CC BY
12
M. N. Kazanskaya
SAPPHO FR. 183 VOIGT (= ALCAEUS FR. 412 VOIGT) AND ANCIENT SCHOLARSHIP ON THE alyiq
Статья посвящена слову каттрцд (катарцд), которое, как сообщают Порфирий и Евстафий, обозначало у Сапфо и Алкея сильный, низвергающийся с высоты, ветер. Разбор античных свидетельств показывает, что это слово упоминается только в контексте толкования понятия эгида александрийскими филологами, которые ссылались на каттрцд при построении лингвистической аналогии; сама же глосса является результатом переосмысления выражения ave/иод кат ород у Сапфо (fr. 47 Voigt). В заключение приводятся аргументы в пользу атрибуции сочинения об эгиде Аристофану Византийскому, как правило, отвергаемой современными учеными.
Ключевые слова: каттрцд (катарцд), Сапфо, Алкей, эгида, гомеровская экзегеза, Порфирий, Евстафий, Аристофан Византийский.
Sappho fr. 183 Voigt (= Alcaeus fr. 412 Voigt) is a single gloss cited by Porphyry, Eustathius, Hesychius, and Theognostus. The word is a hapax legomenon, used by no other author among those whose texts have come down to us. Due to a divergence in the sources, most editors of Sappho and Alcaeus choose to give the fragment under a double spelling: каттрцд (sive) катарцд1.
Porphyry and Eustathius state that Sappho and Alcaeus had used this word for a specific type of strong wind or gale, descending from above, and Hesychius and Theognostus mention it without specifically discussing its meaning. Modern scholarship has not paid much attention to this gloss: the preference is usually given to the spelling катюрцд2, and the word has been explained as stemming from the adverbial катю 3 or as derived from the noun ород with a lengthening
1 Thus, besides Voigt, Lobel and Page (1968: 102) cite both spellings for Sappho fr. 183, but only KaTmpng for Alcaeus fr. 412 (Lobel, Page 1968: 284); Campbell (1990: 418) gives both variants for Alcaeus fr. 412, but only KaTmpng for Sappho fr. 183; Liberman (1999: II, 183) cites both spellings for Alcaeus fr. 412, although an interrogation mark accompanies the reading KaTaprfc. Bergk (1882: 137, 188), as well as Reinach and Puech (1937: 150, 318) retain only KaTapng.
2 See Funaioli (1973/1974) for an argued defence of this spelling.
3 Schwyzer (1990: I, 632) cites KaTmpng among the small group of words formed with the adverbial KaTm- instead of the prefix KaTa-, without
of the root vowel comparable to that of vmapsia (Il. 20, 218)4. Among those who accepted carapng J. Wackernagel may be mentioned5.
For single-word fragments (and especially ones that are hapax legomena) the context is often so difficult to reconstruct, that the accompanying remarks of the grammarians, however scanty, acquire particular importance. The object of this paper is to examine the existing testimonies in order to trace the competing interpretations of the word Kazmpng (carapng) in Antiquity, to explain the fragment's connection with Homeric scholarship on the aiyiq, and to reconstruct the gloss' transformations.
Overview of Ancient Testimonies on Kaxrapn^ (Kaxapn^)
Among the four references to carrnpng (icaxapqq) in ancient scholarship, two appear in commentaries on the Homeric aiyig:
slsyov Ss aiyiSag Tag vvv KaxaiyiSag twv a(poSpmv Kai avvsa-Tpappsvmv wvev^axmv Kai dpa Karapaaaovrmv, oiov Ti Kai sv rovrm Qsmpsnai Isymv o noinrfg^ (citatur Il. 11, 297-298). AlKaiog Ss nov Kai Zanqw Tdv toiovtov avs^ov KaTwpr) Isyovaiv and tov Ka.T<xxpspr\ T^v op^v exeiv (Porhyr. Quaest. Hom. ad Il. 2, 447 p. 47 Schrader)6.
sv Ss Toig Toiomoig Isysi o arndg (scil. ApiaTo(avr)g o ypa^axwdg) Kai oti Td avvsaTpappsvov wvsvpa Kai Kaxapaaaov avs^ov Kaxapn Isyovaiv o AlKaiog Kai ^ Zan(<b Sia Td Ka.T<xxpspr\ op^ffv s/siv (Eustath. ad Il. 603, 35 van der Valk) .
discussing its sense; the gloss is absent from the discussion of this type of formation in Chantraine (1999: s.v. Kara, Kara).
