Sappho 31 and Catullus 51 Wills, Garry Greek, Roman and Byzantine Studies; Fall 1967; 8, 3; ProQuest pg. 167

Sappho 31 and Catullus 51 

Garry Wills 

The Problem 

DESPITE CORRUPTION at certain points, Sappho's famous poem 
preserved by "Longinus" seems clear on its surface. But the 
moment one looks to the implicit ties of part with part, he 

finds different men giving the most divergent reports of them; finds, 
even, that he arrives at different results himself after different read­
ings. Under scrutiny, the components of the poem seem to drift apart, 
or sort ill with one another, despite brilliant efforts to establish their 
coherence.1 The extant strophes move forward in three leaps: "(1) 
That man who (ever) sits by you seems godlike; and (2) this (-TO) it is 
that stuns-or stunned-me; (3) because (yap) whenever I look at 
you, the effects of that glance totally debilitate me." The separate 

1 The most complete bibliography for Sappho's poem, up to the year 1948, is given in 
pp.lO-l2 of Costanza's book (in the list below, under 1950). Much of the literature touches 
only on a Single point or separate verse. For the thought-sequence of the poem as a whole, 
these works seemed to me most useful: 

1816: F. G. WELCKER, Sappho von cincn herrschenden Vorurtheil befreit (Gottingen), with 
supplementary notes=Kleine Schriften II (Bonn 1845) 80-144. 1827: C. F. NEUE, Sapphonis 
Mytilenaeae Fragmenta (Berlin) 27-36. 1913: Ulrich von WILAMOWITZ-MoELLENDORFF, 
Sappho und Simonides (Berlin) 56--61, 75-6.1924: Hermann FRANKEL, "Eine Stileigenheit der 
fruhgriechischen Literatur," NAkG 1924, pp.63-127= Wege und Formen fruhgriechischen 
Denkens 2 (Munchen 1960) 40-96.1931: Bruno SNELL, "Sapphos Gedicht CPalv£Tal fLOL KfjVOS," 

Hermes 66, pp.71-90.1936: Wolfgang SCHADEWALDT, "Zu Sappho," Hermes 71, pp.363-73= 
Hellas und Hesperien (Zurich 1960) 66-77. 1938: Walter FERRARI, "n Carme 51 di CatuIlo," 
AnnPisa ser. IT vol. 7, pp.59-72.1939: Alessandro SETTI, "SuI fro 2 di Saffo," Stltal N.S. 16, 
pp.195-221.1939: Franz TIETZE, "CatuIls 51. Gedicht," RhM 88, pp.346-67.1942: Adelmo 
BARIGAZZI, "L'Ode di Saffo CPalv£Tal fLOL Kfjvos e l"adattamento di Catullo," AttiMilano 75, 
pp.401-30.1945: Lidia MASSA POSITANO, Sajjo (ColI ana di Studi greci II, Napoli), pp.84-115, 
120, 150-4.1950: Salvatore COSTANZA, Risonanze dell' Ode di Saffo 'Fainetai Moi Kinos' da Pin­
daro a Catullo e Orazio (Firenze) 38-85. 1950: Wolfgang SCHADEWALDT, Sappho: Welt und 
Dichtung, Dasein in der Liebe (Potsdam) 98-112. 1955: Denys PAGE, Sappho and A/caeus, 
corr. ed. (Oxford 1959) 19-33.1964: Gunther jACHMANN, "Sappho W1d Catull," RhM 107, 
pp.I-25. 

I shall cite these items by author and page number (distinguishing Schadewaldt's two 
works by date). Though I refer to Schadewaldt's 1936 essay in its first, more accessible locus, 
I use the later versions of Welcker's and Frankel" s monographs. All poems of Sappho and 
Alcaeus are cited from Poetarum Lesbiorum Fragmenta ed. E. Lobel/D. Page (Oxford 1955). 

167 



168 SAPPHO 31 AND CATULLUS 51 

stages are clearly marked, but their interrelations are elusive-so 
much so that, in Tenney Frank's opinion, Sappho "seems purposely 
to have hidden the transition."2 Each of the three stages raises ques­
tions that affect its bearing toward the other two: 

(I) VERSES 1-5: Is this a particular scene, affecting Sappho because 
a particular man has captured the girl's attention? Or does OTTtS 

generalize the congratulatory opening CHappy he, whoever, who sits 
there")? The fact that KijVOS is reinforced by <ljVYJP may point to a 
particular scene; OUrtS can be used of an individual (Kuhner-Gerth II 
pp.399-400). On the other hand, OTTLS must be taken as general, not 
individual, if-to avoid terminal hiatus with a monosyllable (as Lobel 
would have us do, in Ea7T~ofjs M'A'Y] [Oxford 1925] p.lxx)-we follow 
Page's suggestion (21) and read T'I T' luoavEt.3 

(2) VERSES 5-6: What does TO refer to? The whole preceding scene, 
including the man's presence near the girl?4 Or simply to the preced-
. h \,. , "''''..I..' \ \' ., mg prase, YEl\atUas tfLEpOEV, or auv 'f'WVEtUas •.• Kat YEl\atUCXS LfLE-

pOEv?5_in which case the man's presence was merely functional: 
an interlocutor was needed to elicit the sweet voice and laughter that 
can stun Sappho.6 But neither of these alternatives accords naturally 
with what follows. 

(3) VERSES 7-16: Verse 7 introduces a general rule, ciJs ... tow: 

Hwhenever I look at you ... " The subjunctive gives this clause the 
sense of 67ToTav(cf Pindar fr.IOB.7-B Bowra: TaKOfLaL EVT' <Xv tow l7Tatowv 

VEOYVLOV Is 1jj3av).7 This general statement is meant to explain (yap) the 

I Catullus and Horace (New York 19Z8) ZOo 
a On the generalizing effect of'Ti with Ot7'TtS see]. D. Denniston, The Greek Particles' 

(Oxford 1954) 5Z1-3, and E. SchwyzerJA. Debrunner, Griechische Grammatik II (Munchen 
1950) 574-5. 

, It is taken this way by, e.g., Ferrari 6Z, Barigazzi 41Z-3. See also Ernst KaHnka in Wiener 
Branos (Wien 1909) 158, and Carlo Gallavotti, RivFC ZO (194Z) 117. 

S So, for instance, Snell 78, Setti ZIZ-4, Tietze 348. 
8 For Snell and his followers, the man is not merely functional. His presence is explained 

not by the poem itself, but by its occasion (a wedding). This explanation "from outside" 
will be considered later; we must first try to understand the poem from its own terms, if 
that is possible, without recourse to an occasio ex machina for the resolution of difficulties. 

7 One expects Kl in the clause (cf Alcaeus 117b.Z6-7 15 Kl 'T'S SlScp), and Barigazzi suggests 
emending to c,;;s K€ yap a' iSw. But it is better to presume that Kl could at times be omitted 
in Lesbian as in Attic poetry (Soph. OC lZZ5, cf Kuhner-Gerth II p.449), since it is omitted 
in other subordinate clauses with subjunctive at Sappho 16.4 and 98.3. G. L.Ahrens argued 
(RM 6 [1839] 60) that, since this is the best way to understand Sappho's subjunctive, the 
Homeric ws • •• ws which Toll introduced into the text by dividing MS. {3pox1ws into 
{3p&X€' ws must be rejected. The ws .•. ws construction takes aorist indicative in both 



GARRY WILLS 169 

particular reaction of vv.5-6. But it is an explanation that does not fit. 
If 7"6 in v.5 refers to the whole scene called up in the first lines, then 
Sappho is forced to say "I am stunned when I see x with y because, 
whenever I see x, I am debilitated." Her explanation leaves out the key 
point in the scene, the man, whose prominence in the first five lines 
is thus rendered poetically futile. If, on the other hand, 7"6 has a more 
narrow reference to y€lt.ataar; and/or c/>wv€taar;, Sappho is made to say: 
"I am stunned when I hear your voice, because whenever I look at 
you . .. " This explanation not only leaves out the man, but omits the 
point selected from the first scene by 7"6 and made the subject of a 
very strong asseveration by the oathlike 1j fLav ("This it is, and no 
other ... "); it leaves out the girl's voice.s 

There is, then, a problem no matter which way we consider Sap­
pho's movement from the particular scene she limns at the outset 
to the general statement of her reaction whenever she sees this girl. 
For the essential point of the poem is often overlooked: Sappho's 
famous description of her "symptoms" is not a record of her reaction 

clauses; but here we have aorist subjunctive in the subordinate clause and present indica­
tive in the main one. Ahrens would read cOs u£ yap FtSw {3pOX€WS, which makes the enclitic 
p.£ open the main clause. Seidler (RhM 3 [1829] 161) earlier skirted that difficulty by chang­
ing p.£ to u£ so that the subordinate clause runs ws yap dutSw {3pOX€WS U€ (Sappho uses 
€lulSrJV in fr.23.3). For the effect of this on the rest of the sentence, see n.47 infra. 

8 Catullus seems to have grasped this difficulty in Sappho's poem; he circumvents it, in 
his own, by importing SPECT AT et audit (v.4) into the first scene, where Sappho had simply 
inraK0I5n. Thus his simul . .. te aspexi (vv.6-7) has a clear point of reference in the first 
stanza. Snell 80 notes that Catullus omits Sappho's {3pOXEWS and avnKa; but their force is 
retained by the contrast between identidem te spectat and simul te . .. aspexi: "He gazes 
uninterruptedly, while if I catch but a glimpse ... " Snell 78-followed in this by Massa 
Positano 20, Costanza 63-4, lIse Schnelle (Philologus Supp.-bd. 25, Hft. 3 [1933] 17), Otto 
Immisch (SBHeidelberg 1933/34, Abh. 2 pp.5-6) and Vicenzo Bongi (Aegyptus 26 [1946] 98)­
takes identidem to mean that Catullus has a rival with unrestricted access to Lesbia, return­
ing to her "repeatedly." Yet, despite the fact that Snell is here contrasting Sappho and 
Catullus, he seems to apply the adverb to her verb (luSaV€L), not to Catullus' spectat et audit. 
If one considers only the text of Catullus, Snell's position leads one to take the adverb in 
conjunction with the participle sedens (here= veniens et sedens), not with the clause's main 
verbs. But the adjective (or adverb) adversus, which separates the participle from identidem, 
is closely linked to sedens; so that identidem should, for balance, go with the following verbs 
(cf G. Friedrich's Kommentar [Catulli Veronensis liber, Leipzig/Berlin 1908] on the balanced 
apportioning of modifiers in the third stanza of this poem). Thus it is better, with Tietze 
349-50, Ferrari 61, Leiv Amundsen (SymbOslo 12 [1933] 72), Ernst Bickel (RhM 89 [1940] 200), 
and Kroll (1923), to take identidem with spectat et audit, recognizing in it an expression of dura­
tion rather than repetition. The intimacy of the conversation, which Sappho stressed with 
1TAaULOV and V1TaKOV€L (touches omitted by Catullus), is suggested in the Latin poem by 
emphasiS on its extended character. 



170 SAPPHO 31 AND CATULLUS 51 

to the scene she began with. It is a description of what happens 
whenever she sees the girl. "Sappho spricht von der stets eintretenden 
Wirkung, nicht von der besonderendes Augenblicks" (Tietze 348). If 
we ignore the generalizing effect of ws ... t8w and try to consider the 
poem as registering the impact of a single event, we make Sappho 
advance as explanation (yap) what is mere iteration: "This scene stuns 
me because whenever I see it I am stunned." 9 But to grant that vv.7-16 
are the statement of a general law is to face a whole series of prob­
lems concerning its relation with the particulars of the opening-the 
fact that the man is given prominence and then ignored; the clumsy 
transition effected by an ambiguous TO; the fact that the general 
explanation does not explain the sentence to which it is appended. 

