Harvey, A. E._homeric Epithets in Greek Lyric Poetry_1957_CQ, 7, 3-4, Pp. 206-223

7/27/2019 Harvey, A. E._homeric Epithets in Greek Lyric Poetry_1957_CQ, 7, 3-4, Pp. 206-223

http://slidepdf.com/reader/full/harvey-a-ehomeric-epithets-in-greek-lyric-poetry1957cq-7-3-4-pp-206-223 1/19

Homeric Epithets in Greek Lyric Poetry

Author(s): A. E. HarveyReviewed work(s):Source: The Classical Quarterly, New Series, Vol. 7, No. 3/4 (Jul. - Oct., 1957), pp. 206-223Published by: Cambridge University Press on behalf of The Classical Association

Stable URL: http://www.jstor.org/stable/637463 .

Accessed: 03/12/2012 18:29

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

Cambridge University Press and The Classical Association are collaborating with JSTOR to digitize, preserve

and extend access to The Classical Quarterly.

http://www.jstor.org

This content downloaded by the authorized user from 192.168.72.227 on Mon, 3 Dec 2012 18:29:24 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=cup

http://www.jstor.org/action/showPublisher?publisherCode=classical

http://www.jstor.org/stable/637463?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp





http://www.jstor.org/stable/637463?origin=JSTOR-pdf

http://www.jstor.org/action/showPublisher?publisherCode=classical

http://www.jstor.org/action/showPublisher?publisherCode=cup



HOMERIC EPITHETS IN

GREEK LYRIC POETRY

ONEof the ways in which a poet may show his quality is by discriminationandoriginality in his choice of adjectives. Poetry likes to adorn the bare noun; anoun such as 'the sky' calls out for an attribute. But in practicethe poet has totake care to avoid the cliche. He can seldom write 'the blue sky'; even 'theazure sky'has become trite. He has to search for the epithet which will be both

apt and original. One recollects the story that A. E. Housman pondered formonths before the right one came: '... and see the colouredountiesand hear

the larks so high'.Homeric language has the peculiarity that one of its main resources s the

use of epithets which, by virtue of frequent repetition, already have a ring oftriteness or conventionality. I refer to the 'Homeric formula' or 'ornamental

epithet'. There are occasions when it is significant that Achilles was fleet offoot, and on these the epithet irrdasdKV' adds something to the line. But whenthe context contains no reference to running, this epithet is purelyornamentaland conventional. The frequency of these ornamental epithets is familiar toevery reader of Homer. As an element in Homeric style, they may be satisfac-

toryand effective.

Regardedin

isolation, theyretain little

vitality: theyare

'dead'.

This, of course, is by no means a defect, but simply an interestingcharac-teristic, of the Greek epic style. It is not confined to Homer. Hesiod, also

writing in hexameters, draws almost entirely upon the traditional Homericvocabulary,and the innovations,though significant,arefew; and the samegoesfor almost the whole collection of Homeric Hymns, despite the differencesofgeneration among them. Elegy, however, partly no doubt owing to the slightoriginality demanded by the metrical innovation of the pentameter,shows aprogressivedevelopment. Tyrtaeusstill usesHomeric epithets indiscriminately;

but Solon is alreadymore original, and the Theognis collection is on the wholesparingwith ornamentalphrases,even though in certain contexts (suchas anymention of the sea) it is a slave to them.

Now the presenceof these Homeric phrases in post-Homerichexameter andelegiac poetry is perhaps not surprising.By choosing to write in the metre ofHomer, or one as closely related to Homer's as the elegiac, the poet was com-mitting himself to a position of discipleship towards the master, and his taskwas not to find a new language for himself but to invest the one bequeathedtohim with a certain freshness by introducing, here and there, an element oforiginality; and in courseof time the original elements began to preponderate

over the traditionalones. But the survival of Homeric phrases n the completelynew context of lyric poetry is a much stranger phenomenon. Archaic lyricpoetry is often a long way from the epic. It may use entirely differentmetres,it may be personal instead of narrative, and its language is freshand direct.Nevertheless, conventional Homeric epithets occur. Sometimes it seems fairlyeasy to explain their occurrence: if a lyric poet adopts a dactylic metre, ornarrates a heroic story, it is not too surprising if features of epic dialect andvocabulary tend to appear. But there are many more instanceswhich are not







HOMERIC EPITHETS IN GREEK LYRIC POETRY 207

so readily explicable, and which are especiallydisconcerting n that they occurin passages where the diction is otherwise crisp and economical. To put the

problem in its sharpest form: when, in a poem where every word tells (in

Sappho, for example), we come across what I have called a 'dead' expression,an ornamental epithet borrowed from Homer, should we be shocked or

delighted? Is it simply a cliche, and therefore a defect in the poem, or is it aconscious archaism, a judicious borrowingfrom an honoured tradition?

Despite the interest of scholars in the survival of Homeric language intolater literature, I am not aware that the sharpaestheticproblem raisedby thissurvival has been properly examined.' It may be that the problem has no

answer, or that it has been shunned for its associationswith literary criticismrather than rigorousscholarship; yet I believe that it is possible to approachit in a reasonably scientific way. For this purpose I propose to concentrate

upon a part only of the already small field of lyric poetry. The choral Dorianlyric (with the exception of Alcman, whose language is limpid and direct)is not very helpful: the diction of Simonides,Pindar,and Bacchylidestends toaim at an effect of rich profusion,in which a large number of epic words andphrasesoccur together with rarerexpressionsand original coinages; and sincethese poets (despite individual differences)were drawing on all the linguisticresourcesavailable to them, among which Homer naturallyheld an importantplace, it is impossible to subject their practice to rigorous control. Certain

specific instances are illuminating, and will be referred to in what follows;

but a systematic analysis of the poems does not yield any clear result. Thesame applies to tragedy. Here the number of Homeric epithets is in any case

exceedingly small; but the language is often of such a richnessthat it can ap-parently absorb a few conventional expressionswithout any noticeable effecton the style, and although there are isolated instances which lend strikingsupport to the interpretationhere proposedof certaintendenciesdiscernibleinlyric, the writing of the tragic poets is too varied and original to supportanygeneral conclusions. There remains the 'monodic' lyric of Lesbos and lonia(under which head I include, for convenience, Archilochus). These poetswrite in concise metrical forms (and usuallyalso in non-dactylic metres) with

an extreme economy of words and directnessof expression. Any superfluityisat once noticeable; and this makes their practice with regard to ornamentalHomeric epithets easy to control. A simpleprinciplecan be applied: where an

adjective occurswhich is neither predicativenor essential to the sense-whichis ornamental, in fact, in so far as the sentencewould be complete and meaning-ful without it; and where that adjective is already familiar in such a contextfrom Homer, and there is no strikingoriginality in its employment; we cancount this as a gratuitous Homerism, an occurrence of a 'dead' expression.And these instances form the basis of a reasonably objective and systematicstatistic.

But merely to present the result of such an analysisstatisticallywould be toignore both the subtlety and the interest of the problem.A mere list of Homer-isms will have no value in itself unless it can be shown that any noticeablepeculiarities in the distribution of the phrasesare unlikely to be fortuitous,butreflect a more-or-less conscious variation in the practice of the poet. Some

SM. Leumann, in Mus. Helv. iv (1947), pp. Ix9 ff., well describes the phenomenon, butdoes not consider its implications.







208 A. E. HARVEY

evidence must first be produced that archaic poetry had achieved this degreeof self-consciousnessn its choice of vocabulary and its conventionsof diction.

It is easy to observe that the appearance of an epic subjector of epic charac-

ters in a lyric poem tends to be accompanied by Homeric phraseology.Thistendency is particularlymarked in Bacchylides, where the differencebetweenthe vocabularyused in the narrative and the non-narrativepassages sso strik-

ing that no doubt is possibleabout a deliberate intention behind it. To take an

example: the central part of the thirteenth Epinikion relates the storyof Ajaxdefending the Trojan ships from the attack of Hector. There are certaindivergencesfrom the version in the Iliad, but the main lines of the Homeric

story are adhered to, and the passage contains many verbal echoes of Homer.Of the 29 ornamental epithets in these lines (vv. 97-169), 17

can be found inHomer; that is to say, the proportionof traditional to non-traditionalepithetsis 17 to I2. But in the following passage (vv. 175-231), which is concerned with

extolling the victor and offering moral guidance, the ornamental epithets,though only a little less abundant (there are I6), are all, with only two excep-tions, ones that are notto be found in Homer; 2 to I4, therefore, s the propor-tion of traditionalto non-traditionalepithets. An analysisof the fifthEpinikionproducesa closelysimilarresult. Verses 56-1 75 tell the storyof Meleager.Thistime there is no parallel version in Homer, but the subject is treated in theheroic style, and there are a number of verbal allusions to Homer. In theselinesthe proportion of Homeric to non-Homeric ornamental epithets is 32 to 22.