4 This interpretation was first proposed by E. M. Hamm (1957: 88, §171b): "Zu opog gehört vielleicht der Windname cazoopyg als 'von den Bergen herunter (kommend)'"; she is followed by H. Rodriguez Somolinos (1994:
115).
Wackernagel (1889: 41) placed cazapqg avs^og among adjectives formed with the suffix -rfpyg, ultimately derived from apapimcm. The idea is rightly criticised by M. P. Funaioli (1973/1974: 129): "Il Wackernagel [...] lo con-netteva con apapiacm, non si vede quale significato attribuendogli: il confronto con Kairfpyg dovrebbe escludere una tale relazione, non potendosi, mi sembra, parlare di un vento 'ben adatto' o 'provviso' di alchunche".
6 "[The ancients] used to call aigides what we now call hurricanes (katai-gides) from violent blasts that are both twisted up and simultaneously smash down. A phenomenon of this kind is meant by the poet, when he says: [...]. It seems that Alcaeus and Sappho call a similar wind katores, owing to its downward tendency" (translations, here and further, are mine - M. K.).
7 "Among these considerations, the same Aristophanes also says that
Similarities of wording in these two notes are striking: both commentators describe the wind with regard to its direction (катараааю and катюфЕрцд), and vary the word avsjuog in the recurrent expressions аи\Еатра^^Ё\о\ nvsv^a and ор^ц катюфЕрцд. Moreover, Porphyry's and Eustathius' wording suggests that the text(s) referred to contained not only каттрцд (катарцд), but also avs^og, in other words, that каттрцд was an epithet rather than a proper name8. It is evident that the two notes stem from a single tradition, and it has even been suggested that Eustathius was directly using Porphyry's commentary at this point9. But these similarities also set off the differences between the two commentators, which include (a) the already mentioned divergence in spelling, каттрцд in Porphyry and катарцд in Eustathius, (b) the uncertain intonation of Porphyry's reference (ААкаюд дё пои кал Напфт), which is eliminated from Eustathius' account, and most importantly, (c) the fact that Eustathius names explicitly Aristophanes of Byzantium as the source for the parallel. As regards the first point, Eustathius' катарцд has been seen as a lapsus calami under the influence of катараааоу10. The second point, to our knowledge, has not been explicitly remarked upon. Finally, the attribution of the discussion on the Lesbian poets' name for the wind descending from on high to Aristophanes, Aristarchus' predecessor at the head of the Alexandrian library, editor of Homer and some of the lyric poets, has in general been mistrusted, or even ignored by modern scholarship.
This mistrust originated with the first editor of the full collection of Aristophanes' fragments, August Nauck, who examined the reference in Eustathius 603, 28-35, emphasizing that the existence of
Alcaeus and Sappho call the blast that is twisted together and dashes down 'the katares wind', as it has a downward tendency".
8 The fact that the word avspog must be taken as part of the poetic fragment was seen by Bergk (1882: 137, 188); Lobel (1927: 65); Reinach, Puech (1937: 318); Lobel, Page (1968: 284).
M. van der Valk (1976: II, ad 603, 35): "Patet eum Porphyrii scholio usum esse, nam Eustathii verba (l. 34s.) zd avveazpappévov wvsvpa Kai Kazapáaaov religiose, ut ita dicam, exprimunt Porphyrii locum (1. 41, 15s.) zwv... avvsazpappévmv wvevpázmv Kai apa Kazapáaaov (sic!)"; similarly Rose (1971: 163).
10 Thus, Maria Paola Funaioli (1973/1974: 130): "Gli stessi testimoni offrono la soluzione: nella fonte di Eustazio facilmente il Kazapáaaov immediatamente precedente avrá transformato in Kazápr¡q l'originario Kazojpyg. Kazapáaam, voce tecnica nella tarda prosa per l'abbattersi degli elementi, e bensi presente anche in Porfirio: ma qui meno pericoloso perché piu lontano".