These structural problems are crucial to one's interpretation of the 
poem. They gave rise, for instance, to the labyrinthine discussion of 
it as an expression of jealousy. One group of scholars says that the 
man's presence with the girl is the key point and that vv.7-16 there­
fore describe the torture of jealousy.1o The trouble with this view is 
that it does not take seriously the general statement ws ... i8w. 
Some try to escape this problem by assuming that (al) i8w is 

. 1 ,\,,~"'" \ , , ~, 'T'~ , equlva ent to ws yap toW a E"vaVTLav aVTW 7T/\aaLOV LaoaVOLaav, aov TE" 

cptiJVE"taav Kat ylAataav l/L1p0E"vll-an ellipsis not only obscure and 
unlikely in itself, but still open to the charge of circularity ("Your 
sitting with him stuns me because whenever you sit with him it stuns 
me"). 

Others deny that the poem has anything to do with jealousy and 
consider it a simple expression of love for the girl.12 But they cannot 

t Cf Tietze 352. Snell 81 tried to remove the problem of this non-explanatory explana­
tion by calling yap "mehr explizierend ('ja auch') als begrundend." For him, there is no 
break between the particular statement of vv.5~ and the general one introducing the list 
of "symptoms" (vv.7-16). It is all one description ofSappho's reaction to the opening scene. 
Such an interpretation ignores not only yap, but the new start at v.7 and the raised level of 
generalization marked by c.:,s ... tow. No example of his softened yap can be found in the 
intelligible passages of Sappho. She seems always to use it in the full causal sense (as in 16.6 
and 94.8), and that sense is clearly in place here (like Catullus' nam). Cf Setti 215. 

10 It is understood this way by Ferrari 63, Barigazzi 414, Page 28, and G. Perrotta (Saffo e 
Pindaro [Bari 1935] 46-9). 

11 Cf Page 22. Tietze 351 rightly observes that Snell's interpretation of Catullus makes 
him understand te aspexi as te aspexi sedentem cum ilIo. 

11 E.g. Schadewaldt 1936.372 and 1950.99-104, 110-12, Setti 208-14, Bowra (Greek Lyric 
Poetryl [1961] 188), V. Bongi (Aegyptus 26 [1946] 96-101). The debate is complicated by the 
fact that some (like Snell 71-2, 76-8) think Sappho's is not a poem of jealousy but Catullus' 



GARRY WILLS 171 

explain the function of the man who is congratulated in the impressive 
first lines. Even Snell, who can rely on the extraneous factor of a 
nuptial occasion to account for the man's presence, seems grateful 
that the fellow obligingly fades from a love poem where he can be 
nothing but an embarrassment: "Sehr kunstvoll wird von Wendung zu 
Wendung das Bild des Mannes weiter in den Hintergrund gerlickt, 
so dass in dem Satz 7"() fk' .ry fkav ... nicht mehr an ihn gedacht 
wird" (p.78). This gradual recession, which Page (28) also believes 
possible, is accomplished by v.5: that is, we are given four verses in 
which the man's importance is created, then gradually (sehr kunst­
v6llig) diminished, until it imperceptibly disappears! Even if one 
could believe in this kind of "artistry," and even if one were to grant 
Snell's thesis that the occasion of the poem is a wedding, a question of 
poetical economy remains: once Sappho decides to introduce the man 
(whether he is bridegroom, rival or nonentity) into her lines, and to 
give him a prominent role in the opening strophe, then that decision 
should contribute to the total impact of the poem. If he is just an 
embarrassment, to be artfully "phased out," then Sappho has botched 
her job. 

Welcker's Solution 

One of the most ingenious solutions to the problem of thought­
sequence in this poem is that of F. G. Welcker. He maintained that 
taos OEOWLV has a heroic provenance and must mean "godlike in 
power." The drift of the poem would then be: "He is strong as the 
gods who sits near you. That-sitting near you~would surely stun 
me, since merely to look at you affects me so violently." A century 
later, Wilamowitz endorsed Welcker's view of the poem and de­
veloped his suggestion that the man is the girl's bridegroom. He 
also pointed out (p.57) that W elcker' s interpretation gives special 
point to the movement from ~a{veTCf.L ••• KfjVOS taos (Uotatv to 

~a{vofLat ••• TEOVaK'Y}V: "He seems a god, near you and able to bear 

is, while others (like Barigazzi 421-2 and Ernst Bickel, RhM 89 [1940] 198-204) think 
Sappho's is a poem of jealousy and Catullus' is not. A. J. Beattie (Mnemosyne ser.IV, 9 [1956] 
111) tried to make the poem unambiguously express jealousy by suppressing the ue in v.? 
(ws yap £lutSw) and making the man the understood object of dulow. He seems not to know 
he was anticipated in this by H. J. Heller (Philologus 11 [1866] 432), though Heller thought 
Sappho's love homosexual, while Beattie makes it center on the man. 



172 SAPPHO 31 AND CATULLUS 51 

such joy; while I seem almost dead from a mere glance at yoU."13 
Wilamowitz (p.50) found similar responsion of phrases in poem 94, 
between v.4, W~ 8€tv<x 7T€7TOv8<X/L€v, and the answering v.11, K<X~ KaX 

~7T&0'X0/L€v. One might add that, in the same poem, successive strophes 
d .1. ~, \ , (2) d '\' , " , "\ ' en /L€ .,.,LUOO/L€V<X K<X7"€/U/L7T<XV€V v. an '1J /L<XV a <X€KOta <X7TVl\t/L7T<XVW 

(v.5). 
Welcker's reading of the poem offers many advantages. Some of 

these (the new light it can throw on the mutilated opening of the fifth 
strophe, the way it specifies the 7"0 of v.5) I shall return to later. The 
main thing to note is that it gives a rationale to the transitions from 
particular to general: that man (KfjVO~) is only the example that comes 
to hand of anyone's power (o7"n~ ... 7"1., accepting Page's suggested 
7"1.) who can sit near the girl.14 And that act (conversing intimately, an 
idea conveyed by Ev&vno~, 7TA&atOV, v7T<XKovn) would stun Sappho, 

13 In order to emphasize the relation of v.16 to cpalvfTa! ••• Kfjvo>, G. Thomson (CQ 29 
[1935] 37-8) restored the deficient v.16 cpalvoJLat <aiYra>, introducing one solecism, hiatus, 
on the chance that it could be cured by a second one, correptio epica (on which see E. Lobel, 
Ea1Tcpov<; Mt>'TJ [Oxford 1925] Ix). H. J. M. Milne (SymbOslo 13 [1934] 21) proposed cpalvoJL' 
<~wyo to mark the same contrast. Despite arguments against linking v.l with v.17, 
TfOvaKTJV cpalvoJLat is enough to show that cpalvfTa! in v.l can mean OOKEL JLO!; Snell 75 and 
others claim that before Plato it must mean 'have the physical appearance'. Wilamowitz 
(on Eur. HF 804) is Snell's authOrity for this doctrine on cpalvt:u8a! yet the translation given 
in Sappho und Simonides (p.56) is "Der Mann macht mir den Eindruck, Gotterkraft zu haben, 
der ... " (and see his comments, p.55, on Jcpalvt:o in fr.49). As Setti observes (p.21O), Snell is 
restricting the meaning of cpalvt:u8a! to its normal sense with the participle, ignoring its 
wider use (as here) with the infinitive (see also Jachmann 6-7). It would be best, therefore, 
to have an infinitive with cpalvoJLa! in v.16. But the MS. future infinitive makes no sense. Lobel 
and Page accept Hermann's '1T!OE~<; even though the adjective (with TE8vaKTJv dependent 
on it) is clumsy, and this makes the verb less clearly mean OOKW than would cpalvoJLa! with 
infinitive. The best solution to this problem seems to be Beattie's (op.cit. [supra n.12] 108-9) 
to take Tt:8vaKTJv directly with cpalvoJLa! and make oMyw 'mOEQT)v adverbial (= o>'lyov OEr:V). 
Heller (Philologus 11 [1866] 435) emended to dMyw 'Tt OELV to get the same sense. WUamowitz's 
view of v.16 was vindicated when M. Manfredi announced to the XI International 
Congress of Papyrology (Sept. 1965) that a papyru s still to be published quotes Sappho's 
line intact: cpalvoJL' ~JL' aVTff. 

14 Page 20 asks why, if Kfjvo<; means "Any man who sits opposite you is fortunate," we 
have "the addition of the specific 0 aV7)p." The answer of Wilamowitz would be that c!JVT)P 
here means 'husband'; but this depends on the exploded hypothesis that the poem is an 
epithalamion. Besides, fr.lll gives us two uses of r'WrJp in a wedding poem, and the word 
cannot mean 'husband' in either occurrence. A better explanation within the context of 
Welcker's view would be that c!JVT)P juxtaposed to 8€0!u!v is meant to express the antithesis 
avqp wv a~To> 8ELa 1TOtEr:. This concessive use of the noun is clearest when accompanied by 
a participle (e.g. Eur. Bacch. 795 1TPO<; KtVTpa >'aKTl~otJLt 8VT)TO<; WV 8t:cjJ), but it can stand alone 
as well (e.g., ibid. 763-4 Ka1T£VWTt~OV Cpvy9 yvvar:KE<; avopa<; "though women, turned men to 

flight"). Cf. Pind. O/. 11.10, Eur. fr.78l vv.29-31. 



GARRY WILLS 173 

since whenever she looks at the girl, a chain reaction of symptoms 
nearly destroys her. Sappho, too weak to do what the godlike man of 
the first lines does, is dizzied by the mere sight of the girl. In the same 
way the Myrmidons, lesser men than their godlike leader, cannot 
bear the dazzling sight of the armor Thetis brings her son: ov8€ 'TtS 

ETAYJ I aVTYJv €lUtOE€W, <iAN ETpwav (Iliad 19.14-5).15 Under the force of 
admittedly more complex emotions, Penelope is so stunned (BVl-'bS 
I-'Ot €VL uT~Bwut TE8YJTr€v) by the suggestion that the stranger in court 
may be her husband, that she cannot address or interrogate him 
(ovoE TL Trpoucpd.UBaL ovVal-'aL ETrOS QVO' Ep€€uBaL), indeed, cannot even 
look him straight in the eye: OVo' €ls Jma lOEuBaL €VavTtov (Odyssey 
23.105-7). The last phrase shows the force of €VavTtos, "face to face," 
in Sappho's second line, where it is not mere reinforcement of 
7TAd.utov, v.3. 

W elcker' s interpretation is not pushed toward either of the unsatis­
factory poles in the "jealousy" debate. As opposed to (e.g.) Snell, he 
thinks that the man remains important throughout the poem, that 
he is not "phased out" after the inexplicable fanfare that brought 
him before us at the outset. But, as opposed to (e.g.) Page, he denies 
that the only prominent role the man can play is that of a rival. The 
man is, for Welcker, an example of the familiar bliss within reach of 
anyone else but denied Sappho because of her extraordinary vulnera­
bility to the girl's charms. 

Attractive as the Welcker interpretation is, it has been severely 
criticized; by now, it is generally rejected. Some of the objections are 
not so much addressed to Welcker's original thesis as to Wilamowitz's 
embroiderings on it. Setti (203-8), Page (30-33) and Jachmann (9-13), 
among others, have shown how flimsy are the assumptions that 
Lesbian society would not allow intimate conversation between the 
sexes except at or after marriage, that wVYJP means husband, that the 
emotions expressed here are appropriate to an epithalamion. One 
might add the observation that if this is a poem of farewell to the 
bride who leaves Sappho's circle,16 then the iterative 6Js ... LOW 
defeats Sappho's purpose: the poet is reduced to plain silliness if, in 
order to state that she cannot bear separation (seeing the bride no 

15 Aelius Aristeides, Or. 18.4, also seems to think that the symptoms are those of one 
stunned by beauty-if, that is, his 8wqikipov 7(h- 5o/ms is an allusion to Y.11 of our poem 
(cf Neue 35 and Lobel's apparatus to fr.196). 

