The Language of Hesiod in Its Traditional Context 063112750X, 9780631127505

124 65 11MB

English Pages 248 [256] Year 1971

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

The Language of Hesiod in Its Traditional Context
 063112750X, 9780631127505

Citation preview

THE LANGUAGE OF HESIOD IN ITS TRADITIONAL CONTEXT

PUBLICATIONS

OF

THE PHILOLOGICAL XXII

THE LANGUAGE OF

SOCIETY

HESIOD

IN ITS TRADITIONAL CONTEXT

by

G. P. EDWARDS Lecturer in Humanity, University of Aberdeen

PUBLISHED FOR THE SOCIETY BY BASIL BLACKWELL, OXFORD 1971

©

THE

PHILOLOGICAL

SOCIETY,

1971

ISBN 0 631 12750 X Library of Congress Catalog Card Number : 70-129598

To

R. B. E.

PRINTED STEPHEN AUSTIN

IN ENGLAND BY AND SONS, LTD.,

HERTFORD

PREFACE This book studies the language of the three surviving Hesiodie poems and considers their place in the early Greek hexameter tradition. While it has been written primarily for classical scholars, it is hoped that it will also be of interest to linguists in other fields, especially those who are concerned with the relationship between oral tradition and written literature. I have tried to bear in mind the different needs of these readers both in the discussion itself and in giving references. (The abbreviations which have been used for these are explained on p. 210.) The research on which the book is based has extended over several years, first in Cambridge and more recently in Aberdeen. At an earlier stage the work took the form of a dissertation

entitled * Un-Homeric features in the language of Hesiod’, which was approved for the degree of Ph.D. by the University of Cambridge in 1965. At about the same time there began to appear a spate of scholarly books and articles relevant to the subjeot, and this has made it desirable to expand some parts of the work and reduce the scale of others before putting it into print. I am very much indebted to the Philological Society both for undertaking publication and for allowing me time to carry out the revision. In particular I should like to thank the successive Hon. Secretaries for Publications, Professor D. M.

Jones, Dr. N. E. Collinge and Mr. C. J. E. Ball for their patience and helpfulness. The Council of the Society has generously agreed to my suggestion that the Bibliography should serve also as an Index of the works referred to in the text, and I

believe that readers will find this a useful arrangement. It is a pleasure to have this opportunity of thanking Professor W. S. Allen, Dr. J. Chadwick, Professor R. A. Crossland and Professor D. L. Page for valuable comments and suggestions which they have made to me, especially Dr. Chadwick, who supervised most of my work as a research student and gave me much encouragement. For help in obtaining books I am grateful to my colleagues Mr. E. J. P.

Vi

Raven

PREFACE

and Mr. H.

Thomson,

and to the staff of Aberdeen

University Library and the library of Christ's College, Cambridge. I am glad to acknowledge also the assistance which I have received from Mrs. B. L. Farmer and Mrs. D. Bell in their most capable typing of the script, and the resourcefulness and courtesy shown by the Printers and Publishers ın the production of the book itself. It 15 only fair to add that the responsibility for faults and inadequacies must be regarded as entirely my own. Lastly I wish to thank my wife for her unfailing encouragement, advice and practical help. She has ably combined the role of uxor optima with that of eruditissima castigatriz, and itis with sincere gratitude that the book is dedicated to her. G. P. E. King’s College, Old. Aberdeen. July, 1970

CONTENTS CHAPTER Preface I. II.

Introduction . The text of the poems and published work on their language

III.

Traditional poetic language in Hesiod

IV.

Repetitions and formulas

V.

The principle of economy

VI. VII. VIII. IX. X. XI. XII.

‘ Parallels of sound ' Versification and enjambement Un-Homeric word-forms and their dialect sources

101

Traditional alternatives between them

122

and

Hesiod’s

choice

The accusative plural of a- and o-stems

141

Phrases common to Hesiod and the Odyssey

166

Conclusions

190

Bibliography and Index of works cited

2]0

Index of Hesiodie references .

and

principal

Homeric 229

General Index

24]

Index of Greek

246

CHAPTER

I

INTRODUCTION "Ióuev ψεύδεα πολλὰ λέγειν ἐτύμοισιν ὁμοῖα, ἴδμεν δ᾽, εὖτ᾽ ἐθέλωμεν, ἀληθέα γηρύσασθαι. Theogony 27-8.

AMONG the most important achievements of classical scholarship since the beginning of the present century must be counted the growth of a new insight into the complex origins of our earliest Greek literature. In particular, the work of Witte,

Meister,

Parry,

Chantraine

and many

others in this

field 1 has made possible ἃ much decper understanding of the language of the Homeric poems, explaining the artificiality of their dialect and many peculiar features of their style as having arisen in the course of an oral tradition of hexameter poetry, with an elaborate system of formulaic phrases, developing over several centuries. While this work has largely centred upon the Iliad and the Odyssey, it is directly relevant to the study of early Greek poetry other than the great epics, and one of the main purposes of this book is to consider its application to the much smaller corpus of literature which has come down to us as the work of Hesiod ; at the same time we shall examine some

of the featurcs in the language of the Hesiodic poems which are of special interest because they are un-Homeric and distinctive. Of the various hexameter poems ascribed to Hesiod in antiquity,? most have survived only in fragments which are derived partly from citations in ancient authors and partly from papyri.? The best known of these fragmentary poems is 1 Details of their work are given in the Bibliography. For a concise account of progress since about 1900 see E. R. Dodds, ' Homer as oral poetry ’ and L. R. Palmer, ‘ Homer and the philologists ’ in Fifty Years (and Twelve) of Classical Scholarship, 1968, 13-24, and their Appendix on the most recent work, ibid., 38-46.

2 Their precise number is by no means easy to determine: over 20 titles of poems are known, but many of these are hardly morc than names, some of them being merely alternative designations for a single work. For a full discussion see J. Schwartz,

3 Now West,

Pseudo- Hestodeia, 1960, 13-32.

collected in Fragmenta

1967.

Hesiodea,

ed. R. Merkelbach

and M. L.

THE

9

LANGUAGE

OF

HESIOD

the Ehoiai or Catalogue of Women, of which we possess some 140 complete lines from citations, the longest single piece consisting of eleven lines ; papyri have furnished a number of more continuous passages, but many of these are mutilitated and there is often uncertainty about their place and authenticity.t Three of the poems however have been preserved They are the Theogony through a manuscript tradition. (1022 lines), the Works and Days (828 lines) and the Shield of Herakles (480 lines), and there are a number

of reasons for

treating these three poems as a group. Not only are works with these titles attributed to Hesiod by ancient authors, but several of our manuscripts include two of the poems and some have all three.5 The Rainer Papyrus ® of the fourth century A.D. contains fragments of all three poems, and we possess also a titulus from another papyrus 7 of about the same date which reads : HCIOAOY OEOIONIA ΕἸΡΓΑΚΑΙΗΜΕΊΡΑΙ ΑΟΙΠΟ. There is reason to believe that this grouping goes back at least to the first century A.D.® Yet evidence of this kind cannot be pressed as indicating that the poems were composed by the same author, or even at the same period or in the same place. In such matters as these we must turn to the internal evidence of the poems themselves, and here the close study of language is especially important. 4 Apart from the Ehoia? our knowledge is very limited indeed: the Melampodia, with a total of some 25 lines recorded in citations, is the best known of the remaining works of which fragments are preserved. 5 The earliest instance is the codex Parisinus suppl. graec. 663 (eleventh or twelfth century), which, though fragmentary, includes portions of Th. and Sc., while the Florentinus Laur. 32.16 (thirteenth century) has Th., Op. and Sc.;

see further below, p. 11, n. 1.

$ Published by C. Wessely, Studien zur Palaeographie und Papyruskunde, 1901, ii-xxiii ; see now West,

Hesiod. Theogony,

1966, 64.

? Paris. suppl. graec. 1099 ; see U. Wilcken, SPAW HT, 52. 8 See West, HT,

51.

1887, 808, and West,

INTRODUCTION

3

Although the Hesiodic poems are for the most part well removed from heroic epie in their subject-matter, their language is patently similar to that of Homer : it is artificial both in its mixture of dialect forms and in its use of phrases which regularly serve as formulas in the [lad and the Odyssey. On the other hand it is not identical with Homer's: the presence in Hesiod of un-Homeric dialect forms has long been known,? and it has been demonstrated also that Hesiod shows

some independence in his use of formulaic phrases.!? Two obvious questions arise: first, are we dealing with a lind of poetry which is fundamentally different from Homer's in its mode of composition ? It is arguable that some part may have been played by writing in the emergence of the Jlad and Odyssey in their present form, but it will be generally agreed that they are at least composed 1n a style which is basically oral. Can one then detect in the style of the Hesiodic poems— in their language in the broadest sense—un-Homeric features which betray them as written literature, as has been suggested for example of the Works and Days by Carpenter and of the Theogony by West?! Secondly, to what extent are these linguistic differences from Homer to be accounted for simply by reference to Hesiod's provenance ? The Hesiodic poems, on their own evidence, stem from mainland

Greece and not

from Ionia. What un-Homerie features in their dietion—in their language in the narrower sense—arise from this difference in their place of origin ? In the course of considering these main questions, evidence will arise which has a bearing on two further problems about which there exists ἃ wide divergence of opinion, as may be readily illustrated by comparing statements made about Hesiod, many of which are to be found in recent works, or works stillin common use. The first of these concerns the unity of the Hesiodie corpus, and the second the chronological position of Hesiod, especially his relationship to Homer. ? See below, p. 18, n. 27.

10 Notably by P. F. Kretschmer and I. Sellschopp ; see below pp. 19 f.

48. 11 R. Carpenter, Folk Tale, Fiction, and Saga, 1946, 16 ; West, HT', 40,

4

THE

LANGUAGE

OF

HESIOD

As far as the unity of the corpus is concerned, the strongest suspieion undoubtedly is attached to the Shield, which was already believed by some ancient critics not to be by Hesiod, and is today regularly referred to as pseudo-Hesiodic.? The only part of the poem which is sometimes allowed to be Hesiod's own work is the first 56 lines (or 54 lines, following Mazon), which are loosely linked to what follows and which appear once to have formed part of the Ehoiar.!? * Considérons donc comme probable que nous avons là un fragment authentique d'Hésiode' writes Mazon ; but Lesky maintains that even this passage ‘sicher nicht von Hesiod gedichtet ist ’.14 The reasons for doubting the Hesiodic authorship of all or most of the Shield are not in the main linguistic, but we shall see below that linguistic evidence may be adduced to reinforce the other arguments, and it may even be thought decisive with regard to the passage at the beginning of the poem. When’ we turn to the other two poems we find much less agreement among scholars. Large parts of the Theogony have been regarded as interpolated, both in the nineteenth century and more recently. Thus Jacoby prints in his edition only 442 lines in full-size type out of the 1022 in the text, and Schwenn's 'echte Theogonie des Hesiodos' consists of 448 lines.° While similar in length, their reconstructions of Hesiod's original by no means coincide at all points, and each includes between 70 and 80 lines which the other rejects. Mazon, although a relatively conservative critic, nevertheless suspects four passages at the end of the Theogony, namely the battle with the Titans (7h. 687-712), the description of Tartarus (Th. 736-819), the Typhoeus episode (Th. 820-880) 12 Seo Argument A prefixed to the poem (p. 101 in Rzach, ed. min.) ; C. F. Russo, Hesiodi Scutum, 1950, 29 ff.; M. van der Valk, REG Ixxix, 1966,

450-81. 18 Again according to Argument A, ibid.; and for papyrus evidence see Fr. 195 M-W. 14 P. Mazon, Hésiode (Budé ed.), 1928, 121; A. Lesky, Geschichte der

griechischen Literatur?, 1963, 124 —

15 F. Jacoby, Hesiodos,

Theogonia,

1934, 144.

Eng. ed., 104.

1930, 139 ff.; F. Schwenn,

Die Theogonie des

INTRODUCTION

5

and the catalogue of unions between goddesses and mortal men (7h. 965-1022).16 Kirk would follow Mazon in subtracting the last three passages from the text ‘ on the ground that they break a necessary sequence . . . or are totally inconsistent with their context ’,17 while West defends all but a few lines of the

first three passages suspected by Mazon, but believes that there is nothing after Th. 900 which is the genuine work of Hesiod.18 On the other hand N. O. Brown accepts all but the Typhoeus episode as original,!? and P. Walcot is prepared to accept the Typhoeus episode as genuine along with the rest ; he writes : ‘If we allow our text to stand as we possess it, the Theogony has a structure which can be clearly defined. ... We can come to no other conclusion than that the Theogony comprises a self-contained unity.' 39 In the Works and Days also a number of passages have been held by some to be later additions, notably the Proem (Op. 1-10), which was omitted from a text inscribed on lead and seen by Pausanias in Boeotia in the second century A.D.,?' and the Days (Op. 765-828), which is excluded altogether by Wilamowitz from his edition (1928), and 15 still held to be un-Hesiodic by Lesky, who writes: “Man führt das Gedicht häufig unter dem Titel “ Werke und Tage " an, obwohl man den Zusatz über die Tagwählerei nicht unter Hesiods Namen stellen kann.’ 2? This view is shared by Solmsen in his recent and very thorough discussion of the problem.?? But a defence of these and other doubted passages will be found in Mazon's edition,2* where he concludes that none of them need be 16. Budé ed., 13-18.

17 G.S. Kirk, ‘ The structure and aim of the Theogony ' in Hés?ode et son influence, 1962, 63-95, esp. 73.

18 HT, 40, 355-8, 381 f., 398 f. 19 See the Introduction to his translation of the Theogony, 1953, 12. 20 CQ vi, 1956, 206. 21 Pausanias IX, 31; cf. Wilamowitz, Erga, 39 f., and T. A. Sinclair, Works and Days, 1. 22 GGL, 119 = HGL, 100. 23 ἘΠ Solmsen, ‘ The ‘‘ Days " of the Works and Days’, TAPA

293-320. 24 Budé

ed., 76-85.

xciv, 1963,

θ

THE

LANGUAGE

OF

HESIOD

regarded as other than part of the original poem. Sinclair also retains the passages, though he concedes the possibility that the Days may be a very early addition to the poem.?* Walcot similarly is in favour of accepting the whole of the traditional text, and supports the view that ' there is nothing here (sc. in the Days) which compels us to reject Hesiod as their author.’ ?$ A further question remains, concerning the relationship of the Hesiodie poems to each other. The Shield, as we have already seen, raises special problems, the bulk of it bemg regarded as a later composition, with only the Proem (Sc. 1-54 or 56) possibly constituting a genuine Hesiodic fragment in the opinion of some scholars. We have still to ask whether the Theogony and the Works and Days are substantially the work of one and the same poet. Classical writers ascribe both poems to Hesiod, but doubts about his authorship of the Theogony are noted by Pausanias,?? and a wide diversity of opinion is to be found among modern scholars. Beloch writes: ‘ Dass die verschiedenen Hesiod zugeschriebenen Dichtungen denselben Verfasser gehabt haben, ist durch nichts zu erweisen und an sich sehr unwahrscheinlich ' ?8—a statement which contrasts strangely with one by Rose : ‘ The likelihood of the two works having one author seems to amount almost to certainty.’ 39 Both these views have had their supporters. Croiset calls the author of the Theogony ‘le disciple d’Hesiode ’ and ‘le plus remarquable héritier de son génie’, and a similar belief has been held by Bowra, who writes that Hesiod, i.e. the author

of the Works and Days, was ‘ onc of a school of poets ’, and that ‘other works in his manner were attached to his name’ including the Theogony, which was ‘ written by an anonymous poet, who refers to Hesiod as his master ’.3° But most scholars 25 WdD, lvii.

2° REG Ixxiv, 1961, 14.

*7 IX, 31 and VIII, 18.

28. K. J. Beloch, Griechische Geschichte?, 1924, I. i, 312 n. 1. 39 Ἢ, J. Rose, A Handbook of Greek Literature, 1934, 61.

3? M. Croiset, Histoire de la littérature grecque*, 1928, I, 472, 540; Bowra,

Ancient Greek Literature,

1933, 43.

The last statement

C. M.

quoted here

from Bowra rests on an interpretation of Th. 22 ff. which is generally rejected : cf. Mazon, Bude ed., 3 f., W. Jaeger, Paideia?, 1945, I, 430 f., and W. J. Verdenius in Mnemosyne

xi, 1958, 20.

INTRODUCTION

7

of today, including the more recent historians of Greek literature, Flaceliére and Lesky, are prepared to accept substantial unity of authorship, and believe the Theogony to be older than the Works and Days. When we turn to the question of Hesiod's chronological position, it is obvious that the opinions of scholars will depend,

in part at least, on their view of the poems' unity of authorship. Thus Evelyn-White, writing in 1914, speaks of ‘ the middle of the ninth century’ as the probable date of the Works and Days,* adding that ‘ the Theogony might be tentatively placed a century later ', while ‘ the Shield of Heracles may be ascribed to the later half of the seventh century.’ ?? Few scholars would now wish to ascribe any of the poems to such an early date as the ninth century, and most are content to place Hesiod, though not the author of the Shield, in the eighth.?* An eighth century date for the Works and Days 1s put forward, for example, by Croiset and Bowra, both of whom believe in separate authorship for the Theogony, and for the author of both poems by the unitarians Rose and Sinclair, the latter favouring a date “early in the eighth century '.5* But J. A. Scott in a paper delivered in 1930 comes to the conclusion that ‘Hesiod belongs to an age nearer to Solon than to Homer ’,3® which, while no doubt intentionally nebulous, may reasonably be interpreted as supporting a lower date, probably seventh 31 R. Flaceliére, Histoire litteraire de la Grece, 111 = ZGL, 92; so also West, HT, 44.

1902,

94 f.;

Lesky

GGL,

32 This date is suggested by the well-known statement of Herodotus (IL, 53) that Homer and Hesiod lived some 400 years before his own time ; for other ancient opinions see Jacoby, Marmor Parium, 1904, 152-8. 33 Loeb edition, 1914, xxvi. 34 πὶ W. Allen in ‘The date of Hesiod’, JHS xxxv, 1915, 85-96,

ap-

proached the question mainly through the rather inconclusive astronomical evidence, and would place most of Hesiod's life in the ninth century ; but an eighth

century date was already favoured

by Rzach

in RE

VIII,

1912,

1175-7. 35 Croisct, Histoire de la littérature grecque*, 1928, I, 480 ; Bowra, Tradition

and Design in the Iliad, 1930, 262;

Rose,

A Handbook of Greek Literature,

1934, 65; Sinclair, A History of Classical Greek Literature, 1934, 64. 36 T AP A Ixi, 1930, xxvi.

8

THE LANGUAGE

OF HESIOD

century B.C. and certainly not earlier than the late eighth century. A date somewhere in this period now seems to be generally accepted: Wade-Gery and Lesky place Hesiod at about 700 B.c.;37 West would date the Theogony to about 730-700 ».c. and the Works and Days to a little later.?® A slightly lower date is favoured by Kirk, who thinks that there is no clear evidence to place Hesiod earlier than about 680 p.c.?? The Shield needs to be considered separately. Whereas the date proposed for the Works and Days has shifted, as we have just seen, by as much as 150 years, recent estimates of the date of the Shield do not differ much from that put forward by Evelyn-White in 1914. Thus in 1928 Mazon placed its composition between about 590 and 560 B.c.,4° and in the most recent

edition of the poem Russo adheres to an early sixth century date.*! In a thorough discussion of the subject using the evidence of contemporary art R. M. Cook concluded that the Shield 1s likely to have been composed 'in the early sixth century, probably about the end of the first third.’ # J. L. Myres interpreted the same evidence differently, and placed its composition rather earlier than Cook, but still considerably later than the date to which the two major Hesiodic poems are generally ascribed.*? The tendency to lower Hesiod's date is not surprising in view of the corresponding shift in the dates postulated for Yet the relationship Homer during the present century. The between the two poets is by no means wholly clear. opinions of ancient writers provide no reliable evidence about the priority of one to the other, and the confusion is well summarised in the biographical information contained in the 3? Ἢ, T. Wade-Gery,

Phoenix

iii, 1919,

1952, 1; Lesky, GGL, 110 = HGL, 91. 38 HT, 44-6. 39 The Songs of Homer,

19 Budé ed., 124. 41 Scutum,

29, 34.

42 OQ, xxxi, 1937, 213. 53 J HS lxi, 1941, 37.

1962, 283.

83 f., and

The

Poet

of the Iliad,

INTRODUCTION

9

various ‘lives ’.** Among modern scholars the view most widely held is that the Homeric poems are, generally speaking, older than Hesiod:

Whitman,

for example,

speaks

of ‘ the

dependency of the early lyric poets, not to mention Hesiod, upon the Ilad and Odyssey ',*? and Kirk writes that the stylistic evidence ' places the Iliad and the Odyssey somewhat earlier than Hesiod and the earliest Homeric Hymns '.15 But West has recently argued in favour of Hesiod's priority to Homer, holding that “the Theogony may well be the oldest Greek poem we have’, and that ‘the Iliad and Odyssey are both later, at least in their present form.’ 7 Others again believe that the relationship is more complex, and one theory in particular has the support of a fair number of scholars. It is well stated by Finley, who writes : * One plausible view is that the lad took roughly, but not precisely, the form in which we now have it in the eighth century before Christ, more likely in the latter half of the century than in the earlier ; that Hesiod flourished a generation or so later ; and that the Odyssey was composed stil another generation or two after Hesiod.’ 48 This view is held by Solmsen, who thinks that “a good part of the Odyssey may not yet have existed when he composed his Theogony *,*? and by Wade-Gery, who describes the Odyssey as ‘later than Hesiod, whereas the

Iliad

is earlier’.°°

The

most detailed arguments for this theory are those of Inez Sellschopp,5! and since they cannot be lightly dismissed, we shall consider them closely before drawing conclusions about Hesiod's chronological place. These,

then,

are

the

most

controversial

of

the

wider

questions upon which one may hope for light to be thrown 14 Seo Wilamowitz, Vitae Homeri et Hesiodi, esp. 36, 48 f., and Rzach in RE Vill, 1173-5. 45 C. H. Whitman, Homer and the Heroic Tradition, 1958, 86. 46 SoH, 283. 48 M. I. Finley,

^ The

HT, 46; cf. Schwenn, TdH, World of Odysseus, 1956, 32.

66-81.

19 Hesiod and Aeschylus, 1949, 6 and his n. 3. 50 The Poet of the Tliad, 2. 51 Stilistische Untersuchungen zw Hesiod, 1934, 42-81.

10

THE LANGUAGE

OF HESIOD

by the study of Hesiod’s language. Decisive answers cannot always be expected in matters where there is obviously much room for difference of opinion, but it is proper to ask which views are supported, or at any rate not contradicted, by the linguistic evidence. This will be borne in mind

during the discussion, which,

after the preliminary survey of relevant published work (in Chapter II), falls into two main parts. In the first (Chapters III to VII) we shall consider the extent to which traditional linguistic material was used by Hesiod and influenced him in the making of his poems, concentrating especially on topics which have a direct bearing on the question of oral composition. In the second part a number of features in Hesiod's language will be examined in detail, attention being given primarily to points in which Hesiod appears to make innovations or to show some significant difference from Homeric usage. The features here chosen for examination (in Chapters VIII to X) are ones which are of interest from the point of view of dialect, and which seem most likely to illuminate the circumstances of the poems' composition. This is followed by a reconsideration (in Chapter XI) of some phrases common to Hesiod and the Odyssey but used in different contexts, on the basis of which it was argued by Miss Sellschopp that Hesiod's language has influenced that of Homer, so that his chronological place 1s earlier than that of most of the Odyssey. Finally (in Chapter XII) we turn again to the problems of composition, authorship and chronology which have been mentioned above, in order to see what solutions seem most consistent with the evidence of the language of the poems.

CHAPTER

THE

II

TEXT OF THE POEMS AND PUBLISHED WORK ON THEIR LANGUAGE

No INQUIRY into Hesiod's language can be undertaken without some prior consideration of the history of the text. It is hoped that the following outline will provide the background essential to the understanding both of the account of scholarly contributions to the study of this subject in the second half of this chapter, and of the work in the chapters which follow. In the first place it must be remembered that our knowledge of the three surviving Hesiodic poems is primarily dependent on manuscripts which were themselves written at a time much closer to our own than to that of the poet. The mss. of Hesiod are numerous,! but nonc is earlier than the eleventh century, and most of them belong to the fifteenth and sixteenth centuries. The variant readings which they contain show the relationships between them to be very complicated : Rzach, who made a large contribution to their study at the end of the nineteenth century,? ventured to construct stemmata for the mss. of the three poems, believing them to be derived ultimately from a single archetype, probably of the period of Photius (ninth century).

Rzach's account of the ms. tradition

has been revised in some details by more recent editors? and West in particular has shown tireless diligence in examining previously uncollated mss. and in presenting a fuller view of the history of the text.* But many difficulties remain, regardless 1 A recent count lists 284 extant mss. of Hesiod, of which 73 contain Th.,

263 Op., and 62 Sc.: see N. A. Livadaras, ‘Ioropia τῆς παραδόσεως τοῦ κειμένου τοῦ ᾿Ησιόδου (1963) and West’s review of it in Gnomon xxxvii, 1965, 650-5, esp. 653.

? For Rzach's fundamental articles, published between 1893 and 1898, see Bibliography. He summarises his findings in the Preface to his editio maior, and gives an account of work done on the ms. tradition by himself, 6. Kinkel, K. 5100] and others in RE VIII, 1229-35.

3 See Mazon, Budé ed., xxiii; Jacoby, Theogonia, 72; Sinclair, W&D, xlvii-li ; Russo, Scutum, 37-51; West, HT, 52-61. 4 Seo Bibliography for West’s articles in CQ 1961, 1962, and 1964.

12

THE

LANGUAGE

OF

HESIOD

of whether the details of the medieval transmission can be unravelled or not. For even if it were possible, by systematic comparison and criticism of the extant mss., to reconstruct the

text of an archetype of Byzantine or late Roman times, we should still have recovered only a comparatively recent version of the poems. In order to go back further we must turn to such meagre evidence as is available from two other sources, namely citations and papyri. The citations of Hesiod by ancient authors are of special interest, since they represent witnesses to the text independent of the main stream of the ms. tradition. But their reliability varies from one citation to another, and 15 often hard to assess. Many of these testimonia show some divergence from the traditional text, and occasionally they have preserved a true reading which is corrupt in all the mss. ; 7 but there are also examples where they must obviously be rejected in favour of the ms. tradition.? The same is true of the fragments of papyri? for although they are 500 to 1,000 years older than our earliest mss. they are often carelessly written and contain their own peculiar errors. In general they may be said to confirm the text of the mss. Sometimes they are useful in giving a better reading or in explaining a problem in the tradition,!? but in other places they are not acceptable. It must also be remembered that neither the testimonia nor the papyri cover more than a very small proportion of the corpus. There remains an even more basic problem: the earliest 5 $ 7 8 ?

See West’s discussion and examples in HT, 67-9. See Rzach, ed. maior, ix-x ; Jacoby, T'heogonia, 85-7. e.g. Th. 850, Op. 288, 394, Sc. 397. e.g. Th. 116, 739, Op. 241 f. Listed by R. A. Pack, The Greek and Latin Literary Texts from Greco-

Roman Egypt?, 1965, 46-8 ; see also West, HT, 63-7. 10 e.g. Th. b, 28, 32, 34 etc. (seo Jacoby, T'heogonia, 758, and 169a—e, explaining Proclus (see Sinclair, W&D,

82), Op. 174, 293, liii-liv), and Sc. 7,

15, confirming a conjecture by Hermann. 1! e.g. Th. 22, 27, 51 etc. (see Jacoby,

Theogonia,

84), Op.

818

and

the

omission of several lines not otherwise suspect, viz. Op. 260, 325 f., 793, from the Rainer Papyrus, and Sc. 24, 439.

THE

TEXT

OF THE POEMS

13

papyri are still removed by several centuries from the time when the poems were composed. We may be sure that changes in alphabet and orthography took place during this earliest period of the text’s transmission,!? but there is no satisfactory means of determining what more radical alterations may have been introduced in the editing of the poems (for example by the substitution of more readily intelligible expressions for unfamiliar forms of dialect origin 13), or to what extent an early written text may have been influenced by a parallel oral one maintained by rhapsodes, which might perhaps be especially liable to expansion and interpolation. At best we may have the poems substantially as they were first composed ; at worst their language could have been subject to so much later alteration and revision that the recovery of their original form is a hopeless task. But the truth probably lies between these extremes, and the only practical approach seems to be to examine the poems as they stand in the text derived from our traditional sources, before making conjectures about the form in which they were first composed. Fortunately the metre has often safeguarded one reading from the substitution of another, and the transmitted text of Hesiod provides enough material to make a linguistic comparison with other early Greek hexameter poetry both possible and useful. It may be helpful at this point to say something about the printed editions of the Hesiodic poems. These began to appear at the end of the fifteenth century and were produced with especially remarkable frequency during the nineteenth,!? when their profusion reflects not only the increasingly systematic study and comparison of different mss. but also the need 12 cf. T. B. L. Webster,

‘ Notes on the writing of early Greek poetry ',

Glotta xxxviii, 1960, 251-63; d’Homére

13 cf.

’, REG

A.

A. Meillet,

‘Sur une édition linguistique

xxxi, 1918, 277-314.

García

Calvo,

‘ Particularidades

través del texto hesiódico ', Emerita xxxiv,

lingüísticas

recuperables

a

1966, 15-37.

14 The earlier editions are listed in S. F. W. Hoffmann, Bibliographisches Lexicon

der

gesammten

Literatur

der

Griechen

TI,

1839,

247-55;

for

the

nineteenth century see the edition of Goettling-Flach, 1878, Ixxiv-Ixxvi, and Rzach

in RE

VIII,

1235 f.

14

THE

LANGUAGE

OF

HESIOD

which many editors felt to produce texts which distinguished what they believed to be the spurious from the genuine elements in the poems. It is not difficult to single out the editions which were most important in preparing the way for the critical texts of the present century. Goettling (1831 and 1843), van Lennep (between 1843 and 1855) and Paley (1861)

each

made

use

of mss.

which

had

previously

been

neglected, and a critical edition incorporating the finds of all three of them was produced by Koechly and Kinkel in 1870, though they still did not take into account a number of mss. later used by Rzach. This work of Koechly and Kinkel was intended as the first part of a formidable trilogy, the second volume

of which

was

to show

'eadem

carmina

a me

(sc.

Koechly) in varias ex quibus olim conflata sunt particulas resoluta ', to be followed by a third containing the fragments ; although the remainder of this scheme never materialised, the edition of the main poems was a useful one which, along with Goettling's 2nd ed. for the Fragments, formed the basis of Rzach's work on the dialect of Hesiod, published in 1876.15 Perhaps it was this study which led Rzach to see the need for a new edition, based on a more thorough collation of the mss.,

and he first edited the poems himself in 1884.

There followed

his series of articles on the mss.!9 and, in 1902, his admirable

Hesiod? Carmina (editio maior). Here, after the brief introduction summarising his work on the mss. and papyri, we are given on each page beneath the text notes on (a) Homer: loc? similes—lines and phrases from Homer which resemble Hesiod; (b) Poetarum imitationes et loc? stmiles—obvious or possible imitations of Hesiod in the work of later poets ; (c) T'estes—citations of Hesiodic passages by later prose writers; and (d) Varia lectio—a full critical apparatus of the ms. and papyrus readings. Such a burden of information sometimes comes near to ousting the text from the page, but nothing can be said to be redundant, and this edition 15 See below, p. 17. 16 See above, p. 11 n. 2.

THE

TEXT

OF THE POEMS

15

is still undoubtedly our most important aid to the study of the text of Hesiod as a whole. In the same year Rzach published a smaller edition also, with the same text as the larger, but having a more select critical apparatus. The evidence of later papyrus finds was incorporated into a second edition of this smaller work in 1908, and a third edition appeared in 1913 including the readings of further papyri and two new fragments. Most of the more recent editors of Hesiod, while often giving useful commentaries, have been heavily dependent upon the work of Rzach for their text. An exception is provided by West’s Theogony (1966), which is based on a thorough reexamination of the mss. and papyri, including many not available to Rzach. The rest, although they occasionally select a better reading from the alternatives in the mss. than does Rzach,

and

eschew some

of his unnecessary

emendations,!?

rarely make an independent contribution to our knowledge of the text, and sometimes, where they differ from Rzach, need

to be used with care. Thus the edition of the Works by Wilamowitz (Erga, 1928) is full of learning both in its concise notes and in the essay on the poem as a whole; but his inclusion of a trochaic first foot at Op. 22, 132, 270, 372, 550

and 655 does not find favour with many scholars, nor does his emendation εἴοι at Op. 617, though the latter has found its way into the lists of Hesiod's linguistic peculiarities given by Schwyzer 18 and Thumb-Scherer.!? Jacoby's Theogonia (1930) aims at a text showing the poem in the form in which it left Hesiod himself,2° but this leads him to brand many parts as later additions, and he indicates these by smaller type and a complicated system of markings in the margin. The result is a 17 e.g. at Op. 33 and 39;

cf. Sinclair, W& D, ad loc.

18 Griechische Grammatik I, 1953, 108.

19 Handbuch der griechischen Dialekte II, 1959, 8. 20 * Primum Theogoniam edituro munus in eo mihi positum ut, quantum quidem fieri posset, legentium oculis carmen quale ab antiquissimo poeta profectum sit illo qui versibus ge ab ipsis Musis ad veram deorum historiam componendam gloriatur ' (op. cit., 44).

esse videbatur, subicerem tale praeclarissimis impulsum esse

16

THE

LANGUAGE

OF HESIOD

somewhat inconvenient text, though his introduction is useful,

especially the section on orthography and the treatment of testimonia.?! Other useful commentaries are Sinclair’s Works and Days (1932) and Russo’s Scutum (1950)2? The only editions to include all three Hesiodic poems are those of Evelyn-White (Loeb, 1914) and Mazon (Budé, 1928), in both of which a deep indebtedness to Rzach is obvious: EvelynWhite's edition makes no claim to containing a critical text, while Mazon's critical apparatus is so abbreviated as to be almost unintelhgible without reference to Rzach's. While we shall from time to time 1n the following chapters refer to all these editions on specific points, it seems reasonable to take Rzach's third edition as our standard text, noting where the other editors have departed from his reading when this 1s relevant to the feature under discussion. This matter of the choice of text 1s of some importance in the assembling of statistics, which may differ appreciably according to which edition is followed; for example, Rzach omits all ephelcystie -v’s before words believed to have had an initial F, and

in various places departs unnecessarily from the transmitted text. But by keeping an eye on the critical apparatus we can avoid being seriously misled, and it is clear that the choice of one of the other modern editions, or any group of them, would not substantially affect our conclusions. We may now turn to the work that has been published on Hesiod's language. For the most part this has taken the form of fairly short pieces of writing, appearing as pamphlets in the last century and more recently as articles in periodicals, but a small number of more lengthy works of wider scope have also 21 op. cit., 95-106 and 106-35. 22 Sinclair collated three mss. from the British Museum which had not been uscd by earlier editors, * and if they may be taken as a fair sample of the various other late and uncollated mss. our text has not suffered by their neglect ' (Ὁ ὦ}, li). Russo collated for the first time a fifteenth century ms. from Mutina which appears to be alone in preserving the reading κάπρω at Sc. 172. ?3 of. Rzach, Preface to ed. maior, xi-xii ; on his eagerness to ' restore ’ F

wherever possible, see below, p. 133 with n. 27.

PUBLISHED

WORK

ON

HESIOD’S

LANGUAGE

17

been produced. There are in addition many references to Hesiod in books and articles dealing with Greek language and literature, especially those concerned with Homer. The following pages attempt to single out the most important contributions to the study of Hesiod's language, and to show the main lines along which work has proceeded. À convenient starting point is Rzach's ' Der Dialekt des Hesiodos ’ (1876), which deserves special attention both because its author was to become such an eminent Hesiodie scholar and for the strong influence it has had on the subsequent accounts of Hesiod’s language. It presents a brief though systematic statement of the main phonological and morphological features of the language of the Hesiodic poems, and concludes that their dialect is predominantly Ionic but with a number of Aeolic and Doric elements which Rzach lists. On the individual poems he suggests that the Aeolisms peculiar to Hesiod, i.e. those not occurring in Homer also, belong mostly to the Works and Days, and the Dorisms to the Theogony, while the Shzeld follows more closely the pattern of Homeric language.25 Later in the same year as Rzach’s treatise was published there appeared H. Flach’s Das dialektische Digamma des Hestodos, which, though

of the

subject

24 JKPh

of digamma

Supplementband

unsatisfactory

in its treatment

itself,?* contains

an appendix

viii, 353-468 ; the fullest study to appear

prior to this is J. Fórstemann's De dialecto Hesiodea (1863), which however is Other shorter than Rzach’s and is lacking in any positive conclusions. earlier work is listed in the Goettling-Flach edition of Hesiod, Ixxvi f. See also Bibliography under Isler, Ahrens, Krémer, Sachs, Forster, Schneidewind, Fietkau, Ludwich, Treutler, Schneider, Knés, Stolz, Hinrichs, Clemm and

Kausch.

25 DH, 465 f.; the idea of a basically ‘ Doric’ T'heogony and an * Aeolic’

Works and Days was taken to extremes by A. Fick (1887). Fick had already published his celebrated (1883) and the Iliad (1886) ‘in der ursprünglichen gestellt’; he now suggested that the Theogony had the Delphic dialect (HG, 11) and the Works and Days in Aeolia (HG, 43), and he set about restoring them Sce further below, p. 201 n. 53. 26 See below, p. 132 n. 23.

in his Hesiods Gedichte editions of tbe Odyssey Sprachform wiederherbeen first composed in in the dialect of Kyme to their original form.

18

THE LANGUAGE

OF HESIOD

pp. 60-77) entitled * Der aeolodorische Dialekt des Hesiodos ' where the main conclusion drawn about Hesiod's dialect agrees substantially with that of Rzach, though it was reached independently of his work.?’ Flach’s study of the digamma in Hesiod was only one among several on this subject to be published towards the end of the nineteenth century.2? Digamma certainly drew to itself a disproportionately large share of attention,?? but other aspects of Hesiod's language were not neglected, and several studies appeared on topics in Hesiodie syntax °° and in his wordformation and voeabulary,?! including J. Paulson’s useful Index Hesiodeus (1890), which is still the standard word-list for Hesiod although it is based on Rzach's old edition of 1884 and has long been out of date for the Fragments.? Paulson published also a substantial piece of work on Hesiodic metre.?? Most of the studies published in the late nineteenth and early twentieth centuries are not simply descriptive but include also some comparison between Hesiod and other early poetry, especially Homer.®* Convenient surveys of progress in Hesiodie scholarship between 1878 and 1918 were made from

27 The identification of peculiar linguistic features in Hesiod as * Aeolic ' and ‘ Doric’ is much older than Rzach and Flach, going back not merely to Ahrens and other scholars earlier in the nineteenth century but to the ancient commentators

and scholiasts, on wbich see Flach, DDH,

62-6.

28 See Bibliography under Knós, Clemm, Devantier, Solmsen and Paues. ?? cf. the one-sided discussion under the heading ‘ Language’ given by T. W. Allen and E. E. Sikes, T'he Homeric

Hymns,

1904, Ixiii-Ixxiv, which is

substantially reproduced in the second edition (1936), xcvi-cix. 3? See Bibliography under Neumann, Steinacher, Savelsberg,

Seip,

Bauder, Illek, Bolling, Kinzel and Ogden.

31 See Bibliography under Hecht, Parmentier, Cook, Fick, Solmsen and Brugmann.

8? This need has now been met by the thorough Index Verborum

in

Merkelbach and West's Fragmenta Hesiodea (1967). 33 Studia Hesiodea I. De re metrica (1887). For other work of this period on Hesiod's versification see Bibliography under Rzach, La Roche, Solmsen and Ehrlicb.

*4 In addition to works of scholars mentioned in the preceding notes, see also Bibliography under Kausch, Martin and Scott.

PUBLISHED

WORK

ON

HESIOD'S

LANGUAGE

19

time to time by Rzach and may be found in Bursian's Jahresbericht.35 For most of the present century the main focus of attention may be said to have shifted from questions of grammar and vocabulary to more subtle ones of composition and style. P. F. Kretschmer's De sterats Hesiodeis (1913),29 lists all the groups of lines, single lines, and shorter phrases occurring more than once within the Hesiodic poems ; these naturally include many of the Homeric expressions previously noted by Rzach as Homeri ἰοοὶ similes, but there remains a large number of lines and phrases for which there is no place attested elsewhere which might have provided the source, and it is from these repetitions that Kretschmer seeks to establish the probable chronological relationship between the three poems, and in certain instances between different parts of the same poem. There can be no doubt that Kretschmer thought too rigidly in terms of exemplum and imitatio, ascribing the frequency of repeated phrases mainly to the mediocre poet's lack of resource. lad he been writing after Parry, he would have interpreted the evidence differently. But his lists of the repetitions themselves still contain material of interest and importance not to be found elsewhere. The work of Milman Parry, all published within the decade from 1928 to 1937,3® was mostly centred upon Homer, but he says much that 1s of wider relevance, and occasionally discusses Hesiod specifically, as in his article ‘Homer and Homeric

35 Detailed references will be found in EZ VIII, 1237, to which we may add

his later review,

covering the years

1909-1918,

in Bursian,

JB cxcix,

1924, 1-115. 36 Not the famous Paul Kretschmer (1866-1956), but a younger man of the same name, who was born in 1886.

3? e.g. * Interim vel ex imitatione tot locorum Homericorum huius poetae (vel horum poetarum) infirmitas apparet. Vel clarius eam cognoscere possumus, si praeterea observamus, quantopere in ipsis carminibus Hesiodeis—et in iisdem et in diversis—pares et similes loci iterentur ’ (DIII, 7) and * Hune poetam non ita magno ingenio praeditum fuisse apparet ' (DIH, 31). ᾿

38 See the bibliography by A. B. Lord in AJA lii, 1948, 43 f.

20

THE

LANGUAGE

OF

HESIOD

style'.3? Referring to the frequency of repeated phrases which had been shown by Kretschmer, and the presence of so many Homerie expressions as listed by Rzach, he writes: ' It is not the place here to explain the varying degrees of repetition within the Hesiodic poems, nor the use of Homeric phrases. That will be possible only in a longer study in which one will throw aside the idea of imitation, which has weighed so heavily on the early poetry outside Homer, and take up the repetitions as part of a traditional technique of verse-making. One will then learn, I believe, a great deal about the nature of the epic diction, of its use for different subjects, and by different poets or schools of poets, and of its decline.’ 495. Parry himself never produced such a longer study, turning his attention instead to modern examples of oral poetry ; but he was clearly aware of the implications of his work for Hesiod, and points out that ‘the formula is used in Hesiod far more often than it ever is outside of the early epic; and the same thing is to be said for the Shield of Heracles.’ *! Some lines of approach suggested by Parry’s work will be pursued below, especially in Chapters IV, V and VII. Other notable contributions to the study of Hesiod’s language were made at about this time by R. Cantarella in his article * Elementi primitivi nella poesia esiodea ',*? and by Inez Sellschopp, whose dissertation Stilistische Untersuchungen zu Hesiod appeared in 1934. Cantarella discusses proverbs, riddles and other expressions of the same kind in Hesiod which are likely to have their origin in popular tradition. Sellschopp’s book includes some illuminating chapters on epic formulas used by Hesiod, on certain characteristics in his use of adjectives, and on his development of abstract ideas. Although in Chapter XI we shall have reason to argue against Sellschopp’s view of the relationship between Hesiod aud the Odyssey, this forms only a part of her book, which elsewhere contains some 39 40 41 42

IISPh xli, 1930, 73-147. op. cit., 90. ibid. RIGI xv, 1931, 106-49.

PUBLISHED

WORK

ON

HESIOD'S

LANGUAGE

21

of the most useful work done on Hesiod in the period between the Wars. More recently the article * Hésiode et la tradition orale’ by A. Hoekstra * sets out to examine how much Hesiod's style has been influenced by formulaic diction, and in particular to consider the evidence for the presence of traditional formulas which do not occur in Homer. He concludes that Hesiod's stock of formulas 1s largely the same as Homer's but at ἃ more advanced stage of development, and he discusses also certain un-Homeric expressions which are used by Hesiod in a way which suggests that they may have a formulaic origin.** A far less commendable treatment of this subject is ‘ Homer, Hesiod and the Achaean heritage of oral poetry by J. A. Notopoulos, * where there is much that calls for qualification or contradiction.*® Even more recently there has appeared F. Krafft's substantial work, Vergleichende Untersuchungen zu Homer und. Hesiod (1963), in which various aspects of Hesiod's use of traditional poetic material are examined. Krafft compares Hesiodic and Homeric usage of some elements of vocabulary, including words for the human body, verbs of seeing, and abstract nouns, and makes a detailed study of certain lines and passages ; the book also contains useful lists of phrases which are common to Hesiod and Homer or repeated within Hesiod. Questions concerning more formal aspects of Hesiodic 533 Mnemosyne

x, 1957, 193-225.

44 Hoekstra’s book, Homeric Modifications of Formulaic Prototypes (1965), is also relevant to Hesiod in several places, though it is concerned primarily with earlier stages in the development of the hexameter tradition. 55 Hesperia xxix, 1960, 177-97.

46 o.g. he discusses the percentage of ‘ formulaic lines’ in the Hesiodic poems without defining what constitutes a formulaic line: clearly he does not mean a line which consists wholly of formulas, yet if lines are included in which the formulaic element extends to only one or two feet, the resultant

statistics are almost meaningless, since it is absurd to speak of T'^. 1-100 as 80%, formulaic, etc., on such a basis; in the same section (op. cit., 180) Notopoulos' percentages of formulaic lines in the Iliad and Odyssey are

reached by the extraordinary method of taking Parry’s figures for A 1-25 and a 1-25 and multiplying them by four! See further below, pp. 40-4, 201 f.

22

THE LANGUAGE

OF HESIOD

language have also been receiving fresh attention in the last few years. The traditional account of the way in which Hesiod's mixed dialect is constituted has been challenged in an important article by Anna Morpurgo Davies,*? though the earlier view, with some modifications, 1s still maintained by West in his discussion of, Hesiod’s language in the Prolegomena to his edition of the T'heogony.*? H. Troxler has published his study Sprache und Wortschatz Hesiods (1964), containing some sensitive observations on Hesiod’s word-choice and use of sound, as well as an examination of many phonological and morphological features, especially those where comparison with Homer is ofinterest. Hesiod’s vocabulary is included also in the monumental Lexikon des frühgriechischen Epos now being compiled by B. Snell and others, the usefulness of which increases as more parts of it appear; it is already a mine of information for any word with its initial between a and

ἀπό.

Many decades will pass before that work approaches

completion, but meanwhile a concordance to Hesiod is promised by B. Marzullo on the same lines as his revision of the Homerie Concordances of Prendergast and Dunbar, which in spite of certain shortcomings are remarkably useful aids in the study of Homer's vocabulary and formulas. Many other contributions might be mentioned, but an exhaustive résumé is not desirable here, and enough has now been said to acquaint the reader with the context in which this book has been written. Any new work concerned with the language of early Greek poetry will inevitably owe much to its predecessors, and it is proper to acknowledge this general debt to earlier scholars as well as that owed to specific pieces of work to which reference is made in the course of these chapters. 47 *' Doric” features in the language 138-65. 486 HT, 79-91.

of Hesiod’,

Glotta xlii, 1964,

CHAPTER III TRADITIONAL

POETIC

LANGUAGE

IN

HESIOD

A PROBLEM which is fundamental in any consideration of Hesiod’s language is that of distinguishing those features which are specifically Hesiodic from thosc which may be simply traditional. The difficulty here is to establish what was traditional for Hesiod : only by knowing the full rango of hexameter poetry with which Hesiod himself was familiar would it be possible to tell which features he inherited from others and which he himself introduced for the first time. Such knowledge is not attainable, and the only approach open to us is to take what we have of the hexameter tradition outside Hesiod, and then see how much Hesiod shows in common with

this, or with particular sections of it. The aim of this chapter therefore 1s to assess, if only in a very rough way, the degree of likeness which exists between his language—or the language of different parts of his poetry—and that of Homer, and to consider what this implies about the likely extent of Hesiod's linguistic debt to earlier poets. Examination of a specimen passage of the Theogony will illustrate the difficulties which beset any attempt to state the relationship between Hesiodic and Homeric language in statistical terms. In the ten lines beginning at Th. 101 Rzach notes six Homeric phrases in his list of Homeri loc? similes, two of these occurring in a single line (Th. 109), so that all are contained within five out of the ten lines. They include the five line-endings :

Th. 101 θεοὺς ot "Ολυμπον éxyovow | found at E 404, and similarly θεῶν oi "Ολυμπον ἔχουσιν | E 890, N 68, 22 427, t 240, 0 331, 743; θεοῖς ot "Ολυμπον ἔχουσιν | u 337 ; θεοὶ τοὶ "Ολυμπον ἔχουσιν | é 394, c 180 (= Krafft, VUHH, 172

Nr. 246). Th. 104 ἱμερόεσσαν ἀοιδήν | found also στονόεσσαν ἀοιδήν | 2 721 291).

at a 421, σ 304; cf. (= VUHH, 174 Nr.

OF HESIOD

TIIE LANGUAGE

24

106 καὶ οὐρανοῦ

Th.

ἀστερόεντος

| found

© 46,

at E 769,

and without καί at Z 108, T 130, v 113 (=

VUHH,

172

Nr. 260).

Th.

109 οἴδματι θυίων (or θύων) | found at © 234,

(-- VUHH, 119 Nr. 462). Th. 110 καὶ οὐρανὸς εὐρὺς (= VUHH, 172 Nr. 259).

ὕπερθεν

V 230

| found at Ο 36, e 184

The remaining Homeric parallel given ἀπείριτος in the third and fourth feet of occurs in the same position at κ 195 (— But the common ground does not end to Rzach's list :

by Rzach is πόντος Th. 109; this phrase VUHH, 185 Nr. 655). here, and we may add

Th. 101 μάκαράς τε θεούς : the acc. μάκαρας does not occur in Homer, but the nom., gen. and dat. pl. all occur there as epithets of θεοί, θεῶν

θεοῖς, -οἷσι ;

and

θεοῖς at this place in the line Th. 103 δῶρα θεάων |: cf. including δῖα θεάων |

cf. esp. μακάρεσσι

£819 (— VUHH, 168 Nr. 132). θεάων | 9x IL, 28x Od,

7x Il., 26x Od.; cf. also δῶρα θεοῖο |

Y 268, 6 165 (= VUHH,

177 Nr. 406).

Th. 105 αἰὲν ἐόντων | found at y 147, 8 583; cf. also αἰὲν Od. αἰὲν ἐόντας | 4x Od.; Il, 4x ἐόντες Ϊ 4x

(= VUHH, 176 Nr. 369). Th. 106 | ot Γῆς τ᾽ ἐξεγένοντο : cf. | ot Διὸς ἐξεγένοντο E 637 and | of ey

Th.

107

ἐξεγένοντο Y 305 (not noted by Krafit).

| Νυκτός

τε δνοφερῆς : cf.

νὺξ δὲ μάλα dvodepy

v 269 and | νύκτα διὰ δνοφερήν o 50 (= VUHH, 182 Nr. 577). Th. 109 | καὶ ποταμοὶ καὶ πόντος : cf. | kai ποταμοὶ καὶ γαῖα

T 278 (= VUHH, 182 Nr. 560).

This last line (Th. 109) is thus seen to be wholly made up of three phrases, each two feet long, which are found separately in Homer. But even in a line where no parallel phrase from Homer can be adduced, similarities which suggest the influence of the hexameter tradition are to be seen in single word-forms, which often occur at the same point in the line as in Homer. For example, Th. 108, when compared with Homer at this level of word-forms, cannot be said to appear any less

TRADITIONAL

POETIC

LANGUAGE

IN

HESIOD

25

traditional in its language than the lines at which we have just been looking : Th. 108 εἴπατε δ᾽ ὡς τὰ πρῶτα θεοὶ καὶ γαῖα γένοντο. Every word-form here is familiar to us from Homer: εἴπατε is the only one which does not occur in the Jad, but it is found twice in the Odyssey, one of these being at the beginning

of the line in the phrase εἴπαθ᾽ ὅπως ᾧ 198.

The remaining

forms are all common in both /kad and Odyssey, especially in the positions in which Hesiod uses them here: πρῶτα is found 69 times in Homer (38x Il., 31 x Od.), and 47 of these occurrences (30x Il., 17x Od.) are in this position; this means that 68% of the occurrences of the form are found in

this one place, even though there are five other places in the line where a word of this scansion could be used just as easily,! not counting those where the word could be fitted by means of elision or lengthening before a double consonant. Similarly, θεοί is found 200 times in Homer (96x IL, 104x Od.) and 112 of these (51x IL, 61x Od.), ie. 56% are in this place, whereas there are four other places in the line where the same scansional pattern recurs, as well as five places where θεοί may be used before ἃ vowel and so be scanned as two short syllables. The form γαῖα is rather less.common in Homer, occurring 32 times (20x IL, 12x Od.), but again 50% of its occurrences are in this one place (10x Il., 6x Od.), though the same choice is available as for πρῶτα discussed above.? Finally γένοντο occurs 21 times in Homer (12x IL, 9x Od.), and all but one of these occurrences (11x Il, 9x Od.), i.e. 95%,

are at the end of the line, which means that this

1 Itis of course well known that in Homer and in Greek hexameter verse of all periods words of particular scansional patterns tend to be concentrated at partieular positions in the line: see E. G. O'Neill Jr., ‘ The localization of metrical word-types in the Greek hexameter ', YCS viii, 1942, 105-78 ; but

we are here eoncerned primarily with the localization not of word-types but of specific forms ; cf. the next two notes. ? Hesiod's habitual placing of these forms is remarkably similar to Homer's. Thus the heavy first syllable of πρῶτα or γαῖα might theoretically be placed in the first half of any of the six feet of the hexameter,

or in the

second half of any of the first five feet if the -a is elided or lengthened. C

In

26

THE

LANGUAGE

OF HESIOD

form is used by Homer only once in any of its four other possible positions.? Here are some further points of correspondence : πρῶτα 15 found with τά, as in Th. 108, nine times in the [liad and three

times in the Odyssey, seven of these (all in Il.) being in the identical position in the line (= VUHH, 168 Nr. 129). The phrase ws τὰ πρῶτα occurs at θ 268, not in Hesiod's position, although we do find ὥς ot πρῶτα here at A 276. We may also compare τὰ πρῶτα γένηται | at Z 489 and 0 553 with Hesiod's

rà πρῶτα... γένοντο | here.

Out of the 47 places in Homer

where πρῶτα 1s used in this position, in no less than eight 1t 1s followed immediately by θεός or θεῶν. Finally, with the second half of Hesiod's line we may compare : ὕδωρ καὶ γαῖα practice the placing of them is restricted by both poets, but in different ways for the two words. Examples occur as follows : Foot (i) πρῶτα Il. Od. Th. Op So (ii) γαῖα

1 a 5 8 -Ἕ -

2

3

b 2 -

a 1 -

b -

4

a 30 17 δ. 4 -

b =

a = = -

5 b 1 = =

à 2 3 2 -

b —-Ἑ -

6 a& -

1l.

3

-

-

-

2

-

-

-

10

-

5

Od.

1]

-

1]

l

2

1

-

-

6

-

-

Th Op Se

2 -

=

-

—-.

δ 3 =

-Ἕ =

=

= =

8 4 -

-

3 ]

These figures show that in each of the two poets, aud in each poem (leaving aside the Shield), the third foot is the ‘ favourite ’ position for πρῶτα and the fifth foot for γαῖα.

8. G. S. Kirk observes, with regard to the Homeric hexameter, that ‘ the very marked restriction of γενέσθαι, γένηται, γένοντο, γένοιτο to the end of the

verse, which greatly exceeds that of such word-types in general, is not due primarily to the recurrence of formular phrases (which is often a potent factor) ; word-order and the function of verbs provide some explanation, but, in addition, apparently singers did develop some kind of formular attitude to these particular forms, and tended to place them instinctively at the end of the verse ' (YCS xx, 1966, 104). He gives the following totals for the four words: at the verse-end, 53 x Il., 52x Od.; other positions, 6 x Il., 10x Od. The same four words occur in Hesiod as follows: at the verse-end, 5x Th., 6x Op.; other positions, never in Th., once in Op.;

there are no examples from the Shield.

TRADITIONAL

POETIC

LANGUAGE

γένοισθε | H 99, φίλον καὶ ἡδὺ γένοιτο

IN

HESIOD

27

| H 387, ἔπειτα φόνος

καὶ μοῖρα γένοντο | à 24, and fj τὰ πρῶτα πύλας καὶ τεῖχος

ἐσᾶλτο | N 679.

Despite therefore the total absence from

this line of any parallel Homeric phrase, there is nothing in its language, at least at this level of word-forms and the positions which they occupy, which can safely be called ‘ original’ or * Hesiodic ’. It must be emphasised that the passage which we have just examined (Th. 101-110) is fully typical of the Theogony as a whole. As we saw above (p. 23), Rzach noted Homeric parallels in five out of the ten lines (50%); out of the 50 lines from Th. 101-150 he notes parallels in 28 (i.e. 56%); and out of the whole poem of 1022 lines he notes parallels in some 566 lines (1.6. about 55%). This compares with parallels in 330 out of 828 lines for the Works and Days (i.e. about 40%), and in 298 out of 480 for the Shield

(1.6. about 60%).

The figures

must not be too closely pressed, since the lines included in these percentages only rarely are entirely composed of the parallel phrase or phrases which Rzach notes. On the other hand there are many places not noted in Rzach’s list, as we have seen in Th. 101-110, where Hesiod shows signs of a similar dependence upon traditional forms of expression.* The diversity and complexity of these parallels between Hesiod and Homer make it difficult 4 Krafft’s useful list of phrases common to Hesiod and Homer (VUHH, 163-91) includes a good many short expressions not noticed among Rzach’s

Homeri loci similes, e.g. μεγάλου Aids Th. 29, [δεύτερον αὖτε Th. 47;

but

conversely Rzach mentions several parallels omitted by Krafft, whose method of classification causes him especially to miss phrases which are

shifted from one part of the line to another, e.g. | 'Hà τ᾽ 'HéXóv re Th. 19 beside ἠῶ τ᾽ ἠέλιόν re | E 267, M 239, ı 26, v 240, or which are inverted, e.g. νιφόεντος ᾽Ολύμπου | Th. 42, 62, 118, 794 beside Οὐλύμπου νιφόεντος | Σ' 616. Thus in Th. 1-100 Rzach notes Homeric parallels from 61 out of the 100 lines, and Krafft from 62; but Krafft omits what Rzach notes in 7h. 10, 19, 31, 42, 60, 62 and 84; and Rzach omits parallels listed by Krafft from

Th. 30, 51, 53, 61, 65, 72 and 76. Both of them note parallels from Th. 29 and 47, but from different parts of those lines. Yet even a conflation of their lists would be far from exhaustive: many parallel phrases which involve a slightly more complex modification are omitted by both Rzach and Krafft, e.g., from the first few lines of the Theogony, πόσο᾽ ἁπαλοῖσιν Th. 3 beside ἁπαλοὶ πόδες T 92; χοροὺς... . ἱμερόεντας Th. 7 f. beside ἱμερόεντα

28

THE LANGUAGE OF HESIOD

to summarise the position in statistical terms satisfactorily: some phrases extend to a whole line, others are much shorter ; some appear in identical form at the identical place in the line, others at a different place in the line, or with some internal difference, such as a change of case or person or tense, or the

substitution of another word in part of the phrase, which may thereby be adapted to quite a fresh context ; sometimes it is little more than the sound or grammatical structure of ἃ traditional line which seems to have suggested an expression to the poet.* We may however confine ourselves for a short time to a level of comparison which is amenable to statistical treatment. Tapes 1, 2 and 3 (pp. 30f) show simply the proportion of word-forms used by Hesiod which are identical with forms found in the Iliad.* The Iltad does not by any means represent χορόν Σ 603, χορὸν ἱμερόεντα o 194; Δία T’aiyioxov Th. 11 beside Ζεύς 7' αἰγίοχος Θ 287, o 245, Διὸς αἰγιόχοιο | 13 χ Π.,9Χ Od., etc. ; Ἥρην [᾿Αργείην Τῇ. 11 f. beside “Hpn τ᾽ ’Apyein 4 8, E 908. While these are less impressive than the more rigid parallels, one cannot doubt that Hesiod in using them has been influenced to some degree by traditional phraseology, and it may be noted that all of them fall within the revised definition of a ‘formula’ proposed by Hainsworth: see below, p. 43 n. 7. 5 [n addition to the example of this from the second half of Th. 108, mentioned above, see the discussion below in Chapter VI, esp. p. 76 on Th. 105. $ In deciding what constitutes an ‘identical word-form’ it has been found necessary to adopt certain conventions as follows : (i) À word-form is defined as a separate lexical unit, corresponding in practice to the words as divided in the standard texts. Examples of crasis, e.g. κεἰς Op. 44, have been counted as two separate word-forms, i.e. καὶ + eis;

verbs in tmesi similarly

have been counted as two separate word-forms. (ii) A word-form has been counted as occurring in the Iliad if it is found there in precisely the same form as in Hesiod, i.e. with the same case-ending, verbal inflexion etc., except

as is allowed in the following : (a) ephelcystic -v, as being relevant to the phonetic context of ἃ word rather than to its form, bas not been regarded as an integral part of the word-form ; (b) elision has been ignored for the same reason ; (c) variations in breathing have been ignored. (iii) The alternative declensional endings, such as gen. sg. -oıo beside -ov, dat. pl. -σι beside -ns and -ow beside -oıs, have been counted as distinct from one another. Similarly the presence of an augment has been regarded as making a form distinct from the unaugmented form : thus ἐμήσατο Op. 49 has been counted as a non-Iliadio form, although unaugmented μήσατο occurs in the Iliad.

TRADITIONAL

POETIC LANGUAGE

IN HESIOD

29

the total bulk of traditional poetry which might have been known to Hesiod, nor have we the right to assume that Hesiod knew

the entire Iliad as we now

have it, but at least this

approach provides ἃ rough yardstick for testing the consistency of the whole Hesiodic corpus.’ Each line was taken separately in the first instance, but the figures are here tabulated by 50-line groups, in order to show the broad differences which exist between parts of the same poem, as well as from one poem to another. Again the figures must not be pressed in too close detail, but certain conclusions may be drawn : (a) The Shield (having 81% of its forms i common with the Iliad, and an average of 1-3 non-Iliadic forms per line) comes out as the most Homeric of the three poems, and the Works and Days as the least (75-695, 1-7), with the Theogony in between them (78:995, 1-4) but standing rather closer to the Shield than to the Works and Days. This was the same general picture as emerged above (p. 27) from the count of Rzach's lists of Homeri loci similes (Sc. in c. 60% of lines, Th. c. 55%, Op. c. 40%).8 (b) The Theogony is the most consistent in this matter in its various parts, the limits of variation being 73% and 86% of Iliadic forms, which gives a range of 13%, while the Works and Days (68% to 85%) and the Shield (73% to 90%) both have a range of 17%. The average number of non-Iliadic forms per line varies in the Theogony from 1-0 (lines 601-650) to 1.8 (lines 351-400), a range of 0-8, in the Works and Days from 1-1 (lines 101-150) to 2-2 (lines 401-450), a range of 1.1, and in the Shield from 0-7 (lines 401-450) to 2-0 (lines 251-300), a range of 1-3. Again this is in broad agreement with the result which may be obtained by a count of Rzach’s Homeri loci 7 The Iliad, being generally regarded as older than the Odyssey, and containing so much obviously traditional material, seems to provide for our present purpose a more acceptable basis of comparison than the two Homeric poems together ; cf. above, p. 9.

8 Krafft cites Homeric parallels to some 615 lines of Th. (c. 60%) and

349 lines of Op. (c. 42%) ; cf. above, p. 27 n. 4. His work does not embrace the Shield.

30

THE

TABLE 50-line passage beginning at

1

51

101 151 201 251 301 351 401 451

1.

Word-forms found in Iliad (these as percentage of total)

Word-forms not found in Iliad (average number per line)

330

245

(74%)

85

(1:7)

363 344 353 331 336 352 357 348

293 274 278 242 248 262 301 290

(81%) (80%) (1890) (1390) (74%) (74%) (84%) (83%)

70 70 75 89 88 90 56 58

(1:4) (1:4) (1:5) (1-8) (1:8) (1:8) (1:1) (1:2)

(86%) (82%) (81%) (83%) (80%) (80%) (76%) (74%)

48 64 65 57 70 68 77 79

(1:0) (1:3) (1:3) (1:1) (1:4) (1-4) (1:5) (1:6)

335

325 335

1001-1022

128

Total

OF HESIOD

Non-Iliadic word-forms in the Theogony

Total number of word-forms occurring

501 551

601 651 701 751 801 851 901 951

LANGUAGE

254

248 270

349 347 343 339 349 333 322 303

301 283 278 282 279 265 245 224

6,922

5,459

97

(1694)

(76%) (81%)

(76%)

— (78.995)

81

7 65

31

1,463

(1:6)

(1-6) (1:3)

(1-4) (1-4)

similes: if we adhere to the same 50-line groups, the fewest Homeric parallels in the Theogony are found in lines 351-400, where they are noted in 15 lies (30%), the most in lines 951-1000, in 34 lines (68%), giving a range of variation of 38%. In the Works and Days the lower limit is in lines 401—450, with parallels noted in 11 lines (2294), and the upper limit in lines 51-100, with parallels in 35 lines (70%), which gives a range of variation of 4895.9 The figures for the Shield vary between 19 lines (38%) for Sc. 251-300 and 41 lines (82%) for Se. 401-450, giving a range of 44%, (c) The most un-Homeric part of any of the three poems is

? The corresponding ranges of variation worked out on the same basis from

Krafft’s list are 34% for Th. and 50% for Op. Ho records Homeric parallols in only 20 lines out of Th. 351-400, and only 12 lines out of Op. 401-450, and 88 with Rzach these are the lowest extremes in the respective poems.

