Aeolic Reflexes of Labiovelars in Homer Wyatt, William F Greek, Roman and Byzantine Studies; Fall 1975; 16, 3; ProQuest pg. 251

Aeolic Reflexes of Labiovelars 
in Homer 

William F. Wyatt, Jr 

ONE OF THE LINGUISTIC FEATURES diagnostic of the Aeolic dialects 
is the development of the inherited labiovelar consonants 
to labials, even in environments in which a dental is the 

result elsewhere.1 Thus all Aeolic dialects have 7TlJL7T€ for 7TlvTE <five' 
« *penkWe) and 7TECC- /7TETT- for TECC- / TETT- <four' « *kWetw-). There 
are in the Homeric poems a few words containing 7T f14> from labio­
velar consonants before ElL in accordance with the Aeolic rule and 
counter to the Ionic, which would produce T 8 B.2 Since the normal 
dialectal shape of the epic is Ionic, these words have constituted a 
problem: how can words displaying a non-Ionic linguistic form 
appear in an (essentially) Ionic poem? The most popular and well­
known response to this question has been that the poems, though in 
their final shape predominantly Ionic, had passed through an Aeolic 
phase of composition prior to the attested Ionic phase.3 The words 
showing 7T f1 c/> for T 8 8 constitute evidence for the assumed Aeolic 
stage: because there were no Ionic equivalents at hand to replace the 
Aeolic words, they remained in their original form. The theory of an 
Aeolic epic does not rest solely on the aberrant behavior of labiovelar 
consonants in a few words, but I think it is fair to state that it derives a 

1 M. Lejeune, Phonitique historique du mycenien et du grec ancien (Paris 1972) 43-53, esp. 
47-50. O. J. L. Szemen!nyi, SMEA 1 (1966) 29-52, has argued that the labial outcome in 
Aeolic dialects is due not to regular phonetic change but rather to analogical influences of 
one sort or another. His view, though skillfully argued, has not found wide acceptance 
(cf. R. Arena, SMEA 8 [1969] 7-27). For our purposes it matters little how the Aeolic forms 
arose, for their presence in an essentially Ionic poem still requires explanation, regardless 
of their (phonological or analogical) development. All references to Szemerenyi in this 
article are to SMEA 1. 

I P. Chantraine, Grammaire homerique3 I (Paris 1959) Il3-15 [hereafter, CHANTRAINE]. 
3 It is not always clear what "Aeolic phase of composition" is intended to mean in the 

writings of various scholars. There seem to be three possible interpretations of the phrase. 
(1) The Homeric poems are translations into Ionic of poems Originally composed in Aeolic: 
this is the view of A. Fick (Die homerische Odyssee in der ursprnnglichen Sprachform wiederher­
gestellt [Gottingen 1883]) 12-13, and is no longer accepted as is by anyone. (2) The Ionic epic 
tradition is a transposition into Ionic of an originally Aeolic epic tradition. Many have held 

251 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



252 AEOLIC REFLEXES OF LABIOVELARS IN HOMER 

good deal of its support from these words. It is well to point out, 
though, that only five words are in question, and that the majority of 
them are of restricted occurrence: 7Tl>top,a, 'become, be' (65 x ), 
7T{cvpEc/ac 'four' (6x), 7TE>tWP- 'portentous, ominous' (33x). lPfjPEC 
'Centaurs' (2x ), fNpEOpOV 'deep hole' (?) (2x ).4 

Not all scholars have agreed that these words constitute evidence 
for an Aeolic stage in the history of Homeric epic. K. Strunk (Die 
sogenannten Aolismen der homerischen Sprache [Koln 1957] 20-37) en­
deavored to show that the forms in question need not be specifically 
Aeolic, since both Aeolic and other dialects show inconsistencies in the 
handling of labiovelar consonants. His arguments, however, are 
themselves inconsistent both with the attested facts and with linguistic 
method, as has been pointed out by O. J. L. Szemerenyi (JHS 79 
[1959] 192-93) and, at greater length, by E. Hamp (Glotta 38 [1960] 
194-98). Szemerenyi has more recently (SMEA 1 [1966] 31-35) returned 
to the question and has himself argued that the alleged Homeric 
instances do not support the theory of an Aeolic stage in the develop­
ment of Homeric epic. He concludes that all the words in question are 
susceptible of another, better, explanation. His conclusion is, I believe, 
correct, but his arguments have not carried general conviction, so that 
Wathelet (pp.63-91) and Durante (pp.17-62) retain the traditional 
view. In what follows I shall endeavor to support Szemerenyi's 
position with other arguments, in hope of being able to win for it wider 
acceptance. Before proceeding, however, I would point out, as 
Chantraine (1115,509) and Hamp (op.at. 198) have done, that all these 
words are definitely 'epic' and have a special flavor or color of their 

