the ‘potential’ optative expresses epistemic possibility, i.e., what ‘may/might’, ‘can/could’, ‘would’ be the case if some actually expressed, contextually implied or wholly inexplicit condition were fulfilled (negative οὐ, see (4b)): b. τοῦτο (οὐκ) ἂν γένοιτο. ‘This can/could/may/might/would (not) happen’. In the 2nd person the potential optative may make a concession or give a polite command, cf. you can/could/should offer some help; this involves some overlap with the future indicative (see 10.2.1). It also competes with the subjunctive in deliberative questions, cf. how can/could/should we do this? (see 10.2.2). In conditional protases the optative is used to formulate ‘speculations’ about the future (negative μή), and is typically followed by a potential optative in the main clause (apodosis). Thus where (3a) describes a confidently predicted consequence of an anticipated future event, (5a) places an entirely hypothetical eventuality in the future ‘for the sake of argument’ and explores its ‘potential’ consequences: (5) a. εἰ τοῦτο γένοιτο, φύγοιμεν ἄν. ‘If this happened/were to happen, we would flee’. Conditional and temporal clauses containing optatives may also be understood generically (as (3b) is beside (3a)), but the optative then locates an indefinite number of events at random times in the past (see (5b)). The appropriate reading again follows from the temporal reference of the main verb, i.e., future-referring (potential optative) versus past-referring (habitual imperfect indicative):6 b. εἰ/ὅτε τοῦτο γένοιτο, ἐφεύγομεν. ‘If/when this happened, we used to flee’. 6 The imperfect indicative is an all-purpose imperfective past tense; like the present indicative it marks continuousness, progressiveness etc., but contrasts with it temporally (see note 5).

‘high’ and ‘low’ in medieval greek

225

In other subordinate clauses the optative is mainly used as an optional variant of the subjunctive when the verb of the main clause refers to the past. Compare the final- and ‘fear’-clauses in (3c)/(3d) with those in (5c)/(5d): c. τοῦτο ἐποίησε ἵνα διαφύγοι. ‘S/he did this to escape’. d. ἐφοβήθη μὴ (οὐ) τοῦτο γένοιτο. ‘S/he was afraid that this would (not) happen’. Finally, the optative is also used in past-time complement clauses and adjuncts as an optional alternative to the appropriate tense of the indicative. Compare (6a)/(6b) with (1b)/(1c), and (6c) with (1d): (6) a. εἶπεν ὅτι τοῦτο ποιήσοι/ποιoῖ/ποιήσειε. ‘S/he said that s/he would do/was doing/had done this’. b. ἐπεμελήθη ὅπως (μὴ) τοῦτο γενήσοιτο. ‘S/he took care that this would (not) happen’. c. ἦλθέ τις ὃς τοῦτο ποιήσοι. ‘Someone came to do this’.

10.3

Functional Overlaps and Their Consequences

The above summarizes what standard grammars have to say about Attic Greek of the 5th and 4th c. bce. Since both the future indicative and the subjunctive refer to the future, however, the ‘other’ form is sometimes attested beside the ‘regular’ option in most clause types (the main exception is declarative sentences, where subjunctives do not compete with futures). Some of this variation may well have been introduced by later copyists in line with the developments explained below, but some is probably original and works of reference normally treat future/subjunctive variation as a genuine phenomenon (e.g., Goodwin 1890, with many reprints).

226

horrocks

As noted, the optative in future-referring subordinate clauses was mainly an optional variant of the future indicative (see (6a), (6b), (6c)) or subjunctive (see (5c), (5d)), redundantly marking ‘past’ futurity/modality in contexts already so designated by the main verb. This applied equally to conditional clauses in pasttime indirect speech, cf. εἶπεν ὅτι [ἐὰν τοῦτο γένηται, φευξόμεθα/εἰ τοῦτο γένοιτο, φευξοίμεθα] = ‘s/he said that if this happened, we would flee’. The general optionality of the optative must eventually have contributed to a blurring of the distinction between the two types of future-referring conditional (see (3a) and (5a)) even in direct speech.7 It was only in the expression of wishes and in potential function, therefore, that the optative had the makings of a truly distinctive identity. But wishes were a minor category, partly replicated by ‘jussive’ subjunctives, cf. let/may this happen, and the potential optative too had competition, with the future and/or subjunctive in deliberative questions, concessions and commands, and in its declarative role with modal periphrases such as δύναμαι/ἔχω ‘I can’ and ἔξεστι/πάρεστι ‘it is possible’ + infinitive. 7 The generic readings in (3b) and (5b) are not explored further here. ‘Will’ in a conditional apodosis locates an event at a potentially specific time in the future. Since reference to the present or past in an associated protasis requires the use of appropriate indicatives (cf. if he did this/is doing this, he will be happy), the English ‘present tense’ here is taken to refer to the future. The same applies to the Ancient Greek subjunctive, mutatis mutandis, and we might speculate that the conditional context was the source for the ‘tensing’ of a mood that originally denoted inherently atemporal possibilities (cf. the generic use of the subjunctive in the protasis of (3b)). ‘Would’ in apodoses marks a hypothetical consequence that logically follows the fulfilment of a logically preceding hypothetical condition. Consider if he had come/was coming/came, he would be happy, where the tenses in the protases are ‘more past’ than their real-world time reference requires: this non-temporal ‘pastness’ marks logical anteriority to the apodosis, where the ‘future-in-the-past’ in turn marks logical posteriority to the ‘past’ condition. But since would have locates hypothetical consequences in the past, would, by elimination, is taken to locate them in the present or future. The situation in Classical Greek is similar, but since the aorist and imperfect indicatives (with ἄν) were used in apodoses to locate hypothetical eventualities in the past and present respectively, the optative (with ἄν) was taken to refer only to the future. And since the aorist and imperfect were also used in protases to refer to the past and present, the optative there, just like the subjunctive, was also taken to refer to the future (though in Homer the optative has wider scope, suggesting it may once have been used to present hypothetical versions of the past, present or future (Goodwin 1890; Willmott 2007)). Optatives in protases therefore locate speculative conditions ‘in the past’ relative to their consequences, but at random times in the future relative to the time of speaking. Correspondingly, the potential optative in apodoses locates hypothetical consequences ‘in the future’ with respect to the ‘past-time’ context established by the protasis but in the actual future relative to the time of speaking (see Horrocks 1995).

‘high’ and ‘low’ in medieval greek

227

Given the functional overlaps between the subjunctive and the future indicative and the near-redundancy of the optative, it should be no surprise that, with the spread of Greek in the wake of Alexander’s conquests and its subsequent consolidation as the lingua franca of the East by the Romans, major simplifications of the future/modal domain took place, simplifications which, as we shall see, were only ever superficially ‘remedied’ by the Atticists’ efforts to revive the language of the ancients. 10.3.1 Optatives in Hellenistic and Roman Greek During the last three centuries bce, the optative, lacking a clear function of its own, slowly fell out of popular use other than in formulaic wishes and other set phrases, e.g., μὴ γένοιτο ‘God forbid!’, καλῶς ἂν ἔχοι ‘it would be good’. In the New Testament and contemporary documentary papyri, therefore, the optative is already very rare, and in the face of the wholesale substitution of subjunctives and indicatives even the most committed Atticists seemingly lacked the will to try and restore it to its former glory. Even Dionysius of Halicarnassus, the high priest of early Atticism, makes remarkably little use of it, and though it later recovered some ground in literary registers, it was essentially as a mechanical variant of the subjunctive or future indicative in contexts where the use of one or both of these was the contemporary norm. In the middle to lower registers of the Koine, therefore, future indicatives and aorist subjunctives frequently appear where a potential optative might have been used—usually with a potential nuance and sometimes with ἄν.8 This usage perhaps began in deliberative questions and then spread to other main clauses: (7) a. τίς ἐξ ὑμῶν ἕξει φίλον […] καὶ εἴπῃ αὐτῷ […]; ‘Which of you will/can have a friend and say to him […]?’ Lk 11.5

b. μὴ γένοιτο· ἐπεὶ πῶς κρινεῖ ὁ θεὸς τὸν κόσμον; ‘God forbid! For how will/can God judge the world?’ Rom 3.6

8 Editors have in general ‘corrected’ examples with ἄν in more literary texts (e.g., Polybius, Dionysius of Halicarnassus, Plutarch etc) either by deleting the particle with futures or by replacing subjunctives with optatives.

228

horrocks

c. ἀλλὰ ἐρεῖ τις· πῶς ἐγείρονται οἱ νεκροί; ‘But someone will/may say ‘How are the dead raised?’ ’ 1 Cor 15.35

Unsurprisingly, the semantically opaque optative in indirect statements (see (6a)) was also simply dropped and the imperfect indicative was standardly substituted in past-time conditional/temporal clauses expressing indefinite frequency (see (5b)), sometimes with the ‘indefinite’ ἄν previously used only with subjunctives (see (3a)/(3b)):9 (8) ἔθνη ἤτε […] πρὸς τὰ εἴδωλα […] ὡς ἂν ἥγεσθε ‘You were gentiles whenever you were led to the idols’ 1 Cor 12.2

Though the optative is similarly absent from other subordinate clauses in pasttime contexts, the Atticists had a particular fondness for it in future-referring conditionals (see (5a)), and this eventually came to have some impact on general written usage. It is striking nonetheless how often εἰ + optative in a protasis is followed by a main-clause future indicative. Evidently, the optative here was simply taken as a marked variant of the regular subjunctive/future indicative, cf. the future in the protasis of (9a) with the optative in that of (9b): (9) a. ἀμὴν λέγω ὑμῖν, εἰ δοθήσεται […] σημεῖον ‘Truly I say to you, if a sign is/should be given […]’ (sc. ‘may I die’). Mk 8.12

b. εἰς τὴν βασιλείαν τοῦ θεοῦ οὐδεὶς εἰσελεύσεται, εἰ μὴ λάβοι τὸ ὄνομα τοῦ υἱοῦ αὐτοῦ. ‘No one shall/can enter the kingdom of God unless he takes take his (i.e., God’s) son’s name’. hermas Shepherd: Parable 9, 12.4 = chapter 89, 4

9 The expression of indefinite frequency now came to be seen as a function of imperfective aspect and this development paved the way for the eventual triumph of the longer forms of the relevant conjunctions (ἐάν, ὅταν etc) in all functions.

‘high’ and ‘low’ in medieval greek

229

Overall, then, the Koine had little use for the optative. One consequence of its demise was that expressions of the notions of futurity and possibility began to merge functionally, with speakers using the future indicative, and later also the subjunctive (see 10.3.2), to refer to the future with varying degrees of epistemic commitment (e.g. ‘x can/may occur’ as well as ‘x will occur’), leaving the precise determination of intent to context and common sense. 10.3.2 Subjunctives and Future Indicatives in Hellenistic and Roman Greek In the same period sound change largely obliterated the audible differences in regular verbs between the present indicative and present subjunctive and between the future indicative and aorist subjunctive.10 These developments worked with the functional overlaps already noted above to minimise any residual sense of contrast between the future indicative and the (aorist) subjunctive.11 Accordingly, as we move into late antiquity, future indicatives that were formally distinct from aorist subjunctives, e.g., future ἐρῶ ‘I will say’ vs. aorist subjunctive εἴπω ‘I may say’, largely disappear from lower-level writing and aorist subjunctives start to appear with future function in declarative sentences that were once the exclusive domain of the future: (10) a. τότε […] παραδείσου θύρα ἀνοιχθῇ. τότε ὑδάτων ἀπολαύσητε Χριστοφόρων. τότε Χριστοῦ προσηγορίαν λάβητε […] ‘Then the door of paradise will be opened. Then you will profit from the waters that bear Christ. Then you will receive Christ’s name’. cyril of jerusalem, Procatechesis 15, 4th c. ce

10

11

Present indicatives and subjunctives are built to the same stem, while future indicatives and aorist subjunctives are regularly built to a stem ending in /s/. When vowel length distinctions were lost, therefore, destroying the difference between -(σ)ομεν and -(σ)ωμεν, and the pronunciations of -(σ)ει- and -(σ)ῃ- merged, first to /(s)e(:)/ then to /(s)i(:)/, each pair of paradigms became identical in the 1st person plural and in the singular (-(σ)ω having always been common to both). In popular speech analogical levelling of the still distinct 2nd- and 3rd-person plural endings (-(σ)ετε/-(σ)ητε and -(σ)ουσι/-(σ)ωσι) led to the full generalization of the indicative endings, as increasingly reflected in sub-elite spellings in later antiquity (though the better educated always made efforts to maintain the traditional distinctions graphically, and in the 2nd–3rd person plural in speech as well). As the aorist (perfective) subjunctive became indistinguishable from the future indicative and the present (imperfective) subjunctive from the present indicative, the present indicative/subjunctive came to be used as a future too, thereby introducing an aspectual distinction into a domain where previously none had existed.

230

horrocks

b. ἀλλὰ καθὼς ἐμὲ ἠλέησεν καὶ ὑμᾶς ἐλεήσῃ καὶ μισθὸν […] ἀπολάβητε καὶ προσδέξωμαι ὑμᾶς ‘But just as he pitied me he will pity you too and you will receive your recompense and I will accept you’. callinicus, Life of St Hypatius 50.6, 5th c. ce

c. μὰ τὸν κατοικοῦντα ἐν τοῖς οὐρανοῖς, οὐκ εἴπω τινί, ἅπερ εἴπῃς μοι. ‘By him who dwells in heaven I shall not tell anyone what you tell me’. john moschus (ca. 550–619 ce), Spiritual Meadow 45 = migne p. 2900

While the subjunctives in (10a) and (10b) might just be taken to express commands/wishes (‘let x happen’), the use of negative οὐ(κ) rather than μή in (10c) confirms that, by the 6th century, the subjunctive was also used in main clauses as a simple future. This usage was then extended to more literary compositions, as we shall see in 10.4. The loss of a distinctive future paradigm led to various efforts at formal renewal. One option was the transfer of the future function to infinitival periphrases involving the modal/auxiliary verbs μέλλω ‘be about to x’ (a classical option), ἔχω ‘be able’12 and (ἐ)θέλω ‘want’ (see Markopoulos 2009 for a comprehensive treatment of the subject). But these never lost their original modal functions (see the dynamic use of ἔχω in (11a), with some deontic overtones), and sometimes even took on new ones (see the possible epistemic use of θέλω in (11b)): (11) a. ἀλλ’ εἰμὶ Χριστιανὸς καὶ πλέον τούτου παρ’ ἐμοῦ ἀκοῦσαι οὐκ ἔχεις. ‘But I am a Christian and you can/will not hear any more from me’. Martyrdom of Carpus 34.3, 2nd c. bce

b. λέγουσι δὲ ὅτι μέχρι ιε θέλομεν ἐξελθεῖν σῦν θεῷ. ‘But they say we can/may/will leave by the fifteenth, God willing’.13 P.Oxy 1763 r10, 3rd c. ce 12

13

The core meaning of ἔχω was ‘have’ (transitive), but the verb was also used in the Classical period with an infinitive in the sense ‘have (sc. the wherewithal) to x’, = ‘be able to x’, from which an epistemic reading of possibility evolved; in the 2nd person there are deontic overtones of permission etc., cf. you may not come, etc. The preceding context (the author of the letter complains that certain other ships have

‘high’ and ‘low’ in medieval greek

231

In the end, therefore, the development of future meanings for such periphrases served only to replicate the semantic indeterminacy of the monolectic forms they once competed with. Another innovation of the Roman period involved the prefixing of ἵνα to main-clause subjunctives. This began as an elliptical use of the ἵνα-clauses that replaced the classical object plus controlled infinitive construction after verbs of ‘urging’, ‘ordering’ etc., e.g., θέλω σε ἀπελθεῖν ‘I want you to leave’ > θέλω ἵνα ἀπέλθῃς lit. ‘I want that you leave’ (see Horrocks 2010:93–94), with a meaning between a command and a wish. Some of the earliest examples come from the New Testament: (12) a. Κύριε, ἵνα ἀνοιγῶσιν οἱ ὀφθαλμοὶ ἡμῶν. ‘Lord, may/let our eyes open’. Mt 20.33

b. ἀλλ’ ὡσπερ ἐν παντί περισσεύετε, […] ἵνα καὶ ἐν ταύτῃ τῇ Χάριτι περισσεύητε. ‘As you abound in everything, […] so may you abound in this grace’. 2 Cor 8.7

This development guaranteed the final demise of the optative of wish and simultaneously paved the way for the later generalization of ἵνα (soon reduced to νά in allegro speech) as a grammaticalized subjunctive ‘marker’. This was especially useful in main clauses, where there were no modally characteristic conjunctions to help distinguish indicative and subjunctive once their endings had fallen together.14 But the established bifunctionality of the future/modal infinitival periphrases meant that ἵνα/νά + subjunctive soon came to be used as a main-clause future too (see 10.4), a development reinforced by the parallel use of ‘bare’ subjunctives in this role (see the discussion of (10) above). By the early Byzantine period, therefore, efforts at formal renewal in the future/modal domain had left the indeterminacies they were introduced to

14

not yet sailed, thus preventing his own party from leaving, and that he has nothing to do) precludes the reading ‘we want to leave’. Even when formally distinct future indicative forms had disappeared from spoken Greek, the problem of maintaining audible distinctiveness remained for present indicatives and subjunctives, whose domains remained largely distinct.

232

horrocks

clarify stubbornly in place. Though each of the competing realizations of futurity/modality had its own nuances, the extent of the continuing functional and distributional overlaps was such that these key semantic notions were no longer formally distinguished in any clear-cut way.

10.4

Medieval Greek i: Low Registers

Following on from these developments, reference to the future in the middle to late Byzantine vernacular could be made by: (a) νά + subjunctive; (b) one of the infinitival periphrases, in which νά + subjunctive could now be substituted for the infinitive; (c) νά + the ἔχω-periphrasis (a lower-register innovation combining (a) and (b)). But as all three options also had a range of modal uses (including epistemic, volitional, dynamic and deontic), it is often difficult to say what the best English translation should be: (13) a. ὡς δὲ τὸ […] μήνυμαν ἦλθεν […] τίς τῆς καρδίας του τὴν χαρὰν νὰ ἰσχύσῃ καταλέξειν; ‘And when the news arrived—who could/will have the strength to describe the joy in his heart?’ Epithalamium (Constantinople), Folio 7v, 1–4, 12th or 13th/14th c.

b. καὶ τότε δεῦτε πρὸς ἐμὲ καὶ κάτι νὰ σᾶς εἴπω ‘And then come to me and I can/will tell you something’. Spaneas p 208, 12th c.

c. περάσειν ἔχω, Μαξιμού, ὡς διὰ σὲν τὸ ποτάμι ‘I can/will/must cross the river for you, Maximou’.15 Digenes Acrites e 1532, 12th c.

15

Maximou is on the opposite bank determined to cross over, but Digenes insists this is not something a woman should do.

‘high’ and ‘low’ in medieval greek

233

Furthermore, writers may combine two or more of the available options in a single sentence, thus confirming their essential equivalence. For example, in the following extracts from the Chronicle of the Morea we have the ἔχωperiphrasis, the θέλω-periphrasis (with both an infinitive and νά + subjunctive), νά + subjunctive, and νά + the ἔχω-periphrasis used interchangeably: (14) a. τὸ κάλλιον […] ὅπου ἔχομεν ποιήσει […] ‘The best thing we can do […]’; ‘the best thing, which we shall/must do […]’. Chron. Mor. h, 3647, 14th c.

b. κι ἂν εὕρωμεν τὸν βασιλέα […] […] τὸν θέλομεν πολεμήσει ‘And if we find the king, we will/shall attack him’. Chron. Mor. h, 3650–3651, 14th c.

c. ἂν θέλω ἀρτίως νὰ στείλω ἐκεῖ φουσσᾶτα ἐδικά μου […] νὰ τὸν ἐπάρουν εὔκολα καὶ νὰ τὸν ἔχεις χάσει ‘If I (shall) send my own armies there at once […] they can/will take it [viz. your land] easily and you will lose it’. Chron. Mor. h, 4233–4237, 14th c.

We may conclude, then, that the vernacular of the middle and later Byzantine periods largely continues, albeit with some innovative means of expression, much of the formal variability and functional indeterminacy already in place in late antiquity. In other words, we continue to deal with a set of competing realizations of futurity/modality, none of which distinguishes these concepts formally.

10.5

Medieval Greek ii: High Registers

Mastery of Ancient/belletristic Greek clearly required lengthy instruction and considerable personal application (as it still does), but the important question is whether the Byzantines learned it as an autonomous ‘dead language’ (as we must) or rather as a variety of contemporary Greek characterized by distinctive

234

horrocks

grammatical, lexical and stylistic ‘transpositions’. Since the majority of teachers and learners were native speakers, it is hard to believe that the traditional resources of literary Greek were not viewed through the lens of the vernacular, which, even in its most elite varieties, was far more modern than ancient.16 Atticists of the Roman period retained the inventory of traditional forms in a given domain (e.g., modality) but redeployed those that had fallen out of use (e.g., optatives) as variants of those that remained, with little apparent concern for the minutiae of ancient distribution. A priori it seems reasonable to suppose that Medieval Atticists would have done the same. The crucial thing now, therefore, is to establish whether the non-classical innovations in the Byzantine tradition of belles lettres have a vernacular origin. This article begins the process by examining how far Anna Comnena’s use of future indicatives, subjunctives and optatives suggests a ‘modern’ understanding of the future/modal domain.17 Though much of Anna’s literary Greek is apparently classical, other features such as the use of optatives for subjunctives in non-past subordinate clauses reflect subsequent Atticist practice. There are also innovations from late antiquity and the early Middle Ages like the use of subjunctives as futures in main clauses, something which by Anna’s time was sufficiently well attested in ‘respectable’ authors to have been accepted into the learned tradition. Only the most recent innovations, as yet with no validating literary hallmark, are absent, such as periphrases other than μέλλω + infinitive or subjunctives marked with ἵνα/νά. But as we have seen, all realizations of futurity and modality since the classical age were functionally ambiguous and Anna’s own vernacular usage could have been no different, whatever set of forms she might actually have used in casual conversation. In order, therefore, to demonstrate that Anna’s high-register Greek reflects a contemporary understanding of the future/modal domain we need only show that, even if she sometimes uses classical forms in classical ways, she does not differentiate systematically between one set of forms to denote futurity and another to convey possibility. It is important, therefore, to focus on contexts where Classical Attic did not allow overlap. Accordingly, the examples below are all main-clause declarative sentences where subjunctives were impossible and indicatives contrasted with potential optatives. If Anna allows undifferen16 17

We need only look at the ‘vernacular’ literature of twelfth-century Constantinople to confirm that this was so, e.g. the Ptochoprodromica, Spaneas etc. We are of course at the mercy of the text as transmitted and edited (I used KambylisReinsch 2001), but it is assumed here without further comment that recurrent phenomena are likely to be indicative of genuine patterns of usage.

‘high’ and ‘low’ in medieval greek

235

tiated variation between future indicatives, subjunctives and optatives in this context, it is clear that she understood and used the ancient inventory of forms in an essentially modern way.18 10.5.1 The Alexiad We can now examine the use of ‘classical’ future indicatives, subjunctives and optatives in the Alexiad. The examples are taken from Books 1 (including the Prologue), 2 and 11 and involve declarative sentences in which (a) a future indicative appears where Classical Greek might have used an optative, or (b) a subjunctive or optative is used where Classical Greek would have a future indicative (recall that subjunctives could not appear in this environment in Classical Attic). ‘Modal/potential’ future indicatives (15) a. οὐ καλὸς ἂν οὐδὲ σοὶ δόξω […] εἴ γε τοῖς σοῖς ἐπιτάγμασιν […] ὑπείξομαι. ‘I will/can/should not appear loyal even in your eyes, if I comply with your request’. Alexiad 1.16.6

Δόξω here could also be understood as the homophonous aorist subjunctive, but either way the presence of ἄν at least invites a modal/potential reading (though its use to mark potentiality in the Alexiad is at best sporadic). b. οὐ γὰρ ὡς πατρὸς φεισαίμην […] οὐδὲ τὰ κατορθώματα […] παραδραμούμεθα· ἐν ἑκατέροις γὰρ ἀδικήσομεν τὴν ἀλήθειαν.