16 Much is made of this concept by Schadewaldt 1936.366, 1950.113, Tietze 364-6 and 
Milne (SymbOslo 13 [1934] 19). 



174 SAPPHO 31 AND CATULLUS 51 

more), she uses the reflection that she never could bear the effects of 
seeing her. 

But these objections to Wilamowitz's elaborate marriage scene do 
not touch the central points in W elcker' s interpretation. And 
Welcker's view can be maintained even if we object to Welcker's 
motive, the "exoneration" of Sappho, which for him involved the 
demonstration that the man in this poem is not an erotic rival. The 
internal merits of his position are entirely separable from the moral 
bias that Welcker felt while making his way toward that position; it 
has other things to recommend it than one's view of Sappho's rela­
tion to her girls. Indeed, we must keep the consideration of substan­
tive textual points entirely innocent of that debate, or be mired in 
endless untestable surmises-ethical, cultural, psychological-that 
have their origin outside the text and impose themselves upon it. 

If we turn, then, to objections raised against Welcker from the 
words of the text itself, we find that the major one has been a cumula­
tive, now almost canonical, argument against his interpretation of 
taos- 8EOLU£lI. Some say it need not mean 'godlike in power', most say 
that it cannot. This position was formed in three main stages by three 
men, Neue, Dornseiff and Snell. 

" e' tao~ €otatv 

Eleven years after Welcker's interpretation was published, C. F. 
Neue wrote (29-30): <CAt veteres poetae constanter deos aeque ac 
mortales amoris potentiae negant pares esse; neque in verbis quid­
quam reperitur, quod ad tolerantiam pertineat, nullum OVVCX'TCXt, 
{m0p-EvEL, €'TA'T} , sed vocabula taos- 8EOLaw altiorem quendam dignitatis 
et felicitatis gradum ostendunt, in quem ille ascendisse videatur." 
To the first of Neue's objectionsI7 one might answer that not even the 
most erotically unstable gods lose their voice, or sweat, or "almost 
die" from a mere glance at the object of their passion. But Sappho 
who at fr.l.28 summons Aphrodite as a aVP-fJ-cxxOS in her love-wars, 
sees the man who can stand the close onslaught of beauty as a hero, 
godlike in war, where she draws back in cowardice. IS That some gods 

17 Costanza 53-4 thinks this the best refutation ofWelcker. 
18 It has often been observed that Sappho's "symptoms" are largely drawn from Homeric 

descriptions of fear; that Lucretius, though he seems to echo Sappho's poem at 3.152-9, 
does so to illustrate the pathology of terror. Cf A. Turyn, Studia Sapphica (Eus suppl. 6, 
1929) 44-5, 52-5. Tietze 363-6 believed that Sappho chose Homeric phrases deSCriptive of 



GARRY WILLS 175 

are not as appropriately remembered in this context no more invali­
dates the phrase 'tao~ 8EOLaLV than the fact that Aphrodite shuns battle 
would invalidate the term av,rt8EOS used of a great warrior. For that 
matter, gods can at times lose even the dignitas and felicitas Neue 
ascribes to them all. 

To the second of Neue's objections-that the idea of strength should 
be expressed by some word like OVVa'TCXL (lUOaVELV JI9-there are two 
answers: that Homer's words lao8€os or OatfLovL laos have the meaning 
<godlike in war' without any further specification, and that the con­
text of the poem does specify Sappho's phrase in Welcker's sense. 
This second point must be dealt with later, but the first deserves 
some attention here, since Welcker did not support his intuition about 
heroic 'tuoS' 8EotULV with a close look at Homer's language-a situation 
that has not been remedied by any of those who later accepted or 
rejected his view.20 

To judge from Iliad 9.603 luov yap UE 8EcfJ 'riUOVULV 'AxaLOt, it might 
seem that epic practice makes <honored as the gods' the proper 
translation of 'taos 8EOLO'LV. But this verse reflects a special use of the 
adverb laov (a use particularly concentrated in the ninth book) for 
describing the honor given a man: some form of Ttw or 'TLfL~ is present 
in all its uses in the Iliad (5.467; 9.142, 284, 603, 616; 18.82) except one, 
where it is used with a word meaning dishonored (a-m]x8€TO, 3.454). In 
the Odyssey, despite taov at 14.203, the normal adverb is tua, but Tlw 
or 'TL1L~ are still present (1.432; 11.304,484), except at 11.557 and 15.520, 
where other words for honoring are used. 

If we turn from this special use of the adverb to the adjective with 
dative, we find in the Iliad that: (a) Zeus warns the gods that no one 
should claim to be taM €fLol since he is ifJEP'TEPOS' (1.186-7, 15.165-7)­
compare Homeric laocf>apt~€LV and luocf>opo~. (b) The adjective luos (as 

fear, sorrow or anger and applied them to her sad love because that love was "strange." 
But it is safer to see here the influence of the commonplace "Love's delights and terrors 
are like war's" than to search for obscure motives in the psychology of the homosexual. 
In her Priame1 on what is KaAALU-rOV (fr.16), Sappho puts her love over against a totally 
heroic list (ranks of horsemen, foot soldiers, ships), and her mythical exemplum is drawn 
from Helen's (heterosexual) love for Paris-the women's love behind the splendor of man's 
exploits, giving rise to all of them, more dangerous than any of them (cj Schadewaldt 
1950.129, 132-3). 

19 Barigazzi 417 thinks this the principal obstacle to Welcker's interpretation. 
20 Irena Kazik-Zawadzka, for instance, in De Sapphicae Alcaicaeque elocutionis colore epico 

(Wrodaw 1958) 82, simply assumes the heroic meaning without arguing the point. 



176 SAPPHO 31 AND CATULLUS 51 

opposed, say, to iKEAOS'21) is used to compare men with only one god 
-with Ares (usually in the formula taoS' "Ap"'lt, but see 22.132 taoS' 
, EJlVaAt<tJ). (c) One who fights taoS''' Ap"'lt also fights taoS' CUAATl (12.40 
and elsewhere, in this or in extended forms) or AaLAam taoS' (11.746 

etc.).22 (d) The term SaL!-'-oVt taoS' (or avrLBEoS')-again as opposed to 

iKEA"'l Ben TtvL-is nowhere used of a woman, only of warriors. And 
there is no yv"ry ia6BEoS' to go with Homer's ia6BEoS' cPwS'. 

From these facts it seems clear that ia6BEoS' and SaL!-'-OYt taoS' (and, 
one may presume from parallel use, avrLBEoS') mean 'powerful as a 
god' in Homer; and, if context allows us to suppose some heroic note 
in Sappho's taoS' BEOtatY, the words stand close enough to their Homeric 
exemplars to carry the meaning 'strong as god' without dependence 
on other words that say the same thing. 

The next important contribution to the argument against Welcker 
occurred in 1930. Franz Dornseiff, reviewing A. Turyn's Studia 
Sapphica in Deutsche Literaturzeitung 51, wrote (in col. 396) that it is 
not only (as Neue claimed) inappropriate to apply the phrase 'equal 
to the gods' to a lover; one should not look for any specific quality in 
words that refer to the gods generally. One can only take the phrase 
(with Neue) as a reference to the general state of blessedness that 
makes the gods pE'ia ,WYTES': " 'Gottern gleich sein' nicht identisch ist mit 

11 Though rK~AOS, like laos, can be used to describe mode of action (e.g. 11. 24.80), it is 
almost always used of physical resemblance: one "looks like" something (ad. 12.418, 
13.157) or some one (usually somegod,I1. 2.478,19.282, 24.699, Od.17.37,19.54). Thus, when a 
god or man is in disguise, he is LK~A&S "TWL (for gods see II. 4.86, 5.450; ad. 5.54, for men 4.249). 
The word is the exact equivalent of J.OLKWS/~lKV'ia. (compare ad. 5.51 with 5.54, and II. 4.78 
with 4.86). The practice with EtKEAOS is midway between that with laos (usually, perhaps 
always, used to describe mode of action-see n.22 infra) and LKE'\OS (usually of appearance): 
it can be used of physical resemblance (ad. 10.304, 20.88, 22.240), but also of singing like a 
bird (210411) or flitting like a dream (11.207). In the Iliad it is restricted to the formulae .pAOYI. 
~LK"'\OS (or aO"TEpo1Tfj or a?ryfj) and ad (or .pAOYI.) ~tKE'\OS &'\K~V. Considering the normal 
practice with LKE'\OS and the fact that ELKE'\OS" used for mode of action is restricted to comparison 
with .pA6~ and aVS" (never with persons), it seems probable that 8EOE{KE'\OS means that 
Achilles looks like a god-i.e., that it is not to be ranked with la&8Eos but with 8EOELS~S (used 
most often of the handsome Paris and his father). 

22 The brief simile with laos-and-dative is devoted only to the M,\,\'Y/ and the '\a'iAa.p in 
the Iliad with one exception, Thetis' recollection at 18.56 and 437 of the way Achilles shot 
up like a young tree (av€3pafL£JJ €PV~L laos). In the Odyssey, the hero is said to sit silent laos 
civavScp (10.378) or to chatter YP'Y/l. ••• laos (18.27). In all these cases, the comparison is with 
an action. The only use of the adjective that seems to indicate appearance is at 3.289-90 
XEVE KVfLa"Ta ••• laa OPEaaW and 11.243 KVfLa 'lTEpLa"Ta.8'Y/ OVPEL laov, where the obvious trans­
lation is 'high (or big) as a mountain'. But perhaps all the other uses of laoS" point to 'with 
the force (weight) of mountains', as 'with the force of a AaiAa.jJ'. 



GARRY WILLS 177 

iero(}€os, (}€o€ll<.€Aos= einem Gott gleich. Wenn man in etwas tiber die 
Menschen hervorragt, was einer beliebigen Pluralitat von G6ttern 
eigentumlich ist, so ist das nicht eine bestimmte Eigenschaftsqualitat 
wie Schonheit-denn wie bestande da Hephaistos !-oder Wider­
standskraft gegenuber der Verlockung weiblicher Sch6nheit-das 
konnte man doch Zeus oder Apollon nicht nachsagen. Sondern es 
muss auf eine Eigentumlichkeit ihrer Lage gehen, eben: auf ihr 
Gluck." 

The approving use of this passage by Snell (76), Perrotta (Saffo e 
Pindaro p.47) and Jachmann (8) has given it wide currency. Yet it has 
obvious flaws. Dornseiff tells us O€OdK€Ao~ can only mean 'like a 
(specific) god'-though the god is not specified when that adjective 
is used (only) of Achilles, and it is surely fanciful to distinguish 
B€o€lK€A' 'AXLM€v (Iliad 1.131, 19.155) from O€OIS J'TTUlK€A' 'AXLM€v 
(9.485, 494; 22.279, 23.80, 24.486), despite the fact that the plural is 
used in the latter phrase-just as it is in the formula J7TL€lK€Aos 
&BcxvcXToLaL (1.265, 4.394, 11.60). According to Dornseiff, these phrases 
would have to mean 'happy as the gods', a description that hardly 
fits Achilles! The truth is that Dornseiff's whole distinction between 
plural and singular in these phrases is without foundation: aJITl(}€os 
means 'godlike' whether one is thinking of ot B€ot or of the equally 
vague B€os'rts (cf ocxlfLoVL leros). A specific god is not referred to; as 
usual, the godlike is whatever goes beyond human action (cf Aesch. 
Ag. 661-6, Soph. OT 1258-9). Thus phrases like la60€os ~ws mean 
'strong as a god' (i.e., stronger than man), even though some gods are 
not known primarily for their strength; and O€O€LO~S Paris (perhaps, 
as well, B€O€lK€Aos Achilles) is 'godlike in appearance' even though, as 
Dornseiff thought it relevant to observe, Hephaestus is no beauty. 