TRADITIONAL

TABLE

2.

50-line passage beginning at

Total

' '

LANGUAGE

IN

HESIOD

3l

XNon-Iliadic word-forms in the Works and Days

Total number of word-forms occurring

1 51 101 151 201 251 301 351 401 451 501 551 601 651 701 751 801-828

POETIC

Word-forms found in Ihad (these as percentage of total)

366 342 353 341 364 371 375 360 356 342 343 345 358 352 347 357 | 193

282 277. 300 279 294 302 289 268 244 231 236 234 258 274 267 253 144

56,865

4,432

Word-forms not found in Iliad (average number per line)

(77%) (81%) (8595) (82%) (81%) (81%) (77%) (74%) (69%) (6895) (69%) (68%) (72%) (78%) (77%) (71%) (75%) (75-695)

|

84 65 53 62 70 69 86 92 112 111 107 111 100 78 80 104 49

(1-7) (1:3) (1:1) (1:2) (1:4) (1-4) (1-7) (1:8) (2-2) (2:2) (2:1) (2-2) (2-0) (1-6) (1:6) (2.1) (1:8)

1,433

(1:7)

in the middle of the Works and. Days, where the proportion of Tliadic forms falls below 70% in lines 401-600. It will be found that Rzach notes fewer lines containing Homeric parallels in this part of the poem than in any other, the fewest of all in

Total number of word-forms occurring

343

333

Word-forms found in Iliad (these as percentage of total)

277

272

101 151

336 349

284 283

251 301 351

367 348 355

268 204 288

200

170

201

401

451-480

Total

336

343

3,310

(8150)

66

(85%) (8195)

52 66

(73%) (84%) (81%)

99 54 67

(2:0) (1:1) (1-3)

(8592)

30

(1:0)

(896)

249

— (7494)

307

(90%)

2,092

Word-forms not found in Iliad (average number per line)

(81:05)

61

87

36

618

(1-3)

(1-2)

(1:0) (1:3)

(1:7)

(0:7) es

1

δ]

XNon-Iliadic word-forms in the Shield



50-line passage beginning at

3.

~~ E .

TABLE

32

THE

LANGUAGE

OF HESIOD

Op. 401-450, where he finds parallels in only eleven lines (22%), beside an average of 40% for the whole poem. The middle portion of the Shield (lines 201-300) has less of its language in common with the Iliad than the rest, and here too Rzach notes fewer Homeric parallels (in only 41 out ofthe 100 lines, beside an average of 60% for the whole poem). That these particular parts of the poems should show a relatively high degree of difference from Homer is not surprising when one bears in mind their subject-matter. Where Hesiod is talking about farm-carts and ploughs it is not to be expected that his language is going to conform to Homer’s as closely as in a narrative passage. In the Shield similarly this comparatively un-Homeric portion includes the extravagant description of Perseus and the Gorgon’s head, and an even more lurid account of the Fates and the creature Achlys, as well as the peaceful scenes described in lines 270 ff. Yet even in such passages as these it would be wrong to lay too much stress on the differences between Hesiod’s language and Homer’s. In the agricultural part of the Work and Days where, as we have seen, the proportion of non-Iliadic forms is higher than anywhere else in Hesiod, the average number of such forms, when calculated from our 50-line groups, nowhere exceeded 2-24 per line (112 examples in Op. 401-450). When analysed more closely, by groups of ten lines, the highest number of non-Iliadic forms to occur is 27, giving an average of 2-7 per line, in the ten lines beginning at Op. 421. Here Hesiod is concerned with precepts for the autumn season, and at line 420 he begins to give technical advice about types of wood appropriate for making different tools. It would be hard indeed to suggest any Hesiodic passage which is further removed from Homer in its subject-matter, and Rzach notes no Homeric parallels at all between lines 417 and 432. Yet the striking fact is that even in this passage there is abundant evidence to suggest Hesiod's dependence on traditional forms of expression. When each line 15 examined in detail, points of likeness with Homer appear at every turn, sometimes faint but certainly remarkable for their unexpected profusion. (We shall

TRADITIONAL

POETIC

LANGUAGE

IN HESIOD

33

include Op. 420 and 431 here in order to complete the sentences at the beginning and the end of the passage.) Op.

420

τῆμος

| τῆμος

Od.:

ἀδηκτοτάτη

4X Il.,

πέλεται

τμηθεῖσα

σιδήρῳ

3x Od. ; πέλεται in this position, 5x Il.,

2x

cf. esp. | -~~ ὀξύτατον πέλεται E 345, | --~ οἴκτιστον

πέλεται X 76, | -~ ἀριγνώτη πέλεται ζ 108 ; σιδήρῳ 5x Il.,

always at the end of a line : cf. esp. σιδήρῳ [ἐξέταμ᾽ A 485 f. (of a craftsman felling a tree), and the variant ἀποτμήξειε

σιδήρῳ | 2 34 (adopted in Mazon's Budé text ; Monro and Allen read ἀπαμήσειε with Aristarchus). Op. 421 ὕλη, φύλλα

δ᾽ ἔραζε χέει πτόρθοιό τε λήγει

ὕλη N 141 (with enjambement, similarly), V/ 198, e 68 ; cf. φύλλα τὰ μέν τ᾽ ἄνεμος χαμάδις χέει, ἄλλα δέ θ᾽ m | τηλεθόωσα φύει Z 147 f.;

ἐκ... ὕλης πτόρθον κλάσε .

.|

φύλλων ζ 128 f. (the only occurrence of the word πτόρθον in Homer); χέει oceurs in this position at τ 521: cf. also χεῦεν ἔραζε | P 619, o 527, x 20, 85, and κατέχευεν ἔραζε | IT 459; the form λήγει does not occur in Homer, but cf. λήγεις K 164 and ἀπολήγει Z 149, P 565, Y 99, Φ 577, all at the end of a line.

Op. 422 τῆμος ἄρ᾽ ὑλοτομεῖν μεμνημένος ὥρια ἔργα

| τῆμος ἄρ᾽ Η 434, 2 789, 8 401 (= VUHH, 164 Nr. 10); ὑλοτομεῖν is not Homeric, but ὑλοτόμους Ἱ 114, both in this μεμνημένος occurs in this position 4 x Od. (4 oceurrences) ; the adj.

cf. ὑλοτόμοι Y 123 and position (cf. Op. 807); 3 x Il. (4 occurrences) and ὥριος occurs only once in

Homer, in this form and position : ὥρια πάντα | «131; ἔργα] cf. πολεμήϊα ἔργα | 6x Il., plus 7 further instances of ἔργα at the end of a line after an adj. in -a:

17 similar examples

in Od.

Op. 423 ὄλμον μὲν τριπόδην τάμνειν, ὕπερον δὲ τρίπηχυν

| ὅλμον at A 147 (the only occurrence in Homer) is probably the same word:

see Liddell and Scott, s.v. ; τριπόδης does not

94

THE

LANGUAGE

OF HESIOD

occur in Homer, but cf. ἑπταπόδην in this position at Ο 729

(see notes on Op. 424 below):

also | δοιοὺς δὲ τρίποδας

5 129, | γάστρην μὲν τρίποδος 0 431 ; τάμνειν cf. τάμνε(ν) in this position I' 273, 7 209; τρίπηχυς is not a Homeric word,

but

cf.

ζυγόδεσμον.... ἐννεάπηχυ | (

270,

and

ἔγχος ἔχ᾽ ἑνδεκάπηχυ Z 319, O 494. Op. 424 ἄξονα δ᾽ ἑπταπόδην᾽ μάλα γάρ νύ τοι ἄρμενον οὕτω

cf. [θρῆνυν ἐφ᾽ ἑπταπόδην O 129 ; μάλα γάρ in this position Σ 322, y 20, 261, 328, ζ 203, « 171, 180, A 621, w 92; νύτοι: Paulson takes ro: to be the affirmative particle here, but it is equally well understood as the second person pronoun

(= σοι),15 and may be compared with νύ ro: similarly used at A 28, O 440, X 11, and esp. ἐπεί νύ τοι εὔαδεν οὕτως | P 647: cf. also νύ μοι 4 93, νυ σοί Θ 201, and esp. τόθι

γάρ νύ of αἴσιμον ἦεν | o 239;

ἄρμενον in this position

e 254 ; oürw(s) | in this position 4x Il, 6x Od.: cf. esp. κάλλιον odtw(s)

| y 358, 0 543, c 255, τ 128.

Op. 425 εἰ δέ xev ὀκταπόδην, ἀπὸ xal σφῦράν xc τάμοιο | et δέ κε(ν) A 197, B 364 etc., a 289, e 417 ete. ; ὀκταπόδην

is not a Homeric word, but cf. ἑπταπόδην in this position O 729 (see note on Op. 424 above) ; τάμοιο does not occur in Homer,

but cf. ταμόντες

| I' 73, 256 and τάμωμεν

|

94,

T 191. Op. 426 τρισπίθαμον δ᾽ ἄψιν τάμνειν δεκαδώρῳ

ἀμάξῃ

There is no trace of the second element of the compound τρισπίθαμος in Homer, but rpiyAnva (ΚΞ 183, τρίπτυχος A 353,

τριστοιχί Καὶ 473 all occur at the beginning of the line ; for τάμνειν in this position see notes on Op. 423 above; dat.

ἀμάξῃ does not occur in Homer, but cf. ἅμαξαν ἄμαξαι | «241, « 103;

| ζ 260,

δεκάδωρος 1s not a Homeric word, but

cf. képa. . . . ἑκκαιδεκάδωρα πεφύκει | A 109. 19 N.B. Sc. 116, μάλα γάρ νύ οἱ ἄρμενα εἶπεν ; this could depend on Op. 424,

but is more likely to be a traditional phrase, not found in Homer, which Hesiod has adapted to this use in Op.

TRADITIONAL

POETIC

LANGUAGE

Op. 427 πόλλ᾽ ἐπικαμπύλα κᾶλα

IN

HESIOD

35

φέρειν δὲ γύην, ὅτ᾽ ἂν εὕρῃς

For the beginning of the line cf. [πόλλ᾽ ἐπὶ μῆρ᾽ ἔθεμεν y 179, | πόλλ᾽ ἐπιδινεῖται v 218 ; ἐπικαμπύλα and. κᾶλα are not Homeric, but cf. καμπύλα Tó£a | 5x Il., καμπύλα κύκλα | , καμπύλον ἅρμα | each 1x J1., and in this position | -~~ καμπύλα τόξα ı 156, and esp. | --- καμπύλα τόξα

φέρεις & 362 ; φέρειν in this position 5x Il, 2x Od. ; ὅτ᾽ àv is frequent, but there is no example in this position (Ebeling, LH 11,96 f.) ; cf. ὄφρα κεν eüpn | X 192 and eis 6 κεν ἔλθῃς | K 62. Op. 428 ἐς olxov, κατ᾽ ὄρος διζήμενος T) κατ᾽ ἄρουραν

(eis Wilamowitz, with mss.) eis οἶκον (not

at line

beginning)

Z

490,

a 356,

φ

350;

διζήμενος 4X Il., 6x Od., all in this position: cf. esp. ἠέτιν᾽ οὐρήων διζήμενος, 7] τιν᾽ ἑταίρωνΚαὶ 84 ; κατ᾽ ἄρουραν | A 68 (=

VUHH,

190 Nr. 882), and

ἄρουραν

| 4x

Il., 8x

Od.

Op.

429 nplvvov: ὃς γὰρ βουσὶν ἀροῦν ὀχυρώτατός ἐστιν

πρίνινος does not occur in Homer, but cf. ζύγον . . . | πύξινον ‘of box-wood’ 2 268 f.; ὃς yap a 286, p 172: cf. also ὃ γάρ κ᾽ ὄχ᾽ ἄριστον ἁπάντων | εἴη M 844 f. = M 357 f; for this idiomatic use of γάρ after ὅς in Homer see Cunliffe, LHD, s.v. γάρ (4); βουσίν occurs in this position 2x IL., once Od. ; with the end of the line cf. σὺ yàp βασιλεύτατός ἐσσι | I 69, 6 yap πολὺ φέρτατός ἐστιν | A 581, 6 γὰρ πολὺ depraros ἦεν | B 769. Op. 430 εὖτ᾽ ἂν ᾿Αθηναίης δμῷος Ev ἐλύματι πήξας

εὖτ᾽ dv 4x Il, 4x Od., but never at line beginning ; Ἀθηναίης in this position 7x Il, 2X Od.; δμῷος is not a

Homeric word, but the fem. 8j«at, δμῳάς occurs here once 11 Troxler,

SW H,

184, revives the

alternative

separated from καμπύλα : cf. Paley ad loc. gupport to the division.

The

reading

with

Homerie

ἐπὶ (or ἔπι)

parallels lend

36

THE LANGUAGE

OF HESIOD

Il, 8x Od.; with the end of the line cf. ὄμματα πήξας | D 217

ψ 235.

and

the similar-sounding

κύματι

πηγῷ

| e 388,

Op. 431 γόμφοισιν πελάσας προσαρήρεται ἱστοβοῆι

| γόμφοισιν € 248 (the only occurrence of the word in Homer); cf. | ἄκρῃ προσπελάσας ı 285 ; προσαρηρότα occurs in this position E 725, also ἀρηρότα ἡ 45 and àpmpóres B 342 ; ἱστοβοεύς does not occur m Homer, but Boevs ‘strap’ oceurs in the dat. pl. βοεῦσιν at the end of the line 8 426 — o 29],

It would be foolish to insist that every one of the points of likeness listed here must be explained by the dependence of one poet on the other, or even by that of both on a common tradition. Any Greek speaker composing hexameter verse

might use such words as | ὕλη Op. 421 and ἄρουραν | Op. 428 in these positions without ever having heard them so used by an earler poet. Yet all the parallels cannot plausibly be regarded as fortuitous: it is noteworthy that the rare words ὄλμον Op. 423, ἑπταπόδην Op. 424, and γόμφοισιν Op. 431, each of which occurs only once in Hesiod and once in Homer, appear in both 2m exactly the same case-form and at the same point on the line, and a similar fondness for a fixed position is seen in the more loosely related but equally rare pairs, ὑλοτομεῖν Op. 422 beside ὑλοτόμος, δεκαδώρῳ Op. 426 beside ἑκκαιδεκάδωρα, and πρίνινον Op. 429 beside πύξινον. Special attention may be drawn also to the less striking juxtapositions like | τῆμος dp’ Op. 422, μάλα γάρ Op. 424, νύ τοι Op. 424,

| ei δέ κεν Op. 425, and | εὖτ᾽ ἄν Op. 430. At first sight, and taken separately, these seem unimportant ; !? yet it is surely not without significance that we find Hesiod using so many expressions of this kind as the cement which binds together the fabric of his verse, even in this passage which is so far removed from the epic in its subject-matter. When one examines Krafft’s list of parallels between Hesiod and Homer 12 cf. Hoekstra’s remarks criticising Lord and Notopoulos,

HMFP, 15.

TRADITIONAL

POETIC

LANGUAGE

IN HESIOD

37

(VUHH 163-91) one is impressed not only by the numerous longer expressions consisting of noun with epithet, verb with object, etc., but also by these little connecting phrases which can be used in an infinite variety of contexts.!? The extent and complexity of all the correspondences make it impossible to suppose that they could have resulted wholly from chance, that is from independent invention, though a few of them might have done so. Similarly, although we cannot rule out the possibility that some of the Homeric expressions which we now find in Hesiod might have been introduced in substitution for original Hesiodic ones which were unfamiliar and had perhaps become unintelhgible, the similarity with Homer's language 15 too extensive and too deeply rooted in the structure of the Hesiodic poems to have resulted in any large measure from post-Hesiodic alteration to the text. ΑΒ far as the bulk of the phrases common to Hesiod and Homer are concerned, we must either recognise them to be pre-Hesiodic in origin, and so to reflect Hesiod’s dependence on traditional hexameter material, or alternatively resort to the hypothesis that an expression invented by Hesiod himself has been used later by another hexameter poet or poets, so that its appearance in Homer or elsewhere is secondary to that in Hesiod. A number of scholars !* have been prepared to adopt this alternative view in order to account for certain correspondences between the two poets, but even if their arguments were to be allowed, it would still not be possible to explain more than a small fraction of the parallels as having originated with Hesiod. Many of the phrases used by both 18 Here are some obvious examples, for details of which see Krafft's list under the number here given : (a) line-beginnings : 17 αἶψά κε kai, 19 ἀλλ᾽

αἰεί, 20 ἀλλ’ ἔμπης, 21 ἄλλοτε δ᾽ αὖ (αὖτε), 22 ἀλλ᾽ ὅτε δή (δή ῥ᾽), 23 ἀμφὶ δέ ot, 24 αὐτὰρ ἐπεί, 27, 48 δεύτερον αὖ (αὖτε), 35 ἦ τοι μέν, 45 ds δ᾽ αὕτως, 49 ἤματι τῷ ὅτε, 539 δηθὰ μάλ᾽, 543 τοὔνεκ᾽ ἄρ᾽, 547 δηρὸν γάρ, 599 καὶ τότε δή;

(b) line-endings : 372 af x’ ἐθέλητε, 387 αὐτὰρ ἔπειτα, 427 ἤματι κείνῳ, 470 ὅττι τάχιστα, 516 ὡς τὸ πάρος περ, 835 αἴ κε δύνηαι, 847 eis ὅ κεν ἔλθῃ, 878 ὡς γὰρ ἄμεινον; (c) other definite positions in the line: 110 ἐξ ἀρχῆς, 116 ὅττι τάχιστα, 123 μάλα πολλά, 129 τὰ πρῶτα, 131 τὸ πρῶτον, 625 ἔτ᾽ ἢ τὸ πάροιθεν, 631 ὡς τὸ πρῶτον, 638 πολὺ μᾶλλον.

14 See below, Chapter XI, esp. p. 166,

28

THE LANGUAGE

OF HESIOD

Hesiod and Homer show relatively early 15 linguistic features, including nom. sg. masc. in -a, uncontracted gen. sg. in -ào

and -oıo and gen. pl. in -ἄων e.g. μητίετα Ζεύς 4X Th., 1x Op., 2x Sc, 15x IL, 3x Od. (— VUHH, 119 Nr. 454), Διὸς νεφεληγερέταο 2X Th., 1x Op., 6x Il. (= VUHH, 171

Nr. 224), πολυφλοίσβοιο θαλάσσης 1X Op., 6x Il, 2x Od. (= VUHH, 173 Nr. 273), δῖα θεάων 4x Th., 1x Se., TX Il, 26x Od. (= VUHH, 177 Nr. 400). It would not be reasonable to suppose that Hesiod was the first to use any of these phrases: they must be regarded as part of the common stock of early Greek hexameter poetry. At the same time, Hesiod shows himself acquainted also with the hexameter tradition in a later phase, as 1s evident not only from the general Ionic appearance of his language with à 2 5, gen. sg. masc. in -ew and -ov, gen. pl. of a-stems in -έων, -àv, neglected F, etc., but also from particular expressions found also in Homer

which contain some

e.g. κέκλυτέ pev...| ὄφρ᾽

of these features,

εἴπω... Th. 644 f. with Ionic

pev and neglected F1® (5x IL, 3x Od.; = VUHZH, 181 Nr. 535), δεινοῖο πελώρου | Th. 856, Sc. 223 with unresolvable του beside older -ow (1x Il, 2x Od.; = VUHH, 170

Nr. 191). Homer and Hesiod thus appear to stand in one and the same linguistie tradition which had been developed, perhaps over several centuries, for the purposes of hexameter poetry and which was already highly complex by their time. While therefore there is no means of determining the precise extent to which Hesiod 15 indebted to his predecessors, the presence of traditional linguistic material which itself had originated at various dates makes it necessary to approach the study of specific features in his language with very great caution. For instance, a mere count of the places where the 15 There the periods tradition. Mycenaean,

is no need for us here to be concerned with the absolute dates of when these features are likely to have entered the hexameter For an attempt to make use of the chronological categories pre-Migration and post-Migration see T. B. L. Webster, ‘ Early

and late in Homeric diction ', Eranos liv, 1956, esp. 34-8.

16 On ὄφρ᾽ εἴπω see Chantraine, GH I, 134.

TRADITIONAL

POETIC

LANGUAGE

IN

HESIOD

39

poet observes or neglects initial F can achieve little unless each instance is examined in its context for traces of possible influence by the tradition, or alternatively for signs of Hesiodic innovation." It follows also that much of the interest in Hesiod's language centres upon the relatively small number of features where he appears to be departing from traditional forms of expression, since they alone enable us to assess, albeit imperfectly, how far his personal linguistic experience may have caused him to extend or modify the language already available to him in the hexameter tradition. But before turning to these distinctive features in his language we need to give further consideration to the relationship between Hesiod and the poetic tradition 1n which he stands, asking in particular whether his use of traditional language shows any obvious differences from Homer's which might be explained by supposing that the one set of poems has been composed orally and the other with the aid of writing. 17 Thus we cannot regard such phrases as τοῖο ἄνακτος T'h. 859 or βοὸς Ide κταμένοιο Op. 541 (= VUHH, 180 Nr. 499, and 187 Nr. 730) as indicating that Hesiod must have pronounced F in reciting his poems, even though we believe, for other reasons, that F was still pronounced in the Boeotian vernacular of Hesiod's own time: they are in fact evidence only for Hesiod's use of traditional phraseology of an ancient type. See further below, pp. 135-9.

CHAPTER IV

REPETITIONS

AND

FORMULAS

THE presence in the Hesiodie poems of repeated lines and phrases is referred to by Notopoulos in his article + concerned with the subject of Hesiod and oral poetry. He quotes the figures reproduced here in TABLE 4, which he has taken from Kretschmer,? and from these he concludes:

‘ It follows that

23%, of the Hesiodie corpus consists of repetitions of lines or phrases found in other parts of Hesiod's poems. This compares with 33%, in the case of Homer,

for out of 27,853 verses in

Homer, 9,253 are repeated or contain repeated phrases. This proportion of 23% and 33% respectively, a phenomenon found nowhere in subsequent Greek or Latin poetry, is the litmus test of the oral style.’ ὃ TABLE 4. Repetitions in the Hesiodic poems (after Kretschmer and Notopoulos) Repeti- | Repeti- | tions of | tions of | groups | lines | of lines

Repetitions of phrases before the caesura

| Repeti- | Repeti| tions of | tions of | phrases | phrases | after in the the middle | caesura | of the line

| Repeti| tions | occurring | wholly | Total | or partly ina different position

Th. (1,022 lines)

5

38

74

206

7

8

338

Op. (82 lines)

l

9

32

64

1

3

110

Sc. (480 lines)

l

10

22

40

3

3

79

7

57

128

310

11

14

527

Total (2,330 lines)

᾿ i

1 Hesperia xxix,

? DIH, 29.

1960, 177-97 ; see above, p. 21.

* Notopoulos, op. cit., 180.

REPETITIONS

AND

FORMULAS

41

Much is unsatisfactory here. Notopoulos’ comprehensive 23% takes no account of the differences between one poem and another, let alone between different parts of the same poem. By taking Kretschmer's total figure for the three poems together, he completely glosses over the discrepancy between the Theogony (33-195 on this basis) and the Works and Days (13:395). This difference would certainly suggest that an explanation of some, at least, of the repetitions in Hesiod is to be sought in the limited range of the poet’s subject-matter rather than in the practice of oral composition. Kretschmer * points out that the Theogony, concerned as it is with relationships between the gods, is much less varied in content than the Works and Days, where agriculture, seafaring, rules for the treatment of gods and men, autobiographical matter, maxims,

proverbs and stories are all part of the poet's theme. Notopoulos, so far from considering possible explanations, does not mention the difference between the poems. In any case 10 needs to be emphasised that this table from Kretschmer refers not to the number of lines in each poem containing repetitions, but to the number of phrases of various lengths which are themselves repeated. When Kretschmer’s list of the examples themselves is scrutinised, certain lines will be found to contain more than one such repetition : for example,

Op. 214 contains a repetition both in its beginning | ὦ Πέρση, σὺ δέ (as at Op. 213) and in its end φρεσὶ βάλλεο σῇσι | (as at Op. 107). Material from this one line has therefore been included by Kretschmer in both the third and the fourth columns

of his table, and makes

a difference of two to his

grand total. Yet Notopoulos has calculated his figure of 23% by dividing this grand total of 527 by the number of lines in the Hesiodic corpus ! The percentages calculated on such a basis will obviously be misleadingly large ; but from another point of view the above figures are misleadingly small: this particular table from Kretschmer does not give the total number of repetitions found 4 DIH, 80.

42

THE LANGUAGE

OF HESIOD

within the Hesiodie corpus as a whole, as Notopoulos’ conclusion from it implies, but refers only to those occurring within each of the poems taken individually. Kretschmer deals in a separate part of his work with repetitions found between one Hesiodie poem and another, and we need to supplement the figures for repetitions which Notopoulos has quoted with those given later by Kretschmer 5 for Th.-Op., Th.-Sc. and Op.-Sc. correspondences. He enumerates no less than 76 phrases which the Works and Days has in common with the Theogony, and another 75 contained in the Shield and one of the other two poems. But there is à more fundamental confusion which needs to be exposed. Our 'litmus test’, and the reason which led Milman Parry to argue that oral composition had played an essential part in the creation of the Homerie poems, is the presence, not of repeated phrases as such, but of a formulaic system. On the distinction between formulas and repetitions Parry has come cautionary words to say: 'ltisimportant... to remember that the formula in Homer is not necessarily a repetition, just as the repetitions of tragedy are not necessarily formulas. It is the nature of an expression which makes of it a formula, whereas its use a second time in Homer depends largely upon the hazard which led a poet, or a group of poets, to use it more than once in two given poems of a limited length.' $ Important as this last consideration is for Homer, it is even more important for the Hesiodie poems, which taken individually are comparable in length to a single book of the Iliad or Odyssey rather than to the whole of Homer, and when put together make a total equal to less than one tenth of the Homeric corpus. If the formula is ‘une expression qui est réguliérement employée, dans les mémes conditions métriques, pour exprimer une certaine idée essentielle ’,‚” we cannot 5 DIH, 52. 6 HSPh xli, 1930, 122. ? L’epithete traditionelle dans Homére, 1928, 16, on which see esp. J. B. Hainsworth,

The

Flexibility

of the Homeric

Formula,

1968,

33-5.

Much

recent work has concentrated upon this important question of the definition

REPETITIONS

AND

FORMULAS

43

establish how much is formulaic in Hesiod by a simple count of the repetitions within the extant poems, since there are many ‘essential ideas’ which Hesiod does not require to express on more than one occasion in these poems. Generally speaking, the shorter the poem, the smaller the likelihood of the poet's requiring to use a formula a second time. This means that in order to make repetitions serve as an accurate guide to the frequency of formulas in Hesiod, we should need, ideally, to note repetitions not only within a single Hesiodio poem,

nor

even

within

the

Hesiodic

corpus,

but

within

Hesiod's whole repertoire. This ideal is unattainable, but we cannot for that reason forget the principle which is involved. For example, the phrase ῥοδοδάκτυλος ᾿Ηώς occurs only once in Hesiod, at Op. 610;

it therefore falls outside the scope of

Kretschmer's inquiry and is not included in any of the figures or percentages quoted earlier in this chapter. Yet we can scarcely doubt that it is a formula in the full sense of the term, since it is so used frequently by Homer (5x Il., 22x Od.). On the other hand the presence of certain repeated expressions which are included in Kretschmer's figures, e.g. Th. 539 and 541, or Th. 122-3 and 724-5, could easily be explained without recourse to oral composition.? This does not mean that the repeated phrases in Hesiod are of no significance for us in considering the question of oral or written composition. Kretschmer’s figures for Hesiodic of a formula in Homer. J. A. Russo (see Bibliography) has sought to extend the definition to include phrases which contain metrically identical wordtypes showing the same grammatical and syntactical pattern. While such an

approach has some obvious attractions, I find mysclf in sympathy with the observations made by Hainsworth (CQ xiv, 1964, 155-64), Hoekstra (HMFP, 8-20) and Minton (TAPA xcvi, 1965, 241-53), who favour a stricter definition. Hainsworth (FHF, 35-9) defines a formula as a ‘ repeated word-group ’, whether with or without differences in inflexion, metrical value, and word-order, and this undoubtedly has practical advantages for many types of inquiry. But Parry’s definition is a more useful one for our

present purpose, since it will enable us to examine a compact group of interrelated phrases which together below, pp. 45-52). 8 See Kretschmer, DIH, 32.

constitute

a formulaic

system

(see

44

THE

LANGUAGE

OF HESIOD

repetitions are indeed remarkably high ın proportion to the length of the poems within which they are found. Parry observed this in 1930, and wrote that the proportion of repeated phrases in the Theogony ‘is far above anything we shall find for the poetry of times in which writing was beyond any doubt the usual means of composition.’® The point is worth reemphasising, however, since the direct opposite has more recently been suggested by Rhys Carpenter, who, in the context of a brief discussion of the oral poet’s method and its effect on the form of Homeric verse, writes as follows:

' But

Hesiod in the Works and Days does not display these traits of oral composition. He does not repeat himself; he does not use one line to build another; though he is patently familiar with Homer and occasionally echoes him, he does not work with the conventional Homeric tags, epithets, and phrases.’ 10

One can only counter such a statement by an appeal to the facts: Kretschmer !! noted 110 phrases which are repeated within the Works and Days. For the reasons already given (p. 41) it is misleading to divide this figure, after Notopoulos’ manner, by the number of lines in the poem, and so arrive at a percentage, until we have taken into account those lines which contain more than one such repeated phrase. A scrutiny of Kretschmer's list of the examples themselves !? shows that there are ten lines in the poem where this occurs, so that the 110 repetitions which he notes are all contained within 100 single lines. This means that over 12% of all the lines in the Works and Days contain a phrase repeated from earlier in the poem. When we go further and take into consideration phrases repeated from outside the Works and Days, this figure naturally increases. We have seen above (p. 42) how Kretschmer himself counted a further 76 phrases common to the Theogony

and the Works and Days. These are listed by him,}3 and when

material contained in Kretschmer’s two lists is put together, it is found that he quotes 170 lines of the Works and Days as containing a phrase which has been used in the Theogony or ® HSPh xli, 1930, 90. 12 DIH, 19-23,

10 FFS, 16. 13 DIH, 41-8.

ἢ DIH, 29.

REPETITIONS

AND FORMULAS

45

earlier in the Works and Days itself, i.e. a proportion of over 20%. We now refer again to the ideal situation where one might count repetitions within a poet’s whole repertoire (see above,

p. 43).

We

cannot

do this for Hesiod,

but if, for

instance, we were to enlarge the basis of comparison further by supposing that the poet of the Works and Days might have known the Ikad and the Odyssey, as well as the Theogony, the number of lines in his poem containing a repetition would be found to be no less than 385, or 46-5%, of the total number of lines in the poem.!* It is impossible therefore to agree with Carpenter’s assertion that Hesiod “does not repeat himself’ and 'does not work with the conventional Homeric tags, epithets, and phrases’, or to accept his conclusion that the Works and Days differs so fundamentally from Homer in this matter that it must mark the beginning of written literature, let alone that it was composed by the laborious process of inscribing it, not on to paper, but perhaps on to the very sheets of lead which were seen many centuries later by Pausanias ! 15 Perhaps a more instructive approach is to consider some specific examples in the general context of formulaic composition. In the rest of this chapter therefore we focus attention not on the proportion of expressions which happen to be repeated and which might be said to be ' regularly employed ' in the composition of the verse, but on the different ways in which Hesiod expresses an ' essential idea ’ to see if together they make up a formulaic system in Parry's sense. Let us examine a group of expressions meaning ‘the sea ’—an idea which occurs even within this small corpus with sufficient 14 ΤΊ must be emphasised that this figure is calculated purely on the basis of information contained in Kretschmer’s two lists of examples and Rzach’s Homeri loci similes, neither of which can be regarded as exhaustive: cf. above, p. 27 with n. 4.

15 This is suggested in all seriousness by Carpenter, who describes Hesiod as having composed ' as though he had watched each word as he wrote it down' (FFS, 16), and who regards it as plausible that ' the verses were written down in ἃ period before the introduction of Egyptian paper, when stone and clay and metal were the normal repositories of the new ** seratchings " ' (F FS, 15).

46

THE LANGUAGE

OF HESIOD

frequency to make a detailed comparison with Homer possible. The Homeric material has been conveniently collected by Miss D. H. F. Gray,!5 and her findings have been discussed by D. L. Page." In what follows we shall take their figures for Homer,

and

assemble

the

relevant

comparison. There are 384 places where πόντος occur in Homer,

ads,

data

from

θάλασσα,

Hesiod

πέλαγος

for

and

and in 143 of these the noun is ac-

companied by an epithet. Hesiod uses the same four words 52 times, including 23 times with an epithet. The closeness of the correspondence is best seen in terms of percentages (see TABLE 5), though it must be remembered with regard to these TABLE

5.

dds, θάλασσα, πέλαγος and πόντος in

Total occurrences |

Homer and Hesiod

Without

| With

epithet

| epithet

Homer

384

24] 62-89

143 37:295

Hesiod

52

29 55-8%

23 44.297

(One example in Hesiod = 1.90.)

and other statistics discussed below that, where

such small

numbers are concerned, one example represents a comparatively large percentage in the Hesiodic figure. Taking the words individually, ἅλς (3 occurrences without epithet, lx Th., 2X Op.) and πέλαγος (2x Th., once with epithet) are not frequent enough in Hesiod to allow a worthwhile comparison, but the figures for θάλασσα and πόντος are given in TanLEs 6 and 7. We see that Hesiod consistently shows a higher proportion of examples with epithet, though the difference here is hardly significant where such small numbers are concerned. More important than this is the general correspondence of the figures for the two poets, and the fact 19 (Ὁ xli, 1947, 109-13. 11 History and the Homeric Iliad, 1959, 225-7.

REPETITIONS TABLE

6.

AND FORMULAS

47

θάλασσα in Homer and Hesiod

Total Without | With occurrences | epithet | epithet Homer

119

Hesiod

19

101 849%,

18 15-195

15 4 78-995 | 21-194

(One example in Hesiod = 5-397) TABLE

7.

πόντος in Homer and Hesiod

Total Without | With occurrences | epithet | epithet Homer

136

63 46:89.

73 53-795

Hesiod

28

10 35°7%

18 64°3%

(One example in Hesiod = 3:69.)

that, in both poets similarly, πόντος is found with an epithet

much more frequently than is θάλασσα. The same is true when the figures are taken for the Hesiodic poems separately (see TaBLE 8), though the Shield, with only one example, has insufficient material for comparison. We are now ready to examine the various epithets which Hesiod employs with these nouns, to see how closely their use corresponds to a formulaic system like that found in Homer. Miss Gray divides the Homeric phrases into two categories, ‘traditional’ and ‘ individual’, but the classification, as she herself admits,!? is a rough one only, and in choosing between

them she may at times have been too much influenced by the frequency with which a particular formula happens to be repeated within Homer—a circumstance which, as Parry saw,!? is largely accidental. For example, she lists ἀτρυγέτοιο θαλάσσης as an “individual phrase ', since it occurs only at & 204; but its occurrence in the same position in Hesiod 18 CQ xli, 109.

19 See above, p. 42.

48

THE LANGUAGE TABLE

8.

OF HESIOD

Words meaning ‘ the sea’ in the separate Hesiodic poems Total Without | With occurrences | epithet | epithet ars, θάλασσα,

30

πέλαγος, πόντος Th. | θάλασσα

15

50%

11

9 81:895

πόντος

16

4 25%

dds, θάλασσα,

21

πέλαγος, πόντος Op.

15

50%

2 | 18:29 12 75%

14

7

66-795

| 33-3% 1 | 14-395

| θάλασσα

7

6 85.795

πόντος

12

6 50%

6 50%

The Shield has only θάλασσα, with epithet, Sc. 207.

(Th. 413 and 728) and the absence of any other formula of the same metrical value suggest that it may well be just as ‘traditional’ as some of the formulas which occur more frequently. For this reason, instead of collecting the Hesiodic material in the same form as Miss Gray's Homeric list,2° we shall list the phrases according to their grammatical case and their position in the line. Homeric parallels are mentioned here beside each example, for ease of reference in the discussion which follows.

Nom.

(1)

πόντος ἀπείριτος οἴδματι θυίων (θύων most mss.) Th. 109: οἷ. πόντος ἀπείριτος, in this position, κ 195 ; οἴδματι θύων | Φ 234,

Ψ 230.

(2) πόντος ἀπείρων | Th. 678 : cf. ἀπείρονα πόντον | A 350; πόντον ἀπείρονα ὃ 510;

(3)

᾿Ελλήσποντος ἀπείρων | 2 545. ἁλμυρὸς --- Πόντος | Th. 107, 964 : ἁλμυρός is not in Il. ; it occurs in Od. 8x, always ın the phrase ἁλμυρὸν ὕδωρ l.

20 CQ xli, 110.

REPETITIONS

(4)

AND

FORMULAS

49

| πόντος τ᾽ ἀτρύγετος Th. 696 : cf. | πόντον ἐπ᾽ ἀτρύγετον 7X Od.; also | πέμψας ἐπ᾽ ἀτρύγετον πόντον O 27.

Acc.

(5)

ἀτρύγετον πέλαγος ~~ οἴδματι θυῖον | Πόντον Th. 151 f.: for ἀτρύγετος as an epithet of πόντος see (4) above ; for οἴδματι θυίων | sec (1) above. ἰοειδέα πόντον | Th. 844 : the same phrase, in this position, A 107, and earlier in the line, A 298.

ἠεροειδέα πόντον | Th. 873, Op. 620 : the same phrase, in this position, Y 744 and 4x Od. εὐρέα πόντον | Op. 650 : the same phrase, in this position, Z 291 and w 118. οἴνοπα πόντον | Op. 817: the same phrase, in this position, 4x Il., 5x Od. ἀτρυγέτοιο θαλάσσης | Th. 413, 728 :

the same phrase, in this position, & 204. πολυφλοίσβοιο θαλάσσης | Op. 648 : the same phrase, in this position, 6x

Il., 2x

Od.

ἀμαιμακέτοιο θαλάσσης | Sc. 207: no Homeric parallel; ἀμαιμάκετος is used with Χίμαιρα at Z 179, II 328 f. (cf. also Th. 319

discussed below, p. 77), and with ἱστὸν... νηός E 311 ; of these only Z 179 has the adjective in this position.

(13)

Dat.

| πόντου τ᾽ ἀτρυγέτοιο Th. 737, 808 :

(14)

cf. (4) above. | πόντου κυμαίνοντος Op. 390 :

(15)

ef. πόντον --*- κυμαίνοντα | & 229 and 5x πολυκλύστῳ ἐνὶ πόντῳ | Th. 189 :

(16)

the same phrase, in this position, ὃ 354, ζ 204, 7 277. ἐν ἠεροειδέι πόντῳ | Th. 252 :

(11)

the same phrase, in this position, 6x

Od.;

Od.

cf. also

(7) above. εὐρέι πόντῳ | Op. 507 : the same phrase, in this position, 6x Od.; cf. also (8) above.

50

THE

(18)

οἴνοπι πόντῳ

LANGUAGE

OF HESIOD

| Op. 622:

the same phrase, in this position, 6x Od.; cf. also (9) above. (19) | πόντῳ ἐν ἀτρυγέτῳ Th. 241: cf. (4) and (13) above.

Ψ

316

and

While only four of these phrases occur more than once in Hesiod, so that it would scarcely be possible to speak of them as being ‘regularly employed’ if we were dependent for our knowledge upon Hesiod alone, their largely formulaic character is at once suggested by the use of many of them as formulas in Homer, and confirmed by the stringent economy which is disclosed by a comparison of their metrical values. We take these two points in turn. No less than ten (nos. 6, 7, 8, 9, 10, 11, 15, 16, 17, 18) of

these nineteen same case and phrases which Three occur in

expressions are position, and an occur separately Homer with the

found in Homer in exactly the eleventh (no. 1) combines two in the same case and position. noun and adjective in the same

position in the line, but in another case (nos. 4, 13 and 19, all like [πόντον ἐπ᾽ ἀτρύγετον 7X Od., which does not occur

in Hesiod). Two (nos. 2 and 14) are found in Homer both in a different case and in a different position in the line, though there occurs in addition the phrase related to one of these (no. 2) ‘EAAjomovros ἀπείρων |, whose singularly inappropriate

epithet suggests that πόντος ἀπείρων | in this position may have been a traditional phrase. Two (nos. 3 and 5) combine an epithet and a noun not found together in Homer, though both cpithet and noun occur there in other phrases meaning “the sea’. There remains only the solitary example from the Shield (no. 12), where an epithct is found which is nowhere applied to the sea in Homer. No two of these nineteen phrases are ‘ doublets ’ of the same case and motrical pattern, and morcover no phrase exists in Homer which could occupy the same place as any of the nine

phrases which are not found in Homer.

From this point of

view the observance of formulaic ‘economy’

is complete—

REPETITIONS

AND

FORMULAS

δ]

more complete indeed than in Homer, where θαλάσσης eüpvrropoıo stands once in the Iliad and twice in the Odyssey in the space more regularly filled by πολυφλοίσβοιο θαλάσσης 6x IL, 2x Od.? We shall see presently that this apparent economy is less rigid, both in Homer and Hesiod, than our analysis up to this point would suggest, but before exploring that question in greater detail 33 a number of conclusions about Hesiod's use of formulas may be drawn from the evidence which we have already examined. (a) Hesiod is content with the traditional phrases wherever one is available to suit his purpose. He does not try to improve upon them by substituting different ones of his own invention. (b) Even where no complete phrase of the required metrical pattern is available, Hesiod always recombines nouns and epithets which are found in Homer in phrases meaning ‘ the sea’. The only exception is the single example from the Shield, but even here no alternative adjective of the same scansion as ἀμαιμακέτοιο can be suggested from any Homeric phrase for “the sea ’. (c) On the other hand we must not overlook the fact that an element of innovation is present even where a new phrase is closely analogous to an existing one. For example, the extension of a phrase from the accusative to other cases (as in nos. 4, 18 and 19), although at first sight it appears a trivial change, may indicate a later stage in the hexameter tradition, where phrases whose earlier use was restricted to a particular context are now used more flexibly.” (d) Especially illuminating are the places where Hesiod uses as a descriptive epithet an adjective or phrase occurring in Homer predicatively. For example, ἀπείριτος (in no. 1) is a straightforward epithet at 75. 109, but may well be predicative 24 at « 195, where Odysseus tells how from the hilltop 21 See the discussion by Page, HHI, 227. 22 See below, pp. 52-4 and Chapter V. 23 For a discussion of extensions of this kind see Sellschopp’s chapter on the use of adjectives in Hesiod, esp.

SUH, 22-5.

24 Tt is so taken by Gray, CQ xli, 110 note.

52

THE

LANGUAGE

OF HESIOD

he saw that the land was an island, τὴν πέρι πόντος ἀπείριτος

ἐστεφάνωται | ‘encircled by the sea without limit’. Predicative force is further suggested by the combination of ἀπείριτος with the same verb at h.Aphr. 120 ἀμφὶ δ᾽ öpudos ἀπείριτος Eoredavwro | and at Sc. 204 περὶ δ᾽ ὄλβος ἀπείριτος ἐστεφάvoro |, though in none of these passages considered alone is there any imperative reason for understanding ἀπείριτος predicatively, so that the point is arguable. Clearer than this is οἴδματι θυίων in the same phrase (no. 1) and at T^. 131 (no. 5). In Homer the phrase certainly has predicative force: at ® 234 the river Scamander rushes wildly upon Achilles—o δ᾽ ἐπέσσυτο οἴδματι θύων | —, and at V 230 the Thracian sea roars and seethes when the winds return to it, after spending the night fanning the flames of Patroklos’ pyre—o δ᾽ ἔστενεν οἴδματι θύων |. In both passages οἴδματι θύων 15 part of the action described, but not so in the Theogony, where the phrase is used purely as an additional epithet for the sea, conveniently filling the space at the end of the line. It seems reasonable to infer that these expressions originated in their predicative sense, and that the merely descriptive use of them is a later development. We cannot of course assert either that Hesiod himself initiated the later, descriptive usage, or that

234

and Y 230 must have been composed during a period before the descriptive use was known ; but at least we can add the indication suggested by these examples and by the extensions noted above under (c) to other evidence for the relatively late place which Hesiod holds in the early hexameter tradition. But before ending this chapter it must be mentioned that Hesiod is not everywhere so content to depend upon his predecessors as the first two of these conclusions might suggest, and that he does from time to time depart from traditional forms of expression. Here is an example of this, where again Hesiod has to express the same essential idea of ‘the sea’. He 15 talking about how Hekate aids various classes of men, including ‘ those who work the stormy sea ’. The line (Th. 440) runs: καὶ τοῖς, ot γλαυκὴν δυσπέμφελον ἐργάζονται.

REPETITIONS

AND

FORMULAS

53

Here he could easily have used either πόντον, as he did when requiring a word for the sca in exactly the same position at Op. 6T2—EArenev ἐς πόντον φόρτον τ᾽ ἐς πάντα τίθεσθαι--, Or πέλαγος, πουλὺν

as he did at Th. 190—

χρόνον ; but

instead

he

chose

| ὡς φέρετ᾽ ἂμ πέλαγος γλαυκήν,

“the

grey’.

The adjective γλαυκός occurs in Homer only at Π 34, as an epithet for the sca in the phrase γλαυκὴ δέ σε τίκτε θάλασσα, apart from which Γλαύκη is uscd as the name of one of the Nereids at Σ' 39 (so too at Th. 244). We may compare with Hesiod's use of this adjective here Homer's use of ὑγρή as à noun, but at least the latter has a metrical reason, in that it

cannot be replaced by any other word for ‘the sea’ in any of the places where it occurs, namely ἐφ᾽ ὑγρήν | K 27, 2 341, a 97, ὃ 709, e 45, and ἐπὶ τραφερήν τε καὶ ὑγρήν

|

& 308, v 98. Hesiod's γλαυκήν 15 not an isolated instance of his apparent readiness

to innovate,

at least on occasion.

He

elsewhere

employs un-Homeric vocabulary in contexts where Homer seems to have a word of the same meaning, as in ἀληθέα γηρύσασθαι | Th. 28, beside ἀληθέα μυθήσασθαι | Z 382, £125,

p 15, e 342, or he recombines

traditional

words

in

new

phrases, as in ἐπὶ στιβαροῖσι μέλεσσιν | Th. 152, 673, Op. 149, Sc. 76, beside evi γναμπτοῖσι μέλεσσιν | A 669, 2 359, A 394,

v 398, 430, This last common to similar and Homer. If

φ 283. example is striking in that it shows a formula the three Hesiodic poems which differs from a obviously related formula found six times in the phrase evi γναμπτοῖσι μέλεσσι | was already

established in the hexameter tradition before Hesiod’s time,

his own στιβαροῖσι appears to represent an innovation which, once made, has been adhered to in the composition of all three poems. But our difficulty here is that which was mentioned at the start of Chapter III: we cannot assume that the traditional epic poetry known to Hesiod was identical in these matters of detail with that which we now possess in our texts of Homer. In other words, from Hesiod’s own point of view there may in some at least of these instances have been no

δ4

THE LANGUAGE

OF HESIOD

departure from the language with which he was already familiar. Fortunately there are sufficient examples within Hesiod himself to provide ample material for a discussion of this question of economy of expression, and this forms the subject of the next chapter.

CHAPTER THE

PRINCIPLE

V OF

ECONOMY

ONE of the main characteristics which Parry noted in his study of formulas used by Homer is the principle of economy, whereby the poet's language 1s almost wholly free of expressions which have the same metrical value and contain the same idea, so that they might be exchanged with each other.! It must be said at once that Hesiod shows many departures from this principle. Here is a list of places where, saying essentially the same thing in the same metrical context, he expresses himself in two or more different ways. The words "essentially the same thing' must be interpreted broadly, since it will be found that the ideas contained in the alternatives are not always identical; but the cardinal point is that here we have words and phrases which in their present contexts appear to be mutually interchangeable. For ease of reference later, these apparent doublets are listed according to the place where the first member of the pair or group is found. Th. 57 | νόσφιν ἀπ᾽ ἀθανάτων Th. 302 and Op. 169 | τηλοῦ ἀπ᾿ ἀθανάτων Th. 69 γαῖα μέλαινα |

Th. 159, 173, 479, 821, 858 γαῖα πελώρη | (usually Γαῖα : see below, p. 66) Th. 94 καὶ ἑκηβόλου ᾿Απόλλωνος | Sc. 70 -os ᾿Απόλλωνος Παγασαίου | Th.

138 δεινότατος παίδων: θαλερὸν δ᾽ ἤχθηρε τοκῆα

Th. 155 δεινότατοι παίδων, σφετέρῳ δ᾽ ἤχθοντο τοκῆι Th. 148 οὐκ ὀνομαστοί | Th. 310, Sc. 144, 161 οὔ τι φατειόν | (-ós, -dv) 1 This principle has more recently been reaffirmed by Page: 224 f., and the references to Parry given there.

see HHI,

See also Hainsworth, FHF,

23, though his work as a whole is especially valuable for criticising and correcting too mechanical a view of the way in which oral poets made use of traditional expressions.

56

THE

LANGUAGE

OF

HESIOD

Th. 150 = 671 τῶν ἑκατὸν μὲν χεῖρες ἀπ᾿ ὦμων ἀΐσσοντο

Th. 151 ἄπλαστοι Th. 672 πᾶσιν ὁμῶς

|

κεφαλαὶ

Vore / δὲ εκαστῳ

, “ΠΕνΡΤΉΚΟνΤα

Th. 195 | ποσσὶν ὕπο ῥαδινοῖσι (of Aphrodite) Th. 842 | ποσσὶ δ᾽ ὕπ᾽ ἀθανάτοισι (of Zeus) cf. also | ποσσὶ δ᾽ ὑπὸ λιπάροισι 4x Il., 4x

Od., usually

of men but once of Hera (# 186). Th. 218-9

Κλωθώ

te

Λάχεσίν

τε

καὶ

"Ärponmov,

ai Te

βροτοῖσι

γεινομένοισι διδοῦσιν ἔχειν ἀγαθόν τε κακὸν TE ~



>

J

,

Th. 905-6 Κλωθώ τε Λάχεσίν τε kai" Atpomov, αἵ τε διδοῦσι [4

θνητοῖς ἀνθρώποισιν Th.

223

,

~

ἔχειν ἀγαθόν τε κακόν τε

πῆμα θνητοῖσι βροτοῖσι]

Op. 82

πῆμ᾽ ἀνδράσιν ἀλφηστῇσι |

Th. 284 Th. 531

χθόνα μητέρα μήλων | χθόνα πουλυβότειραν |

also χθονὶ πουλυβοτείρῃ | Op. 157, 252 and 510. Th. 300

-ην

Th. 622

-n μεγάλης ἐν πείρασι γαίης |

Th. 317 Sc. 102

σὺν ἀρηιφίλῳ ᾿Ιολάῳ | προσέειπεν ἀμώμητος 'IóAaos |

Th. 324

πυρὸς μένος αἰθομένοιο |

Th,

483

2

ev)

ζαθέης ὑπὸ

κεύθεσι

;

yains

|

Th. 563 πυρὸς μένος ἀκαμάτοιο | Th. 441 καὶ ἐρυκτύπῳ ᾿Εννοσιγαίῳ | also Th. 456 καὶ ἐρίιςτυπον ᾿Εννοσίγαιον | Th. 930 καὶ ἐρικτύπου ᾿Εννοσιγαίου | Sc. 104

καὶ ταύρεος ᾿Εννοσίγαιος

Th. 448 Th. 440

-ην ἐθέλουσά γε θυμῷ | -ων θυμῷ γ᾽ ἐθέλουσα |

Th. 558

τὸν δὲ μέγ᾽ ὀχθήσας

Op. ὅ8

|

τὸν δὲ χολωσάμενος | προσέφη νεφεληγερέτα Ζεύς

Th.625 | , , Koövou ἐν φιλότητι Th, 634 | οὖς τέκεν ἠύκομος ‘Pein | Koövw εὐνηθεῖσα

THE

Th. 636 Th. 712 Th.

713

PRINCIPLE

OF

ECONOMY

| συνεχέως ἐμάχοντο | ἐμμενέως ἐμάχοντο ,

.

Éyetpav

Sc. 261

| ""X" δριμεῖαν

Th. 818 Op. 667

βαρύκτυπος "Evvoctyatoc | Ποσειδάων ἐνοσίχθων |

Th. 923

Th.

944

Op. 52 Op.

57

| μειχθεῖσ᾽

"

|

ἔθεντο |

ἐν φιλότητι

|

θεῶν

βασιλῆι καὶ ἀνδρῶν

Διὸς νεφεληγερέταο

Alta τερπικέραυνον |

218 Δία μητιόεντα |

Op. 91 } νόσφιν ἄτερ τε | κακῶν kai ἄτερ χαλεποῖο πόνοιο πόνων καὶ ὀιζύος P Op. 113 Sc. 351 | --νὐ- τε πόνου καὶ ὀιζύος A

Τ

Op. 678 Op. 682

/

| ἄλλος δ᾽ εἰαρινός | εἰαρινὸς δ᾽ οὗτος

Op. 783 | ἀνδρογόνος δ᾽ ἀγαθή Op. 188, 794 | ἐσθλὴ 8° ἀνδρογόνος

Op. 801

| οἰωνοὺς κρίνας

828

| ὄρνιθας κρίνων

Op.

Sc. ὅ9 Sc.

192

Sc. 846 50.3331)

'Αρην ἄατον πολέμοιο | ἐναρσφόρος οὕλιος "Αρης |

"Agng ἀκόρητος ἀυτῆς | κι



ἐπιόντα "Apnv

προσιόντα

,

ὅς. 495

[αὑτὸς δὲ βροτολοιγὸν

δοκεύσας

Sc. 379 Sc. 412

ὡς ot ἐπ’ ἀλλήλοισι πέσον μέγα κεκλήγοντες ὡς of κεκλήγοντες ἐπ’ ἀλλήλοισιν ὄρουσαν

Sc. 414

| κτεινέμεναι μεμαώς

Sc. 453

| κακτάμεναι μεμαώς

Sc. 424 Sc. 448

Διὸς ταλακάρδιος vids | Διὸς θρασυκάρδιον υἱόν |

At first sight there appears to be ample evidence here that Hesiod violates repeatedly a rule which we find to be generally observed in the Homeric poems. He might even seem to have E

δ8

THE

LANGUAGE

OF HESIOD

gone out of his way in some places to avoid repeating a traditional expression with which he shows himself elsewhere to be familiar, and this is especially striking when he fails to use one of the well-established formulas combining an epithet with a common noun ? or a proper name.? But it would be a mistake to conclude that these examples must necessarily be interpreted as indicating written composition, for there can be no doubt that certain kinds of alternative expression existed even within the oral tradition. The extent to which breaches of the principle of economy occur in Homer is perhaps greater than is generally supposed, and too strong an insistence on the strictness with which the principle is observed 1s in danger of obscuring the real range of choice which is frequently open to the poet. The point may be illustrated from that same system

of formulas

which

we

examined

earlier,

where

the

essential idea “the sea’ is expressed. After noting that, in 143 out of 384 places where ἅλς, θάλασσα, πέλαγος and πόντος occur in Homer, the noun is accompanied by an epithet, Page shows that only 17 formulas account for no less than 128 of these 143 passages. He goes on: ‘ Moreover, in the Iliad, excepting a single line in the Fifteenth Book, the law of economy is strictly observed : each formula is unique, in the sense that it cannot be replaced by any other formula in the same part of the line.’ * The solitary exception referred to is

θαλάσσης εὐρυπόροιο | at O 381, which we have mentioned above. But Homer's language is in fact much more flexible than this would lead us to suppose. While it is true that among these noun-epithet formulas economy is almost complete, there sometimes exist formulas of a different grammatical structure which do provide ‘doublets’. Here are three examples : * e.g. Th. 69/159 etc., 284/531, where γαῖα μέλαινα and χϑόνα πουλυβότειραν may be regarded as the traditional formulas : see below, pp. 65 £., 63. * e.g. Op. 52/273, Sc. 59/192/346, where Ala repmıxepavvov and “Apny darov

πολέμοιο are most obviously traditional: * HAI, 225. 5 See p. δ].

see below, pp. 65, 63, 71.

THE

PRINCIPLE

OF ECONOMY

59

(1) Page ® explains the existence of ἰοειδέα πόντον | A 107, beside μεγακήτεα πόντον | y 158 as allowing for either preceding consonant (before Fio-) or short vowel (before μμεγα-), and goes on: “there is no alternative phrase for “ the sea”, accus., In this position’. This is true so far as the nounepithet formulas are concerned, but there is an alternative of a different structure which has exactly the same metrical value : μέγα λαῖτμα θαλάσσης | found at 8 504, € 174 and ı 260. Tf the word λαῖτμα is to be connected with λαιμός, ‘throat ’,” this

phrase may well have had its origin in some special context, possibly at T 267, where Talthybios hurls the body of the slaughtered boar πολιῆς ἁλὸς és μέγα λαῖτμα |. The notion of the carcase being devoured by the jaws of the sea 1s particularly appropriate here, and is sustained by βόσιν ἰχθύσιν in the next line. But whatever the phrase's origin, the fact remains that for the

author

of the

Odyssey

μέγα

λαῖτμα

θαλάσσης

was

available as an accusative formula for ‘the sea’ in this position, and to that extent a choice between alternatives was open to him,—a choice which he exercised differently in contexts which are for practical purposes identical: προφυγὼν ἰοειδέα πόντον | À 107, beside φυγέειν μέγα λαῖτμα θαλάσσης | 8 504. Hesiod has the same choice, and uses ἰοειδέα rovrov | at Th. 844 and μέγα λαῖτμα θαλάσσης | at Op. 164. (2) A similar but more extensively used example is seen in

the phrase εὐρέα νῶτα θαλάσσης |, which, while formally exhibiting θαλάσσης as a simple genitive without an epithet, is best taken

as a whole, that is as a formula

for ‘the sea’,

accusative, occupying the last three feet of the lme and It thus forms a ‘doublet’ with beginning with a vowel.

ἠεροειδέα πόντον |, and we find, with no important distinction in their contexts: én’ ἠεροειδέα πόντον | V^ 744, 4x Od. (cf. és ἠεροειδέα πόντον | Th. 873 and Op. 620), beside ἐπ᾽ εὐρέα νῶτα θαλάσσης

| 3x

Il., 7x

Od. (cf. Th. 762, 781, 790, 972).

(3) Finally we consider formulas for ‘the sea’, accusative, with initial vowel, and filling the last two feet ofthe line. After * ΠΗ͂Ι, 268 n. 27.

7 See Boisacg, DEG

and Frisk, GEW

s.v.

60

THE

LANGUAGE

OF

HESIOD

explaining that οἴνοπα πόντον | is distinguished from εὐρέα πόντον | by its initial digamma, Miss Gray adds that ἁλμυρὸν

ὕδωρ | (8x Od.) is “needed when water is swallowed or exgurgitated '.5 What she does not mention is that it is by no means invariably used in such a context. For example at ı 227

and

470

we

find

ἐπιπλεῖν

ἁλμυρὸν

ὕδωρ

| where

the

poet could have chosen εὐρέα πόντον |, cf. ἐπιπλὼς εὐρέα πόντον | Z 291, and a further alternative in this position, used in the Iliad and the Odyssey, is provided by ὑγρὰ κέλευθα | : cf.

in

this

same

context

ἐπέπλεον

ὑγρὰ

κέλευθα

| A

312,

ὃ 842, o 474.

It is clear from these examples that alternative expressions can arise from causes which have nothing to do with the use of writing and the more deliberate choice of language which writing makes possible, so that a breach of the principle of economy cannot invariably be interpreted as an indication of written composition. To say this is not to deny that there is a real ‘ principle of economy’ present in the formulas of oral poetry, but we must think of this not as a rigid exclusion of all alternative forms of expression, but rather, in the full context of oral composition, as a natural accompaniment of and aid to fluent verse-making, whereby the poet can use and re-use many expressions from a large stock with which his mind is already furnished, without feeling any need deliberately to create new ones. Αὖ the same time his method does not allow that stock to remain static : * the re-use of a phrase previously restricted in scope in a more general context (as probably with ἁλμυρὸν ὕδωρ | ), or the force of analogy, extending a phrase to

a new grammatical case (as perhaps with ἠεροειδέα πόντον |, from the dative) can at any time bring into being a new expression which may easily have the same meaning and metrical shape as a traditional formula already known to the poet, and this expression may in its turn be re-used as a formula. Parry gives many examples of this in his chapter 5 CQ xli, 111 note. ? ef. esp. Hainsworth's chapters formulas in Homer, FHF, 58-89.

on the modification

and expansion

of

THE

PRINCIPLE

OF ECONOMY

61

entitled * Les formules nom-épithéte équivalentes ’,!° where he demonstrates with especial clarity the important part played by analogy in the creation of these doublets. Returning therefore to our list of Hesiodic examples, we need, if we can, to distinguish between doublets of this sort,

which might have arisen in the course of oral composition, and those which are more deliberate, showing that the poet has leisure to invent new forms of expression and substitute them for traditional ones. Most of the examples from our list belong to the former category, and are explicable in various ways : (1) The alternative expressions are already present in Homer. This is true of Th. 558/Op. 53, the former’s μέγ᾽ ὀχθήσας being by far the more common in Homer, although χολωσάμενος occurs in this sort of context at Y 482, σ 25. The origin of this doublet is perhaps to be sought in an extension from the feminine,

where

the alternatives

would

not be of the same

metrical value ; cf. especially the lines Γ 418 τὴν δὲ χολωσαμένη προσεφώνεε 0t Ἀφροδίτη Q 55 τὸν δὲ χολωσαμένη προσέφη λευκώλενος "Hp, where μέγ᾽ ὀχθήσασα (never found in Homer) is excluded by the metre. (2) The alternative expressions may be explained by reference to their contexts, ie. the traditional expression is avoided on account of some awkwardness which would arise if it were used. From Homer we may quote Parry's explanation " of the

unique στεροπηγερέτα Ζεύς | at II 298, beside the regular νεφεληγερέτα Ζεύς | 22x Il., 8x Od.: the departure from the

traditional

expression

is clearly

here

necessitated

by

the

context, which runs

ὡς 9 ὅτ᾽ ἀφ᾽ ὑψηλῆς κορυφῆς ὄρεος μεγάλοιο κινήσῃ πυκινὴν νεφέλην στεροπηγερέτα Ζεύς. A comparable Hesiodic example may be seen in Th. 37/51, where the lines TA. 31 ὑμνεῦσαι τέρπουσι

ny

e

Atos voov

10 ETH, 218-40.

rris 'OMumov] T Η 5] n ETII, 236.

62

THE

LANGUAGE

OF HESIOD

appear to show a departure from economy, with Th. 51 containing the more traditional form of expression (Aids νόον is found in this position 3x IL, 2x Od.); but an obvious explanation for the substitution of μέγαν for Aids at Th. 37 is to be found in the previous line, which ends with the words

ταὶ Au πατρί |, after which Aids would be intolerable. A similar explanation may be adduced to account for Op. 91/Op. 113/8c. 351. Here the Shield’s πόνου appears to stand closest to the tradition (cf. the expressions πόνον τ᾽ ἐχέμεν kai ὀιζύν

| N 2, 0 529, and πόνος τ᾽ ἔσεται

καὶ olds

| &

480).

But πόνου would not be possible at Op. 91, where the poet is going on to use πόνοιο at the end of the line, and in any case there is no question at this place of the poet’s having in mind the πόνος καὶ ὀιζύς formula ; indeed, κακῶν is quite an obvious word for him to use, cf. Th. 55 | λησμοσύνην τε κακῶν. The

line Op. 113 is then to be explained on the one hand by the influence of the traditional πόνος καὶ ὀιζύς formula (hence not κακῶν), and on the other by the force of analogy with Op. 91 | νόσφιν ἄτερ τε κακῶν (hence πόνων rather than πόνου). (3) The

alternative

expressions

may

be

explained

as

having

originated by analogy with expressions found elsewhere. This influence of analogy may be seen to operate in a variety of Ways : (a) Extension from one grammatical case or person to another. A clear example of this is Th. 324/563, where the former uses the more traditional form of expression (cf. Z 181 f. = Th. 323 f., also πυρὸς κρατερὸν μένος αἰθομένοιο | A 220, and πυρὸς

αἰθομένοιο | 7X Il., 2x Od.). The alternative ἀκαμάτοιο | at Th. 563 15 an extension to the genitive from the familiar ἀκάματον πῦρ

| 7 X Il., 2x

Od.