In the non-narrativepassageson either side (1-55, 176-200) the proportion sI2 to 29.This bald statistical variation in vocabulary can hardly be fortuitous,and

invites us to look for the same tendency elsewhere. And indeed, the earliestchoral lyric we possess,Alcman's Partheneion,onformsobediently.The poemapparently fell into two parts, the first relating the legend of the slaughterofthe Hippocoontids,the second, by contrast,concernedonly with the immediate

present-the details of the religious dance and the personalitiesof the per-formers. Of the first part only a few half-lines survive; but these contain a

catalogue of the Hippocoontidswho met their death. The name of each hero is

accompanied by an epithet, and of the six that are legible, four are traditionalHomeric epithets for heroes

(rroSK<7•,Kopvarrg, foxos 7j•cOGEVy,'

and•dyas).

The concentration of these four epithets into six lines is all the more strikingin that in the remaining ninety lines of the poem there are only four suchepithets altogether. Moreover, the remaining fragmentsof Alcman (with oneimportant exception) are exceedingly sparing of Homeric epithets, but men-tions of Ajax and Odysseus are accompanied by al&S4oLr,oplSvoi-, and

'raAald'pwv,(frr. 76, 77, 80), a reference to Paris by 'EEcdAcWrLavElpLfr. 73).

The tendency, then, must be recognized; but it still needs explanation.Weare no nearer to understandingthe role of what, begging the question,we are

inclined to call 'cliches', even if they do tend to be concentratedin passagesofheroic narrative; for, if they are really clich6s, their presenceis primafacie noless a stylistic defect here than anywhere else. However, a possiblesolution tothis problem is suggested by a characteristic of Bacchylides' language as awhole. Bacchylides uses a language in which many of the adjectives, though

Homer would have said ioXor (pdwv : 4u0'Wv is post-Homeric. But the effect is no moreoriginal for that.








not actually Homeric, are formed on the Homeric model and give the impres-sion of Homeric compounds while retaining some of the freshnessof original

coinages: 'Homerica epitheta audire tibi videris, si nova illa Bacchylidisauri-

bus percipis.'zThe following three phrases,chosen at random from a few linesof Epinikion 5 (73 ff.), will amply illustrate the point: vevpCvA

yvKuayy~(cf.

vvpEV • ~Ly'faXe 125, AtydOYo0Yos,A•ytSwvoS,tc.), XAKEdK~KpaVOVV(Cf.'9

XOaAKO~aprs465),Tpaxbvo

do.rd.v

rlKpbv.i.O-VHomer). Instances of this kindin

Bacchylidesare numerous; and their purposeis clearly to build up a Homeric'tone' which will pervade the diction without robbing it of its originality.Now it is evident that this Homeric tone is especiallycultivated in passagesofheroic narrative,and that the simplest way to reinforcethis tone is to introducea number of phrases which are infact Homeric and unoriginal. It looks as ifthere existed a poetic convention allowing, in heroic contexts, a judicious ad-

mixture of these unoriginal phrases,which contributedto the desiredHomerictone, and which, in these surroundings,did not have the ring of banality, so

long as a reasonable proportion of unoriginal to original phrases was notexceeded-in the case of Bacchylides,about three to two.

This hypothesismay now be tested on the more exacting material of archaic

lyric. A particularly favourable instance is affordedby Sappho, whose poemsappear to be written either in a 'normal' (purely vernacular) style, or in an'abnormal' style, in which Homeric usages are found side by side with the

vernacular."Only one of these 'abnormal'poems is long enough to be of anyuse to us, but this (j 2 = 55 D = 44 L.P.) is an excellent example of Homeric

'tone'. The scene described is the arrival of Hector and Andromache, newlymarried, in Troy. This has no immediate prototype in Homer (though the

phrase rroAv8wpos~AvSpotLdX-q

Z 394, cf. X 471 f., suggests that it may haveoccurred somewhere in the epic tradition), and although there are certainscenes in Homer which may have suggestedsome of the detailshere (e.g. Priam

harnessinghis wagon 2 265 fif.or the gifts offered to Penelope by the suitorsa 292 ff.) it would be a mistake to regard Sappho'spoem as a mere rewordingof a Homeric prototype: a number of the details are anachronistic or Homeric

times, and the manner of narration is lyric rather than epic. However, the

linguisticlicences

occurringin

nearly everyline leave no doubt of a deliber-

ately created Homeric tone, and the adjectival phrases yield the followinganalysis.

(a) Homeric:

'iXvS~yy/EAoS~Q 292, etc.)KAOS •cO&'rov(I 413, etc.)

E'ACKwltSaAvspopLXavcf A 98; AiKwIwrsAxato&assim)

O(9asl

l.'pas(Iliad,passim)

HIaKlas []v(v)cw (?)4 (iEvSwv once in Homer, v 1o9, with tSara)

lJtvpov rdv'rovHesiod, Th. 107;

&APvpov0wp Homer)

SHermann Buss, De Bacchylide HomeriImitatore (Giessen, 1913), p. 27. On Bacchy-

lides' original coinages see E. Eberhard's re-view of this in B.Ph.W. xxxiv (1914), col.

I225.

SThe distinction is carefully formulated

by D. L. Page, SapphoandAlcaeus,pp. 65 f.Diehl's observation, 'carmen conditum

esse ad exemplum Hectoris AtrrpwovQ 265

ss.)', is misleading. Cf. Page, op. cit., p. 71.* Cf. Z 396 f. r

:vcasyvrac HAdnKc~Aieaa,

64yfl7 VroTAaKrfl (i.e. the birthplace ofAndro-

mache). There is no known place, river, ormonument called HA7Kar4except a Pelas-

gian colony near the Mysian Olympus, a

long way to the north-east of Thebes (which

P







25o A. E. HARVEY

Iol&KLA &G1ptaTGacf. o 416 tvpi' dyov-re c0vp4ara vrt)

irrcirp bAos'(Homer, passim)rrrdXAvpvxopov (A265 e7flrEU.)

aa'rivaLs7T'v•=TpdoXLSapiPa,

4tafa dv"rp.omer, passim)

,KEAOL CEOS-(cf. B 478

,KEAoSAld, etc.)

aXw OcEcala (Io times in Homer)Hcov' EKcL/o3ov (;Ar?dAwveK. passim in Homer)... COELKOLS in Homer, passimof heroes)(?TcrpEVI~avs-' -]o~v'pcv)

(b) Original:

,flpav '4vhpoadxav#flpds irst in Hesiod, fr. 218 Rzach)Alyuara XJ um, apyipa avcpLLa irroripea

aiGAos~8vCsEAgthe compound first here: Homer only once uses 75Svofsound0 64)dAos~

d•yvovthe expression is unique)

drrpa'rov pOwovperhaps cf. I 228sa-3p

drr-parov pya)

Ihdov'EAV'pavElA. only here before fifth century)

This analysis may be regarded as a model for all those from which I havedrawn my statistics, and it will be convenient to draw attention at once tocertain points in its arrangement. It will be observed that no account is takenof the frequencywith which a given phrase occurs in Homer, or the relative

antiquity of the passage in which it occurs; indeed some phrasesare placed

under (a) which either occur first only in Hesiod, or else cannot be foundverbatim in archaic epic at all. Thus TroKLh'dGv'pta-a

as no exact prototypein Homer, c/tuvpo rrdrovr one before Hesiod. The criterion which has beenused is necessarilya little subjective, but not, I think, arbitrary.Our concernisnot so much with the adjectival phrase as an entity, but with the adjective it-

self; and the question is not whether a certain phrase occurs once, or infre-

quently, or frequently, in Homer, but whether the adjective which the poetchose was already a familiar conventional one for objects of a certain class.