18

This is not to say that there are no relevant phenomena in subordinate clauses. We might, for example, point to the high volume of future indicatives and optatives in both futurereferring and indefinite/generic conditionals, where they are used with exactly the same value as subjunctives (the standard option throughout the history of Greek): (a) (b)

Alexiad 11.3.2: ἐὰν ταύτην […] μοι δώσεις […] τελέσω τὸν ὅρκον κἀγώ: ‘if you give me this (sc. tent full of money), I too will take the oath’. Alexiad 11.6.3: ἀλλ’ ἐπειδὰν μάχη καὶ πόλεμος παρασταίη … ἀκάθεκτοί εἰσὶν […] εἰ δὲ […] λόχους οἱ πολέμιοι πολλάκις καταστήσαιεν καὶ τεχνικῶς αὐτοὺς μετελεύσονται, εἰς τοὐναντίον πᾶν τὸ θράσος αὐτοῖς περιίσταται: ‘but whenever war and fighting occur, they are unrestrainable […]; but if the enemy regularly sets ambushes for them and pursues them artfully, all their courage turns to its opposite’.

236

horrocks

‘I shall not spare him as (being) my father, nor shall I gloss over his successes; for in either case I should wrong the truth’. Alexiad 1.16.10

Though the conjoined clauses in the first sentence contain an optative and a future indicative, both probably express simple futurity because Anna has just stated that she will reveal all her father’s successes and failures, making subsequent hedging (e.g., ‘I would not spare him, sc. if I found anything wrong’) inconsistent. The future indicative in the second sentence, however, lends itself naturally to a modal reading after the preceding negatives, i.e., as a purely hypothetical consideration of the consequences of doing what you have just stated you will not do: cf. I shall not do x, otherwise I should/?shall do wrong. c. θᾶττον ἂν τὰς ψυχὰς προδοῖεν ἢ πονηρόν τι κατ’ αὐτοῦ μελετῆσαι πεισθήσονται. ‘They would sooner lose their lives than be persuaded to plot any treachery against him’. Alexiad 2.9.4

A clause with optative + ἄν is joined, in a comparison, with one containing a future indicative. Since the two parts of comparisons are normally given parallel readings (cf. they would rather do x than [they would/??will do] y), and since this is a description of customary behaviour, the modal reading is more appropriate in that it naturally allows for an indefinite number of occurrences at unspecified times: cf. they would/?will rather do x than y on any occasion. d. ἐγὼ μὲν τοῦ κάστρου παμμεγέθους ὄντος ἔν τινι τόπῳ κρυβήσομαι. ‘As the fortress is very large, I will/can hide myself in some place’. Alexiad 11.7.4

In this example a future reading perhaps seems more natural at first glance, but the anonymous metaphrast gives κρυβήσομαι an explicitly dynamic/potential interpretation through the use of δύναμαι:19

19

Kriaras’ dictionary of the Medieval Greek vernacular offers no examples of δύναμαι/δύνομαι used as a simple future.

‘high’ and ‘low’ in medieval greek

237

e. ἐγὼ μέν, ἐπεὶ τὸ κάστρον μέγα ἐστί, δύναμαι κρυβηθῆναι ἐν μέρει τινὶ τοῦ κάστρου. ‘Since the fort is large, I can hide myself in a part of the fort’. Metaphrase, section 8

Whatever we decide about the original future indicative, this rephrasing confirms that at least one native speaker of the later Byzantine era assigned a modal reading to it. ‘Future/potential’ (aorist) subjunctives (16) a. ἢν δέ τι […] νεωτερισθείη, αὐτοὶ μὲν πόρρω ἑαυτοὺς τοῦ δράματος ἐξάγωσι ‘But if any revolt occurs, they will keep themselves well out of the business’. Alexiad 1.2.7

The subjunctive in the apodosis of the conditional sentence follows a protasis consisting of ἤν (= ἐάν) + optative, an impossible combination οf elements in Classical Greek, where ἤν requires a subjunctive and the main clause would have a future indicative or perhaps a potential optative. The question, then, is what this was intended to mean. Since Alexius here is trying to persuade a mob to follow his advice, the context demands a confident statement of the damaging consequences of ignoring it, which in turn favours the reading ‘will’ over ‘can/may/would’ and supports taking the optative in the protasis as a traditional Atticizing variant of the normal subjunctive or future indicative (see the discussion of (9a) and (9b)). This conclusion is perhaps supported by Anna’s choice of ἤν, though her overall use of conjunctions with and without ἄν is far from classical. b. πίστεις […] ἐποιησάτην, ὡς […] εἰ εἰς τὸν βασίλειον θρόνον αὐτὸν ἀγάγοι Θεός, ἐκεῖνον ἐς τὴν τοῦ δομεστίκου ἀξίαν ἀναγάγῃ […] ‘The two gave assurances that, if God raised him (Alexius) to the Imperial throne, he would raise the other to the rank of Domestic […]’. Alexiad 2.4.7

Since the conditional sentence is the complement of a verb that refers to the past, the optative in the protasis might be understood as a classically sanc-

238

horrocks

tioned variant of the subjunctive. But in this situation the verbs of both clauses are usually made optative and here the main clause contains a subjunctive, i.e., the ‘original’ future-referring form of the corresponding direct speech (see the subjunctive in (16a)). On balance, therefore, the optative in the protasis is best taken as a free (post-classical) subjunctive variant and the main-clause subjunctive should be read as a simple future. c. κἀγὼ οὐκ ἀφέλωμαί τι […] οὐδέ τινός σοι ἐξουσίας ἐπικοινωνήσω ‘[…] and I will not take anything, nor share power with you’. Alexiad 2.12.2

Here the aged emperor Nicephorus has just begged the young Alexius to become his adopted son. Though in principle the subjunctives could be either future or potential in force (‘become my son and I will/could/would not […]’), this particular context favours a straightforward expression of good intent, i.e., a future reading.20 d. ἄτερ χρημάτων ἴσθι ὡς οὐδὲ φρούριον κατασχεῖν δυνηθῇς. ‘Be sure you will/would never be able to take even a guard post without money’. Alexiad 11.11.6

What was said about (16c) applies equally here, though in this case the metaphrast overtly interprets the subjunctive as a future by using the μέλλω-periphrasis: e. γίνωσκε ὅτι χωρὶς χρημάτων οὐδὲ ἓν καστέλλιον μέλλετε ἐπαρεῖν. ‘Be sure that without money you are not going to take even one little fort’. Metaphrase, section 100

‘Future’ optatives21 (17) a. τὰ […] κατ’ ἐμὲ διηγήματα […] ἐς δάκρυα τὸν ἀκροατὴν συγκινήσειε 20 21

The second could also be taken to be the homophonous future indicative, see (15a). See also (15b) for an additional (probable) example of the optative used as a simple future.

‘high’ and ‘low’ in medieval greek

239

‘But my tale will move the hearer to tears’. Alexiad, Prologue 4.1

Anna has just stated that Orpheus used to move stones with his singing and then adds that, though her own tale will not cause inanimate objects to move in the same way, it ‘will/can/could at least move the reader to tears’. Since the prologue is intended to inspire the reader to press on, a potential interpretation, with its implied ‘if you bother to continue’, would not be optimal and a future reading is therefore probably to be preferred. b. ὡς προϊὼν ὁ λόγος σαφέστερον παραστήσειε ‘[…] as the advancing narrative will relate more clearly’. Alexiad 1.4.1

In referring to later parts of her history Anna normally uses a future indicative: e.g., δηλώσομεν: ‘we shall show’ (1.1.2); διηγήσομαι: ‘I shall tell’ (1.15.3); ὡς ὁ λόγος κατιὼν παραστήσει: ‘as the narrative below will relate’ (11.4.5). The only other optative used in this way in the corpus selected for analysis is: ὡς […] ὁ λόγος […] διὰ πάσης σαφηνείας δηλώσειε: ‘as the narrative will show with all clarity’ (2.2.4). But in all these cases Anna can only be making a confident rather than a hedged assertion about the contents of her own book (cf. it should relate, sc. if I do my job), and the optative is therefore best taken as a future. c. ἐκ τῆσδε τῆς Κορυφοῦς διαπέμπω […] ἀγγελίας τῇ σῇ βασιλείᾳ, ἃς […] οὐκ ἂν περιχαρῶς ἀποδέξαιο […] ‘From this town of Corfu I send news to your Majesty which you will not receive with great joy’. Alexiad 11.12.6

Despite the presence of a ‘potential’ ἄν it is hard to see what implicit condition a truly modal reading of the optative might be made to hinge on.22 Accordingly, the metaphrast unambiguously interprets this example as a future, more specifically as a strong denial with οὐ μή (a construction in which Classical Greek would normally use a subjunctive, see (2d)): 22

It may, of course, have been intended cautiously or even ironically (‘[…] which you may not receive […]’), but this is not what the metaphrast has understood, as (d) clearly demonstrates.

240

horrocks

d. πέμπω σοι μηνύματα ἀπὸ τῆσδε τῆς πόλεως τῆς Κορυφὼ […]· […] οὐ μὴ δέξεσαι ταῦτα μετὰ χαρᾶς. ‘I send you messages from this town of Corfu; you will not receive these with joy’. Metaphrase, section 114

10.6

Conclusion

The examples in (15), (16) and (17), taken in conjunction with the explicit interpretations of the metaphrast, provide good evidence for the essential interchangeability of lassical future, subjunctive and optative forms in declarative sentences, an environment where they could not be substituted freely one for another in Classical Attic, either because their meanings were different (future indicatives and potential optatives) or because the relevant variants could not occur there at all (subjunctives). This might imply a rather random state of affairs (or a series of ‘mistakes’), but as we have seen, the vernacular had long before dispensed with the optative and used its innovative future-referring forms to express both simple futurity and epistemic modality. If the morphological resources of literary Medieval Greek had been reinterpreted to reflect the grammar of futurity/modality in the contemporary vernacular, we would therefore expect to find the attested distribution, i.e., one in which classical future indicatives and subjunctives are functionally equivalent (just like the various vernacular future/ subjunctives), and classical optatives are used as ‘literary’ variants of future-referring forms (in the traditional Atticizing way). The choice of form in any given case may be a matter of whim, some idiosyncratic preference or various contextual and pragmatic considerations (further investigation is required), but the fundamental organization of the future/modal domain is anything but random, whatever the vagaries of its actual realizations. This is no more than a first step towards the recharacterization of the language of Byzantine belles lettres as a variety of Medieval Greek with an underlying grammar closely related to that of the contemporary vernacular. There is therefore a great deal of scope for additional research and some promising areas of investigation have already been suggested in recent work (most importantly Stone 2009, and Hinterberger 2014), including the use of tenses, especially the perfect and pluperfect vis-à-vis the aorist, and the role and status of the middle voice vis-à-vis the active and passive. I plan to engage with these and other issues in the future, but I hope that others too will now feel

‘high’ and ‘low’ in medieval greek

241

inspired to reconsider the high registers of Byzantine Greek as a worthwhile, albeit indirect, source of evidence for the ‘true’ grammar of the medieval language.

References Browning, R. 19832. Medieval and Modern Greek. Cambridge. Davis, J. 2010. ‘Anna Komnene and Niketas Choniates “translated”: the fourteenthcentury Byzantine metaphrases’. History as literature in Byzantium, ed. by R. Macrides, 55–70. Farnham. Goodwin, W.W. 1890. Syntax of the moods and tenses of the Greek verb. Cambridge, ma. Hinterberger, M. (ed.). 2014. The language of Byzantine learned literature. Turnhout. Horrocks, G.C. 1995. ‘On condition: aspect and modality in the history of Greek’. PCPhS 41:153–173. Horrocks, G.C. 1997. ‘Homer’s dialect’. A new companion to Homer, ed. by I. Morris & B. Powell, 193–217. Leiden. Horrocks, G.C. 20102. Greek: a history of the language and its speakers. Chichester. Hunger, H. 1981. Anonyme Metaphrase zu Anna Komnene, Alexias xi–xiii. Vienna. Kambylis, A. & D.R. Reinsch (eds.). 2001. Annae Comnenae Alexias. Berlin. Mackridge, P. 1990. ‘Katharevousa (c.1800–1974): An obituary for an official language’. Background to contemporary Greece i, ed. by M. Sarafis & M. Eve, 35–51. London. Mackridge, P. 2008. Language and national identity in Greece 1766–1976. Oxford. Markopoulos, T. 2009. The future in Greek: from Ancient to Medieval. Oxford. Moravcsik, G. (ed.). 19672. Constantine Porphyrogenitus: de administrando imperio. Washington d.c. Stone, A.F. 2009. ‘The moods and tenses in Eustathian and late twelfth-century highstyle Byzantine Greek’. Βυζαντινά Σύμμεικτα 19:99–145. Wahlgren, S. 2002. ‘Towards a grammar of Byzantine Greek’. SO 77:201–204. Wahlgren, S. 2010. ‘Byzantine literature and the classical past’. A companion to the ancient Greek language, ed. by E.J. Bakker, 527–538. Chichester. Willmott, J.C. 2007. The moods of Homeric Greek. Cambridge.

chapter 11

Confusion of Moods in Greek Private Letters of Roman Egypt? Martti Leiwo

11.1

Introduction

The morphosyntax in Greek private letters of the Roman period shows considerable variation, perhaps even confusion, concerning the use of moods, and to some extent, of tenses. Here I will consider this variation and suggest possible causes. The discussion will proceed from typical linguistic differences between professional scribes and other more private letter-writers, to more detailed analysis of individual language usage. My data consist of private letters written on ostraca (potsherds) and papyri, with examples coming mainly from Mons Claudianus and the Fayyum because those areas offer the most interesting language usage for the present purpose, as the majority of the letters can be dated between 100 and 160ce.1 I am especially interested in detecting idiolectal language use whenever possible and am convinced that ongoing trends in language use can be detected in very small data sets of the kind I use, even if a large corpus is appropriate for showing actual language change.2 My focus is thus at the grass-roots level, not on changes in grammar at a general level, though I believe that the variation seen in idiolects also reflects more universal changes. Several lexical forms have a seemingly faultless appearance, but on closer inspection turn out to be tricky. In what follows, I will focus on such phrases, including the imperative, the infinitive, and the participle, and analyze variation in their external appearance as well as in their syntax. By external appearance, I mean the form as it is written on the ostracon or papyrus, since it is sometimes almost impossible to determine the grammatical or lexical category 1 All papyrological sources can be found at http://papyri.info. I am very grateful to Stefanie Kennell for her revision and copy-editing of my text. 2 In the study of variation and change, large historical corpora are naturally crucial for examining changes in the system. Idiolects, however, can tell us about uncertainties and ambivalence in choosing between different linguistic constructions. This variation may be a symptom of ongoing change.

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_012

confusion of moods in greek private letters of roman egypt?

243

of the form accurately. Obviously, there are multiple reasons for the opaqueness of the forms, because internal variation and contact-induced variation may have similar results although their origin is different. We must also keep in mind that some variation is always due to document type, in other words genre, and some peculiarities are simply faults due to a poor command of Greek generally. The use of the future indicative and the infinitive as directives instead of the imperative further blurs semantic transparency (see Mandilaras 1973:303).

11.2

Forms and Their Context

The forms discussed were chosen for the following reasons. Firstly, the imperative took the place of the participle in some grammatical constructions, thus changing their syntax (Leiwo 2010:101), and furthermore was subject to contactinduced forces in the more Egyptian-speaking environment (see below). Secondly, the infinitive became phonetically obscure, which led to its gradual attrition, even if Hellenistic and Roman Koine levelled the excessive dialectal differences in its morphology that were fundamental to its previous complexity (Langslow 2009:324). The complex situation is immediately evident in the confusion between the spelling of the imperative and of the infinitive, which seems to have been exacerbated by Greek-Egyptian language contact. Mandilaras (1973:309) stated that the infinitive holds its own, although analytical constructions developed into serious rivals. The main cause for its wide use, according to Mandilaras, is the development of the articular infinitive, with or without prepositions. This reason is dubious by itself, however, as the infinitive with an article was already popular in 5th and 4th c. bce Attic, and was also used on its own in all kinds of papyrus texts. Instead, there is a clear difference between professional scribes and individual writers, who moreover can be divided between those who use the infinitive frequently and those who use it only in well-known idiomatic phrases. I will show examples of this variation below. The attrition of the infinitive took a long time and during this period syntactic variation was intense (Leiwo 2012:6–7). Since many infinitives are old datives, formally speaking, the infinitive often gives the goal or result of the activity (Langslow 2009:328). This function is found especially in combination with verbs of going, sending, and giving. It is easy to understand that these infinitives could be replaced by consecutive and final clauses. Here I discuss mainly verbs of denoting, sending, receiving, saying, and knowing, where semantic, pragmatic, and phonetic reasons also played a substantial role in morphosyntax.

244

leiwo

Unfortunately, Mandilaras did not discuss the context.3 To get a clear picture of morphosyntactic variation, one must consider the document’s date and context, both pragmatic and sociolinguistic, individually as well as in terms of its genre, and ask why-related questions. The social context of many Egyptian settlements was multi-ethnic, though the quarry workers at Mons Claudianus, for example, seem to have come mainly from Alexandria (Cuvigny 2006). In other professions and at other sites, the situation was more complicated and the variation was greater, since many of the writers communicated in a language that was technically easier to write than their first language (l1). Their mother tongue, it must be emphasized, was seldom writable at all during the 2nd c. ce, and the Coptic alphabet was not yet used (Layton 2000:1). Only very few Egyptians were able to write the demotic script. In addition, many of the writers, both Greeks and native Egyptians, had far from perfect writing skills in Greek. Yet among the workers in the Roman fortresses between the desert and the Nile were many individuals who could actually write. This fact must be stressed, as it may change our views about writing skills in antiquity.4 For the above-mentioned reasons, morphological analysis can be knotty. The external appearance of some verb forms opens the way to a different syntactic analysis, as can be seen from these three examples:5 (1) καλῶς πυή[σεις], ἄδελφε, πέμψε μοι τὸ τοῦτο ὧν [πέπ]ρακες (δραχμὰς) δ (πεντώβολον). ‘Please, brother, send me the money you got after selling, four drachmas five obols’. O. Claud. 2.243, petenephotes, mid-2nd c.

(2) καλῶς πυήσις πέμψε μοι (δραχμὰς) δ. ‘Please, send me four drachmas’. O. Claud. 2.246, petenephotes, mid-2nd c.

3 See the examples in Mandilaras (1973:299, section 703), where he argues that those examples ‘show clearly the retreat of the imperative (i.e., the aorist imperative) before the jussive subjunctive, or the imperatival infinitive’. All the examples (P.Oxy 1844–1859) are very late, however, from the 6th and 7th centuries, and their genre is extremely polite, even ‘Byzantine’, which makes it impossible to use the imperative in such a written context. 4 As opposed to Harris (1989). 5 All translations are mine if not otherwise indicated.

confusion of moods in greek private letters of roman egypt?

245

(3) κα[λ]ῶς ποιήσις, ἄδελφε, πέμ̣ ψ̣ε μοι ἄρτους δύω. ‘Please, brother, send me two loaves of bread’. O. Claud. 2.284, zosimos, mid-2nd c.

The editors of the ostraca had difficulty in deciding between the infinitive and the imperative. In (1) πέμψε is analysed as an imperative, but in (2) the same form is taken to be either an imperative or an infinitive, and in (3) the same form is again said to be an infinitive.6 I have shown elsewhere (Leiwo 2010:101– 106) that the most popular complements of the idiom καλῶς ποιήσεις in the letters from Mons Claudianus are either the participle, which is standard in literary texts, the imperative, or the imperatival future indicative, whereas the infinitive seems to be rare with this idiom, although it is attested in papyri from other places.7 How should we analyze these forms? In the following pages, I will discuss this in detail, starting with the differences between different kinds of writers.

11.3

Professional versus Idiolectal Usage

Scribes could produce decent standard language, which makes the difference between a literary individual and a more or less professional scribe remarkable. A nice example of a document by a person with a good command of Greek, especially infinitives, is a letter from a woman called Grata.8 She signed one letter to a Successus, who was probably an officer at the quarry (O. Claud. 1.111), and Successus himself signed two or three letters.9 These letters, as well as those of Arteimas (O. Claud. 1.127) and of an unknown sender (130), are all in the same hand, so a scribe was used (O. Claud. 1.126; ca. 107 ce):

6 Edited by A. Bülow-Jacobsen (1, 2) and J. Bingen (3); all subsequent translations and remarks by Bülow-Jacobsen come from the O. Claud. volumes (see Bibliography below). 7 For standards in Egypt, see Vierros (2012:34–35). 8 Bülow-Jacobsen notes that the name can just as well be read as Γράτα⟨c⟩, which would be a minor change before the initial sigma of Successus (O. Claud. 1.113). 9 There is in fact a small archive in his name in O. Claud. 1.124–136. An interesting detail of this archive is that persons with Latin names (Rusticus, Ponticus, Grata, and Domitius) wrote to Successus in Greek, while those with Greekish names (Athenor 131, Agathon 135) wrote in Latin, which reminds us to be very careful in associating a language with a name if we know nothing else about the social context. A Greek or Roman name does not tell us about the l1 of the person, whereas an Egyptian name is not at all typical of native Greek or Latin speakers.

246

leiwo

(4) Γράτα Σουκκέσσωι τῶι ἀδελφῶι πλεῖστα χαίρειν. μετέδωκα Ἀρτειμᾷ ὑπὲρ τῶν ρ̅ μ̅ ἀσκῶν καὶ ἀπεκρίθη μηδὲ εἰληφέναι αὐτοὺς μηδ’ ἑωρακέναι. ἔπεμψα Γέταν πρὸς τὸν σησκούπλικα καὶ αὐτὸς εἶπε μὴ εἰληφέναι. ἐφώνησε δὲ Ἀλέξανδρον τόν ποτε κουράτορα καὶ εἶπε ἐκεῖνος εἰληφέναι καὶ πάλι Καστορᾶτι αὐτοὺς παραδεδωκέναι. [ὁρ]ῶ̣ ν Ἀρτειμᾶς ἃ ἔπεμψες ἐγόνγυσε [λέγ]ων ‘οὐαί μοι τῷ ἐλειδίῳ χορηγῖν̣ [οὐ]κ̣ ἔχω.’ ἀσπάζου Σαβεῖνον τὸν [ἀδ]ελφόν μου καὶ πάντες. [ἐρρ]ῶται ἡμᾶις εὔχωμαι. ‘Grata to Successus, her brother, many greetings. I informed Artimas about the 140 water-skins, and he answered that he had not got them nor even seen them. I sent Geta to the sesquiplicarius and he said that he had not got them. But he called Alexander, the former curator, and that man said he had received them and in turn passed them on to Kastoras. Seeing what you had sent, Artimas said, grumbling, “Oh no, poor me, I can’t supply??” Greet my brother Sabinus and everyone. I pray for your health’.10 The narrative governed by καὶ ἀπεκρίθη (‘and he answered’) καὶ αὐτὸς εἶπε (‘and he then said’) καὶ εἶπε ἐκεῖνος (‘and that man said’), contains a series of positive and negative passive infinitives, but no complement clauses such as ὅτι or ἵνα. Complement clauses are used especially when the indirect object is stated.11 Grata’s scribe was well educated; he had a good literary hand and excellent command of the official standard language, though just this kind of infinitival structure is rare in our sources because letters seldom have long indirect narratives. An identical structure can be found in bgu 8.1873 from

10

11

The ostracon is in two pieces, with the word between the two pieces in line 10 read by Bülow-Jacobsen as ἐλειδίῳ, which he takes to be a diminutive of ἔλεος, ‘pity’, but it is unattested until this date. Daniel (1994:61–62) suggested a correction to ἐλαδίῳ because he regarded abstract diminutives as very rare, and even the meaning of ἐλειδίῳ would be strange here, as the water-skins had already been located. The correction makes sense (‘Oh, poor me, I can’t supply oil’), but the scribe is very competent and would not have made such a spelling mistake. As I cannot verify the reading of line 10 from the photo in plate xv, I prefer to leave the question open. See P.Mil.Vogl 2.50 (108–150 ce): εἰπὲ οὖν Ἥρωνι ἵνα παραβάλλῃ ἐκεῖ.

confusion of moods in greek private letters of roman egypt?