We may include here T'h. 223/0p. 82:

θνητοῖσι βροτοῖσι |

occurs in this position at ἡ 210, and again at Th. 500, as well

as earlier in the line at y 3 and p 386, whereas ἀνδράσιν

ἀλφηστῇσι | at Op. 82, also Th. 512, is not found at the end of the

line

in

Homer;

we

may

however

compare

ἀνδρῶν

ἀλφηστάων | ζ 8, or alternatively the dative ἀνδράσιν ἀλφηστῇσι

THE

PRINCIPLE

OF

ECONOMY

63

itself, but at the start of the line, at a 349 and Sc. 29, so that this

might be included instead among the examples under (c) below. The doublet Th. 713/Sc. 261 is open to a similar explanation : the Shield’s ἔθεντο | is found three times in Il., always at the end of the line, cf. esp. πόνον καὶ δῆριν ἔθεντο | P 158;

ἔγειραν | on the other hand, at Th. 713 and similarly at Th. 666, may be regarded as a simple extension from the

singular éyewe(v) | at the end of the line 5x

Il., 3x

Od.,

cf. esp. πόλεμον δ᾽ ἀλίαστον ἔγειρε | Y 31.

(b) Extension of an epithet or descriptive phrase from one reference to another. Many of Parry's Homeric examples !? belong herc, and from our Hesiodic list we may include Sc. 59/346, where the former contains a traditional phrase (cf. the nom. "Apns

dros

πολέμοιο

| 3x

IL), while the latter’s

ἀκόρητος ἀυτῆς | has arisen by analogy with δεινῆς ἀκόρητοι ἀυτῆς |, of the Trojans, at N 621. A further extension of reference involving the same formula is to be seen in Sc. 459 ᾿Αμφιτρυωνιάδης, δεινῆς ἀκόρητος ἀυτῆς and Sc. 433 τοῖος dp Ἀμφιτρυωνιάδης, ἀκόρητος ἀυτῆς. Another such extension is to be seen in Th. 284/531. Here we cannot doubt that πουλυβότειραν is the regular epithet (3x IL, 1x Od.; also the dat. χθονὶ πουλυβοτείρῃ | 10x Il.,2x Od.), and this makes χθόνα μητέρα μήλων

| at Th. 284

the more surprising. But μητέρα μήλων | is well attested in Homer as an epithet of places ("Irava τε μητέρα μήλων | B 696, Φθίην... μητέρα μήλων | I 479, Θρήκῃ ... μητέρι μήλων | A 222, Πύλῳ ἔνι, μητέρι μήλων | o 226, also Apradinv

... μητέρα

μήλων

| A. xix, 30), and

it is but a

short step from here to Th. 284, where Pegasos flies away from the earth and is described as προλιπὼν χθόνα, μητέρα μήλων |, especially where the meaning of the alternatives is practically identical. An instance of this process in reverse is found at A 770, where πουλυβότειρα, elsewhere in Homer invariably an epithet of χθών, is used as an epithet of a place in the phrase Ἀχαιίδα πουλυβότειραν |. 12 ETH, 224-7.

64

THE

LANGUAGE

OF

HESIOD

A rather different kind of example, but one which is still open to essentially the same explanation involving a shift of The phrase θεῶν reference, is the doublet Th. 923/944. βασιλῆι καὶ ἀνδρῶν | is never found in Homer, who uses either Au Κρονίωνι ἄνακτι

| 4x

Il., or κελαινεφέι Κρονίωνι

| 3x

Il.

Hesiod employs both these expressions elsewhere (cf. Op. 69, Sc. 53), but he may have wished to avoid them here because of the resultant succession of dative singular forms beginning with φιλότητι, though it was open to him to use the genitive Διὸς νεφεληγερέταο | as he does below at Th. 944 in exactly the same context. (Kretschmer’s argument 13 that Th. 944 is shown to be later than 7h. 923 by the genitive is unsound ; either case is admissible, cf. αἰγιόχοιο Atos φιλότητι μιγεῖσα | Th. 920.) But we need not conclude that the phrase θεῶν βασιλῆι καὶ ἀνδρῶν | has been invented for this passage ; it has been used already at Th. 897, not indeed for Zeus, but for the destined son of Metis who is to supplant Zeus and whose birth Zeus forestalls : αὐτὰρ ἔπειτ᾽ ἄρα παῖδα θεῶν βασιλῆα καὶ ἀνδρῶν ἤμελλεν τέξεσθαι, ὑπέρβιον ἦτορ ἔχοντα ... Kretschmer 16 argues that Th. 923 is older here, on the grounds that at Th. 897 it would be more appropriate to speak of this child as ‘ destined to be’ king of gods and men; but this is unnecessary, since in the context of Th. 898 f. no misunderstanding is possible. Once it 1s accepted that the phrase may have been first used in a context, like that of 7h. 897, where

there was no question of its taling the place of an equally appropriate traditional expression of the same metrical pattern, then its use in a context for which an older phrase was available, as at Th. 923, may be readily explained as an analogical extension to a more definite reference. (c) Extension from one metrical context to another.

Here

one

of the phrases forming a doublet may have its origin in a context where no departure from economy is involved, so that the need for the new expression in the first place could have 13 DIH, 38.

14 DIH, 34.

THE

PRINCIPLE

OF

ECONOMY

65

arisen directly from the metre. For example, in the doublet Op. 52/273 we cannot doubt that τερπικέραυνος is traditional, for it is found frequently in Homer in this position as an epithet

of Zeus

whereas and

μητιόεις

there in the

(8x

Il,

7x

Od.

in nom.,

acc.

and

dat.),

is used only once in the whole of Homer, phrase

φάρμακα

μητιόεντα

| ὃ 227,

' slalful

remedies ’. In Hesiod on the other hand we find Διὸς πάρα μητιόεντος | at Op. 51 and 769, as well as Aud μητιόεντι at Th. 286, while τερπικέραυνος occurs only at Op. 52 (acc.)

where μητιόεντα | would be intolerable after μητιόεντος at the end of the previous line. But before we are tempted to see a deliberate breach of economy in these passages, it is worth noting another place where μητιόεις occurs as an epithet of Zeus. The line (Th. 457) runs: Ζῆνά τε μητιόεντα, θεῶν πατέρ᾽ ἠδὲ καὶ ἀνδρῶν. In

this

context,

where

a

final

short

vowel

is

needed,

τερπικέραυνον would not fit, nor would any other of the many epithets which Homer applies to Zeus.15 Faced with this problem an oral poet, just as well as one who used writing, might have applied μητιόεις to Zeus by analogy with the familiar μητίετα Ζεύς, and once it had thus become part of the poet’s total stock, the phrase could have been re-used even where the older τερπικέραυνος would have fitted. It is not necessary to assume that Th. 457 was itself composed earlier than our lines where μητιόεις is found in the last two feet ; all we need to suppose 1s that unriöeıs was first used as an epithet of Zeus in a position where τερπικέραυνος was excluded by a following initial consonant. In other words the explanation of this doublet could lie in the different values of the final syllable, irrelevant when the expression stands at the end of a line, but important in other positions. The same explanation might be snggested for Th. 69/159 etc., where the alternatives differ in the length of their

final syllable.

It is clear that γαῖα μέλαινα | as at Th. 69, is

15 For a convenient list of these see Cunliffe, Homeric Proper and Place Names,

17.

66

THE LANGUAGE

traditional (4x Il.,

OF HESIOD

1x Od.), and γαῖα πελώρη | at the other

references is an innovation ; 16 but the latter may be an extension from a context with the epithet in a different position, as is found at Th. 861 πελώρη καίετο γαῖα | where μέλαινα cannot serve (cf. Th. 505, 731). An alternative explanation, or at least one which would account for the high frequency of πελώρη (5x Th.) beside μέλαινα (once only) in this context, may be detectable in the different meaning of the epithets and the different senses of γαῖα in the Theogony, simply ‘earth’ at Th. 69, but ‘ Earth ', personified, certainly at Th. 159, 173, 479 and 821, and possibly also at Th. 858 ff., though here in the Typhoeus episode the notions of ‘ earth’ and ' Earth ' are hard to distinguish, and all the editors print γαῖα with a small initial letter. We may mention here an example from the Works and Days not included in our list, but where the same process of extension may be seen at work: the phrase γένος μερόπων ἀνθρώπων | occurs at Op. 109, 143 and 180, beside φύλακες

θνητῶν ἀνθρώπων | at Op. 123 and 253.

Whatever reasons

determined Hesiod’s choice here (e.g. variety, avoidance of successive anapaests φύλακες μερόπων, sensitivity to meaning, or mere chance !), his usc of these epithets certainly differs from that found in our texts of Homer.

There, the criterion is

purely and simply a metrical one : θνητῶν is used wherever a double consonant is needed to supply weight to a preceding short vowel, and μερόπων is uscd wherever there is no such requirement. Thus: -V

-V -VC

θνητῶν

ἀνθρώπων

| A 339, 2 404, Y 204, 220, 233,

a 219, e 32, ἡ 247, ı 521. μερόπων

ἀνθρώπων

| A

250,

Γ

402,

I 340,

A

28,

Σ 342, 490, Y 217, v 49, 132. Whether such examples as this must necessarily be interpreted as evidence for writing in the poem’s composition is a question to which we shall return at tho end of this chapter. The possibility of ἃ similar explanation must be admitted 16 On γαῖα πελώρη as a Hesiodic innovation see Troxler,

SW

H, 174-82.

THE

PRINCIPLE

OF

ECONOMY

67

wherever a doublet consists of alternative expressions whose metrical values are not wholly identical, even though the differences between them are irrelevant in the contexts where we now find them. This applies to some of the doublets for which an explanation has already been offered under one of the other heads, as Th. 223/Op. 82 θνητοῖσι βροτοῖσι [ἀνδράσιν ἀλφηστῇσι and Th. 713/Sc. 261 Eyeıpav/&devro, as well as to several of those not yet considered, including Th. 138/ 155 θαλερόν σφετέρῳ, Th. 441/Sc. 104 ἐρικτύπῳ Πταύρεος, Th. 443/446 ἐθελουσά γε θυμῷ "θυμῷ γ᾽ ἐθέλουσα, Th. 636/ 712 συνεχέως ]ἐμμενέως, Sc. 333/425 ἐπιόντα "προσιόντα (with which we may compare Homer's ἐὼν ἐπιόντα O 164, beside ὁ δὲ προσιόντα N 615, both with the participle at the same point in the line as in Sc.), and Sc. 424/448 ταλακάρδιος /Üpaavkdpówv. If more early hexameter material had survived, we might have been able to see how some at

least of these alternatives arose in contexts which were metrically distinct. (d) Extension by analogy n miscellaneous other ways, notably by retaining words in a particular place in the line.

(i) Th. 57/302 and Op. 169. | νόσφιν ἀπ᾽ of Th. 57 appears to be traditional (ef. | νόσφιν ἀπό E 322, Καὶ 416); but the alternative | τηλοῦ ἀπ᾽ might easily have been suggested by τηλοῦ ἐπί + dat. A 712,

V 853, and

| τηλοῦ ὑπέρ

-- gen. v 257.

(ii) Th. 218-9/905-6. Both alternatives involve the use of relatively common words in familiar positions. The former contains βροτοῖσι | at the end of the line (so too 11x J1., 13x

Od.) ; and with the following | γεινομένοισι we may compare | γεινομένῳ ἡ 198. The latter alternative on the other hand has διδοῦσι | (so too B 255 and 4x Od.); and | θνητοῖς ἀνθρώποισι, which occurs twice in the Homeric Hymns, is eomparable

with θνητῶν ἀνθρώπων |, at the end of the line, 5x JL, 4x Od. (iii) Th. 443/446. The first alternative has θυμῷ | in its

commonest place (here 93 x Il., 103 x Od.), as well as ἐθέλουσά ye in the same place as at o 449. Th. 446 on the other hand has the participle at the end of the line, where it is found elsewhere

68

THE

LANGUAGE

OF HESIOD

in Homer, e.g. ἐθελούσῃ | Z 165, B 50, e 155, ἐθελούσης | 2 289. (iv) Th. 625/634. Both εὐνηθεῖσα and ἐν φιλότητι occur at the

end of the line in Homer, ἐν φιλότητι | 3x IL, 1x

Od. (and

earlier in the line 4x 11.), and εὐνηθεῖσα | 2x Il, in the phrases θεὰ βροτῷ εὐνηθεῖσα | B 821 and γυνὴ θεῷ εὐνηθεῖσα | II 176. The influence of these last two lines on Hesiod 1s perhaps to be seen in the rather pointless θεὰ θεῷ of Th. 380, where the whole line runs καὶ Νότον, ἐν φιλότητι θεὰ θεῷ εὐνηθεῖσα. (v) Sc. 379/412. Sc. 412 seem to be traditional here, since the complete line occurs at ΠῚ 430, and the second half of it at & 401; but the alternative expression at Sc. 379 ds oi ἐπ᾽ ἀλλήλοισι πέσον μέγα κεκλήγοντες is reminiscent in its first half of | of πρὶν ἐπ’ ἀλλήλοισι φέρον I” 132, and κεκλήγοντας | occurs at μ 256. (vi) Sc. 414/453. The latter’s | κακτάμεναι μεμαώς is closer to Homer’s | τὸν κτάμεναι μεμαώς E 301, P 8, but the

alternative | κτεινέμεναι μεμαώς may easily have been created on the analogy of ἑλκέμεναι μεμαώς 2 156, | νηχέμεναι μεμαώς ε 375, also | κτεῖναι Z 167 and | κτεινόpev(os) 4x Il. The foregoing detailed discussion has been necessary in order to show how the formulaic style of composition itself may give rise to what Parry calls a ‘surproduction’ of expressions,!? developed by this process of analogy, sometimes from formulas which were originally of unique value, sometimes from other closely related or similar-sounding phrases. Of the remaining examples in our list, several must be regarded as indicating not so much a departure from the economy of a formulaic system, but rather the non-cxistence of any true formula of the appropriate meaning and metrical shape. Extensive and complex as the system of hexameter formulas was, it clearly did not supply the entire necds of a Homer or a Hesiod, so that for certain ideas, both poets are still at a 1 ETH, 227.

THE

PRINCIPLE

OF

ECONOMY

69

stage of experimentation, where alternatives can arise through two original and independent efforts of composition. Here

surely

we

must

include

Th.

150-1/671-2,

Th.

636/712,

Op. 678/682, Op. 783/788, 794, and Op. 801/828. In these examples, neither of the alternatives can be shown to be any more traditional than the other, or any more appropriate in its context, except possibly for the last example, where ὄρνιθας κρίνων may well have been preferred for the sake of its assonance with the phrase ὑπερβασίας ἀλεείνων at the end of the line and of the poem. One of our difficulties in this matter of economy is to know just when two words or phrases are mutually replaceable and when they are not. This is especially true of certain instances in our list, where the poet’s choice may have been directed by the appropriateness of a particular idea in one of the contexts. At Th. 155 (beside Th. 138) σφετέρῳ lays special stress on the closeness of the relationship between Ouranos and his children, and so heightens the horror of his hatred for them (children usually love their parents, but not so much as parents love their children ; at all events ÜaAepóv at Th. 138 seems to put Kronos' hostility towards his father in a better light than

σφέτερον would).

At Th. 195 ῥαδινοῖσιν serves to direct our

attention to the beauty of the newly-formed goddess (cf. καλὴ θεός in the previous line and the tenor of what follows down to line 206);

here. to

ἀθανάτοισιν or λιπαροῖσιν would be flat

Then at Th. 622 μεγάλης emphasises the great distance

the

limits

of the

earth

in μεγάλης

ἐν πείρασι yains,—

a notion which is absent from Th. 300 and 483 where the epithet ζαθέης is used.!? It may even happen that this sensitivity is apparent in one place and not in another: Th. 148 has οὐκ

ὀνομαστοί | beside

Th.

310, Sc. 144 and

161 οὔ τι

dare (ós ) |, i.e.‘ not to be named ’ beside ‘ not to be spoken of’. The latter is not found in Homer, but where the former occurs it is in a strictly literal sense : Κακοΐλιον οὐκ ὀνομαστήν

18 ζάθεος in Homer (7x Il.) is always an epithet of places: p. 63 on μητέρα μήλων Th. 284.

| T 260,

cf. above,

70

THE

LANGUAGE

OF HESIOD

597, 5 19—the name itself is uttered only in a distorted form.

We may contrast Hesiod's οὐκ ὀνομαστοί | at Th. 148, followed at once by the names of the three giants [Κόττος re Βριάρεώς Te Γύης θ᾽... «τὰ Th. 149. This absurdity is avoided when the poet comes to Kerberos at Th. 310, and οὔ τι φατειόν | is used instead.!? Of our original list which included over thirty examples of alternative expressions, we are left with only a handful of instances which contain unexplained departures from a known and otherwise regular formula, and so may point to the deliberate composition which the use of writing makes possible. They are: | (i) Th. 818 βαρύκτυπος ᾿Εννοσίγαιος | beside the expected Ποσειδάων ἐνοσίχθων | Op. 667, 14x Il., 10 x Od. βαρύκτυπος is not

a Homeric

word,

and

wherever

it occurs

elsewhere

(Th. 388, Op. 79, Sc. 318, also at h. Dem. 3, 334, 441, 460) it is an epithet of Zeus. Hesiod certainly has no obvious need for

it here:

in addition to Ποσειδάων ἐνοσίχθων | he has at his

disposal ἐρίκτυπος ᾿Εννοσίγαιος

| (cf. Th. 441, 456, 930) which

could have been used at T. 818 after an ephelcystic -v, . ποίησεν épikrvmos ' Evvoatyatos. (i) Se. 70 ᾿Απόλλωνος Παγασαίου | beside either ἑκηβόλου Ἀπόλλωνος | Th. 94 and 2x Il. (also acc. and dat.) or Ἀπόλλωνος ἑκάτοιο | 2x Il. (i) Sc. 104 καὶ ταύρεος ' Evvootyatos | beside καὶ épirvm (ov) ᾿Εννοσίγαι(ον) | Th. 441, 456, 930. These last two special epithets apparently have reference to local cults of these gods in the Greek mainland, and are totally foreign to the Homeric tradition. This is the only reference to Poseidon in the Shield, but we may note that the poet is elsewhere content to refer to Apollo with a regular traditional epithet (cf. especially Sc. 58, 68, 100).

(iv) Sc. 102 ἀμώμητος ’IoAaos | beside ἀρηιφίλῳ ᾿Ιολάῳ | Th. 317. It is the Theogony which has the expected epithet for 1? On this type of expression as being not descriptive but emotive, cf. the remarks of H. Lloyd-Jones on Eur. Hipp. 875 in JHS Ixxxv, 1965, 169.

THE

this place; metrically

PRINCIPLE

OF ECONOMY

71

lolaos is not mentioned in Homer, equivalent

ἀρηΐφιλος

Μενέλαος

| 8x

but cf. the IL,

1x

Od.

(also in other cases 12x II.) and ἀρηΐφιλος Λυκομήδης | 1x

Il., whereas the epithet ἀμώμητος oceurs only once in Homer, at M 109, and there in the phrase Πουλυδάμαντος ἀμωμήτοιο

πίθοντο |. Perhaps the poet sensed that *dear to Ares? would be a singularly inappropriate epithet for Iolaos in the Shield, especially at this point (cf. the words of Iolaos at Sc. 109 ff.). (v) Sc. 192 ἐναρσφόρος οὔλιος "Apos | beside either "Apns datos πολέμοιο |, as at Sc. 59 (acc.), or "Apns ἀκόρητος ἀυτῆς |

Sc. 346.20 ἐναρσφόρος is untraditional, and οὔλιος is used in Homer only at A 62, where οὔλιος ἀστήρ | appears to refer to Sirius ; cf. however οὖλος " Apns E 461, acc. E 717. The whole phrase seems to be chosen here in order to put Ares in as poor a light as possible, the story being told from the point of view of Herakles. (vi) Sc. 424 Διὸς ταλακάρδιος vids | beside Atds θρασυκάρδιον

υἱόν | at Sc. 448. The epithet ταλακάρδιος appears here for the first time, whereas θρασυκάρδιος occurs twice in Homer, at K 41 and N 343, both times in this place in the line. A possible explanation for the doublet has been suggested above under 3 (c), but the Homeric references do not require the double

consonant initial for the metre, and so lend no support to that There is no motive apparent in the different suggestion. contexts of the two lines which would account for this alternation. It is clear that we must have some reservations in counting even these

six examples

as true departures

from

economy,

for not all the new expressions provide an exact equivalent of a traditional phrase, and it is arguable that some of them could justifiably be included in one or other of the above classes of explicable doublets. Perhaps the most noteworthy thing is the prominence in the list of examples from the Shield containing novel epithets. Attention may now be drawn to a final point which must 20 On these alternatives see above, p. 63.

72

THE

LANGUAGE

OF HESIOD

make us even more cautious in using the presence of these alternative expressions as evidence for the use of writing in the poems’ composition. We know nothing whatever about the circumstances of the first writing down of Hesiod's poetry, or the earliest stage in the transmission of a written text. Yet this is a matter of crucial importance in the interpretation of Their these departures from the principle of economy. explanation may lie in the poet’s use of writing as he composed ; but it may equally lie in a desire to improve upon the text as it was being dictated or written out.?! The novel epithets which we have just observed in the Shield could Furthermore, if these have entered the text in this way. poems were first orally composed and later written down or dictated, either by the poet himself or ἃ rhapsode, casual variations like many of those which we have examined could have been introduced at this later stage without affecting the structure of the poem as & whole. For example, the doublet ἄπλαστοι πᾶσιν ὁμῶς at Th. 151/672 results from the whim of the poet or reciter at the moment of committing the poem to writing ; it is precisely the sort of variation which can be expected to occur in two separate recitations of the same oral The observance of the principle of economy thus poem. depends on the circumstances of a poem's recording as much as on those of its composition. This chapter has done little more than to scratch the surface of the subject of economy, but enough has been said to show (1) that alternative expressions exist quite widely in Homer as well as 1n. Hesiod, (2) that many of these departures from economy could have arisen in the ordinary creative process of oral verse-making, and (3) that others could have entered the text at ἃ date later than the poems' composition. The problem 18 by no means a simple one, and nothing short of a full-scale study embracing Homer, Hesiod and the Homeric Hymns, and considering the evidence of modern oral literatures, can 33 cf. A. B. Lord, The Singer of Tales, 127, on the way in which a dictated

text might be altered if it was felt to be ‘not right’ at the first attempt.

THE

PRINCIPLE

OF

ECONOMY

73

be expected to grapple with it successfully. Meanwhile, our only safe conclusion is that, while the observation of the principle of economy is a good indication of the work of an oral poet, the neglect of that principle need not denote the use of writing in a poem's composition. In the light of this conclusion, the examples of neglect of economy which we have found in Hesiod become less important, whereas the high degree of economy which we noted (in Chapter IV) in his noun-epithet formulas for ‘the sea’ takes on an increased significance.

CHAPTER * PARALLELS

VI

OF

SOUND ’

THE presence in Hesiod of many Homeric phrases, to which we have referred earlier,! need not itself be evidence of oral

composition. The poet might have learnt them orally and yet himself have composed with the aid of writing, or he could even have derived his knowledge of them from a written text. There exists however a less obvious type of parallel which 1s relevant to the problem here, and which certainly makes this latter possibility unlikely. A good example occurs at Works and Days 441: Hesiod has been talking about different kinds of wood and the uses to which they may be put in making tools. This leads him to mention different parts of the plough, and from there to the best sort of team for drawing 1t (Op. 436-40): he recommends a pair of nine-year-old male oxen. Now he turns to the best sort of man for driving them: he should be forty years old and in full strength—rois δ᾽ ἅμα τεσσαρακονταετὴς αἰζηὸς ἕποιτο.

does

not oceur in Homer,

The word τεσσαρακονταετής

though

similar

compounds

found, namely rpı-, zevra-, ἐξ(α)-, érra- and evva-erns.

are But

the numeral τεσσαράκοντα occurs nine times in the Iliad, all in the second book, and all in the same context, enumerating the contingents making up the Greek fleet—rotis δ᾽ ἅμα τεσσαράκοντα μέλαιναι νῆες ἕποντο B 524 and 747, and the same, but with τῷ for rots, B 534, 545, 630, 644, 710, 737 and

759. There is too much in common between Op. 441 and this repeated Homeric line for the correspondence to be accidental. Yet the similarity 15 not to be explained as resulting from the re-use of a traditional phrase, nor even from the deliberate adaptation of an old expression to a new context ; it suggests rather a more subconscious process at work, with the poet’s memory, not so much of sense but of sound and rhythm, playing an important part. ! In Chapter III, esp. p. 27 with n. 4.

‘PARALLELS

OF SOUND’

75

We now list some further examples of a similar kind. Not all these * parallels of sound ’ are so striking as the one which we have just cited, and some might be dismissed as having arisen fortuitously ; but the chance of all or even most of them having so arisen is too small to be credible. Th. 30 καί μοι σκῆπτρον ἔδον δάφνης ἐριθηλέος ὄζον Homer has the phrases μυρικῆς τ᾽ ἐριθηλέας ὄζους | Καὶ 467

and ἔρνος “- ἐριθηλὲς ἐλαίης | P 53. Hesiod keeps the adjective in the same place, though he uses it as an epithet of the tree, not of the ὄζος. But we may compare also καὶ Τιτυὸν εἶδον, Γαίης ἐρικυδέος υἱόν A 576. The only word common to

Th. 30 and X 576 is | «al; yet the structure of these lines is so similar that it seems legitimate to ask here, no less than with the more obvious parallels such as those discussed below in Chapter XI, whether they can be totally unrelated, and if not, which is likely to be the earlier. Certainly there is no need to suppose that the author of A 576 was thinking of the Hesiodic line: his line-beginning corresponds to a pattern familiar in A:

cf. esp.

| καὶ XA@pw

εἶδον A 281,

| kai 4ήδην εἶδον A 298; with the second half of his line ef. ἥβης ἐρικυδέος A 225, Μητοῦς ἐρικυδέος & 327, and from

the Hymns Mains ἐρικυδέος υἱός | h.Herm. 89, 550 (vie |), Anroös ἐρικυδέος vids | h. Apol. 182, h.

Herm. 176, 189, Anroös

δ᾽ ἐρικυδέος vióv | h.Herm. 416, Σεμέλης ἐρικυδέος vidv | ἢ. vii,

1;

with

the

sound

of this

line-ending

cf. also

the

common formula for Diomedes, Τυδέος υἱός | 16x Il., and esp. μεγαθύμου Τυδέος vids | E 25, 235, 335, Καὶ 509 (vie |) ; and for the link between the two halves of the line in A 576 cf. Τιτυόν, Γαιήϊον viov | ἡ 324. Hesiod on the other hand is here ostensibly describing a personal experience, namely his encounter with the Muses ; yet he does so in language which is steeped in traditional forms of expression: for δάφνης

ἐριθηλέος ὄζον | see above;

for the gift of a σκῆπτρον as a

symbol of authority (as commonly in Homer) cf. esp. B 205 f., 1 98 f. where the giver is Zeus; cf. also δῶκε δέ

οἱ σκῆπτρον, of the beggar's staff given by Athene

to

THE

16

Odysseus, v 437.

LANGUAGE

OF HESIOD

No doubt Hesiod could have composed this

line without knowledge of A 576, but if we concede that he

might have known it, this provides a very plausible explanation of their similar sound and structure. Th.

105 (also 21) xAeleve

δ᾽ (ἄλλων v' 21) ἀθανάτων

ἱερὸν

γένος αἰὲν ἐόντων

Again the line-ending is familiar from Homer (see above p. 24), where however the phrase αἰὲν ἐόντες, In whatever case, always agrees with θεοί, -ovs, -ῶν rather than with the tautologous ἀθάνατοι which we find in these Hesiodic lines. But for the structure of the whole line cf. ἐλθεῖν ἐς φυλάκων ἱερὸν τέλος ἠδ᾽ ἐπιτεῖλαι K 56, and esp. χεύαμεν Ἀργείων ἱερὸς στρατὸς αἰχμητάων w 81. Th. 151 and 672 κεφαλαὶ δὲ ἑκάστῳ πεντήκοντα |

As in the example from Op. 441 cited at the beginning of this chapter,

Hesiod

uses here,

in his description

of the

hundred-armed monsters, an expression parallel to one found in the Homeric Catalogue of Ships: ἐρέται δ᾽ ἐν ἑκάστῃ πεντήκοντα | B 719; cf. also ἐν δὲ ἑκάστῃ | πεντήκοντ᾽ ἔσαν ἄνδρες II 169 f., and ἐν δὲ ἑκάστῳ | πεντήκοντα σύες

£ 14 f.

Th. 263 αὗται μὲν Νηρῆος ἀμύμονος ἐξεγένοντο The Iliad has the lines δώδεκα γὰρ Νηλῆος ἀμύμονος υἱέες ἦμεν A 692 and φασὶ σὲ μὲν IlmAnos ἀμύμονος ἔκγονον εἶναι Y 206;

cf. also AyiAjos ἀμύμονος in this position 3x

Il. Nereus is never mentioned by name in Homer, but the “parallel of sound’ here is clear; ef. also τρεῖς παῖδες ἀμύμονες ἐξεγένοντο | & 115, Y 231. Th. 218 τῇ δὲ μιῇ παρελέξατο Κυανοχαίτης |

The form παρελέξατο occurs in “αοδαμείῃ μὲν παρελέξατο μητίετα Ζεύς Z 198 and in the same position in the line at 4) 676, À 242; κυανοχαίτης is found at the end of the line and immediately after its verb at Y 144 and ı 536. Both words occur at Y 224:

ἵππῳ δ᾽ εἰσάμενος παρελέξατο κυανοχαίτῃ,

' PARALLELS

OF SOUND’

17

where the subject of παρελέξατο is Boreas, and κυανοχαίτῃ is an epithet agreeing with ἵππῳ (cf. Sc. 120). Th. 319 ἣ δὲ Χίμαιραν ἔτικτε πνέουσαν ἀμαιμάκετον πῦρ

The word ἀμαιμάκετος is used only once in Od., of the mast of a ship (£ 311), and twice in II. (Z 179, IT 329), both times of the Χίμαιρα. This association must have suggested its use to Hesiod here, but he applies the adjective not to Χίμαιραν but to πῦρ, being perhaps prompted to do so by

the similar-sounding ἀκάματον πῦρ | 1 x Il.,

2x Od.

Th. 696 τοὺς 8’ ἄμφεπε θερμὸς ἀυτμή |

This is the only place where the form ἄμῴφεπε occurs in Hesiod ; in Homer it is found three times, always in this position, in the phrase πῦρ (δ᾽) ἄμφεπε, and two of these occurrences are in the half-line πῦρ δ᾽ ἄμφεπε, θέρμετο δ᾽ ὕδωρ | Σ 348, 0 437. For the last two feet cf. also ἄμπνυτο δὲ θερμὸς ἀυτμή | h.Herm. 110 (ἀνὰ δ᾽ ἄμπνυτο in most mss.). Th. 126 τὸν πέρι χάλκεον ἕρκος eAhAatar ἀμφὶ δέ μιν νύξ The only two occurrences of the form ἐλήλαται in Homer are at II 518, where it stands in this position, following ἀμφὶ δέ μοι χείρ | at the end of 7517 (the passage is quoted below, p. 78, with reference to Op. 214-6), and at n 113, where the line ends

περὶ

δ᾽ ἕρκος

ἐλήλαται

ἀμφοτέρωθεν

| ; cf. also

περὶ δ᾽ ἕρκος ἔλασσε | 2/564. Hesiod's epithet χάλκεον here may well have been partly suggested by the frequent phrase χάλκεον ἔγχος (18x IL,

5x Od.);

cf. esp. διὰ χειρὸς ἐλήλατο

χάλκεον ἔγχος | Ν 595. Th. 189 δεκάτη δ᾽ ἐπὶ μοῖρα δέδασται | The form δέδασται occurs at A 125 and ο 412, ın both at

the line-end ; but for the whole phrase cf. τριτάτη δ᾽ ἔτι μοῖρα λέλειπται | K 253.

Th. 827 θεσπεσίῃς neparfjoıv ὕπ᾽ ὀφρύσι πῦρ ἀμάρυσσεν

The phrase ὑπ᾽ ὀφρύσι occurs in this position 3x IL, 4x Od.; but among these cf. esp. βλοσυρῇσιν ὑπ’ ὀφρύσι O 608, and κυανέῃσιν em’ ὀφρύσι A 528, P 209, where the

178

THE

LANGUAGE

OF HESIOD

preceding -now belongs to an epithet agreeing with ὀφρύσι, not to an independent phrase as in Hesiod. Th. 980 μειχθεῖσ᾽ ἐν φιλότητι πολυχρύσου ᾿Αφροδίτης The epithet πολύχρυσος in Homer is applied once to a wealthy man (K 315) and otherwise only to wealthy cities, namely Troy (Σ᾽ 289) and Mycenae (H 180, A 46, 304). Its application to Aphrodite by Hesiod here and at Op. 521, Sc. 8 and 47 (also at h.Aphr. 1 and 9) is an extension of the Homeric use of χρύσεος with Adpodirn in this position, e.g.

gen. χρυσέης Adpodirns | I' 64, 8 14, so also Th. 975, with ‘golden’ used in a metaphorical sense. Homer only once requires an epithet of this metrical pattern to go with Ἀφροδίτη, and uses Aids κούρης Ἀφροδίτης

| (Y 105);

of.

also φιλοστεφάνου Ἀφροδίτης | at h.Dem. 102. Op. 214-6 ὕβρις γάρ τε κακὴ δειλῷ ῥηιδίως φερέμεν δύναται, ἐγκύρσας ἀάτῃσιν.

βροτῷ" οὐδὲ μὲν ἐσθλὸς βαρύθει δέ θ᾽ ὑπ᾽ αὐτῆς

Hesiod is warning Perses of the dangers of ὕβρις. Neither Rzach nor Krafft notes any Homeric parallel to these lines ; yet cf, from quite a different context, some words of Glaukos after he has been wounded : ἕλκος μὲν yàp ἔχω τόδε καρτερόν, ἀμφὶ δέ μοι χεὶρ ὀξείῃς ὀδύνῃσιν ἐλήλαται, οὐδέ μοι αἷμα τερσῆναι δύναται, βαρύθει δέ μοι ὦμος ὑπ᾽ αὐτοῦ II 517-9.

Op. 388-91 οὗτός τοι πεδίων πέλεται νόμος, οἵ τε θαλάσσης ἐγγύθι ναιετάουσ᾽, οἵ T ἄγκεα βησσήεντα, πόντου κυμαίνοντος ἀπόπροθι, πίονα χῶρον , ναιουσιν.

We are concerned only with line 390, but it is necessary to give the full context in order to show that the words πόντου Kupaivovtos ἀπόπροθι need to be taken together, and contrast with θαλάσσης ἐγγύθι earlier. Homer uses ἀπόπροθι six times (1x IL, 5x Od.); but the curious point here is that in no less than three of these six places ἀπόπροθι is

‘PARALLELS OF SOUND’

79

followed immediately by πίων and a noun signifying a place: cf. εἴ of καὶ μάλα πολλὸν ἀπόπροθι πίονες ἀγροί VU 832, and ὅς κεν ἔχῃσι | δώματά θ᾽ ὑψερεφέα καὶ ἀπόπροθι πίονας ἀγρούς ὃ 156 f., where Eurykleia is the speaker. The precise force of ἀπόπροθι is obscure in both these passages, but is very clear at ı 35 in a similar context where Odysseus 15 the speaker : ὡς οὐδὲν γλύκιον ἧς πατρίδος οὐδὲ τοκήων γίγνεται, εἴ περ Kai τις ἀπόπροθι πίονα οἶκον

γαίῃ ἐν ἀλλοδαπῇ ναίει ἀπάνευθε τοκήων. It seems certain that the origin of this group of phrases is to be found in this last passage or one now lost where the context was similar. Hesiod’s móvrov . . . ἀπόπροθι is an extension of the older adverbial use which is universal in Homer, but the second half of Op. 390 has been influenced by his knowledge of traditional lines. (On Op. 441 see the discussion above, p. 74.) Op. 446 | σπέρματα δάσσασθαι

The only place in Homer where the form δάσσασθαι occurs is in the phrase

| κτήματα

δάσσασθαι v 216, with which cf.

| κτήματα δασσάμενοι y 316, o 13, also | κτήμασι τέρπεσθαι I 400.

Op. 556 χρῶτα δὲ μυδαλέον θήῃ κατά θ᾽ εἵματα δεύσῃ The structure of this line, in a warning after μή, is closely similar to μή πως ἵππους τε τρώσῃς κατά θ᾽ ἅρματα ἄξῃς Ψ 341, but it presents also a “parallel of sound’ with ἀπὸ μὲν φίλα εἵματα δύσω | B 261, χιτῶνά τε εἵματ᾽ ἔδυσαν |

£ 341.

Op. 596 τρὶς ὕδατος προχέειν, τὸ δὲ τέτρατον ἱέμεν olvov

Again the sound and structure have Homeric parallels, and we may quote in particular τρὶς μὲν ὀρέξατ᾽ ἰὼν, τὸ δὲ τέτρατον ἵκετο τέκμωρ N 20, and τρὶς δὲ μεθῆκε βίης" τὸ δὲ

τέτρατον ἤθελε θυμῷ Φ 177. This line of Hesiod was thought by Wilamowitz (ad loc.) to begin with a tribrach ; but while

the vowel of τρίς is normally short (cf. ı 65), the frequency

80

THE

LANGUAGE

OF HESIOD

with which the word stands before a consonant, especially in the emphatic position at the beginning of the line (21x Il. out of 30 occurrences, 9X Od. out of 13), has caused Hesiod to use it here as a heavy syllable in its own right.

A phrase of the type | τρὶς (F-)éreos Op. 173 may have provided a ' bridge ’ to the use of τρίς as a heavy syllable. Op. 624 νῆα δ᾽ En’ ἠπείρου ἐρύσαι πυκάσαι τε λίθοισι

The relationship of this line to Homeric lines on the same subject is obvious, e.g. vja . . . em’ ἠπείροιο ἔρυσσαν | A 485 and z 359; and for the end of the line cf. πυκινοῖσι λίθοισι |

II 212. Less obvious but of greater interest to us here is its relationship with ἃ Homeric passage where Odysseus and Telemachos take vengeance on the wretched Melanthios : they tie him to a rope and ‘ haul him up the high pillar and make him touch the rafters’. The only common thought in this and the Hesiodic line is what is contained in the verb ἐρύω: κίον᾽ ἀν᾽ ὑψηλὴν ἐρύσαι πελάσαι τε δοκοῖσιν x 176, and κίον᾽ av’ ὑψηλὴν ἔρυσαν πέλασάν τε δοκοῖσιν x 193.

Op. 653 Ἑλλάδος ἐξ ἱερῆς Τροίην ἐς καλλιγύναικα Hesiod’s choice of epithets in this line is remarkable.

In

Homer καλλιγύναιξ is used five times, twice with Ελλάδα (B 683, I 447), twice with Ἀχαιΐδα (Γ 75, 258) and once with

Σπάρτην (v 412); ἱερός on the other hand, though used of many places, some of which are on the Greek mainland, is more frequently applied to Troy than to anywhere else.? One is tempted to see Hesiod's sense of humour at work here, deliberately choosing the ‘wrong’ epithets, for he could easily have said ' "Duov εἰς ἱερὴν ἐξ Ελλάδος e)pvyópowo " (cf. H 20, 1 478).

simply be that fluenced less by to him in which which in Homer line (4x Il, 1x

On the other hand the explanation may

Hesiod's choice of epithets has been intheir meaning than by phrases familiar they are used, especially by καλλιγύναικα is always acc. and always at the end of the Od.): cf. esp. Σπάρτην es καλλιγύναικα | * See Ebeling, LH I, 585 f.

‘PARALLELS

v 412:

OF SOUND’

8l

and for ἱερῆς in this position cf. | νηδύος ἐξ ἱερῆς

Th. 460, also | Tipuv® εἰς ἱερήν Th. 292. Op. 815 καὶ ἐπὶ ζυγὸν αὐχένι θεῖναι |

While ἐπιθεῖναι in Homer always stands at the end of the line (4x Il., 1x Od.), the uncompounded θεῖναι occurs only in other positions (2x Il., 1x Od.) ; cf. however—not from τίθημι but from θείνω — ὁ δ᾽ ὑπ᾽ οὔατος αὐχένα θεῖνε | IT 339, “he smote the neck beneath the ear’, and similarly ὁ δὲ φασγάνῳ αὐχένα θείνας | Y 481. It is to be noted that,

although ἐπιτίθημι is commonly constructed with a dative (e.g. κεφαλῇ δ᾽ ἐπέθηκε καλύπτρην x 545), αὐχένι at Op. 818 is only a conjecture by Hermann, adopted into their texts by Kirchhoff, Rzach, Evelyn-White,

and and well lines

Mazon,

Sinclair

Colonna. The acc. αὐχένα 15 found in the older editions, is the reading of all the mss. {10 is correct, Hesiod may have been prompted by the sound of these Homeric to depart from the regular construction.

Sc. 267 | τῆς ἐκ μὲν ῥινῶν μύξαι ῥέον

The word μύξαι does ποῦ occur in Homer, but cf. | τῶν ἐκ μὲν μελέων τρίχες Eppeov βέλεα ῥέον M 159.

x 393, and

| ὡς τῶν ἐκ χειρῶν

Sc. 286 [νῶθ᾽ ἵππων ἐπιβάντες

In Homer ἐπιβαίνω is followed by an accusative only in the phrase Πιερίην δ᾽ ἐπιβάς (-Baoa) & 226 and e 50, where the meaning is ‘to move across’, ‘traverse’. In its more common meaning ‘to climb upon’, ‘mount’, the verb is invariably followed by a genitive:? cf. esp. ὧν ἵππων ἐπιβάς E 328, νύξ᾽ ἵππων ἐπιβησόμενον E 46, II 343, ἵππων ἐπιβαινέμεν E 255, ἵππων | ὠκυπόδων ἐπέβησε © 128 f. and ἵππων ἐπεβήσετο Καὶ 513. The word-order at Sc. 286 helps its unusual syntax to pass almost unnoticed. Sc. 358 συνοισόμεθα πτολεμίζειν

|

The construction of ovuderw with an infinitive does not occur anywhere else (except of course in the impersonal 8 Cunliffe,

LH D. 144, lists 16 such examples from Il. and 18 from Od.

82

THE LANGUAGE

OF HESIOD

construction, but even this is not in Homer). The author must have had in mind συνοισόμεθα πτόλεμόνδε | Θ 400, but the similarity of sound has caused him to confuse the phrase with ἐπιστάμεθα πτολεμίζειν | N 223.

Sc. 378 nos πεδίονδ᾽ ἀφίκωνται | In Homer ἀφικνέομαι is never followed by a word in -δε, though the construction is found at Sc. 38: ἀφίκετο ovde δόμονδε. But cf. the similar-sounding xareßav, πέδιον δ᾽ ἀφίκοντο

| 2 329.

These twenty-four examples vary widely in their nature, but all have this in common : their relationship to traditional phrases consists not so much in their subject-matter—here we are not dealing with expressions of any ‘ certain essential idea ’—, but rather in the sound-patterns which they contain, and which are nearly always associated with a fixed place in theline. One is particularly impressed by the ease with which Hesiod preserves the structure of an epic phrase in a completely fresh context, as at Th. 105 (the gods), Th. 151 and Sc. 267 (supernatural creatures), Op. 214-6 (abstract ideas), Op. 441 and 815 (agriculture), and Op. 624 (seafaring). The Homeric parallel in each case concerns a subject of a military nature, except for the last one, which is about a personal act of violence. If the Works and Days 1s the richest of the three poems in examples of this kind, the fact is perhaps to be explained by its wider variety of subject-matter. From the point of view of language study, those examples are of special interest which show some peculiar feature in syntax, as at Op. 390, 815, Sc. 286 and 358, or in meaning, as at Th. 319, 980 (with Op. 521, Sc. 8, 4T) and Op. 653, or

in prosody, as at Op. 596.* Yet in all these places one is bound to feel that neither the poet nor his audience would be much conscious of the peculiarity, so familiar is the sound of the phrase as a whole.» This is also true of certain inflexional * Some further examples of this kind of prosodical anomaly will be discussed below in Chapter VII, pp. 90-3. 5 The same applies to some phrases noted by Kretschmer where formulas

repeated by Hesiod himself do not properly fit their contexte, as at Th, 555

‘PARALLELS

OF SOUND’

83

innovations which make their appearance in Hesiod. Thus at Th. 119 the need to connect the phrase to what has gone before

prompts the poet to use the ntr. pl. | Τάρταρά τ᾽ ἠερόεντα,

modelling it on the regular masc. sg. Τάρταρον ἠερόεντα found at the beginning of the line in Th. 682 and at the line-end in Th. 721 and © 13.6 The Homeric Ἀπόλλωνα χρυσάορον Ο 256 has the epithet as an o-stem (so too gen. at E 509 and h.Apol. 395), and this is found also in line 3 of the dubious Hesiodic fragment 265 Rz. (= 357 M-W): Φοῖβον Ἀπόλλωνα χρυσάορον, ὃν τέκε “ητώ ; but at Op. 771 Hesiod needs to use the phrase before a consonant, and he modifies it by making the epithet athematic: τῇ γὰρ Ἀπόλλωνα χρυσάορα γείνατο “ητώ (similarly at h.Apol. 123). Conversely the name of Tiryns, which appears elsewhere as nom. Tipvvs, gen. Τίρυνθος (B

559,

Th.

292)

1s found

as an

o-stem

at

Sc.

81, where

Τίρυνθον ἐυκτίμενον πτολίεθρον | may be compared with Κόρινθον ἐυκτιμένας τε KAeovds | B 570. Such examples make one wonder whether ἀγήρω should not after all be retained at beside Th. 540 and Th. 602 beside Th. 585 (DIH, 37 f.). We may mention here also Hesiod's use of μήδεα in two quite different senses in the phrase

φίλου δ᾽ ἀπὸ μήδεα πατρός | at Th. 180 beside φίλου διὰ μήδεα πατρός | at Th. 398 ; in the one place μήδεα refers to the genitals οὗ Ouranos, and in the other to the counsels of Okeanos. The latter is a curious passage : Okeanos is not mentioned elsewhere as offering advice to his daughter Styx, and one cannot help wondering whether the notion behind μήδεα πατρός in the mind of the poet was not rather the scheme of Zeus, who has just been mentioned in Th. 390 ff. as summoning the gods to Olympos in order to place his proposal

to them.

If so, φίλου in T'^. 398 is quite inappropriate, but can be explained

as having originated in the same way as the peculiarities under discussion here. ὁ The idea that Τάρταρα arose here because Taprapdv τ᾽ ἠερόεντα is mctrically impossible, simple as it is, seems preferable to the more elaborate theories of K. Scherling in RE IVA (1932), 2444, attributing Τάρταρα in Hesiod to Orphic influence, and of Schwenn, TdH, 18 f., suggesting that the

ntr. pl. is the older form. A progression from ζόφον ἠερόεντα | (3X 1]., 3 x Od.) to Τάρταρον ἠερόεντα | to | Táprapov ἠερόεντα to | Táprapá τ᾽ ἠερόεντα seems more likely than the reverse order of development. The phrase ráprapa γαίης | Th. 841, which Schwenn regards as ‘der eigentliche Ausgangspunkt' for the idea of the masc. Τάρταρος, may simply have arisen there by analogy with πείρατα γαίης | E 200, 301, 8 563, Op. 168, ἔσχατα γαίης]

Th. 731 and other similar expressions.

84

THE

LANGUAGE

OF HESIOD

Th. 949 in spite of its being ‘ quite anomalous’ (so West ad loc., preferring Peppmüller's ἀθάνατον καὶ ἀγήρων): while there can be no doubt that the original form of this word is ἀγήραος, thematic,’ familiarity with it in a contracted form, as in the phrase ἀθάνατος καὶ ἀγήρως | e 218, might have prompted Hesiod to create an athematic acc. sg. ἀγήρω on the analogy of acc. αἰδῶ beside nom. αἰδώς (as at Op. 324), acc. ἠῶ beside nom. ἠώς (as in mss. at Op. 574); cf. also ἀθάνατοι καὶ ἀγήρῳ Th. 277, with widespread v.l. ἀγήρως. It is noteworthy that a form ἀγήρω does occur twice in JL, in the phrase ἀγήρω

τ᾽ ἀθανάτω ve | M 323, P 444: although this is a dual, the mere sound of this phrase could have been an additional factor in suggesting ἀγήρω to Hesiod at Th. 949. The parallels and near-parallels which we have been examining in this chapter are best explained as arising in the work of someone who had a thorough knowledge of traditional epic poetry, acquired not from perusing a written text at the time he himself was composing, but from having learnt it off by heart, so that its sounds and rhythms were constantly running through his mind and even being uttered aloud through his own mouth. Milman Parry stresses the importance of the voice and ear in memorising words, and makes the point that “as a rule we are unable to recall a single phrase of the book we have read silently.’ 5 While this does not make it certain that Hesiod was himself an oral poct, there is no doubt that thesc 'parallels of sound' might very plausibly be accounted for as ἃ consequence of oral composition. But before drawing conclusions on that question we will assemble evidence from other angles. ? See Chantraino,

G7

I, 54.

5 HSPh xli, 1930, 143. ? ef. Bowra’s brief discussion of somo similar ‘ parallols of sound ' within Homer (in A Companion to Homer, ed. Wace and Stubbings, 32), whore he makes the obsorvation that ' when a bard eomposes without the aid of writing, he has a keon sense of sound as well as of sense, and it is inevitable

that, just as one situation suggests another and provokes a similar form of words, so à sound-soqueneo may suggest another similar in rhythm or in tonal value or in balance of words.’

CHAPTER

VERSIFICATION

AND

VII

ENJAMBEMENT

Tue early history of the Greek language suggests that the development of the dactylic hexameter was prompted to some extent by dactylie rhythms in ordinary speech, and it is also well established that the language of hexameter poetry in its turn was influenced by the exigencies of the metre.? This close interrelationship of metre and language continued to be a characteristic feature of all Greek hexameter poetry, largely because of the dominant influence of the early epic. A. M. Dale rightly speaks of ‘ the essential similarity of all hexametercomposition of all periods, ἃ similarity which greatly outweighs the peculiarities distinguishing one poem or one author from another.’ ? Nevertheless it might reasonably be expected that poetry composed by an oral technique should show features of metre and versification which would distinguish it from written poetry, and we now consider some aspects of Hesiodic versification with this possibility in mind. Metrical studies of Hesiod in the nineteenth century * centred primarily upon two subjects: (a) prosody, ie. the value given to particular syllables for metrical purposes, and (b) the syllabic structure of the verse, 1.6. the relative proportion and distribution of dactyls and spondees in different poets and poems. On the first of these little needs to be said : all three Hesiodic poems, as well as other early hexameter poetry like the Homeric Hymns, adhere closely to Homeric practice,

where many features which might appear irregular can be explained on bistorical grounds ; ? there are however a number of places in Hesiod where, as in Homer, anomalies can be plausibly accounted for by reference to the practice of oral composition, and we shall return to this subject below.* The 1 See Bowra in A Companion to IIomer, ed. Wace and Stubbings, 23 f. 2 See esp. Witte in RE VIII, 1913, 2213-47. 3 Lustrum ii, 1957, 34.

4 See above, p. 18 n. 33. 5 West gives a concise summary of the salient facts, HT, 95-101. ? See below, pp. 90-3.

THE LANGUAGE

86

OF HESIOD

second line of inquiry, which examines the ‘ outer metric ’ of the hexameter line, reveals some distinctive features in the

Hesiodie poems, and it is worth giving a brief summary of the position. Although each of the first five feet of the hexameter line may consist either of a dactyl or a spondee, there are roughly three times as many dactyls as there are spondees in Homer, and this remains true in the Hesiodic poems and indeed in later poetry.” It has been shown however that the Works and Days contains a slightly higher proportion of spondees than the Theogony and the Shield, sharing this feature with certain of the Homeric Hymns, while the Theogony and the Shield stand very close to Homer in this matter.5 But the difference is not great, and it would be unwise to base any major conclusion upon it. Perhaps more significance may be attached to differences in the way in which these spondaic feet are distributed : it is characteristic of hexameter poetry of all periods to have a much higher concentration of spondees instead of dactyls in the first two feet than later in the line. Thus, in Homer and Hesiod alike, at least 35% of all first and

second feet are spondaic ; this proportion usually drops to less than 20% in the third foot, rises to between 25% and 30% in the fourth, and then drops sharply in the fifth foot to less than 1095.19 The infrequency of spondees in the fifth foot is indeed 7 From O’Neill’s Table 38 (YCS viii, 1942, 159), which is based on samples

of 1,000 lines, the following proportions of spondaic feet, as a percentage of all feet, may be calculated: JJ. 24-9, Od. 26-2, Hesiod (i.e. Op. + part of Sc.) 28:1, Aratus 26-7, Callimachus 21-9, Apollonius 22-4, Theocritus 27-7. ® Paulson's figures are: Th. 25-495, Op. 28-196, Sc. 25-594 (DRM, 5). The data which he gives for the Homeric Hymns (DRM, 6 n. 2) provide figures of 29-195 for h. Apol. and 28-295 for h.Aphr. ? At least part of the increase in Op. may be satisfactorily accounted for by the gnomic element in the poem: seo West, HT, 93.

10 See Paulson’stable, DRM, 6 ; his figures differ slightly from those given by West, ΜΊΤ, 93, which

are dorived from Ludwich's work of 1885.

This is

because Paulson based his count on Rzach’s texts of the Iliad (1886) and of Hesiod (1884), and this results in a consistently lower figure for spondees throughout. Exactly the samo pattern is discernible in the details given by O'Neill in his table cited above in n. 7.

VERSIFICATION

AND

ENJAMBEMENT

87

the most striking feature of the distribution, and it is noteworthy how consistent different poets appear to be in this matter! Hesiod shows a proportion of spondaic fifth feet which is higher than the average for either the Iliad or the Odyssey ;?? but while the Theogony and the Works and Days diverge only slightly from the Homeric average—no further in fact than some

individual books of Homer—,

the Shield

stands apart from the rest of Hesiod, its proportion of spondaic fifth feet being well above the highest figure found among any of the 48 books of Homer.!3 Another line of approach, pursued more recently, concerns the distribution of words of particular metrical shapes within H C. H. A. Ludwich in Quaestionis de hexametris poetarum Graecorum spondaicis capita duo, 1866, 9-13, gives figures for various authors and their separate books, from which the following percentages of spondaic fifth feet may be derived: Apollonius, Arg. a 8:995, B 9:095, y 7-395, 8 9:595, α-δ 8-795 ; Oppian, Hal. a 7-595, B 1-695, 8:895,8 1- 295,€ 8:495, a-e 1- 997; Pseudo-Oppian, Cyneg. a 11-295, B 10-895, y 13-195, ὃ 9-395, α-δ 11:295; Aratus, Phaen. 13:395, Diosem. 16: 495. Though based on editions which are now obsolete, these figures, and others which will be cited from Ludwich below, agree very closely with those given for various authors by O'Neill, loc. cit. 12 Paulson's figures are: Th. 6:395, Op. 5-895, Sc. 8:395, beside 4-395 for Iliad A (DRM, 6). Ludwich's figures (see last note) from Bekker’s edition also give 4:39; for A, but show that this is somewhat below the average for the Iliad as a whole (5:195), while that for the Odyssey is 4-895, ; his figures for Hesiod (op. cit., 10) are a little higher than Paulson's, most noticeably in Op., where he finds 56 examples (6:895) to Paulson's 48 (5:895) ; this is because he is using the edition of Goettling, who retains such readings as ἄτῃσιν Op. 352, 'Qpíwvos Op. 598, etc., which Rzach alters. It may be remarked that Goettling's practice here is closer than is Rzach's to that of Monro and Allen in their Oxford Text of Homer. 13 According to Ludwich's count (op. cit., 9) all the books of the Iliad fall within the limits of 3: 89/ to 7-395, the highest figure being in B (7-395) and the second highest in Y (7.095); the corresponding range for the Odyssey is a little wider, from 2-995 to 7-795, this upper extreme being reached in A, while v is second highest with 6-695. The somewhat higher figures which Ludwich gives in his later work (Aristarchs homerische Teatkritik ΤΙ, 1885, 327 f.) are mainly to be accounted for by his inclusion there of examples such as 'Apyeióóvrgsl, which most cditors print with a diaeresis on the second syllable; but this affects his result for the Hesiodic poems correspondingly, and as far as they are concerned the same conclusions emerge, with Sc. well above the highest figure for any Homeric book.

THE LANGUAGE

88

OF HESIOD

the pattern of the hexameter verse. In partieular H. N. Porter ?4 has tried to establish some principles underlying the word-localization earlier demonstrated by O'Neillj* in terms

of cola,

which

are shorter

metrical units

within the

line, and caesurae separating off the cola at three positions, each of which is variable.1® Porter’s analysis shows an essential similarity between the Iliad and the Odyssey with which other texts may be compared. He detects in the T'heogony a relatively high proportion of lines containing a B! caesura, i.e. with a break between the two short syllables of the third foot, and he shows that the first half of the Works and Days agrees in the main, though with some inconsistencies, with the T heogony in its choice of line-types. Porter's other conclusions about Hesiod are of special interest to us here : ‘ The metrical evidence . . . certamly does not support those who place Hesiod before Homer. ... the Theogony consistently and the first half of the Works and Days 1n main outline vary from the Homeric norm in the direction of Alexandrian practice.’ 17 He finds the later part of the Works and Days relatively primitive in its versification, and explains the disparity from the rest as resulting from Hesiod's use of archaic expressions in this part of the poem. ' In the second half (sc. of the Works and Days), where the text reads like a string of proverbs, the traditional element plays such an important role that the statistics of choice of line-type show the influence of the metrical characteristics of earlier folk literature.’ 1? Such a theory is by its nature impossible to prove or disprove, but it serves to remind us that the part of Hesiod which, on com-

parison with Homer, appears to diverge furthest from traditional hexameter verse in its language,!? might be shown to be 14 *'The early Greek hexameter ', YCS xii, 1951, 3-63.

15 See above, p. 25 n. 1. 16 For

a

criticism

of the

76—104. 17 YCS xii, 31. 18 ibid., 30.

1? See above, pp. 30-2.

four-colon

theory

sce

Kirk,

YCS

xx,

1966,

VERSIFICATION

AND

ENJAMBEMENT

89

no less traditional than other parts of the poem if we possessed a fuller corpus of early poetry to serve as our basis of comparison. As far as the Shield is concerned, Porter is cautious about drawing conclusions from the figures, which * present a highly individual pattern. ... There is no consistent agreement with any other text and no conclusions regarding date or authorship can be drawn. I would, however, regard it as improbable that the Shreld and the Theogony were written by the same poet.' 39 The picture which is thus suggested by Porter’s study conforms remarkably closely to that which emerged from the older analysis of the distribution of spondaic fifth feet iu the “outer metric’, even though these are two quite separate aspects of hexameter structure. It is less clear whether one may legitimately draw any more far-reaching «conclusions. On the one hand this kind of approach to poetic habits of composition has a peculiar attractiveness, because the processes concerned are largely subconscious, and the results can be expressed in terms of statistics which arc based upon a relatively large body of material. On the other hand one cannot regard such a phenomenon as the frequency of spondaie fifth feet as though it reflected some natural sequence of an evolution mherent in the development of hexameter verse or of the Greek language generally: while it may at first sight seem reasonable to advance the hypothesis that poets composing orally, with or without some musical accompaniment, might be especially inclined to prefer a regular rhythmical pattern -~~-» | in the fifth and sixth feet, so that an increased frequency of spondaic line-endings could be a sign of written composition, the wider history of hexameter poetry does not lend support to such an interpretation. Valuable as these 20 Y CS xii, 32. 21 Thus O’Neill’s sample passage of 1,000 lines from Theocritus contains only 18 spondaic fifth feet, i.e. less than 2-095, which is much lower than in Homer

(see his Table 38, YCS viii, 1942, 159) ; and for the extreme rarity of

spondaie fifth feet in Latin literature see S. E. Winbolt, Latin Herameter Verse, 1903, 128 f. G

THE

90

LANGUAGE

OF HESIOD

studies are in bringing to light what are possibly distinctive features in the preferences of individual poets, data of this kind tell us nothing certain about the poems' mode of composition. In the main part of this chapter therefore we turn to two fresh aspects of Hesiodic verse, and we shall see that the idea of oral composition does seem to be of help in their elueidation. Both of them, characteristically, are suggested by some work of Milman Parry in the Homeric field. (1) Irregular * lengthening ? and hiatus

One of Parry's major contributions to Homeric studies was to show how certain prosodical anomalies are “not to be explained by alleged metrical rules but by the simple consideration that the singers, when they had the choice between abandoning a formula and violating an established prosodical custom, were forced by their craft to choose the latter alternative.’ ? Certainly the Hesiodic poems provide good illustrations of this phenomenon, which is closely akin to the process already discussed ? whereby syntactical and inflexional innovations were introduced by Hesiod under the influence of the hexameter tradition. One instance of this kind has been noticed in Chapter VI, where the seemmgly lengthened vowel in τρίς at Op. 596 is best explained by reference to Hesiod's adaptation of traditional phraseology ; ?* but the point may now be illustrated with other examples. Here are some further cases of irregular ‘ lengthening ’.

(a) In discussing τραφέμεν

ἀτιταλλέμεναί τε | Th. 480,

Paulson ?5 notes the Homeric

| πρωτίστῳ δίδομεν, εὖτ᾽ ἄν...

B 228, and véxvas τ᾽ ἀγέμεν, ἕτεροι de ..

H 418;

but in both

these places the ‘lengthening ’ is facilitated by the following 22 "These words are taken from Hoekstra’s remarks (H MFP, 9 f.) about Parry's pioneer work in Les formules et la métrique d’Homere (1928), which includes many examples from Homer; some others are discussed by Hoekstra: see HMFP index s.v. ‘ Irregularities (prosodical) ’. 23 See above, pp. 82-4. 24 See pp. 79 f.; so too the shortening in γαιήοχον Th. 15, on which see

below, p. 101. 25

DR M, 110.

VERSIFICATION

AND

ENJAMBEMENT

9]

pause. The anomaly at Th. 480 is better accounted for by supposing that Hesiod, in modifying the formula τρέφον ἠδ᾽ ἀτίταλλον & 202, 303, and ἔτρεφεν ἠδ᾽ ἀτίταλλεν II 191, has

been influenced in his prosody by xopéew ἀτιταλλέμεναί τε | A 250, where there is no unnatural lengthening.

(b) Paulson ?° compares χάριν εὐεργεσιάων | Th. 503 with χάριν ἄνδρεσσι φέροντες | E 874 (χάριν δ᾽ Aristarchus), but it would seem that the expression χάρις μετόπισθ᾽ ὃ 695, x 319, 1s more directly relevant hero.

εὐεργέων

|

(c) δυσηλεγέος ὑπὸ δεσμοῦ | Th. 652 may be compared with δυσηλεγέος πολέμοιο | Y 154 (cf. also ἀμειλίκτων ὑπὸ δεσμῶν | Th. 659). (d) μέγας ὑπὸ δοῦπος ὀρώρει | Th. 703 is akin to various expressions where the syllable there ‘lengthened in arsis’ is heavy in its own right : πολὺς δ᾽ ὑπὸ κόμπος ὀρώρει | 0 380, πολὺς δ᾽ ὀρυμαγδὸς ὀρώρει

| 4x

Il., w 70, dpados Kat δοῦπος

ὀρώρει | 1573 (cf. also ποδῶν ὕπο δοῦπος ὀρώρει | Th. 70 and ποδῶν δ᾽ ὑπὸ δοῦπον ἀκούω | v 10). (6) σμερδαλέον ἵπποισιν ἐκέκλετο Sc. 341 is modelled on

| σμερδαλέον δ᾽ ἵπποισιν ἐκέκλετο T 399, where the δ᾽ makes the preceding closed syllable heavy.?? Several examples of hiatus in the Hesiodic poems are to be explained along the same lines. Krafft 28 has noted some instances of this (Th. 399 beside (2 425 etc., Th. 815 beside Z 227, I’ 451 etc., and Op. 386 beside

Y 833 eto.), and we may

add others : (a) The expression ἐτώσια ἔκφυγε χειρός | Th. 182 is used of Kronos' casting behind him the severed μήδεα of Ouranos : Hesiod is adapting the phrase ἐτώσιον ἔκφυγε χειρός | & 407, X 292, where the singular adjective, agreeing with BéAos,

entails no hiatus. (b) The

Th.

hiatus in κοῦραι πεντήκοντα,

ἀμύμονα

ἔργα ἰδυῖαι

264 results from the juxtaposition of half-lmes which

26 DRM, 113 n. 5. 27 On the @ of some mss. at Sc. 341 see Russo ad loc. and Paulson, DRM, 114 f.

28

VU HH, 57 n. 3.

92,

THE

LANGUAGE

Homer uses without hiatus:

OF HESIOD

cf. | κούρους πεντήκοντα: δύω

A 398, γυναῖκας ἀμύμονα ἔργα Lövias | I 128, ete.