Presumablyfar more epic poetry was current in the time of Sappho than wenow

possess; consequentlya

phrasewhich

alreadysounds conventional in

Hesiod or the Homeric hymns should not be regardedas the more original forthe fact that it does not occur in the Iliad or the Odyssey. hus

aetplua-a, hougha rareword, occursin a closelysimilar context in the Odyssey,nd the Homeric

adjective lTOuKdahich goes with it sounds more ornamentalthan descriptive;the phrase is thereforeplaced under (a). Again, it matters little that

JLKA•is

used alwaysof men in Homer, except once, for Briseis(in a HomericHymn it isalso used once of the Muses). The adjective can have had little descriptive

is at the foot of Mt. Ida). According toDicaearchus (Schol. Venet. A Z 396), HAa-

KI was an epithet of Gjfl?, owing toThebes' propinquity to7' 1HAdciKovKaAo-

/LEVOV por7•

AvKlar. Now it seems unlikelythat Sappho had direct geographical know-ledge of the area: it is more probable thatshe picked up the name from Homer, inwhich case there are two possibilities. Either

d.[. .]vdaw conceals a noun, and HIAala is

correctly used as an adjective (this is barelyconceivable); or else Sappho misdividednro-

IraKfln in Z 397 asm1r

F7IaK2, in which case

the last word in the line will be a purely

ornamental epithet for something ('HAaKr)which never existed.

SH. Frkinkel,G.G.N. I924, p. 64, pointsout that in Homer the phrase rar~ip kLAoswould be used in closer syntactical connexionwith the rest of the sentence; but I cannot

agree with him that in Sappho the phrase isfull of meaning. Cf.

,3Awvromjcjwv, Sappho

fr. 27 a 0o.







HOMERIC EPITHETS IN GREEK LYRIC POETRY e21

force for Sappho, and there can be no doubt that she learnt it from epicpoetry.

To return to the analysis itself: it can be seen that in this poem of Sappho,

where both metre and subject invite the use of Homeric language, the pro-portion of conventional to original epithets is about 14:8--closely similar, in

fact, to the proportion in the narrativepassagesof Bacchylides.The techniqueof building up Homeric 'tone' seems to be much the same in both.

By itself, of course, this proportionalconsistencyis neither an explanationnor ajustification of the presenceof thesephrases n a heroic context: it merelysuggests the existence of a poetic convention. There are, however, instanceswhere a more deliberate intention can be discerned in the use of Homeric

epithets. It has already been observed that in Alcman's Partheneionhe sons of

Hippocoon are each furnishedwith a more-or-lessconventional epithet. Now

the sons of Hippocoon were not a distinguishedcompany: it is exceedingly un-likely that they figured in any considerableepic poem, and their importancewas a local one. Is it not conceivable that Alcmanwished to invest them with a

more heroic stature by giving them some of the stockattributesof epic heroes?This would remain a mere guess were it not for a strikingparallel in tragedy.In the Persae28 ff. the chorus is describingthe glorious army of the Persians,and gives a list of the commanders,each againwith an epic (or epic-sounding)

epithet--pTE4pflcppij,n~Tooxdps-, rotoSda•a(rofo4pos- or

rO-dr-•Homer)

dE'OA3jlatos-, in~rwv Aa7T?p aeOdvYS;nd another list of army commanderstowards the end of the

playshows the same

tendency (967ff., cf.

999 TdAoovalpXIILs(LKopE•ooV).

NOWAeschylus is very sparing of Homeric epithets, and itseems highly probable that in these passageshe was deliberately using wordswith epic associations in order that, in the mouth of the chorus, the Persianofficers might be invested with something of the dignity of epic heroes. The

conjecturethat Alcman was doing the same for the sons of Hippocoon finds, tothis extent, support.

However, one must be cautious in assumingthat a poet as early as Alcmanwas so sensitive towards the use of Homeric language: the evidence is simplyinsufficient.By the time ofAnacreon, on the otherhand, there can be no doubt

at all that words were used deliberatelyfor their epic associations. The follow-ing three examples will make this clear.

(a) Anacreon fr. 88:

ITWAErPflK177, 7 1 j tEl AO~dv~jq~wxwL iACovcaa

JVtAEJ~S~~E~YELS,&OKEIS~ &CL

OtjaEV ELISE'LU C7O~OV...

Amid the direct and simple idioms of this light-hearted poem, one is broughtup shortby v1~cAr,a word of uncertainmeaningwhich has almostdisappearedsince Homer. What is it doing in these gay, intimate surroundings?

The history of the word is difficult. It was Schulze' who firstdiscerned thatbesides the word meaning 'pitiless' to which the popular ancient etymology,2(a)v-EAEos,s appropriate, there is another word, derived from *va~EFir, in-evitable' (cf.

lX.Fo/.aL,to avoid), which appearsmost commonly in the phrase

v7AES(/ap

(a euphemism for death). This sense of 'inevitable' explains one ortwo other Homeric phrases as well, notably the v/Ai

8~/~ucof the Doloneia

(K 443-the chain which cannot be avoided, from which there is no release)

SKI. Schr., p. 375, Qu. Epicae,p. 289, fol-lowed by Boisacq, s.v.

2Et. Mag. 603. 26 vYAegscrd ro0 JAos9 TA.

Cf. Hesychius, s.v.







212 A. E. HARVEY

and perhaps even the puzzling j'Ad"rr~t (P 372), 'sleep which one cannotavoid (indefinitely)'; just as the phrase vy&v;rr c) e •pap becomes preg-nantly paradoxical, 'escaping (on that day) the fate which no one can escape

(for ever)'. Both these senses, 'pitiless' as well as 'inevitable', survive afterHomer; the firstappearsin Theognis II25 •Y•r]Aivid'

(of Odysseusslaying the

suitors) and becomes curiously inverted in Sophocles (Ant. 1197 vyAedsma,unpitied corpse); the second in Hesiod, Th. 217 Molpas~ atlKypar lcrAeolrovovs,and elsewhere.But already in Homer there are instancesof a much vaguer use,which suggest that the two original meanings of the word have coalesced into

somethinggeneraland imprecise.For instance, rAEges an epithetofAchilles:sometimes the translation 'pitiless'is apt enough: in H 33, for instance, Patro-clus throwsthe word at Achilles as a reproach, and adds that he must be theson of grey sea and sheer rocks to have a temper so hard (vios

,&nrv). But

'pitiless' does not render the force of I 496 f.:. . . Ax&AE84oacroOvyV C71Ev"o3IS 71aEXpi,

7EEhiS?'Trop EXEwV 7pElrrTOt &d E xat GEOt(WToL.

Here there is a contrast betweenn-rpe~-r-d

and vr/Aie. The gods may be moved

by entreaty: but Achilles is utterly obdurate, stubborn,unyielding.One mightsay the two meanings of vVrAEIS ave run together. Achilles is so pitilessthat nohuman resourceremains for influencing his purpose, which, as a result, is asinevitable as fate. But to translate it here, one must say 'stubborn'or 'ob-durate'. Similarly, only one sense is suitable in T 229, where the Greeksare

told to bury their dead resolutely,-AeaOviwvExolras.This diversityof meanings is clear proof that already in Homer vA•e?swas

losing its sharpness, and that neither of the original senses was any longervividly perceived; and this dilution of its meaning must account for queerphrases like vr~Aeaowpyjvn H. Herm.385 or

•a'pov yr•~"AGqnvin Hesiod (fr.

215 Rz.) : for here only a general sense of 'baneful' is appropriate.Thereafter,except for the Theognis passage already mentioned and a doubtfuloccurrencein Alcman (i 1o D), the word disappearscompletely until it reappearshere inAnacreon. Aeschylus and Pindar use it to express something like 'stubborn','ruthless',Sophoclesfor 'unpitied'.

What does the word mean in Anacreon? Foals are not 'pitiless', nor is'inevitable' in the least appropriate; what they are, above all, is 'stubborn',and we have seen that vrAEjScould mean this, when used ofAchilles,in Homer.But in Homer it could mean several other things besides,and by the end of thesixth century it must have become an archaism and have lost all sharp defini-tion of meaning. Anacreoncould certainly have found a more incisiveword forstubborn; why did this astonishingly neat and careful writer here borrow adead word fromHomer? Surely there is only one possibleexplanation: that hewas deliberately introducing a heroic overtone into the banter of the poem.The verses are a delicate

mockeryof the

girl:Anacreon taunts her

for herchildish shyness, comparing her to a Thracian foal; but not to anyThracianfoal-this one could figure in a heroic setting, it runs away rl;~AE6.The wordmay convey no precisemeaning,' but its associationsare powerful,and give a

SFor another instance of a word losingits precision, compare Aabcvdo.This occursonce in Homer (1' 583) of a whip, in Sapphoof a plant; and its root meaning is doubtlesssomething like 'pliant'. But even in the

archaic period its vagaries are notable.Stesichorus (x9 D) writes fjaSvob 8' irrqhp-7rov arcovrar, Anacreon (x65 B4) uses theword of horses, to mean 'swift', and Ibycus(58 B4)

ta,,vo3E... Av-roi

,-o EJ•,LyEOELr•XCy.








mocksolemnityto the line. The girl is stubbornwith the epic stubbornnessof anAchilles. 'Yousteadfastlylee me'perhapscatchesthedelicatelyironicintonation.