this view. and I cite only two of the most recent adherents: P. Wathelet. Les traits eoliens 
dans la langue de l'epopee grecque (Incunabula Graeca 37. Rome 1970) [hereafter. WATHELET]; 
and M. Durante. Sulla Preistoria della tradizione poetica greca I (Incunabula Graeca 50. Rome 
1971) 17-62 [hereafter. DuRANTE]. (3) Other scholars. while not positing an Aeolic stage of 
epic composition prior to the Ionic. have nonetheless supposed Aeolic influence at an early 
period of oral epic composition: e.g. M. Parry. HSCP 43 (1932) 1-50. particularly 40-46; 
D. L. Page. History and the Homeric IlilJd (Berkeley 1959) 220-21; C. Whitman. Homer and the 
Heroic Tradition (Cambridge [Mass.] 1958) 60. All of these views have in common the twin 
assumptions that (a) there was an Aeolic epic and (b) that Aeolic forms in the Homeric 
poems were inherited from this Aeolic epic. I shall not distinguish among the various views 
in what follows. intending that by <Aeolic epic' and <Aeolic stage' the reader assume that 
principles (a) and (b) are referred to. and not any specific variety of them. 

t I omit from consideration here 9~C'1rEC'OC and &ClI'fT'OC, o{3~>"oc, c{3EwvfU, though included 
by Wathelet (67-69). Cf. Chantraine. loc.cit .• for these words. and H. Frisk. Griechisches 
etymologisches Worterbuch I-III (Heidelberg 1960-72) s.w. [hereafter. FRISK]. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



WILLIAM F. WYATT, JR 253 

own. They are not words which would appear in everyday discourse 
but are on a higher stylistic and emotional level. 

Our first task is to show that these words cannot have been derived 
in a direct line from an Aeolic epic later translated or transposed into 
Ionic. Those who favor an Aeolic stage in epic composition usually 
maintain that, when the Aeolic epic :was adopted in Ionia, all metri­
cally equivalent words were replaced by their Ionic counterparts: the 
only words to retain their Aeolic form were those which either were 
met~ically different from their Ionic counterparts or had no Ionic 
counterpart. Since by this view Aeolic forms cannot have been 
created by Ionic poets, they must have been preserved intact in old 
formulas (from which they can have subsequently spread) and have 
been handed down over the years from bard to bard. As it happens, 
the word,rl>..cov 'headland' in the formula ,rl>..cov apovp7Jc (Iliad 13.707, 

18.544, modified in 18.547) both (probably) contains an initial labio­
velar (cf Frisk II 873) and does not occur in Ionic, thus fulfilling the 
conditions for the preservation of an Aeolic form.5 In accordance with 
the theory of an Aeolic epic this word should have been preserved as 
*,"IAcov. Since we find TIAcov, the traditional view cannot be correct. 
Furthermore, the word 7TEptTEAAop.at (etymology: Frisk II 869) in the 
traditional phrase 7TEptTEAAoP.EVOVC EVUXVTOVC (Il. 8.404, 418) was not 
retained as *7TEPL7T€A>"op,aL, in spite of the fact that it does not occur in 
Ionic and in spite of the fact, too, that a labial derivative (7TEpt7f>"OP,€VOVC 
EVLavTovc Il. 23.833) existed in the poems to protect the 1T- from Ionic 
corruption. These two words prove that the regular epic development 
of labiovelar before lei was T-, regardless of whether the word existed 
in Ionic.6 Forms with 7T- are secondary, not original. The question now 
facing us is, of course, from what source did the TT- forms enter the 
epic tradition, and why? 

7TEAop,aL 'become, be' (Frisk II 500-01) for expected TEAop,aL (Frisk II 
870-71) has beside it the aorist E7T>..6p.7Jv, and it seems certain that 
Szemerenyi (34) is at least partly correct in holding that the con­
sonantism of the aorist has influenced that of the present. Reducing 

:; On this word most recently cf R. S. P. Beekes, The Development of the Proto-Indo-European 
Laryngeals in Greek Uanua Linguarum Series Practica 42, The Hague 1969) 275-77. 