(c) West 2? rightly says that the hiatus in ce évt φρεσὶ

θυμὸς ἀνώγει | Th. 549 could have been avoided if Hesiod had used μετὰ φρεσί as one might have expected. Certainly Hesiod makes use of this alternative elsewhere: e.g. in expanding the formula

6x Op. (cf. 4x

σὺ

δ᾽ ἐνὶ φρεσὶ

βάλλεο

σῇσιν

| Op.

107,

7 x

IL,

Od., he produces σὺ δὲ ταῦτα μετὰ φρεσὶ βάλλεο σῇσι | 274.39 But at Th. 549, where he is also expanding a formula ἑ θυμὸς ἀνώγει (-ye) | 4x Il., με θυμὸς ἀνώγει (-ye) |2x IL, Od.) the strong association of θυμός with ἐνὶ φρεσί (cf.

evi φρεσὶ θυμός in this position 486 ; also θυμὸς ἐνὶ φρεσί 1x

K 232, d» 386, Q 321, o 165, Il, 1x Od., ἐν φρεσὶ θυμός

4x Il., 1x Od.) has prompted him to use évi in preference to μετὰ (μετὰ φρεσὶ θυμός occurs only at V/ 600). (d)

| πλῆξεν

am’

Οὐλύμποιο

ἐπάλμενος

Th.

855

may

be

compared with | ἦλθες (-θον) am’ Οὐλύμποιο Η 25, 35, before a consonant, and ἐπάλμενος 6x Il, lx Od., always in this position and following a closed syllable. (e) νύκτας τε kai ἤματα, eis ὅ κεν αὖτις | Op. 562 takes the formula νύκτας τε καὶ ἤματα 27 340, (2745, before a consonant,

and places it before eis 6 xe which is especially frequent in this position (cf. Op. 630, H 30, 291, 376, etc.).

(f) τεύχε᾽ ἀρήια, ὄφρα τάχιστα | Sc. 108 shows the same kind of juxtaposition resulting in hiatus.

The usual shape of the

first formula is ἀρήια τεύχεα 2X

2x Od., but τεύχε᾽ ἀρήια

Il.,

occurs in this position, before a consonant, at 7 4; ὄφρα τάχιστα is frequent in Homer at the end of the line (cf. 4 465, E 690, etc.).31

With 29 80 31 with

such

a cavalier approach

to the

normal

rules of

HT, 320. cf. also pera φρεσί Th. 488, Op. 688, beside ἐνὶ φρεσί Th. 239, Op. 531. [t is true that in some of these examples (5, e and f) the hiatus coincides punctuation, but this is a common enough phenomenon where formulas

are juxtaposed

understood recitation.

in Homer

(cf. Chantraine,

GH

in terms of the slight pause which

I, 90), and

would

is most

readily

be present in oral

VERSIFICATION

AND

ENJAMBEMENT

93

versification, Hesiod may be thought lucky in not having had to satisfy any Examiners as to his proficiency in verse composition! Yet it is precisely his familiarity with traditional verse which has led him, in these places, to introduce irregularities. The examples do not suggest a painstaking process of composition in writing but rather the technique of the oral poet, whose complex memory of words regularly associated with particular phrases, or with particular positions in the line, influences his choice of language to such a degree that he is constrained to repeat or adapt a traditional form of expression even where this leads him to infringe conventions which are normally observed in fitting words to the hexameter.?? (2) Enjambement

In his article entitled ‘The distinctive character of enjambement in Homeric verse ’,** Parry draws a distinction between what he terms ‘ Unperiodic’ enjambement and ‘Necessary ' enjambement. He points out the relatively high frequency of the former in the Homeric poems, and explains this as resulting from the ‘ adding’ style characteristic of the His classification places technique of oral composition.**4 every line-junction into one of the following categories : I. No enjambement : the sense is complete at the end of the first line, and an independent sentence begins in the following line. II. Unperiodic enjambement : the sense is complete at the end of the first line, but the second line is tacked on to

it in one of four ways : (1) A free verbal idea is added, using a dependent clause, a participial phrase, or a genitive absolute, e.g. a 4—5. 32 of. Hainsworth's remark that ‘formulaic associations are not only tenacious, as was shown both by their mobility and their modification, but are suggested also by contexts even when the metrical environment is not immediately suitable (FHF, 89). 33 T AP A Ix, 1929, 200-20 (= DCE). 33a cf. Lord, SoT', 54, on the high frequency of Unperiodic enjambement in Yugoslav oral poetry.

04.

THE LANGUAGE

OF HESIOD

(2) An adjectival idea is added, describing a noun found in the preceding verse, e.g. A 1-2. (3) An adverbial idea is added either by a phrase or a simple adverb, referring to an action already mentioned, e.g. A 14-15. (4) There is added, by a coordinate conjunction, a word or phrase or clause of the same grammatical structure as one in the preceding verse, e.g. A 4-5. III. Necessary enjambement : the sense is not complete at the end of the first line. Parry examined four sets of specimen passages from Homer, Apollonius and Virgil, and the figures which he gives ?* are reproduced in TABLE 9. TABLE 9. Absence of enjambement, Unperiodic enjambement, and Necessary enjambement in epic hexameter poetry (after Parry)

Iliad

I

II | III

Odyssey

I

II

III

A 1-100 E „ I , N , P ,

48 50 52 48 48

31 24 18 31 19

a 1-100 € 5, ιν», "NM Do»

35 50 46 43 55

33 30 2] 30 14

32 20 33 27 31

Average

44-8

o

,,

45

Average

27

21 26 80 21 33

28

48.5 | 24-8 | 26:6

$

5

40

32

28

| 26-6 | 28-5

Argonautica

I

II

III

Aeneid

I

II

III

I. 1-100 I. 681-780 II. 1-100 III. » IV. » IV. 880-988

31 98 33 4] 35 3l

18 20 15 14 13 16

51 42 52 45 52 53

I. 1-100 Ill. ,, γι, VII. ,„ IX. ,„ XI. ,„

36 40 37 33 45 38

11 16 12 13 11 13

53 44 50 55 44 49

Average

A more

|

|

968

recent

16-0 | 49:1

study

Average

by Kirk?*

— 38-3

i

makes

use

125

| 49:2

of Parry's

classification in the wider context of sentence-structure and the

34 DOE, 204. 35 * Verse-structure

105-52 (= VSH).

and

sentence-structure

in Homer’,

YCS

xx,

1966,

VERSIFICATION

AND

ENJAMBEMENT

95

position of pauses. In addition to proposing some new terminology and useful subdivisions,?¢ he rightly draws attention to difficulties which arise in allotting certain lines to Parry's categories, in spite of the relatively clear and formal criteria which Parry laid down.?? In these ambiguous cases Kirk himself, who has in mind primarily the effects of pause in the course of recitation, has been guided by ' how a reasonable editor might punctuate the text ’.38 The main result of this, as he recognises,?? 1s to increase slightly, in comparison with Parry, the proportion of line-junctions placed in category III (Necessary enjambement) at the expense of II, and in category II (Unperiodic or ‘ Progressive ’ enjambement) at the expense of I (No enjambement).*? But the difference which this makes is not great, and Kirk’s analysis of the Sixteenth Book of the Iliad provides welcome confirmation of the general picture which emerged from Parry’s sample passages. Thus the book as a whole (867 lines) shows 38:3% with No enjambement, 28-695 with Unperiodic enjambement, and 33-1% with Necessary enjambement ; *! when we divide it into 100-line sections the figure for Unperiodic enjambement nowhere falls below 20% (II 301-400), and it rises as high as

37%, (IT 101-200). 36 VSH, 106-8. 37 Standard punctuation is not a sufficiently accurate guide: see Parry, DCE, 203. 38 Kirk, VSH, 109. 3? VSH, 119. 10 For example, in JT 1-100 Kirk classifies as Necessary enjambement some lines which Parry would probably have regarded as falling into the category of Unpcriodic enjambement ; these are I7 52 and 99 (Parry's II.1), IT 70 (his II.2), and Π 25 (bis II.4) ; he might even have placed IT 25 in his category I (No enjambement). Similarly Π 14, 15 and 64, which are counted as ' Progressive ' enjambement by Kirk, might be considered as showing No enjambement, in spite of the usual punctuation with a comma. See further below, p. 96 n. 43. 41 VSH, 119 with n. 2; for convenience I have here kept to Parry's terminology. 42 These percentages, and those for the whole book, are readily obtained by adding together adjacent figures in column D of Kirk's Table B (VSH, 150 ; the figure 20 given there for lines 751-800 is a misprint for 18).

THE

96

LANGUAGE

OF HESIOD

We now give figures for Hesiod, which have been compiled with Parry's slightly stricter criteria in mind (see TABLE 10). TABLE

10.

Absence of enjambement, Unperiodic enjambement and Necessary enjambement in the Hesiodic poems

Theogony

I

II

III

Works and Days

I

II

III

1-100 101-200 201-300 301-400

29 31 21 38

42 39 65 46

29 30 14 16

1-100 101-200 201-300 301-400

45 30 49 52

26 32 21 27

29 38 30 2]

401-500

41

31

28

401-500

41

29

30

501-600 601-700 701-800

43 30 28

29 35 35

28 35 37

501-600 601-700 701-800

36 43 45

32 25 26

32 32 29

15

7

6

801-900

36

37

27

801-823

901-1000 1001-1022

31 9

36 5

33 8

Average

Average

| 33:0 | 39-1

| 43-0 | 27-2 | 29-7

| 27-9

Shield

I

II

III

1-100 101-200 201-300* 301-400 401-480

32 36 28 39 29

33 35 28 24 21

35 29 42 37 30

Average

| 34-3 | 29-5 , 36-2

*leaving out Sc. 283 and 298, where tbere are special textual problems.

Again it may be conceded that an independent count could yield different figures,*? but any discrepancy so arising would be too small to alter the main pattern, which is quite clear. The high proportion of Unperiodic enjambement in Hesiod 13? Like Kirk (VSH, 119) I have made independent counts of some of Parry's sample passages without arriving at precisely the same figures as those which Parry gives. But the discrepancy is much too small to affect the main lines of our argument here ; the reasons for it are complex:

see Kirk,

loc. cit. The problems raised by the equivocal examples would best be solved by devising a system of punctuation based on more rigorous and formal criteria than those observed in the conventional system. The whole matter merits fuller discussion than is possible here.

VERSIFICATION

AND

ENJAMBEMENT

97

places him firmly beside Homer, these two together contrasting markedly with later poets who are known to have used writing.** Here we are dealing with a feature which is best explained as a direct consequence of the habit of oral composition. Parry makes this point with some force, and his words are worth quoting: ‘Homer was ever pushed on to use unperiodic enjambement. Oral versemaking by its speed must be chiefly carried on in an adding style. ... this need of the oral poet to order his thought unperiodically in word groups has affected the way in which the thought is drawn out from verse to verse : it has made unperiodic enjambement twice as frequent, necessary enjambement twice as infrequent, as in the writers of the literary epic.’ ** His words may be said to apply with equal truth to the Hesiodic poems, or at least to the T heogony and the Works and Days. Α few observations may be made about differences between various parts of the Hesiodic corpus which are noticeable in Table 10. The exceptionally high figure for Unperiodic enjambement in certain portions of the T'heogony, especially Th. 201-300, results mainly from the presence of some long successions of added lines, for example when the poet is listing the names of the Nereids in Th. 240-62. It is to be noted however that Unperiodie enjambement is characteristic of the Theogony as a whole, including its descriptive and narrative passages, not least those whose Hesiodic authorship is often doubted.*$ Thus the Typhoeus episode (T'h. 820-880) shows about 40% of Unperiodie enjambement in its line-junctions, and provides some very good illustrations of the * adding style ', e.g. at Th. 859-66. The same is true of the passage at the end 44 The figures given for Apollonius and Virgil in Table 9 may be reinforced by the following results from some counts of 100-line passages in didactic poetry : Aratus, Phaen. Lueretius I .

. .

. .

I 33-41 29-32

II 15-21 13-20

III 43-51 53-58

45 DOCE, 215; in the last statement Parry has, for the sake of simplicity,

slightly exaggerated the position : cf. the figures in Table 9. 48 See above, pp. 4 f.

98

THE

LANGUAGE

OF

HESIOD

of the Theogony which West 47 suspects of being un-Hesiodic : Table

10 shows

that Th.

901-1022

has about

34%

of Un-

periodic enjambement, and does not differ markedly in this respect from the rest of the poem. If considerations of style and content make it necessary to suppose that these passages are later accretions, the possibility arises that they may be additions from oral sources rather than interpolations in a written text. While the Works and Days does not show such a spectacularly high frequency of Unperiodic enjambement as is found in the Theogony, the frequency for the whole poem (27-2%) 1s still slightly higher than the average which Parry reached from his sample counts of Homeric passages. Indeed, the over-all distribution of line-junctions among the three categories comes remarkably close (within 2%) to that for the 600 lines which Parry examined from the Odyssey. The relatively low figure 48 in Op. 201-300 is to a large extent accounted for by the high incidence of end-stop (No enjambement) in this passage, and since that is undoubtedly a feature of primitive versification, one cannot regard it as evidence supporting the use of writing. To put the point another way, one may expect an oral poet, for reasons connected with his technique of verse-making, to show a markedly lower incidence of Necessary enjambement than will be found in the work of a literary poet. Thus the highest of the figures in column III of Parry’s tables for the ad and the Odyssey (see Table 9) was 33%, and the lowest was 20 or 2194.9 The corresponding range in our Table 10 for the Theogony is from 37 to 14%, and 4 HT, 398. 18 Jt is nevertheless higher than that for several of the 100-line passages examined by Parry (see Table 9) and for II 301-400 in Kirk’s analysis. # This compares with a range of 23%, (II 701-800) to 47% (II 301-400) in Kirk's analysis (see column G of his Table B, VSH, 150). The high figure at the upper extreme results from the unusual frequency of Necessary enjambement in ΠῚ 301-50, and Kirk makes some perceptive comments on the excited style of this passage, and of IT 51-100 (see VSH, 199 f.). The figures for the rest of the book fall much closer to those given by Parry, though the over-all average (33-195) is at Parry’s upper extreme, partly for the reasons given above, p. 95 with n. 40.

VERSIFICATION

AND

ENJAMBEMENT

99

for the Works and Days from 38 to 21%. All these figures contrast sharply with those for Apollonius and Virgil, where seven out of the twelve passages examined by Parry show a figure of over 50%, and where the lower limit of the range is 429.50 The Shield 1n this as in other matters is erratic, and while

the figures for Unperiodie enjambement come within the Homeric limits of Table 9, there is a relatively high frequency of Necessary enjambement throughout, the figure rising to over 40% in Se. 201-300, which is higher than anywhere else in Hesiod. It may be noted that the introductory passage about Alkmene, said to have come from the Ehoiai,°! contains

an unusually low frequency of Unperiodic enjambement (only eight examples in Sc. 1-48); there then occurs a series of ‘adding’ lines with Unperiodic enjambement from Sc. 49-56, which provide the transition to the main narrative. Too much importance must not be attached to counts of this kind when they are taken from a comparatively short passage ; 5? but certainly it 1s untypical of the sort of composition which we find in Th. and Op., and this, taken in conjunction with the very high incidence of spondaic fifth feet in the same passage,9 must be regarded as casting further doubt on the Hesiodic authorship of these lines. To sum up, there exist, within the pattern of versification

which is characteristic of Greek hexameter poetry generally, some distinctive features in Homer and Hesiod which may be plausibly explained by reference to the technique of oral 50 See Table 9.

The same appears to be true of Aratus and Lucretius:

see above, p. 97 n. 44.

51 See above, p. 4. 52 Even passages of 100 lines may VSH,

118,

where

he

points

out

that

be misleading ; cf. Kirk's remarks, the

irregular

distribution

of these

phenomena is not merely a statistical question but is partly connected with the subject-matter and style of particular passages. 53 There are eight cases in the first 50 lines, not counting the ambiguous ᾿Αμφιτρύωνα | in Sc. 2; for the frequency of this feature in Hesiod generally, see above p. 87.

100

THE

LANGUAGE

OF HESIOD

composition. There appears also to be a substantial unity, though not complete uniformity, between the Theogony and the Works and Days from which the Shield stands apart. We may add these points to the rest of the evidence when we come to draw conclusions about authorship and oral composition in a wider context.54 54 See below, pp. 191, 196.

CnaPTER VIII

UN-HOMERIC

WORD-FORMS AND SOURCES

THEIR

DIALECT

THE purpose of this and the next two chapters 1s to consider various linguistic features in which Hesiod's usage shows some difference from that of Homer. We begin with an examination of ἃ small number of un-Homerie word-forms, most of which

are relevant to the problem of dialect influence in the Hesiodic poems—or atleast may appear to be so—on account of some phonological or morphological peculiarity which they contain. The most important ones have been listed in various works on the subject of the Greek language,! and they are treated by several scholars writing on Hesiod.? Detailed points must be discussed ? before any general conclusion can be attempted, and for convenience the arrangement followed here is to take each feature as it arises in the course of reading the poems in the order Th., Op., Sc., giving cross references where necessary. Th. 15 γαϊήοχον This form troubled several nineteenth century editors, and

Goettling read yexoxov with one ms., taking it as a Boeotism and comparing Pindar,

Ol. 13, 81.4

But there is no need to

adduce dialect influence : the form is fully explained, in terms of formulaie modification, by Hoekstra ; ? cf. esp. N 43, Y 34, 1 e.g. Kühner-Blass, AG Li, 28 f. ; Schwyzer, GG I, 108 ; Thumb-Scherer,

HGD II, 8 f. ? Most recently by Mrs. Davies, Troxler and West: see above, p. 22; also by C. Gallavotti and A. Ronconi, ‘ Excursus sulla lingua di Esiodo ' in their La lingua omerica,

1946, 145-56.

3 These observations have however been kept as brief as possible, and where I find myself in agreement with views expressed by others I have felt it sufficient to refer to them without repeating at length what they have said, even though this results in a somewhat uneven treatment of the subject here. The short acc. pl. forms are listed as they occur, but all discussion of them is deferred until Chapter X. 4 See the Goettling-Flach edition ad loc. 5 Mnemosyne

x, 1957, 210 f. ; see also West, 1/7, 157 for many parallels.

THE

102

LANGUAGE

OF HESIOD

Y 584. In the light of our discussion in Chapter VII ® it is worth mentioning a further point: this anomaly could have been avoided without much difficulty ıf Hesiod had felt the need to do so, by substituting, e.g., Ποσειδάωνα βαρύκτυπον ἐννοσίγαιον

| (cf. Th. 818) or Ποσειδάων᾽

χαίτην | (cf. Th. 278, Op. 667,

6,

ἐνοσίχθονα kvavo-

(028) ; but his familiarity

with the formulas //ooeıdawv yavjoyos and γαιήοχος ἐννοσίγαιος in these positions has caused him to use the anomalous γαζήοχον when he wants the name in the acc., instead of working out an alternative. Th. 19 and 371 λαμπρᾶν

Ionic n for à is general in Hesiod, and we even find ' Epueiv Op. 68 beside “Epueiav in Homer with Aeolic a.” On the other hand λαμπρᾶν and a few other forms (cf. below on Th. 135, Op. 534) contain à where Ionic might be expected. Gallavotti and Ronconi think this may be ἃ Boeotism,® but West? is more likely to be right 1n ascribing the whole phrase λαμπράν τε Σελήνην to Attic influence. The chief importance of λαμπρᾶν is that it strengthens our confidence in the reliability of the transmission : !? if there had been any systematic transference of the text from a dialect where ἃ was general into Ionic, on the lines proposed

by Fick," it is difficult to see why λαμπρᾶν should have been exempted. Nor can we ascribe λαμπρᾶν to later Attic interference during transmission !? without supposing that the same thing has happened independently in Hom. Epigr. 3, 3 and other texts.!? The simplest explanation is that the phrase λαμπράν re Σελήνην has been in the text from the first ; and $ See esp. p. 93 with n. 32. * Chantraine, GH I, 20. On this and a small number of other hyper-Ionic forms in Hesiod see West, HT, 80. 8 LLO, 149; ef. Rzach, DH, 396.

? HT, 81. 10 of. above, p. 13. 11 See above, p. 17 n. 25. 1? As is assumed by Troxler, SWF/, 13 See West, HT, 81.

108.

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

103

if it was so spoken by Hesiod, the reason must be that he had learnt it in that form and not in the form λαμπρήν τε Σελήνην OT λαμπράν re Σελάναν. Th. 80 ἔδον and Op. 139 ἔδιδον Although third person pl. forms of this athematic type do not occur in inscriptions outside West Greek areas, their presence here need not be due to West Greek dialect influence, since they might occur as archaisms in any dialect, and in any case closely comparable forms are found in Homer. Nothing needs to be added to Mrs. Davies’ admirable discussion of them.!4 Th. 41 (also 129 West) θεᾶν and Op. 145 μελιᾶν These two genitive plurals in -ἂν are referred to as Dorisms by the ancient commentators and by Goettling-Flach and Wilamowitz ad loce.; but the contraction of -ἄων to -Gv is found not only in West Greek but also in Arcadian, Thessalian,

Lesbian, and in the article in Boeotian. They may therefore be ascribed just as well to Aeolic influence.!* The presence of these forms in Hesiod, instead of the expected -ὧν, can clearly have nothing to do with metrical considerations. It seems likely that the introduction and preservation of θεᾶν at Th. 41 have been helped by the fact that θεῶν there could cause an ambiguity: θεᾶν in this context means ‘of the Muses ’ whereas θεῶν at Th. 44 means ' of all the gods’. In Op. 145 the substitution of μελιῶν for μελιᾶν would not involve such a confusion, but it may be that the frequency of the phrase ἐκ μελέων ‘ limbs’ in epic (17 110, ¥ 689, 880, A 600; also ἐκ μὲν μελέων κ 393, ἀπὸ μελέων H 131, N 672, II 607,

o 354) has prompted the poet to make it quite clear that the word which he is using here is μελιαί “ash-trees’. If this is so, we again come to the slightly paradoxical conclusion that the influence of the language of traditional poetry has caused Hesiod to use an untraditional linguistic form (cf. above on Th. 15 γαϊήοχον).

14 Glotta xlii, 1964, 145-7.

15 See West, HT, 83 f.

104

Th.

THE

60 κούρας

LANGUAGE

(and

the

other

OF

HESIOD

acc.

plurals

listed

below,

pp. 142-5) This might be a West Greek form ; but sce the discussion in Chapter X. Th. 135 Octav and 'Petav ; also ThA. 371 Θεία

These forms with à are well supported by the mss. and

contrast with ‘Pein αἱ Th. 625, 634 (cf. the variants “Peins and 'Peías at © 203). West thinks that Θεία was the regular early form, which may have been influenced by the analogy of Gea ; the choice of ἱΡείαν rather than ‘Peinv is then prompted by the desire for assonance.!9 Attic influence may also have played

some part: at Th. 371 Θείᾷ is in the same line as λαμπράν, on which see on Th. 19 above. Th. 118 λοχέοιο (Aoxeoto Rzach)

The accentuation may be Aeolic, but this cannot be regarded as sound evidence for Hesiod's pronunciation.!” As for the form itself, although it is unique there seems to be no need for Paley's suggestion λόκχοιο or for any other conjecture ; 18 Hesiod 15 not averse to eking out ἃ word artificially with an extra syllable, as in éew !? at Th. 145; 210, 257, 286, Fr. 150.20 M-W, as from

cf. ἐθύνεον Sc. 156, θυνέω, beside θύνω

universally elsewhere.?? In these places the motive is metrical ; but the same process might explain the curious εἰρεῦσαι Th.

38,71

where

the

aim

seems

to

be

merely

to

achieve

assonance with forms in the neighbouring lines.?? Th. 200 φιλομμηδέα

Whether we adhere to the ms. reading with -n- (beside φιλομμειδής Th. 256, 989, 5x Il., 1x Od.) or follow Bergk in 16 HT, 203. 17 See West, HT, 82. 18 See West’s note, HT, 219. 1? For various possible explanations of this form, which is artificial and not of dialect origin, see Chantraine,

GH

I, 76, 82 f.;

Troxler,

SIH,

31;

West, ΠΊ', 208 f.

20 The form

is readily explained

as analogous

to ὁμέλεον : see 2 535

beside Sc. 156. ^. West however rejects this reading : see CQ xii, 1962, 177, and HT,

170.

22 See Troxler, SIH,

89.

88,

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

105

adopting φιλομμειδέα . . . μειδέων (beside μήδεα Th. 180, 188,

4x Od.), it is tempting to see here a confusion of open 8 and close 6 or ei diphthong, such as is found in Boeotian.? But it would be rash to take this line as certain evidence for dialect influence : the poet may simply be distorting the form of one word or the other in order to suggest the supposed connexion between them. The absurdity of such an etymology must surely have been apparent to Hesiod and his audience, to whom μειδάω ' to smile’ and its derivatives were well known (cf. esp. μειδήματα among Aphrodite’s attributes at Th. 205, only five lines below the monstrous φιλομμηδέα as her epithet). Hesiod seems to have derived enjoyment from putting forward unconventional etymologies, and we may compare the novel explanation of [Javédpy at Op. 81.?* A similar distortion of a word for the sake of an etymology is seen in riraivovras at Th. 209 beside the regular riraiww.?? [Ὁ is safer therefore to regard φιλομμηδέα as a punning form, perpetrated perhaps by Hesiod's sense of humour, rather than as evidence for dialect

influence. Th. 267 'Δρπυίᾶς

See on Th. 60 above, p. 104. Th. 321, 825 ἦν The subject in both places is κεφαλαί, so that ἦν is generally taken to be a 3rd pers. pl. form, as is found in most dialects outside Attic-Ionic, whose ἦσαν or ἔσαν is the only form found

in Homer.?*

Rzach

ascribes

it to Doric influence,??

but it might belong to an early stage of any dialect.?* Alternatively it can be plausibly explained as an instance 23 Buck, GD,

25;

Thumb-Scherer,

HGD

II, 21 f.;

cf. West,

HT,

88 (in

his text West reads φιλομμειδέα.. . . μηδέων). 24 See O. Lendle's discussion in Die * Pandorasage ' bei Hesiod, 1957, 63 f.,

though we need not accept his view that the passage is un-Hesiodic. 25 See Troxler, SW H, 33. 26 Ohantraine, GH I, 473. 27 D H, 465; similarly Troxler, SWH, 92, 238. 28 Full discussion by Davies, Glotta xlii, 1964, 141-5. H

106

THE

LANGUAGE

OF HESIOD

of the so-called schema Pindaricum without any reference to dialect influence.” Th. 326 Φῖκα and Sc. 33 Φίκιον

These forms are usually understood as Boeotisms,? as is suggested by the Hesiodic scholia: ᾧϊῖκα δὲ αὐτὴν (sc. Σφίγγα) οἱ Βοιωτοὶ ἔλεγον ; but it appears that the form of the name without initial 2- and without the nasal was quite widely used: Plato ?! is acquainted with both forms, and the word found its way into Latin in the form picus (or ? poz, picis)? "Thus the ascription of these forms to a Boeotian dialect source does not rest on external linguistic evidence but is à plausible inference from their appearance in Hesiod and from the Boeotian context in both these passages of Th. and Sc.

Th. 371 Octav, λαμπρᾶν See on Th. 19 above, p. 102, and on Th. 135, p. 104. Th. 401 pweravarétas

See on Th. 60 above, p. 104. Th. 487, 890, 899 ἑὴν ἐγκάτθετο νηδύν |

Scholars in the nineteenth century attached importance to this variant for ἐσκάτθετο (the reading adopted in all modern editions ?*) as showing ἐν + acc.—the inherited construction retained in many dialects in mainland Greece but replaced in Ionic by eis, és.39 But even if it was the original form in Hesiod, nothing is to be learnt from it about dialect imfluence.3® 20 West, IIT, 84; Davies, op. cit., 141 with n. 3. 39 Schwyzer, GG I, 334, 692; Gallavotti-Ronconi,

LLO,

147 ; Troxler,

SWH, 240. 31 Cratylus 414 d, where he says nothing to imply that the Φικ- form is a Boeotism.

3? Ernout-Meillet, DEL, s.v. picus. 33 34 35 36

See West’s discussion, HT, 88, 256. For the ms. evidence see West, HT, 301. Buck, GD, 107. Davies, Glotta xlii, 1964, 151; West, HT,

84.

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

107

Troxler ?? is right to draw attention to the relationship between

the Hesiodic formula and the Homeric ἑῷ ἐγκάτθετο κόλπῳ | & 223 and similar phrases at & 219, A 614, ψ 223, to which one

should add Op. 27. It seems worth mentioning also the frequent Homeric use of ἐμπίμπλημι + acc., which is obviously not to be ascribed to mainland dialect influence ; cf. esp. the

phrase μεγάλην ἐμπλήσατο νηδύν | ı 296, of the Cyclops consuming the companions of Odysseus. Th. 534 βουλάς See on Th. 60 above, p. 104. Th. 585, Op. 63 x&Xóv, Op. 752 ἴσον

The loss of posteonsonantal F in these words is accompanied in many dialects by compensatory lengthening of the vowel in the first syllable,3® and we find καλός, ἶσος always in Homer and elsewhere in Hesiod. Troxler is inclined to ascribe the Hesiodie forms to Attic influence,?? but West is surely right to cast the net wide1 : *? comparable forms with short vowel, from povF-, evF-, ete., occur in Homer,*! and there are many

examples of them in Sappho and Alcaeus including both καλός (frequently) and itcos.* Th. 653 βουλᾶς

See on Th. 60 above, p. 104. Th. 678 περίαχε

Schulze’s explanation of this form as from *zepiF Faye, aorist, would make it unnecessary to suppose that περί here has undergone apocope or elision.$? On the other hand it 37 SWH, 114 f. 4. HT, 82. 1:2 E, M. Hamm, 13 See Davies,

38 Buck, GD, 49 f. 39 S]VH, 48, 238. 131 Chantraine, GH I, 161 f. Grammatik zu Sappho und Alkatos, 18. Glotta xli,

1964,

149

with

reterences.

In support

of the

interpretation as an aorist she refers to ἐσμαράγησεν in Th. 679 ; but against

it one should note the imperfects ἔφαινον Th. 677 and esp. ἐπέστενε δ᾽ οὐρανὸς εὐρύς | Th. 679, τινάσσετο μακρὸς "Ολυμπος | Th. 680; cf. also the Homeric ἐπὶ δ᾽ ἔστενε δῆμος ἀπείρων | 2 776. The choice of the aor. ἐσμαράγησεν

would seem to be determined by the needs of the metre.

108

THE

LANGUAGE

OF HESIOD

seems likely that by Hesiod’s time ἴαχε, imperfect, was well established in the tradition,** so that there is a strong case for regarding the form as showing περί with -ı elided, which never occurs in Homer but is found in Aeolic and in other dialects including Attic and North West Greek.*5 It seems unwise however to press the importance of dialect influence here, when Homer has such parallel phrases as ἐπίαχον

υἷες

Ἀχαιῶν

| Η

403,

I 50, beside

περίαχε

πόντος

ἀπείρων | Th. 678; cf. also ἐπὶ δ᾽ ἴαχε λαὸς Ἀχαιῶν | N 822 beside Hesiod’s περὶ δ᾽ ἴαχε γαῖα μέλαινα | Th. 69. Th. 133 περοίχεται

The form must be regarded as textually very suspect,*® and now West has replaced it with a new papyrus reading ἐπελήλαται which looks much healthier. It might otherwise belong with Th. 678 περίαχε. Troxler makes the alternative suggestion that περ οἴχεται (with the particle περ) might be read here,*? but this is probably unnecessary. Th. 804 eipäc or cipéág

See on Th. 60 above, p. 104. Th. 825 ἣν

See on Th, 321 above, p. 105. Th. 890, 899 ἑὴν ἐγκάτθετο νηδύν |

See on Th. 487 above, p. 106. Op. 22 ἀρώμεναι and Op. 683 αἴνημι

The athematic inflexion of a contract verb is a feature attested in various dialects : 48 αἴνημι is a clear example and it may plausibly be regarded as an Aeolism.*? As far as 44 Soo.

West,

HT,

83;

and

Chantraine,

GH

I,

examples. 45 Davies, op. cit., 148;

Buck, GD,

79.

$6 See Davies, Glotta xlii, 1964, 148 f. 41 SWH, 231 n. 13. ‘8 Buck, GD,

123.

“ West, HT, 84;

Gallavotti-Ronconi, LLO, 153.

140

for

the

Homeric

UN-HOMERIC

ἀρώμεναι

WORD-FORMS

is concerned,

AND

THEIR

however,

DIALECT

the form

SOURCES

may

have

109

been

partly suggested by similar-sounding infinitives in Homer: cf., in the same position in the line, dAdpevar ᾧ 495, γνώμεναι

d 266, 609. The Hesiodic phrase ἀρώμεναι ἠδὲ durederv | may be compared also with such Homeric expressions as δαμήμεναι nd ὀχέεσθαι | K 403, P TT, δαμήμεναι οὐδ᾽ ὑποείκειν | Y 266, and ξυμβλήμεναι ἠὲ δαμῆναι | © 578. Hesiod's ἀρώμεναι

may therefore be an analogical formation original Aeolism.

rather than

an

Op. 63 κἄλόν

See on 7}. 585 above, p. 107. Op. 139 ἔδιδον See on Th. 30 above, p. 103. Op. 145 μελιᾶν

See on 7}. 41 above, p. 103 f. Op. 208 εἷς

See on Op. 617 below, p. 114. Op. 280 εἰ γάρ τίς x^ ἐθέλῃ

It has been suggested that the word-order here has been influenced by West Greek ; 5° Davies rightly doubts this,

But the

drawing attention to ὅτε τίς ke θάνῃσιν A 218.9

order is readily enough explained from Hesiod’s own habits elsewhere : xe is commonly elided before a subjunctive form of ἐθέλω,

e.g. ὅς κ᾽ ἐθέλῃ

Op.

210,

at κ᾽ ἐθέλῃσ᾽

Op.

268,

αἴ κ᾽

ἐθέλητε Th. 164 (see also Th. 430 ff. ; there are many similar

examples in Homer). And for the earlier part of the phrase we may compare ei γάρ τις καί Th. 98, Op. 321, ei γάρ τοι καί Op. 344. There seems no need to make a mountain out of this molehill. Op. 353 προσεῖναι See on Op. 617 below, pp. 114 f. 50 West, HT, 85.

51 Glotta xlii, 1964, 140 n. 1.

110

THE

LANGUAGE

OF HESIOD

Op. 392 ἀμάειν There 15 an extensive literature relevant to thıs form, from

which the only certain conclusion to emerge ıs that it 18 impossible to tie ıt down to a particular dialect or dialect group. Op. 510 πιλνᾷ and Op. 526 δείκνυ (πίλναι, δείκνυι Wilamowitz) While the middle πίλναμαι is relatively common in Homer and Hesiod, no active form of the verb is found except here, where we may understand it as from *riAvnu ‘ cause to come near ’, the form πίλναι being Aeolic for πίλνησι. The latter is the explanation of Wilamowitz, who takes the form to be a

Boeotism and compares δείκνυι at Op. 526, his own correction for δείκνυ of the mss. Alternatively δείκνυ, without Wilamowitz’s emendation, may be regarded as an Aeolism of a type found in Lesbian ; 58 νῇ at Op. 777 might also belong here, but the reading is uncertain; so too the variant iAva at Op. 510.54 Paley draws attention to an interesting variant δείκνει at Op. 526, and suggests that this might be the result of synizesis in δεικνύει which is found at Op. 451; cf. also the imperative δείκνυε at Op. 502. This preference for thematic verbal endings is a feature of Ionic,55 and it may be best to regard πιλνᾷ similarly, not as an Aeolism, but as from *rıAvaw : cf. πιλνᾷς, the probable reading at h.Dem. 115, though the sense is intransitive there.

GI

52 Meister (17K, 90) thinks that it is a Boeotism, but see Chantraine, I, 355 n. 1. West (HT, 85) would ascribe it to West Greek influence,

but it should be noted that the long -a- is not in fact confined to forms with dicctasis like uevowda : cf. esp. ἀναμαιμᾶει Y 490, and see Chantraine, GH I, 361 f., Frisk, GEW, s.v. δίψα. Troxler (SW H, 102 f.) regards the form as an artificial one following false diectasis; but this also is not without its difficulties, especially if his suggested parallel μηχανάαται Op. 241 is subjunctive. 53 Sce

Buck,

GD,

111;

Gallavotti-Ronconi,

88; West, IIT, 83. 55 See Schwyzer, GG I, 659 f. 55 See Thumb-Scherer, HGD II, 277.

LLO,

154;

Troxler,

SWH ,

UN-HOMERIC

WORD-FORMS

AND

THEIR

Op. 512 μέζεα This word seems indistinguishable ‘ genitalia" (Th. 180, 188, 4x

DIALECT

SOURCES

iu sense

from

11]

μήδεα

Od.), though West thinks that

Hesiod may have learnt it separately ‘as a vernacular term for where miserable animals tuck their tails ’.5° West himself lists μέζεα as West Greek, but this seems too definite an ascription when confusion of ζ and 8 occurs in so many dialects, and even in epic. Not only West Greek but Lesbian and Cypriot also show examples of ζ for δι, 5.58 Alternatively pelea may be connected more loosely with the assimilation (in the reverse

direction)

of £ to 8 or

88, which

is found

Thessalian and Boeotian as well as West Greek. source therefore

cannot

be excluded,

in

An Aeolio

and it seems

safer to

agree with Mrs. Davies ® that we cannot ascribe this form to the influence of any definite dialect.9! Op. 524 ἀνόστεος

It is sometimes suggested that this ‘ Kenning ' ? and others in Hesiod were of local dialect origin, being taken by the poet 5¢ HT, 86. 57 See Chantraine, GH I, 169, on ἀρίζηλος beside δῆλος and on Aristarchus’ reading at B 318. It is relevant to add to these examples a curious Homeric * parallel of sound’: | σχέτλιοί ἐστε, θεοί, δηλήμονες Q 33 beside | σχέτλιοί ἐστε, θεοί, ζηλήμονες ε 118. 58 Buck,

GD, 26;

Lejeune,

TPG, 216.

59 See K. Strunk, ZF Ixvi, 1961, SWH, 48 f. 80 Glotta xlii, 1964, 140 n. 1.

169

f., who

is followed

by Troxler,

61 Whatever the precise phonetic value of ζ in μέζεα, it is evident that we

are dealing here with some kind of palatalization which in its other manifestations, notably r > σ, seems to be less widespread in West Greek than elsewhere: see Lejeune, 7'PG, 54-6. 62 There is a considerable literature on this subject. For its wider aspects see W. Schultz in RE I A, 1914, s.v. Rätsel, 108-11 on ‘ Kenning und Weisheitsprobe ’; on this feature in early Germanic literature, see Bowra, Heroic Poetry, 242 f., and the full study of H. van der Merwe

Scholtz,

T’he

Kenning in Anglo-Saxon and Old Norse Poetry (1927), whose remarks (op. cit., 117 f., 179 f.) on the origin of this type of expression may be compared with those of Sellschopp,

SUH, 37 f., on the Hesiodic examples.

There is no

need to postulate any specially close connexion between Hesiod and oracular poetry with regard to these words: Troxler has a good discussion on this point,

SWH, 25-8.

THE

112

from popular speech. Kleitarchos

as

LANGUAGE

OF HESIOD

But while ἀνόστεος

a Lacedaemonian

name

is attested by

for the octopus,

we

may well be dealing here with a poetic word coined by Hesiod himself, like μελάμπυγος 'eagle' in Archilochus and ἀνθεμουργός ‘bee’ in Aeschylus.°* The suggestion put forward by Troxler, that ἀνόστεος here does not mean the octopus at 811,55 but rather the snail, hibernating in winter, seems to me

extremely attractive :99 ὃν πόδα τένδει is a vivid and accurate way of describing the snail's withdrawal into its shell; the two halves of line 525 cease to be tautologous, since οἶκος 1s the snail's shell and ἤθεα the nooks and crannies into which he is wont to retire;

and his behaviour in winter 15 contrasted

with his more venturesome activity when the weather turns warmer (cf. Op. 571 f.). In the light of our remarks above on λοχέοιο 57 one might further suggest that avóoreos here, instead of being interpreted as from av + ὀστέα ' boneless ',$? could have been intended as from à - νόστος ' one who has no

journey home ' ® and so contain the same thought as φερέοικος at Op. 571. That avóoreos is a ' riddle-word ', perhaps of Hesiod's own making, and not a dialect word, is made probable

by his evident fondness for indirect or cryptic names in this passage: thus he uses τρίποδι βροτῷ (or τρίποδι alone, reading 83 West, HT, 89;

cf. A. B. Cook, ' Descriptive animal names in Greece ’,

CR viii, 1894, 381-5, esp. 382 f., where he gives an interesting list, culled from Hesychius and elsewhere, of local names of this type. 64 Archil. 189; Aesch. Pers. 612; several other literary examples in Cook, op. cit., 382.

95 "There are difficulties in the usual interpretation of the word which are not fully disposed of by T. F. Higham's amusing ' Nature note : autophagy in octopods.

Hesiod vindicated ' (CR vii, 1957,

16 f.) ; nor am I persuaded

that the Hesiodic line throws any light on ancient knowledge of the mysterious reproductive apparatus of the octopus, as is urged by J. Wiesner, JDAI Ixxiv, 1959, 48 f. °° Troxler,

S'W H, 23, says that he is indebted for the suggestion to Prof.

Hermann Koller of Zürich. Curiously enough the samo idea occurred to me independently and I discussed it with Dr. J. Chadwick in 1960. 87 Seo above, p. 104. “5 I would agree however with Troxler that ‘ boneless’ might rcasonably be applied to the snail. 89. cf. Homer’s ἄνοστος ὦ 528 and ἀνόστιμος ὃ 182.

UN-HOMERIC

WORD-FORMS

βροτοί) for “old man’

AND

THEIR

DIALECT

SOURCES

113

at Op. 533, εὐφρόναι for ‘night’ at

Op. 560, ἡμερόκοιτος ἀνήρ for ‘ thief’ at Op. 605, νηὸς πτερά

for ‘oars’ (cf. A 125 = y 272) or ‘sails’

(cf. Fr. 76.7 Rz. =

205.7 M-W) ?? at Op. 628, and πεντόζοιο with αὖον ἀπὸ χλωροῦ

for cutting the nails of the hand at Op. 742 f. As for the animalnames ? we have, apart from ἀνόστεος and φερέοικος, ἴδρις for ‘ant’ at Op. 778, and others may be lurking here also: I would suggest that κεραοὶ καὶ νήκεροι ὑληκοῖται at Op. 529 means specifically ‘deer’, i.e. stags and hinds, rather than woodland

creatures

in general,

and,

more

tentatively,

that

τροχαλόν in Op. 518 may mean some small rodent, e.g. ‘dormouse ', with γέροντα then predicative.”? Two hibernating creatures, the dormouse

and the snail, would make more

suitable company for the young girl quietly resting in her sheltered home (Op. 519-23) than do her present neighbours, the agile old man and the self-devouring polypus. Op. 526 δείκνυ

See on Op. 510 above, p. 110. Op. 534 Eäye Chantraine explains this form as an Atticism,’? but West is probably right to be more cautious, listing it as ' Attic, island Ionic, or Aeolic’.74 The similarity between Op. 534 οὗ τ᾽ ἐπὶ νῶτα ἔαγε and Sappho 2, 9 κὰμ μὲν γλῶσσα éaye, with F apparently observed in both places, has prompted the attractive suggestion that both Hesiod and Sappho have here been influenced by a poetic formula from an Aeolic tradition independent of Homer.?? 70 Or both:

see Sinclair ad loc. and his reference to Gow

in CQ xi, 1917,

116 f. 71 On ἄτριχος ' ? snake’ and dév£a ‘lion’ in the Hesiodic fragments, see Troxler, SWH, 24 f. 72 That is, Boreas ‘ turns the Running One into an old man ’, meaning that the winter makes him fat, sleepy and inactive ; for tbe winter torpidity of the dormouse see Enc. Brit. 16th ed., 1956, XI, 539 s.v. ' Hibernation ', or 11th ed., 1910, VIII, 429 f. s.v.

73 GH 1,18;

‘ Dormouse

so too Schwyzer, GG I, 759.

75 See the discussion of R. Hiersche,

’.

74 HT, 82.

Glotta xliv, 1966,

1-5.

114

THE

LANGUAGE

OF HESIOD

Op. 564 τροπᾶς

See on Th. 60 above, p. 104. Op. 617 ein Mazon, with mss. (eto. Wilamowitz ; elow Rzach,

Sinclair, Colonna) The obscurity of this line arises from the word πλειών. If we understand it, as the Alexandrians did, to mean ‘ year ’,?®

then a verb with the sense ‘go’ 'pass' seems necessary: hence Wilamowitz's emendation to eto: explained by him as a Boeotism for ἴοι (3rd pers. sg. pres. opt. of εἶμι ibo"). Though this is accepted by Schwyzer and others,?" the form is rightly rejected by West."9 Sinclair, while reading εἶσιν, says in his note that he would prefer an optative: ‘If εἴη can be taken as opt. of εἶμι ibo (so Förstemann, Dial. p. 43), then it should

be retained. But Rzach (Dial. p. 453) denies the possibility '.7? Yet perhaps this is after all the right solution, as it may be also in Homer at £ 496.89

It is convenient to consider with this problem two other places in Op. where some confusion of εἰμί and εἶμι appears to be present. At Op. 208 εἷς (2nd pers. sg. indic.) undoubtedly means 'ibis' and is never taken otherwise: but it is to be noted that eis or eis in Homer invariably is from εἰμί ‘ sum’, and only εἶσθα is used as the 2nd pers. sg. of εἶμι ' ibo, as at K 450 and τ 69. The other line which is relevant is Op. 353, where προσεῖναι as from προσ-ειμί ‘sum’ is intelligible in the sense ‘be at hand’, but where the infinitive of προσ-εῖμι ‘go to’ would make even better sense. The line runs τὸν 16 Troxler suggests some possible alternative interpretations which would enable ein to be retained as from εἰμί ‘sum’: see SWH, 186 f.

77 See above, cite εἴοι 88

p. 15 nn. 18 and 19;

Gallavotti and Ronconi similarly

ἃ non-Homeric Aeolism in Hesiod : see LLO,

147.

78 AT, 89. 7 WED, 65. 80 Chantraine,

GH

I, 285,

considers

this possibility,

but

regards

εἴη

at

2139 and £ 496 as being probably from εἰμί ‘sum’,

This may be true of

42 139, but I am hesitant about accepting it at £ 496:

for while

infinitives

of purpose do occur after εἰμί ‘sum’ (see GH II, 301), the sense here surely needs a verb of motion (of. βῆ δὲ θέειν £ 501), and the parallel in thought and construction at H 372 f., with ἴτω, seems to be decisive.

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

115

φιλέοντα φιλεῖν, καὶ τῷ προσιόντι προσεῖναι. In such a context (cf. also Op. 354) the need to take προσεῖναι as precisely antithetical to προσιόντι is so great that Hartel, followed by Flach, emended to προσῖμεν (seo the Goettling-Flach edition ad loc.) Troxler prefers to think that the original reading was προσεόντι and that προσιόντι is a Boeotism which entered the text at an early date.®! But in the light of the above discussion it may be suggested that προσεῖναι should be retained and understood as προσιέναι.83 Op. 663 τροπάς See on Th. 60 above, p. 104. Op. 666, 693 καυάξαις

This form is probably Aeolic in origin; occur in Homer.9?

comparable forms

Op. 675 5ewág See on Th. 60 above, p. 104. Op. 683 αἴνημι

See on Op. 22 above, p. 108. Op. 696 τριηκόντων

This inflected form is paralleled in Lesbian (cf. πέμπων, δυοκαιδέκων in Aleaeus),9* and in the dialect of Chios (cf. πεντηκόντων etc. in inscriptions).®° There are no comparable examples from mainland Greece. Op. 698 τέτορα Various forms of this numeral are found in the inscriptional dialects, and the form used here by Hesiod is attested only in 81 SWH, 10 n. 12, taking up a suggestion made by Haupt in 1876. 82 cf. the opinions of ancient commentators cited in Sinclair’s note on Op. 353. 83 Chantraine, GH

84 Hamm, 85 Buck,

I, 158 f.;

West, HT,

83;

Troxler, SWH,

47.

GSA, 110 § 194. GD, 96;

Thumb-Scherer,

HGD

II, 107, 282;

West, HT,

83.

LANGUAGE

THE

116

OF HESIOD

West Greek, including both Doric and North West Greek.3% It is notable that where Homer requires this scansion he uses the Aeolic πίσυρες (3x Il., 3x Od.). Op. 752 ἶσον See on Th. 585 above, p. 107. Sc. 7 τῆς kat ἄπο κρῆθεν βλεφάρων τ᾽ ἀπὸ κυανεάων

There can be no doubt that κυανεάων is the correct reading, 87 and this has led scholars to assume the existence of a feminine form ἡ βλέφαρος 88 for which there is no other evidence. But a possible explanation ıs to be found at the only other place in the poems where the word βλέφαρον is used. Th. 910 runs : τῶν

kai

ἀπὸ

βλεφάρων

ἔρος

eiBero

Oepkoueváow,

and

the

author of Sc. 7 could have mistakenly regarded the participle as agreeing with βλεφάρων instead of with τῶν (fem. here). This suggestion, anticipating the method which has since been extensively used by M. Leumann,?? was put forward as long ago as 1894 by F. Leo,’ and was favoured by P. F. Kretschmer,® but it has found little support since. It is

rejected by Jacoby and Russo,?? both of whom believe Th. 910 f. to be an interpolation ; but even if this were true,?? it 88 Buck,

GD,

95;

Lejeune,

T'PG,

39;

West,

HT,

87;

fuller discussion

by Davies, Glotta xlii, 1964, 149 f. 8? The indirect tradition now has the support of Pap. Oxyr. 2494: see Fr. 195 M-W. ; the variants in the mss. are attempts to correct the anomaly : see Russo ad loc.

88 So Rzach, DH, 398 f. ; cf. Liddell and Scott s.v. βλέφαρον. 89 Homerische

Worter,

1950.

90 Hesiodea, 13 f. (= Kl. Schr. II, 353 f.). #1 DIH, 54. 9? Theogonia,

187,

and

Scutum,

72;

Xrafft's

discussion

of these

lines

(VUHH, 55 f.) seems to add nothing of substance to the other arguments, and his analysis of the sense is over-subtle. *3 Rzach, Mazon, and Jacoby, following Gruppe, all bracket the lines, but they are present in all the mss., and the context undoubtedly requires some short, descriptive piece about tho Graces after Th. 909 before going on to912: cf. parallel constructions at 903 at . . . (the Horai), 904 3s . . . and 905 αἱ... (the Fates), 913 jv... (Persophonc), and 917 τῇσι... (the Muses). This seems more important than Jacoby's objection : ‘ strophicam compositionem interrumpunt’: granted that a tendency to group lines in triads is

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

117

would not rule out Leo's explanation of kvavedwv at Sc. T: the interpolation might have entered the text at an early date, i.e. before the composition of the beginning of the Shield.?* Russo in his note on Sc. 7 cites a number of examples from other places in Greek literature of -dwy or -&v appearing in gen. plurals where it has no historical justification, but it 1s doubtful whether any of these are parallel to κυανεάων serving as a neuter. Most of them are feminine words in o- or consonantstems, to which the -ἄων ending might readily have been extended by analogy with the numerous feminine a-stems, Viz. νησάων, Σειρηνάων, both quoted by schol. B on Il. T 1, the former occurring also in Callimachus, Del. 66,?9 κυνᾶν

‘bitches’ 856 in Ibycus Fr. 60, and aiyáv in some mss. at Theocritus V, 148 and VIII, 49, clearly ‘ she-goats’ in both places. Apart from these fem. words there is only ἀοιδάων at Xenophanes 5, 4 D, quoted as a masculine by Schwyzer ; 97 but this also may well be a feminine, and Diels has so taken it. The variant ἀοιδᾶν at Euripides, Hipp. 743 and Med. 422 is to be regarded as a false correction of ἀοιδῶν mistakenly understood as a feminine.?9 It therefore appears that present here (see West, HT', 405), we need not suppose that the poet had to adhere to tbis rigidly. West retains the lines without brackets, though he regards the whole passage in which they stand (7h. 900 ff.) as an addition to the poem. 94 Alternatively the autbor of Sc. could have known the line in & similar context elsewhere and misunderstood its construction, though this is less plausible in view of the limited number of contexts in which the doublemeaning might arise. West, while accepting Leo's explanation of Sc. 7, believes that the same poet may nevertheless have composed both passages, and misunderstood the line as he used it at Th. 910 also (HT, 409 £.). This scems to me unlikely, especially as it involves supposing that subsequent readers of Th. have almost universally taken the construction (' grammatically’) in a sense different from that felt by the poet himself (“ ungrammatically ’). I find it simpler to believe that the author of Sc. 1-56 is not the same as the poet of the end of the T'heogony. ?5 The remaining examples in the Homeric schol., ῥοάων and Κρητάων, are best regarded as ordinary a-stems. 98. J, M. Edmonds, Lyra Graeca II, 112. 91 GG

I, 559;

Diels-Kranz, Frag. d. Vorsokratiker, I, 130 and III, 57.

98. See both Barrett and Page in their editions ad loce.

THE

118

LANGUAGE

OF

HESIOD

βλεφάρων ... kvaveawv at Sc. 7 presents a unique problem, and Leo’s solution is an economical and attractive one.

Sc. 33 Φίκιον See on 7}. 326 above, p. 106. Sc. 234 ἐπυκυρτώοντε

This form is quoted as a Boeotism by Thumb-Scherer,?? following Meister, who compares the -ώω inflexion seen in Boeotian Sajucovres.9" But Russo in his edition had already challenged Meister's explanation, suggesting instead that the form may be simply an analogical formation, and comparing it with the Homeric ὑπνώοντας at 42 344 etc.1 It seems very questionable therefore whether this form can be used as evidence for Boeotian dialect in the Shield. Sc. 302 λαγός

See on Th. 60 above, p. 104. It is not easy to draw well-defined conclusions from such a miscellaneous collection. We are hampered partly by uncertainty about the faithfulness of our text to the form in which the poems were first composed, and an even greater handicap is imposed by the inadequacy of the inscriptional evidence for the local dialects of the period which we are considering. Nevertheless enough information is available to allow some general comments to be made about these unHomeric word-forms which have been thought to show dialect features. In the first place it needs to be recognised that several of these distinctively Hesiodic forms can be plausibly explained without reference to dialect influence at all: a sufficient motive for their introduction is provided by analogy with traditional

words

and

expressions,

as at Th.

15,

487,

678,

Op. 22, 280, Sc. 7, 234, or merely the requirements of the 90 HGD II, 9.

100 Hk, 92.

101 On this and other Homeric

-ώω forms see Chantraine,

GH

I, 365 f.

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

119

context, as at Th. 200, or of the metre, as at Th. 178.

This

essential artificiality of Hesiod’s dialect is to be observed elsewhere in the un-Homeric parts of his language no less than the Homeric. Forexample, when we examine the occurrences of two un-Homerie verbs, Bpiaw and ὑμνέω, we find that one always appears in an uncontracted form (Th. 447, Op. 5 (bis)), while the other is regularly contracted (Th. 11, 33, 37, 48, 51,

70); but the explanation is obviously that the contracted βριᾷ, βριῶντα and the uncontracted ὑμνέουσαι, ὑμνέειν etc. are for metrical reasons awkward to use in the hexameter (except by lengthening the -e-, as at Op. 2 üuveiovoa.). Similarly with the uncontracted ἐπικροτέοντα Sc. 308 and contracted πλουτεῦντα Op. 313, both un-Homeric

words, the

choice of the poet is determined by the metre's exclusion of the alternative forms. When the forms which are likely to be artificial are left aside, the dialect forms proper which remain have very mixed affinities. They show remarkably little in common with Boeotian as known from other sources, and we should scarcely connect the poet with Boeotia at all, were 1t not for the autobiographical references in the poems and the later traditions of him. Among the features which have from time to time been cited by scholars as Boeotisms in Hesiod, the only reasonably sure candidate is xa and the adjective Φίκιον, and the use of these forms in their Hesiodic contexts implies little or nothing about the general influence of Boeotian dialect on his language. On the other hand several of the un-Homeric features present in Hesiod are not attested in Boeotian at all, such as τριηκόντων, réropa, and probably the -ἂν gen. plurals

(in Boeotian inscriptions in the article only).'°

Convenient as

102 See above, p. 106. Other proposed candidates as Boeotisms— Ἡσίοδον Th. 22, εἰρεῦσαι Th. 38, διδασκῆσαι Op. 64— are discussed and rejected by West, HT, 87 f.; Troxler connects Hesiod’s use of the dual, of -oıs beside -ovot, of the suffix -φι, and of thematic infinitives in -έμεν with the influence of

Boeotian common 103 See cussed in

in his dialect (SW H, 239) ; but since all these features are relatively in Homer, the suggestion is rather speculative. above, pp. 115 f., 103. The same applies to the features to be disChapters IX and X : see esp. below, pp. 131, 133, 147.

120

THE

LANGUAGE

OF HESIOD

it would be to adopt the straightforward hypothesis that Hesiod's language was that of the Ionic epic occasionally tinged by the influence of his own vernacular,?* this view must be rejected as inadequate. Its inadequacy has been recognised by most scholars, and we have seen that the usual account of Hesiod’s dialect presents a more complex picture. Thus Buck speaks of it as being ‘ substantially the same (sc. as Homer’s), but with some Aeolic

forms

not

used

in Homer,

also

some

Boeotian

Doric peculiarities.’ 1095. In the Prolegomena to his edition of the Theogony West similarly has suggested variety of possible origins for the peculiarly Hesiodie forms, grouping them under the heads ‘Ionic’, ‘ * Aeolic ’ and ‘ West

and

recent a wide dialect Attic’,

Greek ’, as well as certain combinations

of two or more of these.!^9 While many details regarding the place of particular forms within this scheme are open to question, the most serious doubt about these categories themselves revolves round whether it is appropriate to include West Greek among them. In her paper ‘ “ Doric ” features in the language of Hesiod " Mrs. A. Morpurgo Davies has argued that alternative explanations are possible for the features accepted in the past as being of West Greek origin.!°” Her arguments are for the most part irresistible, and it is certainly true that the list of ‘ Dorisms’ collected by Rzach and his predecessors in the nineteenth century includes features which are either shared with other dialects or otherwise explicable without adducing West Greek as their source. 104 of. ' Hesiod wrote in a dialect which is in the main the Ionian of Homer with some admixture of Boeotian’ (P. Harvey, Oxford Companion to Classical Literature, 208). 105 GD, 15; cf. Gallavotti-Ronconi, LLO, 145, and Troxler, SWH, 237 f. 10 HT, 79-91. 107

Glotta xlii, 1964, 138-65.

Part of the confusion has arisen for historical

reasons: nineteenth century understanding of the Greek dialects owed much to ancient scholiasts and commentators who applied tbe naıne ‘ Doric’ more freely than is now acceptable (cf. above, p. 18 n. 27). Hainsworth, TPhS 1967, 62-76, points out the defects of the ancient dialect categories and the unsatisfactory foundations on which they were based.

UN-HOMERIC

WORD-FORMS

AND

THEIR

DIALECT

SOURCES

191

The fact remains that we are unable to explain all the unHomeric dialect forms in Hesiod unless we accept that his language has been influenced by mainland dialects other than Boeotian and Attic-[onic.

The form réropa is obstinate, and

once that is conceded as a West Greek feature the way is open for the explanation of other forms as having originated from the same source.!1?? We shall also argue in Chapter X that the problem posed by the a- and o-stem acc. plurals—the most recurrent of all these distinctive features—appears insoluble without some reference to dialect influence, and short-vowel forms are attested only m Thessalıan, Arcadian, and West

Greek dialects. But first we need to look at another aspect of the difference between Hesiodic and Homerie language, for the understanding of which the whole matter of tradition and innovation, as well as of dialect, 15 of cardinal importance. 108 This is especially true of those features which may be described as possible archaisms in any dialect, e.g. é8ov Th. 30, and which survived longer in West Greek than elsewhere. Other examples are ἀποδρέπεν Op. 611 and τεῖδε Op. 635, though the reading in both places is far from certain : see West, HT,

86 f.

CHAPTER IX

TRADITIONAL ALTERNATIVES AND CHOICE BETWEEN THEM

HESIOD'S

OnE of the characteristics of Homeric language is the availability of alternative forms, of differing value for metrical purposes, which originated in different dialects or periods. It cannot be doubted that these alternatives were a feature of the mixed and artificial dialect in which the Ilad and the Odyssey were originally composed, and their presence is readily explained by reference to the long poetic tradition culminating in the work of Ionian singers, who made use of

linguistic material which had first entered the tradition in its earlier ‘ Achaean ’ or ‘ Aeolic’ phases.! Hesiod makes use of the same alternatives, but the manner in which he exercises his

choice between them has important consequences for our understanding of his relationship to other early hexameter poetry. A good example is provided by the genitive singular of o-stems.

Here we find two endings,

-o1o and

-ov, of which

the former must be the basis of the endings occurring dialects, and the latter is the Ionic contraction from a labie -oto or -oo and therefore belongs to a more stage of linguistic development. Troxler? has recently an account

of Hesiod’s

behaviour

in this matter,

in all disylrecent given

and

has

shown that the proportion of forms with certain -ov is con siderably higher than in Homer, and is especially high in the Works and Days.® This general pattern is consistently reproduced when other similar sets of alternatives are studied, and ! Summary accounts of the Homeric dialect mixture are given by Chantraine in Introduction ἃ l'Iliade (ed. P. Mazon), Ch. IV, esp. 105-115, and by Palmer in A Companion to Homer (ed. Wace and Stubbings), Ch. IV,

esp. 97-106. * SWH,

examples

52-9,

though

his

lists are

in -ov occurring at Th. 362,

not 687,

wholly

accurate:

806, Op. 443, and

they

omit

those in -o.o

at Th. 221 (bis) and 331; but they include Op. 551 ὑψοῦ which is an adverb (Schwyzer,

GG I, 621).

xxii, 1922,

165—71

On this feature in Homer

and Chantraine,

GH I, 45-7,

seo V. Magnien,

MSL

193 f.

? Expressed as a percentage of all o-stem gen. sg. forms, the proportions are:

Jl. 20%,

Od. 24%,

Th. 38%,

Op. 50%,

Sc. 3595.

HESIOD'S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

193

we shall now focus attention on three such instances. But in addition to making a general comparison between Hesiodic and Homerie usage we shall need to examine parts of the material more closely, since we are chiefly interested in considering what the evidence implies about Hesiod's knowledge of traditional hexameter poetry and about the ways in which the language of his own work was influenced by it. The three features to be diseussed are (1) the genitive singular of masculine a-stems, (2) the genitive plural of a-stems, and (3) the presence and absence of traces of initial digamma—a well-worn theme, but one which is of direct relevance here. (1) The genitive singular of masculine a-stems

In both Homer a-stems appears -üo or with the vowel Where initial vowel,

and Hesiod the genitive singular of masculine either with the Aeolic (or archaic) ending Ionic -ew (with synizesis), written -w after a the Ionic ending occurs before a following

as in Πηληϊάδεω

Ἀχιλῆος

A

1, πνοιῇ Bopew

ἀλεγεινῇ & 395, the Aeolic -ào may be substituted, with elision of the final -o, thus reading IImAniada’, Bopea’ etc. But where -ew, -w is found before a consonant or at the end of a

line it has to stand, and Chantraine * quotes over 20 such examples from Homer, though in a few of these the Aeolic ending might be substituted by making some small change in the text. The distribution of these endings in Hesiod and Homer is given in TABLE 11.5 The numbers

are small, but one fact is obvious:

whereas

in Homer it is possible to read the Aeolie -ao ending in over 90%, of the examples (191 out of 211 in IL., 91 out of 96 in Od.), the comparable proportion for the Theogony and the Works and Days is much lower (5 out of 12 in Th., 5 out of 9 in Op., 1.6. The Shield behaves differently, but three around 50%). examples cannot be regarded as sufficient evidence on which to 4 GH J, 70. 5 Troxler’s lists of the examples in Th. and Op. (SWH, 59 f.) need to be completed by the addition of those in -ew at Th. 832 and Op. 71, and in -ao at Op. 99. For details of the examples from Sc. see below, p. 125. The figures for Homer

are taken from Chantraine, GH I, 65, 69 f., 200 f.

124

THE LANGUAGE

TABLE

11.

Forms of the genitive singular of masculine a-stems in Homer and, Hesiod -ao

-εω

before

H.

OF HESIOD

-o

-€o)

a vowel

κω

| before ἃ consonant | Total

|167|

21

3

16

4

211

Od. | m|

13

1

5

0

96

Th. |

5

0

0

6*

1

12

Op |

5

00

0

2

2

9

Sc.

3

0

0

0

0

3

*not counting ’Aidew in Th. 774 ; this line (= Th. 768) is omitted from most mss.: see West ad loc.

base any conclusion, especially when they are contrasted with the low proportion of Aeolie genitive plurals in -àcv which is found in the SAeld.$ Although it could be argued that none of these figures for Hesiod is of much significance when the examples are so few, they may properly be regarded as a part of the cumulative evidence which we are considering in this chapter. It is to be noted that the -do genitives in the Theogony and the Works and. Days belong without exception to forms which occur with an -ào genitive in Homer, often in the same or a similar context. They are, in order of their occurrence : ᾿Ωκεανοῖο βαθυρρείταο Th. 265, cf. BaÜvppeirao μέγα σθένος '"keavoto 195 ; Ais νεφεληγερέταο | Th. 730, 944, Op. 99,

cf. the Y 10,

same

phrase

at E

631,

736,

© 387,

O

154,

499 ;

Ainrao διοτρεφέος Th. 992, cf. Aijrao in other positions, « 137, μ 70, and Atovrrao διοτρεφέος N 427, in this position ; "Hediov .. . Ὑπεριονίδαο

| Th. 1011,

cf. “HeXiov .. . “Ὑπεριονίδαο ἄνακτος | u 176 ; ® See below, pp. 127 f. and Table 12.