247

Bousiris (Herakleopolites), however, but the date (60–61 or 53–52 bce) is c. 150 years earlier:12 (5) καταντήσας οὖν πρὸς τὸν γεωργὸν Παποντῶν περὶ τῶν τεσσάρων ἀρου[ρῶν] εἶπε μὴ δύνασθαι ἑτέρῳ γένει εἰ μὴ ἀράκωι σπαρῆναι, ἐπιδέχεσθαι δ’ αὐτόν, ὥς φησιν, [τοῦτο κατ’] ἄρουραν παρελέγξαι … ‘When coming then to the farmer Papontos asking about the four arourae, the man (sc. Papontos) said that he cannot sow other seed except sweet peas, but allow him, as they say, to examine …’. It is, however, certainly not the idiolect of Grata herself that is represented in the letter, but that of the professional scribe, whereas a more transparent idiolect seems to lurk behind the text of Leontas (O. Claud. 1.129; ca. 107 ce) from the same archive. (6) Λεοντᾶς Σουκέσῳ χαίρειν. γεινώ[σ][κ]ειν σε θέλω ὅτι οἱ ἀσκοί μου πάντες ἄ̣χ̣ρηστο⟨ι⟩ γέγοναν καὶ τὸ ὕδυρ σπανί[ζε]ται. καλῶς οὖν ποιήσεις ἐὰν θέλῃς [πέμ]ψαι μοι ἀσκοὺς καινοὺς καὶ γράψον μ̣ [οι]. ἐὰν δὲ μὴ θέλῃς, διάλεξόν μοι ἐκ τ[ῶν] παλαιῶ̣ ν ἕως Ἐπαφρόδειτος ἀν[αβῇ]. καὶ σχοινείν μοι πέμψον σεβ[ένινον καὶ] διπλοίδα{ν} κ[αὶ τ]ὴν κλείμακα … ‘Leontas to Successus greetings. I wish you to know that my water-skins have all become useless and the water is sparse. Please send me new water-skins if you will, and write to me. If you will not, please choose from among the old ones until Epaphroditus comes up. And send me a rope of palm-fibre and a double cloak and the ladder’ (translation A. BülowJacobsen). 12

An entire example that can be traced using the papyri.info website and searching for ‘#ειπε#’ is bgu 4.1141, l. 49 (14–13 bce): εἶπεν ὁ γέρ̣ ω ̣ ̣ ν̣ μ[ὴ] εἰδέναι αὐτὸν τὸ καθόλον περὶ τούτων μηδέν. For a positive clause, see p.ifao 2.3, l. 12–13 = CdE 46 (1971), p. 133 (1st c. bce–1st c. ce): εἶπε ἐν ἀσφαλεί αὐ[τὰ τε]θεῖσθαι.

248

leiwo

A different scribe wrote the text, but the hand is very accurate and clear, though more documentary than literary, with all letters carefully drawn.13 The letter has some orthographic variation,14 but since it does not include narrative, its register is different from that of (4), consisting only of a series of commands presented with the imperative. We have the phrase καλῶς οὖν ποιήσεις combined with ἐὰν θέλῃς [πέμ]ψαι, where [πέμ]ψαι is again ambiguous. However, the letter continues with the imperatives γράψον, then ἐὰν δὲ μὴ θέλῃς, διάλεξόν μοι, so perhaps [πέμ]ψαι should also be taken as an imperative (= πέμψε, see below). In other respects, the linguistic comparison seems unlikely, since the two distinct registers use different syntax, and as a rule indirect narrative contains more syntactic slots for the infinitive than commands in the imperative mood. In the final section of this interesting letter from Grata (4), the scribe suddenly jumps into direct discourse, introducing it with the present participles [ὁρ]ῶ̣ ν and ἐγόνγυσε [λέγ]ων.15 The use of the nominative λέγων, we observe, seems to be idiomatic to some writers, as for instance Ailouras in O. Claud. 1.161, lines 4–5 (100–120ce): ἔπεμψάς \μ/οι Πανε̣κ̣ῶσιν λέγων ὅτι ἦρκας μου τ̣[οὺς] ἄρ̣τους καὶ τὸν κιθῶνα τέ[θηκας] ἐνέχυρον. The participle λέγων should follow the accusative Πανε̣κ̣ῶσιν: ‘You sent me Panekosis to say that you had taken my bread and impounded the chiton’. Surprisingly, the Latin vae is used as an interjection in Grata’s letter, revealing that it had a role in this speech community, where Latin was also used (see note 9 above). This sudden change to direct discourse seems to have pragmatic or narrative purposes, as the story proceeds from the more remote past to the recent past. The use of direct speech is an interesting phenomenon as a whole, but should be studied separately.16 Scribal standard and idiolectal variation is found in the letters of Lucius Bellienus Gemellus.17 He was a veteran, an ex-legionary, and his name (Bellienus) strongly suggests a person of Italian descent recruited to the legions.18

13 14 15 16

17 18

See the photo in plate xvi. E.g., ὕδυρ (ὕδωρ) and σχοινείν (σχοινίον). The left side of the ostracon is corrupted. sb 18.13867, lines 24–26 (2nd c. ce) contains an interesting mixture of direct and indirect speech: τῇ ἐφαύριον, ὅτε [οὐκ]έ\τι/ ἐνῆ πεῖν, ἐστάθη μοι λέ[γω]ν̣ ὅτι εἰ θέλις μνᾶν κρέος [σοι] ἀγορασθῆναι; λέγω· ναί. ‘On the next day, when there was not even a bit to drink, he stood up, saying to me, “Do you want a mina of meat bought for you?” I said, “Yes”’. Halla-aho (forthcoming) discusses these letters from the scribal point of view. Lucius Bellienus Gemellus, b. 32 ce, settled in the Arsinoite village Aphrodites Berenikes Polis after his legionary career ended. He owned olive groves, livestock, arable land,

confusion of moods in greek private letters of roman egypt?

249

Gemellus had obviously learnt Greek during his long stay in Egypt, but it was perhaps not his original l1, although he was a fluent l2 speaker.19 As evidence for this, I call attention to his consistent habit of always writing the standard Greek πεμψ- as πεμσ-, an assimilation not typical of Greek, though not of Latin either (but see, e.g., Morani 2000:188). Even if this cannot be considered conclusive evidence, the persistence of this spelling has only a few parallels.20 I present examples of the language usage of the scribe and Gemellus himself in three letters dating to 94, 99, and 100ce respectively.21 The first letter (7) is written by a professional scribe in a well-formed uncial hand of literary type and the others, (8) and (9), by Bellienus Gemellus himself, who was sixty-two years old in the year 94. (7) Λούκιος Βελλιῆνος Γέμελλος Ἐπαγαθῶι τῶι ἰδίωι χαίρειν. εὖ ποιήσεις κομισάμενος μου τὴν [ἐ]πιστ[ο]λὴν ἀναγκάσας ἐκχωσθῆναι τὸ ἐν αὐτῶι κόπριον … ‘Lucius Bellienus Gemellus to his own Epagathus, greeting. On receipt of my letter please have the manure there banked up’ (translation GrenfellHunt). P.Fay 110, 1–5

(8) Λούκιος Βελλῆνος Γέμελλος Ἐπαγαθῶι τῶι ἰδίωι χα(ίρειν). εὖ πυήσις διῶξαι τοὺς σκαφήτρους τῶν ἐλαιώνον …

19

20

21

and vineyards in many Fayyȗm villages, including Euhemereia, Dionysias (Qasr Qarȗn), Apias, Senthis, Psennophris, Psinnachis, and Prophetes; see Ast-Azzarello (2013:19). His estate manager Epagathos administered his fields. The archive consists of 83 confirmed documents (28 have only been described, 33 are still unpublished), 3 undetermined, and 1 related. Of the 50 published/described documents, 8 were discovered with the other texts of the archive, but their link with the archive is unclear. The other 42 documents consist of 35 letters, 3 contracts and 4 accounts. For the Gemellus Archive, see Hohlwein (1957) and Ast-Azzarello (2013). The texts: www.trismegistos.org/arch. Gemellus cannot plausibly be said to have been of Graeco-Egyptian descent, though it has been so argued: see Hohlwein (1957, 70–72); P.Oxf 10, 40–41 (E.P. Wegener). The gentilicium Bellienus is quite rare and associated with Italy: Schulze-Salomies (1991:429–430). At Mons Claudianus, a certain Maximus and a curator called Teres show the same persistence (O. Claud. 1.138 and 139, 110 ce; O. Claud. 2.366, see also 374); both belong in the Roman army context. P.Fay 110, 112 and 114 (ed. Grenfell-Hunt). Mandilaras (1973:299) offers the parallel as an example of the aorist imperative used to give an order that demands immediate execution.

250

leiwo

‘Lucius Bellenus Gemellus to his own Epagathus, greeting. Please carry forward the digging of the olive-yards’ (translation Grenfell-Hunt). P.Fay 112, 1–3

(9) Λούκιος Βελλῆνος Γέμελλος Σαβίνῳι τῶι οιεἱῶι χαίρειν. εὖ οὖν πυήσας κομισάμενός μου τὴν ἐπιστολὴν πέμσις μυ Πίνδαρον εἰς τὴν πόλιν τὸν πεδιοφύλακα τῆς Διονυσιάδο(ς) … ‘Lucius Bellenus Gemellus to his son Sabinus, greeting. On receipt of my letter you will oblige me by sending Pindarus, the guard at Dionysias, to me at the city’ (translation Grenfell-Hunt). P.Fay 114, 1–7

The syntactic variation between the participle ἀναγκάσας (7) and the infinitive διῶξαι (8) with the construction εὖ ποιήσεις is in line with the ostraca letters from Mons Claudianus, even if the imperative was far more popular (Leiwo 2010:105). But διῶξαι could be analysed as an imperative as well, by interpreting it as the aorist stem {dioks} + the 2nd person imperative ending {e}, just like πέμψε (see examples (1–3) above, and below). I consider it unlikely here, however, since the infinitive is, as mentioned above, a known variant in this syntactic structure (Leiwo 2010:106 n. 14), and Gemellus seldom had difficulties with the orthography of the grapheme ⟨αι⟩. The only confusion in the letters published thus far, it should nonetheless be underlined, occurs in this letter (P.Fay 112, 15): διω μένφομαί σαι μεγάλως. In this respect, the interpretation of the form διῶξαι remains slightly uncertain. In P.Fay 114 (example (9)) from the year 100, Gemellus wrote the participle κομισάμενος similarly to that in (7) but continued with an imperatival future indicative especially typical of military language: εὖ οὖν πυήσας κομισάμενός μου τἠν ἐπιστολὴν πέμσις μυ Πίνδαρον (‘On receipt of my letter please send Pindarus’). He also wrote an unexpected nonstandard imperative form in line 12 of letter 112 (8): ἐπέχον τῷ δακτυλιστῇ Ζωίλωι καὶ εἵνα αὐτὸν μὴ δυσωπήσῃς.22 This imperative form, with present stem ἐπέχand aorist imperative ending -ον, is interesting because it is the exact opposite of the form πέμψε with aorist stem and present imperative ending.23 Analysing the letters in greater detail makes it evident that Gemellus had learnt some idiomatic expressions (e.g., εὖ ποιήσεις κομισάμενος μου τὴν ἐπιστο-

22 23

The form is a hapax, and can be a simple slip of the pen, but as Gemellus usually does not make such slips, I prefer to take the form seriously. This usage, of course, has as its model the 2nd aorist imperative.

confusion of moods in greek private letters of roman egypt?

251

λήν), but turned easily to commands, as in P.Fay 113: εὖ οὖν πυήσας ἐξαυτῆς πέμσις αὐτὸν ἐξαυτῆς (‘Please then, immediately send him immediately’). Consistent with his habit, Gemellus again used the imperatival future indicative with this politeness idiom.24 His son, Sabinus, however, regularly used the aorist participle with this construction,25 showing unmistakably the social advancement so common among sons of veterans. Sabinus went to school and learned the scribal standard there.26 This also demonstrates how the infinitive/imperative and the future indicative took the place of the participle in this idiom in cases where the letter was written by a person who was literate but not educated in the use of scribal standard language (Leiwo 2010:99–106). It is moreover obvious that the politeness formula εὖ ποιήσεις/καλῶς ποιήσεις became a politeness idiom signifying ‘please’ with no direct syntactic relation to its complement, as in P.Fay 120 by Gemellus: εὖ πυήσις π[έ]μσ[ις] μυ θρ[ί]νακες δύωι καὶ λικμητρίδε̣ς δύωι καὶ πτύ⟨ο⟩ ἕν, ἐπὶ κ[ο]λάζωμαι αὐτῶν εἰς Ἀφροδίτην πόλιν (‘Please send me two forks and two shovels and one winnowing-fan, as I need them at Aphroditopolis’).27 To sum up, we can confidently state that the narrative continues with the nominativus/accusativus cum infinitivo structure in a well-written professional scribal style, but even in a slightly less elaborate style complement clauses become frequent. Also with the imperative and when the person to whom something is to be said is mentioned, complement clauses are favoured, but this variation requires more research.28 I would suggest that the use of the complementary infinitive with verbs of saying is more a matter of register and scribal professionalism than that of diachronic causes, which seem less important during the 2nd c. ce. 24

25 26 27

28

As Grenfell-Hunt comment on P.Fay (1900:262): ‘His (Gemellus’) advanced age is reflected in his handwriting, which tends to become shaky and illegible. But it was perhaps never very good, any more than his spelling and grammar, which are peculiarly atrocious. His sons Sabinus and Harpocration show a better acquaintance with the Greek language, though they are not above reproach’. It is now clear that Sabinus wrote the majority of his letters himself (see Ast-Azzarello 2013:24–25), which means that he had a very good command of orthography and grammar. E.g., P.Fay 121, 122. A full edition of all the papyri belonging to this archive (totaling ca. 100) is in progress: see Ast-Azzarello (2013:19). A detailed analysis of the handwriting is presented by Ast-Azzarello (2013:24–25). Gemellus very consistently writes υ for οι or, less frequently, the other way round. In the development of Greek internal phonology, /y/ and /oi/ were in the process of merging in the Roman period, with both representing the rounded front vowel /y/; see Horrocks (2010:67). See also bgu 1.93, l. 29 (2nd to 3rd c. ce); cpr 7.54, l. 20 (2nd c. ce).

252 11.4

leiwo

Contact Influence?

Using papyrological methods such as analysis of document type, handwriting characteristics, and standards of orthography, writers of private letters can, to some extent, be identified and grouped together. When the linguistic output of the identified writers is analysed, we can see that some of them were very consistent in their constructions and orthography, while others did not care at all and varied their language all the time. In the following section, I will focus on forms of the verb πέμπειν that are infinitives or imperatives, but will also discuss other nonstandard or otherwise exceptional usages. One interesting writer is Ammonis (O. Claud. 4.892, 2–4), who seems to use the infinitive πέμψαι along with the scribal standard usages with the verb form ἐδ]ήλωσας, aorist indicative 2nd person singular: (10) [ἐπεὶ ἐδ]ήλωσας πέμψαι ἀνθρώπους [εἰς τὴν μ]η̣ χανὴν εὐθέως ⟨σ⟩οι πλεῖον̣ ἔπεμ[ψα. ‘Since you let me know to send men for the machine, I immediately sent you more’. Another potential explanation exists, however, since the structure can be analysed as direct speech after [ἐπεὶ ἐδ]ήλωσας, in which case πέμψαι would be an imperative. I prefer the infinitive structure because the clause is easily intelligible as such. But then, with the imperative ἴσθι, which in itself is very exceptional in these letters, Ammonis had to solve a linguistic problem involving an action in the past; instead of a complement clause with ὅτι, he constructed an active perfect participle complement in the accusative (O. Claud. 4.892, 6–7), when the most common usage would have been the complement clause with ὅτι:29 (11) ἴσθι δὲ τὸν κῆπεν τετραυ|ματικότα ἅλιν. ‘Know that the garden has been damaged by salt’. As for the active perfect participle, the editor (A. Bülow-Jacobsen) states that τετραυματικότα ἅλιν cannot stand as read and argues that we must understand

29

Mons Claudianus offers seven examples of ὅτι with a verb of knowing: O. Claud. 1.127, 9; 1.156, 8; 1.173, 3; 2.379, 8; 2.386, 10 (fragmentary); 4.889, 6; 4.889, 9. This is the only letter with the imperative ἴσθι; 2.379, 8 has γίνωσκε ὅτι.

confusion of moods in greek private letters of roman egypt?

253

τετραυματισμένον ἅλι{ν} (‘has been damaged by salt’). It seems clear, however, that the writer did not consider the form ungrammatical; instead, he tried to form a past tense construction, although the active perfect participle of this verb is not attested in the tlg. This construction is unique in the Mons Claudianus data. Confusion again arises with the grapheme of the final unstressed vowel in the word κῆπεν (= κῆπoν), thus supporting its reduction to schwa, very common in Egyptian Greek (Horrocks, 2010:112). A man named Petenephotes composed several letters in Greek. I have analyzed his language elsewhere (Leiwo 2005:248–254; Leiwo 2010:114–118), but would like to suggest a possible new explanation for some of his linguistic idiosyncrasies here. The greatest variation again occurs in the basic verb of letter-writing, πέμπω: the forms πέμψε and πέμψον, πέμψεν, πέμψων are all used in nearly identical contexts.30 While πέμψε can be interpreted both as an infinitive and as an imperative, the others seem to be imperatives. Petenephotes used πέμψε with the idiom καλῶς πυήσις (O. Claud. 2.246,4), but in another letter with exactly the same idiom, having forgotten to write the verb in the first place, he added the aorist participle πέμψας μοι above the line, thus showing that he—while in a correcting frame of mind—knew the standard usage for this politeness formula (O. Claud. 2.245, 4). But he never wrote that form, spelling it as πέμψαι; it might be of some significance that the exact spelling is otherwise attested only once in the Mons Claudianus material.31 This means that forms such as πέμψε, πέμψεν, πέμψον, and πέμψων are used in the same context with the same meaning. Is it possible to find explanations other than phonological confusion of the unstressed vowels e/o? Was the writer familiar with all the endings of Greek present and aorist active imperative and aorist infinitive forms? Can we find explanations for these forms other than bad command of Greek? I suggest that we can add something here from the morphophonology of Egyptian. The supporting evidence can be sought in Coptic even if the written Coptic sources are later.32 The Coptic language had a strong stress accent (Till 1950:13–32; Girgis 1966:73) and its verbs followed a system different from that

30

31

32

O. Claud. 2.243, 3 πέμψε and 11 πέμψον; 2.246, 4 πέμψε, 6 πέμψεν; 2.249, 8 πέμψων. The form πέμψεν is also attested in a letter of Maximos (O. Claud. 2.262, 3): γράφις μυ ὅτι ‘πέμψεν μυ φάσιν ἀπὸ το̃ν ἰς ὖκον’. O. Claud. 4.884,9 ἐὰν δὲ πέμπω ἕνα οὐκέτι ἔχω πέμψαι ἰς τὴν λατομίαν. οἶδες δὲ ὅτι ὄχλον πολ[ὺ]ν̣ δαπανᾶ ἰ[ς ̣ τὴν κατάστ[ασι]ν, and without the occlusive /p/ in O. Claud. 1.166, 6: κα̣ι ̣̀ οὐ γέγονέ μοι ἐν̣ τούτοις πέμσα̣ ι ̣ σοί τι. Egyptian phonological influence on Greek vowel orthography has been studied by Sonja Dahlgren (forthcoming, 2017), who uses Greek loanwords in Coptic as evidence within the framework of l2ws (second language writing system) studies.

254

leiwo

of Greek (Layton 2000:125–127). The strong stress accent reduced the vowels in unstressed syllables to /e/ or /ǝ/, and sometimes to other vowels if the surrounding consonants gave colour to the vowel in question. In such cases, the vowels were written with the letters ⲁ ⲉⲓ (ⲓ) ⲟⲩ ⲟ ⲏ and ⲱ (Girgis 1966:73). The question then is, does this strong reduction represent a situation that already existed in the 1st and 2nd c. ce? I would argue that this rule was already prevalent in speech centuries earlier and that it appeared in Greek in Egypt when individuals with Egyptian as their l1 wrote Greek. I also believe that the internal Greek development that affected vowels and diphthongs, raising them toward /e/ and /i/, was another factor that confused Egyptian l1 speakers, who were taught standard Greek orthography. Together, these influences probably caused those who were not Greek l1 speakers even more difficulties than they did Greek l1 speakers. What, then, can be said about the potential confusion of moods?

11.5

Confusion of Moods?

Greek verbs borrowed by Coptic have only an infinitive form,33 which takes a form resembling the Greek active imperative singular in the Sahidic dialect (Layton 2007:70–71; Till 1961:40) producing borrowed infinitives such as ⲡⲓⲥⲧⲉⲩⲉ, ⲃⲁⲃⲧⲓⲍⲉ, ⲙⲉⲧⲁⲛⲟⲉⲓ, ⲡⲁⲣⲁⲇⲓⲇⲟⲩ etc.34 This type of borrowing of verbs, in almost their root form or another structurally simple form such as the imperative or present tense base, is common. Sometimes the imperative is what is in fact taken over (Wohlgemuth 2009:76, 78–80). In the Sahidic dialect, the imperative indeed seems to be the form borrowed. Furthermore, in most native Coptic verbs, the imperative form was identical to that of the infinitive (Layton 2000:126). Could this have created confusion between the two categories? In the graphemic variants of Petenephotes, the variation in forms even in the same ostracon letter clearly has a phonological basis, as all the endings seem to blend together, apparently representing the Egyptian-influenced phonetic form [ˈpempsə]. The phenomenon present in Petenephotes’s use of unstressed Greek /e, o/ is evident in the Coptic usage of Greek loanwords. An identical confusion with πέμψον/πέμψε can be seen in loanwords such as κεράτιον (carat), καμίσιον (shirt), σάβανον (towel) > ⲕⲉⲣⲁⲧⲉ, ⲕⲁⲙⲓⲥⲉ, ⲝⲁⲃⲁⲛⲉ, as well as in

33 34

For borrowing of verbs in general, see Wohlgemuth (2009:75–76). Other Coptic dialects had different outcomes when the Greek infinitive was borrowed: Till (1961:40); see also Wohlgemuth (2009:80–81).

confusion of moods in greek private letters of roman egypt?

255

names, for example, Ἀντώνιος, Θεόδωρος > ⲁⲛⲧⲱⲛⲉ, ⲑⲉⲟⲇⲱⲣⲉ (Girgis 1966:84– 85; Dahlgren-Leiwo, forthcoming). Though written Coptic is later than my data and though there was a gap in the writing of Egyptian, we have no reason to suspect that the phonetics behind these Coptic loanwords did not exist earlier. During the 1st and 2nd c. ce, no Egyptian script was available or even suitable for private correspondence between individuals. Therefore, it is only natural that Greek was the script widely used. But as l1 phonology has an effect on writing l2, l1 impact on Greek l2 spelling may reasonably be sought (Dahlgren, forthcoming, 2017).35 We can also approach this confusion of unstressed vowels and potential confusion of categories through another verb, κομίζω, which is used with the meaning ‘to receive’. The standard scribal form of the 2nd person aorist imperative is in the middle voice (κόμισαι, κομίσασθε), but many writers use the imperative in the active voice (κόμισον, κομίσατε). This confusion between the active and middle forms is strongest in the letters by Dioscorus, who was a small-scale vegetable transporter in the Mons Claudianus area. Dioscorus himself wrote with a trained hand, but, as Bülow-Jacobsen puts it, had a ‘shaky relationship with Greek syntax, especially with regard to the use of cases’ (O. Claud. 2, p. 43; see also Leiwo 2003:83–86).36 Dioscorus can with confidence be considered an l2 Greek speaker who was well trained in the Greek script but had various difficulties with the scribal standard. A sufficient example is O. Claud. 2.229 (mid-2nd c. ce): (12) Διόσκορος Δράκων καὶ Ερεμη̣ [σις] καὶ Ἀμμωνιανὸς ἀμφοτέροι[ς τοῖς] φιλτάτοις πολλὰ χαίρ(ειν). κόμισον [παρὰ] Πουῶνσις δέσμην καυλίων [εἰς] ἕκαστον. ἐὰν ἐκομισάμη̣ ν τὰ κ[αυ-] λία, γράψον μοι καὶ περὶ τῆς σωτηρίας ἡμῶν. ἐρρωσ[

35

36

Evidence also exists for the lesser importance of vowels in scripts that do not use them or that use them in a way different from sound-based scripts (collected by Cook-Bassetti 2005:1–56). Dioscorus’s correspondence is published in O. Claud. 2.224–242. I have dealt with his typical greetings (Leiwo 2003).