The most complete statement of the case against Welcker's teros 
OlotatV, the one often referred to as decisive, is Bruno Snell's (pp.72-6). 
He relies on a combination of arguments: (a) Domseiff's (and so 
Neue's) assertion that the phrase cannot refer to anything more 
specific than an "Olympian" bliss; (b) the belief that ~CX{V€,TCXL in v.l 
cannot mean 'seems (to be strong as a god)' but must mean 'is re­
vealed (visibly enjoying the bliss of a god beside his bride)'; to which 
he adds the argument (c) that the context is not heroic but epithala­
mial, since an €lKcX'€LV-topos is regular in the makarismos addressed to 

the bridegroom (cf Sappho 115 Tlw a', JJ ~lA€ YcXfLf3p€, KcxAws €lKaaow;) 
and therefore (d) taos BEoLerLv is equivalent to tK€AoL OEOLS (44.22) and 



178 SAPPHO 31 AND CATULLUS 51 

(J€O€LKEAoL5; (44.34), phrases used of the happy couple Hector and Andro­
mache in the poem that describes their wedding. 

We have already considered arguments (a) and (b),23 the ones often 
accepted by those who reject Snell's primary addition to the criticism 
of this poem, arguments (c) and (d). Perhaps Snell (following E. Z. 
Mangelsdorff) is right in saying an €lKa'ELv-topos is a commonplace of 
epithalamia; we have a very slim body of evidence from which to 
draw inferences about Greek epithalamia. All the examples Snell 
gives, outside of Sappho, are in the Roman poets (p.74 n.3), and it is 
especially dangerous to infer Greek practice from Roman in the case 
of marriage poetry.24 Outside Sappho (and Himerius' references to 
her), I find only two Greek instances of any importance, and in both 
the makarismos is not identical with the comparison, as Snell tries to 
make them in tuo5; (JlOLULV OTTL5; • • • In Theocritus' eighteenth poem, 
the comparison brings out the beauty of the bride in order to con­
gratulate the groom: "You happy groom, you have won this bride 
(vv.16-18)-a bride who is like no other mortal (vv.19-20), not 
comparable with any Spartan girl (ovS' ••• TTapLuw(Jfj, vV.21-5) but 
only with dawn, spring, a cypress, a Thessalian steed (vv.26-31)." 
And Aristophanes makes his messenger congratulate subjects be­
cause of their king-bridegroom's splendor: the birds are addressed 
as cL TpLuJ-LaKapLOv ••• rEV05; (Aves 1706-7) because Peisetaerus comes 
to them on his wedding night brighter than a star or than the sun's 
ray (vv.1708-12). In each case one compares x to something as a way 
of congratulating y: cf Euripides, Alcestis 920-1, where Admetus says 
he and his bride were each congratulated (KWJ-L05; ••• dAf3t,wv) because 
of the other's nobility and excellence-

r, ,~ , , , '..I. ' 
W5; €VTTaTpWaL KaL aTT aJ-L'f'oTEpWV 
., " 'f t OVT€5; apLO'TWV O'V~vy€5; € J-L€V. 

Yet Snell would make Sappho's comparison praise the groom as a 
way of congratulating the groom: his €lKa,61uvoS' and J-LaKapt'0J-L€V05; 

are one and the same. 

21 For (b) see n.13 supra. 
It The risks involved in the invention of hypothetical Greek models for Roman epithala­

mia have often been described: cf P. Maas. RE 9.1 (1914) S.V. HYMBNAIOS. col. 132; Wila­
mowitz. Hellenistische Dichtung II (Berlin 1924) 280; L. Perelli. RivFC 28 (1950) 289-312; E. 
Fraenke1.]RS 45 (1955) 7-8. The Greek and Roman attitudes toward marriage were entirely 
different (see H. J. Wolff. Traditio 2 [1944] 91-5). a difference of mentality reflected in their 
respective ceremonies (see Heckenbach, RE 8.2 [1913] S.V. HOCHZIDT, coIl. 2129-33). 



GARRY WILLS 179 

We have true makarismoi in Sappho; but in each case it is a greeting 
or farewell which gives their honorary new title to the yaf113po~ and/or 
vVfU/)(X and the salutation OA{3LE or xaLpE: cf 112 (with Theocritus 
18.29-31), 116, and 117 (with Theocritus 18.49). These makarismoi are 
obviously something different from Sappho's epithalamial compari­
sons at 44.21,34; 105a, 105c, 111.5, 115.2; perhaps 23.7 and 96.4. The 
simple vocative in 115.1 no more makes that fragment a makarismos 
in the strict sense than do the same words at Theocritus 18.9. In the 
same way, we cannot be sure that fr.108 introduces a congratulation: 
Theocritus' imitation of it at 18.38 does not. 

Not only does Snell directly identify the wedding congratulation 
with the comparison; he confuses these with still another alien form, 
the Hphilosophical" makarismos (OA{3LO~ O~ ... ). The wedding greeting 
is a direct address to a specific person in a specific situation, an address 
marked by the vocative and/or by the use of the title 'groom' or 
'bride' (cf Sappho 112, 116, 117, Theocr. 18.16, Eur. Tro. 311) and/or 
by mentioning the gamos which is the occasion for congratulation 
(Sappho 112, Theocr. 18.17, Aristoph. Pax 1333-4, Eur. Tro. 312 and 
fr.781.27-31). The philosophical makarismos, by contrast, is not a direct 
address but a general statement, in which the grounds for congratu­
lation are given, not by using a title or a reference to the occasion, but 
by appending a relative clause that describes the class of people who 
are said to be blessed. The difference between the two types is so 
clearly marked that one need only glance at the examples of both 
facing each other across Snell's pages (74 and 75) to grasp the distinc­
tion. The only case in which they seem to merge is Euripides fr.781.27-
31, of which Snell gives a false impression by quoting only v.27. 
Taken alone, this line does look like a fusion of the epithalamial and 
philosophical makarismoi, of congratulation and comparison, until one 
looks at the rest of the passage: 

cL fl-uKapWV {3uuLA€J~ fL€t,wv ;7"' oA{3ov' 
" () \ <:" 
o~ €UV KTJO€VU€t~ 

\, '()' Kat fLoVO~ a ava7"WV 
yafL{3po~ l>t' a,1Tdpova yaiav 

eva7"o~ vfLvrJuTJ. 

The opening line does not express the conventional oA{3LO~ W~ BEck It 
is a literal statement of fact tailored to this specific marriage: Phaethon 
surpasses even the happy gods in good fortune inasmuch as he (and 

2--G.R.B.S. 



180 SAPPHO 31 AND CATULLUS 51 

not one of them) will marry the goddess.25 The os-clause does not 
present a class but refers to a single situation (like Sappho 112, 
Theocr. 18.17, Aristoph. Pax 1333-4), and the point of the congratula­
tion to the groom is the bride's divinity (OEav) , which allows the 
p.aKapES to be brought into the greeting. 

Once Snell has confused three different types of statement in the 
single sentence cpatVE'Tat P.0L K'TA, he does not notice that their 
inconcinnity works against the very point he is trying to establish­
that Laos OloLaLv must mean OAf3LOS or p.aKapLOs. For the epithalamial 
comparisons do not illustrate the happiness of the ElKa~6fJ-EVoS. The 
Latin examples he adduces are concerned with the bride's or the 
groom's beauty (Cat. 61.16-25, 193-8; Sen. Med. 75-101, where vv.93-
8= Sappho 34). Even the debate on the relative advantages of virginity 
and fertility in Catullus 62 is not about which state is happier, but 
which is more attractive to others (see vv.42-4, 53-5). In the two Greek 
instances outside Sappho, the comparison is also directed toward 
physical beauty and splendor (Aves 1708-12, Theocr. 18.19-31). 

As for Sappho herself, none of the comparative passages Snell deals 
with suggests the meaning 'happy as gods'. The comparison clearly 
turns on the concept of beauty at 96.21-3 OlaLaL fJ-6pcpav ..• E~{awaOaL, 

and at 96.4 tOEaaLKEAavt (Olf!- a' lKlAav apLYvw'TCf, Page: Homer uses 
aplyvw'Tos of theophanies at Iliad 13.72, 15.490, Odyssey 6.108). And 
fr.44.21 LKEAOL OlOLS and 44.34 OEOELKlAoLS, which are part of a descrip­
tion, not a greeting, must refer to the beauty of the couple (ef Aves 
1709-14). Thus Sapphic OlCf (OlOLS) LKEAos has the force of Homeric 
(JEOElKEAos, just as her Laos OEOLaLV can mean aV'Tt(JEoS cpwS.26 And the 

15 The text of fr.781.14-26 suggests that the goddess is a daughter of Aphrodite (so von 
Amim in the Teubner Supplementum Euripideum [1913]), but Wilamowitz emended the 
text (in two different ways) to make Aphrodite herself the bride eef Wilamowitz p.38 and 
Hermes 18 [1883] 415). 

18 Sappho's adherence to Homeric usage in comparison seems not to cover two passages, 
where she echoes Homer's words, but not (so it seems) his meaning. Yet I think first 
impressions are misleading in both cases: 

(1) The first passage is fr.U1.5 Y&./Lf3poHlalpX~Ta, t ZaoS't ~ Ap~v,. Homer's'laoS' ~ Ap"l' referred 
to the power of a hero's onslaught, but Sappho's phrase is followed by the epexegetic line 
uv8poS' /L~&>.w 1TO'\V /L~{'wv. The point seems to be physical appearance (as at Aves 1709-10). 

But G. S. Kirk (CQ 13 [1963] 51-2) suggests an earthier interpretation fitting the obscene 
banter of marriages and explaining the fragment's first lines: the carpenters must heighten 
the hall because the groom comes to his bride as hyperbolically ithyphallic as some of the 
satyrs on Greek vases. Kirk does not refer to the passage that confirms his suggestion, Pax 
1352, where the Chorus says of the groom 'TOV /LEv /L1ya Kat 1Ta~. The antithetical next line 
rijs 8' ~8v TO aVKov leaves no room for doubt about the way in which the groom is /L'yas 



GARRY WILLS 181 

former phrase is no more <happy as a god' than the latter. Therefore, 
even if Snell had established that Sappho's first line contains an 
epithalamial makarismos (and he did not), the line still would not 
mean what he wants it to Chappy as god'), but something else 
(beautiful as a god').z7 

The argument thus far has established only that ZaoS' Olotaw can 
readily enough mean 'godlike in power', not that it must of necessity 
mean that. The phrase is capable of meaning 'heroic' if the context 
calls for this. And the context does. A congratulation necessarily 
recognizes special fortune or achievement. The o/...{3wS' oS' • • • type 
describes a privileged class (e.g., the class of victors described in 
Pindaric makarismoi). This is the kind of congratulation involved 
in Sappho's poem (even Snell admits it is the kind used by Catullus 
in poem 51). But why is sitting near the girl a special privilege? Once 
one excludes Wilamowitz's odd fancy that only husbands ever sat 
near Lesbian girls, two possibilities remain-the man can sit there 

Ka, 1Taxvs. With this interpretation, Sappho 111.5 would keep its Homeric sense: the groom 
is as powerful, as full of prowess, as Ares-a translation that accords well with Lobel's 
suggested Za' for the unmetrical taos. (See now H. Lloyd-Jones' support for Kirk's 
interpretation of av'bpos p.Eya>..w, CQ 17 [1967] 168.) 