This suggestionseems to be confirmedby the rest of the line. Anacreonwants

to say, 'You think I have no skill'.' But insteadof saying, in the contemporaryidiom, SoKECS '

OZ'oKEtLVUTOdv, he uses the Homeric idiom 8oKE'iSF

O•8EV

del•ivatodv (cf. L.S.J., s.v. *dSEwB. I: 'frequent in Homer with neut. Adj.

to express character or disposition', e.g. aypea olSE, dprnalS7, •lAa El8dES,

etc.), and farther on in the poem he keeps up this mock-heroic tone with hisportentous chariot-metaphor lC1 7ipCpLarapdCov.Discreet allusions to Homerare his method of subtly underlining the raillery.

(b) fr. 5: 4 8'-r

yapc dr'Vr'trov

Alrov--r)vCt•v4,tLjvKo•/7V--

AEVK7)YdP--Kaa/I44cra7L,srpos 8' vAA?)VLYoXarKEc.

In the preceding stanza of this poem each noun is artistically coupled with acolour-adjectiveand the two verbs placed together in the clausula-a miracleof neat construction. In the second stanza (quoted here) the composition isagain vigorous and compact; but once more we are brought up short by theone word

eVKTr•70v.It adds nothing to the sense, and looks simply like a piece

of padding.But then we observe that in four out of the seven cases in which Lesbos is

mentioned in the Iliad and the Odysseyt is coupled with the epithetE3KTLp•V•1.Now for Anacreon this form was metrically inconvenient: it could only have

been accommodated in one central position in glyconics; but EU;KL0OS.,lso aHomeric word,2was easier, and was close enough to the original to make theallusion unmistakable.

It looks, then, as if arnv ydp cir' EAKiTOV /A&0ov is a deliberate echo ofHomer-for it is impossible to think that Anacreon,with his astonishingsenseof style and within the rigorouslyconstrictedspace of these tiny stanzas, didnot choose his words deliberately.But why did he want a Homeric echo? Theanswer, again, is that he wished to introducea note of mock-solemnity.What

he had to convey was the fact that the girlwas a Lesbian-in both sensesof theword. One sense-what we should call the 'pejorative'sense-is conveyed bythe gender of the pronoun in the last line, rpds8'

.AAyv

~r~vdXdaKE.The other,the literal one, ran the riskof soundinga triflebanal. So Anacreon, instead ofsaying simply, 'she comes from Lesbos', uses a more elevated phrase-as wemight say, 'she hails from Lesbos'; by so doing he raisesthe girl for a momentout of her contemporarysetting into a more dignifiedworld. For a moment sheassumesthe proportionsof a personagefromheroic times, to whom one mightask the question, 7rs rderv

rE•vavSp6v; and attach more than a passing signifi-

cance to the reply, 'I hail from finely-built Lesbos'.3

(c) fr. 51:

acrol8aPAI'

dEpo-aFoi(

KnPdO•ArOAOp.o oo(45.omer has tro'rap.oio KXLpdOLO (E 4I) and Es wroradpovrpogod&sE 453) . AlirlEL

I cannot agree with Wilamowitz and

Diehl that p' = oio. The following linesmust be understood as an answer to the im-putation that he, Anacreon, is not suffi-ciently skilful. The point is made by B. Snellin Philologus,xcvi (x944), PP. 285 f.

2 B 592.,KT(TOv

Airr.Cf. Soph. El. 707

,va7ror49O7vcv

"rcv

OEoS1jirwvitro. Here the whole passage-the

list of competitors in the Pythian games-istinged with epic vocabulary.







214 A. E. HARVEY

moreover implies something of a heroic effort. These Homerisms are in sharpcontrast with Anacreon'susual style. What is the explanation?

There is no trace of any tendency in Anacreon to give straightforward

descriptionsof battles. This line must therefore be taken as a jest in the tradi-tion of Archilochus (fr. 6) and Alcaeus (fr. 49), who preferred o throw awaytheir shieldsand win immortality through the life of a poet rather than throughthe death of a hero. And how doesAnacreon add salt to hisjest?By the simpleexpedient we are beginning to recognize in his poems: he gives the line amock-heroic tone by using a row of purely Homeric expressions.

This deliberate use of words for their epic associationsis by no means con-fined to Anacreon, but is a recognizable feature of Greekpoetry from the endof the sixth century onwards. There are numerous instances in tragedy. The

word /zEpowrEswas presumably no more comprehensible to Aeschylus than it isto us; yet this detractsnothing from its power when, in describingthe mysteryof the ways of god towards man, he writes the phrase (Hik. 90) tEpdorEacn

Aors:the sad heroic notion of the blind mortality of man resonates in this singleadjective. Or considerthe fine irony which Sophocles achieveswith the phrase(Ajax 374 f.) iv 8'

hAl•crKcYlova• Ka•

KV-Lroi '7ToEUV•roAoLS.Ajax's prowess is of

heroic quality, to be described in epic terms, even when the exploit is only ademented attack on-a flock of sheep. These instances could be multiplied:the technique is established.

Whether it also appears earlier than Anacreon is a more delicate question,

and the evidenceis not decisive; but there is at leastone passagein Archilochuswhich seems to exploit the associations of epic language in a similarway.Consider fr. I Iz2:

'oAA7vKar" cX~;•v OJpLLC-TWOVXE•E•

KAlaS~ E)KcrrTq&'wV~Tra.dS~bPE'vaS.

Each line has its prototype in Homer:

(i) ?433 Aaaltv tir5dyao~rp' ~AvaOElsof Odysseus under the ram);(ii) Y 321

Ka•-'GbO•Av XEEVdXAiv (Poseidon's intervention against

Achilles);(iii) A I15 diraAdv Eo 'f rop d&rrqypaofa lion attacking a stag).

The verbal echoes are so close as to be unmistakable, but the difference intreatment is critical. In Homer the subjects of the phrases are Odysseus,Poseidon, and a lion; in Archilochus exactly the same actions are attributedto-an emotion! This is not merely a new and bold piece of psychologicaldescription:it is a piece of literarybrilliancewhich dependsfor its effect on theadaptation of familiar Homeric phrases to a totally new situation.'

By the end of the sixth century, therefore, and probablya great deal earlier,the lyric poets were fully conscious of the conventional associationsof Homeric

diction. Consequentlyit seems unlikely that they will have used purely orna-mental Homeric epithets indiscriminately, and it is reasonable to think thatsome sort of poetic convention was responsible for the concentration of theseapparent cliches in certaincontexts. One such context (the narrationof an epic

Fr. Io04 XaAei•r&e . . .dcs vycvaw 7TEwrap-

levor80' darrEwv is a striking combination of

the literal and figurative uses of rdlpw inHomer, and might be added as a further

example. Archilochus' particular interest inHomeric words is discernible in 54, 56, I 6D, i86 Bs. See also Hauvette, Archiloque,pp.269-72.








story) has already been sufficiently commented on. Another very noticeableone is the invocation of a god in a hymn. This observation s promptedparticu-larly by Alcaeus' Hymn to the Dioscuri (78 D = B2 L.P.) which, in the short

space of twelve lines, contains no less than six Homeric epithets ([7rayjSe

l?0]•/•Eo, •vp'qavXOova,OKv7ro3wv ITI LT7TW1V,EVCaT&)y7v VCLWV,

Va( L•EAaW(~,pyaA4

VKT&) as against one more-or-less original epithet(Oavdrow

axpvdevros, cf.