HESIOD’S

"Aidao

CHOICE

| Op.

TRADITIONAL

ALTERNATIVES

195

153,

cf. Aldao 15x of line) ; Οἰδιπόδαο

BETWEEN

| Op.

II. (3 at end of line), 18x Od. (5 at end 163,

cf. Οἰδιπόδαο | Y^ 679 (at end of line) ; Βορέαο Op. 506, 547, cf. Βορέαο 4x Il., 3x Od., including the phrase Βορέαο πεσόντος | £ 475 (cf. Op. BAT).

The same is true of the first example in the Shield, Ἀπόλλωνος ἑκατηβελέταο ἄνακτος | Sc. 100, exactly as at A 75; and while the name AAkeiöns does not occur in Homer, the

phrase ἀμύμονος Adkeidao | Sc. 112 is closely parallel to ἀμύμονος Ἀγχίσαο | E 247. The remaining example, φλεγύαο Sc. 134, 1s a word which does not occur in Homer or anywhere else in Greek literature.’ This then is the only -ào genitive in the Hesiodic poems for which we cannot adduce evidence for the possible influence of traditional poetry. Traditional influence is much less apparent in Hesiod's use of genitives In -ew, -w. Out of the eleven examples two are words which never occur in Homer in any form: ἐριβρύχεω Th. 832 and ἀβούτεω Op. 451. There are four instances of -ew genitives which never occur in Homer in an -ew form: Κρονίδεω

Th.

572

and

Op.

ΤΊ, apyéorew

Th.

870, Ainrew

Th. 994 (cf. Homer's apyeoráo A 306, ᾧ 334, Aiyrao κ 137, p. 70). The form Bopéw Th. 870, Op. 518, 553 (all before a consonant), occurs twice in Il. (& 395, V/ 692) and once in Od. (€ 533), but always before a vowel, where Bopéao (elsewhere 4x Il., 3x Od.) might be substituted. This leaves only Aldew Th. 311, 768, which occurs four times in Homer, but three of the examples there are before a vowel (© 16, « 512, 4 17), the

7 On the meaning of this pbrase, see below, p. 182. ? Tt should be regarded as à noun, apparently serving as the name of an

eagle (see Hesychius s.v.), rather than as an epithet as suggested by Liddell and Scott (μόρφνος is almost certainly an adjective at Sc. 134 and not a noun: cf. 2 316). Perhaps it belongs to the same class of descriptive animal-names as those found in the Works and Days (see above, pp. 111-3).

126

THE LANGUAGE OF HESIOD

only example before a consonant being found at y 322 in the passage where Odysseus gives a résumé of his adventures to Penelope,—a passage often suspected as being the work of a late hand.? (2) The genitive plural of a-stems

The text of Hesiod shows five endings in genitive plural forms from a-stems : (1) -ἄων, e.g. Μουσάων Th. 1, μελισσάων Op. 305, θηλυτεράων $c. 4;

(2) -écv, only πυλέων Sc. 246 ; (3) -έων, e.g. ἁπασέων Th. 19, ayopewv Op. 30, εὐχωλέων Sc. 68 ; (4) -àv, e.g. τῶν Th. 39, παρειῶν Sc. 267 ;

(5) -äv, only θεᾶν Th. 41 and perhaps 129, μελιᾶν Op. 145. The most

common

of these, in Hesiod as in Homer,

is the

Aeolic (or archaic) -ἄων (1), which stands closest to IE *-@söm, while the remaining four all show more extensive changes. The ending -έων, with Ionic à > ἡ and then nw > ew, is once

in Hesiod scanned -ἔων (2) but most usually -έων (3) with synizesis. The distinction between this form in -έων (3) and the contracted -ῶν (4) might appear to be largely one of orthography, but a genuine difference in pronunciation is suggested by the fact that the latter 1s found generally after -ı- in Homer 19 and in some Hesiodic examples, namely παρειῶν Sc. 267 and δμῳῶν Sc. 276. Otherwise the contracted -óv ending occurs mostly in pronouns and adjectives, where the existence of the masculine genitive in τῶν may have influenced the form of the feminine ; the Homeric examples are listed by Chantraine,4 and Hesiod provides us with τῶν Th. 39, 549, 910, Sc. 260, στιβαρῶν Th. T15, λιγυρῶν Sc. 278, and αὐτῶν

in all mss. at Sc. 377 and in several at Sc. 237, though Rzach reads αὐτέων in both places. As this last example shows, there ? e.g. by Monro and Bérard ; see Stanford's note on ψ 310—41. 19 See Chantraine, GH I, 65. 11 GH T, 201 n. 1; see also below, n. 15.

HESIOD’S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

197

is some fluctuation in the ms. tradition, but, in the figures given below, Rzach's readings have been followed, since the

main point to emerge—i.e. the lower proportion of -ἄων endings in Hesiod when compared with Homer—is protected by the metre and is not affected by the choice between -έων and -@v when this arises. The remaining ending -áv (6), with contraction of a + w > à, is totally absent from Homer,

and has already been discussed in Chapter VIII. The distribution of these genitive plural endings in Homer may be gathered from Chantraine,'* and we add the figures for the Hesiodie poems.!5 TABLE 12 shows that while the -ἄων ending is still the commonest, the proportion of forms TABLE

12.

Forms of the genitive plural of a-stems in Homer and Hesiod

(D | (2) | (8) | (ὦ | (5) | Total Il.

183

2

2]

15

0

221

Od. | 123

1

19

10

0

153

Th. | 24

0

| 10

4|

1

39

12 A further instance cccurs at Op. 264, where the mss. give σκολιῶν δὲ δικῶν beside σκολιέων δὲ δικέων in Pap. Oxyr. 1090, whose reading Rzach adopts;

cf. West, HT, S3 n. 4.

13 See above, p. 103. 14 GH

I, 201, and for -àv ibid., 65.

15 There are a few doubtful examples, not included in our count, where ἃ masc. form serves as a feminine: αἰθομένων Sc. 275, κυλινδομένων Sc. 378 clearly refer to fem. subjects; νισομένων Th. 71 could agree with ποδῶν, masc., but scems more likely to refer directly to the Muses; ἀργαλέων Op. 92 might be understood as a fcm. form with hyphaeresis, from *apyaleewv (see Lejeune, T'PG, 222 f. for parallels), but it is probably simpler to regard it as ἃ masc. form and include it with the others. Chantraine (GH I, 201 n. 1) quotes similar examples from Homer and suggests that their accentuation may be an Attic feature in the text, thus implying that they may have replaced fein. forms in -έων or -àv. If they were included in the count here, thcy would serve to reinforce the conclusions that will be drawn.

128

THE

LANGUAGE

OF HESIOD

containing it is not so high in Hesiod as in Homer.

The

Iliad’s 183 out of 221 represents 83%, and the Odyssey’s 123 out of 153 15 a little over 80%. The Hesiodie percentages are much lower: 62% in Th., 55% in Op. and 50% in Se. Again we must examine the examples themselves, 1° in order to see how many of the Hesiodic forms in each of the categories are likely to havc been taken over directly from the stock of traditional material which the poet inherited from his predecessors. Of the 24 instances of -ἄων genitives in the Theogony, cight are in phrases which occur in Homer and which almost

certainly did not originate in Hesiod: 969,

1004,

1017;

cf. 7x

IL,

26x

Th. 548, 893, 993; cf. 5x Il.,

Od.;

δῖα θεάων | Th. 376, θεῶν αἰειγενετάων

29x Od. ; γυναικῶν θηλυτεράων

|

Th. 590; cf. A 386, ψ 166. A further six examples occur in Homer in the same -ἄων form and in the same position in the line: | Μουσάων Th. 1, 100, cf. A 604; θεάων | Th. 103, 240, cf. 9x IL, 28x Od.; θεάων, in third and fourth feet, Th. 366, cf. Σ 364, 2 615; and ἱεράων | Th. 1015, cf. B 625, X 323.

Four more are places where a Homcric form in -awy is used, though not in the same place as in Homer: Movodwy, in second and third feet, Th. 36, 93; θεάων, in first and second fect, Th. 965; and βολάων, in third and fourth fcet, Th. 683,

cf. -ácv

p 283. which

εὐεργεσιάων

This leaves six examples of a-stem genitives in are

Th.

not

found

503,

in

Homer:

δυσφροσυνάων

μερμηράων

Th.

528,

Th.

55,

κρατεράων

Th. 683, τιμάων Th. 882, and δερκομενάων Th. 910. Of the six cxamplcs in the Works and Days none occurs in a phrase which !$ Here also changes need to the forms in -à«v should be those in -éwv ought to include Th. 235, though sec below p.

to be made in Troxler's lists (SW, 60 f.) : added the examples at Th. 548, 893 and 993 ; δυσφροσυνέων T. 102 and possibly θεμιστέων 130 n. 18; under forms in -àv it should be

noted that τῶν, fem., oceurs at Th. 549 and 910, as well as at T'h. 39 which

Troxler mentions. But σφέων and χρεέων should be deleted from his list : no form from σφεῖς is ever declined as an a-stem, and while χρεέων (Rzach ; xpeiwv libri) might look as if it is from χρείη, the lattor is never found in epie, whereas tho ntr. s-stem χρεῖος or χρέος is frequent in Homer and oceurs elsewhere in Op.: ef. esp. ntr. pl. aee. χρέα at Op. 404 in a similar eontext to χρεέων at Op. 647. Details of the 18 examples from the Shield are included in the diseussion below.

HESIOD’S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

129

is obviously traditional, but [τάων Op. 826 is found in both JI. and Od. at the beginning of the linc (e.g. J 146, 8 121). The form μελισσάων Op. 305 occurs in a different position at B 87. The remaining four forms are not found in Homer: ἐπιτελλομενάων

| Op.

383, δυσομενάων

| Op.

384,

σβεννυμενάων

|

Op. 590 and κρηνάων Op. 758. The Shield’s nine examples include five which seem to belong to traditional expressions : γυναικῶν θηλυτεράων | Sc. 10 and γυναικῶν... θηλυτεράων |

Sc. 4 (cf. above on Th. 590) ; δῖα θεάων | Sc. 338 (cf. above on Th. 376 etc.) ; and “απιθάων αἰχμητάων | Sc. 178: these two -ἄων forms occur together in the same position at M 128, and αἰχμητάων alone is found at the end of the line a further 7 x Il.

and 2x Od.

Both | τάων Sc. 6 and ἀλλάων Sc. 260 occur in

Homer in the same positions.

This leaves only Τηλεβοάων |

Sc. 19 and the curious cvavedwy | Sc. 7 which has already been discussed in Chapter VIII.!? It therefore appears that out of the 39 Hesiodic examples of a-stem genitives in -awy no less than 27 are forms which occur in Homer, 22 of them being in the same position in the line, including 13 in repeated phrases. It is notable also that all but two of the twelve un-Homeric examples show the -ἄων ending in the sixth foot, where it is particularly frequent in traditional phrases. The Works and Days appears at first sight to contain a lower proportion of Homeric -ἄων forms, but it may be observed that three out of the four un-Homeric examples are middle participles at the end of the line (ef. δερκομενάων | Th. 910, ἐρχομενάων | B 88, ναιομενάων | I 400, βοσκομενάων | E 162 ete.).

In contrast, out of the 14 examples of -έων in the Theogony and the Works and Days, the nearest approach to what might be

a traditional

phrase

is προφερεστάτη

Th. 19 and 361, cf. φιλτάτη ἐστὶν occurrence of ἁπασέων in Homer, found containing masculine forms : H 153, φίλτατοι ἦσαν ἁπάντων |

ἐστὶν

ἁπασέων

ἁπασέων | 0 284 (the only though similar phrases are νεώτατος ἔσκον ἁπάντων | I 586). Another isolated

1? See above, pp. 116-8.

|

130

THE

LANGUAGE

OF HESIOD

parallel is ἐκ κεφαλέων standing in the second and third feet at Th. 828 as at Γ 273. Only πασέων and πυλέων occur more than once in Homer: but πασέων 2x Il., 4x Od. is never found at the start of a line, as at Th. 828, and πυλέων 2 x Il. is scanned

there as an anapaest (so too at Sc. 246), not with synizesis as at Th. 141 and 773. Of the remaining eight examples only two occur in Homer in an -&wv form, and neither of these occupies the same place in the line as in Hesiod: ἀγορέων Op. 30, I 441, and Νυμφέων

Th. 130, μ 318.

The other -éwy forms, all

un-Homeric, are Movoewv Th. 94, óvodpoavvéov θεμιστέων 18. Th. 235, σκολιέων and Gucéov Op.

Th. 102, 264, and

οἰνέων Op. 572. Of the four -ὧν examples in 7^., three are τῶν which is relatively frequent in Homer, but the fourth is the un-Homeric form στιβαρῶν Th. 115. The Shield shows rather more evidence of traditional

influence. The form παρειῶν | Sc. 267 occurs in Homer always in the same position at the end of the line 1 x 7L., 4x Od. (ef. esp. the similar context at r 190 f.) ; αὐτέων | also stands at the end of the line, Sc. 237 and 377, and its only Homeric

occurrence is οἱ δ᾽ ὑπὲρ αὐτέων | M 424, in the same phrase as at Sc. 237 ; δμῳῶν, Sc. 276 is found in the same position at 7 121; πυλέων occurs with the same scansion as in Sc. 246, though in a different position, at H 1 and M 340. This leaves the Homeric τῶν Se. 260 and three forms not found in Homer: εὐχωλέων Sc. 68, λιγυρῶν Sc. 278 and Bawovaéwv Sc. 232. It seems evident from our study of these two sets of endings that, while Hesiod has the metrical alternatives at his disposal and makes use of all of them, the older, uncontracted endings

in -ἄο, -ác become less frequent and many of those which occur are tied to traditional phrases, whereas the use of the younger endings in -ew, -w and -éuv, -dv is extended, in exactly the same way as -ov has been shown to gain ground at the 18 West

reads θεμίστων,

athematic,

against the mss. here:

Wyatt defends the a-stem form θεμιστέων as Hesiodic: 1966, 631 f.

see HT,

see TAPA

234.

xevü,

HESIOD'S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

13]

expense of -o1o. This conclusion is of particular interest when considered in the light of the dialect evidence. As far as Homer is concerned, we find a straightforward opposition between the traditional endings -ào and -ἄων on the one hand and the innovating -ew, -w and éov, -àv on the other, and the

situation there is fully explicable if we suppose that Ionic

ἀοιδοί were working with traditional material containing many archaic features of language, into which they occasionally introduced new forms from their own contemporary speech. But for Hesiod this explanation is not possible : the genitives in -ew and -ἔων belonged to Ionia and not to mainland Greece. Indeed it is to be supposed that Boeotian speakers long after Hesiod's time continued to use the endings -do and -ἄων,19 and it would therefore have been no surprise to find that these uncontracted Aeolic endings were extended at the expensc of the Ionic. Instead, as we see, the reverse happens. Not only that, but the -ew, -ewv endings often appcar in innovating contexts,

while

the

forms

with

Aecolic

-ao,

-ἄων

occur

primarily in traditional expressions. It might with good reason be objected that, where the number of examples is as small as ıt has been here, their evidence ought not to be pressed too far. But one may be confident that the data have not misled us, when the examina-

tion of ἃ very different kind parallel in certain essential conclusions. The subject to deeply complicated, but the reasonable compass if it is directly relevant to the issues

of feature, which is nevertheless respects, points to the same which we now turn attention is discussion can be kept within a confined to matters which are which wc have been considering.

19 Thumb-Scherer, HGD II, 27. 2° In addition to the evidence cited above (pp. 125 f., 129 f.) in support of this point, one may mention those lines where Hesiod, in adapting a traditional expression to a new case or a new reference, introduces forms with thc younger endings (see the examples discussed by Hoekstra, Mnemosyne x, 1957, 205). There are also places where Hesiod's introduction of the Ionic endings appears to have been prompted by a ‘parallel of sound’, e.g. | Νυμφέων

Th. 130 beside B 130, 535, 626 (contrast ζ 123 !), πυλέων Th. 141,

773 beside a 380, B 145, h. Dem. 342.

132

THE LANGUAGE

OF HESIOD

(3) Traces of initial digamma

The phoneme x, which Greek inherited from Indo-European, had already been lost from Ionic by the time of the earliest inscriptions, but it persisted for several centuries in some other dialects, including Boeotian, where 1t is written with the letter F.?4

Our texts of Hesiod, like those of Homer, give no

written indication of it, but where an initial f might be expected from the history of a word, traces of 1t often appear to remain, either where hiatus occurs before initial vowels, or

where an initial consonant or double consonant seems to be required in order to make an otherwise light syllable heavy. There are also many places which contain no such rhythmical irregularity but where F could easily be inserted in the text without affecting the scansion. There remain other places

however where the insertion of ^ would at once upset the metrical pattern.?? Hesiod's practice in thus ‘ observing ' or ‘ neglecting ’ initial F has been studied by a number of scholars, and there is general agreement that his work shows a higher proportion of examples where F cannot be restored to the text than is found in Homer? Thus F. Devantier's very thorough examination of the problem concludes (a) that Hesiod in both the Theogony (with about three “ observances’ of F to every one ‘ neglect ’) and the Works and Days (c. 2: 1) shows neglect of F rather more frequently than is found in Homer (c. 6:1); and (b) that the behaviour of the Shield (c. 5 : 1), which stands closer to Homer in this matter, is largely to be explained by its higher proportion of Homeric phraseology.?4 Similarly Troxler 21 Buck, GD, 46 ff.; Thumb-Scherer, HGD II, 29. 22 See Chantraine, GH I, Ch. IX, esp. 117 f.

?3 [t is true that Flach, writing in 1876, concluded that F was observed in Hesiod

55 times

for every

one

example of its neglect

(DDH,

41), but

this

figure was arrived at only by a distortion of the evidence, counting all equivocal examples as containing F, while suppressing many of the places whicb would preclude F, either by emendation or on the grounds that the lines containing them were spurious. Flach’s approach was well criticised by Devantier : see esp. pp. 29-34 of Part II (1894) of the work cited in next note.

24 Die Spuren des anlautenden Digamma bei Hesiod III, 1897, 51 f.

HES1OD’S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

133

in his recent study reckons that Hesiod observes f in about 71:5% of places (67% when Homeric phrases are left out of account), beside a figure for Homer, derived from Hartel, of about 84-5%.25 Troxler’s work does not include the Shield,

but he finds that the Works and Days contains a higher proportion of examples with neglected F than the Theogony, and this also agrees with the conclusions reached by earlier scholars.?® At first sight the difference between Hesiodic and Homeric practice here may seem relatively slight; but it must be pointed out that the divergence would be more strongly marked if the text of Hesiod were taken as it has usually been transmitted in the mss. Rzach may not have gone so far as some nineteenth century editors in producing a text where initial F is ‘observed’ as frequently as possible, but undoubtedly in his choice of readings and his adoption of emendations he has often been influenced by this consideration.?" While in a few of these places Rzach’s judgment of what first stood in the text might be defended on other grounds, it is evident that the presence in Hesiod of a reading where F is neglected is not in itself a sufficient justification for altering the text. The fact has to be faced that Hesiod in this matter behaves differently from what might be expected of 8 Boeotian poet at the period in question. A consideration of the examples themselves shows that (a) where Hesiod ‘ observes’ an initial 7, the phrase often appears, from comparison with the Homeric evidence, to be traditional, and (b) in many places the ‘ neglect’ of an initial F is found in a context which may be regarded as peculiarly 25 SW

H, 46.

26 West, HT‘, 99, gives a convenient summary of the figures reached by Devantier and Paues. 27 Apart from his omission of ephelcystic -v’s mentioned above (p. 16), as for example at Op. 282 (cf. Th. 232 !), we may note the following places in the Works and Days where he adopts a reading against most —usually all— of the mss.: Op. 137, 316, 364, 407, 409, 412, 428, 443, 494 (cf. & 381 ἢ), 521,

578,

610.

Troxler

(SWH,

36

ff.), following

Rzach’s

text,

naturally

lists all these examples as positive indications of the ' observance ’ of F, and this undoubtedly distorts the total picture.

184

THE LANGUAGE

Hesiodic. here:

OF HESIOD

The first of these points need not be elaborated

the detailed

accounts

of F in Hesiod,

like those

of

Devantier and Troxler, give many examples, and try to take 1t into account in their conclusions.?5 But it is worth illustrating the second point, which serves as a corollary to the first. Here are some examples where we find neglected F in close association with a feature which we have reason to think may be a Hesiodic innovation : (1) in association with an ‘indirect name’: $eÀov

ἐργάζονται

Th.

440,

where

γλαυκὴν δυσπέμ-

γλαυκή

=

πόντος

“the

sea’; 3. ἀνόστεος ὃν πόδα Op. 524 ; τρίποδι βροτοί (or βροτῷ) ἶσοι Op. 533 ; 39 (ii) in the same line as an acc. pluralin -às : εὖτ᾽ àv δ᾽ ἑξήκοντα μετὰ τροπὰς ἠελίοιο Op. 564 ; 31 (111) in association with absence of compensatory lengthening after loss of postconsonantal F: καλὸν εἶδος Op. 63; Suwdek«áumvov: ἴσον Op. 752 ; 32 (iv) in association with un-Homeric vocabulary: iv ᾿Εσπερίδες Th. 275; πρόπαρ ‘Eorepiöwv Th. 518; ξυνήονας ἔργων Th. 595, 601; ἀσχέτου, ὄσσαν ἀγαύρου Th. 832; ἥσυχοι ἔργ᾽ Op. 119; πρωιηρότῃ ἰσοφαρίζοι Op. 490; βίβλινος οἶνος Op. 589; ποτ᾽ οἰνοχόην Op. 144; στίγματα δ᾽ ws Sc. 166; οὐδ᾽ ᾿Ιφικλεΐδην Sc. 111 ; 83

(v) in association with vocabulary used in an un-Homeric 23 Thus Troxler in his list indicates those references where Hesiod is using traditional epic material by enclosing them in brackets (see SWH, 36). The difficulty here however is to know where to draw the line: 6.5. he brackets τοῖο ἄνακτος | T'h. 859 because it occurs in this position 1x IL, 2x Od.; but he does not bracket τοῖο ἄνακτος Th. 493 in the 2nd and 3rd feet, even though there are many comparable phrases in Homer with -oıo ἄνακτος in the same position (M 11, 6 270, £ 366, τ 358) ; nor does he bracket

᾿Ημαθίωνα ἄνακτα the ‘observation’

| Th. 985, though it is clearly analogous, as far as of F is concerned, to the Homeric Κρονίωνα ἄνακτα |,

Ποσειδάωνα

| ete.

ἄνακτα

29 See above, pp. 52 f.

30 See above, pp. 111-3.

31 See below, pp. 152f.,176f.

38 See above, p. 107.

33 In his Index Hesiodeus Paulson marks with an asterisk words not occurring in Homer. Studies of the un-Homeric words have been made in 1866 by Fietkau (see Bibliography) and more recently by Troxler (SIVH, esp. 188-233), neither of whom

includes proper names.

HESIOD’S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

135

sense : Ἵμερος οἰκί᾽ Th. 64, with Ἵμερος personified ; ἐπίκλοπον ἦθος Op. 67, 78, ὥς κ᾽ ἤθεα Op. 699, with ἦθος meaning ‘ behaviour ’ rather than ‘ accustomed haunts ' as in Homer ; 34

(vi) in association with a departure from traditional wordorder: ἐπὶ δ᾽ αἴθοπα πινέμεν οἶνον | Op. 592, beside αἴθοπα οἶνον | Op. 724 (8x Il., 8x Od.).35 What is to be inferred from this about Hesiod’s pronunciation of Fin his own poetry ? It is sometimes supposed that for Homer and Ionic ἀοιδοί generally the sound was optional, that is, that they could choose to pronounce it either in the archaic manner, e.g. Fotvos, or as in their own daily speech, e.g. οἶνος, according to the metrical needs of the context.?9 Tf this were so, it would be natural to assume that ἀοιδοί from other parts of Greece who pronounced F in their native dialect would do so wherever possible in poetical recitation; thus Schwyzer writes: ‘Mindestens Rhapsoden, deren Mundart F noch besass, werden es wohl beim Vortrag der Gedichte gesprochen haben.’ 3” It appears certain however that this situation did not prevail at any rate by the time that Hesiod was composing.?® For if F was still generally pronounced in hexameter 31 See Troxler, SWH, 168. 35 See Hoekstra, HM FP, 27, with his reference to Severyns.

36 It is to be noted that Chantraine, while he writes that ‘ tout se passe comme si les derniers aódes avaient disposé d'un double jeu de formes, les unes plus archaiques, les autres plus récentes' (in Mazon, Introduction ἃ lIliade, 104), is careful not to imply that F was still pronounced in the older formulas ; see his discussion,

scholars Stanford between probably

op. cit. 102-4,

and

cf. GH

I, 152-7.

Other

however assume that F was still present in pronunciation: thus writes that ‘in Homer’s era the sound was in a transition stage pronunciation and non-pronunciation ', and suggests that F was written in the oldest Homeric texts (The Odyssey of Homer I, lvi f.) ;

cf. Schwyzer, GG I, 229, where it is said that for Homer F was a ‘ mitgeführte

Altertümlichkeit ’; the same implication is present in Palmer's discussion, where F is described as ‘a traditional “‘ poetical ’’ colour on the palette of the Homeric artist’ (CH, 101). Sturtevant, PGL, Graeca, 47, do not commit themselves on the point.

65,

and

Allen,

Vox

37 GG T, 220 ; cf. K. J. Dover's discussion of the problem of digamma in Tyrtaeus, Archiloque (Fondation Hardt Entretiens x), esp. 192 f. 38 The view adopted here is in broad agreement with that of Meister, HK, 199-202,

to

which

West

refers,

though

his own

brief discussion

(HT,

91,

99 f.) seems not to draw a sufficiently clear distinction between F as a spoken sound and F as the historical cause of various prosodical irregularities.

136

THE

LANGUAGE

OF HESIOD

poetry, and neglected only where new or modified expressions had entered the tradition after the loss of F from spoken Tonic, there seems to be no sufficient reason either to account

for the relatively high proportion of ‘ neglects’ to ‘ observances ' in Hesiod when compared with Homer, or to explain why so many examples of neglected F should occur, as we have just seen, in peculiarly ‘ Hesiodic ' contexts rather than in phrases which might have been originally coined by Ionic ἀοιδοί.

Moreover, if F for Hesiod still had consonantal force wherever

the metre allowed in poetic recitation, it is very difficult to account for some of his adaptations of traditional phrases. Thus the stem ἠχ- admits the restoration of an initial F wherever it occurs in the Homeric poems,? as in δώματα

(F)nyjev7a | ὃ 72;

yet this phrase appears to have been

known to Hesiod without a F, since he makes use of the hiatus

in his own adaptation of it:

δόμοι

ἠχήεντες

| Th. 767.40

Similarly éAikexp, ἑλικῶπις might always be pronounced with initial F in Homer; ! but the formula ἑλικώπιδα κούρην, which occurs there only in the phrase δόμεναι (F)eAuccómióa. κούρην | A 98, is treated by Hesiod at Th. 307 as having an

initial vowel:

ἀνομόν θ᾽ ἑλικώπιδι

κούρῃ

|.?

A further

example is provided by the perfect participle of εἴκω 43: a glance at a list of the Homeric occurrences makes it clear that this form had taken its place in the hexameter tradition at a time when it was pronounced with an initial consonant. The nom. sg. ἐοικώς 19x

Il.,

8x Od., invariably stands at the end

of the line after a short final vowel, which almost always 44 belongs to a dat. sg. in -«: 39.1.6, in ἠχή Chantraine,

GH

and ἠχήεις,

νυκτὶ ἐοικώς

though

not

in

the

| A 4T, vit ἐοικώς compound

δυσηχής : see

I, 140 witb n. 2.

40 Likewise there is shortening before the hiatus in τέρπεται ὃν κατὰ θυμόν | Op. 358 beside ἔλπετο ὃν κατὰ θυμόν | N 8, ete.; see Hoekstra, Mnemosyne x, 1957, 207. 4 GH Y, 132. *? θ᾽ is omitted by a few mss., but it seems necessary in the context; the other instances of the formula are equivoeal: TA. 298, 998, Fr. 75.15 M-W. *? Other forms of this verb in Homer are less stereotyped in use, and include a few examples where F cannot be restored to the text: see GH I, 129. 44 Exceptions at ® 600, v 80.

|

HESIOD'S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

127

B 20, etc., reflecting an older νυκτὶ ξεξοικώς | etc. Similarly acc. sg., 1x IL, bx Od., and dat. sg., 1x Od., always occur

after -V, e.g. παῖδα ἐοικότα E 800. In the plural, ἐοικότες 14 x IL, 4x Od., characteristically follows a dat. pl.: παισὶν

ἐοικότες B 337, φύλλοισιν ἐοικότες Β 800, etc., and while the

transmitted texts give ephelcystic -v in these places, it is not difficult to suppose that, as in the singular, the usage goes back to a time when F had consonantal force, e.g. παισὲ FeFowóres, which may thus be brought into line with μῦθοί ye ἐοικότες y 124, the only place where anything other than a dat. pl. precedes ; so also acc. pl. σιάλοισιν ἐοικότας 1X Od., ntr. pl. νιφάδεσσιν ἐοικότα lx

Il., dual σύεσσιν ἐοικότε 1x Il.

forms always oecur in the same positions in in bth-6th foot, all other cases in 3rd-4th keeps them only in these positions ; but merely in phrases of the traditional type,

These

the line (nom. sg. foot), and Hesiod he uses them not as -ovrı éoucds |

Sc. 215, -ovow and -eoow ἐοικότα 45 Th. 584, 834, but also in

οὐδὲν

ἐοικός

[46

Th.

295,

and

γυναῖκες

ἐοικότα

τέκνα

Op. 235, where there can be no initial consonant. Hesiod’s readiness to use the participle after a closed syllable here strongly suggests that he already knew the Homeric phrases παισὶν

ἐοικότες,

νιφάδεσσιν

ἐοικότα etc., in their forms with

ephelcystic -v. These examples cannot easily be explained if we take the view that f was still pronounced as a consonant in recitation wherever it was possible to do so. If however we follow Parry in believing that ' the digamma was lost in the diction of early Greek heroic poetry neither sooner nor later than it was lost in the daily speech ’ of the Ionian singers, 7 the situation in Hesiod at once becomes intelligible. The traditional style in which he was composing would then not be one in which F was occasionally neglected but usually observed in pronunciation, but rather one which 55 Rzach 46 Rzach Hesiod into West's text

8379. K

suppresses the ephelcystic -v in both places. reads οὐδέ with some slight ms. support, which would bring line with the Homeric οὐδὲ &owe(v) | 5X 1L, 5x Od.; but see and note ad loc.; cf. also οὐ μὲν ἔοικεν | B 233, οὐ yàp ἔοικεν

47 Lg x, 1934, 132.

138

THE LANGUAGE

OF HESIOD

was so dominated by the influence of spoken Ionic that ἀοιδοί no longer pronounced f even in contexts where it had once been present.5$ Digamma now survived only in its metrical effects, either in traditional phrases coined at a time when it was still a living consonant, or in new phrases containing a word which the poet used as though it still contained the sound, following the analogy of its use in traditional phrases. However familiar Hesiod was with the pronunciation of F as a consonant in his everyday speech, as far as the composition of his poetry was concerned the influence of his own vernacular in the matter appears to have been negligible. This conclusion may appear to fly in the face of the seemingly high proportion of ‘ observances ' to ‘ neglects ’ found both in Homer and in Hesiod, even in the Works and Days, but it must be remembered that the kind of evidence which hexameter poetry provides in support of a consonantal pronunciation of f is far from clear-cut : for example, the maintenance of & final long vowel or diphthong in hiatus, which counts as evidence for 'observed' f, is by no means an infrequent phenomenon even before an initial vowel where no F is restorable.*” In contrast, the evidence for ‘ neglected’ F is unequivocal, since before an initial consonant 1t 1s impossible for elision to occur or for ἃ preceding long vowel to be shortened 9? or for a final short vowel plus consonant to remain 48 In the pronominal

forms é, of it is possible that F retained some

consonantal force longer than elsewhere, and Hoekstra thinks that here it might still have been pronounced in Homcr’s time or until very shortly before it (HMFP, 43 f.); but see Kirk’s comments in Gnomon xxxviii, 1966, 738, offering an alternative explanation for the very frequent ‘ observation ’ of Fin these words. 4% Instances at A 24, 30, 39, 42, 74 etc. ; cf. Chantraine, GH I, 88-90, and for Hesiod

vowel

Devantier,

SADH

I, 9-12.

can occur before an initial vowel

Similarly

where

hiatus

after a final short

there was never a F:

GH I, 90-2, SADH I, 20-3; and for ‘lengthening’ plus consonant cf. GH I, 103-5, SADH I, 7-9.

cf.

of a final short vowel

5° I do not find it possible to accept A. García Calvo’s suggestion that a final open diphthong before an initial consonant could in certain circumstances be counted by Hesiod as a short syllable (Emerita xxxiv, 1966, 19-26). His theory conveniently obviates the crux at Th. 48, but the other examples which he proposes to introduce into the text look most improbable.

HESIOD’S

CHOICE

BETWEEN

TRADITIONAL

ALTERNATIVES

139

a light syllable. Thus Hartel’s frequently-quoted figures for Homer, l which give the impression that for every one example of ‘neglected’ F there are over five where it is ‘ observed ’ cannot be interpreted as an indication of the extent to which F was still pronounced as a consonant by the poets who created and first recited the Iliad and the Odyssey. The truth is rather that these poets were repeating forms of expression invented at a time when F was still pronounced as a consonant and affected the metrical pattern of the phrase. The frequency with which such forms of expression are repeated within Homer merely reflects, as Parry showed, the fact that he and other

ἀοιδοί composed in a style which was exceedingly conservative, in which ‘ to have given up the traditional phrase wherever the loss of the digamma now caused hiatus or failure to make position, would have been to destroy the diction almost entirely.’ 52 Hesiod composed in the same conservative style, but it is evident that he was thoroughly familiar with it in its Ionic dress. His work belongs to a stage of development where the neglect of F, which must be regarded as a linguistic innovation when compared with much of the phraseology at the poet’s disposal, was already firmly established in the hexameter tradition.°® All the features which we have been examining thus lead to the somewhat paradoxical conclusion that the language of the mainlander Hesiod shows a more decided preference for Ionic 51 Sitzungsberichte der k. Akad.

Schwyzer,

GG

I, 229;

Palmer

der

Wiss. zu

in CH,

101;

Wien lxxviii,

Troxler,

1874,

74;

SWH, 46;

cf.

West,

HT, 99. δὲ Tg x, 1934, 132. 53 To the evidence of words which once had initial F may be added that of

several forms in which contraction of adjacent vowels has followed what must be the complete loss of an intervocalic F which once stood between

them.

Some contractions of this type occur in Homer

(cf. Chantraine,

GH I, 28-38), but the following Hesiodic examples are notable as containing a contraction which is rarely, if ever, found there: ῥεῖ Th. 84, Σιμοῦντα Th.

342,

ἄθλος

Th.

800,

Op.

656,

κεραμεῖ

Op.

25, βασιλῆς

Op.

248,

203,

λουτρῷ Op. 753; and with synizesis βοῶν Th. 983, βασιλέων Op. 261. Παοσειδέων Th. 732 is probably not to be included here, in spite of Corinthian IIorE8aFów : see Chadwick, TPAS 1954, 1.

140

THE

LANGUAGE

OF HESIOD

forms of expression than does that of the Ionian Homer. While this same point has been made in different ways by a number of scholars,5? it is nevertheless one which tends to escape notice in the summary accounts of Hesiod’s language.** Yet a satisfactory explanation of it is necessary if we are to have a proper understanding of Hesiod’s place in the early hexameter tradition.°® It may also be relevant to the highly complex problem which is to be discussed in the next chapter. 54 Hoekstra,

Mnemosyne

x, 1957, 201 f.;

id. HMFP,

26;

Troxler,

SWH, 234 f. 55 Such as those mentioned above, p. 101 n. 1. 56 We shall return to this question in our conclusions below, pp. 201-3.

THE

ACCUSATIVE

CHAPTER

X

PLURAL

OF

A-

AND

O-STEMS

THE feature which we are about to examine is of special interest, since it occurs several times in Hesiod but is totally absent from the Iliad and Odyssey.! In Homer the acc. plural forms from a- and o-stems, in -as, -ovs, are always treated as

ending in ἃ heavy syllable, whether they stand before a consonant or before a vowel. In addition to these forms in -äs, τους, the Hesiodic poems contain also forms which are required to end in a light syllable, in -ds (eight certain examples and two possible ones) and -os (one example). This peculiarity has long been known, but it has recently been the subject of further study in articles by Mrs. A. Morpurgo Davies and Wilham F. Wyatt Jr., who have reached conclusions which differ both from each other's and from the older explanation of the forms as the result of ‘ Doric’ influence on Hesiod’s language. Before turning to the Hesiodic examples in detail it will be convenient to give a brief outline of the history of these acc. plurals, which are shown to have originated from Common Greek *-ans, *-ons both by the evidence of the Greek inscriptional dialects and by that of comparable forms in other IE languages. Forms in -avs and -ovs are attested in Cretan and Argive, where the group -vo-, -vs regularly remained unchanged;

but

in most

dialects

the

-v- was

lost, and

the

syllabie weight was maintained either by lengthening the preceding vowel (as in Attic-Ionic -às, -ovs), by diphthongization with an i-closure (as in Lesbian and Elean -a«s, -o«), or by combination with a following initial consonant (as in Cretan 1 Textual variants provide a few possible examples in Homer, but none of them is likely to be correct : see pp. 156 and 618 f. of the articles cited in next note.

2 Mrs. Davies, ‘ ‘‘ Doric ’’ features in the language of Hesiod ’, Glotta xlii, 1964, 138-65 ; Wyatt, ‘Short accusative plurals in Greck ' , TAPA xcvii, 1966, 617-43 ; these articles will be cited in this chapter as ‘ Davies ’ and

* Wyatt '. 3 See Schwyzer, GG I, 556, 559.

142

THE LANGUAGE

OF HESIOD

τὸς kadeoravs but rövs ἐλευθέρονς with -vs retained before an initial vowel). In certain dialects there appears to be loss of -v- without any such maintenance of syllabic weight, and forms

with short vowel

occur in Thessalian,

Arcadian,

and

a number of West Greek dialects ; * they are found also ın several poets who either were West Greek speakers themselves or came to have close association with places where West Greek dialects were spoken.* There are 112 examples of acc. plurals from a-stems in the Hesiodie poems and 131 from o-stems. Most of these stand either immediately before a word beginning with a consonant or at the end of a line, and in these cases it 15 not possible to determine whether the vowel which they contain was pronounced long or short. Only in the comparatively small number of examples which occur before à vowel is a syllable required which must be light or heavy in its own right. These examples will first be listed together with comparable material from Homer and other sources where relevant. TABLES 13 and 14, which are set out opposite, enumerate the occurrences of all a- and o-stem acc. plurals according to their phonetic context. (1) A-STEM ACC. PLURALS (a) -as V- scanned as a light syllable Th. 60 ἐννέα κούρας ὁμόφρονας, cf. τέσσαρες... ὁμόφρονες h.Herm. 195 (κούρᾶς at v 66, 77). Th. 267 ἠυκόμους θ᾽ ' Aprvías ᾿Αελλώ 7’. Th. Th. T Th.

φῶτες

401 μεταναιέτας εἶναι l. 534 ἐρίζετο βουλὰς ὑπερμενέι, of. ἐπεύθετο βουλάς [δ 677, 412, συμφράσσατο βουλάς | A 537, 540, ὃ 462. 653 διὰ βουλὰς ὑπὸ ζόφου, οἵ. διὰ βουλάς | O 71,0 82,

A 276, 437 (βουλάς at Y 154, V 78). 4 Buck, GD, 68; Thumb-Scherer, HGD II, 61 (Thessalian), 118, (Arcadian) ; Thumb-Kieckers, H@D I, 105, 175, 199 (West Greek).

126

5 “ The phenomenon is the common inheritance of poets within the West Greek family’ (Page, Aleman, the Partheneion, 132). Examples occur in Tyrtaeus, Aleman, Stesichorus, Simonides, Epicharmus, Empedocles and Theocritus : see Troxler’s list, SIVH, 74-6. Most of the literary examples are

discussed by Wyatt.

THE ACCUSATIVE TABLE

13.

PLURAL

OF A- AND

O-STEMS

143

A-stem acc. plurals in Hestod Th. | Op.|

Before init. vowel, ' light ’

6*

|

Sc. | Total

3|

0

9

| 7t|

2|

6

15

Before init. cons.

351

| 15,

9

59

At line-end

12

| 11]

6

29

60

| 31 | 21]

112

Before init. vowel,

‘heavy’

Total

* Including eipas Th. 804, but excluding räcas Th. 184 v.l. f Excluding orıßapas Th. 675 v.l. t Including πάσας Th. 184. TABLE

14.

O-stem acc. plurals in Hesiod Th. | Op. | Sc. | Total

Before init. vowel, ‘light’

0

0|

1

]

Before init. vowel, ‘ heavy ' | 12

6|

7

25

27.

8

74

Before init. cons.

39*

At line-end

M)

Total

65

5| 12 |

a

| 38 | 28 | 181

*Excluding éovs (several mss. at Th. 401; others have eoö, and Rzach reads €o after Brugmann), and νόμους (van Lennep’s plausible conjecture at

Th. 14 which is adopted by West).

Th. 804 εἴρας és ἀθανάτων. The mss. here have εἰρέας ἀθανάτων, but és or ἐπ᾽ seems to be required (cf. the construction at Th. 802), and Rzach adopts Hermann’s emendation, which seems preferable to Ruhnken’s etpats; eipas has therefore been counted as having a light final syllable, especially since the same would probably be true of εἰρέας if that reading were retained : see Davies, 152 n. 1 and West, HT, 377.8 * Another possible example from Th. has not been counted as belonging

here : at Τῇ. 184 most mss. give | πάσας δέξατο and this is followed by all the editors; but one ms. has πᾶσας ἐδέξατο and others preserve traces of this reading.

144

THE

LANGUAGE

OF HESIOD

Op. 564 {μετὰ τροπὰς ἠελίοιο l. cf. ὅθι τροπαὶ ἠελίοιο | o 404. Op. 663 Op. 675 Νότοιό τε δεινὰς ἀήτας |, cf. ἀνέμοιο δὲ δεινὸς anrns | O 626 (δεινᾶς at 2 479). At O 626 there is an alternative reading ἀήτη, and it is not possible to tell from the other occurrences of the word (& 254, 6 567, « 139) whether Homer is using it as a masc. or fem. ; sec Wyatt, 628 f. In Hesiod however cf. Op. 645 κακὰς ἀπέχωσιν ἀήτας |. There are no examples in the Shield.’ (b) -as V- scanned as a heavy syllable Th. 53 | τὰς ἐν Πιερίῃ, of. | τὰς ἐν Πηρείῃ B 766 (so Monro and Allen, but most mss. read {Περίῃ here). Th. 220

παραιβασίας

ἐφέπουσιν

|, cf. ὑπερβασίας

ἀποτίσῃ

|

cf. διὰ κρατερὰς

ὑσμίνας

|

x 168, and Op. 828 below. Th. 502

| Opaví8as, ovs.

Th. 631

Th. 112 jos κρατερὰς ὑσμίνας |. p 4o and κατὰ κ. 5. 6X Il. Th. 663 ἀνὰ κρατερὰς ὑσμίνας |,)(not found in Od.). Th. 615

| πέτρας

ἠλιβάτους,

cf.

| πέτραι 7 ἠλίβατοι

II 35,

v 196.9 Op. 645 ye κακὰς ἀπέχωσιν ἀήτας

|, cf. kaxäs before a vowel

in this position, M 113, B 316, ψ 332. Op. 828 ὑπερβασίας ἀλεείνων |, cf. ὑπερβασίας

ἀποτίσῃ

|

x 168, ἁματροχιὰς ἀλεείνων | Y 422, and Th. 220 above. Sc. 13

Sc. 107 Sc. 188 Sc.

188

Sc. 292

| ἐς Θήβας ἱκέτευσε, cf. | és Θήβας, ὅτε Καὶ 286.

| σὰς és χεῖρας, cf. φίλας ἀνὰ χεῖρας H 130. χρυσέας ἐλάτας, cf. xpvoeias + vowel X 598. cf. δρεπάνας ev xy. ey. | ἐλάτας ἐν χερσὶν ἔχοντες |,4 27 551, πελέκεας ἐν y. δρεπάνας

ex. | 114. Sc. 479 κλειτὰς ἑκατόμβας |, cf. κλειτὰς ἑκατόμβας | H 450, M 6, ἀγακλειτὰς er. | n 202.

ΤΑ further example occurs in the Hesiodie Catalogue (Fr. 55 Rz. = 150.15 M-W): ἰδὲ Σκύθας ἱππημολγούς |.

* Also at Th. 675 is another possible example, where στιβαράς, agreeing with πέτρας, is a variant in the phrase στιβαρῇς (-ais) ἐν χερσὶν ἔχοντες | ; for

parallel examples in this context see Sc. 188, 292 below.

THE

ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

145

(2) O-STEM ACC. PLURALS

(a) -os V- scanned as a light syllable There are no examples in Th. or Op. Sc. 302 ὠκύποδας Aayós ἥρευν |. Most mss. give Aayovs or λαγώς, but λαγός is protected by the metre:

see Rzach and Russo ad loc.

(b) -ovs V- scanned as a heavy syllable Th. 7 χοροὺς

Th. 8

ἐνεποιήσαντο

|, cf. χορούς

+ vowel A. xxvii, 18.

καλούς, ἱμερόεντας, cf. | καλοὺς ἀμβροσίους E 177. cf. θεούς, ot "Ολυμπον

Th. 101 Th. 391

θεούς, ot "Ολυμπον ἔχουσιν θεοὺς és μακρὸν "Ολυμπον

|;

ἔχ.

|

E

404,

θεούς

‚+ vowel in this posi-

tion 9x Th. 229 ψευδέας re Aóyovs AudıMMoyias τε |.

Il., 2x

Od.

But the reading V/eóed re Adyous 7’ A. is preferred by several editors including Jacoby and West, q.v. Th. 231 ἐπιχθονίους ἀνθρώπους |, cf. ἐπιχθονίων ἀνθρώπων | 4 T^.

45, 391

n 94. Th. 470

4 x Od. | ἀθανάτους ἐκάλεσσε,

| τοὺς αὐτῆς, cf.

cf. | ἀθανάτους

+ vowel ὃ 378,

τοὺς αὐτοί (-τή, -τός) Φ 31, κ 213,

A 574. Th. 501 πατροκασιγνήτους

0Aowv,

cf. marpokactyvgrov:

ὁ δ᾽

£ 330, with -ov scanned as a heavy syllable ! Th. 636 πλείους ἐνιαυτούς |; of. περιτελλομένους ev. | Θ 404, 418, περιπλομένους Ev. | V^ 833. Th. 935 | δεινούς, οἵ τ᾽ ἀνδρῶν. Th. 995 | τοὺς πολλοὺς ἐπέτελλε, cf. πολλούς + vowel 3x

Il, 2x Od., though not in this position; ἐπέτελλε E 320. Op. 39 δωροφαγους, ol τήνδε.

cf. | τάων

ds

Op. 74 χρυσείους ἔθεσαν.

Op. 100

κατ᾽

ἀλάλησθαι

ἀνθρώπους

ἀλάληται

|, οἵ.

| o 276, v 206, and ἀνθρώπους

Kar’

ἀνθρώπους

+ vowel in this

position £2 535, £ 29, ἡ 32, τ 395. Op. 278

| ἐσθέμεν

ἀλλήλους,

ἐπεὶ

οὐ, οἵ. ἀλλήλους

+ vowel

146

THE LANGUAGE

OF HESIOD

20x Il, 20x Od., including examples in this position at M 105, O 708, « 453. Op. 487 βροτοὺς én’ ἀπείρονα γαῖαν |, cf. βροτούς + vowel in this position 2 464, also βροτῶν ἐπ᾽ ἀπείρονα γαῖαν | H 446, p 386, τ 107. Op. 509 δρῦς ὑψικόμους ἐλάτας τε, cf. δρῦς ὑψικόμους ταναήκεϊ Y 118.

Sc. 94 χαλεποὺς ἐπετέλλετ᾽ ἀέθλους |, cf. χαλεποὺς ἐπετέλλετ᾽ ἀέθλους | A 622. Sc. 109 | δίφρους ἐμπελάσαντες, cf. δίφρους + vowel Y 352. Sc. 194 ὡς εἰ ζωοὺς ἐναρίζων |, cf. ζωούς + vowel ® 27, 102, ὑπερφιάλους évapitow | ® 224, ἀλλήλους ἐνάριζον [1 530, A 337, & 24, P 418, τοὺς ἐνάριζε E 155.

Sc. 249 λευκοὺς ἀραβεῦσαι ὀδόντας |, cf. λευκοὺς δ᾽ ἐπέρησεν ὀδόντας | E 291. Sc. 293) , , th cf. ἐν ταλάροισι φέρον, in this

Se. m: Sc. 337

ταλάρους ipe | ov?’

ἵππους

on. Σ 568.

ἑλέειν, cf. ἵππους

+

vowel

26x

Il.,

once Od. It is already to be observed that no significance can be attached to the higher incidence of ‘light’ examples from a-stems which is found in the Theogony. This was one of the features—probably the most important one—which led Rzach to make the statement : ‘ Wie wir sehen, ist es die Theogonie, die die meisten dieser Dorismen enthält ’ ;? but to base such a conclusion on this feature is mistaken, since it is clear

(see p. 143) that the total number of a-stem acc. plurals in the Theogony is almost exactly twice that found in the Works and Days, and the proportion of ‘light’, ‘ heavy’ and ‘ un-

certain ' examples is very much the same in both poems.1? ? DH,

465;

cf. abovo, p. 17 with n. 25.

10 This same point has been well made by Mrs. Davies (154 f.), but it bears repetition, especially since Troxler's discussion of the subject still asserts that the -as forms belong primarily to the Theogony (SWH, 73). The seemingly raro incidence of -ás forms in Op., whose language in other respects appears to be less traditional than 7'À., is even used by Troxler as a reason for discounting the possibility that this feature might have originated in Hesiod's spoken dialect (SW H, 77). But such an argument is erroneous.

THE

ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

147

The first ànd most obvious way of accounting for the presence in Hesiod of these acc. pl. forms with short vowel is that they reflect mainland dialect influence. This is the basis of the traditional explanation of the forms as a ‘ Doric’ or “West Greek’ feature, as maintained by many scholars

including Rzach and West.!!

There are two main stumbling-

blocks to this approach: (a) Hesiod was not a Dorian but a Boeotian of Aeolic parentage. Boeotian is not one of the dialects which show short acc. plurals in inscriptions, and none of those which do so appears likely to have influenced the dialect of Hesiod himself.!? (b) The second difficulty lies in the complete absence of light o-stem acc. plurals from Th. and Op., the solitary example in the Hesiodic corpus occurring in the later portion of the Shield, regarded as being pseudoHesiodic on other grounds.!? If we seek to explain the forms by reference to Hesiod’s own speech, we seem obliged to suppose that he regarded -as as short but kept -ous as long ; but in the inscriptional dialects the development of *-ans to -as or -as and that of *-ons to -ovs or -os are parallel to each other and share a common phonological explanation ; we have no reason to suppose that there was any dialect where *-ans would give -as but *-ons give -ous or -ws. One way of circumventing both these difficulties is to suppose that Hesiod regarded the a-stem acc. plurals as short without connecting this feature with any of the Doric and other dialects where syllabic weight was not maintained. Thus A. von Blumenthal !* suggested that Hesiod (and Tyrtaeus) might have felt the -as accusatives to be short on the analogy of the short-vowel acc. plurals in -ds which the Greek language already employed in i-, u- and consonant-stems.!® It is not 11 Rzach,

DH,

465;

West,

HT,

85.

The

label ‘ Doric’

is given to this

feature in the scholia : see esp. schol. on Th. 267. 12 See Davies, 155 f. 13 See above, p. 4. 14 Hermes Ixxvii, 1942

103 f.

15 We need not however accept von Blumenthal’s suggestion that Hesiod regarded the long forms as a metrical lengthening ; forms in -às and -ovs, ubiquitous in the /liad and Odyssey, were part of the stock of traditional material at the poet's disposal, and could be used by him without any sense of their being artificial.

THE

148

LANGUAGE

OF HESIOD

difficult to see how analogical forms might arise for a-stems, especially where words with -ds and -as endings found themselves in juxtaposition, e.g. πυκινὰς κλονέουσι φάλαγγας | Th. 935, καὶ ἐμὰς φρένας Op. 55, καὶ ποιμένας ἀγροιώτας | Sc. 39. In all of these examples the a-stem word may be pronounced with short vowel like that of its neighbour without upsetting the metre. This would perhaps be a sufficient explanation of the -ds form at Th. 60: ἐννέα κούρὰς ὁμόφρονας. Α similar process of analogy could have influenced other a-stem aec. plurals which do not stand 1n close proximity to an athematic form in -as but which occur in similar contexts : for example we may compare Ipatas τέκε καλλιπαρήους | Th. 270 with Χάριτας τέκε καλλιπαρήους | Th. 907, where some analogical levelling in pronunciation may easily have taken place.!$ Against von Blumenthal’s view both Mrs. Davies and Wyatt!" have raised the objection that, if analogy with athematic stems was involved, one would expect to find examples of lengthened accusatives in -as from the athematic declension. This objection has less force than might at first appear: the 'licence ' of a shortened form of a-stem acc. pl. is evidently of more use to a hexameter poet than would be that of a lengthened form of the athematic acc. plural. Metrical considerations alone make it easy to say when a vowel in a closed syllable must be short, whereas it 1s often very hard to be sure that such a vowel must be long, even when the syllable is heavy and an initial vowel follows.18 Mrs. Davies’ second objection concerns the presence in the Shield of λαγός, which, 16 Phrases of this type seem to mc to offer a morc plausible ‘ bridge ’ than either

ὑσμίνας Ἐὐσμῖνας

and

the other words

listed

by Troxler,

SW1I,

78,

or forms like εὐπατρίδας Πεὐπάτριδας (from εὐπατρίδης and εὔπατρις) mentioned by Wyatt, 620 n. 6. 17 Davies,

157;

Wyatt,

620 n. 6.

18 Thus Homer uses οἰκῆας ἄλοχόν re Z 366, νῆας ἄλαδ᾽ ἑλκέμεν B 165 = 181, νῆας αἱρησέμεν J 260, κατηρεφέας ἰδὲ σηκούς & 589, but no one would suggest that these aro long-vowel forms of the athematic acc. plural. They belong in fact to a fairly common type of ‘ lengthening in arsis ’ (see Monro, GHD,

348 8 375 (3);

Chantraine,

GH

formulaic modification (see Hoekstra,

I, 104), or are explicable in terms of

HMFP, 125 on tho νῆας examples).

THE ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

149

if von Blumenthal’s view 1s followed, has to be regarded as an artificial form based on the analogy of the a-stem forms with short vowel which are themselves analogous to the athematic acc. plurals. Such a process would admittedly be complicated, but would not be without parallel: the gen. pl. -orum of Latin o-stems 15 created by analogy with the -arum of a-stems, which 1s itself an analogical extension from the pronominal ending. The other criticism which Mrs. Davies puts forward is that ‘it seems useless to invent a new ad hoc explanation in order to account for a morpheme which has already a philological justification.’ !? The trouble here however is that this ‘ philological justification’ extends to o-stems as well as a-stems, and von Blumenthal's suggestion has the undoubted advantage over the ‘ Doric’ theory of explaining why Hesiod might have used acc. plurals from a-stems in -ἂς without using -os from o-stems. On balance therefore there scems much to be said in favour of von Blumenthal’s approach, though we shall see below that, taken alone, it does not account for all the facts. Recently an alternative hypothesis has been put forward by Wyatt, who in his lengthy and closely reasoned article ?? argues that the -ds acc. plurals originated in the masc. a-stems by a process of analogy with the short nom. sg. in -a of the type ἱππότα, whence they were extended first to masc. a-stems in -äs (— Ion. -ns), then to feminines in -ἃ and lastly to feminines in -à (— Ion. -n). He believes that this development took place in Boeotian or North West Greek, perhaps specifically Delphian, and that it spread, through the medium of Hesiod's poetry and the responses of the Delphic oracle, to all non-Ionie poetry. It lics beyond our present scope to assess all his arguments in detail, but the following points may be made : (1) Wyatt appears to exaggerate the importance of masc. a-stems in the development of the short acc. plurals: he expresses on p. 621 his opinion that this is the only place 19 Davies,

157 n. 3.

20 See above, p. 141 n. 2.

150

THE

LANGUAGE

OF HESIOD

where they can have originated, and presents as much evidence as possible in favour of the hypothesis in the following pages ; on p. 640 he writes: ' From the fact that the Doric accusative first appears in masculine a-stems we may conclude that it originated there. But our earliest evidence is in fact the Theogony and Works and Days, which give us the fem. à-stems κοὐρᾶς, BovAds (twice), possibly eipäs, tpomds (twice) and δεινᾶς, beside one fem. á-stem ‘Apmvids and one maso. a-stem peravaiérás. One must conclude that the -as forms were established in all a-stems well before Tyrtaeus with whom Wyatt begins.4 Even in Tyrtaeus there are only two maso. a-stems out of three examples, and the reader may be less impressed by the fact that δημότας 1s à masc. a-stem than by its occurrence in a line which includes at least two athematie accusatives in -às which would provide a likely occasion for levelling in pronunciation : πρεσβυγενεῖς τε γέροντας, ἔπειτα δὲ δημότας

ἄνδρας

(Fr. 3a.5 Diehl — 4.5 Hudson-Williams,

who reads c peofvyevéas).?? (2) Much of Wyatt’s argument rests on supposed grammatical speculation by Hesiod or his predecessors about what the acc. pl. of ἱππότα and other such forms might be. He has to suppose that ‘ Hesiod created a *ueravaıera ’ (p. 625) and ' Hesiod came to feel that there existed a form xoópa beside

κούρη, the accusative plural of which was xovpas ' (p. 630), etc. But this is sheer conjecture: the nom. sg. κούρη appears three times in Th. and twice in Sc., while κοῦρα (κῶρα) does not occur ?! Tyrtaeus is generally given a mid-seventh-century date (e.g. by Lesky, GGL, 139 f. = HGL, 118 £.).

2 Comparison

with

the

Homeric

phrases

κεκμηότας

ἄνδρας

2x

Il.,

ἡγήτορας ἄνδρας | 2X Il., Boropas a. |, Θρήϊκας d., θηρήτορας a., Παίονας a. each once in Il., scems to reinforce the likelihood of some analogy with the athematic acc. pl. Another factor, pointed out by Mrs. Davies (162), is the impossibility of fitting the δημότας and δεσπότας of Tyrtaeus and the μεταναιέτας of Hesiod into dactylic verse without shortening the final syllable. The initial consonant group of Σκύθας in the Catalogue might impose a similar difficulty (though cf. Homer's Σκάμανδρος, σκέπαρνον :

Chantraine, GH I, 110).

Πέρσας in Simonides (seo Wyatt, 625) seems to be

the only example of a masc. a-stem acc. pl. in -äs to which this does not apply, whereas most of the fem. words with -äs in Hesiod and later might easily be fitted into the verse in a form with -às.

THE

ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

151

before Theocritus, and then only as a vocative.?3 Hesiod's BovAäs raises special problems, calling forth a very complex

hypothesis (pp. 631-3), and the explanation of 8ewds also seems

highly

feminine,

contrived:

but knew

‘Hesiod

knew

also that Homeric

that

ἀήτη

was

concord with it was

queer (sc. in δεινὸς ἀήτη ν.] ἀήτης at Ο 626). I believe that he felt that somehow Homer had slipped up and had incorrectly made of ἀήτη a masculine noun’ etc. (p. 629). Did the mind of Hesiod or of anyone at Hesiod’s date really work along such lines as these ? ! (3) Wyatt's view that the presence of short acc. plurals in later Greek poets is to be understood as a literary feature, resulting from their imitation of Hesiod and indeed from their wish ‘ to align themselves in the Hesiodic tradition as against the Homeric and the Ionic ' (p. 642) must fail to carry conviction. He has to postulate (p. 637) an ad hoc explanation for λαγός in the Shield. His conclusions about the -ds and -os forms in Theocritus (pp. 637-40) are particularly unsatisfactory, and quite unnecessary in view of the short forms attested inscriptionally in Doric dialects including Coan. For these reasons Wyatt’s theory, though often ingeniously presented, does not appear to be justified by the evidence or to account for all the facts. We shall see later in this chapter that there are also other peculiarities in Hesiod’s use of these acc. plurals which cannot be explained by postulating the influence of a dialect which had light a-stem acc. plurals but kept a heavy syllable in its o-stems. A quite different approach to the problem is made by Mrs. Davies,?4 in whose opinion ‘ it is very difficult to accept any hypothesis based on the possibility that the short accusatives belong to the spoken language of Hesiod.’ ?* Instead, she explains their presence as arising through modifications in the use of hexameter formulas, and in particular through the 23 Greek a-stems inherited an IE voc. in -a, but this was for the most part ousted by the nom. in -à (Ion. -5) : see Schwyzer, GG I, 558.

24 See esp. her pp. 152-65. 26 Davies, 158.

THE

152

LANGUAGE

OF HESIOD

transference of a-stem acc. plurals from preconsonantal position (where she believes that the vowel would have been pronounced short, and the weight of the syllable maintained by the following initial consonant) to prevocalic position, especially before a feminine caesura.?9 This metrically useful device she thinks was then extended to words otherwise unusable in the acc. pl. like μεταναιέτας, and to analogical

adaptations of formulas, as in Hesiod's τροπάς ἠελίοιο beside the nom. τροπαὶ ἠελίοιο in Homer.?' The short-vowel acc. forms are therefore, in her view, essentially artificial, and she

cites as a parallel case the use by Homer of és before à vowel— a form which is ‘ metrically convenient but linguistically not justified —beside

the

expected

eis + vowel

and

és + con-

sonant. ‘The aoidoi of the Greek mainland went a step further and applied the same procedure to the short accusatives of the first and possibly of the second declension.’ 38 26 This suggestion would be more convincing if any of the Hesiodic examples illustrated it when compared with ἃ Homeric parallel, but unfortunately none of them does so; the nearest approach is βουλάς which occurs in Homer not preconsonantally but at the end of the line; see notes on Th. 534 and 653 above, p. 142. 517 This last example had already been explained in much the samc way by Hoekstra in Mnemosyne

x, 1957, 217.

28 Davies, 163f. But the parallel is inexact: as far as the acc. plurals are concerned, Greek inherited the endings *-ans, *-ons, heavy syllables which provide the basis for the forms found in all dialects. On the other hand the inherited form of the preposition was *en, and the form with the extension -s (probably by analogy with ἐξ : Schwyzer, GG II, 456 with n. 5), giving évs, eis or és, is confined to Pamphylian, Lesbian, Attic-Ionic and Doric ; forms of the preposition without -s continued to be used in historical times in

Arcado-Cyprian,

Thessalian,

Boeotian

and

North

West

Greek.

An

* Achaean ' or ‘ Aeolic ’ stage in the development of epic must therefore have included a form of this preposition which had the metrical value of a light

syllable when standing before an initial vowel. This being so, it would not be surprising if Ebcling's list, which Mrs. Davies cites, of almost 60 instances of és -+ vowel were to include some very old epic phrases. A formula such as πρόπαν ἦμαρ és ἠέλιον καταδύντα | 3x 1l., 6x Od. need not be a new creation by Ionic ἀοιδοί but could be merely an adaptation of an ancient formula with ἐν. That such a usage of ἐν + acc. existed at an early stage in thc epic tradition is made probable by the presence in Homer of the stereotyped ἐνῶπα,

ἐνδέξια

SW 1H, 115).