(2) At 115.2, answering her own question about the groom, Sappho says tJp1TaK£ {1pa'btvtp 
UE p.a>..urr' EiKaa'bw, which resembles II. 18.56 and 437, where the verb used makes it seem 
that Achilles' rapid growth is all that is described. But the passages have a larger context, 
illustrated by Homer's other comparisons of dear ones to an ;pvos. In Iliad 18 Thetis laments 
the fact that, after all the care she lavished on Achilles, she made the mistake of sending 
him off to the war (Opl..paaa <purov cUs yOVJIcp MwfjS ••• l1Tt1Tpol.T}Ka ••• p.aXT}aop.£JIOv, 18.57-
9,438-40). In the five Homeric passages that compare a human being to an ;pvos, the shoot 
is one that has been specially cared for. In four places the verb -rp'¢XU is used of this 
care (I!. 17.53, 18.57 and 438; ad. 14.175). In four places the plant is raised in a special, 
a sacred, spot: yOVJI0 MWfjS" (II. 18.57, 438), 'A1TOMwvos 1Tapa {1wp.0 (Od. 6.162), xwptp lv 
Oi01TO>"tp, &0' &A£S" ava{1l.{1pOX£JI v'bwp (II. 17.54: cf Sappho's sacred grove, fr.2.5-8). The 
€PVOS is an honored, even holy, plant, which adds to the pathos of its obliteration in 
four of the five passages (II. 17.57-9, 18.57-9,438-40; ad. 14.178-82). Sappho's groom is like 
the precious €PVOS, just as her bride is like the hyacinth of 105c, where the pathos ofloss is 
expressed, or like the inaccessible apple of 105a. Aeschylus describes his marriageable 
young suppliants as specially tended and guarded fruit in Danaus' "antepithalamion," 
Supp. 996-1005. Thus, in its full context, Sappho's tJp1TaK£ ••• EiKaa'bw is, like Homer's 
€PVE£ laoS", probably not so much a physical description as an expression of love for a 
treasured plant, like the hyacinth and the apple. So far as her fragments allow us to form 
a judgement, Sappho seems to draw on the stock of heroic comparisons-laoOEos, laos· ApT}£, 
OEfj lKI.>"T}, €PVE£ lao;--for use in new contexts, but not with new denotations. 

27 Snell 72 relies on Himerius Or. 1.16 to establish the fact that there is an ElKa'££v-topos 
in Sappho; but this, too, works ultimately against his thesis, since Himerius says that the 
comparison emphasizes the groom's heroic prowess (see Lobel fr.105b), not the happiness 
that Snell would find in the comparison with a god in v.l. 



182 SAPPHO 31 AND CATULLUS 51 

because he is a successful lover (and jealous Sappho is excluded from 
her company), or the man can sit there because he is able to endure 
her dazzling proximity and Sappho is not. Since Sappho spends the 
rest of her extant lines expressing the thought that she cannot stand 
the girl's overwhelming presence, the second alternative must be 
the right one. The state of Sappho is contrasted with the man's and 
this contrast defines the nature of his privilege. Thus a heroic meaning 
for the phrase Laos- t)£otaw is not only likely in itself (from Homeric 
parallels), but, as Setti (211) argues, sustained by the whole drift and 
order of the poem, by its total context. Therefore Wilamowitz was 
justified in taking o'Tns- here as «ein Mensch der Art dass er ... " 
(p.58, cf Kuhner-Gerth p.399). 

Sappho, Verses 5-6 
Despite all the advantages ofWelcker's reading of fr.31 and despite 

the ungrounded nature of the usual argument against it (the meaning 
of Laos elotGLv), there is one argument, less frequently made,28 that 
invalidates the whole thesis as it was originally advanced. Welcker 
(99) interpreted the poem this way: HDer Mann, der dir nahe sitzen 
und ruhig verweilend deinem sussen Gesprach und Lachen zuhoren 
kann, scheint mir wie ein Gott-nicht bloss glucklich, wie Hor. od. 
I 1.30, sondern auch eine starkere Natur als ich Weib: mir wtirde es 
gewiss (denn der Aorist hat diesen Nachdruck) das Herz erschut­
tern ... " That is wishful thinking. The simple aorist does not yield 
the sense he wants. 

It is hard to say what the aorist in v.5 (€1TTOaLa€v) does signify. Some 
have called it Hgnomic,"29 but this is not the kind of sentence where 
that occurs. Furthermore, to state a general rule here would antici­
pate, and so weaken, the general statement 6JS- ••• tSw, again making 
the argument circular ("This always stuns me, because whenever . .. "). 
Welcker was right in wanting this cardinal sentence to accomplish a 
transition from the particular to the general. Most scholars simply 
treat the aorist as a present marking the sudden onslaught ofSappho's 
disorientation. so The fact that the word is rarely used in the present 

18 Mainly by Turyn (supra n.18) 10. 
It E. Kalinka, op.cit. (supra n.4) 157, and Longinus, On the Sublime, ed. D. A. Russell 

(Oxford 1964) 101. 
80 Cf Carlo Gallavotti, RivFC 20 (1942) 106, and W. W. Goodwin, Syntax of the Moods and 

Tenses of the Greek Verb (Boston 1890) 18 § 60. Catullus uses the present (eripit). 



GARRY WILLS 183 

gives some credit to this view,31 but even so the verb sits oddly with 
the subsequent present tenses (or perfects used as present). Page (19) 
translates €1T7"OatUEV and imaDEDpofLT}KEV in exactly the same way. The 
aorist is troublesome in any view of the poem. But that does not 
justify Welcker's broad construction of it. 

If the aorist could mean what Welcker thought, it would remove the 
troublesome ambiguity of 7"0 in v.5. The o7"n~ of v.2 is virtually 
equivalent to El (as in most of these makarismoi: cf. Aleman 1.37-9 
o 8' oAfJtO~. oan~ EtJ1>pWV afLlpav 8£a1TA€KE£ aKAav7o~). Since Welcker 
makes €1T76aWEV apodotic, his 76 would stand for the protasis,32 
which naturally resumes the elements in the "semi-protasis" intro­
duced by (5Tn~: "He is blessed inasmuch as he sits there. If I were to sit 
there, it would stun me." But since the simple aorist does not mean 
'would stun', 76 cannot contain the suppressed protasis, and we can­
not make the pronoun resumptive only of the elements in the 077£~ 
clause. It is vague, and therefore inclusive. It should refer to the 
whole preceding statement-not only to the verbs in the subordinate 
clause, but to the main verb 1>aLvE7a£. This makes the most probable 
reading of our text not "Sitting there would stun me," where 76= 70 
i~avE£v ( EfLl), but "The fact that the man is godlike enough to sit there 
stuns me" (76= 70 8' athov laoBEov Etva£ 1>w7a).33 

Thus-even though we recognize a heroic meaning for tao~ Blo£u£v 
-the simple aorist puts us back on the old merry-go-round with 
respect to jealousy, the transition from the man to the girl, the 
passage from particular scene to general statement. The man's 
power is, as Welcker claimed, contrasted with Sappho's powerless­
ness; but the specific point of vv.5-6 seems to be the fact that the man 
has the power-i.e., is a successful rival. And this still makes vv.7-16 
swing abruptly and illogically (under cover of a non-functioning y&p) 

31 I find the present only at Theognis 1018; but Aeschylus uses the compound p.ETa7TToIE'" 

(Supp. 332). Alcaeus used the aorist of the verb as a true past tense at fr.283.3; so probably 
did Sappho in fr.22.14. 

32 For a relative pronoun to express a protasis, see Goodwin (supra n.30) 173 §472 on 
OVTW, etc. in the apodosis, and the examples at Kuhner-Gerth II p.483 §577.3. 

33 So, correctly, Gallavotti (supra n.30, p.1l7) and C. del Grande (Euphrosyne 2 [1959] 186). 
Most of those who seek a referent for T6 earlier in the sentence than the preceding parti­
ciple ,),EAaluas- find it in the subordinate and separated verbs luMv€1 Kal ••• v7TaKovEI. But 
why presume that a reference so distant will be so selective? Reference to the whole first 
sentence, as a single statement, is far more likely. 



184 SAPPHO 31 AND CATULLUS 51 

from the man (and symptoms of jealousy) to the girl (and symptoms 
of love). The promise that Welcker's interpretation held out, the 
promise that it could resolve these problems, is belied. His effort 
founders on the intractability of vv.S-6. Or, more precisely, it must 
founder there if previous restorations of v.5 express Sappho's mean­
ing. 

So far I have been dealing with Lobel's reconstruction of v.5-TO 
p.' .ry p.av for the meaningless 'TO p.~ €p.av of all HLonginus" manu­
scripts save one. The reading of that maverick codex, despite its 
pointless pleonasm (-T<~ p.o, €p.av), could at least be approximately 
construed; it was accepted by Toup, Blomfield, Neue, and only 
slightly altered (to T6 p.0' Tav) by Stephanus, Ursinus, Vossius. But the 
Teubner editions gave widest currency to two emendations, Ahrens' 
TO o~ Ep.av (Bergk, Hiller 1897, Crusius) and Schneidewin's T6 fLm fLav 

(Hiller 1890, Diehl 1925). Since 1936, however, when Diehl put Lobel's 
suggestion in his second edition, .ry fLav has been universally 
accepted. 

There is a fatal objection to this popular expedient: .ry fL~V always 
introduces its asseverative clause (see examples, Denniston pp.3S0-1), 

as in Sappho's only use of it in an understandable context, fr.94.5 (no­
thing can be made of 7Jp.av ..• at fr.99 col. ii 2S) and in Alcaeus 344.1. 
At most, the relative T6 might delay the asseveration; but p.( ot) should 
not intervene to delay it further. 

If we are not to accept.ry p.av, what is the explanation of MS. fL~ EfLav? 
It may be useful to compare here what happened to one manuscript 
of Aristotle 1367a14 (= Sappho 137.7): although the best witnesses give 
al8<hs KEV U€ OUK €lX€v ()p.fLaTa, Page (lOS) reports that the Dresden 
codex reads alows KEV U€ OU KaV X€V ()fLfLaTa, where the fiddling with 
various forms of K€ suggests that the av in KaV arose from a gloss of (tv 

for K€. If, in the Top.T/€p.av of 31.S, the two final letters represent a 
similar use of av to gloss KE-in this case extruding the original word 
-then T6 ••• KEV ••. E1TT6a,u€v would give Welcker's sense to the 
passage, explain the aorist, specify the pronoun T6, and remove the 
difficulty of position in Lobel's .ry p.av, along with all the difficulties 
of logical connection which we have discussed. A possible criticism is 
that we expect an imperfect tense for the contrary-to-present-fact 
condition (ef Sappho 63.7, 137.5). But this could be explained by the 
rarity of the present stem of brT6a,ua (see n.31), a more likely 
account of the aorist than that offered by other hypotheses; apodotic 



GARRY WILLS 185 

use of the verb would keep it from jarring with the tense of precedent 
and sequent verbs.34 

If we seek to make sense of the rest of the line-fL7JEfL-several 
alternatives present themselves. They are, in order of ascending likeli­
hood: 

(1) 'TO fLot KEV (cf Schneidewin 'TO fLOL wxv): this gives a weak and 
asyndetic opening to a very strong statement. Besides, when K€ is 
separated from a verb in the indicative, it usually gravitates toward a 
modal word (as in the next three choices): cf Goodwin, Moods and 
Tenses p.72 §219. It is true that we need a fLot in this sentence (cf Lobel 
'AAKCdov MI.A7J [Oxford 1927] p. lxxxv); but it could have occurred, 
very fittingly, in the next line (KapStav fL' EV GT~OeaLv), and been can­
celed there when the end of v.5 was corrupted to EfLlxv. 

(2) 'T6 fL' ~ KEV is possible: there is more latitude in delaying ~ than 
in the postponement of ~ fL~v. But this is not normal except with 
a vocative or an exclamation, and ~ introducing an apodosis is one of its 
special uses (Denniston p.281, iii). Thus what was said of Lobel's 
emendation applies here, though not as severely: 'TO, as relative­
implying-protasis, might precede the asseverative, but not fL( at) as 
well, which this reconstruction must retain in v.5 to prevent 
hiatus. 