6 One can, to be sure, rescue the theory of an Aeolic stage of composition by assuming 
that both'TtAcov and 'TT£PLTEAAop.aL existed in Ionic at the time of the introduction of Aeolic 
epic but disappeared prior to our written records. At this point it is preferable to reach for 
Occam's razor. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



254 AEOLIC REFLEXES OF LABIOVELARS IN HOMER 

all forms of a stem to one single shape is a frequent linguistic occur­
rence, and where there is a labial-dental conflict in Greek, it does in 
fact seem that the labial outcome triumphs: cf >..€l7TW. >"€{7T€t (for 
*'A€{T€t); and E7TOP.C1.t. E7T€Tat (for *€T€Tat), though the instance of 
8€{vw : E7T€cpVOV « *ghWen- : *ghWn-, cf Frisk I 657-58) should caution us 
against assuming that all such irregular paradigms are leveled in 
favor of the labial, especially when different aspect stems are involved. 
Szemerenyi apparently assumes that an earlier *Tl>..op.at : E7T>"€TO was 
analogically leveled in favor of 7Tl>..op.at : E7T>"€TO within the Ionic 
dialect or within the history of epic. He is perhaps correct in this, but 
the facts that 8€{vw did not pass to *4>€{vw, -Tl>">"op.a£ did not pass to 
*-1Tl>">"op.a£, and that neither *Tl'Aop.a£ nor *1T€>..op.a£ occurs in Ionic, 
encourage us to look in another direction, while at the same time 
maintaining the essence of Szemerenyrs explanation. If *kwelomai (in 
its two root shapes *kWel- and *kwI-) had been part of the epic tradition 
prior to the elimination of the labiovelars, it must of necessity have 
appeared at one time as *Tl>..op.a£ : E7T>..6p.TJv;7 it is not likely that 
E7T>..6p.TJv alone was inherited and that 1T€>..op.at was a new, purely epic 
creation formed by analogy. One wonders, then, what motive there 
can have been for the Homeric poets to remodel *TE>"op.a£, the more 
so since (7T€P' )TE>..>..op.a, was not remodeled to *( 7T€P' )7TE>">"op.a£. 8 

We can cite as a possible motive the fact thatTE>..op.a, was unknown 
in Ionic, was supported by no analogies, and was therefore isolated in 
epic. Its isolation cannot be the sole reason for its reshaping, however, 
for the remodeling of an attested word on the basis of its aorist is a 
far from obvious solution, one resorted to only when communication 
is likely to be seriously impaired or when the effort required to re­
member the correct form exceeds the amount of informational yield. 
Both of these factors may have played a role in the reshaping of 
*TE>"op.at, but it seems to me that there is another factor which 
accounts both for the reshaping and for the Aeolic appearance of the 
word.7TE>"op.a£ is not an epic word only, for it occurs frequently also in 

7 Forms of T'o..op.m occur in some Doric dialects: Cyrene T'tvrw. (SEG IX 3.38, 72.18, 86: this 
latter inscription is Buck no.115=C. D. Buck, The Greek Dialects [Chicago 1955]); on this 
form if. Szemerenyi, Synape in Greek and Indo-European and the NatKre of the Indo-European 
Accent (QIulderni Napoli 3,1964) 165-67; Drerus T'0t.0p.m ICr I.ix 1.46, 63; cvvro..Ec8w. 1.69-70 
(=Buck no.120). Hierapytna T'o..ET'w. ICr ill.iii 4.68. 

8 In the case of this latter word we can posit that the common avarl,uw 'rise' (of any 
heavenly body) preserved '7TE(J'T'IMop.w. from analogical change. Such is, however, not 
necessary: the form remained because there was no reason to change it. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



WILLIAM F. WYATT, JR 255 

the poems of Sappho (50.1, 81eb)3, 79.5, 94.26 L-P) and Alcaeus (E2.6, 
237.2,250 L-P) in the same meanings that it has in epic. In the Aeolic 
poets it does not give the appearance of being a peculiarly poetic 
word, and it thus stands a good chance of being relatively colloquial. 
It is most natural to assume, then, that epic poets borrowed the shape 
7TiAoJl-a, from Lesbian Aeolic. In so doing they will have both regu­
larized the paradigm, thus rendering it more clear, and will have 
incorporated an easily comprehensible word for one that could have 
caused difficulties in comprehension. They may, indeed, have felt that 
nIi\.op.aL was the <correct' epic form and TI.AoWn some sort of mistake 
in the tradition. 7TEAop.a" though a Lesbian form and hence an Aeolism, 
represents a surface Aeolism in the poems-an Aeolism, that is, 
which replaces a form nonexistent in Ionic with an existent (though 
not Ionic) word. It is doubtful, however, that the poets thought in 
terms of Ionic and Aeolic dialects, or even of Ionic and Lesbian or 
local and foreign. Rather they probably thought in terms of existent 
and nonexistent words. 