AflosKPVdESin Homer). Now Alcaeus is the most careless of all the lyric poetsin his use of Homeric epithets, and this poem in any case is not one of his best(Wilamowitz regarded it as frankly inferior to the 'Homeric hymn' to theDioscuri); consequently the remarkableproportionof unoriginal to originalepithets may simply be a piece of carelessness. But the incidence of rAwovdiKeWv,

OWV,•VdK7)SKVUpdS, HcqOV 'rrTEppplrav, and

McakaAlyeLan invocations of

Alcman, of

aly7dXo

and K

arKO/LOLMoLaaL n invocations of Sappho, of

rar"LEydAowlos in another hymn of Alcaeus, and of•,•afr7

A E avO) IraTAdLoSand diypywvOtlppv n an invocation of Anacreon, suggests that the propersolemnity requiredfor addressinga god was achieved, as often as not, by a freeuse of Homeric epithets. This is not just a matter of the epithetsapplied to thedivinities themselves.These, in lyric as elsewhere,are fairlystereotypedin anycontext,' and original coinages (such as Sappho's flpa&&vav' A4bposlrv) areless common than traditional attributes. What is noticeable in the context of

hymns is the prevalence of Homeric phrases in general: the substantives,whatever they are, tend to be qualifiedby conventionalepithets. The analysesbelow will

amplyillustrate

this;here it

is merely worth adding that thisconcentration appears to be neither accidental nor confined to lyric poetry.The great prayer in the parodos of the OedipusTyrannus(151-255) contains, inthe space of sixty lines, some ten expressionsof a more-or-less conventionalcharacter. In the rest of the play there are barely any at all?.

Another area in which thesephrasesarecommon,particularly n the Lesbianpoets, is the description of nature. The unusual poem of Alcman (fr. 58),describing the Sleep of Nature, has four of them in seven lines (pdpwvKopvcal,,MA'twayada, E'v EVvOCETropqupas Aotds, lwvcv

TavvTrEp•ywOv)as well as several

other Homerisms; and the vivid metaphor in fr. 59

'Plras~ pos~cv'oov iAaVVKTSo 1LEAwa•S CTrjpVOV

is built out of two common Homeric phrases. In Sappho descriptions of, or

comparisons with, nature contain the following:8•mp

bXpov, AEL'LWv~3do7oroT,KAO& WKEES cTrpOL'OL, rTKVa 7Tpa,3c KdAav aTElvvv. ..

&pyvpiav, 'bEwov Ei OS (Of

the moon), OJaaoav c~Apv'pav,vOe~wvptOaGAwv.This concentration, again, can

hardly be accidental, and it is not very difficult to suggest an explanation.The descriptionof natural scenes demandsconsiderable inguistic resources;itis not as simple as writing love poems or political verses and cries out for an

elaborate poetic diction. Only one ready-made poetic diction was availableto the lyric poets, that of Homer; and it is perhaps not surprisingthat theydrew upon it more liberally than usual when faced with the task of describingthe beauties of nature.

There are two more limited instances which it may be convenient (evenI have therefore allowed a separate

category for them in the analyses below.a

Soph. Ant. I115 f. is another example.

3The context of these four is also that of a

hymn.Here, as elsewhere, some overlappingof the categories is inevitable, but immaterial.







216 A. E. HARVEY

though the reasonsfor them are probably quite different) to regardas specialcases of the descriptionof nature. They concern the sea and the earth. First,the sea. In archaic poetry it is remarkably rare to find the sea referred to

without either an elaborate periphrasisor an ornate adjectival phrase. In thewhole Theognis collection, which is not unduly free with ornamentalepithets,I know of pnly one instance where 'the sea' is mentioned without at least a

qualifying adjective. Solon, a tidy writer, confronts us unexpectedly with

Irdvrov,roAvKVlzovosTpY,•TOLOfr. I. 19) and Archilochustwice fills halfa line

with Homeric arabesques: fr. 7. 3 Kard KOvLaroAVUAolfloLo aAro-rjsand fr.

I2. I,roA~e

cAdS vrTAd)EyaarvxThese instances are all from elegiac poems;but the tendency in lyrics is equally marked. The sea occurstwice in Alcman,with the phrasesrropvup'aCrahdsnd E'l KtzLaros~vOos'; t occurs twice also in

Sappho, with

Odaaaoav

E'

dl\.ul'pav

and

J(vpov rd'rov; of the nine mentions of

the sea in Alcaeus, six involve ornamental expressions;and none of the threeinstances in Anacreon is unadorned. Out of all these phrasesonly one shows

any originality,Alcman's w Kv'laros~cvOo"0:he rest are purely conventional.We can hope for no explanation of this strange phenomenon; but conventional

descriptionsof the sea seem to be an integral element in the poetic diction ofthe time, and we cannot assume that they produced a banal effect. We must

simply accept this as a literary convention of the period.The second case is in some ways even more mystifying: I mean the continual

recurrence of the phrase 7y pL'aeva(or, less frequently, x0&v 2AaLva)-more

mystifyingbecause it is farfrom clear what the phrasemeansorwhy, alreadyinHomer, it became a commonplace. The commonest application of the word

pAas~ n Homer is to ships-not, probably, because they were treated withblack pitch2 (for they were sometimes

tAzro?dP70)but because, from a dis-

tance, they look black. In the same way Apollonius Rhodius explains that

Corcyrawas called 'Black' (M\~cva) because, seen from the sea with its darkforests, it looksblack (4. 567-9). And indeed Aristotle's account of blackness

(De coloribus,. I) treats the word ~dAasas if it simply denoted deep shadow.But this clearly will not help for

y7 2'AaLva.The Greeks did not always thinkof the land as seen from the sea, and if one is on land the landscape, particu-

larly in Greece, appearsanything but black. Perhaps, therefore,the colour ofrich soil is intended by dAas-.This seems to suit phrases such as 4EAa'vf

KErLEvov'd'lXiOov (Arch. 58. 2) and especially the remarkable line in Theognis

(878) ... dy& OOav&vaz tlhAawv'~EopaL.his applicationof dAas since thenotion of 'pitch-black's surely strange to the archaic period) can perhaps bejustified by the sense of opacitywhich it has in connexion with water, whetherruffled by the wind (cf. qpit pdAawva)or deep and still (Theognis 959 KP'pvLEadvvspor--the water beforet has been stirred up with mud).3But if Sapphocan write (27a D), 'Some say that a fleet of ships is the fairest thing on the

black arth', he clearly cannot have been thinking of the soilwhen she used the

expression .teaava yata. I am inclined to think that there must at some stagehave been some deep religious association behind the word,4 which was

SI cannot believe that in this verse

EvfrAOKaC•oulso goes with dadr.The name of

some divinity must have preceded.2

Hesych., s.v. yAavwavqe]

s aZ faOEae xa

rlacodXploro0.

SCf. Suidas, s.v.[AcvvSpoo*

paOLEra

KUplow S, KaOdpov tSaros.

SThere may be traces of this in the mythi-cal and vague personage MdAawva Paus. I0o.6. 4) who seems to have chthonic associa-tions, and in the cult-title pi.Aawa

Atjlnjr'qpin

Phigalaea (Paus. 8. 5. 8; 8. 42. x). Cf. O.Gruppe, Gr. Mythologie, p. 103 n. Io, alsoSchol. Eur. Or. o094, Callim., fr. 52 Pf.








doubtless forgotten even by the time of Homer, but which continued to makethe adjective a regular concomitant of the word ys, even when the earthwas personifiedas a goddess (Solon fr. 24. 4 f.-):

/lrj77l /tryvarq 8aqimvwv 'O~vCLrtwv

?p~rray1j7CAatL'(

....

But it is probablyuseless to probefurther or an explanation; it is sufficienthereto note that the phrase occurs so frequently in lyric poetry (particularly in

Sappho) that it must once again be regardedas a properelement in the poeticdiction of the time. By contrast it does not occuronce in Aeschylusor Sophocles.They, perhaps, found it banal: the lyric poets evidently did not.'

These three categories (heroic contexts, hymns, nature) will be found toabsorb by far the greater part of the Homeric epithets in the lyric poetry

I have chosen to examine, and those that remain over are relativelyfew. Thisremainder cannot usefully be discussedin general terms; nevertheless,before

proceeding to the detailed analyses, there are a few individual cases to be con-sideredand certain qualificationsto be introducedwhich affect the compilationof the analyses and still further reduce the number of unexplained epithets.

(a) The German word Lieblingswiirters the most convenient label for thefirstqualification to be made. Every period in the historyof poetry has certainfavourite words of affection and appraisalwhich occur with the frequency ofclich6sbut are yet such an integral part of the contemporarypoetic idiom thatit would be a mistake to regard them as trite or dead. 'Soote' in Chaucer,

'doux' in Ronsard, 'fair' in Elizabethan lyrics, are all instances of Lieblings-wirterof the period; and Greek lyric is particularlyrich in words of this kind.