(3) 'TO fLrXv KEV (adversative fL~v, Denniston p.334): this would bring 
out the contrast between the man's power and Sappho's powerless­
ness, and Sappho seems to have written fLrXv KE at fr.70.8. Though TO 
fLrXV KE I KapSLav may seem dysphonious, see fr.137.5-6. 

(4) 'TO S~ KEV (cf Ahrens 'TO 80 ;fLav): this gives a normal opening 
for an apodosis (Denniston pp.224-5), and therefore normal juxta­
position with KI.. Furthermore, the joining of S~ to the pronoun T6, 

implying a protasis resumptive of elements in the OTTLS clause, resem­
bles other uses of S~ with resumptive pronoun (Denniston p.226). 
Thus I think the most probable restoration of the sentence: TO 8~ 

I <;;: , ,. '0 " 35 KEV KapoLav fL EV GT7J €GLV €1TTOaLG€V. 

U And this interpretation does not make J1TT6cau£v mark a single coup, against normal use 
of the verb to mark an induced state of confusion. Cf. Theognis 1018, where the present 
participle shows that the "fluttering" process goes on while one gazes, 1TTOtw/Lat 8' euopwv. 

The same thing is indicated by the use of the perfect passive (Aesch. PV 856 etc.) to describe 
an enduring condition. 

35 There is no single rationale for the transmissional errors that afflict this line. The 
emendations that are palaeographically tidy, those of Lobel and Ahrens, are ungram­
matical. But part of the story may be the confusion of uncial A for M, giving risc to s.n) 



186 SAPPHO 31 AND CATULLUS 51 

Catullus, Verses 5-6 

It might be asked why Catullus 51 does not reflect Sappho's apodo­
tic Kl. It is true that Latin is not as flexible in the omission of the 
protasis (see Gildersleeve-Lodge §593); nonetheless, other things 
being equal, Catullus' quod followed by the simple present indicative 
should (like Sappho's text according to Lobel) refer to the whole first 
statement of the poem, not simply to the verbs in the relative clause. 
But Catullus has taken extraordinary pains that other things not be 
equal. A series of interconnected changes thrusts the subordinate 
verbs into a prominence that makes quod naturally refer to them: 

(1) Catullus removes the first part of his sentence from what follows 
by interjecting a parenthetical line not found in Sappho (ille, si fas est, 
superare divos). 

(2) He brings the two verbs in the relative clause together (again 
departing from his model) and, having retarded the development of 
the sentence's thought with his parenthetical second verse (and by 
changing luoav££ to a subordinate concept, sedens, in v.3), saves the 
verbs for climatic position in the adonius, where they stand alone: 
spectat et audit. 

(3) Furthermore, spectat is introduced (in place of Sappho's second 
verb luocfv€L) to point a contrast between identidem te I spectat and 
simul te I ... aspexi, between the power of the other man to gaze and 
the weakness that overcomes him at a glance. 

(4) The same contrast is pointed by the proleptic misero that, con­
trasted with par . .. deo, helps specify the following quod: Catullus is 
made miser by the same thing that makes the other man beatus (par 
deo does not have the epic ancestry of Sappho's phrase, and the idea 
that power is what makes the man happy as a god must be brought 
out by other words). 

(5) While misero is contrasted with par deo in sense, its proleptic 
prominence at the outset of the second statement puts it in marked 
opposition to the opening word of the first statement, Ille, while 
mihi and mi pick up misero in succeeding lines (6 and 7) much as Ille is 
echoed, anaphorically, in the ille and qui of vv.2 and 3. 

(6) In place of Sappho's Kapo{av .•• (Klv) E7TTOaLU€V, Catullus wrote 

for 8";, and EMAN might be an attempt to make sense of the remaining KAN (cf. KUv in 
Arist. cod. Dresd.) or AN. 



GARRY WILLS 187 

omnis eripit sensus, where omnis is contrasted with the separate inter­
ruptions of sensation in vv.7-12. This parallels the contrast between 
identidem spectare (et audire) and aspicere ( ... nihil < vocis) etc.): "Gazing 
intimately at her ravishes all my senses, because even a glance dis­
rupts particular senses." Note that the nam has full causal force, a fact 
granted even by those who deny it to Sappho's yap. 

(7) Finally, by heightening KapSlav, •• (KEV) €7TT6aLU€V to omnis 
eripit sensus, Catullus sacrifices (by anticipating it) the climax Sappho 
saves for the end of her list (T€(JVaW(JV ..• cpalvoJLat) in order to make 
the superlative statement hypothetical in effect: "Doing what he is 
dOing-looking at her, and listening, close up, and long-takes away 
all my sensation (i.e. would do so if I were imitating him), since even 
a glance makes me aware of a series of physical disorders." Every 
means has been used to contrast the two situations of Catullus-as 
(really) aspiciens and (hypothetically) identidem spectans-and to mark 
the latter situation as unthinkably audacious, impossibly risky.36 

Every device Catullus invents here represents a departure from 
Sappho's technique in order to preserve her meaning.37 Though he 
lacked certain of the resources at her disposal-e.g., a phrase for 
'equaling gods' that expressed heroic power, an economical use of 
apodotic aorist that would leave its protasis implied-he puts his 
passage together in such a way that his quod refers to what the man is 
doing (rather than the fact that this man is doing it) and makes his 
eripit hypothetical in effect (as "unthinkably" superlative). So successful 
has he been in retaining Sappho's concept (as Welcker discerned it) 
that even those who, like Snell and Schnelle and Amundsen, deny 

86 Ilse Schnelle began the analysis of Catullus' contrast between synthetic summary 
statement in vv.5-6 and analytic list in vv.7-12; cf pp.17-23 of her "Untersuchungen zu 
Catulls dichterischen Form," (supra n.8). L. Amundsen (SymbOslo 12 (1933) 73), Setti 215, 
and especially Massa Positano 94-5 have developed Schnelle's inSights, but without suffi­
cient emphasis on the way this contrast unites with the poet's other changes to make the 
"synthetic" statement hypothetical in effect. 

87 This is true even of the whole strophe he omits. Sappho's fourth strophe extends her 
list of symptoms and climaxes it with n8vaK'l]v ••• c/Jalvop.a£, Since Catullus sacrificed this 
effect in order to achieve his superlative statement at vv.5-6, he cannot sustain and cap the 
long list as she did. He shortens his list and describes the symptoms he retains in a rhetoric 
progressively more complex, culminating in the delayed sed, the chiasmus of the four 
verbs and subjects, the enal/age and alliteration of his third stanza. He thus arranges the 
only kind of ascent still available to him and, at the same time, prepares the way for the 
rhetorical topoi of his last stanza. See the fine analysis of the third stanza in Schnelle 18, and 
Otto Immisch's criticism of the view that there is a total break in tone between the third 
and the final stanza (SBHeidelberg 1933/4, p.10). 



188 SAPPHO 31 AND CATULLUS 51 

that Welcker's interpretation fits Sappho 31.5-6, believe that it gives 
a proper account of Catullus 51.5-6. 

The Sapphic Style 

One of the reasons for the eclipse of Welcker's interpretation has 
been the popularity of Hermann Frankel's discussion (1924) of Sap­
pho's style as a simple "link-on" technique (reihende Stil). On the 
basis of this analysis, many contrast the logical structure of Catullus' 
poem with the naive, loosely-joined, "timeless" poiesis of Sappho.38 

The clear articulation between what I have called the poem's three 
"leaps" is, by these critics, dissolved or blurred lest Sappho display a 
structural sense of which they think the archaic style innocent. Snell, 
for instance, claims that Sappho's naivete makes it impossible for her 
to use the relative pronoun TO in v.5 more ambitiously than as a link 
with what immediately precedes it;39 and, in the same way, he would 
dissolve the major junction that marks her passage from hypothesis 
Cit would stun me") to explanation C yap mehr explizierend als 
begriindend," p.B!). 

But Frankel's brilliant and seminal discussion, which in any case 
needs revision in terms of recent work on oral technique,40 over­
simplifies the style of Sappho. Schadewaldt demonstrates how com­
plex is her dramaturgy of the emotions. His prime example is fr.94, 
with its three marked temporal strata-the time when Sappho 
laments the girl's departure, the remembered time of the departure 
itself, and the many remembered times called up at that departure 
to comfort the girl (and resummoned to console Sappho in the same 

88 See especially Snell 81-90, Massa Positano 89-91, H. Frankel 50-1 and his Dichtung und 
Philosophic 2 (Munchen 1962) 212; her poems "stehn unter dem Zeichen der absoluten und 
direkten Gegenwartigkeit." 

89 Snell also argued that reference back to vv.l-4 is precluded by the enclosed character 
ofSappho's strophes, for proof of which he used fr.l (p.78 n.2). But contrast Frankel's own 
analysis of that fragment (pp.48-9). 

'0 See, for instance, Frankel 79-80, where a psychological explanation is elaborated for 
what seems in some of his instances a mechanical problem, the harmonizing of oral 
formulaic patterns. Frankel also subsumed under a general heading sets of phenomena 
which may have independent and more limited origins: for instance, he does not allow as 
cardinal a role to the Priamel (pp.68-9, 90) or to hymn devices (p.43 n.2) as some scholars 
now would. What he considers as one large (and largely unconscious) cultural trait of the 
archaic age must, in many cases, be considered as separate devices-oral formulae, Pria­
meln, hymn-phrases, deliberate antithesis (e.g. Sappho 1.21-4), etc. 



GARRY WILLS 189 

terms she used upon the girl).41 Here is a memory-within-a-memory 
technique of considerable sophistication; and it has certain structural 
parallels with fr.31. There is, for instance, a farewell scene (v.2) fol­
lowed by a statement of the effect of that scene upon one (vv.3-5). 
More important, there is a general statement (KcXA' €1TaaXOjL€v, v.II) 
followed by a long list of good things meant to prove the generaliza­
tion. The list, fragmentary as it is, obviously contains more items 
than the catalogue of symptoms in fr.31, and it is a simple paratactic 
account; each element in the extant part of the text is introduced by 
Kat (compare the seven uses of OE in 31.7-16). These, the most emo­
tional, detailed memories in the poem, have been placed within a 
careful historico-dramatic perspective-just as the famous conversa­
tion with Aphrodite in fr.1 (a conversation carried on "reihenweise")42 
is placed in the past and reduces the intensity of the poem's demand 
(vv.3-4) by making it part of a recurrent pattern. 

Thus, although Sappho's vivid lists-of questions (1.15-20) or pro­
mises (1.21-3) or requests (1.25-8) or luxuries (94.12-29) or symptoms 
(31.7-16)-are given with paratactic immediacy, each of them is con­
tained within a temporal and causal framework that uses complex 
transitions to prepare one for the "naive" listing Ccf the y&p clause at 
fr.94.8, which gives the reason for taking comfort, a reason that will 
be supported by the list in vv.12ff). There is no reason to blur fr.31 
into a naive flow of impressions lacking the forceful connections of 
Catullus' poem. Just the opposite: Sappho seems always to view her 
most passionate moods or moments through some aperture of 
historical control. The restless, yearning figure is removed geo­
graphically in fr.96, mythically in fr.16.43 Even Frankel admits that 
fr.16 is a "mehrschichtige Gedicht," a description he would deny to 
her other poems.44 

Thus in fr.31 we do not witness Sappho paling, sweating, fainting at 
the sight of a particular scene. These reactions are offered as an 

41 Schadewaldt 1936.363-5; 1950.115. 
42 Frankel 49-50 on the questions and 43 on the requests of the Aphrodite-hymn. 
43 See Schadewaldt 1936.372; 1950.12-3, 129-30. 
" Dichtung und Philosophie 2 (supra n.38) 212. Schadewaldt and Page seem to have a more 

adequate grasp of Sappho's style than do those who find in her nothing but a "naive 
immediacy." Schadewaldt wrote (1936.371) "Sie scheut die direkte Ausserung ihres eigen­
sten Gefiihls und sucht ihr Tiefstes auf eine sehr zarte mittel bare Weise auszudriicken." 
And Page 86: "Wherever the evidence suffices for a judgement of her art, we find it to be 
the expression of a reflective, self-critical, and self-dramatizing personality." 