(Schwyzer,

GG

I,

619;

Leumann,

HW,

37

f.;

Troxler,

THE

ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

153

Objections to Mrs. Davies' solution have been raised by Wyatt, but they are not insuperable. To his first question (“How did the short forms of the accusative plural come to be used before a vowel ?’) one might answer that Hesiod is particularly fond of inverting the word-order within a traditional phrase and of shifting phrases from one part of the line Modification to formulas does not always to another.” preserve their entire metrical shape, as Wyatt himself admits (p. 618), though he stresses that 1t ordinarily does so. In any case, it is precisely this motive of preserving the metrical shape of Homer's τροπαὶ ἠελίοιο | which could, perhaps subconsciously, have prompted Hesiod to create τροπὰς ἠελίοιο |. We know that the transference of ἃ phrase from one grammatical case to another occasionally gave rise to forms which could never have had an existence in their own right in any dialect, e.g. εὐρέα πόντον after εὐρέι πόντῳ.39 It is true that the example of τροπᾶς 1s the only one for which we have a parallel in the nominative attested im Homer,? but it is conceivable that some of the others might have arisen similarly and that the ‘ models’ for them have not survived. One may

compare the prosodieal anomaly μέροπες ἄνθρωποι | 27 288, where the short -es has to be scanned as a heavy syllable, the explanation for which is to be found in the commoner μερόπων ἀνθρώπων |.32 To Wyatt's second question ( Why do such forms not occur in Homer ?’) we could answer that Homer was, generally speaking, more conservative than Hesiod in the maintenance of syllabic weight. For example, Hesiod uses both καλός and ἴσος which never appear in Homer with a short vowel in the first syllable.33 His γαιήοχον is a further example of a shortening never found in Homer, and the very fact that this anomaly 29 See above, p. 27 n. 4, with some examples from early in the T'heogony. 30 Seo Chantraine, GH

I, 97.

31 Two further examples might be more loosely related to other phrases in order to explain the metrical peculiarity : see notes on Th. 60 and Op. 675 above, pp. 142, 144.

32 7 x Il.,2x Od.; cf. Chantraine, GH I, 104 f. 33 See above, p. 107. L

154

has from

THE

been

most

LANGUAGE

plausibly

the modification

OF HESIOD

explained *4 as resulting

of a traditional formula,

and

directiy indeed

from 105 transference from one case to another, must be thought to add weight to Mrs. Davies' argument. Wyatt’s third question (‘Why do they occur only in a-stems ?") raises a more complex problem. Part of the answer may be contained in Mrs. Davies’ note (p. 157 n. 2), where she suggests that some short -os acc. plurals could have once been present in the text but have been eradicated during the transmission of the poems through being corrected to a singular in -ov—something which would have been impossible for a-stems with an acc. sg. in -yv. But it often happens that plurals simply cannot be substituted for singulars without doing violence to the sense, and one must agree with Wyatt that this does not seem a satisfactory explanation. I believe that à more plausible answer is to be found by making use of von Blumenthal's suggestion mentioned above : the existence of many acc. plurals in -às from athematic stems might favour the creation of a similar acc. pl. for a-stems, whereas there was no such factor to help the creation of -os in the o-stem accusatives. In any case the presence of Aayós in the Shield, though isolated, appears to weaken the force of this objection. We have seen in an earlier chapter 35 that various prosodical anomalies in Hesiod appear to have arisen directly from his adaptation of traditional phrascology, and this naturally predisposes us to favour an explanation of the light acc. plurals on the lines which Mrs. Davies has suggested. With the very minor modification just mentioned, her theory seems to account for the facts more successfully than the traditional explanation of the forms as resulting from dialect influence on Hesiod's language. But there is another approach to the problem which appears never to have been pursued, and which has the advantage of using a comparatively extensive body of evidence rather than the small number of examples listed at 34 By Hoekstra: see above, p. 101. 85 Chapter VII, esp. pp. 90-3.

THE

ACCUSATIVE

PLURAL

OF

A-

AND

O-STEMS

155

the beginning of this chapter. It has the effect of reinstating the traditional view. Up to this point we have considered only those instances where a- and o-stem acc. plurals stand before a vowel, and which can therefore be described as containing either a long or à short vowel, according to the requirements of the metre. There remains à much larger number of examples where the vowel is of uncertain quantity, since the word stands before an initial consonant or at the end of a line. Now if Hesiod regarded the acc. plurals of a- and o-stems as having a final syllable which was heavy in its own right, it is reasonable to expect that he would use them in the same way as any other word of the same phonetic structure, ie. with ἃ final long vowel plus consonant. If on the other hand he regarded them as naturally light in their final syllable, we might expect to find some reluctance to place them before a word with an initial vowel where the metre requires the final syllable to be a heavy one. In other words, does Hesiod place a disproportionately large number of a- and o-stem acc. plurals before a consonant ? An answer to this question may be reached by analysing the distribution of several groups of words with the ending -VC according to their phonetic contexts. It might be expected that such words would occur before an initial vowel or an initial consonant in 10ugh proportion to the number of words in the author's total vocabulary beginning with a vowel or a consonant, but it should be mentioned here that the distribu-

tion is probably not a completely random one. In his examination of Homer's traditional epithets, Milman Parry detected a preference for -VC C- where a heavy syllable is required

rather than

-VC

C-, though

epithets ending in -VC

are

frequently used before an initial vowel.?9 It is possible that in 36 Parry, ETH, 113 f., where he compares dva£ ἀνδρῶν ᾽Αγαμέμνων, dva£ ἀνδρῶν ᾿Αγχίσης, etc., with βοὴν ἀγαθὸς Μενέλαος, βοὴν ἀγαθὸς Διομήδης, and κρείων ᾿Αγαμένων, etc. with ξανθὸς Μενέλαος, etc. and κρατερὸς Διομήδης, etc. ; and he offers this explanation: ‘Pour garder le mouvement rapide du rhythme,

les aédes ont cherché à éviter, surtout à la fin du vers, comme

remarqué Platt, la syllabe longue à la fois par nature et par position.’

l'a

166

THE

LANGUAGE

OF HESIOD

Hesiod similarly there may be a tendency to avoid -VC C-, resulting in an apparent preference for placing -VC before a vowel; but there is no need for us to pursue this point for our present purpose, since, whether the distribution itself is a random one or not, it is reasonable to expect that different

types of word which all end in -VC will show much the same distribution as each other. Three sets of words have been examined with a view to establishing the average proportions 1n which the final syllable

-VC is place before a vowel, or before a consonant,?? or at the end of a line: (a) -ων gen. pl. of all declensions ; (b) -ys gen. sg. of a-stems ; (c) -ν acc. sg. of a-stems. The corresponding figures for a- and o-stem acc. plurals have already been given in Tables 13 and 14 above, but are repeated for convenience here as (d) and (e).?® These endings are found to occur as enumerated in TABLE 15,?? and when the figures there are converted into percentages, as in TABLE 16, the abnormal distribution of the acc. pl. forms in Th. and Op. becomes obvious at once. It will be seen that whereas over a quarter of the words which end unequivocally in -VC (a, b and c) stand before a vowel (Th. 26:395, Op. 29-995, Sc. 30-695, with 23% as the lowest limit of

variation in any poem for any of the three word-types examined), the proportion of -ἂς and -ovs acc. plurals placed before a vowel is much smaller than this in Th. and Op. (Th. 16%, Op. 12: 195, with the limits of variation 7-1% and 18:595). The discrepancy can be shown even more strikingly if we leave aside all the examples which occur at the end of the line, and compare only instances of final -VC standing before a vowel or a consonant with Hesiod's treatment of these ?' Including a supposed initial digamma ; this affects only a very few examples. 38 The ten ‘light’ examples arc necessarily omitted, since we are here concerned only with Hesiod’s use of -as and -ovs where the metro requires or allows ἃ heavy syllable. 39 Rzach’s text has bocn followed throughout. The adoption of alternative readings (as by West at T'h. 106 and 280) would change the figures slightly, but not enough to affect the conclusions drawn from them below.

THE

ACCUSATIVE

TABLE

15.

PLURAL

Words with final

|

A-

AND

O-STEMS

ΟῚ

lineend

| +V

|

157

-VC : Occurrences Sc.

Op.

Th. +V

OF

+C

line| end

| +V

| +C

line| end

(a) -wv

61

74

69

41

6]

41

27

47

33

(b) -ης

94

61

35

18

27

19

14

6

15

(c) -nv

49

|

126

38

33

40

28

18

20

13

a+tb-+c|

144 |

261

142

92

128

88

59

73

61

(d) -as

7

35

12

2

15

11

6

9

6

(e) -ovs

12

39

14

6

23

5

7

8

12

d-+e

19

74

26

8

42

16

13

17

18

TABLE

16.

Words with final -VC : Percentages

(a) -ov | 29-9

|

+C

|

end

| 36-3 | 33:8] | 46:9

| 26:9

line-

line-

lineἜΝ

Sc.

Op.

Th.

-+C |

ἜΝ

|

28-7

| 42-7

end

+V

|

-FC | end

| 28-7 | 25-2 | 43-9

| 28-1 | 42-2 | 29-7

(30:8

| 40-0 | 17-1 | 42:9

(b)-ns

| 26:2

(e) -o»

| 23:0 | 59-2 | 17-8 | 32-7 | 39-6

| 27-7 | 35-3 | 39-2 | 25-5

| 41-6

| 28-6 | 30-6 | 37-8 | 31-6

a-b-rc | 26-3 | 47-7 | 26-0 | 29:9 (4) -ας | 13-0 | 64-8 | 22-2

(e) τους] d+e

18.5 | 60:0 | 21-5 | 16-0

|

7-1 | 58-6 | 39-3 | 28-6 | 42:9 | 28-6

| 15:8 | 71-1 | 13:2 | 25-9 | 29-6 | 44-4

! 62-2 | 21-9 | 12-1

63-6 | 24-2 | 27:1

35:4

137-5

acc. plurals. TABLE 17 shows the proportions which may be established from the data assembled in Table 15. (The figures there mean that for every once that Hesiod used a gen. pl. των before a vowel in Th., he used the same ending before a consonant 1-21 times, etc.) The proportions are fairly consistent for all the word-types examined, except for the marked difference in the figures for -ἂς and -ovs acc. plurals. [Ὁ is

158

THE LANGUAGE TABLE

17.

OF HESIOD

Proportion of occurrences, -VC Th.

Op.

V- : -VC CSc.

(a) -wr

1:1:21

| 1:1-49

| 1:1-74

(b) -ns

1:1:79

| 1:1-50

| 1:0-43

(c) -nv

1:2:57

| 1: 11:21

! I1:1-11

a+b+e|1:1-81

| 1:1-39 | 1:1-24

(d) -as

1:5-00 | 1:7-50 | 1:1-50

(e) -ovs

1:3-25

| 1:4-50 | 1:1-14

d 4- e

1:3-89

1:5-25 | 1:1-31

therefore possible to give this clear answer to the question posed above (on p. 155) : in the T'heogony and the Works and Days there is evidence of ἃ reluctance to use a- and o-stem acc. plural endings as heavy syllables before an initial vowel, and Hesiod places a disproportionately large number of them before à consonant.

This conclusion is confirmed by an examination of Homer's practice 1n the same matter. The occurrences of -as and -ovs endings in four books were counted, and the figures given in TABLE 18 were obtained. In general they support the view that Homer, unlike Hesiod, used the acc. plurals of a- and o-stems

in the same way as other words ending in -VC, since both the percentages and the proportional figures, with one exception, are closely similar to those for -wv, -ns and -nv in Hesiod (cf. Tables 16 and 17 above). The result for Iliad A demonstrates a limitation of the method. For whereas a count of all words with final -VC (excluding variable words such as οὐκ, ἐγών) in this book gives the following result : occurrences,

186 + V

percentages, 32% +

217 +C

V

3795 --C

184 at line-cnd,

31%

at line-end,

proportional figure of -VC V- : -VC C-, 1:1-17, the figures for a- and o-stem acc. plurals are seen to diverge widely from this mean. How can we be certain that the

159 O-STEMS AND AOF PLURAL ACCUSATIVE THE

Δ. 11:1 99. 0:1

LT-T:I

00-1:1

09. Ῥ:1

5πο-

(LT TIavL

19) sensy Ir

G6

GE

GS

c6

GV

8

ΟἹ

GV

66

ς

et

96

65

δ

6

ST

cr

L

6

cE

ey

Ol

3

0%

Ol

SD-

sao-

sp-

Sno-

g$90uolInoo()

(gi wav],

sodeyuooled

= (QT ATAVY,

sp- [euorjaodoaq FG

£9

9

9

At

0L-0:1

LG

el

L

G

O+

-

81 CG

F

e

pus -euiT

09-1:1

ος CL



I

At

-

95 0

c

6

+

0° 7:1

εξ cE

9 9

pue -9Ul]

-

IF LY

95 0

ΔῈ

00-€:1

sI

$8 6I

oT

ΔῈ. 111

I» | 96

pus -Oul|

9A0q? :J9)

9A0q* 0)

OT

At

F

'

Qr

N

ATAVL

pue -oui[

D

"SU

At

a

pup -

pus | Dd -QUIT Te30,

49wt0H ur sqpanjd. "900 08-0

160

THE LANGUAGE

OF HESIOD

discrepant figures in Hesiod (Tables 16 and 17 (d) and (e) above) are not similarly the rcsult of chance distribution as they must be in Iliad A ? 4° One can best answer such a question by pointing out a very important difference between the Hesiodie and Homeric evidence in this matter: in Iliad A the low proportion of -ovs accusatives before an initial vowel stands in 1solation, both

from the data provided by the other Homeric books examined and from the comparable figure for a-stem acc. plurals in the same book as well as the other books. There 1s no reason to suppose that the figure is of any significance for the question of dialect, or that it is anything other than the result of ἃ purely random distribution.*! With Hesiod on the other hand we are concerned not with one piece of evidence in isolation but with three, all of which point in the same direction, and in all of

which the Theogony and the Works and Days show a remarkable accord : (a) the presence of nine a-stem acc. plurals which must be scanned as light syllables in -às (plus the single o-stem example in -os from the Shield) ; (b) the very small proportion (Th. 13-095, Op. 1-194) of a-stem acc. plurals occurring as heavy syllables before a vowel, as compared both with Hesiod's use of other

-VC endings (Tables 16 and 17) and with Homer's use of -as acc. plurals (Table 18) ; 1° [ am indebted to Mr. D. A. Hester who, in correspondence subsequent to the meeting of the Mycenaean Seminar in London on May 13th 1964, drew attention to Bernoullian methods of determining whether figures obtained from what are in effect small samples are likely to be statistically significant. This approach shows that, even after taking into account the distortion which follows upon the necessary exclusion of the ‘ light ’ forms, the figures in Table 16 do reflect a significant discrepancy in Hesiod's use of these acc. plurals. But it has not been thought necessary to include details of the working here, and most readers will probably be more readily persuaded that a real discrepancy exists by the reasons which are given in the remaining pages of this chapter. *! The number of examples is very small, with the result that one occurrence represents 5% in the percentage figures for both a- and o-stems from this book.

THE ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

161

(c) the small proportion (Th. 18:595, Op. 15-8%) of o-stem acc. plurals occurring as heavy syllables before a vowcl, similarly compared. Previous discussions of the problem have concentrated wholly on the first of these pieces of evidence, which is more immediately obvious than the other two. But in choosing between the proposed explanations, we must seek a solution which accounts not only for point (a) but also for (6) and (c). The adaptation of traditional exprossions, suggested by Mrs. Davies as a likely way of explaining the occasional shortvowcl forms, does not provide a satisfactory reason for the abnormal distribution of tho much larger number of remaining examples: if Hesiod used -as and -ovs regularly as heavy syllables, shortening them only as an artificial device when adapting a traditional phrase, why should he place such a high proportion of these endings before a consonant * But if we were to suppose that Hesiod regarded the vowel of the acc. pl. ending -as as short for the reason put forward by von Blumenthal, or for the different reasons expounded in Wyatt’s theory, this would still leave unexplained his abnormal distribution of the acc. pl. of o-stems. The influence of a dialect where the vowel in both thesc endings was short sccms to be the only adequate explanation of this phenomenon. We thus come back to tho old view of the short-vowel accusatives as being of dialect origin. It was observed earlier *? that two obstacles stand in the way of accepting this explanation. The second of these, arising from the absence of short o-stem acc. pl. forms from Th. and Op., appears to be illusory in the light of this further evidence. The first difficulty however still remains: short-vowel forms are attested only in Thessalian,

Arcadian,

and

West

dialects,

Greck

work of lyric poets with West Greek connexions. have

to suppose

either

(a) that

the

dialect

and in the

We therefore of Boeotia

in

Hesiod's time included the short-vowel forms and that these were later replaced by the long-vowel forms attested in 42 See above, p. 147.

162

THE

LANGUAGE

OF HESIOD

inscriptions ; or (b) that the short-vowel forms were familiar to Hesiod because they were already a feature of poetry, now lost, which existed on the Greek mainland in his time. The first of these alternatives 1s not impossible, though 1t raises à conjecture which we have no other evidence to support. The second may appear more plausible, and has the advantage that it relates Hesiod's use of these forms to their later presence in West Greek lyric poetry. If this were accepted as the explanation, it would certainly be enough to account for the occasional use of a short-vowel acc. pl. form, but it may be doubted whether, taken alone, it fully explains the whole set of abnormalities ** which we have found in the Theogony and the Works and Days. Other factors therefore may be involved here besides the existence of forms in -as and -os in West Greek and other dialects. Perhaps a tentative explanation might be suggested along the following lines: we are dealing with a highly complex situation in which a speaker of a mainland Greek dialect is consciously using an artificial but predominantly Ionic dialect for the composition of hexameter poetry.5? It is possible that what would most strike the speaker of a dialect such as Boeotian, hearing these acc. pl. endings pronounced by an Ionian, would be not the length of the vowel but its quality. He would be aware that as a general rule a in his own dialect corresponded to ἡ in Ionic, while & was the same in both. Thus when he himself came to recite, using the Ionic forms, he would substitute ἡ for the à of his own dialect in many places, including the singular forms of a-stems, e.g. in such a phrase as ἄλλης μὲν λώβης Ν 622; but he had to remember not to do so in the acc. plural, e.g. in πολλὰς δὲ καὶ ἄλλας 88. The position with regard to o-stems is rather different from this, but is analogous in one important respect : here also Attic and Ionic maintained a threefold opposition where only a twofold one existed in most dialects, as is made 43 See above, p. 142 with n. 5. 41 i.e. the threo points listed above, pp. 160 f. 15 See Chapter IX, esp. pp. 130 f., 139 f.

THE

ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

163

clear in TABLE 19.15 A non-Ionic speaker wishing to maintain an opposition observed by Ionic between long vowels not kept distinct in his own dialect (ie. 2 and 3 in the Table) might treat this as an opposition of quantity rather than of quality, even though to do so involved a loss of opposition (in his wouldbe-Ionie pronunciation) between two vowels which he regularly kept distinct (ie. 1 and 2 in the Table).47 This would be especially easy if short-vowel acc. pl. forms already existed in West Greek poetry which was known to him. TABLE

19.

Opposition of primary and secondary long vowels Common Greek

(1) short a

Ionie

Boeotian etc.

ntr. pl.

&

à

*.d

ἄλλα

ἄλλα

(2) secondary long a (= lengthened 3)

fem. acc. pl. *-ans

ἃ ἀλλᾶς

à

(3) original long a

fem. gen. sg. *-ds

7) ἄλλης

(1) short o

masc. nom. sg. *-Ós

o ἄλλος

(2) secondary long o (=lengthened 6)

masc. acc. pl. *-óns

|

ov ἄλλους

adverb

|

e

(3) original long o

*.68

ἄλλᾶς o ἄλλος

ω

ἄλλως

ἄλλως

If then the poet of Th. and Op. had come to pronounce, in recitation, both a- and o-stem acc. pl. endings as light, how do we account for the fairly large number of places where -as or τοὺς occurs before a vowel as a heavy syllable ? The answer must be that we should expeot to find this occurring mostly where he is using a traditional Ionic phrase, and this is certainly borne out by a high proportion of the examples.*? 46 See Lejeune,

7'PG, 202 f.;

Buck, GD, 28 f.

47 The process would be further complicated, and perhaps also assisted, by any qualitative differences between long and short vowels, though on these see Allen,

Vox Graeca, 59 f. with n. 1, 68, 73.

48 See the notes on parallels in the above lists, pp. 144-6.

THE

164

LANGUAGE

OF HESIOD

It must also be remembered that it is not uncommon for closed syllables containing a short vowel to be treated as heavy even before an initial vowel : if Homer could say ‘ πατροκασίγνητον" ὁ δ᾽ ἐπιζαφελῶς ᾿ (ζ 330), why should Hesiod not say “Ἐπατροκασιγνήτος ὀλοῶν (Th. 501)? 4 Another possibility is that some at least of the apparently heavy acc. pl. endings could have resulted from interference with the text at a postHesiodic stage in the transmission of the poems. Thus several of the -as or -ovs cxamples stand before a relative pronoun ot, οὕς, which might have replaced an older τοί, τούς.

The

phrase θεούς, of "Ολυμπον ἔχουσιν occurs in Homer as well as at Th. 101, but we may compare ἀθανάτους μάκαρας, Tot "Ολυμπον

ἔχουσιν at Sc. 79, where τοί is necessary to make position (cf. relative τοῖσιν, τά making position similarly at Op. 172, 294) : if, as we believe, Hesiod was reluctant to use these acc. endings as heavy syllables in their own right, he might easily have recited his poems

saying τούς, not οὕς, at Th. 502, and τοί,

not of, at Th. 935 and Op. 39.°° What conclusion can be drawn about the Shield? If it were not for the solitary Aayós at Sc. 302, it seems that the poem could be completely left out of this discussion: it shows no examples of a-stem -ds and no abnormality in its distribution of either the a-stem or the o-stem acc. pl. endings.°! If the author of Sc. himself regarded these endings as light syllables, we have to assume that he was so strongly under the influcnce of traditional material that he always 4° Seo also the p. 148 n. 18. 50 It is possible be reflected in the HT,85; Troxlor,

athematic

aec. pl. examples

from

Homer

cited above,

that this kind of interference during transmission may also problematic variant δήσας for δῆσε δ᾽ at Th. 521 (see West, SWH, 77). ‘The -as of the aor. participle is derived from

*.avs « *-ants, and its history in the dialects is similar to the acc. pl. in -as < *-ans (Thumb-Scherer, HGD II, 61, 126). But the example is isolated

and the reading very doubtful. Tho distribution of these participles in -as throws little light on the problem: it seems to be slightly abnormal in Th. (4X +V,12x +C), but this is hardly sufficient evidence to suggest that Hesiod might not have regarded the -as as a heavy syllable ; in Op. and Sc. these participles oceur with almost equal frequency before a vowel and a consonant (7 x +V,8x +Cin Op.; 3x --V, 4x +C in Sc.). 51 See Tables 16 and 17 above.

THE

ACCUSATIVE

PLURAL

OF A- AND

O-STEMS

165

used the Ionic forms in -as and, except once, -ovs before a

vowel, and did not treat either of these as in any way different from other endings in -VC. While such an interpretation is not impossible, especially in view of the high proportion of traditional phraseology in this poem, perhaps a more plausible explanation is that the author of the Shield himself normally used these endings as heavy syllables, as Homer and many others did, and the isolated Aayos requires some special explanation : either it is a genuine dialect form used earlier by another poet and borrowed by the author of Sc., or it is an analogical form which arose in the manner described by Wyatt,*?? or something has gone wrong with the text at this point, e.g. a dual has been replaced by a plural.5? The most likely hypothesis therefore is that Hesiod himself, but probably not the author of the Shield, felt these syllables to be light and regularly pronounced them as -ἂς and -os in his own poetic recitation. This does not necessarily mean that he normally spoke a dialect in which short-vowel forms were current,5* and in any case he clearly knew and used hexameter material which included long-vowel forms. Such a view would account both for the presence of the short-vowel forms themselves and for the apparent reluctance of the poet to use these acc. plurals before vowels where a heavy syllable was needed, although he did not refrain from doing so in traditional phrases or in phrases analogous to traditional ones. The complete absence of short o-stem forms from Th. and Op. must then be explained as due to the strength of the tradition in which -ovs could never be short, while the introduction of the -as forms

was facilitated by the presence in earlier poetry of acc. plurals in -ds from i-, u- and consonant-stems. 5? Wyatt, 637, though I do not find his explanation plausible. 83 This may perhaps seem an unlikely conjecture, but (a) duals are not uncommon in the description of Hcrakles’ shield, and ὠκύποδε λαγὼ ἤρευν would scan whereas our ms. readings do not; (b) it would balance kapxapóBovre κύνε in the following line: cf. the antithesis of the dogs and the hares in Sc. 304 ; (c) a dual appears to have been replaced by a plural in almost all the mss. at Sc. 172 f. (see Russo ad loc.). 54 This nevertheless would perhaps be the simplest explanation: see below p. 194 with n. 21.

CHAPTER PHRASES

COMMON

TO

XI

HESIOD

AND

THE

ODYSSEY

WHILE it is generally agreed that Hesiod presupposes the existence of heroic poetry of the Homeric type, some scholars ! have claimed to show that certain parts of the [lad and the Odyssey are dependent upon Hesiod. Many, perhaps all, of these allegedly post-Hesiodic passages in Homer are open to alternative explanation,? but we saw earlier ? that a theory which has won a considerable following is that Hesiod's chronological place is later than the lad but earlier than the Odyssey. The credit for this must be given mainly to Inez Sellschopp, a major part of whose dissertation,* published in 1934, 1s devoted to a persuasive study of lines and phrases common to the Odyssey and the Hesiodic poems.* If her arguments are accepted, virtually the whole of the Odyssey— all, that is, except for Books ı, x, part of A, and p—were composed later than Hesiod. If on the other hand they can be shown to be unsound, we are still free to draw our own con-

clusions about Hesiod's chronological position from the linguistie evidence which has been discussed in the foregoing chapters. Miss Sellschopp's method is to examine various lines and phrases which Hesiod and the Odyssey have in common, and by a comparison of their contexts and their use of language to ! E. Bethe, Homer. Dichtung und Sage II, 1929, esp. 303 ff.; TdH,

esp. 66-81;

H. Munding,

Hesiods Erga in ihrem

Schwenn,

Verhältnis zur Ilias,

1959, 173 and passim. ? See Lesky, GGL, 110 = HGL, 91; the arguments of Munding especially are highly questionable: see J. Kerschensteiner, Gnomon xxxiv, 1962, 1-7. 3 See above, p. 9. * SUH, 42-81. 5 i.e. Th. and Op. * In his brief discussion of this subject (see above, p. 9) Solmsen writes of Sellschopp : ‘I have found no flaw in her arguments’ (H&A, 6 n. 3). He himself however is careful to avoid adopting too rigid a view, and he has more recently argued against Hesiodic priority in the case of Th. 92 beside 0 172 (see below, pp. 168 f.).

PHRASES COMMON TO HESIOD AND THE ODYSSEY

167

determine where possible which of the two is original.” She recognises the need to make allowance for phrases of a formulaic type which are unlikely to have been composed for either of the passages in question: “Solche Verse brauchen für keine der hier zu vergleichenden Stellen erst geprügt zu sein, sondern kónnen altes, ganz allgemein benutztes episches Gut sein.’ 5 She lists eleven Hesiodic lines belonging to this category, and then turns to the places where she believes that the order of composition can be established by comparison of the passages in Hesiod and the Odyssey where they now stand. Generally speaking, ἃ short phrase may fit ἃ wider variety of contexts than a whole Ime or longer passage, so that the longer a parallel 1s, the greater is the likelihood of our being able to establish which of the passages containing it is the earlier. It follows that the pillars of Miss Sellschopp's argument are the parallel expressions extending over a whole line ; ° The shorter parallels 10 cannot be made to bear so much weight, since they may be transferred more easily from one context to another, so that the priority of a particular passage 1s much harder to establish. Out of these longer parallels, Miss Sellschopp cites five as demonstrating the priority of Hesiod to the relevant place in the Odyssey. Are any of these convincing ἢ (1)

Th. 27 ἴδμεν ψεύδεα πολλὰ λέγειν ἐτύμοισιν ὁμοῖα T 203 tone ψεύδεα πολλὰ λέγων ἐτύμοισιν ὁμοῖα

While Miss Sellschopp 11 is right to say that T. 27 is firmly tied to its context by anaphora and antithesis, she assumes too readily that the presence of these figures proves that the expression must have been originally composed for this place. At τ 203 also the line is closely integrated with its context, and the words ἐτύμοισιν ὁμοῖα are more important 7 The method is essentially the same as the one which Kretschmer applied to repetitions within the Hesiodie poems in order to establish their order of composition : see esp. D/H, 33. ® These are discussed in her Ch. III (SUH, 42-65) 8 SUH, 42. 10 i.e, those considered in her Ch. IV (SUH, 65-81). 11 SUH, 43.

1θ8

THE

LANGUAGE

OF

HESIOD

there than in Hesiod for what follows: Odysseus’ story is false, but it has been told so convincingly that Penelope is now overwhelmed with grief (τ 204 ff). The crux of Miss Sellschopp’s argument is that r 203 is shown by its * abundantes ' participle λέγων with take to be dependent on an older line;

but ἴσκε does not mean

‘say’ here, so that

λέγων is not pleonastic but essential to the meaning and structure !? of the line. Once this is accepted, there is no reason for supposing that Th. 27 has provided the model for T 203, and indeed the reverse may seem more likely.

(2) 1}. 92 0 172

} αἰδοῖ μειλιχίῃ, μετὰ δὲ πρέπει ἀγρομένοισιν

The relationship between these two lines cannot be considered in isolation from that between the larger passages in which they are contained.!* Miss Sellschopp 15 adopts the view of Wilamowitz 1° that the Odyssey passage is dependent on Hesiod, adding nothing herself to Wilamowitz's arguments. But these are not convincing, as has been shown by Solmsen, who in his detailed re-examination of the problem 17 himself advances a strong counter-argument from the feebleness of the second half of Th. 92 after θεὸν ὡς ἱλάσκονται | Th. 91, whereas at 0 172 the words lead on to the climax θεὸν ὡς

1? See Chantraine, RPhx,

1936, 351 ; id., GH I, 317 n. 1 ; Stanford ad loc.

The suggestion by M. van der Valk that the existence of ioxw ‘ to speak ' is proved by the related (ev)iorw (Textual Criticism of the Odyssey, 1949, 116 n. 7) is mistaken: cf. Frisk, GEW s.vv. év(v)éro and too. 18 West (HT, 163) believes that the Odyssey line is ‘ the less satisfactory of the two as Greek ’, and says that if toxe is meant in the proper sense ‘ as-

similate ', then ὁμοῖα is superfluous ; but while ioxe ψεύδεα πολλὰ ἐτύμοισιν would be an acceptable construction, toxe ψεύδεα πολλὰ λέγων ἐτύμοισιν would be very odd indeed. The words ἴσκε and λέγων belong together in sense, and ψεύδεα πολλά depends as much on λέγων as on ἴσκε : cf. διαπρήξαιμι λέγων ἐμὰ κήδεα £ 197. 14 This has been much

discussed : see West,

HT,

183.

15 SUH, 49 f. 16 Die Ilias und Homer, 477-9.

17 The *' gift’’ of speech in Homer 1-15, esp. 11-13 on Wilamowitz.

and Hesiod’,

TAPA

Ixxxv.

1954,

PHRASES

COMMON

εἰσορόωσιν | in 6 173:

TO

HESIOD

AND

THE

ODYSSEY

169

‘It looks as though Hesiod, having

broken away from the Odyssey passage (at Th. 87) to say that the eloquent king can even settle large disputes and having then gone on to explain why the kings are called éyédpoves, found himself forced to invert the order of the two remaining lines of the Odyssey passage since this was the only way in which they now could be worked in—and they probably appeared to him altogether too good and pertinent to be omitted.’ 1% The problems posed by these two passages are complex, but it is clear from the discussions of them by Solmsen and more recently by Krafft !? that there are strong reasons for doubting Hesiodic priority. (3) Th. 228 *Yoyivag te M&xag τε Φόνους τ᾽ ᾿Ανδροκτασίας τε A 612 Sopivat τε μάχαι τε φόνοι τ᾽ ἀνδροκτασίαι τε

In A the context is a description οὗ Herakles’ sword-belt on which these ὑσμῖναι etc. are in some way represented, whereas in Th. Hesiod includes the line in a list of baneful personifications when enumerating the children of Eris. Miss Sellschopp 39 assumes that the poet of A is using the words in their personified sense, and therefore is unlikely to have taken the second half of the line from the Iliad, where μάχας τ᾽ ἀνδροκτασίας re | occurs, without personification, at H 237 (acc.) and 2 548 (nom.). But it is unnecessary to suppose that the words are personifications in the Odyssey passage: 2: what is being described at A 612 is the scene portrayed on the sword-belt rather than a group of creatures. There is no notable difference between the Iliad and the Odyssey in the sense of the phrase, and it is the Theogony which provides the first example of its use with a personified meaning. In any case the authenticity of the Nekyia, in whole or part, has often been questioned, and λ 612 belongs to a passage whose lateness has been suspected 18 op. cit. 13. 18 VUHH, 69 f., where he comes down firmly on the side of the priority

of 0 170-3 over Th. 84-92. 20 SUH, 52 f. 21 See Krafit, M

VU HH,

67 f.

THE

110

LANGUAGE

OF HESIOD

even by critics who accept the bulk of the book as genuine.” Thus even if the priority of Th. 228 to A 612 could be proved, it would not mean that any major part of the Odyssey was dependent upon Hesiod. (4) Th. 373 ἀθανάτοις τε θεοῖσι, voi οὐρανὸν εὐρὺν ἔχουσι

ὃ 479 ἀθανάτοισι θεοῖσι, τοὶ οὐρανὸν εὐρὺν ἔχουσι

Miss Sellschopp rightly points out that both halves of the line occur separately in the Iliad, and infers from this that ‘Hesiod hóchstwahrscheinlich sich seinen Vers aus zwei Formeln, die er in der Ilias fand, selbst zusammenstellte und

ihn nicht als Ganzes aus 6 übernahm.’ 38. One might reasonably expect that, if Hesiod could have composed this line without knowing it from the Odyssey, the poet of the Odyssey might have done the same thing without knowing it from Hesiod ; but instead we find that Miss Sellschopp later includes T^. 373 in her list 24 of lines where she believes that Hesiod probably provided the model for the corresponding Homeric place. But to show that Hesiod could well have composed this line out of Ilhadie formulas is not the same thing as showing that his line provided the model for 8 479. The poet of the Odyssey was obviously familiar with the formula ot οὐρανὸν εὐρὺν ἔχουσι | which occurs 16 times in the poem, including twice in lines identical with 6 479, namely A 133 and y 280, and several times in lines 25 which are closely similar.?$ (5) Op. 317 αἰδὼς δ᾽ οὐκ ἀγαθὴ κεχρημένον ἄνδρα κομίζει ρ 941 αἰδὼς δ᾽ οὐκ ἀγαθὴ κεχρημένῳ

ἀνδρὶ παρεῖναι

Wilamowitz once wrote that ‘ o 347 aus Hesiodos Erga 317 ?2 e.g. Stanford in his edition of the Odyssey, Vol. I, 381, where he writes :

‘My own opinion is that all, except perhaps 565-627, is probably genuine.’ See further R. Merkelbach,

Untersuchungen

zur Odyssee,

1951,

177, and the

references given by van der Valk, T'CO, 229 n. 3. 23 SUH, 54: for the Homeric occurrences see Krafft, Nr. 72, and 173 Nr. 276a. 24 SUA, 64, and again at the start of her Ch. IV, SUH, 65.

VUHH,

165

25 cf. esp. a 67, ζ 243, u 344, « 200, 211. 26 See further Krafft,

VU HH,

46 n. 4, where he draws attention to other

formulaic phrases in which Hesiod appears to replace a dat. in -oı with the shorter form of the ending plus a particle.

PHRASES

COMMON

TO HESIOD

AND THE

ODYSSEY

171

entlehnt : das kann niemand leugnen ; Hesiod hat auf seinen Vers 500 zurückgewiesen ' ; 2” but a year later in his edition of Op. he abandoned this conviction, describing the relationship of the two passages as ‘nicht sicher bestimmt ’.28 Miss Sellschopp’s treatment of the lines and her discussion of αἰδώς and ἀναιδείη * contain nothing to show that Op. has provided the model for p here, and Krafft has drawn attention to some strong reasons for rejecting her view.®° The relationship between the two passages has also been discussed at length by K. J. McKay, who concludes that Hesiod here * changes the function of οὐκ ἀγαθή in an old proverb (and also the verb, with a display of bitter irony). ?! We may infer either that Hesiod knew the proverb from this Odyssey passage or that both he and Homer drew upon ἃ common stock of proverbial material. These then are the five Hesiodic lines which ‘ sich sämtlich deutlich als vor den entsprechenden Versen der Odyssee (τ 203, 0 172, A 612, ὃ 479, p 347) gedichtet herausgestellt haben,’ ® and Miss Sellschopp now turns to the shorter parallels. Some of her arguments here are at first sight more convincing, but she often fails to consider the whole of the evidence, and the basic difficulty with even her best examples is that they concern phrases which are short enough to be usable in a very wide range of contexts, so that their evidence is particularly open to the objection that both of the passages 27 Die Heimkehr des Odysseus,

1927, 154 n.

28 Erga, 79.

29 SUH, 59, 96-9. 30

VUH H,

74-6:

both

the

construction,

with

οὐκ ἀγαθή

predicative,

followed by a dependent infinitive (cf. A 589), and the more obvious sense at p 347, ie. ‘a poor man must not be ashamed to beg’ (cf. p 352), would appear to rule out dependence on Op. 317 (where οὐκ ἀγαθή is part of the subject, as at Op. 500), and lead Krafft to a conclusion which is the direet

reverse of Miss Sellschopp’s. 31 4JPh

Ixxxiv,

1963, 27;

see also his p. 17 n. 2, showing that indic.

κομίζει, and not the variant κομίζειν, must

composed it.

32 SUH, 65.

have stood in the line as Hesiod

172

THE

LANGUAGE

OF HESIOD

in question may be dependent on a third now lost. Fourteen phrases are discussed, and we will consider each of them. (6)

Th. ὃ and Th. ὃ

160 δολίην δὲ κακήν τ᾽ ἐφράσσατο τέχνην | 529 αὐτίκα δ᾽ Αἴγισθος δολίην ἐφράσσατο τέχνην 547 δολίης δ᾽ οὐ λήθετο τέχνης | 455 οὐδ᾽ ὁ γέρων δολίης ἐπελήθετο τέχνης |

Miss Sellschopp’s argument 33 rests on the idea that τέχνη in its oldest sense (as at J” 61) means a skill or craft, but at ὃ 529 it means the plot, the dark plan, by which Aigisthos murdered Agamemnon. She sees the transition to this sense in Th. 160, where Gaia prepares the sickle of adamant with which to castrate Ouranos: similarly at 6 455 it is not a physical skill that 1s meant, whereas such a meaning 18 present 1n the Prometheus story at Th. 547. But (a) at Th. 160 δολώ) τέχνη looks forward not to the fabrication of the sickle but to the whole story of the trick played on Ouranos by Gaia and Kronos (cf. δόλον at Th. 175); this is certainly how the translators have understood it.* (b) Equally at Th. 547 the phrase δολίη τέχνη does not refer merely to Prometheus’ skill in dressing up the bones, but needs to be taken in the context of the whole story, in which the same words occur at Th. 540, where δολίῃ ἐπὶ τέχνῃ 35 means ‘ for the purpose of a cunning trick! (for this use of ἐπί cf. o 44), and at Th. 547 and 560, where they refer to Prometheus' deceitfulness and not to any physical skill. (c) If we are to look among the surviving lines for one where the transition to the later sense may have been effected, then it is surely to be seen in ὃ 455, where δολίη τέχνη refers to the magical ability of Proteus to transform himself into other shapes at will. This is much closer to a physical art or skill than any of the other contexts where 33 SUH, 69. 834 e.g. “a crafty device’ (Mair), ‘a crafty and an evil wile’ (EvelynWhite), “an evil and cunning stratagem’ (Brown), ‘une ruse perfide ct cruelle’ (Mazon). 35 The same phrase occurs at Th. 555 which seems to have arisen directly rom 7h. 540:

see Kretschmer,

DIH,

37.

PHRASES

COMMON

TO HESIOD

AND THE

ODYSSEY

173

the phrase occurs, and it would be much more plausible to detect the origin of the phrase here than in the Hesiodic passages. (7) Op. θ2δ

πάντοθεν, ὄφρ᾽ ἴσχωσ᾽ ἀνέμων μένος ὑγρὸν ἀέντων

ε 478 8 τοὺς μὲν ἄρ᾽ οὔτ᾽ ἀνέμων διάη μένος ὑγρὸν ἀέντων

Here Miss Sellschopp exaggerates on the one hand the irrelevance of the words ὑγρὸν aevrwv in the Odyssey and on the other their appropriateness in Hesiod. She writes of the former: ‘der Wind wird ausgedrückt mit der hesiodischen Wendung, obwohl ὑγρὸν hier gar nichts besagt, da die Feuchtigkeit ja als ὄμβρος nachfolgt. Wir dürfen also wohl annehmen, dass Hesiod die Vorlage für diese Verse gebildet

hat.’ 3°

But the epithet ὑγρὸν ἀέντων is not so redundant as

this implies : Odysseus is concerned about the kind of shelter he must find for the night, and reflects that 'early in the morning a cold wind blows from the river’ (e 469); but he finds the double olive-bush which 1s so thick that the ἀνέμων μένος ὑγρὸν àévrov will not penetrate it (e 478). Miss Sellschopp is also wrong to say that ὄμβρος (e 480) duplicates the idea of ὑγρόν in 478: Homer is there concerned to show how thick and impenetrable Odysseus’ shelter is, be it by sun (e 479) or rain (e 480).

Pouring rain is not the same thing as

blown mist. Of the passage in Op. Miss Sellschopp has this to say: * Man soll auf jede Weise sein Schiff schützen, denn vom Meer, den feuchten Winden und dem Regen droht ihm Fäulnis und Schädigung. Die einzelnen Teile soll man unter Dach und ins Warme bringen ; Feuchtigkeit ist also die eigentliche Gefahr für das unbenutzte Schiff.’ 3° This is doubtless true of Op. 626 ff., but in the lines up to and including 625 it is the violence of the winds which is the leading idea : blasts of all 36 Also at τ 440, which we may reasonably (so Sellschopp).

87 SUH, 72.

38 ibid.

take as secondary to e 478

174

THE LANGUAGE

OF HESIOD

kinds of winds rage (Op. 621) in this season, and that is the time to avoid having your ship at sea (622 f.); draw it up rather on the shore, and pack it round with stones on all sides,

ὄφρ᾽ ἴσχωσ᾽ ἀνέμων μένος ὑγρὸν aevrwv (624 f.). The purpose of the stones is to prevent damage in storms, not to keep out damp. We conclude that, although the phrase ὑγρὸν aevrwv 18 appropriate as ἃ description of November winds, it 15 unlikely to have been created by Hesiod expressly for Op. 625. If any particular surviving passage is its original place, e 478 would appear to have a strong claim, though the phrase is used elsewhere as a formulaic epithet for wind,? and Homer may simply be making very felicitous use of an older expression. (8) Op. 508 μέμυκε δὲ γαῖα xal ὕλη | € 398 ὡς ᾽᾿Οδυσῇ᾽ ἀσπαστὸν ἐείσατο γαῖα καὶ ὕλη

Miss Sellschopp 49 15 troubled by the fact that Odysseus at

e 398 is described as being able to see γαῖα καὶ ὕλη, but when he swims closer he meets with only rocks and cliffs (e 404 f.) ; she concludes that the use of the phrase here is secondary, being known to the poet from Op. 508. Krafft's discussion *! shows well that this difficulty is more imaginary than real: he points out, among other things, that ὕλη 1s shown elsewhere in the story to be a feature of this coast, and that an observer approaching such a shore as the poet has in mind might easily see ὕλη from some distance out to sea but not from the foot of the chff. It may be added that the contexts at e 398 and e 404 f. are not identical: in the former, γαῖα καὶ ὕλη is what cheers Odysseus after two days in the sea, ie. the prospect of land, also of shelter, warmth and sustenance, all of which woodland provides, and even timber for ἃ boat (cf. e 243-57) ; but at e 404 f. Odysseus is concerned only with finding à landing-place, and the cliffs seem to present a hopeless obstacle. 89 At ἡ. vi, 3, and at Th. 869 in the Typhoeus episode where its use seems inappropriate : seo Kretschmer, DIH, 53; Sellschopp, SUH, 72 f.

© SUH, 73.

41 VUHH, 131.

PHRASES

COMMON

TO

HESIOD

AND

THE

ODYSSEY

175

If one of our two places has to be chosen as the source, it might be argued that the phrase is so deeply rooted in the context at e 398 that it must be original there,*? and taken up by Hesiod, who then elaborates the idea of the groaning of ὕλη in Op. 509-11. But again it would be unwise to rule out the possibility that the phrase is secondary in both Homer and Hesiod. (9) Th. 661 ἀτενεῖ τε νόῳ xai ἐπίφρονι βουλῇ | ῥυσόμεθα γ 128 ἕνα θυμὸν ἔχοντε νόῳ καὶ ἐπίφρονι βουλῇ [ φραζόμεθα

Miss Sellschopp *? contends that an expression which sounds ‘voll und ursprünglich ’ at Th. 661 has been taken from there by the poet of y 128. But an examination of the contexts suggests that this is most unlikely. The Hesiodic passage concerns the aid given to Zeus by the hundred-armed giants, who say that they will defend his power ἀτενεῖ τε νόῳ καὶ ἐπίφρονι βουλῇ. Yet the contribution which these giants have to make is in fact one of brute strength : Zeus himself already has all the wisdom he requires (cf. Th. 656), and does not need their νόος and βουλή. Mair accurately translates ‘ with steadfast mind and wise counsel', but the expression is strangely stilted in this place,** and West finds it so unsuitable that he reads: πρόφρονι θυμῷ, against the mss., on the strength of a papyrus fragment showing θυμῷ instead of βουλῇ. At y 128 on the other hand, both νόος and βουλή are to the point : Nestor is describing in his garrulous fashion how at Troy he and Odysseus used to be of one mind together and advised the Achaeans ‘ with understanding and wise counsel '. The presence at Th. 661 of the epithet ἀτενεῖ with νόῳ, 4? Thus Krafft, loc. cit., thinks that γαῖα καὶ ὕλη is not formulaic in the Odyssey passage, but that it is likely to be so at Op. 508, where γαῖα does not form a subject for μέμυκε in the same literal sense as ὕλη does: of. Wilamowitz ad loc. 43 SUH, 68 f. 44 N. O. Brown, with more thought for the context, resorts to ' with unflagging energy and loyal hearts ', though elsewhere in the poem he has rendered ἐπίφρονα βουλήν literally by ' prudent purpose’ (Th. 122) and * prudent understanding ’ (Th. 896).

LANGUAGE

THE

176

OF HESIOD

balancing ἐπίφρονι with βουλῇ, to which Miss Sellschopp draws attention, certainly does not indicate that the line ıs the original home for the phrase and the source of its use elsewhere.

The formulaic character of this half-line, without

the first epithet, may be seen not only in Th. 122 δάμναται ἐν στήθεσσι νόον kai ἐπίφρονα βουλήν 45 but also in the closely related νόον καὶ ἐπίφρονα μῆτιν | 7 326, μένος kai ἐπίφρονα βουλήν | Th. 896, μένος καὶ θυμὸς ἀγήνωρ | Y 174, μένος καὶ

ἀγήνορα θυμόν | A 562.

While no amount of discussion could

ever sort out these lines into their chronological order of composition, it is clear that Th. 661 could not be given the place at the head of the lıst.*®

(10) Op. 564 εὖτ᾽ ἂν δ᾽ ἑξήκοντα

ol μετὰ τροπὰς ἠελίοιο

663 ἤματα πεντήκοντα o 404 ᾽Ορτυγίης καθύπερθεν, ὅθι τροπαὶ ἠελίοιο

There is no doubt that in the Odyssey here ὅθι τροπαὶ ἠελίοιο means a place, whereas the Hesiodic lines refer to a time of year, 1.6. the solstice, as also at Op. 479 and commonly in later

Greek literature ; 47 but Miss Sellschopp 48 assumes too readily that the place meant at o 404 is the West, τροπαί being understood as the turning of the sun in its daily course, so that Homer's use of the phrase contains a sense never found elsewhere. The interpretation of o 404 raises various problems,*® but it is clear that τροπαὶ ἠελίοιο there also may equally be understood as the turning of the sun in its yearly course, 1.6. the solstice, so that

ὅθι τροπαὶ

ἠελίοιο

(sc. εἰσιν, φαίνονται)

signifies the place where this annual phenomenon

ooccurs.9?

15 Th. 122 is regarded by Sellschopp—though not by Schwenn !—as post-Hesiodic. 46 cf. also π 242 yetpas τ᾿ αἰχμητὴν ἔμεναι xai ἐπίφρονα βουλήν where ἐπίφρονα may well be masc. and not an epithet of βουλήν : see Stanford ad loc. If so, it might be argued that 7 242 provided the transition from eridpwv as an epithet of persons (cf. V 12) to its use with the abstract βουλή. *? Seo Liddell and Scott s.v. τροπή. 48 SUH, 70. 49 See Stanford ad loc.; H. L. Lorimer, Homer and the Monuments, 80-2; van der Valk, TCO, 111 f.

50 See P. Waltz’s argument cited in Stanford’s note.

1950

PHRASES

COMMON

TO HESIOD

AND

THE ODYSSEY

177

There need therefore be no essential difference between Homer's meaning of τροπαὶ ἠελίοιο and that found in Hesiod and elsewhere : the notion of place, after all, is contained in ὅθι

rather than in rporai,5! which by itself significs something which happens, 1.6. a turning round. It is probable that we are dealing here with a formula which is older than any of the passages in which it has been preserved.*? (11) Th. 166 = 172 πρότερος γὰρ ἀεικέα μήσατο ἔργα | υ 394 πρότεροι γὰρ ἀεικέα μηχανόωντο |

There is not much to choose here: in both places the purpose of the phrase is to indicate that those who are about to suffer—the suitors in Od. and Ouranos in Th.—have brought upon themselves the punishment that is coming to them. Miss Sellschopp 5? argues that γάρ fits the Homeric context less obviously than the Hesiodic one, but it is easy to exaggerate this difficulty. J. D. Denniston points out that this particle is not always used in a way that is logically precise, and he writes: ' Compression of thought is often the source of difficulty, and formal exactitude can then be achieved by supposing an ellipse.’ 54 In the passage at the end of v, the sequence of thought may be traced as follows : the suitors had an excellent dinner (v 390 f.) but nothing could be more horrible than the ‘ supper’ which Athene and Odysseus were preparing for them (392 f); then the ellipse—sc. ‘ this terrible punishment was going to be exacted from them '— because they did wrong first (394). No doubt, as Denniston says, to make the ellipse explicit in this way is “ psychologically somewhat misleading ', in that no such idea was consciously in the poet's mind ; and it is certainly very clumsy : much better is Butcher and Lang’s old-fashioned ‘ for that ’ instead 51 For the same construction cf. o 175:

ἐλθὼν ἐξ ὄρεος, ὅθι οἱ γενεή τε

TÓKOS T€.

52 See above p. 152 with n. 27, and cf. ávroAai ἠελίοιο | u 4. 53 SUH, 71. 54 The Greek Particles, 61, where he gives examples, including some from Homer, and cites Verrall’s remark on E. Med. 573, that ‘ the use of ydp is regulated by the substance of the thought, and not by its form.’

178

THE

LANGUAGE

OF HESIOD

of a simple ‘ for’, which somehow catches the train of thought exactly. In any case the idea contained in these lines is a commonplace. Miss Sellschopp cites 4 67, 72, 236 and 271 as Hesiod's source for the idea : all have πρότεροι in this same position, and 4 236 and 271 are particularly close in sense, with πρότεροι ὑπὲρ ὅρκια δηλήσαντο

| used by the Greeks, assuring

one another that the Trojans will certainly be punished for having broken the truce. But there are other Homeric parallels: for the same sentiment cf. δὸς τείσασθαι ὅ με πρότερος κάκ᾽ ἔοργε | I’ 351 (where Menelaos prays for revenge on Paris), and for the same line-structure, with γάρ in this position, and the same verb as in v 394, cf. τῶν δ᾽ ἄλλων μή τις Ἀχαιῶν | πευθέσθω: πολλοὶ yap ἐμοὶ κακὰ μηχανόωνται π 133 f.

Here, in the mouth of Telemachos, these words are

motivated by a concern for his own safety, but for the poet and his audience, who know that Odysseus is at last returned and that the suitors’ days are numbered, they cannot fail to carry a hint of the retribution that is to come. The same idea is

contained in ἀτάσθαλα μηχανόωντο | A 695 of the Epeians in Nestor’s tale, who are later fittingly punished by him, and in αἵ περ πρόσθεν

ἀεικέα

μηχανόωντο

| y 432, when

Odysseus

summons the disloyal servant women who are now to pay the penalty of death. There is therefore no need to suppose that the Theogony has provided the model for v 394.55 (12) Op. 315, 335, 646 ἀεσίφρονα θυμόν |

$ 302 ἀεσίφρονι θυμῷ | Again there is little to choose between one context and the other. Hesiod applies the phrase to his delinquent brother Perses, while in $ it is used of the unfortunate Centaur Eurytion, punished for committing follies under the influence of strong drink. Miss Sellschopp 59 feels that ἀεσίφρονι at 55 Krafft assumes VUHH, 101 n. 4. 5€ SUH,

the

reverse

relationship,

but

without

71 with n. 119, where she refers to Wilamowitz,

discussion :

HdO,

77;

but

Wilamowitz there admits the possibility that both Hesiod and Homer may

here be dependent on older material, and a year later in his edition he wrote of Op. 315 ‘ Hier nach ¢ 302 ' (Erga, 79).

PHRASES

COMMON

TO

HESIOD

AND

THE

ODYSSEY

179

$ 302 is betrayed as second-hand by the way in which it

duplicates the idea already expressed in φρεσὶν ἧσιν ἀασθείς

$ 301; but one might just as well interpret this as evidence for the phrase's originality in 6,5” since ἀεσίφρονι is thus tied more closely to its context than it is in Op., moral blindness being deliberately emphasised in & 296, 297, 301 and 302. But the phrase may well be formulaic in origin. For closely similar expressions cf. ἄφρονι θυμῷ | ¢ 105, h.Aphr. 286 ; πρόφρονι θυμῷ | m 257, Th. 536; πρόφρονι θυμῷ earlier in the line ὦ 140; θυμῷ | πρόφρονι Θ 40, X 184; κρατερόφρονα

θυμόν | Op. 147. The adjective ἀεσίφρων is found twice in II. (Y 183, V 603), and its first use with θυμός may be older than either Φ or Op. (13) Th. 117 Tai? εὐρύστερνος, πάντων ἕδος ἀσφαλὲς αἰεὶ | ἀθανάτων 58 ζ 42 Οὔλυμπόνδ᾽, ὅθι φασὶ θεῶν ἕδος ἀσφαλὲς αἰεὶ | ἔμμεναι

Miss Sellschopp’s argument °° begins from the assumption that ἀσφαλές in this phrase must mean the same as it does at O 683 (the only other place where the adjective occurs in Homer), where ἀσφαλὲς αἰεί | is used of a skilled horseman leaping from mount to mount, and she translates it there ‘constantly’, ‘incessantly’ (“unablässig’). She therefore understands the phrase at Th. 117 as referring to Earth as the constant, unalterable setting for the successive generations of the gods, with whom Earth’s own permanence is thus contrasted. This in her opinion provides the transition to its use at ¢ 42, where no contrast with more transient forms of existence is in question, and she concludes that the author of £

has taken over the phrase ready-made from Hesiod. But is it true that ἀσφαλές at O 683 means ‘constantly ' ? By derivation the word means ‘ without falling’, * without slipping ’, and it can be applied either to a moving subject or to 8? This view is taken by Krafft,

VUHH,

129.

58 Th. 118 is bracketed by Rzach and most editors, but is retained by West: see HT, 194; but the point does not much affect our discussion.

59 SUH,

73.

180

THE

a stationary one.©

LANGUAGE

OF HESIOD

Thus when uscd of the skilled horseman at

O 683 it means ‘ without faltering ', but at ζ 42, Th. 117, 128,

all with ἕδος, it means ' not slipping ’, ‘ firmly set’. Any idea of continuity in time is absent from ἀσφαλές in either meaning : if such an idea is present in these passages it is to be found in αἰεί, not in ἀσφαλές, but when there are so many formulas of this type with other adjectives,® it is obvious that the notion of time in ἀσφαλὲς αἰεΐ is very much subordinate to that of steadiness. This invalidates Miss Sellschopp’s argument from the alleged contrast between what is permanent and what is impermanent, and onc might argue that ἀσφαλὲς αἰεί is more fully in place at £ 42, where it is followed by the description in ζ 43-6 of that magically serene ‘other world’ which is the θεῶν Edos, than it is in the Theogony. The phrases ἀθανάτων ἕδος E 360, O 456, and θεῶν ἔδος E 367, 868, h. Apol. 109, are always found in close association with the name of Olympos,® and £ 42 stands close to this, with Οὐλυμπόνδ᾽ at the start of the line, making it most unlikely that Homer is here dcpendent on Hesiod, who applies the expression first to Gaia at Th. 117 and then to Ouranos at Th. 128 (a line which is suppressed by Jacoby, Sellschopp and Schwenn). (14) Op. 90 πρὶν μὲν γὰρ ζώεσκον ἐπὶ χθονὶ ςὔλ᾽

ἀνθρώπων

n 307 δύσζηλοι γάρ τ᾽ εἰμὲν ἐπὶ χθονὶ φῦλ᾽ ἀνθρώπων

In writing that, for the understanding of Odysseus’ speech at ἡ 303-7, ‘es... völlig belanglos ist, dass wir, die Stämme der

Menschen,

auf der Erde wohnen ’,® Miss Sellschopp is in-

$9 These two applieations are well illustrated by the adverb ἀσφαλέως in Homer (a) of moving subjects : with θέει N 141 of a boulder rolling downhill, with θέεν

v 86 of the

Phaeacian

ship taking Odysseus

back to Ithaea,

and with ἀγορεύει 0 171 of the eloquent king’s unfaltcring speech ; but also (b) of stationary things which remain firm or steady : ὡς (sc. like a tombstone) μένον ἀσφαλέως P 436 of the horses of Patroklos refusing to budge after their master's death, and aAA Euev’ ἀσφαλέως p 235 of Odysseus holding his ground when kicked by the insolent Melanthios. ?' e.g.

ἐμμενὲς

ἄφθιτον αἰεί | 3X

αἰεί

| 3X

IL, 2x

Od.,

νωλεμὲς

αἰεί | 4x

Il., μείλιχον αἰεί |, amos αἰεί |, each 1x

Il, 9x

Od.,

Il.

** cf. also ἕδος Οὐλύμποιο $2 144, h. Dem. 341, and ἕδος νιφόεντος ᾽Ολύμπου ἢ. xv, 7.

933 SUH,

74.

PHRASES

COMMON

TO

HESIOD

AND

THE

ODPSSEY

181

sensitive to Homer's art. Odysseus is explaining to Alkinoos why he had not allowed himself to be brought back to the palace by Nausikaa and her party : the reason which he gives is that he was afraid of Alkinoos taking offence when he saw a stranger with his daughter (n 305 £.). To confess this at once appears to involve the implication that he had regarded his host as likely to be ill-humoured and resentful of his presence, but Odysseus tactfully avoids this embarrassment by making it clear that he would regard such behaviour as ‘ only human’ : the jealous reaction which he had feared was the reaction to be expected from anyone in the same circumstances, so that Odysseus' earlier hesitation could not be taken by Alkinoos as implying a reproach against himself.5* The effect of ἐπὶ χθονὶ φῦλ᾽ ἀνθρώπων 15 to introduce the widest possible generality into the remark, and this is reinforced by the first person εἰμέν, showing that Odysseus does not regard himself as any different from the rest of mankind in the matter. Homer achieves all this in the space of a single hexameter. We cannot therefore dismiss out of hand the possibility that the poet of ἡ himself composed this quasi-proverbial line expressly for this place, making use as he did so of two formulaic phrases which are widely attested: ἐπὶ χθονί, in

this position, 10x 7L, 10x Od., and φῦλ᾽ ἀνθρώπων | 1x IL, 2x Od. (besides 7 307), 6x Hymns; cf. also φῦλα γυναικῶν | Alternatively the whole line could be genuinely 2x Il. proverbial, or based on a similar proverb concerning human nature from which the poet has taken the second half of the line only. We certainly need not suppose Op. 90 to be the

source;

and the recurrence of ἐπὶ χθονὶ φῦλ᾽ ἀνθρώπων

|

elsewhere 85 makes it even more likely that the complete expression was a formula in its own right, at least for Hesiod and his successors. 64 7 believe this is a stronger argument against Sellschopp's view than is Krafft's (VUHH, 107 n. 1), suggesting that the idea of 7 307 is anticipated by Nausikaa's words at £ 273 f.: her remark there concerns ordinary people among the Phaeacians, whereas that of Odysseus arises out of his reference to

Alkinoos personally (σοί in ἡ 306). ss Th, 556, Sc. 162, Fr. 23(a).25, 30.11, 291.4 M-W.

182

THE LANGUAGE

OF HESIOD

(15) Op. 547 } Βορέαο πεσόντος | ξ 475 Opinions are divided about the meaning of πεσόντος in this phrase. Some believe it to be ‘ when the wind drops’, i.e. at the subsiding of Boreas,99 but others take it in the opposite sense of ‘ when the wind assails i.e. at the onset of Boreas.’ Miss Sellschopp prefers the second alternative,°® and in this I agree with her: the whole gist of the passage in € seems to demand it, and her reference to an Ζέφυρος μέγας αἰὲν éQvOpos in the parallel situation at £ 458 is surely decisive.9 We therefore have to face the fact that πεσόντος of Boreas at & 475 does not mean the same as in the phrase ἄνεμος πέσε ‘the wind subsided’ at 7 202, again of Boreas. But why is it any more likely that this phrase originated in Hesiod rather than in Homer? Miss Sellschopp’s answer is presumably contained in her reference to Hesiod as living with mountains to the north of his home, and therefore, like the

inhabitants of modern Trieste, personally acquainted with the rigours of Boreas ; but the idea of Boreas as a rough wind and a bringer of wintry weather is a commonplace, of which there are many examples in Homer, e.g. O 170 f., T 357 f. Moreover the sense ‘ assail' for πεσόντος at £ 475 is in fact less extraordinary than Miss Sellschopp implies. For while πίπτω, without a preverb or prepositional phrase, is not used elsewhere in Homer or Hesiod in this sense of ‘ assail’, the compound ἐμπίπτω is so used, of Boreas, both in Hesiod (Op. 511) and in Homer (where, perhaps significantly, the verb is tm tmesi): ἐν δὲ πυρῇ πεσέτην Y 216, of Boreas and Zephyros attacking 66 These

include

Butcher

and

Lang,

Merry,

Monro,

Stanford

and

Rieu

in their translations or notes at this place in Od., Mazon in his Budé Hesiod, Ebeling (LH II, 186) and Cunliffe (LHD, 330). 9? This mcaning is adopted for Op. 547 by Goettling, Paley, Mair, EvelynWhite, Wilamowitz, Sinclair, and by Liddell and Scott s.v. πίπτω.

88 SUH, 74 f. 99 Krafft however (VUHH, 133 f.) argues for the meaning ‘ subside’ at both Op. 547 and £ 475. I would not follow him here or through the rest of his argument, but his conclusion is the same as ours, that it is not possible to

establish the priority of either passage.

PHRASES

COMMON

TO

HESIOD

AND

THE

ODYSSEY

188

Patroklos’ pyre and blowing up its flames. A similar use of πίπτω, this time with σύν as its separated preverb, occurs in the description of the storm at e 295 f.: σὺν δ᾽ Εὖρός τε Νότος τ᾽ ἔπεσον Ζέφυρός τε δυσαὴς | καὶ Βορέης αἰθρηγενέτης.

These examples suggest that our Βορέαο πεσόντος | could have its origin in a place now lost where the object was expressed

in a prepositional phrase, e.g. with | πόντῳ Ev ἀτρυγέτῳ in the same or the following line. But this offers no basis for establishing any chronological relationship between £ and Op. (16) T'h. 400 ἔθηκε θεῶν μέγαν ἔμμεναι ὅρκον | B 377 γρηὺς δὲ θεῶν μέγαν ὅρκον ἀπόμνυ |

Here we may grant two of Miss Sellschopp’s points, 79 namely (a) that the great oath of the gods consisted of Heaven, Earth, and especially Styx, and (5) that we never hear of men using the same form of oath as the gods. What need not be conceded 15 that the author of B wishes to imply that Eurykleia uses the gods' great oath, i.e. that she swears by Styx, etc. More reasonably θεῶν μέγαν ὅρκον is to be understood as ‘ an oath by the gods ’.”! Miss Sellschopp does not mention the most elaborate account of an oath made by ἃ mortal in Homer,

at I” 276-80, where Agamemnon

invokes Zeus, Sun,

Rivers, Earth, and avenging gods of the underworld : in the passage leading up to this, θεῶν undoubtedly serves as an objective genitive dependent on ὅρκια πιστά at I' 249 and 269. Alternatively we may suppose that the poet of B has taken the phrase θεῶν μέγαν ὅρκον, always used elsewhere of an oath which ἃ god swears, and has applied it to Eurykleia either without reflecting that it is inappropriate to use it of a mortal, or—which seems more plausible—having misunderstood or reinterpreted the phrase's meaning. But if so, neither the word-order nor the sense at 7}. 400 lends support to the view that it is Hesiod's line which has provided the model. Far 70 SUH, 67. 71 See the notes ad loc. by Merry and Riddell compares of θεῶν ὅρκοι at Xen. Anab. II, 5, 7: TCO, 54 f.

and by Stanford, who so also van der Valk,

184

THE LANGUAGE

OF HESIOD

more likely is a formulaic origin, with B 377 preserving the words in their traditional place,7? whereas Hesiod’s θεῶν

μέγαν ἔμμεναι ὅρκον | shows a departure from this order.” As far as the sense is concerned, Hesiod provides no obvious transition to the employment of θεῶν μέγαν ὅρκον of a mortal at B 377. Both at Th. 400 and at 784 and 805 he is talking about Styx as being the great oath of the gods. If a transitional passage is sought, we need one where the gen. θεῶν after μέγαν ὅρκον could be taken either possessively or objectively. Such an ambiguous usage is suggested by x 299, where Hermes tells Odysseus to extract from Circe à μακάρων μέγαν ὅρκον that she will not harm him. No details of this oath are given in x 345 f., but there can be little doubt that μακάρων here was intended to mean ' of the gods’ in the sense that Circe was a goddess and could use the gods’ oath.?* It is easy to see how, in such a context as this, where the nature of the oath is not

specified, the genitive after ὅρκον could have originated 1n one sense (possessive), have been later understood in another (as

objective), and so re-used. (17) Th. 70 ποδῶν ὕπο δοῦπος ὀρώρει | a 10 ποδῶν

δ᾽ ὑπὸ δοῦπον ἀκούω

Ϊ

This is a better example for Miss Sellschopp’s purpose than the earlier ones, for once we have agreed with her that ὑπό (ὕπο) is to be taken with ποδῶν at 7 10,75 then the obscure force of the prepositional phrase after ἀκούω betrays the line's wording as secondary. Other interpretations of ὑπό are possible which do not involve this difficulty,"9 but to take 72 The phrase μέγαν ὅρκον occurs in this position 3x IL, 5x Od., 4x Hymns, including θεῶν μέγαν ὅρκον ὄμοσσεν (-ca) | h. Apol. 83, h. Herm.