256

leiwo

‘Dioscorus to Dracon and Eremesis and Ammonianus, all his friends, many greetings. Receive from Pouonsis a bunch of cabbage for each. If you receive the cabbage, write to me, also about your health. Farewell’. As regards Dioscorus’s knowledge of Greek verb morphology, it is interesting to see that he does have a certain command thereof, using the correct 1st person middle aorist indicative form ἐκομισάμη̣ ν, for example. Unfortunately, he used it where the information needed in the letter would require the 2nd person plural ἐκομίσασθε. This wrong form derives from the formulaic language, since Dioscorus uses the 1st person forms correctly in O. Claud. 2.233 and 234, though both texts are fragmentary. In contrast, he never uses the middle aorist imperative of the same verb and always writes this seemingly active voice form as κόμισον, whereas in other verbs he shows variation of the middle vowels of the type mentioned above: φρόντισέν μοι (= φρόντισον 2.225, 14), πέμψε (2.227, 11), and πέμψον (2.232, 8). He seldom uses infinitives at all; when he does, they appear in well-known idiomatic phrases like πολλὰ χαίρειν. One interesting example of another kind of usage does exist, however: ἔλαβαν ἡμῶν […] ἔχωμεν λαβεῖν (O. Claud. 2.232, 5).37 Unfortunately, the context is fragmentary and the exact meaning escapes us, being something like ‘I have received your […], we have received the […]’. In any case, this usage recalls the Modern Greek periphrastic formation of the active perfect indicative, of the type έχω λάβει (‘I have received’); it would be tempting to see an early stage in the development of this periphrastic perfect here. All in all, the active form κόμισον is attested fourteen times and the scribal standard κόμισαι twice as often (twenty-eight times) in the Mons Claudianus data.38 A simple explanation for these results is that some writers used the verb in the active voice to mean ‘receive, have, take’. A contact-induced explanation for the whole confusion of unstressed vowels of the type described here is that they were difficult for Egyptian l1 speakers, and the weakness of the vowels is reflected in the written record, thus generating forms that could suggest morphological variation of the type πέμψον ~ πέμψε, where the present tense ending replaces the aorist ending. As this change did in fact take place, and the active aorist imperative ending -on was replaced by the active present tense ending -e, for example, ἄκουσον > άκουσε (Modern Greek), internal reasons of language must have operated as well. This does not

37 38

To be understood ἔλαβα ὑμῶν […] ἔχομεν λαβεῖν. The diacritical markings are of course made by the editor. A search on all values via papyri.info produced 36 hits for κόμισον against 133 for κόμισαι, indicating that Dioscorus used the variant κόμισον more often than any other letter-writer.

confusion of moods in greek private letters of roman egypt?

257

mean that contact-induced reasons had nothing at all to do with the merging of unstressed /e/ and /o/ in Egypt, however. This fusion was certainly influenced by Egyptian, and multicausality must be accepted for increasing the confusion between imperatives in -e and in -on, thus simplifying verbal morphology. The syntactic and semantic context of these forms does not, however, suggest that morphological confusion was usual between infinitives and imperatives of the type πέμψε = πέμψον (imp) and πέμψαι (inf), where the writer did not recognise which was the imperative and which the infinitive. On the contrary, the majority of the writers did recognise the difference. A clear enough example of this is O. Claud. 2.275 (125–175ce), which was written in a good hand: (13) Ἀπολινάρις Σονσναῦτι το̃ ἀδελφο̃ χα(ίρειν). κόμισε διὰ Ληωνίδου (δραχμὴν) α (τριώβολον) καὶ ἀγόρασόν μοι τεμάχια καὶ δὸς Ἀχιλλᾶτι ὀνηλάτῃ ἐπὶ εἰς Αἴγυπτον θέλο πέμψε. ‘Apollinaris to Sonsnaus, his brother, greetings. Receive one drachma and three obols from Leonidas and buy me slices (of fish) and give them to Akhillas, the donkey-driver, because I want to send them to Egypt’. Apollinaris uses the form κόμισε as an imperative (= probably κόμισαι, not κόμισον), but the form πέμψε cannot syntactically be anything other than an infinitive. In contrast to this is the usage of Petenephotes (O. Claud. 2.246): (14) Πετενεφώτης Οὐαλερίῳ τῷ ἀδελφῷ πολλὰ χαίριν. καλῶς πυήσις πέμψε μοι (δραχμὰς) δ ἵνα λάβο τὸ πιττάκιν καὶ ἠσέλθω τῇ κδ. πέμψεν αὐτὼ μετὰ Παχωμι. πέμψε̣ν καὶ πιττάκιν μοι ἄγναφον. μὴ ἀμ[ε-] λήσις {σοι}.

258

leiwo

‘Petenephotes to Valerius his brother, many greetings. Please send me 4 drachmas, so that I can get the pass and can arrive on the 24th. Send it with Pachomis. Send me a blank pass as well.39 Do not neglect this. Farewell’. Here the non-standard syntax allows both the imperative and the infinitive in line 4 and, as I have shown, the imperative is very common in this position with this politeness idiom (Leiwo 2010:102). Petenephotes diligently writes both the imperatives in lines 6 and 8 in a similar way in this letter, πέμψεν (= [´pempsǝn] = πέμψον), though in 247, 7 he spells the imperative as πέμπψε. All this clearly shows that generalizations to explain the potential confusion of the infinitive and imperative are not tenable.

11.6

Concluding Remarks

Special attention has been devoted to detecting idiolectal language use whenever possible. The data show clearly that ongoing language change is transparent in the private letters analysed here. I suggest that this conclusion holds true for a much larger corpus as well, and the same peculiarities appear again and again in similar linguistic contexts. I focused on phrases that include forms which have an opaque morphological appearance and require more contextual syntactic analysis than has usually been conducted up to now. Multiple reasons for the opaqueness of the forms, as well as of the syntax, obviously exist, because internal variation and contactinduced variation may have similar results, although they differ in origin. The confusion is nevertheless stronger when both forces are at work and simplification occurs more rapidly. The aorist imperative ending {on} gives way to the present imperative ending {e}, and the infinitive and imperative become identical both in pronunciation and in script. Scribes could produce good documentary-standard Greek, but the experienced writers working with small-scale businesses in the Mons Claudianus area were subject to language contact, as many of them were certainly of Egyptian origin with Greek as their second language. At the same time, Greek was obviously their only script, because alphabetic writing for Coptic had not yet emerged and Latin was used only sporadically.

39

This translation was suggested to me by Stefanie Kennell. Bülow-Jacobsen translated ‘an unfulled (new?) pittacium’.

confusion of moods in greek private letters of roman egypt?

259

The infinitive is represented in the letters, which have a register suitable for its use, for example when using indirect narrative, and which are written by a competent scribe. The infinitive is also used in idiomatic clauses, for instance, greetings and farewells, and with very common verbs such as (ἐ)θέλω taking an infinitive complement. My data show, however, that in clauses where the writer could easily have chosen between the infinitive and other syntactic constructions without having ready-made phrases containing the infinitive, the infinitive usually gives way to other constructions. The imperative is very common in private letters, and as the most basic Greek verb form, it was taken up by Sahidic Coptic to be the infinitive mood for borrowed Greek verbs. Identical forms for the imperative and the infinitive in Sahidic Coptic and the borrowing of Greek verbs in the imperative mood increased Egyptian l1 speakers’ confusion about the imperative and the infinitive when they were writing Greek. Combined with the strong stress accent, which affected unstressed vowels by reducing them toward schwa, the situation was tricky and caused spelling difficulties. Writers also had difficulty spelling the aorist imperative forms, especially of frequently-used verbs like πέμπω and κομίζω/κομίζομαι. The confusion is strongest in writers who were undoubtedly of Egyptian descent. This uncertainty in spelling, however, did not cause categorial disorder that could be spotted everywhere; rather, the infinitive and the imperative seem to preserve their basic grammatical difference, even if in some clauses distinguishing between the two forms is hardly possible. This opaqueness is very conspicuous in complements of the politeness formula καλῶς ποιήσεις.

References O. Claud. 1–4: Mons Claudianus. Ostraca Graeca et Latina. 1. 1–190. 1992. Ed. and tr. J. Bingen, A. BülowJacobsen, W.E.H. Cockle, H. Cuvigny, L. Rubinstein, and W. Van Rengen. Cairo: Institut Français d’Archéologie Orientale. Mons Claudianus. Ostraca Graeca et Latina. 2. 191–416. 1997. Ed. and tr. J. Bingen, A. Bülow-Jacobsen, W.E.H. Cockle, H. Cuvigny, F. Kaiser, and W. Van Rengen. Cairo: Institut Français d’Archéologie Orientale. Mons Claudianus. Ostraca Graeca et Latina. 3. Les reçus pour avances à la ‘familia’, (417–631). 2000. Ed and tr. H. Cuvigny. Cairo: Institut Français d’Archéologie Orientale. Mons Claudianus. Ostraca graeca et latina 4. The Quarry Texts. O. Claud. 632–896. 2009. Ed. and tr. A. Bülow-Jacobsen. Cairo: Institut Français d’Archéologie Orientale.

260

leiwo

Ast, R. & G. Azzarello. 2013. ‘New perspectives on the Gemellus Archive. Sabinus and his correspondence’. Das Fayyûm in Hellenismus und Kaiserzeit, ed. by C. Arlt & M.A. Stadler, 19–28. Wiesbaden. Cook, V. & B. Bassetti (eds.). 2005. Second language writing systems. Clevedon. Cuvigny, H. 2006. ‘L’organigramme du personnel d’une carrière impériale d’après un ostracon du Mons Claudianus’. Chiron 35:309–353. Dahlgren, S. Forthc. 2017. Outcome of long-term language contact: transfer of Egyptian phonological features onto Greek in Graeco-Roman Egypt. PhD diss. Helsinki. Dahlgren, S. Forthc. ‘Egyptian transfer elements in the Greek of Narmouthis ostraca’. Proceedings of the 27th International Congress of Papyrology, Warsaw 2013. jjp Suppl. Dahlgren, S. & M. Leiwo. Forthc. ‘Confusion of mood or phoneme? The impact of L1 phonology on verb semantics’. Post-classical Greek: intersections of linguistics and philology, ed. by I. Serzant & D. Rafiyenko. Berlin. Daniel, R.W. 1994. ‘Neither do they put new wine in old skins’. zpe 101:61–66. Girgis, W.A. 1966. ‘Greek loan words in Coptic’. bsac 18:71–96. Halla-aho, H. Forthc. ‘Scribes in private letter-writing: linguistic perspectives’. Beyond free variation, ed. by J. Cromwell & E. Grossman, Oxford. Harris, W.V. 1989. Ancient literacy. Cambridge, ma. Horrocks, G.C. 20102. Greek: a history of the language and its speakers. Chichester. Hohlwein, N. 1957. ‘Le vétéran Lucius Bellienus Gemellus, gentleman-farmer au Fayoum’. ÉdP 8:69–91. Layton, B. 2000. A Coptic grammar with chrestomathy and glossary. Wiesbaden. Layton, B. 2007. Coptic in 20 lessons. Leuven. Leiwo, M. 2003. ‘Both and all together? The meaning of ἀμφότεροι’. Arctos 37:81–99. Leiwo, M. 2005. ‘Substandard Greek: remarks from Mons Claudianus’. Ancient Greece at the turn of the millennium, ed. by N.M. Kennell & J.E. Tomlinson, 237–261. Athens. Leiwo, M. 2010. ‘Imperatives and other directives in the Greek letters from Mons Claudianus’. The language of the Papyri, ed. by T.V. Evans & D.D. Obbink, 97–119. Oxford. Leiwo, M. 2012. ‘Introduction: variation with multiple faces’. Variation and change in Greek and Latin, ed. by M. Leiwo, H. Hallo-aho & M. Vierros, 1–12. Helsinki. Mandilaras, B.G. 1973. The verb in the Greek non-literary papyri. Athens. Morani, M. 2000. Introduzione alla linguistica latina. Munich. Schulze, W. & O. Salomies. 1991. Zur Geschichte lateinischer Eigennamen. 1904 reprint with a list of corrections. Zurich. Till, W.C. 19612. Koptische Dialektgrammatik mit Lesestücken und Wörterbuch. Munich. Wackernagel, J. 2009. Lectures on syntax with special reference to Greek, Latin, and Germanic. Translated by D. Langslow. Oxford. Wohlgemuth, J. 2009. A typology of verbal borrowings. Berlin. Vierros, M. 2012. Bilingual notaries in Hellenistic Egypt: a study of Greek as a second language. Brussels.

chapter 12

Register Variation and Tense/Aspect/Mood Categories in Ancient Greek: Problems and Perspectives Andreas Willi

Δεῖ δὲ μὴ λεληθέναι ὅτι ἄλλη ἑκάστῳ γένει ἁρμόττει λέξις. οὐ γὰρ ἡ αὐτὴ γραφικὴ καὶ ἀγωνιστική, οὐδὲ δημηγορικὴ καὶ δικανική. ‘One must not forget that a distinct style is appropriate for each genre; for the style of writing is not identical with that of debating, and the style of assembly speeches is not the same as that of lawcourt speeches’. aristot. Rh. 1413b3–5

∵ 12.1

Introduction: Some Ancient Views

By the remark chosen as the motto of this paper, Aristotle reminds his students and readers—ourselves included—that an awareness of register variation1 formed an essential part of ancient theorizing about language in action. Considerable attention was paid not only to the selection of the mot juste, but also to questions of grammar and syntax. Already Protagoras, it is said, disapproved of Homer using an imperative in Il. 1.1 μῆνιν ἄειδε θεά, as he meant to formulate a prayer (Aristot. Poet. 1456b15–18 = Prot. test. 80a29); Dionysius of Halicarnassus censured Thucydides’ fondness for nominalized adjectives (Dion. Hal. Thuc. 31); and Aristotle advised that writers of prose avoid asyndeta (Aristot. Rh. 1 The term ‘register variation’ is here preferred to what is sometimes also called ‘variation in style/genre’. Since ‘register’ more specifically refers to ‘the form (or signifiant) plane of an utterance or text, which corresponds to genre as the content (or signifié) plane’, it has an advantage over ‘style’ as one may also ‘speak of the “style” of an author or even an epoch’, and ‘the peculiar “style” of a poet like Aristophanes arises precisely from the mixture of “registers” belonging to different communicative situations’ (Willi 2010b:298, with further literature).

© koninklijke brill nv, leiden, 2017 | doi: 10.1163/9789004315358_013

262

willi

1413b14–21), which are, however, perfectly appropriate at the end of oratorical epilogoi, in phrases like εἴρηκα, ἀκηκόατε, ἔχετε, κρίνατε (Aristot. Rh. 1420a6– 8). Matters of word order, too, were touched upon, as when Dionysius found that figures of speech such as antitheses were more suitable in ceremonial/epideictic speeches than elsewhere (Dion. Hal. Isoc. 2);2 and, leaving aside the theoreticians, an otherwise unknown Ariphrades made fun of the tragic poets’ license to use prepositions in anastrophe, as in δωμάτων ἄπο for ἀπὸ δωμάτων, which no-one would ever say in normal speech (Aristot. Poet. 1458b31–1459a2). The last-mentioned example points paradigmatically to the value of parody if we want to find out something about ancient perceptions of generically conditioned linguistic variation beyond ‘academic’ circles. When a classical comedian invents, say, a parodic law, treaty, or decree, and therein uses a ‘jussive’ infinitive with accusative subject as in (1) or (2) below, to be contrasted with the citation of an actual law in (3)3—i.e., a construction that is otherwise rare in comedy4—, this may simply confirm our intuition that the accusative-subject jussive infinitive was indeed (at least at the time of its being parodied) a stylistic feature peculiar to legal texts in officialese. However, without observing its peculiar distribution in comedy, how could we be sure that the construction was not in more widespread use outside the written and formally oratorical domains, and hence for the Greek ear potentially devoid of the ‘stately’ overtones we assign to it, or how would we know that oral proclamations of an official nature, as in (4), adopted it too? (1) Ψη. χρῆσθαι Νεφελοκοκκυγιᾶς τοῖσδε τοῖς μέτροισι καὶ σταθμοῖσι καὶ ψηφίσμασι καθάπερ Ὀλοφύξιοι. (Decree-seller:) ‘The Cloudcuckoolanders are to use these measures and weights and decrees, just like the Olophyxians’. aristoph. Av. 1040–1042

(2) ἀποδοῦναι δ’ ὅσα ἔχομεν ἀλλήλων, ἡμᾶς μὲν τὰς Θρᾴττας καὶ Ἀθερίνην τὴν αὐλητρίδα καὶ Σηπίαν τὴν Θύρσου καὶ τοὺς Τριγλίας καὶ Εὐκλείδην τὸν 2 This fits into the wider context of ancient ideas on ‘natural’ word order, as expounded by Pseudo-Demetrius’ Περὶ ἑρμηνείας ([Dem.] Eloc.) or Dionysius of Halicarnassus’ Περὶ συνθέσεως ὀνομάτων (Dion. Hal. Comp.); cf. now de Jonge (2007; 2008:251–328). 3 The restriction is significant, for ‘jussive’ infinitives with nominative/vocative subject show a different pattern of distribution and may have a different origin: see Bers (1984:169–181), after Wackernagel (1926:266–268). 4 At least if we distinguish from it the ‘optatival’ infinitive of prayers: cf. Wackernagel (1926: 267); Bers (1984:182–183); and Willi (2003:33–34).

register variation and tense/aspect/mood categories in greek

263

ἄρξαντα καὶ Ἀναγυρουντόθεν τοὺς Κορακίωνας καὶ Κωβιοῦ τοῦ Σαλαμινίου τόκον καὶ Βάτραχον τὸν πάρεδρον τὸν ἐξ Ὠρεοῦ ‘We are to give back what we have got from each other, we (the Athenians?) the Thracian women and Atherine the flute-player and Sepia the daughter of Thyrsus and the Triglians and Euclides the former archon and from Anagyrous the Coraciones and the offspring of Cobius of Salamis and Batrachus the assessor from Oreos’. archippus fr. 27, from a treaty between the fishes and the Athenians

(3) Πε. ἐρῶ δὲ δὴ καὶ τὸν Σόλωνός σοι νόμον· νόθῳ δὲ μὴ εἶναι ἀγχιστείαν παίδων ὄντων γνησίων. ἐὰν δὲ παῖδες μὴ ὦσι γνήσιοι, τοῖς ἐγγυτάτω γένους μετεῖναι τῶν χρημάτων. (Peisetaerus:) ‘But in addition I will quote to you the law of Solon: “The illegitimate child is not to have the right of next of kin as long as there are legitimate children; and if there are no legitimate children, those nearest of kin are to share in the inheritance”’. aristoph. Av. 1660–1666

(4) Πε. ἀκούετε λεῴ· τοὺς ὁπλίτας νυνμενὶ ἀνελομένους θὤπλ’ ἀπιέναι πάλιν οἴκαδε, σκοπεῖν δ’ ὅ τι ἂν προγράφωμεν ἐν τοῖς πινακίοις. (Peisetaerus:) ‘Hear, o people: the infantry is now to take their arms and go home again, and to watch out for what we announce on the noticeboards’. aristoph. Av. 448–450; cf. aristoph. Ach. 172, 1000–1001, Vesp. 937–939, Pax 551–555

12.2

Synchrony and Diachrony

A case like that of the accusative-subject jussive infinitives in laws, decrees, and proclamations illustrates, moreover, that we must not look at register variation from an exclusively synchronic perspective.5 If we assume—as is at least

5 Compare the pertinent remarks by van der Eijk (1997:101–102) on the development of Greek

264

willi

plausible—that the accusative-subject jussive infinitive originated through the ellipsis of a verbal phrase such as ἔδοξε ‘it was decided (that)’, the moment when the ellipsis first occurred (or better: when ‘elliptical’ accusative-subject jussive infinitives were no longer confined to follow-up sentences after a first sentence introduced by ἔδοξε) marks a terminus post quem for the construction to become stylistically marked; though of course we are not dealing here with the birth of a register feature ex nihilo (or from simple linguistic conservatism within a given register), but with the productive transformation of a phraseological peculiarity—the use of introductory ἔδοξε + AcI in laws and decrees—into a syntactic one. A combination of these two aspects, the genetic and the productive-transformational, can be seen at work when we consider, for example, one of the best-known register markers of Ancient Greek: tmesis. Like prepositional anastrophe in tragedy, tmesis fossilizes an ancient syntactic state of affairs, in this case one where a specific group of originally independent adverbial particles were not yet firmly attached (as preverbs) to verbs. Interestingly, this archaism is not only a feature of high poetic registers—triggering the semi-parodic ξύμ μοι λάβεσθε in Socrates’ address to the Muses in Plato’s Phaedrus (5)—but also, at least until classical times, of conversational language (6): (5) ἄγετε δή, ὦ Μοῦσαι, εἴτε δι’ ᾠδῆς εἶδος λίγειαι, εἴτε διὰ γένος μουσικὸν τὸ Λιγύων ταύτην ἔσχετ’ ἐπωνυμίαν, ‘ξύμ μοι λάβεσθε’ τοῦ μύθου, ὅν με ἀναγκάζει ὁ βέλτιστος οὑτοσὶ λέγειν ‘Well then, o Muses, whether you are called ‘clear-sounding’ (λίγειαι) because of your kind of song or because of the musical race of the Ligyans, aid me with the story this excellent man here makes me tell […]’. pl. Phdr. 237a

(6) Στ. οἴμοι. τί οὖν δῆθ’ ὁ κακοδαίμων πείσομαι; ἀπὸ γὰρ ὀλοῦμαι μὴ μαθὼν γλωττοστροφεῖν. (Strepisades:) ‘Help! What’s going to happen to me, poor guy? I’m lost if I don’t learn to twist my tongue’. aristoph. Nub. 791–792

scientific discourse in medical texts (e.g., with regard to word-formational features like the rise of nosological terms in -ῖτις or -ώδης).

register variation and tense/aspect/mood categories in greek

265

However, while colloquial tmesis is fairly restricted in scope (with the preverb and verb only ever being separated by particles or enclitic words),6 poetic tmesis develops an independent, ‘artistic’ productivity when verses like Il. 8.90 καί νύ κεν ἔνθ’ ὁ γέρων ἀπὸ θυμὸν ὄλεσσεν ‘there then the old man would have lost his life’ are generated. Here the tmesis is neither due to ἀπὸ θυμόν simply being a prepositional phrase of its own, before preverbation (like, e.g., παρὰ νηυσὶ κορωνίσι in Hom. Il. 18.338 τόφρα δέ μοι παρὰ νηυσὶ κορωνίσι κείσεαι αὔτως ‘so long you will lie by the curved ships as before’), nor is it triggered by a leftdislocation of the adverbial particle to a phrase-initial specifier position, which might have been permissible in early Greek (thus accounting for the colloquial type as well as tmeseis like Hom. Od. 19.592 ἐπὶ γάρ τοι ἑκάστῳ μοῖραν ἔθηκαν ‘for they have given his share to everyone’).7