190 SAPPHO 31 AND CATULLUS 51 

explanation of her general inability to stand the girl's bright presence. 
And even this explanation, though it has reached such a desperate 
climax ('TE(}vaK7]v ••• cpCXLYOJLCX£), leads into a contrasting, cooler state­
ment, presenting a fourth "leap" in the poem at just the point where 
"Longinus" breaks off his quotation: "All, however, can be 
borne ... "45 There is no reason for "Longinus" to quote lines which 
no longer illustrate his point (the aVJLf3cx{yOVTCX ••• JLcxv{cx£s 7TCX()~JLCX'Tcx); 
but the corrupt v.17 shows that what went before must act as foil to 
some kind of counter-statement. Even Frankel, whose thesis it is that 
the poem lacks "Gliederung und Stafferung," remarks that "Unser 
Text bricht an der Stelle ab, wo die Sprecherin begonnen hat aus 
einem gewissen Abstand auf das Ereignis zu reflektieren" (Dichtung 
und Philosophie2 p.200). Actually, as we have seen, there was already 
a withdrawal from the specific experience in dJs ... tow, and the fifth 
strophe marks the recession to an even larger perspective, one from 
which the first generalization can be modified and corrected. At the 
same time, v.17 carries Sappho back from her past symptoms to a 
present mood of resolution. 

Furthermore, the second generalization not only broadens the first 
one but reverses it. The obvious completion of her line of thought is 
that she must brace herself to bear all the symptoms that batter her 
in the girl's presence, because she means to enter that presence, to bear that 
obliterating proximity. Wilamowitz and others think Sappho must 
bear the girl's marriage or departure; others, that she must bear the 
loss to a rival. But even if the occasion for the poem presumed by each 
of these schools could be established, such a conclusion would not fit 
the poem as Sappho has shaped it to this point. The danger to be 
overcome, the disorienting influence she has so vividly described, is 
not the pain of loss, of some impending absence, but the unendurable 
joy of her presence, the barrage of her charms. Thus, as Setti realized 
(pp.217-8), v.17, put at the end of the list, is resumptive: Sappho is 
telling herself that all these symptoms can be borne, and therefore 
that she means to expose herself to their cause (the laughing girl's 
nearness).46 The symptoms have been exaggerated, either seriously 

" It is safer. with Frankel and Snell. to accept this translation of TOA/LaTOV than. with 
Wilamowitz 56, to assume that it means the same thing as TOA/LaT£ov. See Eva-Maria Hamm, 
Grammatik zu Sappho und Alkaios (AbhBerl 1951) §142.5. Attempts to decipher the line 
beyond aAAa 1Tav TOA/LaTOV, E7T£t ••• have so far proved futile. 

46 If TOA/LaTOV means (as it may) 'can be ventured' rather than 'borne', then Sappho's 
determination to approach the girl would be even more forcefully stated. 



GARRY WILLS 191 

or playfully, to serve as foil to her final declaration-to describe 
the awe with which she does, at last, approach the girl, or to give a 
humorous note of trepidation to the mustering of her forces for love's 
encounter. Compare the way she summons her divine ally in the 
Aphrodite-hymn.47 

If Sappho's poem is one that brings her into the girl's presence, then 
her first lines may have not only a verbal reminiscence of the Iliad 
(9.190) but a reference to one whole scene: Achilles is discovered 
singing, while Patroclus sits opposite him and listens intently (ol oloS' 

EvavT[os ~aTo aLwTTfj); and he is discovered in this pose by the ambas­
sadors sent to negotiate with him, who approach with understandable 
trepidation Ccf vv.182-4), just as Sappho approaches the girl. Similar 
reference to a whole scene or situation in Homer may be found in fr.l, 
where Sappho's playful conversation with Aphrodite resembles the 
conference of cronies when Odysseus and Athena meet again at 

47 If cpwvas is correct, then EtK£L= tK£L, as in Epicharmus 35.13 (Kaibel), Hesych. s.v. 
EiKEL: cf Hamm, Grammatik p.I26, who also suggests (p.127) that the same verb be read in 
114.2 (it may also be indicated at 20.13 €lKD. With pi as the terminus of this verb's anion, 
argued Seidler (RM 3 [1829] 160), rpwvas ova' EV €T' £LK€L must mean "no voice reaches me" 
(from you). Lobel (CR 43 [1929J 136) and Page 23 agreed, and therefore accepted Daniels­
son's pi rpWV£LU(at) (where, however. dK£L is of indeterminate meaning: =1TapdK£L?) to 
keep the traditional translation "I can no longer speak." Yet Page's objection to "Your 
voice no longer reaches me" (= "I can no longer hear") is weak. He writes that it is "an 
unsuitable sense here." But listening (v1TaKovEL) constitutes part of the man's privilege and 
felicity in the first lines; and it would be more effective for Sappho to contrast her power 
with the man's in this respect, at vv.7-8, since her very next line makes the point that she 
cannot speak. It might be objected that "I cannot hear" comes to much the same thing as 
€1T£PPOP.{J£LUL O' ((KOVaL, vv.1l-2. There are two answers to this: first, the two are not so 
obviously tautological as a directly juxtaposed repetition of the statement "I cannot 
speak"; and second, the tautology is only apparent, since the later statement concerns 
ringing in the ears as a sign of violent pounding of the blood, something one experiences 
entirely apart from the strain oflistening for a dizzyingly sweet voice. Catullus stresses the 
internal pounding of his pulse by adding Sliopte to sonitu. 

Seidler thought he could keep rpwvas and make vv.7-8 mean "I cannot speak" simply by 
removing the terminalp.€ from the main clause. But "the voice no longer comes" should 
still mean that Sappho cannot hear; ETL makes better sense if a glance at the girl has 
broken offSappho's sense of hearing. Nonetheless, Seidler's emendation improves grammar 
(see n.7) and sense: U€ ending the first clause suggests that rpwvas beginning the main 
sentence is "your voice," the rpwva of the girl who is «au rpwv£Lua. Hermann (Wiener 
Jahrbb. 54 [1831] 109, 112) and Heller (Philologus 11 [1866] 434) asserted that lli& following 
a negative must restate the preceding clause affirmatively (thereby fixing the sense of v.8 
as "I cannot speak"). But such a polar expression is out of place in this paratactic list. The 
lli& can be progressive, as at Aleman 1.71, where it varies the list of items beginning with 
repeated ovO€ (Denniston p.22). Catullus, in the same place, has an odd use of sed (post­
positive, his only example), and it too is progressive in sense. 



192 SAPPHO 31 AND CATULLUS 51 

Ithaca (ad. 13.278-310, especially vv.293, 301, 303). Sappho takes 
intimate moments from the man's world of Homer and looks at 
them from a woman's point of view, as she did when she reversed the 
heroic Priamel (cf. ad. 14.222-8, Tyrt. fr.9) by taking Helen's point of 
view in fr.16. 

To summarize this discussion of lines 5-6 and 17 in fr.31 (and to 
relate it to an over-all interpretation of the poem) I offer below what 
seems to me the most reasonable text of Sappho's fragment presently 
available. 

..J.' , ~., 8' 'l'aW€Ta£ JLO£ K'1VOS LGOS €otGW 
" H ~ ,,, I 

€JLJL€V WV'YJP. OTTLS €vaVT£OS T€ 

• , 0;:. I \ \ I -;-O:-..J.. ' 
T £Goav€£ Ka£ 7Tl\aGLOV aov 'l'WVEL-

4 aas v7TaKOVef, 

\ \,., \~I 

Ka£ Y€l\a£Gas LJL€PO€V. TO 0'1 K€V 
0:- I ., '8 " KapotaV JL €V GT'1 €GW €7TTOaLG€V. 

• , • I~ R I ..J.. 1 ws yap €£GWW fJPOX€WS G€. 'l'wvas 

8 avo' EV ;T' €ZK€L. 

Y.\\'" '\~' " \1 
t.tI\I\ aKav JL€V YI\WGG a7T€ay€. 1\€7TTOV 

8' aVTLKa xpip 7TVP V7TaO€OpbJL1'JK €V • 
., I ~"~,.."" .,., , 

07T7TaT€GGL 0 avo €V OP'1JLJL • €7TLppOJL-

12 {3€£G£ S' aKovaL. 

\ 0;:. ~ I ..,~ .1. ~ .. 1 ~ \ 
KaO O€ JL LOpWS 'l'VXPOS €X€£. TpOJLOS O€ 

~ " \ I ~\ 1 
7Tawav aypEL, XI\WpOT€pa O€ 7To£as 

" 8 I ~., \ I ,~, 
€JLJL£. TE vaK'1V 0 • OI\LyW 7TLO€V1]V. 

16 ,/... I .,,,, )/ 

'l'awoJL EJL aVTf[-. 

'\ \ ' "\ ,\ t \ , al\l\a 7Tav TOI\JLaTOV €7T€1. KaL 7T€V'YJTa 

I follow the Lobel-Page edition of 1955 for (a) slight changes to correct 
metre or dialect; (b) established emendations that appear in all 
modern editions (Neue's &8v 4>wvelaas at vv.3-4, Buttman's yeAalaas 
at v.15, Hoffmann's opT)JLJL' at v.l1); and (c) Lobel's division of Codex 
p's lliaKtXV at v.9. I depart from the Lobel-Page text in the following 
instances: 
2-3 T€ T' laMvEL Page, for TOL'aveL P. 5-6 T6 8-r] K£V ••• 1-" €v Wills, 
for T6 JL~ €JLav MS. 7 ELal8w Seidler, for at8w P; fJpoxEws MS; a€ Seidler, 
for JL€ (supra n.7); 4>wvas Seidler, for 4>wvas P.47 9 YAwaa' chr€ay€ 



GARRY WILLS 

Beattie, for yAwaaa €ay€ P: cf Galen's use of a7Tayvva8aL in Oribasius 
46.6.3 (Bussemaker-Daremberg). 13 KaS U IL' Ahrens, EX€L Page, for 
EKaS€ IL' ISpws tPvxpos KaKx€aTaL P. 15' mSd'T} v Ahrens, for 7TtS€VU'T}V P, 
interpreted in Heller's and Beattie's sense (see n.13 supra). 16 II 
(n.l3). 

Catullus, Verses 13-16 

193 

Welcker's interpretation of lines 5 and 6 in Sappho greatly reduces 
the contrast presumed to exist between her poem and Catullus', a 
contrast between unstructured, almost random reactions on the part 
of Sappho and consciously wrought, Hellenistic self-examination in 
Catullus. An interpretation that brings the poems so close together 
suggests the possibility of further resemblance,48 and, in particular, 
brings up the question of Catullus' last stanza, which seems to depart 
entirely from his exemplar. 

Most critics approach Catullus' last stanza head on by way of its 
most famous problem, the meaning of otium. The far reaches of poli­
tical speculation,49 of cultural history,50 of moral philosophy51 have 

48 The most frequent use of one poem to explain the other has, naturally, been the 
attempt to fill the lacuna in Catullus (v.S) by consulting Sappho. Most attempts at restora­
tion add vocis to nihil (v.7) on the model of cpwvas otiS' EJJ-vocis in ore (Ritter), Lesbia, voris 
(Friedrich). pectore voris (Pleitner), vocis amanti (Meissner), tum quoque vocis (Lenchantin). 
Those who. like Lobel and Page. accept Danielsson's emendation of the Sappho text, will 
see no cogency in this. But, as we saw in n.47. there is reason to keep the MS reading of 
Sappho with the sense 'I cannot hear'; and the same reasons make us expect the same 
sense in Catullus. Since he was careful to make aspexi correspond with spectat, one expects 
that, in a poem so logically antithetical (see n.56 infra), he would balance audit with some 
reference to his own inability to listen identidem. Thus the sense of the lacuna is best sought 
in some restoration like vocis amatae, "the voice I long for does not reach me." For the loved 
voice that is no longer super mi, compare the imago of Aen. 3.489 which is mihi sola super 
(Le., superstes). 