Mostly, however, they are words of a slightly sensuous connotation (dflpds,rraA~ds,/LpEPdE,tc.) and do not occur in the same contexts in Homer; more-over the fact that Homer uses EparTEwosnd the lyric poets the slightly different

(though metrically no more tractable) formOpa-o's

which only occurs once in

Homer) indicates clearly that these Lieblingswirterere very much alive andwere in no sense a loan from an already existing poetic vocabulary. In the mainthese words (aflpo, &yavds, parods,

.poELS,~cfIEpo, .LEpdELSs,Eprdo, Xaples, are

the principalones) either do not occurin Homer or are used therein a different

way,2 and consequently do not concern us; but it would be rash to concludethat there are no instances of words which are favourites both to Homer and

to the lyric poets. I suspect that one instance of this is the word ALyvpdo's,hich,togetherwith the compound AtydwvoS, accounts forfourprima-facieclich6sin

Sappho and seems to have been her favourite word for characterizingsounds.

(b) Certain words for colour, pattern, material, etc. (e.g. zav6ds,ropblpeoS,

TrroLrKlAos,ptato), occur rather frequently in both Homer and the lyric poets,and there is a temptation to rank them among the 'dead' expressions.But thiswould probablybe a mistake. The visualsenseof the lyric poetswas strong,andthe

objects theydescribe are often rendered more vivid

bythese

epithets.Moreover,words for colour and brilliance were a standardresourcefor addinglustre to a scene. When Alcaeus writes Adc4rpatKvviat, or Alcman TroLKL2o

' At least until the fifth century. Pindar(0. 9. 50) and Bacchylides (i3. I53) eachonly use the phrase once, and rrlAMavrovKVaviaGOovs

a-rpovPindarfr.33b Sn.) looks

like a conscious elaboration on it.

a Cf. Max Treu, 'Von Homerzur Lyrik'(Zetemata, xii [x955]), PP. 175 if., for avaluable discussion of the lyric poets' use ofthese words.







218 A. E. HARVEY

8pdK(JV rrayxp;aLO,these poets are using words to a deliberate effect: it is easyto distinguish this from the degenerate use of lustre-wordswhich can be ob-served in a purely conventional phrase like 4'Aad, 'iry (Theognis 985). Par-

ticularlystrikingis the sensibility towardscolour shown by Anacreon,who wasfond of placing two or more contrasting colours in close apposition, e.g.

fr. 2 KalNvz'qa&KvavWTL8Esm

1Topc vpEl'7' 'r Aqgposr?

fr. 5oaailpq

SEI-CLErropdbvpl7"

cdo.,wvypVyOKdyqSEpWE

srVyL TOLKLAOOc~fCflaACp

fr. 53 ["EpoW].. . tp' icrVc YEvELovinro'rrdAov XPv(o9adVYWWv -rTEpty/Wv a~d7aE-

TrapcLTrIrE-rat.

In this last example the juxtaposition of colours is particularly forceful:3'greying heek... gold-shiningwings'; and in general this acute colour-sensi-bility accountsfor a numberofexpressionswhichwe mightotherwisebe temptedto regard as purely ornamental. When Anacreon writes of the 'dark-shielded

Ialysians' ('IlqAvulovsobs Kvavaaurrl'afr. I5), of a girl 'with golden raiment'

(XpUVo'rreTrAE~ovp7 fT.91), and of 'golden-helmeted Athene' (XpvuaohdovHIIa-Ad8osOx. Pap. 23216) he clearly intended to convey vivid visual images.Thereis nothing conventional or banal about the language in which he did so.

(c) Anacreon, in fact, was a consummate stylist, and this may be the placeto mention another of his characteristic devices. He is fond of buildingperiodsin such a way that each noun is attended by an adjective, forming a kind ofsyntactical symmetry through successive verses. The opening of fr. 5 (quotedabove) is a good example; the following is another:

fr. 44 7roALol.vLtAv

87 Kpd7oa'O,ap77TEAEVKdV,

Xapka 8 OVKE7) 7Tpa,YlPa..AEO& 8o

)8ES,

yAvKEpOo 8'O"K 7"

.OM$S

LdrOUpdoOSETrrat.

Here each noun has its adjective. Four of the six--ioAol,AEvKO•dV,

7paEos,

r'oMd's---aressential to the sense; but the

remaining two, XapEraaaand

yAvKE-po3, are purely ornamental,and are, moreover, unoriginal loans from Homer.Yet their presence is justified--if justification is necessary-by the formalstructureof the lines. Each of the two nouns had o have an attendantadjective;but the sensewas alreadyfull, there was no roomfor a new idea, and a stronglyoriginal expression would have overloaded the period. These two slightlycolourlesswords were all that the verses could take.

(d)Itisimpossibleforusnow to graspthe associationscarriedby theword lEpdS,

SThis is probably not an indication of a

particular colour so much as a suggestion of

rich, varied colour in general, cf. Weber,Anacreontea,. 65, L. Deroy in ?t. classiques,xvi (1948), pp. 3 f. B. Marzullo even argues(Maia, 1950, pp. 132 ff.) that

rropopcdais an

old cult-title forAphrodite meaning 'marina'.3 The ball of the Phaeacian dancers in

o 372 wasnopuvpri

(= versicolor ?), andthis may also have been in Anacreon's mind.Eustathius, ad loc., says aprvvvwv (7r rrop-#uvplv)

:rdyE":it is more the brilliance of the

ball than its colour which is suggested, thefact that it was coloured rather than its

specific hue.3 Aeschylus, it has long been believed(Schol. P.V. I28), learnt the ionic metre fromAnacreon: did he also learn this trick withcolours from the same poet? Cf. Hik. 529 f.

Al*vY8" ElslaAE 7rop upoLL TdrvsEAavd•uy.rav, Eum. I81 if.

/s• KL,Aa Ol

orrr9r vv

apy7)077,VOLvw,xpUacA7roP u Odw/LC

orioptdL-

ucvov,

fdv.i;ir'

)Ayoury4Aav' d"'

dvOprorrwvd pdv.








a favourite word of the lyric poets and one probably rich in meaning. Whenit doesnot mean ceremonially 'sacred'(asin

fpas• AoAt'asAlcaeus G2. 35 L.P.,cf. JAoAvypvLEpo'v . Sept. 268) they use it mainly with some natural scene or

object: e.g. iAmropcvposlap ~pv~sAlcm. 94, lEp~rvvK7,-r'sStes. 6, cpav ~pdvasXKvdvos Sim. 20o--andthey doubtlessconvey a good deal by it (for this use ispost-Homeric). On the other hand, the wordseems to have none of this powerwhen used, as in Homer, of cities.Phrases ike O(94flas l'pas (Sappho 55. 6),"IAeov pav (Alcaeus 74. 4), BafldAwovos-as (Alcaeus 82. Io),

ZEpaYpdt' l4avaiv

(Timocreon I. 3, cf. Soph. Aj. 1221I), have the ring of clich6s, and I have listedthem accordingly.

The principles by which the following analysesare composedshould by nowbe clear. The first four categories in each correspondwith the main pointsdiscussed in this article; all instances occurringin dactylic poems are placedtogether; and one or two special categories are added for individual poetswhere this seemsjustified by what has alreadybeen said. Square brackets in-dicate that the same instance has been listed under more than one category:some overlap is inevitable, and it is not always clear which of two possiblecategories is the correct one. The brackets are simply there to show my per-sonal opinion and to prevent the phrasefrom being counted twice. A questionmark indicates doubt about the text, the attribution, or the context. All re-ferences are to Diehl, unless qualified by B (= Bergk, P.L.G.4) or L.P.

( = Lobel-Page, Poetarum esbiorumragmenta). have not normally thought it

necessaryto

give references to lines within the fragments.Archilochus

A. (i) Heroic context: None.

(ii) Hymnodic: None.

(iii) Divinities: ApEU /LLaLY Vov (31), iraS bputnK ov Aos (51), ?fnrrLo

HoLELscv (i17), dAh•TppoSdyv4 ( 119).

(iv) Nature: (a) The sea fla0B0 rvros~56), 0acaaors 4XC'6raKt/Gara

(74).(b)

i/EhaCv dl x0ov0 (58).(v) Dactylic poems: (12 in about 40 lines) 0a4ELal tEvSdvaL, IoA'orlovov

i~pyov(cf. "EpLST. A 73, etc.), SEo'rrra 80ovplKvrAO(3), O VI~dS,otvov dpvOpdv (5 A), &Trosc&6~tiV-ov

cf. ALLt;ovL•r•oO 463) (6),

K,?EaoTovdoVra,

,roAv#AolaLPoLoOALoa7s, aV77KEC0oLLKaKOO(7),KaGapOLcL V EV E/ILaoULv IO), 2UTAOK4LOU, 7TOALS cLAOS

EYV7rE aYEc?cTL,

yAvKEpOV VdTOV (I 2).

dv0pWhroLUL.. ~wvio'u'occurs in a deliberate paraphrase of Homer

(68). It is difficult to know whether AVaqLEA•ir0o0S

(118) was a con-

ventional expressionby this time or not.