12.3

Register and Dialect

This caveat, that distinctive register features are born, live a life, and may die—exactly like any other linguistic feature—and that we must not therefore mix up distinct diachronic layers, has a spatial counterpart. If linguistic conservatism can be a determining factor in the genesis of a register, this is just another way of saying that generic conventions are often paramount. But conventions may change not only over time, they may also differ from place to place.8 If it has been observed, for example, that ‘the use of the 6 Cf. Wackernagel (1928:172–174); Morpurgo Davies (1985:86–88); and Willi (2003:250). 7 Cf. Hajnal (2004:esp. 156–161), who then argues, against Horrocks (1981), that Homeric Greek is not more archaic than Mycenaean Greek in this respect: the absence of tmesis in the Mycenaean texts is rather a register feature of a more technical genre where fixed ‘phrasal’ and ‘compound’ verbs are often needed (as the preverb modifies the verbal meaning itself). On the poetic productivity of tmesis see also Haug (2010:92–97), with statistics even showing an increase from the Iliad to the Odyssey. Note, however, that the use of Il. 13.306 τὸν καὶ Μηριόνης πρότερος πρὸς μῦθον ἔειπε by Haug and of Il. 23.68 στῆ δ’ ἄρ’ ὑπεὶρ κεφαλῆς καί μιν πρὸς μῦθον ἔειπεν by Hajnal as examples of artificial tmesis is more problematic than it seems at first sight: while in metrical terms *τὸν καὶ Μηριόνης πρότερος μῦθον προσέειπε ‘would be an equally good hexameter’ (Haug 2010:95), the ‘Muster’ cited by Hajnal (2004:160), Il. 22.329 ὄφρά τί μιν προτιείποι ἀμειβόμενος ἐπέεσσιν, suggests that only *τὸν καὶ Μηριόνης πρότερος μύθῳ προσέειπε would have been grammatical. In other words, from a syntactic point of view we are not dealing with real tmesis of προσέειπε in either verse, but rather with a— in Haug’s example artificially discontinuous—prepositional phrase with ‘anastrophe’ (this terminology of course being anachronistic). 8 Instructive, in this respect, is the transition from Ionic to Attic as the ‘main’ dialect of Greek

266

willi

modal [i.e. phrases like we would appreciate vs. we will appreciate, a.w.] may be one of the forms which distinguishes British from American business letter writing’ (Jenkins-Hinds 1987:332), there is no reason to believe that similar diatopic differences did not also exist in ancient Greece. Given the limited geographical spread of our literary data and the consequent danger of mistaking an individual author’s stylistic preferences for a generic convention in the dialect used by him/her, if that dialect is otherwise underrepresented in our corpus, pertinent observations can be made most easily, if at all, in epigraphic genres. For instance, the use of a potential optative with the modal particle κα (~ Att. ἄν) as an (early) alternative to the accusative-subject jussive infinitive of laws, decrees, and treaties appears to have been restricted to Elean:9 (7) […] αἰ δὲ βενέοι ἐν τἰαροῖ, βοί κα θο̄άδοι καὶ κοθάρσι τελείαι, καὶ τὸν θεαρὸν ἐν τα⟨ὐ⟩τᾶι. αἰ δέ τις παρ τὸ γράφος δικάδοι, ἀτελε̄ς́ κ’ εἴε̄ ἀ δίκα, ἀ δέ κα ϝράτρα ἀ δαμοσία τελεία εἴ-ε̄ δικάδο̄σα […] ‘[…] but if he were to have intercourse in the sanctuary, he would pay as a penalty an ox and a complete purification, and the thearos in the same way. But if someone were to pass judgment against the written rule, the judgment would be void, and the decree of the people passing judgment would be valid […]’. IvO 7 = dge 412 = minon no. 4

(8) […] τοῖ ζέ κα θεοκόλοι θ[άρο]ς ἔα ὐτοῖ καὶ χρε̄μ⟨ά⟩τοις, ὄ τι [αὐτ]ο̃ γα εἴε̄ ποτ’ ἀλάθειαν· αἰ δ’ αἰλότρια ποιοῖτο, ⟨πε⟩ντακατίας κα δαρχμ⟨ὰ⟩ς ἀποτίνοι κατὰ ϝέκαστον θεθτμόν, ὄ τι ἀδίκο̄ς ἔχοι καὶ πο⟨ι⟩οῖτο ἀδίκο̄ς γα […] ‘[…] but the theokolos could be confident both for himself and for the money, as far as it would be really his; but if he were to lay claim to

literary prose in the late 5th c. bce (or later, depending on genre: see van der Eijk (1997:100– 101) on medical prose). On this phenomenon, and its cultural conditioning, see now Willi (2010a). 9 Slotty (1915:97, 124) also refers to Arcadian, but the context is different in the examples cited at Slotty (1915:127–128). See further Minon (2001:311–321; 2007:2.453–465), also on the origins and chronology of the Elean usage and its subsequent replacement by the more usual infinitive or imperative. Minon (2008:189) highlights yet another difference between the conventions in Elean laws and elsewhere, viz. the use of present-stem forms in hypothetical subordinate clauses (vs. aorist-stem ones in Gortyn: cf. Bile-Lallot 2008:14).

register variation and tense/aspect/mood categories in greek

267

that of others, he would pay five hundred drachmae according to every regulation, as it would be unjust and he would lay claim unjustly […]’. IvO 4 = dge 411 = minon no. 9

Similarly, the regular occurrence of a present-tense verb ἀντίθειτι or ἀφίειτι— rather than a corresponding aorist—in manumission inscriptions from Boeotia seems to be a regional peculiarity: (9) θιός· τούχα ἀγαθά. Ϝαστίαο ἄρχοντος Βοιωτῦς, ἐν δὲ Λεβαδείη Δόρκωνος Δωΐλος Ἰρανήω ἀντίθειτι τὸν ϝίδιον θεράποντα Ἀνδρικὸν τῦ Δὶ Βασιλεῖι κὴ τῦ Τρεφωνίυ ἱαρὸν εἶμεν […] ‘God, good fortune. Under the archonship of Wastias among the Boeotians, and of Dorcon in Lebadeia, Doilos son of Iranaeus dedicates his own servant Andricus to Zeus the King and to Trophonius, to be sacred […]’. ig vii 3083 = dge 509, Lebadeia, 3rd c.10 (10) [Ε]ὐρυμείλω ἄρχοντος ἀφίειτι Σάων Ἄτταν ἐλεύθερον ἐναντία τῶ Ἀσκλαπιῶ κὴ τῶ Ἀπόλλωνος· ϝίστορες Ἀντι(μέ)νων Ἄσιος Ἀθανόδωρος Εὔφραστος. ‘Under the archonship of Eurymelus, Saon releases Atta to be free before Asclepius and Apollo; witnesses are Antimenon, Asius, Athanodorus, Euphrastus’. ig vii 1779 = dge 492, Thespiae, 2nd c.11 To be sure, for these somewhat unusual Boeotian data we are free to hypothesize that ‘[i]n the case of immediate and unconditional emancipation, the use of the present tense might imply a continuous sanction on the part of the master of the freedom of a former slave’,12 but even such an account does not alter the basic fact that similar manumission documents from other areas, such as Delphi (11), commonly show aorists like ἀπέδοτο, or also ἀνέθηκε, and that the aorist equally appears in a few pertinent texts from Boeotia itself if they are written in Koine Greek (12):13

10 11 12 13

Cf. e.g., dge 511, 512, 515–517, 528. Cf. dge 491. Claflin (1905:68), who signals the divergent usage in other regions. But even in Koine documents ἀνατίθησι prevails: cf. e.g., ig vii 3357, 3359, 3362, 3364, etc.

268

willi

(11) ἄρχοντος [Ν]ικοβούλου μηνὸς Βουκατίου, ἐπὶ τοῖσδε ἀπέδοτο Νεοπάτρα Ὀρθαίου Δελφὶς τῶι Ἀπόλλωνι τῶι Πυθίωι σώματα γυναικεῖα δύο αἷς ὀνόματα Ζωπύρα, Σωσίχα, τιμᾶς ἀργυρίου μνᾶν ἕξ, καθὼς ἐπίστευσαν Ζωπύρα, Σωσίχα τῶι θεῶι τὰν ὠνάν, ἐφ’ ὧιτε ἐλευθέρας εἶμεν […] ‘Under the archonship of Nicobulus, in the month Boukatios, Neopatra wife of Orthaeus, from Delphi, released to Apollo Pythios two female persons, whose names are Zopyra and Sosicha, at the price of six mnai, as Zopyra and Sosicha had entrusted to the god the purchase-money, in order to be free […]’. dge 33514 (12) ἄρχοντος ἐγ Χαιρωνείᾳ Ἀρίστωνος, μηνὸς Ἀγριωνίου πεντεκαιδεκάτῃ, Θέων Θεομνήστου Φανατεὺς ἀνέθηκεν ἐγ Χαιρωνείᾳ ἱερὸν τῷ Σαράπι τὸν ἴδιον θρεπτὸν ὃν εἶχε οἰκογενῆ, ᾧ ὄνομα Σώσων, τὴν ἀνάθεσιν ποιούμενος διὰ τοῦ συνεδρίου κατὰ τὸν νόμον Χαιρωνέων […] ‘Under the archonship of Ariston in Chaeroneia, on the 15th of the month Agrionios, Theon son of Theomnestus Phanateus dedicated in Chaeroneia his own slave born and raised at home, whose name is Soson, to be sacred to Sarapis, carrying out the dedication through the council in accordance with the law of the people of Chaeroneia […]’. ig vii 3376; cf. ig vii 3378, both on stones bearing dialectal manumission acts with ἀντίθειτι on the side

Thus, we do indeed observe in Boeotia a noteworthy and—as far as we can tell—register-conditioned15 tense-aspect phenomenon, but this does not mean that we can deduce from it any rule, however generously formulated, about tense-aspect usage in the entire corpus of Greek manumission decrees.

12.4

Register-Internal Variation

Despite these sobering premises, the study of register-conditioned variation in Ancient Greek remains possible, even when fine-grained grammatical cat-

14 15

Cf. dge 336–347. Note that ‘normal’ dedications in Boeotian use ἀνέθε̄κε, ἀνέθεικε, ἀνέθεαν, ἀνέθιαν etc., e.g., dge 477, 481, 482, 507, 508 etc.

register variation and tense/aspect/mood categories in greek

269

egories such as tense and aspect are concerned.16 One reason why such investigations are not fatally affected by the theoretical need for an unrealistically extensive corpus of chronologically and geographically homogeneous sources lies in the fact that ‘typical’ register features are unlike ‘typical’ dialect features or even ‘typical’ features of diachronic variation, in one crucial respect: they are hardly ever exceptionless, but almost by definition surface as tendencies, not rules. This is true even in those comparatively rare cases where they are strictly phonological or morphological in nature. If we were to discover, say, an inscription from Thessaly in which 70 percent of all o-stem genitive singular forms ended in -οι, but 30 percent in -ου, as in Attic-Ionic, the dialectal ‘purity’ or ‘authenticity’ of such a text would be called into question immediately. But if we discover that the language of Greek epic regularly uses a genitive in -οιο next to a genitive in -ου, nothing stops us from treating this latter type of variation as a perfectly acceptable feature of a ‘pure’ and ‘authentic’ epic register (the only concession being that we then start calling it a Kunstsprache). In other words, as we try to formalise the grammar of a given register we expect internal variation to occur and we are not inconvenienced when things do not turn out to be black and white only. By way of illustration, let us briefly consider two further examples relating to tense-aspect phenomena. It would be absurd to claim that each and every oracular response in Ancient Greek had to be formulated with a presenttense verb instead of a future one, and that whenever we do find a future tense, this indicates that the oracle is spurious or improperly framed. But it is a valid and interesting observation that a number of oracular or oraclelike statements, whether real, literary, or parodic, show an otherwise fairly uncommon17 futuric use of the present tense, by which the anticipated event is, as it were, removed from the temporal axis and presented as a matter whose validity is not constrained by notions as ephemeral as the actual speaker’s human reference time:18 16

17

18

And not just when the conditions are as ideal as with Xen. Ages. 1.6–3.1, an ‘encomium’ based directly upon the ‘historiographical narrative’ of Xenophon’s Hellenica: on the divergent tense-aspect usages in these two sources see the exemplary study by Buijs (2007). The same praesens pro futuro can however be used for rhetorical effect (e.g., Thuc. 6.91.3): cf. Kühner-Gerth (1904:1.138–139) and Schwyzer-Debrunner (1950:273), with further examples. Cf. Wackernagel (1926:161): ‘Es wird etwas für die Zukunft ausgesagt, aber der Seher oder die Seherin sieht das Ereignis vor Augen’; see also Hdt. 3.57.4, 8.77.1–2; Aristoph. Equ. 1020, Av. 985; Paus. 4.20.1, 9.14.3, 9.17.5, and compare Pind. Ol. 8.42; Aesch. Ag. 126, pv. 171, 767, 948; Aristoph. Equ. 127, 176–177, 1087; Eupolis fr. 195. The responses collected in ParkeWormell (1956) contain no further examples.

270

willi

(13) πέμψαντες γὰρ οἱ Ἀθηναῖοι ἐς Δελφοὺς θεοπρόπους χρηστηριάζεσθαι ἦσαν ἕτοιμοι· καί σφι ποιήσασι περὶ τὸ ἱρὸν τὰ νομιζόμενα, ὡς ἐς τὸ μέγαρον ἐσελθόντες ἵζοντο, χρᾷ ἡ Πυθίη, τῇ οὔνομα ἦν Ἀριστονίκη, τάδε· ‘ὦ μέλεοι, τί κάθησθε; λιπὼν φεῦγ’ ἔσχατα γαίης δώματα καὶ πόλιος τροχοειδέος ἄκρα κάρηνα. οὔτε γὰρ ἡ κεφαλὴ μένει ἔμπεδον οὔτε τὸ σῶμα, οὔτε πόδες νέατοι οὔτ’ ὦν χέρες, οὔτε τι μέσσης λείπεται, ἀλλ’ ἄζηλα πέλει· κατὰ γάρ μιν ἐρείπει πῦρ τε καὶ ὀξὺς Ἄρης, Συριηγενὲς ἅρμα διώκων. πολλὰ δὲ κἆλλ’ ἀπολεῖ πυργώματα […]’ ‘The Athenians, who had sent envoys to Delphi in order to consult the oracle, were ready to do so; and after they had performed the customary rites at the sanctuary, and as they were going into the hall and sitting down, the Pythia, whose name was Aristonice, gave the following response: “O you wretched ones, why are you sitting around? Leave, and flee from the hindmost houses of the land and from the high peaks of the wheel-shaped city. For the head does not remain fast, nor the body, nor the feet beneath, nor indeed the hands, nor is anything left of the middle, but things are dreary; for fire and fierce Ares throws it down, chasing the Syrian-born chariot. And many other towers will be destroyed […]”’. hdt. 7.140

(14) Ἀλ. πῶς δῆτά φησ’ ὁ χρησμός; εὖ νὴ τοὺς θεούς, Δη. καὶ ποικίλως πως καὶ σοφῶς ᾐνιγμένος· ‘ἀλλ’ ὁπόταν μάρψῃ βυρσαίετος ἀγκυλοχήλης γαμφηλῇσι δράκοντα κοάλεμον αἱματοπώτην, δὴ τότε Παφλαγόνων μὲν ἀπόλλυται ἡ σκοροδάλμη, κοιλιοπώλῃσιν δὲ θεὸς μέγα κῦδος ὀπάζει, αἴ κεν μὴ πωλεῖν ἀλλᾶντας μᾶλλον ἕλωνται.’ ‘(Sausage-Seller:) So how does the oracle speak, then? (Demosthenes:) Well, by the gods, it is formulated neatly and wisely in riddles: “But when a leather-eagle with crooked claws snatches with his jaws the blood-drinking idiot dragon, then indeed perishes the garlicsauce of the Paphlagonians, and the god gives great fame to the tripesellers, unless they rather choose to sell sausages” ’. aristoph. Equ. 196–201

register variation and tense/aspect/mood categories in greek

271

Or, to turn to the second example while staying within the sphere of religion, it has long been observed that the incidence of aorist imperatives vis-à-vis present ones is significantly higher in Greek prayers than in ordinary conversations between human beings.19 Again this does not mean that addressing a god with the present imperative would be in any way unusual, but the tendency as such still calls for an explanation. According to Duhoux, it is because ‘l’ homme est l’inférieur social de la divinité. Lorsqu’il s’adresse à ses supérieurs, il tend à exercer une pression basse et emploie dès lors le ponctuel, où la volonté s’ exprime de façon minimaliste, grâce à la non considération des diverses phases du développement de l’action’.20 I am not21 sure this is correct. To the extent that there is any such sociolinguistic dimension involved at all, I would rather suspect that the aorist imperative, being more ‘focal’,22 is also potentially more peremptory, and hence more suitable to convey a certain urgency of the request. The social factor will then operate exactly the other way round: because it can be taken for granted that a human being is ‘socially inferior’ to the god, i.e., not in a position to impose his or her will through any kind of sanction, the peremptory or urgent tone of the aorist imperative pragmatically underlines the immutable dependency relationship.23 Incidentally, this will also explain the even more substantial predominance of aorist stems among 19

20

21

22

23

Cf. Kieckers (1909); Bakker (1966); Moorhouse (1982:218–219) (on Sophocles); Fanning (1990:380–382) (on the New Testament); Duhoux (2000:245–246); and Willi (2003:31–32) (on Aristophanes). Duhoux (2000:245); cf. Ruijgh (1985:36): ‘Le locuteur n’est pas en position de demander que la personne socialement supérieure ou le dieu commence immédiatement l’action en question’. No longer, that is: in Willi (2003:32) I had accepted Duhoux’s explanation and tried to account for the distribution of cletic imperatives vs. χαῖρε (cf. below) in this way, but failed to recognise that, while it may be appropriate to formulate crucial requests in a non-pressurising manner to one’s social equals in order to give them the possibility of refusing without seeming impolite, such questions of ‘secondary’ (im)politeness do not arise where the relative social positioning is not ambiguous or negotiable anyway. Thus Sicking (1991:esp. 154–170); cf. also Pulleyn (1997:221–226). This is not to say that other factors may not also play a role: Ruijgh (1985:30–38), for example, shows that the present imperative can often have an ‘inceptive’ value (‘begin to x, be x-ing!’). Like Sicking, however, I do not think that this proves temporal over aspectual primacy in the choice of aspect stem: much rather, the present imperative might seem to presume, somewhat impolitely, that the god will act, although this is far from certain. Cf. Scott (1907) and, with many interesting remarks, Post (1938: esp. 37–38), according to whom the aorist imperative is applied when there is ‘[s]ubmission, pretended or real, and supplication, including all attempts to soften resolution or to remove hostility’, whereas

272

willi

cletic imperatives (because the presence of the divinity is the sine qua non for any further assistance to become possible) and, vice versa, the typical occurrence of a present stem in the imperative χαῖρε in prayers: to be sure, the divinity’s joy about a sacrifice or the like increases the chances of other requests being granted, but it is not in itself essential to the petitioner.

12.5

The Functional Dimension

What all the material adduced so far has demonstrated is the interdependence of linguistic variation and register. On the one hand, where the system (langue) offers a choice as to how a given idea may be expressed, the consideration of what is appropriate in the target register may determine which of the available options is selected: a seer formulating an oracle is more likely to adopt the praesens pro futuro than a weather forecaster. On the other hand, by choosing a particular variant a speaker may consciously or unconsciously insert a given utterance into a particular register: by using a genitive in -οιο or the noncolloquial type of tmesis, an Athenian of the classical age would have made his/her utterance sound ‘epic’ or ‘poetic’ no matter what the actual content was like. Up to this point, however, our focus has been primarily phenomenological. We have simply asked (a) what the distribution of a given grammatical feature (such as tmesis or the accusative-subject jussive infinitive) was across the range of linguistic registers of Ancient Greek, or (b) what noteworthy grammatical features a given register contained (e.g., the register of manumission inscriptions). Meanwhile, a somewhat different approach is suggested by the last two examples concerning the imperative in prayers and the oracular present. Moving beyond the mere description, we should also ask from a functional point of view what a feature f achieves, so that it and a specific register r come to interact in the way they do. Such questions are not always easy to answer, especially when the force of convention has become strong. In 5th-century tragedy, for instance, prepositional anastrophe may no longer have had any recognisable function other than contributing its share to the de-automatising differentiation of the tragic Kunstsprache from ‘ordinary’ Greek. In other cases, however, a more tangible form-function relationship can be uncovered, and this may then contribute to a deeper understanding both of the grammatical feature under

the present imperative is used when either ‘[f]acts are in control’ or ‘[t]he speaker is assumed to be in control’.

register variation and tense/aspect/mood categories in greek

273

investigation and of its selection (or deselection) as operated by a given register. In the remainder of this paper, this ‘functional approach’ to register variation will be illustrated with a case study based on a particularly well-known and much-discussed tense-aspect phenomenon.24

12.6

Mapping the Greek Perfect: A Hypothesis

It is commonly acknowledged that the Greek perfect descends from an IndoEuropean perfect formation that used to be stative-intransitive.25 When the core meaning of the verbal root involved a patient or experiencer as the principal participant of the eventuality, such a stative-intransitive perfect would typically stand next to a medio-passive present or aorist (type γέγονα ~ γίγνομαι/ἐγενόμην). But even when the core meaning of the root presupposed an agentive participant, the perfect could still be stative-intransitive, and thus contrast with an active present or aorist (type τέτοκα ‘be mother’ vs. τίκτω/ἔτεκον ‘give birth’). However, the more the perfect was perceived as an integral part of the verbal paradigm, the more such a situation was unsatisfactory. Hence, already at some point in the prehistory of Greek, formally distinct mediopassive perfects began to be created, to match the medio-passive presents or aorists (type γεγένημαι ~ γίγνομαι/ἐγενόμην).26 Vice versa, following its present or aorist the type τέτοκα could acquire a direct object—perhaps starting with semi-adverbial neuter plural direct objects (e.g., ἐσθλὰ ἔοργε: ‘is a well-doer’). This stage has already been reached in Homeric Greek, and it may be much older than that. However, the proliferation of perfects in syntactically transitive 24

25

26

For some interesting examples involving clause-structure syntax (nominativus pendens, anacolutha), see again van der Eijk’s (1997:102–111) study of the register of Greek medical texts. The need to pay attention to (pragmatic) function as well as form is also brought out well by Slings’ (1997; 2002) demonstration that ‘rhetorical’ figures such as chiasmus and anaphora may, but do not have to, be features of ‘literary’—as opposed to ‘spoken’— registers. Contrary to wide-spread opinion (see e.g., the quotations listed in Sicking-Stork 1996:125), the Indo-European perfect was not intrinsically ‘nactostatic’ (expressing a state resulting from a (dynamic) past eventuality); where it was, this was epiphenomenal (as shown, for example, by the existence of ‘intensive’ perfects like μέμυκε ‘bellows’). For a preliminary exploration of this issue, see Willi (2007:42–43). The starting-point for this development may have been among the participles and/or pluperfects: see Chantraine (1927:47–70), and cf. the predominance of passive participles among the perfect forms of Mycenaean (Duhoux-Dachy 1992:222–224) and Homeric Greek (Schlachter 1907–1908:221–222).

274

willi

constructions of this type has not yet begun in Homer; nor has the old perfect formation, which could now be regarded as formally active (i.e., opposed to the new medio-passive perfect), already been integrated into the active paradigm of the verb fully enough to trigger the large-scale creation of new active perfects, on the basis of either the medio-passive perfect stem (act. τέτραφα after med.pass. τέτραπται, from τρέφω) or the aorist stem (act. πέπεικα after aor. ἔπεισα). Two of the most eminent scholars of Ancient Greek, Jacob Wackernagel and Pierre Chantraine,27 argued that this largely28 post-Homeric rise of the transitive active perfect can be described in semantic terms as the transition from a perfect in which the state of the verbal subject is central to a so-called ‘resultative’ perfect focussing on the state of the verbal object. In more recent literature, this view has been criticized, with opponents stressing that even in classical times it is still the state of the subject that is at stake when a perfect is used.29 To be sure, this may at times be a matter of textual interpretation more than anything else, but even so it seems to be true overall; and yet, it is equally true—and measurable in purely statistical terms—that something has changed between Homeric and Classical Greek. There are instances in which Homer uses what may be called a ‘resultative’ aorist where a perfect would seem appropriate later on:30 (15) ἔχθιστος δέ μοί ἐσσι διοτρεφέων βασιλήων· αἰεὶ γάρ τοι ἔρις τε φίλη πόλεμοί τε μάχαι τε. εἰ μάλα καρτερός ἐσσι, θεός που σοὶ τό γ’ ἔδωκεν.