49 Passerini (Stltal 11 [1934] 52) extracted from Hellenistic political thought a definition 
of otium as 'Tpvcp~, which leads to the hybris (nimiumque gestis) that destroys all forms of 
government, whether monarchy (reges) or democracy (urbes). As Tietze noted (362) this 
slights oligarchy. and therefore can hardly reflect Hellenistic categories; but Ferrari 
67-9, E. Paratore (Catullo 'poeta doctus' [Catania 1942] 144) and Bongi (Aegyptus 26 [1946] 
107-9) accepted Passerini's semi-scholastic exegesis of the stanza. 

50 CJ. Tietze 353-4 on the shifts in meaning the word underwent in the transition from 
the Republic to the Principate. 

61 Tietze (354-62) thinks the stanza draws on an ethical teaching concerning voluptas as 
an expression of libido, and that the poem moves, painfully, toward the readiness for 
renunciation expressed (painfully) at Cat. 76. Ernst Bickel (RhM 89 [1940] 210), Barigazzi 
425-6 and Jachmann 19-25 accept a version of this position. as do Baehrens and Kroll in 
their editions. P. Giuffrida (L'Epicureismo nella letteratura latina II [Torino 1948] 245-65) 
thinks the fourth stanza is an orthodox Epicurean denunciation of anything that prevents 
aTcxpagtcx. 



194 SAPPHO 31 AND CATULLUS 51 

been traversed in the search for an answer to this problem, as have 
the poet's psychological heights and depths, the "spiritual autobio­
graphy" of Catullus as fancy reconstructs it.52 Yet no scholarly con­
sensus has been reached. 

Perhaps the way to solve the problem is to choose a new point of 
entry into it. The nature of otium can be judged from its effect, which 
Catullus describes by calling it moles tum. Is this only a vague word for 
malaise and moral degeneration? So many have thought: Catullus is 
giving us his "tragic flaw" in the word otium, a flaw that critics connect 
with his self-rebukes for loving Lesbia. The flaw keeps him from his 
proper neg-otium-his law studies according to Wilamowitz (59), civic 
virtue according to Passerini and others (see n.49), the renunciation of 
Lesbia according to Tietze and Bickel (n.51). 

But nothing in the rapturous love poem prepares us for this sudden 
self-reproach. The break that occurs at v.13 is, in this interpretation, 
too abrupt. The final stanza becomes a foreign thing added on, not 
an integral part of the poem as it has developed to this point. Further­
more, the comparison of his own distress with the fall of kings and 
cities is comically pretentious if this is a bit of straightfaced moralizing. 
Self-contradiction is, admittedly, a Catullan theme; but the amo and 
the odi must be fused, or painfully intermingled, if they are to be 
expressed artistically. The moralizing approach to this poem makes 
Catullus append a lifeless, impersonal odi to a vivid and lengthy amo; 
and the two seem so disjunct that critics are driven to desperate 
expedients-attributing them to different speakers,53 different per­
sonae,54 even to different poems.55 But, as we shall see, certain formal 

62 The most famous attempt to argue from the course of Catullus' affair with Clodia/ 
Lesbia is that of Wilamowitz 58-9, who finds in this poem the young man's first declaration 
of his love. Snell 71 would, like Tietze and others, place it later in the hypothetical history. 

63 Kalinka (supra n.4, p.163), Immisch (SBHeidelberg 1933/4, pp.13-7) and Gallavotti (AeR 
[1943] 14) think the last stanza is spoken by Lesbia, who gives her young admirer some 
altruistic advice. 

U The last stanza is spoken by Catullus' genius or alter ego according to Birt (Phil%gtls 
63 [1904] 445-6). It is spoken by the real Catullus (as opposed to the not-yet-involved 
"translator Catullus" of the first three stanzas) according to A. Goldbacher (WS 21 [1907] 
113). Kranz (Hermes 65 [1930] 237) takes a position resembling Goldbacher's. R. Katich: (Ziva 
Antika 8 [1958] 31) thinks Catullus speaks through two "translator-selves" so juxtaposed 
that they reveal the feelings of the "real" Catullus. 

66 This view takes two forms: that the last stanza strayed into place accidentally (the 
view, in the last century of Bergk and Ellis, among others; in 1961, of Fordyce) or Fried­
rich's suggestion (supra n.8, p.237) that the fourth stanza is an addition, a kind of post­
script by way of palinode, written after disillusionment had set in. 



GARRY WILLS 195 

ties with the rest of the poem. show that we possess the last stanza in 
situ. Is there a way to make better sense of it in that setting than the 
moralists have so far done? 

Return to that word molestum: what is it that impedes or afflicts 
Catullus? EVidently the same thing that makes him miser (v.5), and 
that misery is defined by contrast with the beatitude of the par deo. 
It is not love that tortures Catullus, but his debility in the presence of 
Lesbia. He does not flee from her but seeks her presence, only to find 
that he collapses at a glance and cannot enjoy the continued gazing 
and listening of the man first described. We have noticed already how 
proleptic misero (5), picked up by mihi (6) and mi (7), is set off from 
the opening of the first three lines (Ille ... ille ... qui). The anaphora 
in the first stanza is even more strikingly opposed in the line open­
ings of the last: Otium ... otio ... otium. The negotium suspended by 
Catullus' languor is that of the man in the first stanza; he is beatus, 
Catullus miser; he plies his senses effectively (spectat et audit), Catullus 
loses all sensation. The poem ends with the reverse image of the man 
who opened it-the pining distant lover set far off, contrasted with 
the man who sits near to Lesbia and to her laughter.56 

So much the structure of the poem, its mere shape and sounds, can 
tell us about the last stanza. But two questions remain. Why does 
Catullus call his weakness otium, which suggests deliberate remissness, 
rather than by some word for involuntary paralysis? And why does he 
indulge the heroics of linking his failure to the fall of kings and cities? 
The answer to both questions is to be sought in the same place. The 
first one explains, I think, the unwillingness of most critics to trust the 

56 The structural resemblances of the first and fourth stanzas were traced by Ferrari 70: 
the last stanza echoes the first not only in the anaphora of the first three lines in each, 
but in the climactic arrangement of the first and second lines (par . .. superare and Otium ... 
moles tum ... otio exsultas nimiumque gestis), and in the pacing of the phrases (lines 1 and 2 
are end-stopped, while 3 is enjambed, in each stanza). See also n.3? on the preparation for 
the last stanza accomplished by the escalation of artifice in vv.9-12. Tietze 367 found an 
ironic response to Ille . .. ille ... qui in the iterated unhappiness of vv.13-15. But he and 
Ferrari both sought in the last stanza a renunciation of the ideal offered by beatus in the 
first. The technical contrasts are far more telling if we hold that in the fourth stanza. as in 
the second and third, Catullus is yearning toward the state of the man in the first, echOing 
it from his pole of powerlessness. E. A. Fredrickson (TAPA 96 [1965] 161-2) realizes how 
important is the contrast between the poet's situation in the last stanza and the joy of ille 
in the first; but he thinks the word sedens points to a distinction between the otium of a love 
satisfied and that of unsatisfied love-an aimless distinction, not otherwise attested, and 
obscured by the mention of "kings and cities" (which have nothing to do with the satisfac­
tion or frustration of love). 

3--G.R .B.S. 



196 SAPPHO 31 AND CATULLUS 51 

poem's own antitheses as guides to the meaning of otium. It is assumed 
that miser refers to what Carullus cannot help (debility in Lesbia's 
presence, of jealousy), and otium to what he can do something about 
(breaking off the affair, refusing to see her, etc.). But Sappho's poem 
turns back on itself with the assertion that the apparently involun­
tary symptoms can be borne; and it seems idle to debate whether 
ceasing to love at all or bearing love's ecstasies is more within the 
scope of the will. When the poet braces himself to overcome a weak­
ness, it should be the weakness analyzed in the poem (the one that 
makes him less able to bear Lesbia's presence than the model listener 
of the first lines), not some weakness imported from other poems 
(loving Lesbia after a series of degradations). 

But otium says more than mere ·weakness'. It is a moral term, 
related to duty. Why is Carullus bound in duty to bear the ecstasies of 
love? Posed that way, the question suggests its own answer: Catullus 
is talking about a lover's code-one that embraces suffering and con­
demns desertion under trial. This is his «heroic code" (a point empha­
sized by reges and urbes). Love is his negotium, and he must be fit for 
all its encounters. 

Otium is properly used of remissness in war (cf. Tac. Ann. 14.39, 
Rist. 4.70), of the warrior who is segnis. And love-as-war is a topos as 
old as Sappho (fr.1.2S), a topos latent in the heroic language with which 
she opens the very poem Catullus is translating. Though the com­
parison is used throughout classicalliterature,57 the best summary of 
its possibilities is Ovid's Amores 1.9 (Omnis amans militat). There he tells 
us (vv.27-8) that the lover and the warrior are miser in precisely the 
same way, and he gives a list of heroic parallels-Achilles, Hector, 
Agamemnon, Mars himself-to prove the point. But the poem's 
conclusion brings us closest to Catullus, since it presents love's activity 
precisely as a renunciation of otium: 

Ipse ego segnis eram discinctaque in otia natus; 
Mollierant animos leetus et umbra meos; 

61 See the examples in K. Preston, Studies in the Diction of the Sermo Amatorius in Roman 
Comedy (diss. Chicago 1916) 50; R. Pichon, De Sermone amatorio apud Latinos elegiarum scrip­
tores (Paris 1902) 201-2; and J.-M. Andre, L'Otium dans La vie moraLe et intellectuelle romaine 
(Paris 1966) 421-5. Andre (p.12) even argues that the original sense of otium was martial, not 
commercial; it meant 'an intermission in militia', soldiers' 'leave' or 'peacetime service' 
(cf Propertius 4.4.79). 



GARRY WILLS 

Impulit ignavum formosae cura puellae 
Iussit et in castris aera merere suis. 

Inde vides agilem nocturnaque bella gerentem. 
Qui no let fieri desidiosus, amet! 

197 

Catullus' poem closes, then, not with the wretched morality of disgust, 
but with the playful moralism of erotic poetry. He must have a heroic 
ardor and energy, great enough to bear love's most excruciating bliss. 
After all, no greatness has been won or kept by those who did not 
overcome their pusillanimity; Otium et reges prius et beatus I perdidit 
urbes. Kings and cities must rule; he must love (cf Sappho fr.16). 

This interpretation has the added advantage of making Catullus' 
"break" at v.13 resemble Sappho's (at v.17) in substance as well as 
form. Many have noticed that the poems agree in the way they de­
part from the list of symptoms with a strong push toward moral 
resolution-Sappho with a turn from description of lost control to a 
statement that control is possible, Catullus with the self-exhortation 
of his little sermon on otium.58 Both steady themselves for encounter 
with the sense-bereaving object of their passion. They tread an 
elaborate, courtly path of approach to their lovers, all the while say­
ing that such a privilege and disabling pleasure is beyond their power. 
Each uses as foil a man who is less affected than they are by this bril­
liance, but whose feat they mean, in the long run, to equal. The poems 
shape the same thought, hers dancingly, his as a thing baroquely 
sculpted. 

THE JOHNS HOPKINS UNIVERSITY 

February, 1967 

68 Cf Neue 35, Barigazzi 424, and R. Lattimore, CP 39 (1944) 184-5.