the treatise nspi aiyiSog is recoreded by no other source, and that a large part of the passage is not consistent with what we know of Aristophanes' method. Eustathius indeed refers to Aristophanes in his discussion of the Homeric aiyig, crediting the grammarian with two arguments, (a) the idea of allegorical nature of the aiyig, and (b) the parallel from Sappho and Alcaeus:
o Ss ypa//a.TiKdg ApiaTo(pavr\g ISiov Ti avyypa//a nspi aiyiSog nposvsyKwv alla ts Tiva Isysi ov navv yla(vpa Kai oti "O/npog Toig naOsai Kai Toig Si' avtwv anotslov/svoig npay/aaiv o/wvv/ovg Tivag s(iam Sai/ovag, siSwlonoiov/svovg /vOiK&g, s(' mv Kai Td anotslovv, fyovv napaaKsvaatiKov, vosiTai Kai Td anotslov/svov KaKdv svapy&g. "Epwg yovv Kai nlovtog Kai spig Kai vfipig Kai Ssi/og Kai (OPog Kai KvSoi/dg Kai Ppovtj Kai aatpanj Kai alla /vpia o/mvv/mg Srflovai Tf)v siSmlonoii'av ts Sai/ovimSr) Kai Td sKsiOsv av/wtw/a. sv Ss Toig ToiovToig lsysi o avTdg ktI. (Eustath. 603, 2835 in Il. 5, 738)11.
Noting that the list of abstract notions allegorized as deities is assembled not only from various Homeric passages, but also from later poets, such as Hesiod and Pindar, Nauck expressed a doubt as to whether the interpretation of the aiyig as a personification was not more likely to stem from the philosophical tradition, rather than from the grammatical, and suggested that it might have been Aristoteles' name that was misspelled as Aristophanes1 , admitting, however, that
11 "For the grammarian Aristophanes, in a treatise of his own published on the subject of the aegis, expresses some not very subtle ideas, and in particular that Homer appoints to emotions, and to the acts resulting from them, some deities bearing the same name, whose images he fashions in the mythological manner and to whom he ascribes not only the action itself, that is, the capacity to cause it, but also evidently the resulting evil. Thus, Love and Wealth, and Strife and Outrage, and Fear and Terror and Uproar, and Thunder and Lightning, and a thousand others through their homonymy (with the phenomenon) point not only to the process of personification in the guise of deities, but also to the resulting features." The same idea is formulated in very similar expressions in Porphyry (Quaest. Hom. B 447, 20-21), although its source remains unnamed.
12 "Sed haud scio an omnis haec observatio a consueto Aristophanis more dissideat ac philosophi magis ingenium sapiat quam grammatici" (Nauck 1848: 272). The suggestion that it might have been Aristoteles is indeed plausible from the point of view of palaeography, as the two names do tend to be confused: thus, in schol. b(BCE3E4)T in Il. 20, 252: pyai yap fapiaTo(avr)gf • „yvvij avSpdg snupOovwTspov, /s/^i/oipotspov, (piloloi-Sopov, (ilonlrfKTiKWTspov., /allov apiSaKpv ", whereas the quotation comes from Aristoteles' Historia Animalium (608b8; cf. schol. bT in Il. 2, 447c
this doubt was based on conjecture, and that evidence to the contrary, confirming Eustathius' attribution, could also be found in other scholia13. Nauck's treatment of the reference to Sappho and Alcaeus is much briefer: he speaks of the uncertainty of the spelling катарц, and insufficient evidence as regards the sense of the gloss.
August Nauck's scepticism proved contagious: after him V. Rose, L. Cohn and later R. Pfeiffer have discarded much more vehemently the attribution of Пгрг агугёод to Aristophanes of Byzantium14. An even more extreme approach is found in Slater's edition of Aristophanes' fragments designed to replace that of Nauck. As Slater (1986) limited himself to fragments that he considered certain, he omitted to make any mention of the reference, despite the explicit Eustathius' explicit mention of Aristophanes.
The two shorter references to каттрцд appear in Hesychius (к 91) and in Theognostus (Canones 248), who cite the word is for its curious form, giving no information about its source. Hesychius cites the gloss, explaining its meaning as "катю psnrnv": this should be viewed not as a lexicographical definition, but as a rather unsophisticated attempt at etymology, where the verb is chosen for its initial letter p-15. The second reference, though containing no
where the previous scholium in II. 2, 447c shows that Aristophanes, not Aristotle was meant, and schol. T in Il. 24, 340 where Aristotle's name has been corrected to ApiaTotpdvyg by Rose 1971: 167).
However, A. Nauck's and V. Rose's identification of Porphyry's source with Aristoteles, because the philosopher's name will later appear in the note on Gorgon's head on Athena's aegis (Quaest. Hom. E 738, 2) does not seem convincing. Two aspects in particular seem to suggest that these two notes stem from different traditions: (a) the fact that the note on Il. 2, 447 avoids direct association of the aegis with a shield, which is central for the argument in Quaest. Hom. E 738, 2; and (b) Porphyry's polemic statement in Quaest. Hom. B 447, 1-4 that the aegis is primarily Zeus', not Athena's weapon.