27 28

29 30

Wackernagel (1904); Chantraine (1927:119–145). Though not entirely: sometimes it is difficult to draw a clear semantic line between Homeric and later transitive perfects (pace Wackernagel 1904:5; cf. Chantraine 1927:11–13, e.g., on Hom. Il. 17.542 λέων κατὰ ταῦρον ἐδηδώς). Transitive ὄπωπα, for example, is fairly common in Homer where ‘it describes (or characterizes) a person as one who has (or happens to have) seen (or not seen) something’, just as it does later on (Sicking-Stork 1996:155). Cf. McKay (1965); Rijksbaron (1984); Sicking-Stork (1996:esp. 146–150); and Duhoux (2000: 428). In (16) ἔρεξεν contrasts with the preceding ‘perfect of persistent situation’ ἐσθλὰ ἔοργε, as it describes Odysseus as someone who has just done an extraordinary deed (by silencing Thersites). Note also the apparent ‘avoidance’ of a perfect (e.g., ἀπέκτονε τὸς ἄνδρας), and the (repeated) use of a nominal phrase instead (e.g., φονές̃ ἐσστι τõν ἀνδρõν) in a 5thcentury inscription from Mantineia in Arcadia (ig v 2 262), where the context would be ideal for a ‘responsibility’ perfect to be used (cf. below, also on the semantic similarity between agent nouns and perfects).

register variation and tense/aspect/mood categories in greek

275

‘To me you are the most hated of all the god-reared kings: for you always like strife and wars and fights. If you are very strong, it is some god who has given this to you’. hom. Il. 1.176–178; note that the focus is on ‘some god’, as opposed to Agamemnon himself, as the provider of strength.

(16) ὦ πόποι, ἦ δὴ μυρί’ Ὀδυσσεὺς ἐσθλὰ ἔοργεν βουλάς τ’ ἐξάρχων ἀγαθὰς πόλεμόν τε κορύσσων, νῦν δὲ τόδε μέγ’ ἄριστον ἐν Ἀργείοισιν ἔρεξεν, ὃς τὸν λωβητῆρα ἐπεσβόλον ἔσχ’ ἀγοράων. ‘Wow, Odysseus is indeed a doer of good things, initiating excellent plans and organising war; but now he has accomplished this as his masterpiece among the Achaeans, banishing this aggressive slanderer from the assembly’. hom. Il. 2.272–275

In the first place this is undoubtedly an issue of diachronic variation.31 However, there is also a register dimension to it. If we ask, from a synchronic point of view, and for Classical Greek only, what the added value of a perfect such as γέγραφε is in comparison with an aorist ἔγραψε, we can say that, although the actual ‘writing’ act referred to may be exactly the same, the perfect ‘individuates’ less than the aorist. On a scale ranging from highly individuating, or eventuality-focused, Tolstoy wrote Anna Karenina (in the 1870s) via intermediate Tolstoy has written Anna Karenina (in his exquisite style) to least individuating Tolstoy is the writer of Anna Karenina, the classical aorist would be located towards the left (or verbal) end, and the perfect towards the right (or nominal) one.32 Hence the remaining stative nuance of the perfect makes it particularly

31

32

Since this is a functional replacement (aorist → perfect), I do not think that Rijksbaron (1984:412–413) can be right when he suggests that the rise of the ‘resultative’ or transitive active perfect has something to do with the rise of the concept of the individual. Significantly, the parents are still oἱ τεκόντες in Classical Greek (not: *τετοκότες), and by a similar process of fossilization dedicatory inscriptions and artists’ signatures retain the older resultative aorist (ἀνέθηκε, ἐποίησε, ἔγραψε; see Wackernagel 1904:6–9). Cf. Sicking-Stork (1996:148): ‘In many cases … the appropriate rendering of a perfect is a noun rather than a verb form’. For the contrast between ‘individuation’ and ‘typicalization’ (‘nominalization’) see Willi (2003:121), quoting (in translation) Lehmann (1982:68): ‘A propositional act involves, apart from reference and predication, a series of further operations. These operations find their expression in categories which help to constitute the

276

willi

suitable for the expression of lasting achievements: the perfect, unlike the classical aorist, cannot be cancelled, so to speak.33 And that in turn predisposes the perfect for rhetorical exploitation in utterances and genres where such achievements, be they positive or negative, have to be highlighted to make the text effective. The prime example of such a genre is oratory, especially lawcourt oratory.34 If, say, Euphiletus killed Eratosthenes, that is a fact, which even Euphiletus may not be able to deny. But one can still present this fact in different ways: Εὐφίλητος ἀπέκτεινε Ἐρατοσθένην ‘Euphiletus killed Eratosthenes’ merely records what happened; whereas Εὐφίλητος ἀπέκτονε Ἐρατοσθένην ‘Euphiletus is the killer of Eratosthenes’ stresses Euphiletus’ lasting responsibility for the act, and this may be crucial for a punishment to be meted out.

12.7

A Qualitative Approach

How, then, can we test this hypothesis, that the use of the perfect—and in particular the transitive active perfect—was determined, at least in part and at least in the Classical era, by considerations or conventions of register? Both a qualitative and a quantitative approach are conceivable. The former would have to highlight in the extant corpus of Greek oratory passages in which the perfect appears to be preferred over the aorist in some significant way: for example because there is a dense cluster of perfects, which are not paralleled easily in other texts from the same period, or because a perfect contrasts with an aorist that refers to the same eventuality or an eventuality of a similar kind.

33

34

sentence: mood, aspect, tense, diathesis, person, case complements. If they are neglected, the thought is generalized or “typicalized” and loses its individual character. Typicalization goes hand in hand with nominalization. The categories listed above as constituents of a proposition gradually disappear in the process of nominalization’. In Sicking-Stork’s (1996:145) terms, the perfect is characterized by a lack of the feature ‘control’: ‘At some moment in the past, the Agent was at liberty to decide whether he would perform a certain action or not, but now that he has done so he cannot go back on what he has done’. Cf. Sicking-Stork (1996:148), who also cite two loci classici from tragedy for such a ‘lasting responsibility’ reading, Aesch. Eum. 587–588 (Χο. τὴν μητέρ’ εἰπὲ πρῶτον εἰ κατέκτονας.— Ὀρ. ἔκτεινα· τούτου γ’ οὔτις ἄρνησις πέλει. ‘Chorus: First tell us if you are the killer of your mother.—Orestes: I killed her; there is no denying that’. Orestes ‘only’ refuses to be considered a ‘killer’.) and Soph. Ant. 442 (Κρ. φής, ἢ καταρνῇ μὴ δεδρακέναι τάδε;—Ἀν. καὶ φημὶ δρᾶσαι κοὐκ ἀπαρνοῦμαι τὸ μή. ‘Kreon: Do you admit, or do you deny being the culprit?—Antigone: I admit having done it, and I do not deny it’. Antigone refuses to be considered guilty, despite Cleon’s decree.).

register variation and tense/aspect/mood categories in greek

277

A good (and standard) example of the former occurs in Lysias 14, Κατὰ Ἀλκιβιάδου λιποταξίου: (17) σκέψασθαι δὲ χρή, ὦ ἄνδρες δικασταί, διὰ τί ἄν τις τοιούτων ἀνδρῶν φείσαιτο; πότερον ὡς πρὸς μὲν τὴν πόλιν δεδυστυχήκασιν, ἄλλως δὲ κόσμιοί εἰσι καὶ σωφρόνως βεβιώκασιν; οὐχ οἱ μὲν πολλοὶ αὐτῶν ἡταιρήκασιν, οἱ δ’ ἀδελφαῖς συγγεγόνασι, τοῖς δ’ ἐκ θυγατέρων παῖδες γεγόνασιν, οἱ δὲ μυστήρια πεποιήκασι καὶ τοὺς Ἑρμᾶς περικεκόφασι καὶ περὶ πάντας τοὺς θεοὺς ἠσεβήκασι καὶ εἰς ἅπασαν τὴν πόλιν ἡμαρτήκασιν, ἀδίκως καὶ παρανόμως καὶ πρὸς τοὺς ἄλλους πολιτευόμενοι καὶ πρὸς σφᾶς αὐτοὺς διακείμενοι, οὐδεμιᾶς τόλμης ἀπεχόμενοι, οὐδενὸς ἔργου δεινοῦ ἄπειροι γεγενημένοι; ἀλλὰ πεπόνθασί τε ἅπαντα καὶ πεποιήκασιν. οὕτω γὰρ διάκεινται, ὥστ’ ἐπὶ μὲν τοῖς καλοῖς αἰσχύνεσθαι, ἐπὶ δὲ τοῖς κακοῖς φιλοτιμεῖσθαι. ‘Judges, you also have to consider this: Why should anyone spare such men? Because they have been unfortunate in their dealings with the city, but are otherwise well-behaved people who have led an orderly life? Have not most of them kept company with whores, others slept with their sisters, some have got children from their daughters, some have played the mysteries and mutilated the herms, committed impiety against all the gods and crimes against the whole city, behaving unjustly and unlawfully towards other people in their capacity as citizens, and even in their relationships with each other, not refraining from any kind of recklessness, having become experts in every horrible action? No, they have both experienced and done everything; for that is how they are: they feel ashamed of right behaviour and take pride in wrong behaviour’. lys. 14.41–42

For the latter, meanwhile, one may point to a passage from Lysias 10, Κατὰ Θεομνήστου, a text notable for the degree of metalinguistic reflexion displayed: (18) [6] ἴσως τοίνυν, ὦ ἄνδρες δικασταί, περὶ τούτων μὲν οὐδὲν ἀπολογήσεται, ἐρεῖ δὲ πρὸς ὑμᾶς ἅπερ ἐτόλμα λέγειν καὶ πρὸς τὸν διαιτητήν, ὡς οὐκ ἔστι τῶν ἀπορρήτων, ἐάν τις εἴπῃ τὸν πατέρα ἀπεκτονέναι· τὸν γὰρ νόμον οὐ ταῦτ’ ἀπαγορεύειν, ἀλλ’ ἀνδροφόνον οὐκ ἐᾶν λέγειν. [7] ἐγὼ δὲ οἶμαι ἡμᾶς, ὦ ἄνδρες δικασταί, οὐ περὶ τῶν ὀνομάτων διαφέρεσθαι ἀλλὰ τῆς τούτων διανοίας, καὶ πάντας εἰδέναι ὅτι, ὅσοι ἀπεκτόνασί τινας, καὶ ἀνδροφόνοι εἰσί, καὶ ὅσοι ἀνδροφόνοι εἰσί, καὶ ἀπεκτόνασί τινας. πολὺ γὰρ ἂν ἔργον ἦν τῷ νομοθέτῃ ἅπαντα τὰ ὀνόματα γράφειν ὅσα τὴν αὐτὴν δύναμιν ἔχει· ἀλλὰ περὶ ἑνὸς εἰπὼν περὶ πάντων ἐδήλωσεν. [8] οὐ γὰρ δήπου, ὦ Θεόμνηστε, εἰ μέν τίς σε εἴποι πατρα-

278

willi

λοίαν ἢ μητραλοίαν, ἠξίους ἂν αὐτὸν ὀφλεῖν σοι δίκην, εἰ δέ τις εἴποι ὡς τὴν τεκοῦσαν ἢ τὸν φύσαντα ἔτυπτες, ᾤου ἂν αὐτὸν ἀζήμιον δεῖν εἶναι ὡς οὐδὲν τῶν ἀπορρήτων εἰρηκότα. [9] ἡδέως δ’ ἄν σου πυθοίμην (περὶ τοῦτο γὰρ δεινὸς εἶ καὶ μεμελέτηκας καὶ ποιεῖν καὶ λέγειν)· εἴ τίς σε εἴποι ῥῖψαι τὴν ἀσπίδα, ἐν δὲ τῷ νόμῳ εἴρηται, ‘ἐάν τις φάσκῃ ἀποβεβληκέναι, ὑπόδικον εἶναι’, οὐκ ἂν ἐδικάζου αὐτῷ, ἀλλ’ ἐξήρκει ἄν σοι ἐρριφέναι τὴν ἀσπίδα, λέγοντι ὅτι οὐδέν σοι μέλει; οὐδὲ γὰρ τὸ αὐτό ἐστι ῥῖψαι καὶ ἀποβεβληκέναι· [10] ἀλλ’ οὐδ’ ἂν τῶν ἕνδεκα γενόμενος ἀποδέξαιο, εἴ τις ἀπάγοι τινὰ φάσκων θοἰμάτιον ἀποδεδύσθαι ἢ τὸν χιτωνίσκον ἐκδεδύσθαι, ἀλλ’ ἀφείης ἂν τὸν αὐτὸν τρόπον, ὅτι οὐ λωποδύτης ὀνομάζεται. οὐδ’ εἴ τις παῖδα ἐξαγαγὼν ληφθείη, οὐκ ἂν φάσκοις αὐτὸν ἀνδραποδιστὴν εἶναι, εἴπερ μαχῇ τοῖς ὀνόμασιν, ἀλλὰ μὴ τοῖς ἔργοις τὸν νοῦν προσέξεις, ὧν ἕνεκα τὰ ὀνόματα πάντες τίθενται. [11] ἔτι τοίνυν σκέψασθε, ὦ ἄνδρες δικασταί· οὑτοσὶ γάρ μοι δοκεῖ ὑπὸ ῥᾳθυμίας καὶ μαλακίας οὐδ’ εἰς Ἄρειον πάγον ἀναβεβηκέναι. πάντες γὰρ ἐπίστασθε ὅτι ἐν ἐκείνῳ τῷ χωρίῳ, ὅταν τὰς τοῦ φόνου δίκας δικάζωνται, οὐ διὰ τούτου τοῦ ὀνόματος τὰς διωμοσίας ποιοῦνται, ἀλλὰ δι’ οὗπερ ἐγὼ κακῶς ἀκήκοα· ὁ μὲν γὰρ διώκων ὡς ἔκτεινε διόμνυται, ὁ δὲ φεύγων ὡς οὐκ ἔκτεινεν. [12] οὐκοῦν ἄτοπον ἂν εἴη τὸν δράσαντ’ ἀφεῖναι φάσκοντα ἀνδροφόνον εἶναι, ὅτι ὁ διώκων ὡς ἔκτεινε τὸν φεύγοντα διωμόσατο; τί γὰρ ταῦτα, ὧν οὗτος ἐρεῖ, διαφέρει; καὶ αὐτὸς μὲν Θέωνι κακηγορίας ἐδικάσω εἰπόντι σε ἐρριφέναι τὴν ἀσπίδα. καίτοι περὶ μὲν τοῦ ῥῖψαι οὐδὲν ⟨ἐν⟩ τῷ νόμῳ εἴρηται, ἐὰν δέ τις εἴπῃ ἀποβεβληκέναι τὴν ἀσπίδα, πεντακοσίας δραχμὰς ὀφείλειν κελεύει. ‘[6] Now, judges, he will perhaps not put forward any of these things in his defense, but rather say to you what he already dared to say to the arbitrator, namely that it does not fall under the rules about forbidden expressions if someone says “to have killed the father”: for the law does not forbid this, but instead does not permit to call someone a “homicide”. [7] But judges, I believe that we do not have disputes about words but about their content, and that everybody knows that all those who have killed someone are also homicides, and all those who are homicides have also killed someone. It would be a big task for the legislator to write down all the words which have the same meaning; instead, by using one, he indicated all of them. [8] For, Theomnestus, if someone called you a “father-striker” or a “mother-striker”, you would then think that he is liable to you, but if someone said that you had been hitting her who gave birth to you or him who begat you, you would believe that this person should not be punished since he had not used any forbidden expression. [9] I would like to ask you this (for in these matters you are an expert and have experience both in terms of doing them and talking about them): If

register variation and tense/aspect/mood categories in greek

279

someone said that you had thrown away your shield—when in the law it says “If someone says ‘to have cast away’, let him be liable”—would you not sue him, but accept “to have thrown away the shield”, saying that it does not matter? After all, “to throw away” and “to have cast away” is not the same. [10] Similarly, if you were one of the Eleven and someone charged someone else by saying that his cloak had been taken away from him or his dress taken off him, you would not receive the case, but dismiss it in the same way because the term “clothes-stealer” was not used. Nor, if someone were caught abducting a child, would you say that he is a “kidnapper”, since you will fight about words and not pay attention to the deeds for which everybody uses the words. [11] So, judges, consider also the following: It looks to me as if this person has never gone to the Areopagus, out of laziness or weakliness. For you all know that, when they judge murder trials in that place, they do not take the party-oaths by using that term, but by using the one with which I have been slandered: the prosecutor swears that the other person killed, and the defendant swears that he did not kill. [12] Would it then not be odd to release the one who is guilty by saying that he is a “homicide”, whereas the prosecutor had sworn that the defendant “killed”? In what way is this different from what this person is about to say? And you yourself brought a charge for defamation against Theon who had said that you “have thrown away” your shield. Yet, in the law nothing is said about “to throw”: only when someone says “to have cast away” the shield, the law makes him owe five hundred drachmai’. lys. 10.6–12

By explicitly equating the formulations (τὴν ἀσπίδα) ῥῖψαι and (τὴν ἀσπίδα) ἀποβεβληκέναι (10.9), the speaker plays down not only the potential semantic difference between the two (which is his primary aim), but also the tense-aspect contrast. However, even his biased wording does not conceal the fact that the actual defamatory statements which are or were at stake in the current and previous lawsuits referred to, or which are envisaged by the laws quoted, are formulated with the perfect,35 not the aorist (see 10.6 presupposing a formulation with ἀπέκτονε, 10.9 with the law using ἀποβεβληκέναι, 10.12 with the actual wording ἐρριφέναι, in contrast to 10.9 ῥῖψαι). The aorist, or alternatively the imperfect, are used instead in the speaker’s hypothetical sketch of situations that, in his

35

Or a noun: cf. Lys. 10.6 and 10.8 on laws proscribing as defamatory the terms ἀνδροφόνος and πατραλοίας/μητραλοίας.

280

willi

view at least, could be regarded as equally defamatory (10.8 εἰ δέ τις εἴποι ὡς τὴν τεκοῦσαν ἢ τὸν φύσαντα ἔτυπτες, 10.9 εἴ τίς σε εἴποι ῥῖψαι τὴν ἀσπίδα).36

12.8

A Quantitative Approach

Turning to the quantitative approach, which may be more decisive when we want to make wholesale claims about register-based preference patterns, things are more complicated. In order to substantiate our hypothesis, we would ideally need statistical information on the usage of perfects vs. aorists—in particular, but not exclusively, transitive active perfects vs. transitive aorists— not just in Greek oratory, but also (to demonstrate the supposed divergence between registers) in a variety of other, contemporary genres. Such statistics do not exist, and to compile them in a meaningful way—making sure, for example, that only like and like are compared—would go well beyond the scope of the present paper.37 We must therefore limit ourselves to a representative sample study, looking at the distribution of one high-frequency verb which, thanks to its semantic generality, is as widely attested in oratory as it is in most other 36

37

Quite naturally, in Lys. 10.10 the perfect is also used in such a hypothetical case, but the controversy imagined there is different in that it does not revolve around a defamatory statement. In 10.11–12, where reference is made to the wording of the oaths sworn by the prosecuting and defending parties before the Areopagus tribunal, the non-use of the perfect is equally expected: the oaths have to assert whether or not x committed a specific murder, not whether x is a murderer. Wackernagel (1904) does not contain statistics, nor does Chantraine (1927) for the matters that are of interest here (cf. Chantraine 1927:129, about the 4th c. data: ‘On ne saurait donc prétendre énumérer tous les parfaits résultatifs que la langue a formés’.). SickingStork (1996:187–245) list all the perfects in Aeschylus, Sophocles, Euripides, Aristophanes, Thucydides, and Lysias, plus some of Xenophon’s works; but they are not contrasted with the frequencies of other tense-aspect stems, nor are totals for sub-categories of the perfect given, let alone set in relation to text length. Cloud (1910) does the latter, but only for the various orators (and individual speeches), and active and medio-passive perfects are not distinguished (as they should be, since the share of the former increases over time: see Duhoux (1995) for Antiphon, Thucydides, Lysias, Isaeus, Demosthenes, and Duhoux (1996) for the non-Attic dialects). Perhaps the most interesting outcome of Cloud’s counts is that, while there is much internal variation from speech to speech, the overall frequency of perfects is similar in all the orators except Antiphon and Andocides (with noticeably fewer perfects: 2.90 and 2.87 perfects ‘per page’, respectively) and Dinarchus (with substantially more of them: 7.91; contrast Lysias 4.17 ~ Isaeus 3.46 ~ Isocrates 4.14 ~ Lycurgus 4.13 ~ Aeschines 4.09 ~ Hyperides 4.01 ~ Demosthenes 4.75). On the whole, this is in line with the findings presented below.

register variation and tense/aspect/mood categories in greek

281

genres: ποιέω ‘to make’. In fact, the limited statistical data cited by Chantraine suggests that πεποίηκα is one of the most common ‘resultative’ or transitive active perfects in classical times across genres.38 The results of this sample study are presented in Table 12.1, whose columns give: i.

the total number of words (in thousands, rounded from the figures in the Perseus database, www.perseus.tufts.edu) in the works of each of the authors named to the left, ii. the total number of active perfect indicatives, infinitives, and participles of ποιέω (established by a lemmatized search on the Thesaurus Linguae Graecae database, www.tlg.uci.edu), iii. the ratio of col. ii : col. i, which indicates how common perfects of ποιέω are in relation to the overall length of text, iv. as a cross-check, the total number of active aorist indicatives, infinitives, and participles of ποιέω (again based on a lemmatized tlg search), v. the ratio of col. iv : col. i, which shows the relative frequency of aorists of ποιέω, vi. the ratio of col. ii : col. iv, which is of interest as the transitive active perfect naturally competes with the aorist, and finally, for additional reference, vii. the total number of occurrences of forms of ποιέω in the authors selected, and viii. the ratio of col. vii : col. i, quantifying the overall frequency of ποιέω. Col. viii confirms that, while ποιέω is indeed a favourite lexeme of the orators, the verb cannot be said to be distinctly oratorical: the figure for Herodotus is similar to that for Lysias and Demosthenes. The most striking feature here is perhaps the fairly low frequency of ποιέω in both Plato and Aristophanes, who may go together as they both—though each in his own way—portray everyday conversation. For the issue under consideration, however, this is of little relevance. Instead, cols. iii and vi are revealing. To start with, col. iii illustrates the growth of the transitive active perfect per se, in each genre (broadly defined). In historiography, Xenophon has many 38

Cf. Chantraine (1927:123–138): note, for example, his figures for Herodotus vii (2 πεποίηκα, out of 10 ‘resultative perfects’ in total) and Aristophanes (15 πεποίηκα, exceeded only by 18 δέδρακα and 17 εἴρηκα, both of which would have been less suited for a sample study, the first because of its restricted distribution across genres, the second because it is not always as clearly transitive as is πεποίηκα).