The same explanation does not work so neatly in the case of 7TlcvPEC 
<four' (7TlcvpEC Od. 5.70, 16.249, 7Tlcvpac It 15.680, 23.171, 24.233, Od. 
22.111). Indeed this word constitutes a problem both for those who 
posit an Aeolic epic and for those who do not. The form occurs as such 
only in Homer, appears in no Aeolic dialect, and therefore cannot 
without further argument be labeled Aeolic. All the Aeolic dialects 
agree in showing an e-vowel in the first syllable, and all show -ss­
(-tt-); 7TlcvPEC cannot have entered the epic tradition from a living 
Aeolic source.9 There the matter may rest so far as the theory of an 
Aeolic epic is concerned. There is, however, the further problem that 
7TlcvPEC cannot be Ionic either, since Ionic always has 'TECC- and never 
has a u-vowel. It must therefore be a compromise form, and must 
replace an earlier form of identical metrical but different phonological 
shape. Szemerenyi (34) holds that 7Tlcvpac is an approximation to 

e rucvp£c has been frequently assumed as an Aeolic (Lesbian) form on the basis of 
]YPE:EKAI~EKOTO:E (IG xn 282.3), which occurs in a clear numerical sequence. The 
r~storation rUc]vp£c looks convincing in the editor's transcription into standard Greek letters, 
but his representation of what actually appeared on the original reveals that the left hand 
margin is not even. One can then restore either IIE:E or IIE:E:E, but given the unanimous 
testimony of the other Aeolic dialects in showing 71'£CC- and Hesychius' 71'lccvp£c, it seems all 
but certain that 71'£CC- is to be restored here. Balbilla (0. Hoffmann, Die griechischen Dialekte 
II [Gottingen 1893] no.177.6 p.1Z8) has 71'lcvpa., and this is to my knowledge the only evidence 
for ,c- in any Aeolic dialect (unless the Homeric form be included). 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



256 AEOLIC REFLEXES OF LABIOVELARS IN HOMER 

neighboring Lesbian of an earlier (regular Ionic) *Tlcvpac or *Ttcvpac, 
the Ionic development of the accusative *kWeturns. He is surely correct 
in his assumption of an approximation to Aeolic, and his *kWeturns, 
amply supported from the Indo-European point of view (cf. Skt. 
catura~, accusative plural masculine), economically accounts for the 
presumed early Ionic form. The only objection to his view is that 
nowhere else in Greek is any development of a form *kWeturns 
attested. Hence I would set beside-not against-his explanation the 
possibility that 1Tlcvpec is an accommodation to Aeolic either of 
TlTOPEC or of an Ionic phonological development of it, e.g. *Tlcvpec or 
*TlCVPEC « *TlTVPEC < *TlTOPEC). TlToPEc is required for Greek, since 
the West Greek dialects have it, and for epic, since Hesiod has TlTop(a) 
(Erga 698).10 The operative consideration is not, however, what the 
replaced form was, for we shall probably never know. What is 
important is that 1TlcvPEC is, as Szemerenyi holds, an approximation 
by an Ionic poet to a (presumed) Aeolic word. We must posit an early 
epic form of <four', the first two syllables of which were short, and 
which probably derived either from *TlTOPEC or *TlTVpac. Why the 
epic poets replaced their own unclear form with what must have been 
the equally unclear 1Tlcvpec is nonetheless a mystery, a mystery for 
which I have no convincing solution. The best I can do is to suggest 
that Ionic epic poets found themselves singing a nonexistent word to 
which there corresponded in neighboring Lesbos a near phonological 
and metrical equivalent. Feeling that the Lesbian form, though not 
an exact metrical equivalent, was nonetheless likely to be closer to 
the old epic form than the form they had learned, they adopted it.ll 

1TEAWp- seems also to have been borrowed from an Aeolic source, 
this time Thessaly. Szemerenyi (32-33) has a different solution, one 
that seems less satisfactory. He is inclined to deny (1) that 1TEAWP- and 
TEAWP- are the same word and (2) that 1TlAWp is connected with Tlpac 
<ominous sign' « *kwer-, cf. Frisk II 878): he holds that 1TlAWP Hcan be 

10 Cf. G. P. Edwards. The Language ofHesiod in Its Traditional Context (Philo!. Soc. Pub!. 22. 
Oxford 1971) 115-16 with the references to earlier work there cited. 