B. Unexplained: V#NEGoat (twice: 56 A, Pap. Ox. 213I3), dKElr•UL v]7)vu

(Marmor PariunmBI

35, v. Philol. xcix, 1955, PP. 4 ff.), AaLPd& 7r-Epd(92 b), ? Op7)KES aKpOKOLOL ([79 a]), (4fl'vayAa4v (Pap. Ox. 23Io0. 35)*

Alcman

A. (i) Heroic context: 0ro8•K7, Kopvu-cv, foxovo7r••E.,WY,

plCyvY (I), bal8LLoS

A'fa•(76), Swpl 5vuar (77),

'OmvuaoosraAaualpovos (80), 'EAAcLS

WI-LavEdpa(73)-(ii) Hymnodic: rdAwvWKEOWV,-AV86EKTSKvSpdS2), llcov TrEp~Lppt;Tav35).







220 A. E. HARVEY

(iii) Divinities:M.ca&KpoKdoTIAo (34*).

(iv) Nature: oiwvv Tavvywrep4ywv58), opos, vOdovAa,VvK7,c9ptaiva•

(59).(a) The sea ropvpda&sdcAd 58).

(b)5c'acva

ya&a 58).(v) Dactylic oem: otca e TroLILEVESv pE GXovUwe37).

(vi) ?Lieblingswirter: jaa Alyta (7),,AlytEa2p,4v (io).

(vii) Doubtful: A4p'ra 0,ESjS, &

KaA'aJV9poS~ AyqucoXdpa,Ciriovrraydv dEOAd-copov (a simile for Hagesichora) (I),' ?vrAhE avayKc (IIo),

?dydpwXoc 122 B).2

B. Unexplained:dnpyw 2EpdCrvas7), Zap&lwvir' bKpevI3), rr3p CFLov57).

Sappho

A. (i) Heroic ontext:~IoAwvoK'WV (27a), ]roC lacdA'qTES28), fourteenin 55 (seeabove, pp. 209 f.), &uXarayais,PpodrraxvvAitwov65 A).

(ii) Hymnodic: avovcyvov, [ $wptbXpov],[AElWvIidrfloros.] (2 L.P. = 5D.), atyodXw(86 L.P.),

Ka.AKot1.oMoiuaa (90).

(iii) Divinities:XpvwoodCav''qpds&ra9), [cHdov' ~KfoAov5], d~yvaXdcpres(Io3 L.P.),

poSorrc0.XEEsya XdpTES (57), XCpcES,pLdKapaL80),

Xpvuoir8,&os (1o3 L.P.).

(iv) Nature: Koc WKEES~aTpOi-oOL,rlKVa TT'pa

I), E~dvvovE8o0scf. O 555)(4), jBpo808oKrvUos njva, G&GXauavi' dczpav, 1OToavG4Loos acpov-pace(cf. loAvavijs &353) (98).

(a) The sea [•chAvpovrruvrov55)],[G.Aauuav

Er'' cAp5pav98)].(b) ya tCLa

•va(three times: I. Io, 27a 2, (?) 31. 6).

(v) Dactylicpoems:7TolLEYESvepES,xd#zac

4 rer- r-p/VPov

civOos(i i7).(vi) Lieblingswiirter:&

A•yywrvos(30 L.P.), Acypav

XEAVtwav(65 A), Acyipav

[doalsav(io3 L.P.), irvos . . . yAV'KVSGes 67), Aos~- yA'KEPiov,Alyvpua' ci [ (70), yA'K- lla7TEpI 14).

(vii) Stylistically accountableor:ITolA&O•z •icaar] (I 7), S&fvL

EAdva ElCTK7V

(35), cav'pwv VEKULOv (58), AwrivoLS SpouOEYTaS OX0OLS4XPO'TOS

(97), Xpvloiwe dvOdt4LowLv152). The unoriginalityof the wordsdoes not detract from the

visual effectivenessof these expressions.(viii) Doubtful:"EposBAvauqJjs.(see under Archilochusi) (i37).B. Unexplained:dvav A4ypav (25), av8GEOvpclaAdwc (98 L.P.), p dyav dE

"OA[vtywrov27 L.P.).

Alcaeus

A. (i) Heroic ontext:"IAoVLpay,, Aglov~'v:av Aa'r[pa I•dAwv5s74), 'EAbs-a

Apyedas, evadrpwrovAdXo, 'AK'wnrTsNI, L.P.), vittdav waAhlav76),[Kdr orvola . .] (QI, L.P.), [pdEavas. o'vos] (73).

(ii) Hymnnodic:r ,LEyaAOAlto (i),

(Q o,(qEST ,S

PavEOva, ,K,,1Jd•,WVxIt is possible that these heroic phrases

are used of the girls deliberately, as in thepassages of Anacreon discussed above, pp.2II ff.

Alcman and Alcaeus used this word

differently, Eust. II. 314. 41.See above, p. 217.favOd' may have implied exceptional

beauty as well as colour: Eust. in 11.432. 27ewasvor 8S idCr wapd

TOoLwa7alo(S rd

favOo'v.

s Alcaeus probably understood 11 x48-9rc S cal

Az"'rol7SwvliTayE

•Vy'viKEar

rrrov •Edv0Ooval BaAlov,r d•LcamwolljorrEriaOrlvs giving the colour, not thenames, of Achilles' horses.








IT rTwOV,•evt ,yWV

vav,vat •

tEalvaW,apya\Ea CvVV'KT 78-79 =

B 2, L.P.), d0avd-rwvaKdpwv,Ed8e~ovctvoror,' cpyaAdas~#yar(G I, L.P.).

(iii) Divinities:E•ir',AAo~

'Ipe, XPVaoKo/d Ze4ypc (8), Alos i ar'ydxw (11),

dcyva,Xcp,'reS12), ?MolacrvyAa[(T I. ii, L.P.)

(iv) .Nature2 edtwSeoa[. yA]avt;av DXpov6wpd/•Edoeaa[

(F I, L.P.), 4pordveyo (98), cf. yis cdir ieppCrWV(I35).

(a) The sea ALvpovwdro'vrov5I), wrop plav OdAaciav 77), rwoAilar

Kv.L' AoS(Io9/II o, 27),Ka- ofvowa... (Q I, L.P.).

(b)t'AaevaxOdv(twice: 73, G 2. 29, L.P.).

(v) Stylistically ccountableor:Arfla3Se AKecire•r•oL, XO OEo(Tara

(G 2.

32 ff., L.P.): that these two Homeric epithets were used deliber-

ately is suggestedby the presence n the same lines (which describe

a ritual beauty contest) of the Homeric verb rwAev and thetmesis ?repi84flpitue.But we can hardly guess at the reason.

(vi) Doubtful:flA~Xpcodiv4jtwv322), ]rov E/A\LSea (35. 25), (probably used

as a substantive,cf. 98)dyeppwxo••4

Z 79, L.P.), rpave wdrAw? vel

"Ipav)(42).

B. Unexplained:way Cpav 83), Bap'Awovorpas(82), v&"2peAalvw46 A),

<bKv.[.. ~. vaa (?d KvdAovs (H 28. 20 L.P.),sowdaevr'A4poVa

(twicein 73), voflo'dpwAvar~(43), otvov AaOucdSea(96).

AnacreonA. (i) Heroiccontext:None.

(ii) Hymnodic:NiCya&,KvavdLnLCS2), JAanL1dhAeavOi ira Zdo's~,ypoiv

OrPpaYI).(iii) Divinities:KaAAlCopO

Koopa,dLs (63).

(iv) Nature: vefipdvveoGqAtayaAaO'tvrv(39) (Homer uses yaaOqvods and

(a) The sea iroAbvKta (I7), K/Lc rrnVwinov,xOvoevPap. Ox. 2321•),? iwo'v O iojra iopfvpdotcU

KLc•zaLt? Anacr. Pap. Ox. 239221.

I 7).