282

willi

table 12.1 A sample study on the use of active perfects in Classical Greek

Author

Hdt. Thuc. Xen. Pl. Aristoph. Antiph. And. Lys. Isocr. Dem. Is.

i. Words (k)

ii. pf. act.

iii. ii:i

iv. aor. act.

v. iv:i

vi. ii:iv

vii. ποιέω

viii. vii:i

185 150 312 574 102 20 20 60 119 294 35

10 5 32 64 18 5 6 36 49 185 20

0.05 0.03 0.10 0.11 0.18 0.25 0.30 0.60 0.41 0.63 0.57

231 75 311 175 46 14 36 84 124 327 42

1.25 0.50 1.00 0.30 0.45 0.70 1.80 1.40 1.04 1.11 1.20

0.04 0.07 0.10 0.37 0.39 0.36 0.17 0.43 0.40 0.57 0.48

1218 534 1737 1887 354 52 139 403 879 1973 304

6.58 3.56 5.57 3.29 3.47 2.60 6.95 6.72 7.39 6.71 8.69

more perfects than Herodotus and Thucydides,39 whereas in oratory Lysias, Demosthenes, and Isaeus have more than Andocides or Antiphon. Isocrates, it will be noticed, has a substantially lower figure than, say, Lysias. In isolation one might be tempted to interpret this as an indication that lawcourt oratory was more prone to use transitive active perfects than epideictic oratory, but this is where col. vi becomes relevant. Here Lysias and Isocrates have very similar figures, meaning that each of the two was about as likely as the other to make use of a perfect or an aorist in reference to a past eventuality described with ποιέω. In other words, the real growth of the transitive active perfect, as effected at the expense of the aorist, is shown in col. vi, with its peak in Demosthenes’ work.40 Here, Andocides, not his elder Antiphon, ranks lowest among the orators, but we can only speculate about the reasons: if it is meaningful at all, the outcome may have less to do with Andocides’ aristocratic background than with the fact that, of all the orators listed, he was the only non-professional who would also depart from the ‘standard rules’ of the genre in other respects. But leaving aside

39

40

In later historiography, the perfect becomes even more prominent, to judge by the statistics given in de Foucault (1972:132) for Polybius; de Foucault also illustrates the strength of the ‘resultative’ perfect in the same author (134–137). But note that, whilst the divergence between Demosthenes and Andocides is significant (p < 0.05), the one between Demosthenes, and, say, Lysias, Antiphon, or Plato is not.

register variation and tense/aspect/mood categories in greek

283

Andocides, col. vi drives home our main point: oratory—including epideictic oratory—was the ‘macro-genre’ most actively contributing to the emerging boom of the perfect, surpassing by far the authors of literary prose, and even, though by a lesser margin, conversational speech (if we can trust the Platonic and Aristophanic data for this). So, bearing in mind the preliminary nature of the data, we can make a case for the crucial spread of the ‘resultative perfect’ (or whatever name one chooses to use) in the late 5th and early 4th c. bce being not simply a matter of diachronic variation, as conventionally presented, but also a matter of register-related variation, with oratory and everyday conversation heading a more general development that was ultimately to result in the Hellenistic perfect-aorist merger.

12.9

Conclusion

Notwithstanding its limitations and shortcomings, it is hoped that the small case study just presented demonstrates how much we can potentially learn about ‘hidden’ historical developments in a ‘dead’ language like Ancient Greek as soon as we start paying attention to distributional patterns across registers. If it has been stressed at the outset that register variation and diachronic variation must not be mixed up, we now see, more positively, how much they are intertwined. To believe that the one can be studied in depth without the other being studied as well would be as serious a mistake as to believe that we cannot properly do either because of the complexities on the ground. The fact that we sometimes have to content ourselves with approximations does not invalidate the undertaking as such. In order to appreciate the specific—the individual utterance or word—we first need an idea of the general. This is what variationist linguistics seeks to provide, and why the discipline is as indispensable to our dialogue with the past as it was for the ancient rhetoricians’ didactic response to the present.

References Allan, R.J. & M. Buijs (eds.). 2007. The language of literature: linguistic approaches to Classical texts. Leiden. Bakker, E.J. (ed.). 1997. Grammar as interpretation: Greek literature in its linguistic contexts. Leiden. Bakker, W.F. 1966. The Greek imperative: an investigation into the aspectual differences between the present and aorist imperatives in Greek prayer from Homer up to the present day. Amsterdam.

284

willi

Bers, V. 1984. Greek poetic syntax in the Classical Age. New Haven. Bile, M. & J. Lallot 2008. ‘Les oppositions aspectuelles dans la “loi de Gortyne”’. In: Hodot & Vottéro 2008:7–27. Buijs, M. 2007. ‘Aspectual differences and narrative technique: Xenophon’s Hellenica & Agesilaus’. In: Allan & Buijs 2007:122–153. Caragounis, C.C. (ed.). 2010. Greek: a language in evolution. Essays in honour of Antonios N. Jannaris. Hildesheim. Chantraine, P. 1927. Histoire du parfait grec. Paris. Claflin, E.F. 1905. The syntax of the Boeotian dialect inscriptions. Baltimore. Cloud, F.L. 1910. Use of the perfect tense in the Attic orators. Norristown, pa. dge = Schwyzer, E. 1923. Dialectorum Graecarum exempla epigraphica potiora. Leipzig. Duhoux, Y. & F. Dachy. 1992. ‘L’aspect verbal: du mycénien à l’indo-européen’. Mykenaïka, ed. by J.-P. Olivier, 215–237. Athens-Paris. Duhoux, Y. 1995. ‘Les diathèses du parfait en attique littéraire aux ve–ive s.’. Eos 83:43– 45. Duhoux, Y. 1996. ‘La fréquence décroissante du médio-passif au parfait depuis le ve s. avant j.-c.: une caractéristique de l’attique et de la koiné’. hs 109:53–72. Duhoux, Y. 20002. Le verbe grec ancien: éléments de morphologie et de syntaxe historiques. Louvain-la-Neuve. van der Eijk, P.J. 1997. ‘Towards a rhetoric of ancient scientific discourse: some formal characteristics of Greek medical and philosophical texts (Hippocratic Corpus, Aristotle)’. In: Bakker 1997:77–129. Fanning, B.M. 1990. Verbal aspect in New Testament Greek. Oxford. de Foucault, J.-A. 1972. Recherches sur la langue et le style de Polybe. Paris. Hajnal, I. 2004. ‘Die Tmesis bei Homer und auf den mykenischen Linear B-Tafeln: ein chronologisches Paradox?’ Indo-European perspectives, ed. by J.H.W. Penney, 146– 178. Oxford. Haug, D. 2010. ‘The language of Homer’. In: Caragounis 2010:87–100. Hodot, R. & G. Vottéro (eds.). 2008. Dialectes grecs et aspect verbal. Nancy-Paris. Horrocks, G.C. 1981. Space and time in Homer: prepositional and adverbial particles in the Greek epic. New York. ig =1873–. Inscriptiones Graecae. Berlin. IvO = Dittenberger, W. & K. Purgold. 1896. Die Inschriften von Olympia. Berlin. Jenkins, S. & J. Hinds. 1987. ‘Business letter writing: English, French, and Japanese’. tesol Quarterly 21:327–349. de Jonge, C.C. 2007. ‘From Demetrius to Dik: ancient and modern views on Greek and Latin word order’. In: Allan & Buijs 2007:211–231. de Jonge, C.C. 2008. Between grammar and rhetoric: Dionysius of Halicarnassus on language, linguistics and literature. Leiden. Kieckers, E. 1909. ‘Zum Gebrauch des Imperativus Aoristi und Praesentis’. if 24:10–17.

register variation and tense/aspect/mood categories in greek

285

Kühner, R. & B. Gerth. 1904. Ausführliche Grammatik der griechischen Sprache. ii: Satzlehre. Hannover. Lehmann, C. 1982. ‘Nominalisierung: Typisierung von Propositionen’. Apprehension: Das sprachliche Erfassen von Gegenständen, ed. by H. Seiler & C. Lehmann, 66–83. Tübingen. McKay, K.L. 1965. ‘The use of the Ancient Greek perfect down to the second century a.d.’. bics 12:1–21. Minon, S. 2001. ‘Emplois et valeurs de l’optatif dans le dialecte éléen’. Les modes dans les dialects grecs anciens, ed. by R. Hodot, 297–322. Nancy. Minon, S. 2007. Les inscriptions éléennes dialectales (vie–iie siècle avant j.-c.). Genève. Minon, S. 2008. ‘L’aspect dans l’énoncé de loi éléen’. In: Hodot & Vottéro 2008:171– 191. Moorhouse, A.S. 1982. The syntax of Sophocles. Leiden. Morpurgo Davies, A. 1985. ‘Mycenaean and Greek language’. Linear b: a 1984 Survey, ed. by A. Morpurgo Davies & Y. Duhoux, 75–125. Louvain-la-Neuve. Parke, H.W. & D.E.W. Wormell. 1956. The Delphic oracle. ii: The oracular responses. Oxford. Post, L.A. 1938. ‘Dramatic uses of the Greek imperative’. AJPh 59:31–59. Pulleyn, S. 1997. Prayer in Greek religion. Oxford. Rijksbaron, A. 1984. ‘Het Griekse perfectum: subject contra object’. Lampas 17:403–419. Ruijgh, C.J. 1985. ‘L’emploi “inceptif” du thème du présent du verbe grec: esquisse d’une théorie de valeurs temporelles des thèmes temporels’. Mnemosyne 38:1–61. Schlachter, L. 1907/8, 1908/9, 1909. ‘Statistische Untersuchungen über den Gebrauch der Tempora und Modi bei einzelnen griechischen Schriftstellern’. if 22:202–242, 23:165–204, 24:189–221. Schwyzer, E. & A. Debrunner. 1950. Griechische Grammatik. ii: Syntax und syntaktische Stilistik. München. Scott, J.A. 1907. ‘Prohibitives with πρός and the genitive’. CPh 2:324–330. Sicking, C.M.J. 1991. ‘The distribution of aorist and present tense stem forms in Greek, especially in the imperative’. Glotta 69:14–43, 154–170. Sicking, C.M.J. & P. Stork. 1996. Two studies in the semantics of the verb in Classical Greek. Leiden. Slings, S.R. 1997. ‘Figures of speech and their lookalikes: two further exercises in the pragmatics of the Greek sentence’. In: Bakker 1997:169–214. Slings, S.R. 2002. ‘Figures of speech in Aristophanes’. The language of Greek comedy, ed. by A. Willi, 99–109. Oxford. Slotty, F. 1915. Der Gebrauch des Konjunktivs und Optativs in den griechischen Dialekten. i. Göttingen. Wackernagel, J. 1904. ‘Studien zum griechischen Perfectum’. Programm zur akademischen Preisverteilung, 3–24.

286

willi

Wackernagel, J. 19262. Vorlesungen über Syntax mit besonderer Berücksichtigung von Griechisch, Lateinisch und Deutsch: erste Reihe. Basel. Wackernagel, J. 19282. Vorlesungen über Syntax mit besonderer Berücksichtigung von Griechisch, Lateinisch und Deutsch: zweite Reihe. Basel. Willi, A. 2003. The languages of Aristophanes: aspects of linguistic variation in Classical Attic Greek. Oxford. Willi, A. 2007. ‘Of aspects, augments, aorists—or how to say to have killed a dragon’. Greek and Latin from an Indo-European perspective, ed. by C. George et al., 34–48. Cambridge. Willi, A. 2010a. ‘Attic as the language of the Classics’. In: Caragounis 2010:101–118. Willi, A. 2010b. ‘Register variation’. A companion to the Ancient Greek language, ed. by E.J. Bakker, 297–310. Chichester.

Index locorum Achilles Tatius (Ach. Tat.) 2.24.3 171 Aeschylus (Aesch.) Agamemnon (Ag.) 126 269n3 Eumenides (Eum.) 587–588 276n2 Prometheus bound (pv) 171 269n3 767 269n3 948 269n3 Aesopus (Aesop.) 4.5

171

Ammonius grammaticus Περὶ ὁμοίων καὶ διαφόρων λέξεων (Diff.) 19.1–4 158n2 Andocides (And.) 1.119.3 Anna Comnene Alexiad 1.1.2 1.2.7 1.4.1 1.15.3 1.16.6 1.16.10 2.2.4 2.4.7 2.9.4 2.12.2 11.3.2 11.4.5 11.6.3 11.7.4 11.11.6 11.12.6 Prologue 4.1 Metaphrase 8

64

239 237 239 239 235 236 239 237 236 238 235n1 239 235n1 236 238 239 239 237

100 114

238 240

Apocalypsis Enochi 97.11 171 Apollonius Dyscolus De adverbio (Adv.) 142.9 158n2 144.1 159n De constructione libri quattor (Synt.) 353.6–354.9 159n Apollonius Sophista 125.8 158n2 Archippus fr. 27

263

Aristophanes (Aristoph.) Peace (Pax) 259 203 543 60 551–555 263 Thesmophoriazusae (Thesm.) 1050–1051 161 The Acharnians (Ach.) 172 263 303 68 1000–1001 263 The Birds (Av.) 115 160 448–450 263 656 173 985 269 1040–1041 262 1572 203 1660–1666 263 The Clouds (Nub.) 633 203 791–792 264 1296 203 1298 203 1299 203 1352 202, 202n2

288 The Knights (Equ.) 127 176–177 196–201 1020 1087 The Wasps (Vesp.) 397 671 704 731–732 937–939 Wealth (Plut.) 742–744

index locorum

269n3 269n3 270 269n3 269n3

203n 202n2 60 181 263 61

Callimachus (Callim.) Epigrams (Epigr.) 17.1 170 Callinicus Life of St Hypatius 50.6

230

Chronicle of the Morea (Chron. Mor.) 18–19 67 240–241 66 329 66 3647 (h) 233 3650–3651 (h) 233 4233–4237 (h) 233 Cicero (Cic.) Letters to Atticus (Att.) 10.16.3 176 Constantine Porphyrogenitus De administrando imperio chapter 1 220n2 Council of Chalcedon 2.1.1.66 169 Cyril of Jerusalem Procatechesis 15

229

Demades (Demad.) Fragments (Fr.) 78.16 169 Demosthenes (Dem.) 18.156 127–128 18.189–190 175 19.124–125 176 23.117 179 25.44.5 180 49.47 161

Digenes Acrites 1532 (e)

232

Diodorus of Sicily (Diod. Sic.) 1.67.3 148 1.77.3 161n2 10.18.2 170 19.19.4 148 Dionysius of Halicarnassus (Dion. Hal.) Isocrates (Isoc.) 2 262 Thucydides (Thuc.) 31 261 Epithalamium (Constantinople) Folio 7v, 1–4 232 Eupolis Fragments ( fr.) 195

269

Euripides (Eur.) Hecuba (Hec.) 394–395 161n2 Helen (Hel.) 1674 62 Ion 567–568 166 Iphigenia in Tauris (it) 27–28 122 Medeia (Med.) 1320 179

289

index locorum Evagrius Ecclesiastical history (he) 68.25 169 Herennius Philo a.35

158

Hermas Shepherd 89.4

228

Herodianus Grammaticus Partitiones (Epim.) 270.12–16 159n Philetaerus (Philet.) 255.1–3 159n Herodotus (Hdt.) 1.2 1.8.6 1.17.3 1.41.2 1.110.3 1.122.3 1.129 1.185.6 1.196.5 2.93.4 2.157.1 2.161.4 3.11.2 3.13.3 3.15.2 3.16.2 3.43.2 3.57.4 3.71.5 3.80.4 4.13 4.31.2 4.118.4 4.139.1 4.151 4.156.1 4.190.1 5.17.1 5.24.1 5.37.2 5.56

23n2 176 14 160, 161n1 19 14 23n2 16 16 19 57 14 145 146 144 64 15 269n3 160, 161n2 175 23n2 68 73 11 23n2 126–127 65 126 100 15n 127n

5.98.1 6.68 7.8.1 7.50.2 7.140 8.6.2 8.61.2 8.74.2 8.76.2 8.76.3 8.77.1–2 9.22.3 9.51.3 9.99.3 Hesiod (Hes.) Theogony (Theog.) 7–8 87 102 102–103 103 254 418 420 437–443 438 442 443 Works and days (Op.) 5 6 7 20–24 45 185 213–218 215 222–224 238–243 245–247 288 289 289–290 292 312 325 347 354–356

15n 23n2 15n 160, 161n1 270 17 73 13 12 15n 269n3 68 13 15

94n1, 95 92n2 92n2 87 92n2 92n2 92n2 92n3 86 92n2 92n2 92n2 92n2 92n2 92n2 93 92n2 92n2 84 92n2 94 87 96 92n2 92n1 94n1, 95 92n2 92n2 92n2 91 96

290

index locorum

Works and days (Op.) (cont.) 360 92 379 92n2 401 92n2 498–500 97 508 92n1 512 92n1 580 92n1 678–681 94 721 92n2 740–741 91 762 92n2 Homer (Hom.) Iliad (Il.) 1.176–178 1.415–416 4.11 4.442–443 9.698–699 13.306 13.452 14.84–85 15.596–599 16.689–690 17.177–178 17.542 18.367 18.338 22.329 22.438–439 23.68 23.692–693 24.253–254 24.582–586 Odyssey (Od.) 2.172–173 3.76–78 4.356 11.70 11.439 12.156–157 13.358 17.250 18.399–400 19.592

275 177, 178 142 142 180–181 265n2 58n2 161, 161n2 11 93n 93n 274n2 161n2 265 265n2 75 265n2 142 167, 181 11n 143 11n 95n 57n 77 11n 57n 19 143 265

Inscriptions cirb 731 dge 335 336–347 411 412 477 481 482 491 492 507 508 509 511 512 515–517 528 I.Didyma 529 I.Eph 27 2211b 2226 2265 2299b 2304 2311 2435 2446 2476 3216 4337 I.Kalchedon 16 I.Smyrna 223b ic iv 41 iv 43 iv 46 iv 47 iv 72 iv 73 ix 15 ig i3 680

202n1 268 268n1 267 266 268n2 268n2 268n2 267n2 267 268n2 268n2 267 267n1 267n1 267n1 267n1 201 202 201 202 202, 201 201 201 201 201 201 201 201, 202 202 201 202 200 200 200 200 200 200 200 58

291

index locorum i3 1118 ii/iii2 1358 ii/iii2 1363 iv 72 v 2 262 vii 1779 vii 3083 vii 3357 vii 3359 vii 3362 vii 3364 vii 3376 vii 3378 ix 2 517 xii 4 94a xii 4 94b xii 4 100 xii 4 152 xii 4 274 xii 4 275 xii 4 276 xii 4 277 xii 4 278 xii 4 279 xii 4 280 xii 4 281 xii 4 290 xii 4 298 xii 4 302 xii 4 304 xii 4 306 xii 4 307 xii 4 310 xii 4 311 xii 4 313 xii 4 315 xii 4 318 xii 4 319 xii 4 324 xii 4 325 xii 4 326 xii 4 328 xii 4 332 xii 4 335 xii 4 341 xii 4 343 xii 4 350

58 211 211 200 274n4 267 267 267n4 267n4 267n4 267n4 268 268 73 195–196, 197 196 192n2, 193 191 205, 206, 207–208, 208– 209, 210, 213, 215 206, 209 205, 206, 210 206 196, 205, 206, 208, 208n, 209, 210, 215, 215n4 206, 209, 215 206, 208, 209 206, 207, 212, 215 197 193, 194, 196, 210–211 192, 194 192, 193, 195, 198 192, 193, 194 210 199 198 192 192, 195 192 192, 195, 198, 199 192 193 192, 195 193, 199 210, 214n 191, 198, 200n4 194 194 198

xii 4 358 xii 4 359 xii 4 364 xii 4 887 igdop 14 iospe i 24 iv 342 IvO 4 7 lsam 41 lscg 7 20 96 lss 10 mama 8 544 8 545 8 552 8 553 seg lvi 636 tam ii 55

210 210 210, 214 192n2 201 201 202n1 267 266 211 211 211 211 211 201 201 201 201 74 201

Isaeus (Is.) 8.3

64

John Moschus Spiritual Meadow 45

230

Lucianus (Luc.) Nigrinus (Nigr.) 14.1 Soloecista (Sol.) 15.1 15.1.28–15.2.2 Symposium (Symp.) 20.4 Lysias (Lys.) 1.14 1.17

23n3 171 183, 185n 23n3

70 70n

292 Lysias (Lys.) (cont.) 1.42 3.32 4.3 10.6 10.6–12 10.8 10.10 12.16 12.40 12.74 13.41 14.41–42 19.52 28.5 29.13

index locorum

77 72 166 279n 279 279n 280n1 71 77 70 70 277 76 70 202n2

Martyrdom of Carpus 34.3 230 Moeris Grammaticus 205.16–18 159n New Testament (nt) 1 Corinthians (1 Cor) 4.8 7.26 7.40 12.2 15.35 2 Corinthians (2 Cor) 8.7 11.5 Acts 7.25 10.43 11.28 12.15 15.11 23.12 23.30 24.9 25.4 25.19 26.26 Epistle of James ( Jas) 2.14

170 25n3 25n1 228 228 231 26 25n3 25n6 25n3, 52n 25n1 25, 50n3 25n6 26 25n1 25n3, 50n3 25n3 25n3 25n1

Hebrews (Heb) 4.1 11.5 11.24 John ( Jn) 4.17 5.39 12.29 17.21 20.2 21.25 Luke (Lk) 2.26 12.51 19.18 24.23 Mark (Mk) 6.25 8.12 9.26 10.42 Matthew (Mt) 3.9 5.29 6.29 8.8 9.18 13.17 16.11 20.33 Revelation (Apoc) 5.7 Romans 3.6 3.8 3.28

26 26 25n3, 50n3 45 25n1 26n2 45 45 26n5 25n6 45 45 25n3 22n2 228 45 25n3 25n3 22n2 45 22n2 45 45 45 231 42, 50 227 25n3 25n3

Orpheus (Orph.) Argonautica (a.) 1159

169

Ostraca O. Claud. 1.111 1.113 1.124–136 1.126 1.127 1.129

245 245n3 245n4 245 245, 252n 247

293

index locorum 1.138 1.139 1.156 1.161 1.166 1.173 2.225 2.227 2.229 2.232 2.233 2.234 2.243 2.245 2.246 2.249 2.275 2.284 2.366 2.374 2.379 2.386 4.884 4.889 4.892 Papyri bgu 93 214 248 371 451 530 531 584 594 595 665 830 843 846 923 1078 1095 1097 1141 1563 1776

249n3 249n3 252n 248 253n2 252n 256 256 255 256 256 256 244, 253n1 253 244, 253, 253n1, 257 253n1 257 245 249n3 249n3 252n 252n 253n2 252n 252

251n5 53 53 36 53 53 53 53 53 53 53 53 53 53 53 53 53 53 247n1 53 27

1873 2246 2493 Chr.Mitt 80 Chr.Wilck 237 cpr 1.103 7.1 7.52 7.54 15.2qtpl 15.8 P.Alex.Giss 37 P.Amh 68 130 P.Ashm 24 P.Athen 29 62 P.Bacch 24 P.Bad 35 90 P.Berl.Moeller 4 7 9 P.Brem 6 P.Col 215 P.Congr xv.15 P.Corn 49 P.Dura 18 19 P.ifao 2.3 P.Fay 109 110

246 27, 53 53 53 53 53 39 53 251n5 28, 29, 32n1, 53 53 53 53 53 53 53 53 30, 53 31, 32n1, 34, 35, 47, 53 53 53 53 53 53 53 29, 29n2, 30, 34, 53 53 30, 35, 38, 39, 50n3, 53 53 247n 44, 53 249, 249n3

294 P.Fay (cont.) 111 112 113 114 120 121 122 123 P.Flor 61 P.Fouad 14 21 P.Fuad.Univ 31 P.Hamb 4 29 30 P.Haun 28 P.Iand 104 P.Koeln 56 P.Lips 105 P.Lond 897 1168 P.Mert 12 P.Mich 121 194 203 236 263 344 635 656 10779 P.Mich.Mchl 25 P.Michael 15 P.Mil.Vogl 2.50

index locorum

53 249, 249n3, 250 251 249, 249n3 251 251n2 251n2 53 53 30, 53 53 53 53 53 53 53 53

P.Muench 3.1.118 P.Oslo 50 P.Oxf 10 P.Oxy 37 44 73 286 295 298 300 325 393 745 985 1027 1143 1155 1480 1672 1763 2190

53 53 53 27, 53 53 29n1, 33n1, 36, 53 30, 31, 53 53 53 53 53 53 53 53 53 53 246n2

2198 2339 2352 2353 2725 2756 2757 3057 3070 3264 3274 3357 3503 10240 P.Princ 188 P.Ryl 2 P.Sarap 90 P.Stras 274 334

42n1 53 249 53 53 53 53 44, 53 53 43, 53 53 42n1 53 35, 39, 50n3, 53 53 39 41, 53 36, 46, 48, 53 23n1, 53 230 24n2, 37, 46, 48, 53 42n1 53 23n1 53 53 53 53 47, 53 53 53 37, 53 53 53 53 53 53 47, 53 53 53