518 ; cf. also θεῶν μέγαν ὅρκον ἐνεῖκαι | T'À. 184, which is regarded as part of an interpolation by Jacoby, Sellschopp and Schwenn, so that it enter into Sellschopp's discussion ; but whether ‘ echthesiodisch ' is still evidence for the traditional word-order in this formula. 18. cf. on ἐπὶ δ᾽ αἴθοπα πινέμεν οἶνον | Op. 592, above, p. 135 with *4 Circe’s promise is identical with Kalypso's in e (x 343 f. =

where the ‘ oath of the gods’ Heaven, and Styx (ε 184-6).

does not or not, it n. 35. ε 178 f.),

is certainly uscd, i.e. she swears by Earth, 75

SU I, 70.

76 See Merry ad loc., and esp. Chantraine, GH II, 138 f.

PHRASES

COMMON

TO

HESIOD

AND

THE

ODYSSEY

185

ποδῶν ὕπο together seems reasonable, especially in view of ποδῶν ὕπο σείετο ὕλη | & 285 and the simple verb in δοῦπον ἀκούσας | K 354, « 556 and δοῦπον ἄκουσα | μ 202.

To postulate the prior existence of ποδῶν ὕπο δοῦπος ὀρώρει | which was then conflated with δοῦπον ἀκούω | by the poet of 7 10, provides a neat resolution of the difficulty, and to this extent Miss Sellschopp’s suggestion is an attractive one. But we may surely question whether Th. 70 is the only place in which the poet of the Odyssey could have found the expression.

The

formula δοῦπος ὀρώρει

| occurs in the Iliad

(I 573,

M 289), and the complete phrase ποδῶν ὕπο δοῦπος ὀρώρει |

though not attested in Homer, would be usable in contexts frequently recurring in epic, such as descriptions of a galloping

horse or of men going into battle: σμερδαλέον

κονάβιζε ποδῶν

αὐτῶν

cf. αὐτὰρ ὑπὸ χθὼν |

τε καὶ ἵππων

at B 465 f.

Since the two halves of the expression are attested separately in the Iliad at these positions, the half-line may well be formulaic and older than Hesiod in its origin.?? (18) The next example concerns

the phrase ἐξ ἀρχῆς, which

is found four times in the Odyssey and seven in the T'heogony. Miss Sellschopp 78 begins by drawing attention to the fact that where the word ἀρχή occurs in Homer it means the beginning of some particular thing or happening, except m the phrase ἐξ ἀρχῆς a 188, B 254, X 438, p 69, where the sense is merely

‘from of old’, with no particular beginning in mind, i.e. no point in time where some specific new thing began. She then asks how it has come about that this watered-down meaning occurs only in this phrase and only in the Odyssey, and she claims to find the answer in Hesiod's frequent use of ἐξ ἀρχῆς still in its precise sense, namely ' the beginning of the generation of the gods ’ (Th. 45 and 115), ‘ the beginning of their life ’ 77 Apart from T. 70 it is attested in the Hhoiat (Fr. 48 Rz. = 158 M-W), though various suggestions have been made which would alter it there: see both Rzach and Merkelbach-West ad loc., and Liddell and Scott, main text

and Addenda s.vv. &bove, p. 91. 78 SUH, 66 f. N

ὑπόδουπος

and

rodfos;

cf. also

Th.

703

discussed

186

THE LANGUAGE

OF HESIOD

when speaking of the children of Gaia and Ouranos (7}. 156), and ‘the beginning of evil’ (7h. 512). She concludes that ‘im echten Hesiod’ ἀρχή in ἐξ ἀρχῆς stil has its precise meaning, and by his frequent use of the phrase Hesiod provides the transition to the weakened sense found in the Odyssey. The neatness of this pattern is blurred somewhat by the presence in Th. of ἐξ ἀρχῆς in its weakened sense also, at Th. 203 and 452, and similarly am’ ἀρχῆς at Th. 425. Miss Sellschopp avoids this complication by supposing that these lines are interpolated ; but their authenticity 15 by no means universally suspected : 9 Rzach, Mazon, Evelyn-White and West do not bracket any of them, and the only editor to doubt the genuineness of all three is, predictably, Jacoby, but he 1s equally suspicious of Th. 45, 115 and 512 which are required by Miss Sellschopp’s argument to be genuinely Hesiodic. But even if it is granted that ἐξ ἀρχῆς originated in the way which Miss Sellschopp suggests, can we be certain that Hesiod himself was the originator ? Might not such a phrase have been so used long before the Theogony took its present shape ? It is àn expression likely to be found useful m a large number of contexts,?! and once established ıt would be available for use in

its more basic sense even after the weakened sense had developed. This view would account for the presence of ἐξ ἀρχῆς in both senses in the Theogony more satisfactorily than does a theory of interpolations. (19) Th. 46 oi τ᾽ ἐκ τῶν ἐγένοντο θεοί, δωτῆρες ἐάων 0 325 ἔσταν δ᾽ ἐν προθύροισι θεοί, δωτῆρες ἑάων

The same objection can be raised against Miss Sellschopp’s argument that 0 325 must be later than Th. 46.9? She says that ” This last example is much less plausible than the other three ; a more convincing one might be Th. 408 ‘ the beginning of her life ’ when speaking of Leto, but presumably Miss Sellschopp omits it as belonging to an allegedly spurious verse:

cf. Schwenn,

7'dH,

144, but sec West ad loc.

80 Seo West’s discussion of the * Hymn to Hekate’, ZT, 276-80. 81 On the likely existence of a pre-Hesiodie tradition of stories about the origin of the gods, either in prose or verse, sco Kirk,

HT, 12 f. $2 SUH

74.

SAT,

90, and

West,

PHRASES

COMMON

TO HESIOD

AND

THE

ODYSSEY

187

in the Odyssey passage the phrase θεοί, δωτῆρες ἑάων is merely formulaic

and bears no relation to the context, whereas at

Th. 46 it has special force, the new generation of gods being contrasted with the old one of Gaia and Ouranos. But can we be certain that Hesiod orıginated this expression, which occurs also at Th. 111, 633 and 664 ? It is relevant to compare the formulaic δῶτορ ἑάων | used in the same Odyssey passage (0 335) and at A. xviii, 12 and A. xxix, 8, in all of which the

words are addressed to Hermes. In any case, how integral is the context at 0 to the rest of the Odyssey ? This episode of Ares and Aphrodite is full of linguistic oddities, 8° and its tone is reminiscent of some of the Homeric Hymns. Even if we decide that the story is ‘fully worthy of Homer in lighter vein ',** how integral is the phrase to the episode as a whole ? We are dealing here with an example of a ‘ departure from economy ’ within Homer, since he elsewhere uses θεοὶ peta ζώοντες | Z 138, ὃ 805 and θεοὶ, roi "Ολυμπον ἔχουσι | ξ 394, σ 180.

A casual variation like this could have been introduced to the text at a relatively late stage, and reflect a whim on the part of the recorder of the poem in writing.5? It therefore cannot be pressed as proving anything about the date of the poem's composition or its relation to the T'heogony. We now draw together some of the points which have emerged from this discussion, and summarise what appear to be the chief reasons for rejecting the view that the language of Hesiod in these places has influenced that of the Odyssey. First to be dismissed are those examples where one may question specific points in her understanding of the Greek which are essential to Miss Sellschopp's argument: the most obvious instances of this are the lines which contain ἴσκε (no. 1 above) and ἀσφαλές (no. 13), but there are four others (nos. 3, 6, 10 and 16) where she has followed an interpretation which seems unacceptable, and a further two (nos. 2 and 5) 84 Stanford, ibid. 83 See Stanford on @ 266. applied to similar variations in Hesiod, we which arguments the 85 cf, above, p. 72.

188

THE

LANGUAGE

OF HESIOD

where recent discussions of the passages in question must raise serious doubts about her view of their relationship. Miss Sellschopp has assumed too readily that these parallels are to be explained by supposing Hesiod to have provided the source for the poet of the Odyssey, without giving due weight to features in her examples which do not support such an interpretation and which might even suggest the opposite.®® A certain amount of prejudice is evident in some of her arguments

(see esp. nos. 4, 6, 7, 8, 9 and 14).

One ought to

make a much bigger allowance for the possibility of both places being dependent on a common source elsewhere (this is true to some extent of all the examples, but especially of nos. 4, 6, and 9 to 19 inclusive).

It is clear that this kind of

problem needs to be approached in a more methodical way, with basic principles laid down which will reduce the role of subjective judgment to the minimum. Two such principles are established by A. G. Lee in a paper 37 concerned with a similar problem in Latin. He writes: ‘ When trying to decide the priority as between two parallel passages in different authors, (i) we ask whether the passages are so closely connected that one author must be imitating the other. (11) If probability rules out both a common source and pure coincidence, then we ask whether there is any small misfit in either of them, whether both are equally organic constructions. If we find anything awkward or inorganic, then the probability is that the passage containing it is the imitation. 99 When we approach our Hesiod-Odyssey parallels with this in mind, we find that, even if we are prepared to give a positive answer to the first question, there are only two of the nineteen examples where it is possible to speak of something ‘ awkward or inorganic’ in the Odyssey passage, namely θεῶν ὅρκον at B 377 (no. 16) and ποδῶν ὑπό at π 10 (no. 17). We may infer that these lines contain a secondary use of an older phrase; but it is much 8$ [n over half the examples (nos. 1-3, 5-8, 11-14) as good a case—if not a better one—can be made out for the priority of the Odyssey over Hesiod. 87 * On [Tibullus] III, 19 (IV, 13)’, PCPAS

88 op. cit. 6.

elxxxix, 1963, 4-10.

PHRASES

COMMON

TO HESIOD

AND

THE ODYSSEY

189

more difficult to find a line of which we may say with any degree of conviction : * Here such-and-such a phrase is being used for the first time; this is its original place. Miss Sellschopp has demonstrated that B 377 and « 10 each contains an (mitatio ; she has not shown, nor is it possible to show, that the corresponding Th. 400 and 70 provided the exemplum.® This brings us to a final point which is of great importance. Miss Sellschopp's discussion adopts too rigid a view of the process of composition by which both the Odyssey and the Hesiodic poems were made. It is notoriously difficult to trace literary influences even where we know that the works in question were composed by writing and so had a fixed text from the start.? If one or both of them should belong to an oral tradition, the difficulties are infinitely greater, and if there is any possibility of a chronological overlapping of those traditions, 1.6. that both poems could have existed in some form of oral tradition before either of them became fixed in a written text, then they are insuperable. For, given two poems A and B, now in a written text, however well a word or phrase fits its

context in A, it is impossible to prove that it was invented for that place at the moment when the text of A became fixed. We can never rule out the existence of an older place X, which provided a common source for both A and B at the lines in question, so making their chronological relationship impossible to determine. This remains true even if X was only an older version of A. If an oral tradition has played any major part in the early history of the Hesiodic poems, then the common phrases which form the substance of Miss Sellschopp's argument—even if the examples themselves were much more convincing than we have found them to be—cannot provide ἃ satisfactory foundation upon which to build a theory of Hesiod's priority to the Odyssey. 89 Τί is convenient to use these terms of Kretschmer (see above, p. 19), though obviously when applied to the early Greek hexameter tradition their connotations are misleading. 90 See

for

example

1956, 465-80, esp. 465.

L.

Herrmann,

' Catulle

et

Lucréce',

Latomus

xv,

CHAPTER

XII

CONCLUSIONS

WE arc now in a position to turn again to the problems which were outlined in Chapter I and to reconsider them in the light of the various features in Hesiod's language which have since been discussed. First,

was

Hesiod

an oral poet or did he compose with

the

aid of writing? The most obvious consequence of the practice of oral composition in early Greek poetry was the development of ἃ highly complex system of formulas; but one cannot assume that all poetry which makes use of these formulas, even when that use 1s extensive, was itself orally composed. It has been shown by more than one scholar that literate poets may adopt a style which includes repetition of formulas on the Homeric model.! We therefore need to make a distinction, as G. S. Kirk has recently emphasised, between natural composition in a formular tradition and deliberate, selfconscious composition in a formular style? Kirk himself singles out three further signs of truly oral composition besides the extensive use of formulas: these are (1) the observance of the principle of economy, (2) the naturalness of formular extension and articulation, and (3) traditional details of rhythm and enjambement.? Let us consider how Hesiod stands in respect of these features. In general he appears to stand in substantially the same position as Homer. His readiness to repeat phrases within his own poetry and his obvious familiarity with a large stock of memorised expressions * show him to be following the habits of an oral poet in much the same way as Homer does. Departures from the principle of economy undoubtedly exist, but 1 See Hoekstra, HMFP, 17 with n. 2 on Quintus Smyrnaeus Arion vi, 1967, 320 f. on Nonnus.

2 also * *

; D. Young,

* Formular language and oral quality’, YCS xx, 1966, esp. 174; 161-3 on the Batrachomyomachia. op. cit., 174. See above, Chapters III and IV, esp. pp. 36-8, 43-5, 50.

see

CONCLUSIONS

191

these would seem to be accounted for better by the working of analogy within an oral tradition than by the deliberate exercise of choice in favour of a novel expression, which might be made possible by the use of writing ; 5 certainly in his use of formulaic expressions for ‘ the sea’ Hesiod is quite as strict as Homer in the observance of the principle.? As for the naturalness

of formular

extension

in Hesiod,

this has been

sufficiently illustrated both in the discussion of economy 7 and in the many examples of ‘ parallels of sound ’, which show the poet building upon traditional material. The ‘ parallels of sound ' especially suggest not a conscious process but rather the work of voice and ear, involving the poet’s memory of sound and metre no less than that of the sense and ideas of traditional poetry.9 And with regard to rhythm and enjambement it is striking how Hesiod stands firmly beside Homer in contrast to later poets, such as Aratus, who are known to have used writing.? What reasons, then, might be adduced for supposing that Hesiod was literate ? Many of those scholars who believe that writing played some part in the composition of the /lad and Odyssey do so because of their feeling that poems of such length and complexity could not have been created and preserved by an oral tradition alone.!9? But it is clear that this particular reason cannot be invoked as an argument for the written composition of the Hesiodic poems: none of them poses a problem with regard to its length, which in each case 1s 5 6 7 8 ® 10

See See See See Sce See

Ch. V, esp. pp. 60 f. Ch. IV, esp. pp. 50 f. Ch. V, esp. pp. 62-8. Ch. VI, esp. pp. 74, 84. Ch. VII, esp. pp. 87, 96 f. Bowra, Homer and his Forerunners. 1955, esp. 9 f.;

Lesky, GGL,

53-6 = HGL, 37-9; Adam Parry, ' Have we Homer’s Iliad ?’, YCS xx, 1966, 177-216. For the opposite view see Page, HHI, 261; Kirk, SoH, 98-101; M. W. Edwards, TAPA xevii, 1966, esp. 177-9. Adam Parry’s

remarks about the fast rate of change in an orally transmitted text (op. eit., 188 f.) are not supported by the remarkable examples of verbatim memorisation of oral poetry in Scotland, to which D. Young draws attention, Arion vi, 1967, 285, 296 f.

102

THE

LANGUAGE

OF HESIOD

more closely comparable to that of a single book of Homer than to the Iliad or Odyssey as a whole ; in this respect they comprise convenient units for oral recitation. As to complexity, they contain exceedingly little in the way of internal crossreferences, and the loosely-knit and unsystematic structure of the Theogony and the Works and Days is perhaps more consonant with the theory of oral composition than with the laborious care associated with planned writing.!? Nor does the Shield, consisting largely of narrative and description, appear to contain anything which would be beyond the competence of an oral poet. Granted that writing was known and used in the period when these poems were composed, there seems to be nothing in their content which might cause an experienced reciter to feel any need to aid his memory by writing them down before performing them. But perhaps there are special reasons for suggesting that the two longer Hesiodie poems must have had a written origin. In the first place we are conscious of the fact that the Works and Days, addressed as it 1s to Perses, belongs to a particular social situation which includes the poet's relationship with his brother and with the βασιλεῖς. Secondly, we have the autobiographical information contained in various parts of the Works and Days and in the opening passage of the Theogony, whereby the poet himself obtrudes into his poetry in a way quite alien to the Homeric epic ; the notion of a post-Hesiodic oral perpetuation of a fixed text, complete with these personal allusions,

seems

exceptionally

improbable,

celebrated

unless

person.

11 See Notopoulos' diseussion, HSPh

Hesiod

Thirdly,

had

some

been

an

of the

Ixviii, 1964, 1-18, esp. 11, where he

shows that the average length of Serboeroatian and Cretan oral poems is of the same order as we find in Hesiod. 12 Contrast however the remarks of Carpenter eited above, pp. 44, 45, and West, HT, 40 and 48.

15. West makes this point well (HT, 40), but the argument is not so strong as it may appear at first sight : see Solmsen's review of HT in Gnomon xl, 1968, 326. I would regard it as very likely, however, that the name of Hesiod (T'À. 22) entered the poem when the text was first written down : for similar examples from other literatures, see Bowra,

ΠΡ, 407 f.

CONCLUSIONS

193

subject-matter, especially in the more technical parts of the Works and Days, may seem to be unlikely material for oral poetry. But none of these three reasons is conclusive: the first two certainly need not compel us to think that the poems were composed with the aid of writing; they might equally have been composed orally and recited by the poet on many occasions before being recorded in a written text,!^ even if the personal references which they contain suggest that this recording may have taken place in the poet’s own lifetime. And while the third reason, namely the abstruse nature of some of the subject-matter, might be the most likely of the three to persuade us that Hesiod used writing as an aid to his composition, judgment on this point depends very much upon one's view of the capabilities and limitations of the oral poet. In the light of our discussion of an exceedingly technical passage of the Works and Days,!? it may be doubted if this provides a sufficiently cogent reason for believing that writing must have played a part in the composition of the Hesiodic poems. In Hesiod as in Homer then, we are dealing with poems whose language and style show features which are best explained by reference to an oral tradition which these poets inherited and themselves practised. An important problem therefore arises concerning the relation between the poems as they now exist and the form in which they were composed and recited by their authors. This brings us to the matter of dialect influence in the Hesiodic poems. It seems true to say that the most significant difference between their language and that of the Iliad and the Odyssey lies not so much in the conspicuous but often isolated un-Homeric word-forms,!9 but in the increased proportion of relatively late Ionic features!" 14 Such a reconstruction would accord with B. A. van Groningen's view that the Hesiodic poems were intended for public recitation rather than private reading : see La composition littéraire archaique grecque, 1958, esp. 270, 282 (Th.), 301 f. (Op.), 121-3 (Sc.).

15 Seo pp. 32-6. 16 i.e. those discussed in Chapters VIII and X. 17 See Ch. IX, esp. pp. 131, 139 £., and further below, pp. 201-3.

THE LANGUAGE

194

OF HESIOD

While all three poems belong to mainland Greece and it is reasonable to suppose that they were composed by and recited among speakers of mainland dialects, the signs of mainland dialect influence are less extensive than one might expect. Nor ean they be associated with any particular locality or dialect source : definite Boeotisms are few or non-existent, and

it appears that the traditional scholarly view of this aspect of Hesiod's language is broadly correct, and that the un-Homeric elements which it contains originated in several different dialects.18 That these elements should be so diverse in their affinities 1s at first sight surprising but is not difficult to explain, since the presence of any given form may have resulted not from the influence of the poet’s own spoken dialect but from his knowledge of poetry, not known to us, of the mainland and of Aeolia.

As

far as

eastern

Aeolic

features

are

concerned,

Hesiod’s own testimony that his father migrated to Askra from Kyme (Op. 636) is sufficient to account for their presence, even

if it is impossible to determine whether such a form as τριηκόντων 19 came into his language by way of his father's day-to-day speech or through his knowledge of earlier Aeolic poetry. But there seems to be no reason which obliges us to suppose that these dialect influences should be confined to East Greek : ?? while it is true that West Greek dialects played no part in the formation of the language of the Homeric epics, North West Greek and Doric speakers had by Hesiod’s time been established for centuries in most of mainland Greece, and

Boeotia was an area which came strongly under their influence.?! One wishes that we knew more about the spoken 18 19 2° last

See pp. 119-21. See p. 115. cf. pp. 120 £., 161 f. I should here dissent from the view implied by the paragraph of Mrs. Davies' article, Glotta xlii, 1964, 165.

? On Scherer,

the West Greek element in the Boootian dialect, see ThumbHGD II, 15 ff. The usual view that Boeotian is an Aeolic dialect

strongly infiltrated from West Greek may be inadequate: see R. Coleman's analysis, T'PAS 1963, esp. 118 f., where he suggests that Boeotian might be regarded as having been from the beginning a bridge dialect between Aeolie and Dorian.

CONCLUSIONS

195

dialects of Greece at Hesiod's date, but in the absence of fuller

evidence we must be content with the conclusion that he uses the complex and already highly evolved language of hexameter poetry, predominantly Ionic in its dialect, into which there occasionally intrude elements not found in Homer and apparently derived both from the Aeolic of the eastern Aegean and from dialects in mainland Greece. Limited as these peculiarly Hesiodic elements are, they may nevertheless help us in considering the problem of the unity of the corpus.?? Although the Theogony and the Works and Days are for the most part widely separated in their subjectmatter, we have found nothing in their language to suggest that they must be of independent authorship.?? Indeed, a close similarity between them 1s seen (a) in the presence of some of the same un-Homeric features, and (b) in the direction, and occasionally also the extent, of their divergence from Homeric practice in the exercise of choice between alternative forms. The main details which have been discussed above may be briefly summarised. (a) Un-Homeric features common to both poems are



(i) a-stem gen. pl. forms in -ἂν Th. 41, *129, Op. 145 ; ?4

(ii) loss

of

postconsonantal

F

without

compensatory

lengthening in κἄλόν Th. 585, Op. 63, ἴσον Op. 752 ; 35

(ii) a-stem acc. pl. forms in -às 6 or 7x Th., 3x Op.;?? (iv) the high proportion of both -as and -ovs acc. plurals which stand before a consonant rather than ἃ vowel.??

?? See pp. 4-7. 23 Rzach's conclusions about this in DH (see above, p. 17 with n. 25) are not borne out by the examples discussed in Ch. VIII and Ch. X: see csp. p. 146. 24 See p. 103. 25 See p. 107. One may mention here also the loss of intervocalic F, followed by contraction, in a number of words which are found uncontracted in Homer : see p. 139 n. 53. 26 See pp. 142-4. 27 See p. 158.

THE LANGUAGE

106

OF HESIOD

(6) A common divergence from Homeric practice in choosing between alternatives is seen in (v) a higher proportion of contracted o-stem gen. sg. forms in -ov beside the disyllabic τοιο, -oo ; ?? (vi) ἃ higher proportion of masc. a-stem gen. sg. forms in -ew, -« beside the alternative -ἄο ; 39

(vii) a higher proportion of a-stem gen. pl. forms in -écv, -@v beside the alternative - ἄων ; 30 (viii) an increased ‘ neglect ' of initial 6.31 The Shield does not share in this conformity : of the unHomeric features listed here (i-iv) it shows none, the only hint of something similar being contained in the isolated form Aayos at Sc. 302.? In the features where the other Hesiodic poems consistently show an increased preference for the developed Ionic forms, the Shield’s behaviour is erratic: it gives much the same picture as the Theogony with regard to -ov beside -oto (v) ; it contains no examples at all of -ew, -w (vi), but has a very high proportion of Ionic -έων, -öv (vii); and in the matter of initial F (viii) it stands closer to Homer than to the Theogony or the Works and Days. This confused pattern doubtless results partly from the fact that the Shield is a relatively short poem, and the paucity of examples makes it unwise to attach too much significance to some of these points (especially vi and vii); but taken all together, and in conjunction with the evidence of the metrical features discussed in Chapter VII,®* they give strong support to the view that the Shield is not the work of the same author as the Theogony and the Works and Days. This would appear to apply to the opening passage of the Shield no less than to the rest : admittedly this contains the dialect form Φίκιον at Sc. 33 in common with the Theogony’s d'ixa,?* but this name may have been 28 29. 30 31 3? 33 34

Soo See Seo Seo See Seo See

p. 122 with n. 3. p. 123. p. 128. pp. 132 f. pp. 164 f. esp. pp. 87, 89, 99. p. 106.

CONCLUSIONS

197

known in various parts of Greece, and in any case its presence in the Shield might be explained by the poet’s knowledge of it in other verse, perhaps in the Theogony itself.?9 And if we are right in thinking that the author of Sc. 7 is a poet who has misunderstood Th. 910,36 it does not seem possible that the beginning of the Shield can be by Hesiod. The poem may nevertheless have a complex history, and we cannot rule out the possibility that substantial passages of it might have existed orally before it came to take its present form.?? As far as concerns the various passages in the poems which have sometimes been suspected of being un-Hesiodic,** it must be admitted that there is nothing much in the linguistic evidence either to confirm these suspicions or to contradict them. The short acc. pl. forms, which seem to be the most distinctive characteristic of Hesiodic morphology, are confined to passages which are not suspect, with the possible exception of εἴρας Th. 804 in the description of Tartarus, but the reading there is far from certain.?® It may be mentioned that there is no notable increase in the proportion of ‘ heavy’ -às or -ovs accusatives in the suspected parts of the poems which might have caused us to hesitate in accepting them as Hesiodic ; * but here one is clutching at straws. If these passages are interpolations, they may have entered the text at a very early stage, since their language does not obviously distinguish them from those parts of Hesiod which are held to be genuine, and they show the same familiarity with traditional phraseology and versification.*! Such a suggestion receives some support from the presence in all the suspect passages of ‘ parallels of 35 That the author of the Shield knew Th. and Op. is shown to be probable by Kretschmer's examples of common phrases (D/H, 41-50). 36 See pp. 116-8. 37 See further M. van der Valk, ‘ Le Bouclier du pseudo-Hésiode ', REG Ixxix, 1966, esp. 453 f., 466 f.

38 See pp. 4-6. 39 See p. 143. 19 Examples of -às 4- vowel occur at Th. 712 and Op. 828, and of -ovs + vowel at Th. 935, 995, but this does not represent a higher proportion than elsewhere, and in any case the position is complicated by the existence of possible models in Homer for the phrases in which most of them stand :

see pp. 144 f.

41 See esp. pp. 97 f.

THE LANGUAGE OF HESIOD

108

sound ' similar to those found elsewhere in the poems.** We must also remember that the identification of particular passages as interpolations often rests on an assumption that the poems were composed according to a logical and consistent plan. Here the whole question of oral and written composition is clearly of crucial importance. It is proper to ask whether the sort of plan which is to be expected in a poem belonging to this early, oral phase of Greek literature might differ from that found in ἃ mature written style. The possibility arises that some at least of the suspect passages are not interpolations in the sense of being insertions into a written text: they may have entered the poem at the time of its recording, whether in the poet's own hand or in someone else's, or they may even go back to an earlier stage when the form of the poems was still being developed in successive recitations by their author himself. It is perhaps worth saying a little more with regard to the end of the Theogony. While the ‘linguistic argument ’ is not the only one which West adduces for separate authorship here, he nevertheless believes that it 1s a much stronger one than most linguistic arguments applied to problems of authorship in Hesiod ’.*? Yet of the five formulas which he cites as common to Th. 901-1020 and the Catalogue but not occurring in the rest of Th., those with ἄκοιτιν |—his (a) and (b)—are very

closely related to expressions of common hexameter stock : cf. esp. 1" 53, 1 397, ® 479, ὦ 537, e 120, n 66, A. Dem. 79. The same is true of (c) μιχθεῖσ᾽ ἐν φιλότητι: of. B 232, H 302, & 314, 331, 360, ε 126, ψ 219, h. Aphr. 263, ἢ. xxxii, b, as

well as of (d) μιγεῖσ᾽ ἐρατῇ φιλότητι |, which is an expansion of φιλότητι μιγεῖσα

| τ 266.

In fact the only distinctive features

in these four formulas are the use of θέτο in (b) and of ἐρατῇ in (d), and neither of these recurs in the Catalogue whereas we may compare from other parts of the Theogony ἄλοχον θέτο Μῆτιν | Th. 886, and épards as an epithet of an abstract noun at Th. 65, 70 and 879. As for (e) διὰ χρυσέην Ἀφροδίτην |, it includes a formula common enough in Homer which reappears 42 See pp. 77 f., SI.

^ HT

398.

CONCLUSIONS

199

at Op. 65 as wellas Th. 822: cf. esp. παρὰ χρυσέῃ Ἀφροδίτῃ] θ 337, 342. To get this into perspective we need also to look at the rest of the picture. One of the reasons why there are * no less than four formulae relating to marriage and reproduction ’ 44 which are repeatedly used in Th. 901-1020, but not elsewhere in the Lheogony, is that the subject crops up most frequently in this part of the poem. Tt must be pointed out that there are other formulas on this same subject which the passage shares with parts of Hesiod where the authorship is not disputed, among which the following have no direct Homeric parallel: (a) Γραίας (Χάριτας) τέκε καλλιπαρήους | Th. 270, 907 ; (b) ἱερὸν λέχος εἰσαναβαίνων (-Báca) | Th. 57, 939; (c) ἐν φιλότητι | γείναθ᾽

Th.

961

f. and

| yewar’

. . . ev φιλότητι

| T^. 1012

beside ἐν φιλότητι... | . . ἐγείνατο Th. 406 f. and | γείναθ᾽ ... ev. φιλότητι | Th. 374; (d) | γείναθ᾽ ὑποδμηθεῖσα Th. 314, 962; and (e) πολυχρύσου Ἀφροδίτης | Th. 980, Op. 521.4% West's conclusion that ‘ it is evident that we are here reading a poet with a different formulaic vocabulary from Hesiod's ’ 46 seems to be very much exaggerated. This does not mean that the end of the Theogony is necessarily Hesiodic; but this argument from language should not be counted as evidence which condemns it. Finally we come to the difficult question of Hesiod’s Here it must be frankly recognised chronological position. that as far as establishing an absolute date for the poems is concerned, the contribution which the linguistic evidence can make is very limited indeed. The main reason for this is the paucity of inscriptional material from the eighth to the sixth centuries B.c., as a result of which we have no firm chrono-

logical framework to which our early literary texts might be related. A further difficulty is raised by the obviously large 44 West, ibid.

45 On this formula see p. 78. The common ground between 7'h. 901-1020 and the rest of the Theogony is in fact very extensive indeed, as may be gathered from a study of Kretschmer’s list of parallel expressions, D/H, 9-19. 46 AT, 398.

200

THE LANGUAGE OF HESIOD

number of traditional elements in the language of the Hesiodic poems which are of no value in determining the date at which the poems reached substantially their present shape. The problem with Hesiod is thus very similar to the one which has to be faced in considering the ‘early’ and ‘late’ linguistic features in the Iliad and the Odyssey: one may be confident that Homer and Hesiod lived in a period after the loss of initial F in some parts ofthe Greek world, and after the emergence of such forms as the Ionic genitives in -ew, -έων and -ov, and those

with contraction following upon the loss of intervocalic F ; but this only provides us with a vague terminus post quem, which itself cannot be dated with any precision.*? As for a terminus ante quem, it 1s hard to see how the linguistic evidence can be of any help at all when one is dealing with poetry composed in this traditional style with its highly complex dialect mixture. Where the evidence of language has more to offer is in deciding Hesiod's relative date. Here also great caution is necessary, since the proportion of relatively ‘early’ as opposed to ‘late’ forms depends not on the date of composition but on the extent to which the poet is able to utilize traditional material, either in complete lines and phrases or in single word-forms.*® This is the reason why West has more than once challenged the idea ' that a relative chronology of epic poems can be established from comparison of the degree of linguistic development they show ',*? and he has cited the Shield as a “control case '.9 But perhaps this is unduly pessimistic: the ' observation’ of F in the Shield is only one aspect of its language, and other linguistic features would certainly point towards a relatively late date.9 West’s pro47 See Kirk's discussion, Museum Helveticum xvii, 1960, esp. 201-3. 18 cf. pp. 37-9. 19 Gnomon xxxv, 1963, 11. 50 OR xv, 1965, 159.

51 See above, pp. 196 f. One may mention here G. M. Bolling’s study of the frequency with which participles occur in each book of the Iliad and in the Hesiodic poems: whereas the figures for Th. and Op. fall well within the limits of the range

of variation

found

within

the

Jliad, the

Shield

stands

above the upper limit (The Participle in Hesiod, 1897, 470). With this result compare the Shield’s behaviour in the matter of spondaic line-endings, above, Ῥ. 87.

CONCLUSIONS

201

posed alternative method of establishing the sequence of the various poems by narrowing down their absolute dates, while excellent as an ideal, is impracticable where so little in Homer, and still less in Hesiod, is objectively datable with the requisite degree of precision.9? Nevertheless one must agree with West that the relative proportions of these ‘early’ and ‘late’ linguistic features in Hesiod and Homer cannot be used as an index of the relative dates of the poems' composition. A more certain clue to Hesiod’s relative date is perhaps provided by the dialect of the poems, and here we refer again to the predominantly Ionic appearance of his verse and the increased proportion of relatively late Ionic features which it contains. While various explanations for this are possible in theory,9? the only feasible one is to suppose that Hesiod did not compose 1n the dialect which he spoke but in the traditional language of hexameter poetry, which now, superficially at least, showed more in common with spoken Ionic than with any other dialect. This for Hesiod was the appropriate language in which to compose his own verse. Why then should Hesiod, living in mainland Greece, and of Aeolic parentage, be so thoroughly familiar with Ionian hexameter poetry ? First let us look at Notopoulos’ answer to this question. He sees the epic as originating in mainland Greece before the Ionian migration and surviving in two streams, ' the Ionian in Homer and the mainland stream in Hesiod and the other early epic poets ’, and he adds that ‘ the two streams join later in the 52 See further below, pp. 203-6. 53 Fick’s view of a post-Hesiodic Ionization of the poems insuperable

difficulties:

see

West,

HT,

80 n.

2, and

above

130 f., 137 £. At the other extreme, G. H. Mahlow suggested an Ionian who composed in his mother-tongue (Neue griechische Sprache und Dichtung, 1926, 256 f.), but this is with the autobiographical references in the poems (Th.

runs against pp.

37,

102,

that Hesiod was Wege durch die hard to reeoncile 23, Op. 635-40,

649-51) and does nothing to explain the features discussed in Chapters VIII

and X. Mahlow grossly underestimated the extent of the common ground between Hesiod and Homer when he wrote that ‘die "Epya xai “Hyepat haben mit der heroischen Epik ausser dem Metrum nicht das geringste gemein ’ (loc. cit.). O

202

THE LANGUAGE OF HESIOD

mainland '.5* He clearly regards Hesiod as composing without direct knowledge of the Homeric poems themselves and at a date which may be earlier than that of Homer, and he believes that ‘ the Homeric repetitions within Hesiod are the basic oral vocabulary of the Achaean epic’.®> But even if we were to place Hesiod at such a high date as the middle of the eighth century, Notopoulos’ theory would mean that the mainland and Ionian traditions had been developing separately for several generations—perhaps as long as two hundred years. It is surely inconceivable in these circumstances that the common ground between Hesiodic and Homeric language could be so extensive aS was apparent in our earlier chapters (especially III, IV and VI). It is also inconceivable that the distinctively Tonic features of the epic dialect, seen in both Homer and Hesiod, had already developed and become part of the hexameter tradition before the Ionian migration, or alternatively that they developed independently on the mainland and in Ionia after the separation.°° We may accept the notion of a widespread ‘epic kowne’*’ as the traditional language of hexameter poetry in which early and late forms existed side by side ; what seems unnecessary and untenable is Notopoulos’ belief that this survived independently in mainland Greece and in Ionia, to be joined again at a date later than Hesiod and Homer. The alternative is to suppose that Ionian epic poetry was known on the mainland in the time of Hesiod, and was recited

in its Ionic form even by mainlanders. The Homeric poems suggest themselves as the most obvious representatives of this Ionian tradition, but clearly they need not have been the only Ionian poetry which Hesiod could have known,5® nor can we assume that Hesiod knew them in the form in which they have 54 Hesperia xxix, 1960, 197. 55 jbid.; the same view is maintained in his later article, HSPA Ixviii, 1964, esp. 42 f. 5° See further the pertinent criticisms made by Hoekstra, HMFP, 25 f. 57 Notopoulos, op. cit., 196. $8 cf. Kirk, SoH, 285, on evidence for the Cypria and Aithiopis being in circulation on the mainland in the first half of the 7th century.

CONCLUSIONS

survived in our written texts.

203

At the same time, the most

economical hypothesis may be that the Iliad and the Odyssey already existed and were known on the Greek mainland by Hesiod’s time in a form recognisably the same as that in which we know them today: certainly the contrary has yet to be proved, and as far as concerns the suggested dependence upon Hesiod of major parts of the Odyssey, the onus rests upon those who believe in Hesiod's priority to put forward much stronger arguments than have so far been forthcoming.© Nor has it been shown that Homer has any features which presuppose a more advanced stage of linguistic development than we find in the Hesiodic poems : in fact the opposite would seem to be true. In any case it may reasonably be thought that only an exceptional flowering of the Ionian epic, such as Homer represents, can account for the pre-eminence of Ionic as the language of hexameter poetry in Hesiod's time, and we may well believe ‘ that it was the prestige of the heroic epos which led Hesiod to adapt the metre and diction of Homer to his own matter. 9! On this view Hesiod and Homer may be said to belong to one and the same tradition, and both the linguistic features in which we have compared the two poets and the further argument from dialect point to Hesiod's late place, relative to Homer, in that tradition.

Since this account of Hesiod's chronological relationship to Homer contradicts the opinion recently expressed by West in the Prolegomena to his Theogony, it is desirable to indicate briefly the reasons for rejecting his view.9? The core of West's argument is as follows : Hesiod is certainly to be dated between 750 and 650 s.c., since an earlier date is excluded by geographical and other references, and a later one by imitations in Archilochus and other lyric poets. A more precise date within this period is suggested by Hesiod’s remark (Op. 654 ff) 59 See above, p. 9. 80 See Ch. XI, esp. pp. 187-9. 61 J, A. K. Thomson in CH (ed. Wace and Stubbings), 4. 88 See esp. HT, 46 f. As far as the subsidiary argument from the opinions of the ancients is concerned, see above, pp. 8 f. with n. 44. 68 HT, 40-3.

THE

204

LANGUAGE

OF

HESIOD

that he took part in the funeral games of Amphidamas at Chalkis, which may even have been the occasion for the composition of the Theogony; there is some evidence to associate the death of Amphidamas with the Lelantine War, and ‘ as it was a cavalry war, it presumably took place before the

encounter

Euboea

in

recalled

(by

by

implication)

Archilochus (fr. 3.4-5), in which slings and bows are no longer From this West arrives at a rough dating of used'.94 730-700 s.c. for the T'heogony. There are of course many points at which this reasoning may be questioned, but the most crucial one concerns the Lelantine War : there can be no certainty about its date or duration, and even if Hesiod’s Amphidamas is the same as the one said to have met his death in it, we have no idea to which part of the war he belongs.*° As for West’s belief that the Ikad and Odyssey ‘ both admit elements that archaeology shows to be not older than ο. 700’, one must stress that the archaeological evidence does not justify such a categorical statement with regard to any of the 94 HT,

43;

this matter of cavalry receives further emphasis

at HT,

45,

where it is related to Th. 439; but iaajeco in that line as transmitted would appear to mean not men at war but charioteers in a race (cf. V 262-87, Sc. 305 f.), the idea being characteristically tacked on to that of contests in Th. 435-8. West’s transposition of Th. 439 is unnecessary and spoils this sequence of thought. From the linguistic point of view it may be added that δ᾽ αὖτε needs no precedent of parallel structure to justify it: ef. Op. 814. $5 The dates which have been proposed for the war range from the late eighth to the early sixth centuries B.C. : see D. W. Bradeen, TAPA lxxviii, 1947, 223 n.

1; A. W.

Gomme,

A Historical Commentary

on Thucydides, I,

1956, 126. Beloch’s very late dating (c. 570 2.0.) is still maintained by Ernst Meyer in Der kleine Pauly ITI, 1969, s.v. * Lelantion Pedion ’ ; but most scholars would now reject this, and the period around the end of the eighth century is generally favoured : see esp. H. Bengtson, Griechische Geschichte,

1950, 85 n. 6;

W.

G. Forrest, Historia vi, 1957,

160-2;

J. N.

Coldstream, Greek Geometric Pottery, 1968, 369 f. But there seems no reason to place the outbreak of the war earlier than c. 710, and we cannot assume that all hostilities ceased as early as 700 n.c. (occupation of both Chalkis and Eretria continued without interruption into the seventh century : see Coldstream, op. cit., 368). And even granting the identification of Amphidamas (though on this see Kirk, SoH, 283), it might well have been some piece of epic poetry, and not the T'heogony, which Hesiod recited at Chalkis : see Walcot, REG Ixxiii, 1960, 36-9.

CONCLUSIONS

205

elements which he adduces in its support. In any case, isolated references are not sufficient for dating the monumental composition of the Homerie poems : their text was probably open to rhapsodic elaboration for a considerable period. It is not unreasonable to suppose that by 700 s.c. the Iliad and Odyssey existed in substantially their present form, but without certain portions thought to be later than the main composition,9? just as West believes that an original Theogony once existed without the last section of our present text and a few other short passages. Miss Gray summed up the position well when she wrote that “nothing which can be shown to be postgeometric is organically incorporated in the poems. ... It seems established that the epic tradition admitted innovations freely until some date not far removed from 700 s.c. and then ceased to do so. 58 Thus both the historical evidence for 66. West’s unhappily dogmatic footnote (HT, 46 n. 2) will persuade no one familiar with the discussions of these elements which have been published since Miss Lorimer’s book of 1950. (a) On hoplite tactics see T. B. L. Webster, From Mycenae to Homer, 1958, 214 (‘ Hoplite equipment and tactics need no longer be dated as late as 700’); G. S. Kirk, ‘ Objective dating criteria in Homer',

Museum

Helvelicum

xvii,

1960,

194 ; id., SoH,

186-8,

where

he

discusses the Homeric passages referred to by West (‘ they may still belong to the 8th century ’); A. Snodgrass, to whom West refers, is in fact very cautious about the interpretation of the Homeric passages (see esp. Early Greek Armour and Weapons, 1964, 179) : heis in any case using the literature to try to throw light on the archaeological evidence and not vice versa. On the possibility of an eighth century date for (5) Agamemnon’s shield with the

Gorgon

blazon

see Webster,

MtH,

213;

Kirk, ODCH,

196, SoH,

186.

On the absence of any convincing archaeological parallel for (c) Odysseus’ clasp, see P. Jacobsthal, Greek Pins, 1956, 141 ; Kirk, ODCH,

196, SoH, 185.

(d) On Theano's supplieation Miss Lorimer was following the arguments of Bethe in aseribing this passage to an Athenian source of the mid sixth century; but there is no need to place the reference so late as this on archaeological

grounds:

see Webster,

MtH,

212, Kirk,

ODCH,

196, SoH,

186; the new Bronze Age evidence from Keos has radically altered the picture here: see D. H. F. Gray in Fifty Years (and Twelve), 1968, 46 (new Appendix) beside her earlier discussion, ibid., 29. 67 Thus even Miss Lorimer, who favoured a seventh century date for the elements discussed in the last note, nevertheless held that the composition of the Odyssey should not be brought lower than 700 n.c.: see HM, 508 ff. and esp. 515 n. 2. 88 Fifty Years (and Twelve), 30.

206

THE LANGUAGE

OF HESIOD

dating Hesiod and the archaeological for Homer would seem to be inadequate as a basis for West’s conclusion 9? that Hesiod is the earlier. The fact is that the instruments available to us here, while they may be of help in finding the right century, are too blunt for the more delicate operation of determining the precise decade. It makes much better sense, in the light of the relationship between the language of the poets, to place Homer somewhere in the late eighth century and Hesiod a little later, perhaps in the early part of the seventh. It is to be expected that future work will lead us to modify these conclusions in some places, and to add fresh details. While the foregoing chapters have ranged over various aspects of Hesiod's language in seeking to clarify his relationship to the poetic tradition in which he stands, there remain a good many others which we have not touched. It seems appropriate in these last pages to indicate some of the main directions in which the study of this subject might be pursued further. First to be mentioned is the need to examine the language of Greek hexameter poetry over as wide a range of material as possible. The history of the formulaic style did not begin with Homer and it does not end with him. We saw in Chapters III to VI how this style of poetry, so far from remaining static, continued to be developed by the creation of new phrases often involving extensions by analogy in unexpected directions which have little or nothing to do with the ideas which the words contain. A full study of this matter is precluded by the complete absence of spoken texts, but despite this handicap an outline history of the formulaic style might emerge from a careful examination of the language not only of Homer and Hesiod but also of the Homeric Hymns, fragments of the Cyclic epics, Orphica, and other relevant material In the light of such a diachronie study it might be possible, for $? T am sorry that Dr. West soems to have come under fire quently in this chapter. Nothing said here should be allowed to fact that in HT he has given us a superb edition, and this book's is obvious from the numbor of times his work has been referred

rather freobscure the debt to him to.

CONCLUSIONS

201

example, to formulate the ‘law of economy ' in such a way as to take account of mutually replaceable expressions like those discussed in Chapter V. Secondly, more study is needed of the processes involved in what we may call, for want of a better term, ‘ imitation ’.

One of the most instructive things to emerge from the discussion in Chapter XI is the very fact that such contrasting views may be taken of the same set of parallel phrases. Even though a re-examination of the material left us unable to accept the theory that these phrases must be original in Hesiod and secondary in the Odyssey, the fact remains that they are remarkably well integrated with their Hesiodic contexts, and in several places the poet’s thought seems to have been moulded by the language which he is using rather than the reverse.”° The process is thus very similar to that which we saw at work in the creation of some of the more striking ‘parallels of sound ’ in Chapter VL?! Such a mode of composition makes it difficult, if not impossible, to identify the original home of any particular phrase ; the most we can hope for is to be able to pick out places which are almost certainly not that original home on account of something ‘awkward or inorganic '."? More often than not there is nothing which betrays that a phrase is being used at second hand, but it would be well worth studying various passages of early hexameter verse to ascertain how frequently there is present some positive evidence of a secondary usage. Thirdly, as far as Hesiod himself is concerned, a descriptive

grammar of the poems would be useful, and might well bring to light some interesting features in his language which we have not here examined. A description of Hesiod's syntax and sentence-structure, which could then be compared with Homer's, might also confirm or modify the conclusions which we have drawn about the poets' relationship from other aspects of their language. A reliable text is essential to serve as a basis 70 e.g. at Τῇ. 277 f. (seo pp. 167 f.), Op. 317 (pp. 170 £.), Op. 508 (pp. 174 f.).

"1 e.g. at Th. 319 (p. 77), Op. 214-6 (p. 78), Op. 441 (p. 74). 72 See p. 188.

208

THE LANGUAGE

OF HESIOD

for such a piece of work, particularly for those parts of it which involve statistical counts. We have seen how Rzach’s text has certain disadvantages in this respect, but the appearance of Solmsen's new Oxford Text might well provide a suitable opportunity for a descriptive account of the language of the poems. Finally, there is need for a more thoughtful use of statistics. One approach which should be tried is the counting of certain features, such as those studied in Chapter IX, within smaller units, breaking down megaliths like ‘ Homer’, ‘ Jiad’, and

‘ Odyssey’ into single books or even into passages of 50 or 100 lines. A basic difficulty which often faces a worker in this field 1s that of assessing the significance of what he finds in, for example, one of the Hesiodic poems or one of the Homeric Hymns. It is often easy enough to obtain a comparable figure for the occurrence of a given feature in the whole of the Iliad or the Odyssey, but what 15 needed 1s not only the total or average figure but also a series of figures which will show how widely it is possible for a passage of a certain length to vary from the over-all average. To know this range of variation is often as important as knowing the mean or average for the whole poem. Many features simply involve mechanical counting, and this sort of work is already being helped by the computer.

The main reason why there is so much contradiction in the statements which are made by scholars about Hesiod les not so much m any inherent obscurity of the material available for study but rather in human fondness for drawing too large an inference from too small a body of fact. We make no claim to having avoided this pitfall here, and time will doubtless show where our own conclusions need to be rejected or qualified. But there is reason to hope that generally agreed solutions will eventually be reached to some at least of the problems which we have been discussing. For in philological scholarship as elsewhere, the Muses may often speak ‘ falsehoods that wear the guise of truth’, but they know also how to tell the truth itself, ‘ when they so choose '.

INDEXES

BIBLIOGRAPHY AND

INDEX

OF

WORKS

CITED

This list includes only work to which reference has been made in the text. Abbreviations for periodicals follow the system of the Linguistic Bibliography compiled annually by the Permanent International Committee of Linguists (CIPL). Titles of periodicals on non-linguistic subjects, and of a few journals which are no longer current, have been abbreviated in the footnotes but are given here in fuller form. Special abbreviations which have been used for books and articles are given in the left hand margin. AHRENS, H. L. De Graecae linguae dialectis. Göttingen, 1839-43 : 17, 18. ALLEN, T. W. ' The date of Hesiod ', JHS xxxv, 1915,

85-06 : and 1904; 1936 :

7. SIKES, E. E. The Homeric Hymns. London, 2nd ed. (with W. R. Halliday), Oxford, 18.

see also Monro, D. B. ALLEN, W. S. Vox Graeca. Cambridge, 1968 : 135, 163.

Barrett, W. S. Euripides. Hippolytos, edited with Introduction and Commentary. Oxford, 1964 : 117. BAUDER, W. De generis neutrius pluralis cum verbo construendi vi et usu, praecipue apud. Homerum et Hesiodum. Leipzig, 1877 : 18. Betocu,

K.

J.

Griechische

Geschichte

I.

2nd

ed.,

Berlin, 1924-26 : 6, (204). BrwersoN, H. Griechische Geschichte (Handbuch der Altertumswissenschaft III.4). Munich, 1950 : 204. BrrnHg, E. Homer. Dichtung und Sage I-III. Leipzig, 1914-27 ; II, 2nd ed., 1929 : 166, (205). BLUMENTHAL, A. VON

Tyrtaios', 154, 161.

Hermes

‘ Prosodisches bei Hesiod und

lxxvii,

1942,

193

f.:

147-9,

BIBLIOGRAPHY

DEG

HP

CH

AND INDEX

OF WORKS

CITED

Bo1sacq, E. Dietionnavre étymologique grecque. 3rd ed., Paris, 1938 : 59. BorurNa, G. M. The Participle in Hesiod Hopkins Univ.). Washington, 1897 : Bowra, C. M. Tradition and Design

211

de la langue (Diss., Johns 18, 200. $n the Iliad.

Oxford, 1930 : 7. —— Ancient Greek Literature. London, 1933 : 6. —— Heroic Poetry. London, 1952 : 111, 192.

Homer and his Forerunners (Andrew Lang Lecture at the Univ. of St. Andrews). Edinburgh, 1955 : 191. Ch.I * Metre ' and Ch. II ‘ Style ' m A Companion to Homer (ed. Wace and Stubbings), 19-37. London, 1962 : 84, 85. BRADEEN, D. W.

‘ The Lelantine War and Pheidon of

Argos’, TAPA Ixxviu, 1947, 223-41 : 204. Brown, N. O. Hesiod’s Theogony, translated with an Introduction.

BRUGMANN, K.

New York, 1953 : 5, 172, 175.

‘ Homerisch ἐννῆμαρ, ἐννήκοντα und

hesiodisch &vvderes ’,

GD

IF xx, 1906-7, 225-8 : 18.

Buck, C. D. The Greek Dialects. Chicago, 1955 : VIII 105-16 passim, 120, 132, 142, 163. Burcuer, 8. H. and Lane, A. The Odyssey of Homer, done into English prose. London, 1879 : 177 f., 182. Elementi primitivi nella poesia CANTARELLA, R. esiodea ^, Rivista indo-greca-italica di filologia etc. xv, 1931, 105-49 : 20.

FFS

CARPENTER, Homeric Berkeley, CHADWICK,

R. Folk Tale, Fiction, and Saga 1n the Epics (Sather Classical Lectures xx). 1946 : 3, 44 f., 192. ‘Mycenaean: a newly discovered J.

Greek dialect’, CHANTRAINE, P.

GH

T PAS 1954, 1-17 : 139. Review of Sellschopp,

SUH

in

RPh x, 1936, 351 : 168. Ch. IV * La langue de l'Iiade ’ in Introduction à l'Iliade (ed. Mazon) 89-123. Paris, 1942 : 122, 159. — — Grammaire homérique. I, Phonétique et morpho-

212

BIBLIOGRAPHY

AND INDEX

OF WORKS

CITED

logie, Paris, 1942; II, Syntaxe, 1953: 1; on Hesiodic forms, 104, 110, 113, 115; on Homeric sounds and forms, 38, 84, VIII 102-18 passim, IX 122-39 passim, 150, 153 ; syntax and vocabulary, 114, 168, 184 ; versification, 92, 138, 148, 150, 153.

CLEMM, W. “Kritische Beiträge zur Lehre vom Digamma im Anschluss an dessen Wiederherstellungsversuch bei Hesiod’, Curtius Studien ix, 1876, 407-57 : 17, 18. COLDSTREAM, J. N. Greek Geometric Pottery. London, 1968 : 204. CoLEMAN, R. ‘The dialect geography of ancient Greece ', TPAS 1963, 58-126 : 194. CoLoNNA, A. Hesiodi Opera et Dies. Milan, 1959: 81, 114. Coox, A. B. ' Descriptive animal names in Greece’, OR viii, 1894, 381-5 : 18, 112 bis. Cook, R. M. ' The date of the Hesiodic Shield ’, CQ xxxi, 1937, 204-14 : 8. CnorsET, M. Histoire de la littérature grecque I. Homére — la poésie cyclique —Hésiode. 4th ed., Paris, 1928 : 6, 7.

LHD

Cuntuirre, R. J. A Lexicon of the Homeric Dialect. London, 1924 : 35, 81, 182. — — Homeric Proper and Place Names. London, 1931: 65. Datz, A. M. ' Greek metric 1936-1957 ’, Lustrum ii, 1957, 5-51: 85.

Davies, A. M.

* * Doric” features in the language of

Hesiod ’, Glotta xli, 1964, 138-65 : 22, VIII 101-11 passim, 116, 120. X 141-54 passim, esp. 151-4,

SADH

161, 194. DENNisTON, J. D. The Greek Particles. 2nd ed., Oxford, 1953 : 177. DzvaNTIER, F. Die Spuren des anlautenden Digamma bei Hesiod.

I, Jever,

1878 ; II, III, Eutin,

1897 : 18, 132 bis, 133, 134, 138.

1894,

BIBLIOGRAPHY

AND

INDEX

Diets, H. and Kranz, W. sokratiker I-III.

OF WORKS

CITED

213

Die Fragmente der Vor-

5th ed., Berlin, 1934-37 : 117.

Dopps, E. R. Ch. I. ii. ‘ Homer as oral poetry’ in Fifty Y ears (and Twelve) of Classical Scholarship (ed. Platnauer) 13-17, 38-42. Oxford, 1968 : 1. Dover, K. J. Ch. V. ‘ The poetry of Archilochus ’ in Archiloque (Fondation Hardt Entretiens x, by J. Pouilloux et al.), 181-222.

Dunzar, H.

LH

Geneva, 1964 : 135.

A Complete Concordance to the Odyssey of

Homer. Oxford, 1880. Rev. ed. by B. Marzullo, Hildesheim, 1962 : 22. EsELING, H. Lexicon Homericum. Leipzig, 1880-85 ;

35, 80, 152, 182. Epmonps, J. M. Lyra Graeca I-III (Loeb Classical Library) London, 1922-27 : 117. Epwarps, M. W. ‘Some features of Homeric craftsmanship ’, TAPA xcvii, 1966, 115-79: 191. EnnLicH, H. ^| Die epische Zerdehnung’, RAM xiii,

1908, 107-126 : 18. EncrcropaEpıa Britannica, 11th ed., Cambridge, 1910; 16th ed., London, 1956: 113.

DEL

ERNOUT, A. and MEILLET, A. Dictionnaire étymologique de la langue latine. 3rd ed., Paris, 1951 : 106.

Loeb

EVELYN-WHITE,

ed.

and Homerica (Loeb Classical Library). London, 1914: 7, 16, 81, 172, 182. Fick, A. Die homerische Odyssee in der ursprünglichen Sprachform wiederhergestellt. Göttingen, 1883 : 17. —— Die homerische Ilias nach ihrer Entstehung

HG

——— Hestods Gedichte in ihrer ursprünglichen Fassung und Sprachform wiederhergestellt. Göttingen,

H. G.

Hesiod,

the Homeric Hymns

betrachtet, etc. Göttingen, 1886 : 17.

1887 : 17, 102, 201

See also next two entries.

* Die ursprüngliche Sprachform und Fassung der hesiodischen Theogonie ’, Beiträge zur Kunde der indogermanischen Sprachen (= BKIS) xu, 1887, 1-37 : (102, 201).

214

BIBLIOGRAPHY

AND

INDEX

OF

WORKS

CITED

Fıck, A. ‘ Zur Sprachform und Fassung der griechischen Epen ’, BAIS xvi, 1890, 1-28 (mostly on Sc.) : (102, 201). —— Die griechischen Personennamen. 2nd ed. with F. Bechtel, Góttingen, 1894 : 18.

—— ‘Die griechischen Verbandnamen (ethnika) ’, BKIS xxvi, 1901, 233-65 : 18. Frgrkav, H. De carminum Hesiodeorum atque hymnorum | quatuor magnorum | vocabulis non Homericis.

DDH

Königsberg, 1866 : 17, 134.

Fıntev, M. I. The World of Odysseus. London, 1956 : 9. FLACELIERE, R. Histoire lttérawe de la Gréce. Paris, 1962 : 7. FrAcH, H. Das dialektische Digamma des Hesiodos. Berlin, 1876 : 17 f., 132. See also next entry. ‘Die neuesten Arbeiten über das Digamma bei Hesiodos',

Jahrbücher

für

klass.

Philologie

cxüi,

1876, 369-75 : (18). see also GOETTLING, C. FÖRSTEMANN, J. De dialecto Hesiodea. 17, 114.

FÖRSTER, R. Hesiodeo.

GEW LLO

Halle, 1863 :

De usu coniunctionis πρίν Homerico et Breslau, 1863 : 17.

Forrest, W. G. ‘ Colonisation and the rise of Delphi ’, Historia vi, 1957, 160—75 : 204. Frisk, H. Griechisches etymologisches Wörterbuch. I, A-Ko Heidelberg, 1960; IIin progress: 59, 110, 168. GALLAVOTTI, C. and RONcoNI, A. La lingua omerica. 3rd ed., Bari, 1955 : 101, 102, 106, 108, 110, 114, 120.

Garcia Catvo, A. “Particularidades lingüísticas recuperables a través del texto hesiödico ’, Em xxxiv, 1966, 15-37 : 13, 138. GOETTLING, C. Hesiodi carmina.

Gotha,

1831;

2nd ed., 1843 ; 3rd ed. by H. Flach, Leipzig, 1878 : 14; text, 87, 101, 115; commentary, 115, 182 ; introduction, 13, 17.

101,

103,

BIBLIOGRAPHY

AND

INDEX

OF

WORKS

CITED

215

Gomme, A. W. A Historical Commentary Thucydides. Vol. I, Oxford, 1956 : 204.

on

Gow, A. S. F. ‘ Miscellaneous notes on the Works and Days ^, CQ xi, 1917, 113-18 : 113. Gray, D. H. F. ‘ Homeric epithets for things ’, CQ xli,

1947, 109—21 : 46-8, 51, 60. — — Ch. I. v ‘ Homer and the archaeologists ’ in Fıfty Years (and Twelve) of Classical Scholarship (ed. Platnauer), 24-31, 46-9.

FHF

GSÁ

Oxford, 1968 : 205 bis.

GRONINGEN, DB. A. van La composition littéraire archaique grecque. Amsterdam, 1958: 193. HAINsworTH, J. D. ‘Structure and content in epic formulae : the question of the unique expression ’, CQ n.s. xiv, 1964, 155-64 : 48. -- The Flexibility of the Homeric Formula. Oxford, 1968 : 28, 42 f., 55, 60, 93. —— ‘ Greek views of Greek dialectology ’, T PhS 1967, 62-76 : 120. Hamm, E.-M. Grammatik zu Sappho und Alkavos. 2nd ed., Berlin, 1958 : 107, 115. Harte, W. ‘Homerische Studien

Ixxvii, 1874, 7-88 : 139. Harvey, P. The Oxford

III’,

Compamion

SHOAW

to Classical

Literature. Oxford, 1937 : 120. Hecut M. 'IVoAa bei Hesiod ’, Philologus xlv, 1886,

380 f. : 18. —— Die griechische Bedeutungslehre. 18. Herrmann,

L.

Catulle

et Lucréce’,

Leipzig, 1888 : Latomus

xv,

1956, 465-80 : 189. HIERSCHE, R. ‘ Zu Sappho 2, 9D κὰμ μὲν γλῶσσα éàye “Die Zunge ist gebrochen "', Glotta xliv, 1966, 1-5: 113.

Hicuam, T. F. * Nature note : autophagy in octopods. Hesiod vindicated ’, CR n.s. vii, 1957, 16 f. : 112.

Hinricus, Aeolicis.

G.

De

Homericae

Berlin, 1875 : 17.

elocutionis

vestigiis

BIBLIOGRAPHY

216

Horxstra, A.

OF

INDEX

AND

WORKS

CITED

‘ Hésiode et la tradition orale ', Mn x,

1957, 193-225 : 21, 101, 131, 136, 140, 152, 154. HM F P—— Homeric Modifications of Formulaic Prototypes. Amsterdam, 1965: 21, 36, 202 ; on formulas, 48, 90, 135, 148, 190; on Hesiod, 21, 135, 140, 202 ; on Homeric language, 90, 138, 148.

Horrmann, S. F. W. Bibliographisches Lexicon der gesammten Literatur der Griechen I-III. 2nd ed., Leipzig, 1838-45 : 13. HupsoN-WiLLiaAMS, T. Early Greek Elegy. Cardiff, 1926 : 150. Intex, F. ‘Der Dual bei Hesiod’, Zeitschrift f.d.öst. Gymnasien xxxix, 1888, 97-102 : 18.

——

Über den Gebrauch der Präpositionen bei Hesiod.

2 pts., Brünn, 1888-9:

18.

Zur Syntax des Hesiod. Brünn, 1890 : 18. Ister, M. Quaestionum | Hesiodearum specimen. Berlin, 1830 : 17. JACOBSTHAL, P. Greek Pins. Oxford, 1956 : 205. JacoBy, F. Das Marmor Parıum. Berlin, 1904: 7. ——— Hesiodi Theogonia. Berlin, 1930: 15 £., 116; text and transmission, ll, 12 ter, 16, 145 ; lines of suspect authorship, 4, 116, 180, 184, 186.

JAEGER, W. Paideia : the Ideals of Greek Culture I-III (translated by G. Highet). Oxford, 1939-45 : 6. KauscH, E. Quatenus Hesvodi in Theogonia clocutio ab

exemplo Homeri pendeat. Königsberg, 1876 : 17. —— Quatenus Hesiodi clocutio αὖ exemplo. Homeri pendeat. Elbing, 1878 : 18. KERSCHENSTEINER, J. Review of Munding (q.v. below) in Gnomon xxxiv, 1962, 1—7 : 166. KriNKEL, G. see KoECHLY, A. KiNzEL, J. ‘Die Kopula bei Homer und Hesiod ’, Jahresberichte des Staatsgymn. in Máhrisch-Ostrau 1907-8 and 1908-9 : 18. Kincuuorr,

A.

Hesiodos’

Berlin, 1889 : 81.