13 "Sed haec omnia fateor coniecturis niti, quibus firmiter insistere eo minus
licuerit, cum Aristophanis nomen tueri videatur locus Schol. A, 4: ^ Sinlr, oti nolspoio Tspag Tdv siSmlonoiovpsvov nolspov, Tdv noirTiKov tov evspyovpevov nolspov —. oi Ss atrcpanrfv (aoi Tf)v spiSa (l. Tf)v aiyiSa) (spsiv, mg Kai ApioTofidvrg• (roi yap (K, 5.) ,,<bg S' OTav aaTpdrnrfl Tsv/mv f\ nolvv, r\s noOi nTolspoio" (Nauck 1848: 273).
15 See Rose (1971: 163), Cohn (1881: 287 n. 4), Pfeiffer (1968: 285-286). 15 The verb psnm is not frequent in Hesychius, who uses it to define ¿Tspopsneg (s 6572), iaopponov (i 979); o&ppsnrfg (o 959); xapaipsn&g (x 136), in the entry nsnoiwTai (n 1488) and for some derivatives from penm. As regards the adjectives mentioned above the etymology is correct, which is not the case for KaTmpng.
information as to the sense and use of the word, is more interesting: Theognostus, grammarian of the 9 century A.D., mentions caxmpng among examples of adjectives and names ending in -mp^g:
ta Sia tov qphe Sia tov q /syalov ypa(si Tjv napalijyovaav• oiov, KaTmpr\g- Aimpng ovo/a Kvpiov avtmprg vsaamprg• vlrmprg^ vsmpng I AvKwpsog ysviKj and rfg AvKwpsvg svOsiag sK(spo/svr ovaa TsTiiprfTai (Theognost. Can. 248) .
The material which appears in this late source in all probability goes back to a much earlier period. We find two similar, though much shorter, entries in De accentibus preserved under the name of Arcadius17 which represents an epitome of Aelius Herodianus' De prosodia catholica, and it is highly likely that Theognostus' entry, as well as Suda's erroneous lemma "Katmpng: ovop.a cvpiov" (Suda c 1116)18, was actually derived from Herodianus. Indeed, August Lenz in his edition of Herodianus in the Grammatici Graeci went so far as to reconstruct a single entry, adding to Arcadius' examples those cited by Theognostus:
sti ta Sia tov qphe [Papvvovtai], vsmpng, avtmprg [Arcadius, 135], Katmprg, vlr\mpr\g, */saamprg [Theognost. Can. 248 where /saamprg is a conjecture], Aimpng ovo/a Kvpiov. AvKmprg [conflation of Arcadius, 28, and Theognostus]. i Ss AvKwpsog ysviKj napa Kalli/a/m and rfg AvKwpsvg svOsiag sK(spsTai [Theognost.] (Hdn. De pros. cathol. 3, 1, 72).
The choice of examples points to the first centuries A.D. as the most probable date for the original source: besides the name Avcwpyg that both Pseudo-Arcadius and Theognostus explicitly connected with Callimachus (see Hymn. 2, 19), the adjectives
16 "Words ending with -mpng, spelled with omega, carry the stress on the penultimate syllable: thus, KaTmpng, the proper name Aimpng, avTmpng 'acting simultaneously', vsaampng [unattested and probably corrupted], ЬХцшрцд 'woody', vsmpng 'fresh'. The genitive AvKmpsog, derived from the nominative AvKmpsvg, is [also] preserved".
17 "Eti та Sia tov qphe [fiapvvsTai] Aimpng AvKmpng, onsp KaXXi/iaxog o^vvsi (Arcadius, 28); sti Kai та Sia tov qphe • vsmpng vsmpsg, avTmpng avT&psg (Arcadius, 135).
18 KaTmpng: ovo^a Kvpiov results no doubt from a misunderstanding of KaTmpng• Aimpng ovo^a Kvpiov which the compiler of the lexicon divided into two entries (Aimpng, Suda S 1248, and KaTmpng, Suda к 1116), as was noted by Funaioli (1973/1974: 129 n. 1).
аЬттрцд and ЬХцтрцд are found in Hellenistic poetry19; similarly, the name of the Homeric hero Аюзрцд was subject to scholarly inquiry, as the following entry in the Etymologium Magnum shows: "Аютрцд: оуора кьрюу. ц уар кара тц\ Sia жровкт\ каг то орт, дшорцд [...] ц тд and тоь ьХц ЬХцюрцд жарауюуо\, оьтюд каг and тцд Аюд уЕУжцд Аклрцд- 'А1&ркод аХкщод ьюд' [Il. 17, 429]" Etym. Magn. 280)20. Among these examples, only уеюрцд remounts to Sophocles (El. 901 and O. C. 730), whereas the others point to Hellenistic or Roman times.