11 It is interesting that in most instances Homer provides contextual evidence to show 
that frlCVPEC means <four', or at least that a number is involved: (1) the number four is 
inevitable (Od. 22.111). Here TiccEpa occurs in the previous line. and there are only four 
people to be equipped with armor. (2) The number four is implied. though not inevitable. 
At II. 15.68Off reference is made to a team of horses. probably four (Leaf ad 15.683): that 
teams offour horses were known in the Homeric heroic world is dear from 11.699-700 and 
8.185 (if. Leaf ad loc.). For the yoking offour horses in the Homeric poems if. J. Wiesner. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



WILLIAM F. WYATT, JR 257 

suspected of <Eastern' provenience." One cannot argue this last point, 
though it seems reasonable enough, but the first two are not likely 
to be correct. All the evidence we have points to the correctness of the 
equation 'TrlAWP= -r/.AWP Hesychius has the two glosses -r/.AWP· 'TrEAW­
pWVJ fLaKp6v, fL'ya; TEAwptoC fL'yac J 'TrEAWpWC. Though it is true that we 
cannot know the basis on which Hesychius excerpted these words, we 
must admit that he or his sources were in a better position than we 
are to judge the correctness of a semantic equation between rare and 
poetic words. It is not enough to say that Hesychius may have been 
misled. The other bit of evidence, cited also by Szemerenyi, is a 
second-century inscription from Egypt, the first line of which reads :12 

CTij8L Aaov KaTEvama TEAWpLOV E-V Tpt680tct 

To be sure, the poem is queer, full of peculiar expressions, and the 
first line, with its Aaov for Aaav, is difficult and scarcely GreekP This 
fact does not, however, affect the equation 'TrEAWptOC= TEAWpLOC. In 
fact this poem provides good evidence that, at least in second-century 
Egypt, TEAWptOC, known from whatever source, was equated with 
7TEAwptOC. Szemerenyi points out that TEAWPWC here cannot mean 
<immense, gigantic'. He is surely correct, but again this point scarcely 
affects the issue, since it seems quite likely that the poet of this more 
or less mystical and religious poem (about the death of a sacred snake) 
was recalling Odyssey 11.594: 

descriptive of Sisyphus' attempts to roll a rock up a hill. All the 
evidence we have, such as it is, points clearly to the identification of 
TEAWptOC and 'TrEAWpLOC.14 

"Fahren und Reiten," Archaeologia Homerica IF (1968) 22-23, with references to earlier work. 
At II. 23.171 it is likely that we are to think of one horse to a side of Patroklos' funeral 
pyre: numbers are mentioned in succeeding lines. (3) A number, though not four, is 
inevitable: Od. 16.249 7T{CUp£C Kal £iKoCL CPWT£C €aCLv. Amongst the gifts given by Priam to 
Achilles we find at II. 24.233 two tripods and 7Ticvpac Mf37JTac. (4) In 5.70, though numbers are 
mentioned of other phenomena, there is no necessity that there were four springs in 
Calypso's cave. Four is, however, likely. Thus Homer seems to have been at some pains in 
each case to explicate this yAwcca for his hearers. 

12 W. Peek, Griechische Vers-Inschriften I (Berlin 1955) 1313. 
13 One will compare for the form and scansion of A&ov the equally peculiar scansion of 

va6v 'temple' as an iamb in Cretan inscriptions, e.g. Ier I xxm.3.9, 11-12 (Phaistos), xVI.24.2 
(Lato), xVII.21.2 (Leben a). 

14 The only bit of evidence that might suggest that the 7T- of 7Ti)..wp is original and not from 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



258 AEOLIe REFLEXES OF LABIOVELARS IN HOMER 

Szemerenyi glosses 7TlAwp as <giant' and hence denies connection 
with ·dpac. His gloss is, however, incorrect. Hesychius glosses 7T€Awpa 

as 8"f}pla, fHdJ.La'Ta, T€pa'Ta, c"f}J.LE'ia J.LE'yaAa, and comes much closer to the 
truth. 7T€AWp- in Homer does frequently enough refer to things that 
are large or huge, but the primary Homeric meaning of the word is 
<ominous, portentous', not <huge'. It refers to people, things and 
situations which one hopes one will never encounter, and is therefore 
admirably suited for etymological connection with T€paC, as in the 
Hesychian gloss. IS There are two passages in the Homeric poems in 
which 7TEAwp- and 'T€pac are closely connected: 

II. 5.741 
742 

Od. 15.161 
168 

EJI O€ 'TE rOpYELTJ KEcpaA-rJ, OEtJlO'io TTEAwpov 
"', "', A" " OELJI"f} 'TE CJ.LEpOJl"f} 'TE, "-J LDC 'TEpac aLYWXOW 

, , , , "" ,I.. I " I, 
aLE'TOC apYTJJI XTJJla 'f'EPWJI OJlVXECCL TTEIIWpOJl 
,,~ , "'. ",I.. 8' , So ' • ~ "f} JlWLV 'TOO E'f'TJJlE EOC 'TEpaC (IE COL av'TctJ 

In neither case is the equation TTEAwp-= 'T€pac assured, for 5.742 is more 
likely to have referred originally to the aegis described than to the 

a labiovelar is the Pamphylian personal name (in the gentive) 1T€>.wpav (E. Schwyzer, 
Dialectorum graecarum exempla epigraphica potiora [Leipzig 1923] 686a5b). This name, how­
ever, is more likely to be connected with Apollo. with the loss of a- common in Pamphylian. 
Cf. A. Thumb, Handbuch der griechischen Dialekte2 , rev. A. Scherer, II (Heidelberg 1959) 179. 