295

index locorum P.Tebt 286 302 409 P.Turner 18 P.Vind.Bosw 1 P.Vind.Tand 10 P.Yale 77 psi 1109 sb 3924 5230 5231 5233 5236 5275 5761 6705 7258 7268 7341 7356 7461 7463 7737 8070 8947 9120 9271 10967 11017 11127 11585 11625 11644 12172 13222 13303 13590 13867 StudPal 4, pg. 58–78

53 27–28, 31, 32n1, 34, 35, 50n3, 53 53 53 35, 38

upz 1.60

166

Pausanias (Paus.) 4.20.1 9.14.3 9.17.5

269n3 269n3 269n3

Phrynichus (Phryn.) 93.193 159

53 53

Pindarus (Pind.) Olympian Odes (Ol.) 8.42 269n3

35 53 53 29 31, 32, 53 35, 53 53 53 53 53 53 27 53 53 27, 53 53 29, 30, 32n1, 35, 47, 53 53 53 53 53 53 53 53 53 43n2, 44, 53 53 53 34, 53 53 248n4 53

Plato (Pl.) Apology (Ap.) 32c 33b Charmides (Chrm.) 157c Crito (Crit.) 44d Phaedo (Phd.) 118a Phaedrus (Phdr.) 117c 237a Protagoras (Prt.) 338a Symposium (Symp.) 175a 178b The republic (Resp.) 377e 586c

72 23n2 68 168, 181 161n1 69 264 202 203n1 23n2 23, 23n2 23n2

Plautus Miles Gloriosus (Mil.) 911 175 Pliny the Elder Natural History praef. 26 Plutarch (Plut.) Camillus (Cam.) 14.3 Moralia (Mor.) 5.4

58n1

23n3 23n3

296

index locorum

Romulus (Rom.) 28.4–5 23 Themistocles (Them.) 18.6 23n3 20.1 23n3 Ptolemeus De differentia vocabulorum (Diff.) 391.8–9 158n2 Septuagint Exodus (Ex) 16.3 Job 30.24 Psalms (Ps) 118.5

170 182 170

Sophocles (Soph.) Ajax (Aj.) 738–739 127 Antigone (Ant.) 442 276n2 927–928 164 Electra (El.) 126–127 164 Oedipus at Colonus (oc) 956 202 Philoctetes (Phil.) 1363 175 Spaneas p 208

232

Thucydides (Thuc.) 1.26.1 1.67 1.89.2 2.3.4 2.23.3 2.25.4 2.31.1 2.92.1 3.3.3 3.15.1 3.22.3 3.22.8 3.33.8 3.104.4 4.6.2

111 23n2 104 20, 20n1 103 103 105 62 145 146 145–146 20 20n1 78n 124

4.9.2–4 4.31.1–32.1 4.34.1 4.68.6 4.83 4.105.1 4.108.2 5.50.4 6.45.1 6.50 6.65.2–3 6.91.3 6.96.3 6.98.2 7.17.4 7.26.1–2 8.17.1 8.17.2 8.17.3 8.66.1 8.87.3 8.93.1 8.98.3 Xenophon (Xen.) Agesilaus (Ages.) 1.6–3.1 1.38 Anabasis (An.) 1.2.21 1.2.22 1.3.3 1.3.5 2.1.4 2.3.19 3.1.19 3.2.17 7.6.23 Cyropaedia (Cyr.) 8.1.26 8.7.26 Hellenica (Hell.) 2.3.39 4.3.1 4.3.13 5.1.18 Memorabilia (Mem.) 1.2.46 7.2.1

105–106 106 57 73 23n2 122 73–74 76 78 109–110 107–108 269n2 20n1, 21n1 108 20n1, n2, n3 107 107n 107n 107n 61 20n1 107n, 108 107n

269n1 77 77 147 147 202n2 167 79 69n2 65 175 76 69n2 65 78 79 173 164 77

Index nominum et rerum Aarts 135 accomplishment 102, 104, 108, 112, 113, 114, 121, 125, 126, 133, 139, 142, 145, 148, 154, 197, 215 achievement 112, 113, 114, 115, 121, 122, 123, 126, 133, 139, 142, 148, 150, 152, 154, 197, 216 actionality 64n2, 131, 132, 141, 154 activity 46, 49, 50n3, 92, 96, 112, 113, 119, 121, 123, 124, 133, 138, 142, 145, 148, 154, 199 Aeschines 280n2 agent 145, 194, 195, 196, 198n2, 200, 276n1 agent noun 274n4 agentive 200, 201, 273 Aktionsart 101, 104, 111–115, 120–124, 129, 131, 132–136, 137, 139, 140, 141, 143, 145, 146, 147, 153–155, 196n3 Alexandria 244 Alexiad 220, 235–240 Amigues 20, 21 Ammonis 252 anastrophe 262, 264, 265n2, 272 Andocides 280n2, 282 Anna Comnena 220, 234 anterior(ity) 22, 23, 36, 38, 39, 46, 51, 51n2, 61–62, 63, 64, 66n1, 67, 69, 72, 86, 87, 94, 95, 97, 121, 122, 140, 150, 226n1 Antiphon 280n2, 282 aorist 39–40, 53, 56, 57, 60, 84, 104, 109, 118, 119, 120, 123, 124n1, 124n2, 127n1, 137, 141, 143, 150, 152, 152n1, 153–155, 273–276 complexive 123, 123n3 confusion between aorist and future forms 46–47 co-occurrence with the imperfect 148 effective 123n2 empiric 87–88, 96, 97 functional competition with the perfect 280–283 functional merger with the perfect 49– 52, 283 gerund 66, 67 gnomic 84–98, 129, 215n3 origin of 86–90 synchronic value of 90–93 imperative 60, 198, 244n1, 249n3, 250, 253, 255, 256, 258, 259, 271

indicative 43–45, 48, 58, 59, 59n1, 75, 85, 92–93, 95–96, 97–98, 100, 101–102, 105, 109–110, 111, 128, 142, 144, 145, 148, 226n, 240, 267, 279, 280–282 infinitive 22, 23, 24, 25, 26, 35, 36–41, 43, 45, 47, 48, 52–53, 140, 197, 198, 253 passive 39, 50n3 ingressive 123 loss of morphological distinctions between aorist and perfect 41–43 optative 71 participle 63, 64, 66, 68, 251, 253 relationship with Aktionsart 120–124 subjunctive 60, 227, 229, 229n1, 229n2, 235, 237–238 tragic 129 Aphrodisias 202 Apollinaris 257 Arcadian 266n1 Ariphrades 262 Aristophanes 261n1, 281 Aristotle 261 Arteimas 245 aspect 56–59, 62, 84, 92–93, 111–124, 131–132, 132–136 aspect frequencies in private letters 40 aspectual and/or temporal distinctions 22, 23, 52–53 aspectual composition 197 aspectual opposition 141, 154 choice 138, 141, 154 default 144 generic 215 grammatical 111, 118, 120, 121, 131 habitual 28, 59, 71, 71n1, 91, 92, 132, 134, 135n, 137, 138, 143, 145, 212, 215, 223, 224 imperfective 56, 111, 118, 121, 127n1, 141, 153, 154, 215, 222, 223, 228n iterative 65 lexical see Aktionsart Modern Greek 137–140 morphological 134 morphology 121, 126 diachrony of 136–137 neutrality 57, 221

298 perfective (aoristic) 49, 56, 63, 97, 98, 118, 121, 123, 127n1, 141, 153, 154, 215, 222, 223 progressive 58, 59, 134, 135n, 137, 223n, 224n resultative 56 Atticism 185, 186, 227, 234 Atticists 152–153, 154, 227, 228 of the Roman period 234 Medieval 234 Atticistic influence 23 Atticistic norm 48 Atticizing variant 237 augment 39, 41, 42, 66n3, 88–90, 143, 161, 162, 167, 168n1, 171 auxiliary 140, 140n, 161, 165, 165n1, 167, 172, 173, 183, 184, 185, 186, 230

index nominum et rerum Coptic 253, 254–255, 258 alpabet 244 Sahidic 254, 259 Coseriu 3, 3n1 counterfactual(ity) 71n1, 71n2, 80–81, 89, 144, 158–187 Cretan 200 Croft 3, 4 cumulative 198 custom 214

declarative 80, 164n, 177, 178, 179, 180, 184, 222, 225, 226, 229, 234, 235, 240 infinitive see infinitive Delphi 267 Demosthenes 280n2, 281, 282 demotic script 244 Bache 134 desiderative Bakker 88–89 illocutionary force see illocutionary Bary 92 force Bechtel 76 sentence 163–173 Boeotia 267–268 desire 158n2, 159, 163–173, 178, 184, 185 borrowing 254, 259 De Smet-Verstraete 134, 135 Bousiris 247 diachronic 109n, 131, 161–162, 167, 173–187, Bybee 59, 120–121, 189, 203, 204 216, 251, 265 variation see variation calendar 205, 206, 207, 208, 209, 210, 211, 212, dialect 21, 75, 243, 265–268, 269 213, 214, 215 Coan 189–217 discourse 210, 211, 214, 215 dialectal sub-groups 202 cancellation/cancellability (of implicatures) Didyma 202 176, 176n, 177, 178, 276 diglossia 219 Carlson 212 digression 106, 107, 194, 207, 212 Chantraine 10, 19, 26n5, 41, 42, 142, 273n3, Dinarchus 280n2 274, 274n2, 281, 281n Dionysius of Halicarnassus 227, 261, 262n1 Chronicle of the Morea 233 Dioscorus 255–256 cline 187 direct speech 10, 45, 69, 70, 71, 75, 78, 238, Cognitive Linguistics 101, 111–124, 129, 134 248, 248n4, 252 complement clause 22, 43–45, 48, 225, 246, discourse conditions 194, 211 251, 252 documentary (non-literary) papyri 22–53, confusion 227 morpho-syntactic 36n3, 36n4, 39, 41, 42n, durative 56, 57, 59, 64, 66, 69, 77, 112, 113, 114, 50, 242–259 145, 153, 197, 197n, 199 phonetic/phonological 162–163, 171, 173, 183, 250, 253, 256, 259 Egyptian 253, 254, 257 construal 101, 111, 115–119, 121, 124n2, 125, 129, script 255 134 -speaking environment 243 default 67, 70–71 Greek see Greek non-default 67, 70–71 Elean 202, 266, 266n temporal 66, 71 inscription 201

index nominum et rerum Ephesos 201 Eustathius of Thessalonike Evans 150

221

Fanning 148, 150 Fayyum 242, 248n6 final clause 9–21, 60, 72, 223, 243 formulae/formulaic language 59n1, 132, 143, 150, 191–197, 199, 201–202, 205–211, 227, 251, 253, 256, 259 Fraenkel 76 functional merger 41, 42, 49, 50, 51 functional overlap 222, 225–232 functional reanalysis 220 future 10, 18, 56, 57, 59, 60, 91, 136, 138, 197, 203–205, 216–217, 219–241, 269 deontic 192, 199, 200, 202 futurity 221 futurus pro imperativo 178–179 imperatival 177, 179, 245, 250, 251 in the middle to late Byzantine vernacular 232–233 indicative 9, 191–205, 212, 221–222, 224, 227, 228, 229–232, 235–237, 239, 240, 243, 251, 269 infinitive 22, 24, 25, 26, 35, 36, 43, 45–47, 49, 51, 52, 59 stative 51 non-stative 51 modal 189–217, 235–237 optative 59, 238–240 participle 59, 63 perfect see perfect periphrastic 49, 52, 52n, 57n, 136, 165n1, 230, 231, 232, 233, 234, 238 reference 10, 12, 13, 15, 18, 91, 222–225, 226, 228, 232 -referring forms 240 -referring modality 221 García Ramón 76 generic 215, 224, 226n, 235n aspect see aspect sentence 212, 213, 214 gerund 58, 66–67 Goodwin 9, 10, 19 Gortynian inscriptions 199, 200 grammatical polysemy 216–217

299 grammaticalization/grammaticalized 61n, 66, 71–72, 134, 165n3, 172, 173, 174, 177, 179, 180, 182, 183, 187, 231 grammatical, lexical and stylistic ‘transpositions’ 233–234 Grata 245–248 Greek Attic 21, 74, 76, 220, 225, 234, 235, 240, 243, 265 comedy 203 belletristic 233 Classical 23, 23n1, 48, 52, 66, 136, 141, 142, 148, 165, 202, 219, 274, 275 Egyptian 253 Hellenistic and Roman 22–53, 227–232 Homeric 88, 141, 265n2, 273, 273n3, 274, 274n2 Ionic 74, 265 Koine 74, 148, 201, 227, 229, 243, 267 Medieval 66, 219–241 vernacular 221 Modern 23, 41, 57n, 58, 66, 67, 71n1, 71n2, 131–155, 159, 165n2, 204, 256 Mycenaean 265n2 New Testament 23, 25, 35, 45, 48, 49, 50, 51, 148, 165 Greeks 244 habits of conduct 212 habitual see aspect Halasana 195 Hauptzweck 9, 10, 11, 12, 14, 17, 18 Heraclean tablets 201 Herakleopolites 247 Herodotus 9–21, 127, 149, 281, 282 Hesiod 84–98 historiography/historiographical texts 132, 140, 144–153, 281, 282n Homer 10, 17, 19, 71, 75, 76, 88, 89, 98, 141–144, 154, 171, 176, 177, 181, 184, 185, 261, 274 Humbert 12, 87 Hyperides 280 idiolect 242, 242n2, 245–251, 258 idiosyncrasy 240, 253 illocutionary force 163, 164, 169, 172, 173, 177–179, 183, 185 desiderative 163, 164n, 173

300

index nominum et rerum

particle 158–187 potential 203 type 190 imperative 172, 178, 180, 189–217, 242–259, 261, 266n, 271, 271n3, 271n4, 272 imperfect 26, 57–59, 89, 90n2, 100–129, 136, 143, 144, 145, 147, 148, 150, 152, 203, 215n3, 226n, 228, 279 creating a temporal framework 103–105 discourse-structuring 105–107 habitual 143, 224 oblique 72–80, 81 of continuing relevance 107–111, 124– 128 of likelihood 122–123 implicature 62, 122, 142 conversational 88, 98, 174–177, 178, 179, 180 conventional(ized) 174, 177, 179 inception of state 198 incrementality 132, 133, 134 indirect object 246 individuation 275n2 infinitive 56, 59, 152–153, 167, 172, 182, 184, 185, 194, 196, 205, 208, 209, 211, 214, 215, 232, 242–259 articular 243 aorist see aorist declarative 22–53, 78 dynamic 22–53 future see future imperatival 96, 202 jussive 262, 262n2, 264, 266, 272 loss of 136, 138 optatival 262 perfect see perfect present see present provisional 197, 199, 200, 201, 207 intentional 212 Isaeus 280n2, 282 Isocrates 280n2, 282 juridical language Kühner-Gerth

197, 202

9, 12, 19, 20

l1 (language/speaker) 244, 245n4, 249, 254, 255, 256, 259 l2 (language/speaker) 249, 253n3, 255

language contact Greek-Egyptian 243, 253n3, 258 Latin 58n1, 61, 69, 71n2, 175–177, 245n4, 248, 249, 258 Leiwo 4–5 Leontas 247 list 194, 211 form 206, 211 structure 205 Lucius Bellienus Gemellus 248–251 Lycurgus 280n2 Lysias 277, 280n2, 281, 282 Mandilaras 23, 35, 40, 48, 51, 243–244 manumission inscriptions 267, 272 matrix 68, 69, 212–213 McKay 92 meaning postulate 174–177 metaphrase 221 modality 163, 164, 169, 189–217, 221–225, 226, 232, 233, 234, 240 deontic 71, 160, 162, 179, 180, 186, 189–192, 196, 198–204, 208–209, 212–213, 215–216, 230, 230n1, 232 epistemic 71, 72, 160, 162, 180, 186, 189, 202, 204–205, 224, 229, 230, 230n1, 232, 240 epistemological 72 future-referring see future Mons Claudianus 242, 244, 245, 249n2, 250, 252, 253, 255, 256, 258 mood 56, 72n1, 128, 137, 159, 163, 164, 165, 167, 168, 170, 172, 173, 182–184, 185, 189, 191, 205, 206, 209, 210, 211, 212, 226n, 242– 259, 261–283 Mopurgo Davies 76, 191 Moulton-Turner 148 narrative 100, 102, 103, 105, 109, 111, 132, 140, 144, 148, 150, 154, 155, 246, 248, 259 Nebenzweck 9, 10, 11, 12, 14, 17 necessity 174, 189, 196, 198, 199, 200, 201, 202, 203, 204, 208, 209, 216 negation 167, 168, 170, 172, 176, 179, 180, 201, 201n2 New testament 23, 24, 25–26, 32, 33, 35, 39, 45, 48–49, 50, 51, 52, 148, 150, 227, 231

index nominum et rerum Nicetas Choniates 221 nominal style 205 nominative 167, 172, 205, 206, 212, 248, 262n2 nominativus/accusativus cum infinitivo 166, 168, 251 non-factual character 212 non-factuality 204, 213 non-quantized 197, 198, 199 non-verbal clauses 205 construction 209 syntagma 207, 210n1, 215 official documents 24, 34, 46, 47–48, 49 Olbian decree 201 omnitemporal 84–98 optative 9–21, 59, 71, 89, 141, 163, 182, 223– 225, 226, 226n, 227–229, 234, 235, 235n, 236, 237, 238–240 future see future iterative 71n1, 124n2 loss of 136 oblique 9, 18, 21, 69–72, 74, 75, 76, 77, 78, 78n, 79 of wish 231 potential 9, 19n, 92n2, 224, 226, 226n, 227, 228, 234, 237, 266 orthography 219, 250, 251n1, 252, 253n3, 254 o-stem genitive 269 Pantikapaion 201 parody 262 participle 56, 59, 63–67, 67–69, 162, 168, 171, 172, 242, 243, 245, 250, 251, 273n3 aorist see aorist future see future perfect see perfect present see present particle 9, 85, 90, 106, 107, 165, 165n3, 227n, 264, 265, 266 illocutionary see illocutionary path of development 204n1 patient 198n2, 200, 273 patientive 201 pattern of occurrences of events 213 perfect 26, 40–41, 53, 56, 60–62, 90n3, 119, 136, 137, 140, 140n, 150, 152, 153–154, 240, 273–276, 276–280 active 274, 275n1, 276, 281–282

301 anterior 51, 61n, 61–62, 66n1, 140, 150, 152, 152n confusion between aorist and perfect 39 functional merger with the aorist see aorist future 61, 62 indicative 35, 44, 45, 48, 141–142, 256 infinitive 26, 34–36, 39, 40–41, 42–43, 46, 47, 48, 49, 50, 51, 52, 281 loss of morphological distinctions between aorist and perfect 41–43 medio-passive 274, 280n2 participle 63, 64, 66, 68, 252, 253, 281 periphrastic 152, 256 pluperfect (past perfect) 61, 71n2, 75, 77, 89, 144, 152n, 240, 273n3 resultative 51, 61n, 274, 281, 282n, 283 stative-intransitive 273 perfectivity see aspect performative use 165–167, 190, 191, 203 status 202 periodic 210, 214 periphrasis 58, 136, 137, 152, 226, 256 periphrastic future see future periphrastic perfect see perfect Petenephotes 244, 253, 254, 257, 258 Plato 281 Plutarch 23 Polybius 148, 151, 282n Porter 25, 26, 150 Porter-O’Donnell 39–40 possible/possibility 163–165, 169, 174, 183, 185, 189, 196, 199, 201, 202, 203, 204, 208, 209, 216, 222, 223, 224, 226n, 229, 230n1, 234 posterior(ity) 22, 31, 36, 46, 51, 63, 226n pragmatic dimension 217 pragmatically inferred 203 prayer 261, 271, 272 present 26, 39, 56, 57n, 60, 63, 115, 116n, 118, 120, 122, 123, 124n2, 127, 137, 141, 153, 155, 186, 250, 256, 266n, 269, 273 gerund 67 imperative 198, 199, 250, 253, 258, 271, 271n4, 272 indicative 44–45, 74, 74n1, 75, 78, 78n, 80, 85, 86, 91, 92, 93, 128, 142, 205–216, 229, 229n1, 229n2, 231n, 267 non-stative 44

302 infinitive 22, 25, 26–33, 35, 37, 47, 51, 197 stative 25n3, 39, 47, 48, 49, 52 non-stative 25, 28, 29, 31, 32, 36, 38, 43, 46, 48, 50, 50n3 participle 63, 64, 65, 66, 68, 69, 69n2, 128, 248 praesens pro futuro 269–270, 269n2, 272 praesens pro perfecto 128n, 129 scheduled 212–216 present or past reference 204 private letter 24, 32, 33, 34, 39, 40, 46, 47–48, 51, 242–259 Protagoras 261 punctual(ity) 56, 57, 59, 64, 66, 114–115, 126, 132, 153, 154, 197 purpose clause 9–21 quantized 197, 198, 199 question status 202 reduplication 41, 42, 43, 47, 89 reference time 61n, 197n1, 269 register 219–241, 248, 251, 259, 261– 283 relative tense marker 9–21 theory 63, 64, 65, 66, 67–69, 72, 80 relative time reference 59–62, 63–67, 69, 80, 81 relevance current 140 continuing see imperfect reported speech 69, 70, 71, 72, 77, 81 restrictor 212–213 Rijksbaron 10, 12, 22, 63, 92, 95, 100, 105, 275n1 Roman fortresses 244 Ruipérez 75, 76, 120 schwa 253, 259 scribe 242–259 semantic change 173n2, 177 path 216 sentence type 177, 178, 180 simultaneity 22, 31, 51, 60, 63, 64, 65, 66, 67, 72, 79, 103n1, 122

index nominum et rerum speech act 177–179, 183, 190 Stahl 87 stativity 24, 50, 51, 150 non-stative verb 27, 28, 43, 49, 50n3 state 24, 28, 46, 49, 50n3, 60, 61, 63, 64, 69, 92, 112, 113, 119, 123, 124, 128n, 133, 139, 142, 144, 145, 148, 153, 154, 155, 198, 215, 273n2, 274 stative verb 32, 33, 48, 49, 50n3, 92 stress accent 253, 254, 259 subevent 124n2, 197, 198n2 subjunctive 9–21, 59, 60, 87, 89, 91, 94, 124n2, 137, 169, 182, 201, 219–241 future/potential (aorist) subjunctive 237–238 Successus 245, 245n4, 246, 247 telicity 111, 113, 128, 136, 155 atelic predicate/verb/state of affairs 112, 119, 120, 121, 123, 124, 133, 153, 199, 215 processes 215 telic accomplishment 215 achievement 216 predicate/verb/state of affairs 102, 112, 113, 119, 120, 121, 123, 124, 128, 133, 153, 197, 198 Telmessus 202 tense-backshifting 56–81 terminativity (terminative) 87, 91, 97, 132, 136, 155 Thorley 24, 49 Thucydides 10, 20, 21, 101–111, 149, 152n, 153, 261, 282 time of the event 197n1 reference absolute 59–62 relative 59–62, 63–67, 69, 80, 81 timeless 85n2 tmesis 264, 265, 265n2, 272 tragedy 264, 272 typicalization 275n2 typological 81, 203 Van der Auwera-Plungian Vandaele 9, 12

189

index nominum et rerum variation contact-induced 243 dialectal 75 dimensions of 3 domains of 3 genre-determined 221, 262 grammatical 191 idiolectal 248 internal 243, 258, 280n2 levels of 4 modal/mood 10, 209 morpho-syntactic 3, 243, 244, 250, 256 orthographic 248 register 261–283 register-internal 268–272 scribal 248 Vendler 112, 132, 133, 134, 199

303 verbs of (immediate) perception 56–81, 139 Verdenius 87 Verkuyl 134, 137 voice active 66, 255, 256 medio-passive 39n1, 66, 273, 274, 280n2 middle 240, 255, 256 passive 50n3, 51n1, 240, 246, 273n3 Wachter 191 Wackernagel 87, 171, 179n2, 274 West 87 Willi 89 Willmott 10, 19 wish 71, 162, 163, 165n1, 165n2, 166, 178, 180, 181, 182, 184, 222, 223, 226, 227, 230, 231