While these shorter references in the grammatical tradition tell us little of Sappho and Alcaeus, but they do show that the scholars of late Antiquity regarded каттрцд as the preferable form; they also have a bearing on the approximate dating of the discussions of the form каттрцд in the grammatical tradition.
Катюрп^ (катару) and the Homeric scholarship on the aiyi<;
When examining the two larger notes in Porphyry and Eustathius containing a definition of каттрцд (катарцд), one must keep in mind that these occur in commentaries on the epic aegis, and that the gloss had in all probability been first cited by Homeric scholars. In order to reconstruct the place and function of the Lesbian gloss, it is necessary to briefly summarize ancient theories concerning the агугд.
The question of the exact nature of the aegis, epic attribute of Athena and Zeus, has been written about extensively in modern times21 and was a debated problem already for the ancient readers, owing to its multiple functions and conflicting properties. The first explicit interpretation of the aegis remounts to Herodotus who states that the aegis as a goat-skin mantle on Greek archaic statues of Athena was taken over from the article of clothing worn by Lybian
19 The adjective ammprg is once attested in Callimachus (fr. 671 Pfeiffer), and vlrmprg, also a hapax legomenon, appears in Nicander's Theriaca: Toig Sf xrpapa Koila Kai vlrmpsag svvdg / Ksivmosig... (Nic. Ther. 55).
20 "A proper name, either constructed from the prefix Sid and the verb opm, Siaoprg [...], or as vlrmprg is a derivative of vlr, thus Aimprg is a derivative of the genitive form Aiog: 'the mighty son of Diores' (Il. 17, 429)".
21 The modern scholarship on the aegis is immense, and we shall give only a few references. An analysis of proposed etymologies and the Homeric usage, with references to literature may be found in the article, remarkable for its strictness and conciseness, written by M. van der Valk, s.v. aiyig in LfgrE (1955: I, 253-255). Among later references, see Fowler (1988: 103113); Deacy, Villing (2009) with literature; most recently Janda (2014: 460-476).
women (Hdt. 4, 189). For a while the idea of агугд as stemming from аг<д seems to have been the predominant theory. However, the etymology was contested by Alexandrian scholars, who proposed to connect агугд with the verb агааю22, seeing in it an instrument used by Zeus to provoke storms. The scholia show that this second interpretation gained progressively in authority23, possibly because it seemed more poetic and better in keeping with the majesty of Zeus and Athena, than the goatskin; one of the most important arguments that regularly recurs in the scholia is that the current Greek uses the word катафд for hurricanes, which is interpreted as a prefixed derivative of агугд24. In their efforts to weaken the association with аг<д the scholiasts could go far: for example, an allegorical commentary on the god Pan explains his goat feet as a symbol of hurricanes25. However, the "meteorological" explanation could not account for all the aspects of the Homeric aegis, and a more complex theory was invented by Dionysius Sidonius, who distinguished two types of aegis (see the exegetical scholium to Il. 15, 229b).
Keeping this context in mind, we can now return to the question of the double spelling каттрцд (катарцд), and of possible origins of the gloss. It should be noted that the concurrence of several words beginning with кат- in Porphyry' and Eustathius' explanation of the word used by lyric poets for the katabatic wind may not be accidental. In scholia such agglomerations of paronyms often indicate that the scholiast (or the hypomnemata he used) is explaining the word's etymology, or even that he has conflated into one note several etymologies that his predecessors had proposed26.
22 This usage of aiyig, though rare, is attested: cf. Aesch. Choeph. 592; Pherecrates, fr. 117 (Kassel).
23 Besides Porphyry and Eustathius, cf. schol. A in II. 17, 594a; schol. Gen. in Il. 1, 205; the idea also influenced the Roman scholars (see Servius on Aen. 8, 354). It is Aristarchus who is usually credited with theory of the unique aegis that served as attribute to Zeus (see Lehrs 1882: 190; Schwartz 1881: 437).
24 Cf. schol. A in Il. 2, 148a; schol. A in Il. 4, 167b; schol. A in Il. 17, 594a; schol. in Od. 15, 293; Et. Gud. 36, 20 Stephani).