Mahnlieder

an

Perses.

BIBLIOGRAPHY

ODCH

Kirx, G. 8. MH

SAT

SoH

AND

INDEX

OF WORKS

CITED

217

'Objective dating criteria in Homer’,

xvii, 1960, 189-205 : 200, 205 quater.

—— Ch. II ‘The structure and aim of the Theogony ' in Hesiode et son influence (Fondation Hardt Entretiens vii, by K. von Fritz et al.), 63-95. Geneva, 1962 | — — The Songs 202, 204, 205 — Review of

: 5, 186. of Homer. Cambridge, 1962 : 8, 9, 191, quater. Hoekstra, ΠΗ͂ F P in Gnomon xxxviii,

1966, 7877-40 : 138. ‘The structure of the Homeric hexameter ', Yale Classical Studies xx, 1966, 76-104 : 26, 88.

VSH

| ——

‘Verse-structure

and

sentence-structure

in

Homer ’, ibid. 105-52 : 94-6, 98 f.

‘Formular language and oral quality’, «bud. 153-74 : 190 bis. Knös, O. V. De digammo Homerico quaestiones,

I-III, in Upsala Univ. Ärsskrift 1872, 1873, 1879 : 17, 18. KozcHLy, A. and Kınkeı, G. Hesiodea quae feruntur carmina. Leipzig, 1870 : 14. VUHH Krarrt, F. Vergleichende Untersuchungen zu Homer Gottingen, 1963: 21, 116, 182; on und Hesiod. 170; 91, of formulas, modification Hesiodic bis, 169 Homeric passages compared with Hesiod,

171, 174 f., 178, 179, 181, 182; DIH

lists of parallel

phrases, III 23-39 passim, 78, 170. Krerscumer, P. F. De iteratis Hesiodeis. Breslau, Hesiodic repetitions 3, 19, 167, 189; 1913: discussed, 43, 64 bis, 82 f., 116, 172, 174 ; lists and

tables, 40-5, 197, 199. Kromer, W. De articuli vi atque usu apud. Hesiodum. Miessen, 1841 : 17.

AG

Künner, R. and Brass, F. Ausführliche Grammatik 3rd ed., Hannover, der griechischen Sprache Y. 1890-92 : 101. La Rocug, J. ‘ Untersuchungen über den Vers bei

218

BIBLIOGRAPHY

AND

INDEX

OF

WORKS

CITED

Hesiod und in den homerischen Hymnen ’, WS xx,

1898, 70-90 : 18, (85). Lee, A. G. ‘On [Tibullus] III, 19 (IV, 18)’, Proc. Cambridge

TPG

Philol. Soc. clxxxix,

(207). LEJEUNE, M.

1963,

4-10:

Traité de phonétique grecque.

188,

2nd ed.,

Paris, 1955 : 111 bis, 116, 127, 163. LENDLE, O. Die‘ Pandorasage ’ bei Hesiod. Würzburg,

1957 : 105. LENNEP, D. J. van Hesiodo Theogonia. Amsterdam, 1843 : 14, 143. —— Hesiodi Opera et Dies. Amsterdam, 1847 : 14. —— Heswdi Scutum Herculis (ed. J. G. Hulleman). Amsterdam, 1855 : 14.

Leo, F.

Hesiodea.

Göttingen, 1894; = Ausgewählte

kleine Schriften (ed. E. Fraenkel, Rome,

GGL HGL HW

1960), II,

343-63 : 116-8. Lesky, A. Geschichte der griechischen Literatur. 2nd ed., Bern, 1963 : 4, 5, 7, 8, 150, 166, 191. ——— A History of Greek Literature (translated by J. Willis and C. de Heer). London, 1966 : as for GGL. LEUMANN,

M.

Homerische

Wörter.

Basel,

1950:

116, 152. Lippett, H. G. and Scorr, R. A Greek-English Lexicon. 9th ed. rev. by H. Stuart Jones and R. McKenzie, 185.

Oxford,

1940:

33, 116,

125,

176, 182,

LivADARAS, N. A. 'loropía τῆς παραδόσεως τοῦ κειμένου τοῦ ᾿Πσιόδου. Athens, 1968 : 11. Luoyp-Jonses, H. Review of Barrett, Euripides. Hippolytos 1n J HS Ixxxv, 1965, 164-71 : 70. Lorp,

A. B.

‘Homer,

Parry,

SoT

1948, 34-44 : 19. — — The Singer of Tales. (36), 72, 93.

HM

Lorimer, H. L.

and Huso’,

lii,

Cambridge, Mass., 1960 :

Homer and the Monuments.

1950 : 176, 205 bis.

AJA

London,

BIBLIOGRAPHY

AND INDEX

OF WORKS

CITED

219

LupwicH, (C. H.) A. Quaestionis de hexametris poetarum Graecorum spondaicis capita duo. Halle, 1866 : 17, 87 ter.

—-—— Anistarchs homerische Textkritik I, II. 1884-85 : 86, 87. McKay K. J.

Leipzig,

‘ Ambivalent αἰδώς in Hesiod’, AJ Ph

Ixxxiv, 1963, 17-27 : 171. Macnien, V. ‘Deux questions

homériques..

I.

Génitifs en -o10 et -ov ', Mémoires de la Soc. linguistque de Paris xxii, 1922, 165-71:

122.

ManrLow, G. H. Neue Wege durch die griechische Sprache und Dichtung. Berlin, 1926: 201. Marn, A. W. Hesiod. The Poems and Fragments, done into English prose. Oxford, 1908 : 172, 175, 182. Martin, 8. ' Quatenus Hesiodeae rationis vestigia in carminibus Homericis reperiantur! Programm d. Studvenanstalt Sperer, 1889 : 18.

Bude ed.

MARZULLO,

B.

G. L. Mazon, P.

Hesiode (Bude series).

see DUNBAR,

H.

and

PRENDERGAST,

Paris, 1928:

16;

text, 81 114, 116; translation, 172,182 ; commentary, 6, 8, 11 ; on lines of suspect authorship, 4, 5 bis, 116.

Introduction à VIliade (Budé series). Paris, 1942 : 122, 135. —— Homére. Iliade (Bude series). 4 vols., Paris, 1937-38 : 33. MEILLET, A. ‘ Sur une édition linguistique d'Homére ’, REG xxxi, 1918, 277-314 : 13. see also ERNOUT, A.

HK

Meister, K. Die homerische Kunstsprache. Leipzig, 1921: 1, 110, 118, 135. MERKELBACH, R. Untersuchungen zur Odyssee. Munich, 1951 : 170. and West, M. L. Fragmenta Hesiodea. Oxford, 1967 : 1, 18, 185. see also SOLMSEN, FRIEDRICH.

220

BIBLIOGRAPHY

AND

Merry, W. W.

INDEX

Homer.

duction, notes, etc.

OF

WORKS

CITED

Odyssey, edited with Intro-

2 vols., Oxford, 1870-78 : 182,

184. and RrpDELL, J. Homer’s Odyssey. Books I-X11, edited with notes, ete.

Oxford, 1876:

183.

MERWE SCHOLTZ see SCHOLTZ. Meyer, Ernst Article s.v. ‘ Lelantion pedion’ in Der kleine Pauly. Lexikon der Antike (ed. K. Ziegler and W.

1969 : 204. Minton, W. W.

‘The

formula’, TAPA

GHD

Monro, D. B.

Sontheimer)

III, 551.

fallacy

of the

Stuttgart,

structural

xcvi, 1965, 241-53 : 48.

A Grammar of the Homeric Dialect.

2nd ed., Oxford, 1891 : 148.

——

Homer’s Odyssey.

with notes, etc.

and

ALLEN,

Books XIII-X XIV, edited

Oxford, 1901:

T. W.

126, 182.

Homeri

Opera

(Oxford

Classical Texts). I, II. /kas, 3rd ed., 1920 ; III, IV. Odyssea, 2nd ed., 1917, 1919; V. Hymni etc.,

1912 : 33, 87, 144. Munping, H. Hesiods Erga in ihrem Verhältnis zur Ihas. Frankfurt, 1959 : 166. Myres, J. L. ‘ Hesiod's “ Shield of Herakles’”’ : its structure and workmanship’, JHS lxi, 1941,

17-38 : 8. NEUMANN, C. De primariis optandi, iubendi, vetandi enuntiatis apud. Homerum comparato usu Hesiodeo. Varel, 1883 : 18. NoroPouros, J. A. ‘ Homer, Hesiod and the Achaean

heritage of oral poetry', Hesperia xxix, 1960, 177-97 : 21, 40-2, 44, 201 f. —— ‘ The Homeric Hymns as oral poetry ; a study of the post-Homeric oral tradition ', AJ Ph lxxxiii,

1962, 337-68 : (36).

——

‘Studies in early Greek oral poetry’,

Ixviii, 1964, 1-77 : 192, 202. Ocpen, C. J. De infinitivi finalis

HSPh

vel consecutivi

BIBLIOGRAPHY

constructione

AND

INDEX

OF

WORKS

apud. priscos poetas

CITED

Graecos.

991

New

York, 1909 : 18. O'NEimLL, E. G. ‘ The localization of metrical word-

types in the Greek hexameter. the

Alexandrians’,

Yale

Homer, Hesiod, and

Classical

Studies

vii,

1942, 105-78 : 25, 86 bis, 87, 88, 89. Pack, R. A. The Greek and Latin Literary Texts from Greco-Roman Egypt. 2nd ed., Ann Arbor, 1965 : 12. Pace, D. L. Euripides. Medea, edited with Introduction and Commentary. Oxford, 1938: 117. — — Aleman.

HHI

| ——

The Partheneion.

142.

History and the Homeric Iliad (Sather Classical

Lectures xxxi).

CH

Oxford, 1951:

Berkeley, 1959:

46, 51, 55, 58, 59,

191. Parey, F. A. The Epics of Hesiod. London, 1861 ; 2nd ed., 1883 : 14, 35, 104, 110, 182. Parmer, L. R. Ch. IV ‘ The language of Homer’ in A Companion to Homer (ed. Wace and Stubbings), 75-178. London, 1962 : 122, 135, 139. ——— Ch. I. iv ‘ Homer and the philologists ' in Fıfty Years (and Twelve) of Classical Scholarship (ed. Platnauer), 17-24, 42-6.

Oxford, 1968 : 1.

PARMENTIER, L. Etudes historiques sur la formation des mots dans la langue grecque. 1. Les substantifs et les adjectifs en -es dans la langue d'Homére et d'Hésiode. Liege, 1889 : 18. Parry, A.

‘ Have we Homer’s Iliad ? ^, Yale Classical

Studies xx, 1966, 177-216 : 191. L’epithete traditionnelle dans Homere. Paris,

ETH

Parky,M.

DCE

1998 : 1, 19, 42, (55), 60 f., 63, 68, 155. — — Les formules et la métrique d’Homere. Paris, 1928 : 1, 19, 90. ‘The distinctive character of enjambement in Homeric

verse’,

TAPA

Ix, 1929, 200-20:

1, 19,

93-9. ‘Studies in the epic technique of oral versemaking. I. Homer and Homeric style’, HSP xli,

222

BIBLIOGRAPHY

AND

1930, 73-147 ; ‘II.

INDEX

OF

WORKS

CITED

The Homeric language as the

language of an oral poetry’, HSPh xlüi, 1932, 1-50: 1, 19 f., 42, 44, 47, (55), 84. Parry, M. ‘ The traces of the digamma in Jonic and Lesbian Greek ', Lg x, 1934, 130-44 : 1, 19, 137, 139. see also LORD, A. B.

DRM

Pauzs, A. V. De Stockholm, 1897 : PaursoN, J. Studia further parts) in 1887 : 18, (85), 86 ——

RE

Index Hesiodeus.

digammo Hesiodeo quaestiones. 18, 133. Hesiodea I. De re metrica (no Lunds Univ. Arsskrift xxiii, bis, 87, 90, 91 bis. Lund, 1890:

18, 34, 134.

PauLvy, A. Wissowa, G. and others (eds.) ZealEncyclopädie der klassischen Altertumswissenschaft. Stuttgart, 1894 onwards ; see RZACH,

SCHERLING,

SCHULTZ, WITTE. For Der kleine Pauly see MEYER. PLATNAUER, M. (ed.) Fafty Years of Classical Scholarship. Oxford, 1954; 2nd ed., Fifty Years (and Twelve) etc., 1968 : 1, 205. Porter, H. N. ‘ The early Greek hexameter ', Yale Classical Studies xii, 1951, 3-63 : 88 ter, 89.

PRENDERGAST, G. L. A Complete Concordance to the Iliad of Homer. London, 1875. Rev. ed. by B. Marzullo, Hildesheim, 1962 : 22.

Rıev, E. V.

Homer.

The Odyssey (Penguin Classics).

Harmondsworth, 1946 : 182 RONCONI, A. see GALLAVOTTI, C. Rose, H. J. A Handbook of Greek Literature.

London,

1934 : 6, 7. Russo, C. F. Hesiod? Scutum. Florence, 1950: 16; Scutum text, 91, 116, 145, 165 ; on transmission, 11, 16 ;

on date and authorship, 4, 8 ; explanatory notes,

116, 117, 118.

‘Russo, J. A.

“A closer look at Homeric formulas ’,

TAPA xciv, 1963, 235-47 : 43. ——— * The structural formula in Homeric verse ’, Yale Classical Studies xx, 1966, 217—40 : 48.

BIBLIOGRAPHY

DH

AND INDEX

OF WORKS

CITED

Rzacn, A. ‘ Der Dialekt des Hesiodos ’,

klass.

Philologie,

223

Jahrbücher für

Supplementband

viii,

1876,

353-468 : 14, 17f., 114, 116; on dialect forms in Hesiod, 17, 102, 105, 146, 147; on differences between the poems, 17, 146, 195.

— —

‘Studien

heroischen

zur

Technik

des

Verses’, SBÖAW

nachhomerischen

xcv,

1879,

681-872 :

18, (85). “ Neue Beiträge zur Technik des nachhomerischen

Hexamcters ’, SBÖAW c, 1882, 307-432 : 18 (85). —— 'Hoó8ov rà σῳζόμενα. 86.

Leipzig, 1884:

14, 18,

—— Homeri Iladis carmina. Prague, 1886-87 : 86. —— ‘Zur ältesten Überlieferung der Erga des Hesiodos ' in Symbolae Pragenses 165-94. Vienna, 1893 : 11, 14. —— ‘ Die Sippe des Codex Messanius der hesiodischen Erga ' in Serta Harteliana 209-23. Vienna, 1896 : 11, 14.

——

‘ Die handschriftliche Überlieferung der hesiodi-

schen Theogonie ', WS xix, 1897, 15-70 : 11, 14.

——

‘Neue

handschriftliche

Studien

Erga ’, WS xx, 1898, 91-118:

——

‘Die

handschriftliche

zu

Hesiods

der

pseudo-

11, 14.

Tradition.

hesiodischen Aspis ’, Hermes xxxii, 1898, 591-625 :

11, 14.

——

Hesiodi carmina (editio maior).

Leipzig, 1902 :

14 f., 16 ; on transmission, 11, 12 ; list of Homer? loci similes, 14, 19, 23 f., 27, 29-32, 45, 78 ; for text etc., see next entry. Hesiodi carmina (editio minor). Leipzig, 1902 ;

2nd ed., 1908; 133,

137,

145,

3rd ed., 1913: 156;

variants

15;

text, 16, 127,

and conjectures,

in

Th., 104, 137, 143 bis, in Op., 81, 87, 114, 127, 128, in Sc., 126, in Frag., 185 ; bracketed lines, 116, 179.

—— Article s.v. ‘ Hesiodos in Pauly-Wissowa, VIII, 1912, 1167-1240 : 7, 9, 11, 13, 19.

RE

224

BIBLIOGRAPHY

AND

INDEX

OF

WORKS

CITED

RzacH, A. ‘Bericht über die Publikationen zu Hesiodos für das Jahrzehnt 1909-1918 ', Bursian’s

Jahresbericht über die Fortschritte d. kl. Altertumswissenschaft cxcix, 1924, 1-115 ; 19. Sachs, A. De digammo eiusque usu apud Homerum et Hesiodum. Berlin, 1856 : 17. SAVELSBERG, H. De modorum usu Hesiodeo. 1886 : 18. SCHERER, A. see THUMB, A.

SCHERLING, K. Article s.v. ' Tartaros' Wissowa, RE IV A, 1932, 2440-5 : 83. SCHNEIDER, P.

Aachen,

in Pauly-

De elocutione Hesiodea commentatio I

(no further parts). Berlin, 1871: 17. SCHNEIDEWIND, E. T. De casus locativi vestigiis apud Homerum atque Hesiodum. Halle, 1863 : 17. SCHOLTZ, H. van DER Merwe The Kenning in Anglo-Saxon and Old Norse Poetry. Utrecht, 1927 : 111.

SCHULTZ, W.

Article s.v. ‘ Rätsel ’ in Pauly-Wissowa,

RE I A, 1914, 62-125: 111. SCHWARTZ, J. Pseudo-Hestodeia.

Recherches sur la

composition, la diffusion et la disparition ancienne TdH GG

d'oeuvres attribuées à Hésiode. Leiden, 1960: 1. SCHWENN, F. Die Theogonie des Hesıodos. Heidelberg, 1934 : 4, 9, 83, 166, 176, 180, 184, 186.

SCHWYZER, E. Griechische Grammatik I-III (Handbuch der Altertumswissenschaft II. 1) Munich, 1953-60 : on dialect forms in Hesiod, 15, 101, 106, 110, 113, 114 ; other grammatical points, 117, 122,

135, 139, 141, 151, 152 bis. ScoTT, J. A. A Comparative Study of Hesiod and Pindar (Diss, Johns Hopkins Univ.) Chicago, 1898 : 18. — — ‘The vocative in Homer xxiv, 1903, 192-6: 18.

——

and

Hesiod ', AJPh

‘The comparative ages of Homer and Hesiod ’,

TAPA lxi, 1930, xxvi : 7.

BIBLIOGRAPHY

AND

INDEX

OF

WORKS

CITED

995

SEIP, O. De participis et infinitivi apud Hesiodum usu. Giessen, 1886 : 18.

SUH

SELLSCHOPP, 1. Stilistische Untersuchungen zu Hesiod. Hamburg,

1934:

3, 9, 20 f., 51, 111, XI

166-89

passim. SIKES, E. E. see ALLEN, T. W.

W&D

SINOLAIR, T. A.

Hesiod.

Works and Days.

London,

1932 : 16 ; text, 15, 81, 114 ; on transmission, 11, 12, 16; on passages of suspect authorship, 5, 6 ; explanatory notes, 113, 114, 115, 182.

— — A History of Classical Greek Literature.

London,

1934 : 7.

SNELL,

B. and others

Epos.

Fasc.

1-6,

Lexikon des frühgriechischen

a to ἀπό,

Göttingen,

1955-69 :

22. Snoperass, A. Early Greek Armour and Weapons. Edinburgh, 1964 : 205. SoLMsEN, FELIX ‘Zur Lehre vom Digamma’, KZ xxxii, 1893, 273-88 : 18.

——- ‘ Eigennamen als Zeugen der Stammesmischung in Böotien ', FAM lix, 1904, 481-505 : 18.

‘Die metrischen

Wirkungen

anlautender con-

sonantengruppen bei Homer und Hesiod ', RAM lx,

1905, 492-504 : 18. H&A

SOLMSEN, Friepricu Hesiod and Aeschylus. N.Y., 1949 : 9, 166.

——

Ithaca,

‘ The “ gift ” of speech in Homer and Hesiod ’,

TAPA

lxxxv, 1954, 1-15:

168 f.

‘The “ Days ” of the Works and Days’, TAPA xciv, 1963, 293—320 : 5. — — Review of West, HT in Gnomon xl, 1968, 321-9 :

192. —

—,

MERKELBACH,

R.

and

West,

M.

L.

Hesiod

(Oxford Classical Texts). Forthcoming, 1970 : 208. STANFORD, W. B. The Odyssey of Homer, edited with Introduction

etc.

2 vols.

London,

1947-48:

on

digamma, 135 ; on passages of suspect authorship,

220

BIBLIOGRAPHY

126,

170,

AND

187;

INDEX

OF

WORKS

explanatory

CITED

notes,

168,

176

bis,

182, 183. STEINACHER, J. Die Syntax des hesiodeischen Infinitivs mit stetem vergleichenden Rückblick auf Homer. Vienna, 1885:

18.

Sronz, F. Die zusammengesetzten Nomina in den homerischen und hesiodischen Gedichten. Klagenfurt, 1874 : 17. Strunk, K. ‘Sprachliches und Prosodisches zur mykenischen Orthographie ’, I F Ixvi, 1961, 155-70 : 111. SruBBINGS, Εἰ, H.

see Wack, A. J. B.

PGL

SruRTEVANT, E. H. The Pronunciation of Greek and Latin. 2nd ed., Philadelphia, 1940 : 135.

CH

'THoMsoN, J. A. K.

Companion 1-15.

HGD

‘ Homer and his influence’ in A

to Homer

(ed. Wace

and

Stubbings),

London, 1962 : 203.

ΤΉΜΒ, A.

Handbuch der griechischen Dialekte.

ed., Vol. I rev. by E. Kieckers, Heidelberg,

2nd 1932 ;

Vol. II rev. by A. Scherer, 1959 : on dialect forms in Hesiod, 15, 101, (114), 118; on Boeotian, 105, 131, 132, 194 ; other dialects, 110, 115, 142, 164.

SWH

TREUTLER, J. Homerica Jovis epitheta comparantur cum Hesiodeis. Liegnitz, 1867 : 17. TROXLER, H. Sprache und Wortschatz Hesiods. Zürich, 1964: 22; on textual variants and conjectures, 35, 108, 115, 164 ; on language, 66, VIII 101—20 passim, IX 122, 123, 128, 133-40 passim,

TCO

X 142, 146, 148, 152. VAN DER MERWE SCHOLTZ see SCHOLTZ. VAN DER VaLK, M. (H. A. L. H.) Textual Criticism of the Odyssey.

——

Leiden, 1949 : 168, 170, 176, 183.

* Le Boucher du pseudo-Hésiode',

1966, 450-81: 4, 197.

VERDENIUS,

CH

W.

J.

‘Hesiod

Theogony

REG Ixxıx, 35’, Mn

1958, 20-4 : 6. Wace, A. J. B. and Stusgings, F. H. (eds.)

xi,

A Com-

BIBLIOGRAPHY

AND

panion to Homer. 139, 208. Wave-Gery,

H.

INDEX

OF

WORKS

CITED

997

London, 1962 : 84, 85, 122, 135, T.

‘Hesiod’,

Phoenix

ii,

1949,

81-03 : 8. —— The Poet of the Iliad. Cambridge, 1952 : 8, 9. Warcor, P. ‘ The text of Hesiod’s Theogony and the Hittite Epic of Kumarbi’, CQ n.s. vi, 1956, 198-206 : 5. —— * Allusion in Hesiod’, REG lxxii, 1960, 36-9 :

204. —— ‘The composition of the Works and Days’, REG lxxiv, 1961, 1-19: 6. Wesster, T. B. L. ‘Early and late in Homeric diction ’, Eranos liv, 1956, 34-48 : 38.

MtH

—— From Mycenae to Homer. London, 1958 : 205 ter. —— ‘Notes on the writing of early Greek poetry ’, Glotto xxxii, 1960, 251-63 : 13.

WESSELY, C. Studien zur Palaeographie und Papyruskunde 1, 11. Leipzig, 1901-2 : 2. West, M.L. ‘ Hesiodea ’, CQ n.s. xi, 1961, 130-45 : 11. ‘More notes on the text of Hesiod ’, CQ n.s. xii, 1962, 177-81: 11, 104.

——

Review of Hésiode et son influence (see above

under Kırk, G. 8.) in Gnomon xxxv, 1963, 10-15 : 200.

‘The medieval and renaissance manuscripts of Hesiod's Theogony ’, CQ n.s. xiv, 1964, 165-89 : 11. Review of Krafft, VUHH in CR n.s. xv, 1965,

158 f. : 200. —— Review of Livadaras

(q.v. above) in Gnomon

xxxvii, 1965, 650-5 : 11.

HT

Hesiod. Theogony edited with Prolegomena and Oxford, 1966 : Commentary. 84, VIII 103-6 passim, 114, 137, X 143 bis, 145, 156, 164, mission, 11, 12 bis ; on lines of 5, 7, 98, 117 bis, 179, 186, 198, written composition, 3, 192;

15, 186, 206 ; text, IX 124, 127, 180, XI 175; on transsuspect authorship, 199 bis ; on oral or on date of Hesiod,

228

BIBLIOGRAPHY

AND

INDEX

OF

WORKS

CITED

8, 9, 200 f., 203-6 ; language, 22, 92, VIII 101-21 passim,

IX

133,

135,

139, X

147, XI

168 bis, 186,

XII 201 ; versification, 85, 86 bis, 117. West, M. L. see also MERKELBACH, R. and SOLMSEN, FRIEDRICH. WmurirMAN, C. H. Homer and the Heroic Tradition. Cambridge, Mass., 1958 : 9. WizsNER, J. ‘ Die Hochzeit des Polypus ’, Jahrbuch d. Deutschen Arch. Inst. Ixxiv, 1959, 35-51: 112. WILAMOWITZ-MOELLENDORFF, U. von Vitae Homeri et Hesiodi. Berlin, 1915 : 9. - Die Ilias und Homer. Berlin, 1916: 168.

HdO Erga

——

Die

Heimkehr

des

Odysseus.

Berlin,

1927:

170 £., 178. —— Hesiodos Erga. Berlin, 1928: 15; text, 35, 110, 114; on metre, 15, 79; dialect, 103, 110, 114 ; explanatory notes, 5, 15, 171, 175, 178.

WiLCKEN, U. ‘Die Achmim-Papyri in der Bibliothéque Nationale zu Paris’, Sb DAV 1887, 807-20: 2. WINBOLT, S. E. Latin Hexameter Verse. London, 1903 : 89. Wirre, K. ‘Zur homerischen Sprache’, I, II Glotta i, 1909, 132-45; III, IV Glotta ii, 1910, 8-22;

V-XI

Glotta

m,

1912,

105-53;

XII

bid.

388-93 ; and further articles on Homeric language

in Glotta iv, 1913, 1-21, 209-42, and v, 1914, 8-57:

——

1.

Article

s.v.

'Homeros:

Sprache’

in Pauly-

Wissowa, RE VIII, 1913, 2213-47 : 1, 85. Wyatt, W. F. ‘Short accusative plurals in Greek ’, TAPA xcvu, 1966, 617-43 : 130, X 141-65 passim,

esp. 149-51. Young,

D.

‘Never

blotted a line?

premeditation in Homer 1967, 279-324 : 190, 191.

Formula

and Hesiod',

Arion

and

vi,

INDEX

OF

HESIODIC

HOMERIC

LINE ] 3 5 7-8 10

PRINCIPAL

REFERENCES

Theogony Th.1-100

AND

LINE

PAGE

62 64

21 n., 27 n.

65

126, 128 2'" n. 12n. 27n., 145 2'" n.

69 10

119

74

71 72

11

28η.,

12

28n.

15

90 n., 101, 103, 118

19

27n. bis, 102, 104

21

76

84

6

87

]12 n., 119 n., 192 n.

91

201n. 1, 207 n., 208 12 n., 167 f.

92 93 94

22-5

22

23 27-8 27

76

79 84-92

98 100 Th, 101-50

28 29 30

12n., 53 27 n. bis 27n., 75f., 103, 121 n.

31

2" n.

101-10

32

12n.

101

33

119

102

34 36 37

12n. 128 61f.,119

103 105 106

38

104, 119 n.

107

39

126 bis, 128 n.

108

41

103, 126, 195

109

49 44 45 46 47 48

27 n. bis 103 185, 186 186 f. 27n. bis 119, 138 n.

111 115 116 117 118 119

Th.51

12n.,27n.,61f.,119

53 55

27n., 144 62, 128

57

δᾶ, 67, 199

60

27 n., 104, 142, 148, 153 n. 27n.

61

122 128 129

130 131-2 135 138

PAGE 27 n. bis 135 27 n., 198 55, 58, 65 £., 108 91, 119, 184 f., 189, 198 127 n. 27 n. 143 27 n. 126, 129 169 n. 27 n., 139 n. 169 168 166 n., 168 f. 128 55, 70, 130 109 128 27 23-7 145, 164 128 n., 130 128 28 n., 76, 82 156 n. 48 28 n. 48, 51, 52 187 185, 186 12 n. 179, 180 27 n., 179 n. 83 175 n., 176 & n. 180 103, 126, 195 130, 131 n. 49, 52 102, 104 55, 67, 69

930

INDEX

OF HESIODIC

LINE

145 148 149 150-1 Th. 151 152 155 156 159 160 164 166, 172 173 175 178 130 182 184 188 189 190 195 200 Th. 201-300 203 205 209 218-9 220 221 223 228 229 231 232 235 239 240-62 240 241 244 Th. 252 256 263 264 265 267

AND

PAGE

104 55, 69 f. 70 56, 69 72, 76, 82 53 55, 67, 69 186 55, 58, 65 f. 172 109 177 55, 58, 65 f. 172 104, 119 83 n., 105, 111 91 143 105, 111 49 53 56, 69 104, 119 97 186 105 105 56, 67 144 bis 122 n. 56, 62, 67 169 f. 145 145 133 n. 128 n., 130 92 n. 97 128 50, 51 63 49 104 76 91 124 105, 142

PRINCIPAL

HOMERIC LINE 270 275 277 278 280 284 286 292 295 293 300 Th. 302 307 310 911 317 319 321 324 326 331 342 351-400 361 362 366 371 373 374 376 380 388 390-6 391 398 399 400 Th. 401 405-6 408 413 425 430-9 435-8 439 440 44]

REFERENCES PAGE

148, 199 134 84 76, 102 156 n. 56, 58, 63, 69 n. 65 81 137 136 56, δῦ, 136 55, 125 56,

n. 69 67 69 f. 70 f.

49, 77, 82, 207 n. 105 56, 62 106 122 n.

139 n. 29, 30 bis 129 122 n. 128 102, 104 170 199 bis 128, 129 68 70 83 n. 145 bis 83 n. 91 183 f., 189 106, 142, 143 199 186 n. 48, 49 186 109 204 n. 204 n. 52 f., 134 b6, 67, 70 bis

INDEX

OF

LINE

443,446 447 452 456 457 460 470 479 480 483 487 488 493 500 Th. 501 502 503 505 519

HESIODIC

AND

PAGE

56, 67 bis 119 186 56, 70 bis 65 81 145 δῦ, 58, 65 f. 90 f. 56, 69 106, 118 92n. 134n. 62 148, 164 144, 164 91,128 66 62, 186 bis 134

591 598 531 534 536 539 540 541 547 548

164n. 128 56, 58, 63 107, 142, 152 n. 179 43 83 n., 172 43 172 128 &n.

549

92, 126, 128 n.

555

82n.,

556 558

181n. 56, 61

563 572 584 585 590 595 Th. 601-50 601 602 622

625 631

HOMERIC LINE

518

560

PRINCIPAL

633 634 636 644-5 652 653 656 659 661 663 664 666 671 672 073 675 677 678 679 680

172 n.

172

56, 62 125 137 83 n., 107, 195 128, 129 134 29 134 83n. 56, 69

56, 68, 104 144

682 683 687-712 687 696 Th. 703 712 713 715 721 722-5

REFERENCES PAGE

187 56, 68, 104 57, 67, 69, 145, 194 38 91 107, 142, 152 n. 175 91 175 144 187 63 56, 69 56, 69, 72, 76 53 143, 144 & n. 107 n. 48, 107, 108 bis, 118 107 n. 107

n.

83 128 bis 4 192n. 49, 51, 77 91, 185 n. 57,67, 69, 144, 197 57, 63, 67 126, 130 83 43

726

77

728 730

48, 49 124

731

66, 83 n.

732 733 736-819 737 739 741 762 767 768 773

774 781

92]

139n. 108 4 49,51 12n. 130, 131 n. 59 136 124, 125 130, 131 n.

124 59

232

INDEX

LINE

784 789 790 794 800 Th. 802 804 805 800 808 815 818 820-80 821 822 825 827 828 832 834 841 842 844 850 Th. 855 856 858 859-66 859 861 869 870 873 879 882 886 890 893 896 897-9 899 Th. 901-1022 903 904 905-6 907 909

OF

HESIODIC

AND

PRINCIPAL

PAGE

184 & n. 17 59 27 n. 139 n. 143 108, 143, 197 184 122 n. 49, 51 91 57, 10, 102 4, 97 55, 58, 65 f.

199 105 Tl 130 bis 123 n., 125, 134 137 83 n. 56 49, 59 12 n. 92 38 55, 58, 65 f. 97 39 n., 134 n. 66 174 n. 125 bis 49, 59 198 128 198 106 128 & n. 175 n., 176 64 106 5, 98, 117 n., 198 f. 116 n. 116 n. 56, 67, 116 n. 148, 199 116 n.

HOMERIC

LINE 910-11 910

REFERENCES

PAGE

116 116, 117 n., 126, 128 &

913 917 920 921 923 930 935 937 939 941 944 946 948 949 Th. 951-1000 953 961-2 962 964 965-1022 965 969 970 972 975 950 933 985 989 992 993 994 995 998 999 Th. 1004 1005 1009 1011 1012 1014 1015 1017 1018

116 116 64 198 57, 56, 145, 198 199 198 57, 198 198 84 30

n., 129, 197 n. n.

64, 198 70 bis 148, 164, 197

64, 124, 198

198 199 198, 199 48

198 59 78 78, 139 134 104 124 128 125

82, 198, n. n.

& n.

145, 197 136 n. 198 128 198 198 124 199 198 128 128 198

199

INDEX

OF

HESIODIC

AND

Works and Days LINE

Op.1-10

PAGE

PRINCIPAL

HOMERIC

LINE 157

163

REFERENCES

PAGE 56

195

5

164 1 59

22

15, 108, 118

169

25

139 n.

30

126, 130

173

33

15n.

174

12n.

39

15n., 145, 164

180

66

Op. 201-300

98

2,5 27

44,49

119 107

28n.

168 83 n.

169 a-e

172

51-100

30

908

51

65

210

52 53 55 63 64 65 67 68 69 71 74

57, 58, 65 56, 61 148 107, 134, 195 119n. 199 135 102 64 128 π., 125 145

78 79 8l 82 90 91 92 99 100

Op. 101-50

135 70 105 56, 62, 67 1801. 57, 62 127 n. 123 η., 124 145

913 214-6 235 241-2 241 248 252 253 260 261 263 264

164 80

114

109 41 78, 82, 207 n. 137 12n. 110 ῃ. 189 η. 56 66 12n. 139n. 139 n. 127 n., 130 109 15 57, 58, 65 41, 92 145 109, 118 133 η. 12n.

12n.

29

293

41, 92

294

109

66

113 119

57,62 134

123 132 137

66 165 133n.

139

103 66

12n.

268 270 273 274 278 280 282 288

107

143

55, 67

Op. 305

164

126, 129

313 315 316 317 321

119 178 133 n. 170f., 207 n. 109

324

84

325-6

12n. 178

108, 126, 195

335

179

344

109

149

53

352

87n.

153

125

353

114f.

145 147

233

234

INDEX LINE

354 358

364 372 383 384 386 388-91 390

392 304 Op. 401-600 401-50

404 407 409 412 417-32

420-31 428 436-40

441 443 446 451 479 487 400 404

500 Op. 502 506 507 508 509-11 509 510 511 512 518 519-23 521 524

OF HESIODIC

AND PRINCIPAL HOMERIC

PAGE

115 136 n. 133 n. 15 129 129 91 78 f., 82 49 110 12 n. 31 29, 30 bis, 32 bis 128 n. 133 n. 133 n. 133 n. 32 32-6 133 n. 74 74, 76, 82, 207 122 n., 133 n. 79 110, 125 176 146 134 133 n. 171 & n. 110 125 49

174, 175 n., 207 n. 175 146 56, 110 182 111 113, 125 113 78, 82, 133 n., 199 111-3, 134

525

112

526

110

529

113

531

92 n.

LINE

533 534 541 547 550 551 553 556 560 562 564 571 572 574 578 589 590 592 596 598 . 605 610 611 617 620 621-4 622 624 625 628 630 63540 635 645 646 647 648 649-51 650 Op. 653 655 656 663 666 667 672 675

REFERENCES PAGE

113, 134 102, 113 39 n. 125, 182 15 122 n. 125 79 113 92 134, 144, 152 f., 176 112 bis, 113 130 84 133 n. 134 129 135, 184 n. 79 f., 82, 90 87 n. 118 43, 133 n. 121 n. 15, 114 45, 59 174 50 80, 82 173 f. 113 93 201 n. 121 n. 144 bis 178 128 n. 49 201 n. 49 80, 82 15 139 n. 144, 152 f., 176 115 57, 10, 102 53 144, 153 n.

INDEX

OF

LINE

HESIODIC

AND

PRINCIPAL

PAGE

678

HOMERIC

LINE

57,69

24

PAGE

12n.

682 683

57,69 108

688

92n.

693 696

115 115

698

115 f.

53

64

135 135 113

58 59 68

70 57, 58, 63, 71 70, 126, 130

107, 134, 195 139 n.

76 79

653 164

699 Op. 724 742-3

744

29 33

REFERENCES

39

41 49-56

134

752 753 758 165-828 769 771 777 778 783 788 793 794 Op. 801 814 815 817 818 826 828

70

12n., 129 5 6 8 110 113 657,69 67,69 12n. 57,69 57,69 204 n. 81, 82 bis 49 12n. 129 57, 69, 144 bis, 197

Shield of Herakles LINE

PAGE

81 94 100 Sc. 102 104 107 108 109 111 112 116 120 134 144 156 161 162 166 172 178

63 1060, 196 148

78, 82 99

δῦ, 70

83 146 70, 125 56, 70 f. 56, 67, 70 144 92 146 134 125 34n. 77 125 55, 69 104 55, 69 181 n. 134 16 n., 165 n. 129

188

144 bis & n.

192

57, 58, 71

194

146

Sc. 1-56 2

4, 6, 99, 117 n. 99 nn

4 6 7

126, 129 129 12 n, 116-8, 197

223

38

8

78,82

232 234

130 118 bis 126, 130 126, 130 bis

146

129,

Sc. 201-300 204 207 210 215

10

129

13

144

15

12n.

237 246

19

129

249

32, 99 52 48, 49, 50, 51 104 137

235

236

INDEX

OF HESIODIC

LINE 251-300 257 260 261 267 275 276 278 283 286 292 293 296 298 Sc. 302

304 305-6 308 318 333 337 338 341 346 351 358 377 378

PRINCIPAL

PAGE 129, 30 104 126, 129, 130 57,63, 67 81, 82, 126 bis, 130 127 π. 126, 130 126, 130 96 81, 82, 104 144 & n. 146 146 96 145, 164 f., 196 165n. 204 n. 119 70 57,67 146 129 91 57,58, 63, 71 57,62 81 f., 82 126,130 82,127 n.

Iliad

A

AND

LINE

PAGE

1-25 75 98 125 276 312 350 485 528 637 540

2] n. 125 136 11 20 60 48 80 77 142 142

LINE

581 589 604

B

40 87 130 165 181 228 232 233 261

HOMERIC LINE 379 397

Sc. 401-50

REFERENCES PAGE 57,68 12n.

29, 30

412 414 424 425

57, 68 57,68 57, 67, 71 57, 67

433 439

63 12n.

448 453

57, 67, 71 57, 68

459

63

479

144

Hesiodic Fragments Rzach

Merkelbach and West

PAGE

— — 21b.15

23a.25 30.11 75.15

181 n. 181 n. 136 n.

55 — 48 — 76.7 180.4

150.15 150.20 158 195 205.7 291.4

144 n., 150 n. 104 185 n. 4 n., 116 n. 113 181 n.

265.3

357.3

83

PAGE 35 171 n. 128

LINE 318 465-6 524 535 659

PAGE 111] π. 185 74 131n. 83

144

570

129

625

83 128

131 n.

626

131 n.

148 n.

683

80

148 n. 90

719 147

76 74

198 137 n.

766 769

144 35

79

821

68

LINE 53 61 217 245 269 273 276-80

OF HESIODIC AND PAGE

198 172 36 .183 183 34, 130 183 24 178 61

PRINCIPAL HOMERIC LINE

382 489 490

PAGE

53 26 35

49 125 53 76 178 63

178 34 178 178 33 125

287

28n.

27 n.

400 404

$82 145

71

24 71

36 24 24 180 91 28 n.

187 33 49 bis, 77 62 76 49, 60 34 148 n.

237

144 71 129 81 27 n. 185 84 130 35 35 130

28 n. 178

146 68 145 67 81 180 180 23, 145

REFERENCES

PAGE,

418

145

456 494

180 34

50 69 209 397 400 44] 447 573

108 35 84 198 79 130 80 91, 185

4] 56 84 253 286 354 416 467

71 76 35 77 144 185 67 75

62 68 147

71 35 33

62 136 n. 79 101 33, 180 n. 82 71 124 77 67 63 27 108

Gl

INDEX

76 145 56 91 104 47, 49 107 101 8] 49 185 91 198 198

238

INDEX OF HESIODIC AND PRINCIPAL HOMERIC REFERENCES LINE

345 360 387 395 401 407 480

10)

PAGE 33 198 133 n.

123, 125 68 91 62 49 24 142 67 83 58 77 144, 151 179 34 ter

9 53 144 68 91 80 61 49, 71 81 81 68 71, 78 68 75 63 741 180 n. 84 34 33 53 68 148 n. 153 77 128

LINE

LINE

PAGE

535 551 564 568 589 598 603 616

104 n. 144 77 146 148 n. 144 28 n. 2" n.

92 267 399

277 n. 59 91

3l 34 105 154 174 183 206 224 231 305 481 490

63 101

177 195 224 234 334 379 479 495 600

79 124 146 24, 48, 52 125 137 n. 198 109 136 n.

76 184 292

33 179 91

341 422 482 584 603 679 692 744 832 833

£2

33 55 139 140 144 268-9 270 316 329 359 537 545 548 615 721 776

78

91, 142 176 179 76 76 76 24 8l 110 n.

33, 144 146 33

79 144 61 102 179 125 125 49, 59 79 91, 145 111 61 114 179 180 35 34 125 82 53 198 48, 169 128 23 107

n. n. n.

n.

50

α

LINE

PAGE

1-25

2] n. 185 63 131 n. 23

188 349 421

β

145 254 377

131 n.

185 183 £., 188 f. 36

182 24, 48, 52 204 n. 50

n.

Odyssey

380

114 118 123 216 230 262-87 316

PAGE

62 102 137

INDEX LINE

128 147 158 358

OF

HESIODIC AND

PAGE 175 f. 24 59 34

PRINCIPAL LINE

204 2734 330

111 n.

198 198 59 184 184 24 84 174 36 34 183 68 174 174 173 173

n. n.

f.

f.

PAGE 49 181 n. 145, 164

198 11 144 62 180 f.

136 112 n. 65 49 172 142 170, 171 59 bis 48 172 24 142 9] 19 187 60

HOMERIC

θ

82

142 169 180 n. 166 n., 168 f., 171 26 129 186 f. 187 199 199 91 34, 77 62 34 26 27 n.

REFERENCES LINE

512 556 107 125 133 218 220 250 276 323 386 394 437 438 562 576 612

939

PAGE 125 185 49, 59 bis 113 170 109 62 91 142 128 128 53 142 185 176 75 f. 169 f., 171 107 146 177 n. 125 124, 125 124 185 130 62 136 n. 180 n. 144 27 n. 24 53 80 f. 53

53 168 n.

62 179 f. 180 33 131 n.

33

49, 77 79 23, 187 182 125,182 f. 114

240

INDEX

OF HESIODIC AND

PRINCIPAL

HOMERIC

PAGE 92 130 34 182

PAGE 114 n. 125 24 177 n. 34 28 n. 145 36 144, 176 f. 60

REFERENCES LINE

PAGE

19 166 219 223 272 280 322

69 f. 128 198 107 113 170 126

ω

Homeric Hymns

91, 184 f., 188 f.

h. Dem. 3

79 115 334 341 342

178

130 176 n.

179 80 142

441

φ

53 185 180 n. 128 170 f. 171 n. 61 172 23, 187 28 n. 34 23 53

76 49 112 n.

81

118 528

24 105

198 288 296-301 302 362

X

ψ

168 176 193 319 432 12

460 h. Apol. 83 109 123 h. Herm.

110

195 518

h. Aphr. 120 263 286 À. vi, h. xv, h. xviii, h. xxix, h. xxxiii,

3 7 12 8 5

70 198 110 70 180 131 10 70 184 180 83 77 142 184 52 198 179 174 180 187 187 198

n. n.

n.

n.

n. n.

GENERAL abstracts, 20, 21, 82 aecentuation, 104, 127 n.

Archilochus,

accusative, see ‘ short ' acc. pl. forms * Achaean

phase

of epic,

21,

152 n., 202 adaptation of traditional see modification

122,

phrases,

‘adding ' style, 93, 97, 99 adjectives, H.’s use of, 20,

51 ; see

also epithets Aeolic,

17 f.,

VIII

102-20

passim,

123, 126, 131, 194 f. ; — phase of epic, 122, 152 n.; see also Boeotian, Lesbian, Thessalian

Aeschylus, 112 Aithiopis, 202 n. Aleaeus, 107, 115 Aleman, 142 n. alternative expressions,

my’; 103,

— 111,

INDEX 112, 203, 204

Ares, epithets of, 71 Argive, 141 Aristarchus, 33, 91, 111 n. artificial forms, 104, 110 n., 149, 152, 153 artificiality of language, 1, 3, 119, 162 assimilation, 111 assonance, 69, 104 bis Attic, 102, 104, 107, 108, 113, 120 Attie-Ionic, 105, 121, 141, 152 n., 162 autobiographical references, 41, 119, 192-4, 201 n., 203 f.

Batrachomyomachia, see ' econo-

forms

and

inflexions,

122-65,

and

see INDEX

OF GREEK Amphidamas, 204 analogy, in forms, 104 bis, 109, 118 bis, 147-50 ; —, in phrases,

60 f.,

62-8, 134 n., 191, 206 anaphora, 167 animal-names, 111-3, 125 n. anomalies, prosodical, 79 f., 82, 90-3,

101 f., 135 n., 153 antithesis, 115, 165 n., 167 aoidoi, 131, 135, 137-9, 152 ; see also

oral composition, recitation, rhapsodes aorist, 107 ; — participle, 164 n. Aphrodite, epithets of, 56, 69, 78,

109 n.

Boeotian, 39 n., VIII 101-21 passim, esp. 119 £, 131, 132, 147, 149,

152 n., 161-3, 194 caesurae, 88, 152 Callimachus, 86 n., 117

Catalogue of Women, see Ehoiai Catullus, 189 n.

cavalry, 204 n. Chalkis, 204 Chian, 115 Chimaera, epithet of, 49, 77 chronological place of H., 7-9,

tion citations, 1 f., 12, 14, 16 Coan, 151 cola, 88

compensatory lengthening, labie weight

105 apoeope, 107 Apollo, epithets of, 70

composition,

Apollonius Rhodius, 86 n., 87 n., 94,

166,

199-206 ; see also ‘early’ and ‘late’ features, order of composi-

see

oral

see syl-

composition,

order of composition, writing

Aratus, 86 n., 87 n., 97 n., 99 n., 191

contraction, 103, 119, 122, 126, 131, 139 n., 195 n., 200

Areadian, 103, 121, 142, 152 n., 161 archaisms, 88, 103, 105, 106, 123,

Corinthian, 139 n. Cretan, 141 bis;

99

126,

131;

see also

guistic features

'early"

lin-



oral

192 n. Cyclic epics, 202 n., 206

poetry,

242

GENERAL

INDEX

Cypria, 202 n. Cypriot, 111, 152 n.

equivalent expressions, see ‘economy’ etymologies, 105 Euripides, 70 n., 117

daetyls, 85 f. date, see ehronologieal plaee

exemplum,

deer, 113

formula, definition of, 28 n. 4, 42 f. formulaie style, 26 n., 60, 68, 190 f., 206 ; — system, 1, 42, 43, 45-52,

Delphian, 17 n., 149 dialeet of H. poems, 3, 17 f., 22, 11821, 193-5, 201-3 ; see also vernaeular of H. and the names of the various dialeets.

dietation, see reeording dieetasis, 110 n. digamma, initial, 16-18, 60, 113, 132-9, 196, 200 ; —, intervocalic, 139 n., 195 n., 200 ; —, posteonsonantal, 107, 134, 195 ; —, pronuneiation of, 39 n., 135—9

Dorie, 17 f., 22, 103, 105, 116, 120, 141, 146 f., 151, 152 n., 194 ; see also West Greek. dormouse,

191 ; see also juxtaposition, modifieation fragmentary poems, Hesiodie, 1 f., 18 ; see also Ehoiai Gaia, epithets of, 65 f., 179 f. genitive, objeetive/possessive, 183 f.; gen. pl. of a-stems, 126-31, 196 ; gen. sg. of mase. a-stems, 123-6, 130 f,

196 ; —

of o-stems,

122,

130 f., 196 gnomie

element

in

Op.,

86 n. ; see

also proverbs

113

dual, 84, 119 n., 165 ‘early ’ linguistie features, 38, 136, 152n., 200-2; see arehaisms

122, also

60, 73, 190f.; —, departures from,

51-3, 55-73, 187, 206 f. editions of H., 13-6, and see BIBLIo-

GRAPHY Ehoiai, 2, 4, 99, 198, and OF REFERENCES Elean, 141 elision, 107 f., 138

see INDEX

enjambement, 33, 93-9, 190 f. ephelcystie -v, 16, 133 n, 137

2, 37, 46 f., 166-89, 201-3; see also ' late ' passages in Homer and INDEX or REFERENCES humour, H.’s sense of, SO, 105 Hymns, Homerie, 75, 78, 85 £., 187,

Ibyeus, 117 imitatio, 19 f., 188 f., 207 imperfeet, 107 f.

Epicharmus, 142 n. Epigram, Homerie, 102 of,

hibernation, 112 f. Homer, date of, 8 £, 2046; —, relationship of H. to, 7-9, 18, 20-

and see INDEX OF REFERENOES hyper-Ionisms, 102 n. hyphaeresis, 127 n.

ellipse, 177 Empedocles, 142 n.

ehoiee

Hekate, 52, 186 n. Herodotus, 7 n. Hesiod, name of, 119 n., 192 n. Hesyehius, 112 n., 125 n.

hiatus, 91 f., 132, 136, 138, 139

*eeonomy’, observanee of, 50f.,52-4,

epithets,

19, 189

influeneed

by

indireet names, infinitive,

—-, use

innovation,

of, with

ence, 75, 77, 78

ehange

of refer-

52 f., 111-3,

125 n.,

134

metre, 65, 66, 80 f. ; —, deseriptive/predieative, 51 £., 171 n. ; —, inappropriate, 50, 83 n. 5, 174 n. ;

121 n. ;

forms

of,

109,

119 n.,

ineonstruetions, 81, 114 n. linguistic,

in H.,

51-4, 66, 82-4, 134 f.

3, 39,

GENERAL

interchangeable expressions, see “economy ' interpolations, 'see suspect authorship Ionian epic, 137, 139, migration, 38 n., 202

202f.;



epithets, 65, 66 ; —, — forms, 104,

119, 122, 150n. 152; —, — tenses, 107 n. ; as protecting text, 13, 127 metrical effects of F, 132, 135 n., 138

metrical

Tonic, 17, 38, 102, 106, 110, 113, 120, IX 122-39 passim, 162 f., 193-6, 200-3 ; see also Attie-Ionic, hyperJonisms irregularities, see anomalies, prosodical, and syntax

243

INDEX

position

of

words

and

phrases, changed, 27 n., 50, 64-7, 83, 151-3 ; —, retained, 23-7, 36,

50, 67 f., 74-82, 88, 124 £., 198 £., 136 f., 153 modification of formulaic phrases,

juxtaposition of formulas, 91 f. ; — of forms in -as, 148

by change of case, 27 n., 50, 51, 62 f., 101, 152-4 ; —, — of person, 63 ; —, — of position, 27 n., 50, 64-7, 151 f. ; —, — of reference, 60, 63f.; —, by other changes,

Kennings,

83 f.,

111-3 ; see

also

indirect

"names

91-3,

107,

136£,

148 η.,

170 n.

Mycenaean, 38 n.

koine, epie, 202

Kyme, 17 n., 194 Lacedaemonian, 112 ‘late’ linguistic features, 38, 122, 130 f., 139, 193, 200-2 ; see also innovation

‘late’ passages in 169 f., 187, 205 n. Latin language, ture, 89, 188, Lelantine War, length of poems,

Homer,

126,

Lesbian, 103, 110, 111, 115, 141, 152 localization, 25 f., 88 ; see also metri-

cal position Lucretius, 97 n., 99 n., 189 n. lyrie poets, 9, 101, 107, 113,

115,

117, 142, 161 f., 203

metre,

149,

Oppian, 87 n. oraeular poetry, 111 n. oral

composition,

generally,

60 f.,

65, 72 f., 93, 97, 189; —, in H., 21, 40-5, 84, 93, 97, 99 f., 190-3, 198 oral poetry, modern, 20, 72, 93 n., 191, 192 n. oral transmission, 13, 191, 192 order of composition, 7, 19, 167, 188 f., 197; see also Homer,

relationship

of

H.

to

of lines, 169, 204 n. of words, see word-order

2, 11-16

by scholars,

168, 172,

176, 179 £., 182, 187 memorization,

116,

oath, 183 f. octopus, 112

order order

meaning, 82, 134 f. ; H.'s sensitivity to, 66, 69f., 71; —, variously understood

108,

opposition of vowels, 162 f.

106, 149 ; — litera189 204 2, 42-4, 191 f., 196,

208 lengthening, see anomalies, prosodical, and syllabic weight

manuscripts,

Nonnus, 190 n. North West Greek, 152 n., 194

84, 191, 192

18, 85-90 ; —, as influencing

palatalization, 111 Pamphylian, 152 n. papyri, 1 £., 4 n., 12-15, 108, 116 n., 127 n., 175 parallels

of

sound’,

28,

74-84,

111 n., 131 n., 191, 197 f., 207

244

GENERAL

Pausanias, 5, 6, 45 personification, 66, 135, 169 Pindar, 101 Plato, 106 pleonasm, 168, 179 Poseidon, epithets of, 70, 76, 101 f.

INDEX

* surproduction ' of expressions, 68 suspect authorship, in H., 4-6, 97 f., 99, 105 n., 196-9 ; see also ‘ late’

passages in Homer syllabic weight, loss of, 107, 134, 142, 153; —, maintenance of, 141 f.,

153

priority, see order of composition pronunciation, 39, 104, 105, 111, 126, 135-9, 148, 150, 162-5 prosody, see anomalies, prosodical,

synizesis, 110, 123, 126, 139 n. syntax of H., 18 ; peculiarities in, 81, 82, 106 f.

and syllabic weight testimonia, text of H.

proverbs, 20, 41, 88, 171, 181

pun, 105

BIBLIOGRAPHY

punctuation, 92 n., 95, 96 n. Quintus Smyrnaeus, recitation,

72,

92 n.,

190 n. 95,

135,

137,

162 f., 165, 192 f., 198 recording of oral poem, in writing, 72, 187, 189, 192 n., 193, 198 repetitions, 19, 20, 40-5 rhapsodes,

13, 72, 135, 205 ; see also

aoidoi

under

the

also

chief

editions Theocritus, 151 bis

86 n., 89 n., 117, 142 n.,

Theogony,

as

poem,

17,

supposedly 146;

—,

‘ Doric’

end

of,

5,

198 f. ; similarity of its language to Op., 195-7 Thessalian, 103, 111, 121, 142, 152 n., 161 tmesis, 182 f.

tradition, see manuscripts, oral composition, oral transmission

riddles, 20, 111-3

Sappho, 107, 113 schema Pindaricum,

see citations poems, 11-16 ; see

transmission, reliability of, 11—13, 102, 154, 164 ; see also citations, manuscripts, oral transmission,

106

Scottish oral poetry, 191 n. ' sea ’, expressions for, 45-53, 58-60,

73, 191

papyri transposition, see order of lines

Serbo-Croatian, 93 n., 192 n. Shield of Herakles, authenticity of, 4, 6, 196 f. ; novel epithets in, 50,

tribrach, 79 trochaic first foot, 15 Tyrtaeus, 142 n., 147, 150

70 f. ; peculiarities of versification in, 87, 99; as possibly an poem, 20, 197 ‘short’ acc. pl. forms, 101 n., 140, 141-65, 195 f., 197 Simonides, 142 n.

oral 134,

unity of corpus, 4—7, 195-9

singers, see aoidoi snail, 112 f.

variant readings, 11, 14, 16, 106, 110, 171n.; see also BIBLIOGRAPHY

sound, use of by H., 22, 74, 84, 191 ;

see also

assonancc,

' parallels

of

sound '

spoken language, see vernacular of H. spondees,

85—7, 89, 99, 200 n.

Stesichorus,

142 n.

un-Homeric forms, 101-21, 139 n., 141-65, 193 f. ; — passages, 30-3 — vocabulary, 53, 70, 134 n.

under the chief editions vernacular spcech of H., 39, 138, 147, 151, 165, 201 n. versification,

see

position, writing

Virgil, 94, 99

metre,

oral

111, com-

GENERAL

vocabulary of H., 18, 21, 22, 53, 70, 134 f. vowels, confusion of, 105, 162 f.

West Greek, 103, 109—11, 116, 120 f., 142, 147, 161-3, 194; see also Doric, North West Greek word-formation, 18, 104, 112 word-order, 81, 109, 135, 153, 183 f. Works and Days, ‘ Days’ of, 5f.;

high proportion of spondees in, 86 ; — of -ov gen. sg. in, 122 ; —

INDEX

245

of neglected F in, 133 ; similarity of its language to Th., 195-7; varied subject-matter of, 32, 41, 82 writing, use of, in composition, 44 f.,

58, 70, 72 f., 93, 97, 190-3, 198 Xenophanes, 117 Xenophon, 183 n. Yugoslav oral poetry, 93 n., 192 n. Zeus, epithets of, 61, 64 f., 70, 83 n.

INDEX à/1, 102, 104, 106, 113, 162 f.

ἀγαύρου, 134 ἀγήρω͵-ων, 83 f. ἦθλος, 139 n. αἴνημι, 108

OF

GREEK δεκαδώρῳ, 84, 86 δεύση δύσω, 79 δήσας, 104 n. διδασκῆσαι, 119 n. δοῦπος, 91, 184 f.

ἀκάματον πῦρ, 62, 77 ἁλμυρὸν ὕδωρ, 48, 60 bis ἅλς, 46-8, 58 f.

(F-)& 138 n.

ἀμάειν, 110 ἀμαιμάκετος, 49, 51, 77

ἔγχος pkos, 77 ἔδιδον, 108

-äv gen. pl., 103, 119, 127, 195 avocreos, 111-3, 134

ἔδον, 108, 121 n. ἕεις, 104 ἐθύνεον, 104

éaye, 113

-äo/-ew, -w, 123-6, 130 f., 196 ἀπείριτος, 48, 51 f. ἀποδρέπεν, 121 n. ἀπόπροθι, 78 f.

( F- )εῖδος, 134 ein, 114

( F-)etxos, ( F-)é( F)owa, 136 f.

ἀρηΐφιλος, 70 f.

εἶμι

“Αρπυίας, 142, 150

εἴοι, 15, 114 εἴρας, -péas, 143, 150, 197

ἀρχή, 185 f. ἀρώμεναι, 108 f.

eiui, 109, 114

εἰρεῦσαι, 104, 119 n.

-ds/-as acc. pl., 134, 141-65, 195, 197

εἰς, ἐν and és, 152

ἀσφαλές, -écs, 179 f. ἄτριχος, 113 n.

(F-)Aur-, 136 -euev in thematic infin., 119 n.

ἄφυζα, 113 n. -awv/-Ewv, -àv, 126-31,

ἐν + acc., 106 f. 196

βασιλῆς, «λέων, 139 n. βίβλινος, 134 βλέφαρος, 116 BovAds, 142, 150, 151 βοῶν, 139 n. βριάω, 119

γαῖα in hexameter, 25 γαῖα πελώρη, 65 f.

yavjoxov, 101 f., 153 γάρ, 35, 177 f. γένεσθαι in hexameter, 25-7 γηρύσασθαι, 53 γλαυκή, 52 f., 134 γναμπτοῖσι )΄στιβαροῖσι, 53 γόμφοισιν, 36 bis

8/t, 111 δείκνυ, 110 δεινάς, 144, 150, 151

ἐνὶ dpeci/ pera φρεσί, 92 ( F-)é£-, 134 ἐπί

+ dat., 172

ἐπιβαίνω, 81 ἐπικροτέοντα, 119 ἐπικυρτώοντε, 118 ἐπίφρονι βουλῇ, 175 f. ἑπταπόδην, 34, 36 (F-)épy-, 134 ἐς, 152 ( F-) “Eomepides, 134 εὐφρόναι, 113 fF, see digamma

in GENERAL

t/8, 111 ζάθεος, 69 ἠεροειδέα πόντον, 49, 59, 60 (Ε- ἦθος, 135 ἡμερόκοιτος, 118 ἦν, 105 f.

INDEX

INDEX

‘Hoiodov,

θάλασσα, 46-53, 58 f, θεᾶν, 103, 195 Θεία, -av, 104, 106 θεῖναι θεῖνε, 81

130 n. 25 188 μερόπων a., 66

ἴδρις, 118

241

GREEK

ξυνήονας,

119 n.

ἥσυχοι, 134 (F-)nx-, 136

θεμιστέων, 128 n., θεοί in hexameter, θεῶν ὅρκον, 183 f., θνητῶν ἀνθρώπων

OF

134

( F-)oi, 138 n. ot, οὕς ἡτοί, τούς, 164 οἴδματι θυίων, 24, 48, 49, δ2

(F-Joix-, 135 (F-)oivos, 134, 135

( F-Joivoxönv, 134 -o:0/ -ov, 122, 130 f., 196 -oıs/-oroı, 119 n., 170

ὄλμον, 33, 36 -os/ -ovs, 141-65, 195 f., 197 (F-) daca, 134 2 οὐκ ὀνομαστός [οὔ τι φατειός, 69 f.

ἱερός, 76, 80 f.

Ἵμερος, 135 ἱππήεσσι, 204

Ilavöspn, 105

ἴσκε, 168 ἴσον, 107, 134 ( F-), 153, 195

πέλαγος, 46-53, 58

( F-) ᾿Ιφικλεΐδην, 134

περίαχε, 107 f. περοίχεται, 108 πιλνᾷ, 110

καλόν, 107, 134, 153, 195 καμπύλα, 35 καυάξαις, 115 κεραμεῖ, 139 n. xepaot, 113 κούρας, 142, 148, 150

κυανεάων, 116-8, 129 κυανοχαίτης, 76 f.

Aayós, 145, 148 f., 151, 154, 164f., 196 λαῖτμα, 59 λαμπράν, 102 f., 106 λουτρῷ, 139 n.

räcas, 143 n. πεντόζοιο,

113

πίπτω, 182 πλουτεῦντα, ποδῶν ὕπο, πολύχρυσος,

f. 119 184 f., 188 78, 199

πόντος, 46-53, 58-60 Ποσειδέων,

139 n.

πουλυβότειρα, 63 πρίνινον, 35, 36 προσεῖναι, 109, 114 f.

πρωιηρότῃ, 134 πρῶτα in hexameter, 25

πτερά (νηός), 113

Aoxéovo, 104, 112 μέζεα, 111

μελιᾶν, 103, 195 μερόπων ἀνθρώπων θνητῶν à., 66 μετὰ φρεσί,&i φρεσί, 92 μεταναιέτας, 142, 150, 152 μήδεα, 83 n., 111

μητέρα μήλων, 63, 69 n.

μητίετα Ζεύς, 38, 65 μητιόεις, 65 νήκεροι, 113 NMjos/ Νηρῆος, 76

ῥεῖ, 139 n. *Peíav, 104

Σιμοῦντα, 139 n. Σκύθας,

144 n., 150 n.

στιβαροῖσι γναμπτοῖσι, 53 στίγματα, 134 συμφέρω,

81 f.

Táprapa, 83 τεῖδε, 121 n. τένδει, 112

248

INDEX

OF GREEK

Teooapakovraerns, 14

ὑλοτομεῖν, -μος, 33, 36

reropa, 115 f., 119, 121 τέχνη, 172 f. Τίρυνθος, -vvs, 83 τιταίνω, 105

ὑμνέω, 119

τοί, τούς "οἷ, οὕς, 164

φερέοικος,

112, 118

-di, 119 n. Dixa, Φίκιον, 106, 119, 196 f.

τριηκόντων, 115, 119, 194 τρίποδι, 112, 134 τρίς, 79 £., 90

φιλομμειδής7-μηδής, 104 f. φλεγύαο, 125

τροπαὶ ἠελίοιο, 176 f. τροχαλόν, 113

χάλκεον éyxos/x. ἕρκος, 77 χρεέων, 128 n. xpvodopa/-ov, 83

ὑγρὰ κέλευθα, 60

ψεύδεα

τροπάς, 134, 144, 150, 152, 153

ἐτύμοισιν, 1, 167 f.

ὑγρή, 53 ὑληκοῖται, 118

(F-) ὥς, 134