15 The identification of 1T€>'WP- and Ttpac goes back to H. Osthoff, ArchR W 8 (1905) 51-68, 
particularly 51-54. And the definition of 1Tt>.WP- as 'ominous, portentous' is amply sup­
ported by the text of Homer. He uses 1T€>'WP-
(1) of portent scenes. In addition to the passages cited in the text the word is used also in the 
description of the portents at Aulis at II. 2.321. 'Huge' in general seems inappropriate. 
(2) of otherworldly and frightening persons and events. In all these cases something super­
natural and possibly threatening is involved. 

Underworld: Orion (Od. 11.572), rock lifted by Sisyphus (11.594), Gorgon (11.634), Hades 
(fl. 5.395). Circe's island: a deer (Od. 10.168), humans transformed by Circe (10.219). Gods: 
Hephaistus (fl. 18.410), Ares (7.208). Cyclops: Od. 9.428, 257, 187, 190 (8avpn.). Scylla: Od. 
12.87. 
(3) Only when used of mortals and inanimate objects is the notion 'huge' possible. In most 
cases, however, the accessory notion of 'portentous, ominous' is present and even 
prominent. 

(a) Agamemnon (fl. 3.166): Priam asks Helen to name the man who, not so big as others. 
is nonetheless more 'Y€papoc 'majestic' than the rest. Ajax: Helen identifies Ajax for Priam 
(3.229); is compared to Ares (7.211); at 17.174 Hector refers to Ajax in response to Glaukos 
who had chided him with cowardice; at 17.360 Ajax protects the corpse of Patroklos. 
Achilles: 21.527 describes Priam's watching Achilles rage over the battlefield; 22.92 Hector 
awaits Achilles (and death). Hector: at 11.820 Hector is irresistible. Periphas (5.842, 847), 
an Aetolian. killed by Ares just before Ares himself is wounded by Diomedes. 

(b) frxoc 11.5.594 of Ares attacking. KVpn.Ta: sent by Zeus (Od. 3.290). T€Vx€a: Il. 10.439 the 
weapons of Rhesus are like those of the immortals; 18.83 the weapons of Achilles which 
had been given to Peleus on his wedding day. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



WILLIAM F. WYA IT, JR 259 

Gorgon's head of the previous line;16 and 15.168 probably refers both 
to the eagle and the goose of 161 and not the goose alone. Nonetheless 
there is a connection, if not of grammatical agreement, at least of 
proximity. Once, however, ·dpac is used unequivocally in apposition 
with a noun which is earlier described as 7rO,WPOC : 

II 12 201 ' \ ~ ./~' " , \'" , •• aL€'TOC V'f'L7T€'TTJC, €7T apLC'T€pa /\aov €€pywv. 

202 ,/,. I ~ I ,/,. I " '\ 
'f'0LvTJEV'Ta UpaKOV'Ta 'f'EpWV ovvx€cn 7T€/\WPOV. 

208 
209 

T - ~'" " .. ~ '" "'/" PW€C 0 €ppLYTJcav, 07TWC LOOV aLO/\OV O'f'LV 

, " .If" " K€Lf.L€VOV €V f.L€CCO£Cl, "" wc 'T€pac a,y£oxow. 

The semantic connection between 7T€AWP- and'T€pac is thus clear, and 
an etymological connection is by no means excluded. Whether or not 
the words are of IndO-European origin is a separate question. If they 
are related, they must be descended from a root *kwer-. 

Weare thus left with two options: either we know nothing of the 
origin of 7T€AWP-, or it shows a labial development of an old labiovelar. 
The former is a course of despair, and it is perhaps best to attempt to 
account for the latter. T€AWP- occurred in various areas of the Greek 
world, as is proved by the Hesychius glosses and the Egyptian epigram, 
and it can be presumed to have occurred at one time in early Ionic. It 
seems not to have occurred in the Ionic of Homer's day, for no such 
form is attested in any extant document. An earlier epic *kW el6r- must 
of necessity have developed to *TEAWP-, and because it is not attested 
in Ionic, it must have been isolated in epic and perhaps subject to 
misunderstanding. We must assume that the epic poets learned of 
a 7TEAWP- synonymous with their isolated *T€AWP- and adopted it for 
epic. 