(v) Dactylicpoemss(6 in 23 lines): wo'eptov aKpvo'irca,dyAaa 8S6pa 96),7arp&S0sabds, jypd adta'ra (102), dyhawv KdcJLov 107), S&aaxiorE,

rrowiolo (III).(vi) Stylistically accountable or: eiicd-rov Adoiov (5-v. p. 213 above),

7o-raLo0KaAApdoV5i -v. pp. 213 f. above), xapleova (y,' yAvKEpOD

fld-rov (44--v. p. 2 i8 above).Colours: a lapy rrop#vpEl (5), 'I'7valove -robs KvavrdcrrnSasI5--cf.

AEVKanornSX 294), favOfEipVItndA(16),XpvadcreirAE

KO;P1 (91),

aIt is possiblethat a secondsubstantive

stood in v. 2 and went with ei~vscAov.nHomer thisadjective s used only of islands;but Alcaeus' use of it is not much moreoriginal for that. KvsaAl4cvOiov,however,has a certainfreshness v. 6).

a Cf. also sub 'Incertumutrius auctoris'

L.P.: LSKWQrtTovIO), Trda?riptv 9v0o16).SM. Leumann, Homerische Worter, p. 340,

arguesthat this phrasederivesfroma mis-

division of dPAixwpovdcv•L•v

n some passageof the Homeric poems now lost.

4 See above, p. 220o, n. 2.s Out of conservatism, I attribute no more

than the following elegiac fragments to Ana-creon: 96-99, Io1, Io0, Io7, Io8, IIo-I12.Cf. Weber, pp. 31 ff., Wilamowitz, S.u.S., p.1o7, C. A. Trypanis in C.Q. xlv (1951), pp.3' ff.







222 A. E. HARVEY

?1dAawvavdveV? Anacr. Pap. Ox. 23221. 6).Ironic:' 7rp4yaLLEALsrla21).

B. Unexplained:opv .orv

(25),,OdAeBcavop",v

(37), ?Z?

v 1EPE#7 8opor

(? Anacr. 7).

Further exegesis is unnecessary.The number of conventionalphraseswhichcannot be accommodated in these few and clearly defined categories is verysmall indeed--only in Alcaeus do they exceed five; so that, even ifa substantial

margin of error is allowed for (and I make no claim to rigorousobjectivity)these analysesseem to sustain the conclusion that ornamentalHomericepithetswere not used indiscriminately by the lyric poets. Those phraseswhich remain

'unexplained' either have overtones which we can no longer hear, or else are

simplyclich6s, and the

productof careless

writing.It may be useful, in conclusion, to compare these analyses with that ofsome poetry of which the date is uncertain. First, the Polycratespoem attri-buted to Ibycus (Pap. Ox. I790= 3 D). This, though it containssome eighteenunoriginal epithets in forty-two lines, shows no great statistical divergencefrom, say, a narrative passage in Bacchylides: the proportionof Homeric tonon-Homeric epithets is again about 3 to 2. But the phrasesthemselves,whichfurnish thisstatistic,are somewhat peculiar. In the firstplace, there is a kind ofentassementf conventional epithets which is unlike anything else in archaicGreek. For instance, Homer commonly writes HptdctosoAap3av43aoand also

darv UdyaHp~d4Loeo:here the poet runs the two together into AapSavliaHpetdolo

tdy'd&rv,and then, not content with this, he adds rrIEKAEE'Ss an

equivalent to Homer'srrepLKAVUrrV

~rv,and throws in Afleovorgood measure.The resultant rigmarole,AapSavi&lapudpLoeody' arlv rrEpUK(AESoA8Lov,

as

little to recommend it. Or again, Homer has both d'yar EAadveosAil'as nd

TE;aaudvoso•AKLQO

AarS:this poet rolls the two expressions into one with

t'yas TeAaivdVosaKLCOSAiar. In the second place, the phraseswhich are notborrowed directly from Homer are either trite or infelicitous. For triteness,consider(a) Sipw rroAdvftvov

EYXoV7E•:Homer has 8ipw gXELV,roAiv;tvoSccursin

H. Hom. 26. 7; while the phraseas a whole is simply a periphrasis or 'fighting',

and is further amplified by lrAELOVKaCa 8aKpVdEV'a.b) ~EVwadravHadpw:the word is alreadyin Alcaeus (N I. 5, L.P.), the notion a commonplaceby thefifth century. For infelicity consider the only two phraseswhich are original,

iAdcnpqovtpap divvvyov and lerEpd(avov dpEr7v: both these contain ambi-

guities. avavvyornormallymeans, not 'unnameable' (as here) but 'unnamed',i.e. unknown- a most undesirable overtone in thispassage;and tirepdbavosiselsewhere always pejorative ntil Plato. It may have been permissiblefor thepoet to pressthese wordsinto new uses; but he can hardlyhave preventedthe

SAnacreon fr. 21:

<c) MEy'rrijsg8'

fL#Adcpwv SiKa

&7bCLpVE

arGvoTrral ray&E Ka& 7ptya rtvil CtL,-

,16.a.Cynulcus (Athen. 67I f) comments c ydp

-rmyMyovwCicavos ciroroi-and indeed thequestion much exercised the grammarians.

Hephaestion even wrote a book called UepLTroO iap" 4vaKploVrL Avylvov ao7EdCvov, and

many learned theories were advanced to ex-

plain why Megistes wore such an 'odd' gar-

land. But surely it is just as odd that heshould have drunk rpc5, which means eithernew, unfermented wine, or the dregs, thelees of wine; neither of which could bedescribed as

FLEA•71y8.here is one perfectly

simple answer: Anacreon is mockingMegistesfor wearing a rustic wreath (Tivapo- S8adypOlKwv ELVaL AEyEC GTTE?avwLa vjY A'OV

Athen. 1.c.) and drinking cheap raw wine.

cAd,#pwvand

•LEAL77S3rare used with irony.








old meanings from being present to the hearers'minds. And in these passagesthe old meanings are disastrous.

The genuine poems of Ibycus are unfortunately too scanty to provide a

reliable standard of comparison.However, one may observethat his adjectivalphrases,though fulsome, are original: his descriptionof nature in spring (fr.6),for instance, contains no clich6s. In the fragmentswith certainty attributed to

him, the Homeric epithets are distributed as follows. Three qualify heroic

personages: AEvK&TTOVSdpoVSOf the Molionidae (2),yAavKm•TrTaacadvSpav

( 6), and ? dvotaKrUv "Opqyv 7); the simileof thehorse(7) bringswith it obv

OXEpa Goo'a' and 9EpErvyos ?'wos,deOAoodpos~cf. the similar phraseology in

Alcman); and ravvalrEpos- 9) and ravtrr-epos.i o) are used in descriptionsofbirds.That is to say, in about sixty lines thereare altogetherseven conventional

epithets, most of which admit of stylistic explanation.

Corinna

It would be wrong to suggest that the thesis of this article providesany newcriterionfor deciding between the alternativedatesproposedby ProfessorPagefor Corinna-viz. the middle of the fifth century or the end of the third.Neverthelessit is worth observingthat her diction is not only distinctlyclassicalin the sense that (apart from four ra Ae

yd•~va)it contains only five words

which do not appear somewhere else in Greekliteratureby the end of the fifth

century, but shows the same sort of discrimination in the use of Homeric

epithets that characterizes archaic lyric.That is to say, if her language is in factnot archaic but archaizing, the work has been done by a singularly delicateand well-trained hand. The following analysis (in which allowances must bemade for the particularlyfragmentarynatureof thematerial)follows thepatternof the others; and it should be noticed that the two discernibleLieblingswdrter

(vKav•parc-v

(.4), Lyov'pavMovp7a (I 5)) are also favouritesof archaic lyric.

A. (i) Heroic context: ? <dKoUnodpw5 A = PSI I174); [04flav] l=r-dovAov(5 B).

(ii) Hymnodic:None.

(iii) Divinities:dyKv•pdLraoKpdvw, cdaclpaPla (4), Als. flaaE1Js, Iderw

/LywvHo~Tlov (5).(iv) Nature: ? AcL8ovros.ovao'7pdow.

B. Unexplained:AoidacL EQ, Aerca 7rpav (4), KaLXdpO OVd (I8),? yiav EpoJ[xopov (Pap. Ox. 2370o. 8).

A. E. HARVEY

Harvey, A. E._homeric Epithets in Greek Lyric Poetry_1957_CQ, 7, 3-4, Pp. 206-223

Documents

Transcript of Harvey, A. E._homeric Epithets in Greek Lyric Poetry_1957_CQ, 7, 3-4, Pp. 206-223