We cannot of course know whence Homer or his predecessors 
borrowed the Aeolic form of the word: it may also have been current 
in Lesbian Aeolic, but we have no evidence from any source to 
indicate that it was. We do, however, have evidence for Aeolic 
provenience of the root, though not perhaps in its Homeric meanings. 

1& The Gorgon's head in this passage is notoriously a late feature, possibly seventh 
century. Cf Leaf ad loc.; H. L. Lorimer, Homer and the Monuments 190 (London 1950); 
G. S. Kirk, The Songs of Homer (Cambridge 1962) 186; M. L. West, Theogony (Oxford 1966) 
46 n.2. The passage, prior to its elaboration, will have contained only lines 738 and 742 
(modified for grammatical agreement): 

738 ap.q,~ 8' u.p' ~J'O'c", fJ&>.ET' alyl8a 8vccavOEccav 

742 8m'1]v TE q.I.Ep8v1]v TE, .1uk Tlpac alytOXOtO 

For a similar expansion of a passage if. Ziva Antika 21 (1971) 55-64. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



260 AEOLIC REFLEXES OF LABIOVELARS IN HOMER 

Baton of Sinope (FGrHist 268 F 5) refers to a Thessalian festival, the 
II€Awpux, dedicated to Zeus IIEAwp, which he says is similar to the 
Roman Saturnalia and was celebrated in his own day. It is true that 
F. Jacoby in commenting on the passage (FGrHist InA Comm. pp.208-
211) is strongly inclined to deny any sort of authority to Baton's 
account. He may be correct in his main contention, but he has not 
provided any reason why Baton should have come up with the 
epithet IIEAwp.J7 It seems most reasonable to assume that Baton was 
correct in assigning the epithet to Thessaly and that the epic poets 
knew of the words 7T€AWP- from a Thessalian source and replaced their 
own difficult *T€AWP- with an actually occurring form. 

<p~P€c <Centaurs' (It 1.268, 2.743) also stems from a Thessalian 
rSource. The word quite clearly belongs amongst the sure Aeolisms in 
the Homeric poems, is the Aeolic development of earlier *ghWeres 
(Frisk I 671-72), and is the counterpart of Ionic *e~p€c. Szemerenyi 
(34) hesitantly suggests that <p~P€c may be a borrowed word, like 
KEvTavpo£, but this suggestion is unlikely to be correct (Arena, SMEA 
8 [1969] 9-11): it is most natural, though not inevitable, to retain the 
traditional connection. How, then, did <P~PEC make their way into the 
Homeric poems? Szemerenyi (34-35) maintains that Hit is quite 
possible that the story of the CP~PEC, first sung by local bards in Thes­
saly, was received into the mainstream of epic poetry, naturally with 
the local form of the name." He is quite possibly correct, but he is 
forced on this explanation to assume a Thessalian (epic?) poetic 
tradition for which we have no other evidence. It is rather more likely 
that Homer (or one of his predecessors) had from the tradition the 
phonologically correct ~~PEC, which corresponded to nothing in 
Ionic if referred to human(oid)s and which could not refer to animals 
in the passages in question. Both Sappho (cp~p[ 19.2 L-P) and Alcaeus 
(cpijpa 01(b)3 L-P) have CP1}P, but in neither case does the context 
indicate that the word refers to Centaurs. We cannot rule out the 
possibility-even the likelihood-that <P~PEC= <Centaurs' was known 
on Lesbos and adopted by the Homeric poets from Lesbian. Since the 
word is striking and has elicited comment by both ancient and 
modem scholars (cf Leaf ad Il. 1.268), it is rather more likely that it is 
a yAwcca in Homer.iS Somehow the epic poets-and we must not 

17 The Thessalian connection of the term is further strengthened by the notice in Schol. 
T to II. 16.176 that the giant ru>.wp was killed by Poseidon in the Spercheius river. 

18 So, from a different point of view. K. Witte. Glotta 5 (1914) 48 n.2. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



WILLIAM F. WYATT, JR 261 

imagine that they were isolated from contacts with the mainland­
learned that to the difficult *B~p€c of epic there corresponded the 
Thessalian form (/>~P€C, a word which referred to the equally Thes­
salian Centaurs. Hence, since there was no metrical cost and great 
poetic gain, they substituted the <correct' CPfjpt:c for the incompre­
hensible *(8Hjp€c. <Pfjp€C is an Aeolism but a secondary one, and 
Szemerenyi (34) is correct in maintaining that cp~P€C "does not guaran­
tee an Aeolic stratum." The most that it can do is guarantee an aware­
ness on the part of the Homeric poets of words and names current 
elsewhere in the Greek world.19 

Szemerenyi (33) has said what needs to be said about {3€pdJpov < deep 
hole (1)': there is no linguistic reason to consider it an Aeolism. There 
is, however, perhaps a thematic reason for considering it a borrowing 
from Aeolic or an adaptation to Aeolic of an earlier Ionic form 
(*8lp€()pov? cf Frisk I 219). The word occurs twice in the poems. At 
Iliad 8.13-14 Zeus threatens any god who aids either the Greeks or the 
Trojans: 

" <\ \ <1.1. ''1' ' " Tj ILLV €I\WV pL'f'W €C .1 apTCZpov Tj€pO€VTa 
A \ , \ ':r f3 'I) • \ () " f3' I) TTjI\E JLal\ ,'IXL a LeTOV V1TO X OVOC ECTL EP€ POl'. 

And at Odyssey 12.93-94 the Scylla is described: 

" \, '\ I:' II:' JLECCTj JL€V TE KaTa C7TELOVC KOLI\OLO O€OVK€JI 

19w 8' €gtCX€L K€r/>aAd.c OELJlOLO {1€pNJpov. 

In both cases mythical and terrifying places are described, and 
f3lp€l)pov thus falls into the same category as 1T€AWPUX and rp~PEC. We 
might therefore be tempted to label {1lp€l)pov Aeolic, if by that term 
we intend remote and frightening things. I can, however, point to no 
Aeolic source for the word in Homer and therefore adopt Szemerenyi's 
solution. 

U The battle of the Lapiths and Centaurs is referred to also at Od. 21.295-304 but without 
mention of tlJiipf:c. In that passage, however, the word ,pp.qv occurs twice (,pp£vac 297, ,ppf:clv 
301), and it is possible that Homer used this word because he had made an etymological 
connection between ,pp..]v and ,p..Jp. nnp{Oooc also makes his appearance twice in the same 
passage (296, 298), as well as at II. 1.263, a passage in which q,iipf:c also appear (268). All this 
suggests both that Homer was thinking of the word tlJiipfiC in the Odyssey passage as well, 
and that he connected nf:'ptoooc and tlJiiPf:C etymologically (cf my Metrical Lengthening in 
Homer [Incunabula Graeca 35, Rome 1969] 199-200) and that he felt that ,pp.qv was related to 

both. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke University, Department of Classical Studies 



262 AEOLIC REFLEXES OF LABIOVELARS IN HOMER 

Thus we see that forms in Homer displaying the Aeolic develop­
ment of the labiovelar must be secondary and cannot in any way 
support the assumption of an Aeolic stage of epic composition. I have 
also suggested motives which might have induced the Homeric poets 
to replace the inherited Ionic forms of certain words with Aeolic 
forms. These motives may seem insufficient, and I add here further 
general considerations which contributed to the poets' willingness to 
include Aeolic words and forms. First of all, Aeolic, at least in the 
Lesbian variety, was spoken nearby and was known to Homer and 
his hearers. Secondly, some of the heroes, Achilles included, lived in 
areas in which, during the archaic period, an Aeolic dialect was 
spoken, and some of the action reported in the Iliad was situated near 
Lesbos. Furthermore, Lesbian was spoken in roughly the same area 
in which Troy was located. In addition to these thematic reasons there 
were also linguistic: many archaisms of Homeric epic-such as K€ for 
av, apocope of prepositions, -ow of the genitive singular masculine­
actually occurred in contemporary Aeolic dialects, particularly 
Thessalian. This sharing of archaisms with Aeolic dialects called forth 
in the mind of the Homeric poets the equation of archaisms with 
Aeolic forms generally. Thus, in cases-such as those discussed in this 
paper-in which an archaic Ionic form was (a) no longer present in 
Ionic, and (b) had a metrically equivalent Aeolic form, the poets 
adopted the Aeolic form on the grounds that in so doing they were 
< correcting' the epic tradition.20 

BROWN UNIVERSITY 

April, 1975 

10 I hope to develop this notion in greater detail elsewhere. For the moment, cf. my 
"Homer's Linguistic Ancestors," 'E1nCTT//-WvLK9 'E1I£TT/pk B£ccaAovlK'I'Jc 14 (1975) 133-47. 

Copyright (c) 2003 ProQuest Information and Learning Company 
Copyright (c) Duke lJniversity, Department of Classical Studies