25 näv KaXsizai, ozi änäzwp ¿azi, ацрггоу Se tov navzög, oOsv Kvpiwg npoaayopsvszai näv. avvoiKsiovzai Se Kai щ цорщ avzov тф nspis/ovzi- Kai zd psv zwv Kspäzwv anopipnpa ffXiov Kai asXr\vr\g pqvioKovg qaaiv sivai [...] zf)v Se avpiyya zwv ¿v тф KÖaptp Kvsvßäzmv ¡лщцаы sivai. zd Se aiyinovv zäg KazaiyiSag Kai and zwv vs(p&v yivopsvag nspi zdv aspa uszaßoXäg (Schol. vet. in Theocr. 1, 34).
6 To cite one example, scholium to Od. 2, 315 contains a plethora of similarly sounding words without an explicit statement of purpose: ptfriq
When seen under this light, the presence of the expression cazrnpsptf t-v op/u-v and the participle caxapaaaov in both commentators suggests that Eustathius' carapng is not a lapsus calami, but a variant spelling, adapted to one of the proposed etymologies (carapng from caxapaaarn), and existing on a par with cazmpng, form linked with the adverb carni consistent with the wind's downward motion (opu-carmtpsprfg).
As fluctuations of a> and a are not normally associated with the Aeolic dialect, we are obliged to reconstruct a distinct form that could have given birth to the two spellings in the scholarly tradition. And the best candidate is actually attested among Sapphic fragments. As has already been noted by Nauck27, avsjuog Karwpqg is very close to an expression found in Sappho, fr. 47 (Voigt):
"Epog S' ¿rivals /oi (psvag, mg avspog Kar opog Spvaiv ¿/nsTwv28.
Recent editors (Lobel and Page, Voigt, Campbell) are certainly right to print car opog without the sign of elision29: in continuous writing, KATOPOS, this expression is the most economic explanation of the divergent readings caxmpng / carapng, and also the form that allows us to trace different strands of grammarians' argumentation.
Although there is very little doubt that in the surviving fragment 47 KATOPOS meant "down the mountain", the group could easily be reinterpreted as a single word, especially since the preposition was written as the monosyllable car, thus becoming an epithet of the
Ss KaiI^¿vog opyj svanoOsmg Kai sppovog^ Za(slig ovvxolog i ptfvig [...] mg ovv napa td aOsvog ^¿vog, ovtw Kai xolog ^¿vwv ^¿vog Kai ptfvig (Schol. in Od. 2, 315f Pontani). However, the comparison with a much shorter note on the first verse of the Iliad shows that this paronymy was due to the etymology of /iijvig: /ijvig napa td /svw /ijvig mg svog r\vig (schol. A b (BCE4) T in Il. 1, 1c).
27 "Tametsi scripturae Kataprg fidem non praestiterim, Sapphonis tamen indicari puto fr. 44 p. 609 Bergk, ubi nunc legitur avs/og Kat' opog Spvaiv s/nsawv" (Nauck 1848: 273).
2 "Love shook my heart like a wind rushing downward on oaks down the mountain". As in Maximus of Tyre the poetic excerpt is woven into the texture of the sentence, rather than presented as a separate quotation, there have been several reconstructions of Sappho's text: thus, Bergk (1882: 103) suggested ("Epog Savt' stiva^sv e/oi (psvag,) / avs/og Kat' opog Spvaiv s/nsawv, and Reinach - Puech (1937: 228, fr. 44) in their edition read ¿rivals taig / (psvag mg avs/og Kat' opog Spvaiv s/nsawv.
29 For the apocope of prepositions in Lesbian poets, especially Kat and ov, see Hamm (1957: 110-111, § 195a) and Bowie (1984: 108-109).
previous word: avsjuog *Kaxopoq. That it was this form that the grammarian (whom Porphyry uses without naming and whom Eustathius identifies as Aristophanes) referred to becomes evident if we consider the larger context of the two commentaries. In Porphyry, whose detailed account, despite occasional lacunae in the development of thought that have to be filled in from the quotations and the larger context, seems to reproduce quite accurately an earlier treatise on the aegis, the mention of Kazmpng follows the statement that the epic aiyig has a counterpart in a much more current word Karatyig, "hurricane"30. This placement surely suggests that the Lesbian gloss appeared as a member in an analogical ^proportion, where the fourth (unprefixed) term would have been ovpog, poetic word used since Homer for a fair wind or breeze31. The ancient grammarians were well aware of the fluctuations in the rendering of the Attic-Ionic diphthong ov in the text of Sappho and Alcaeus: the best example is, of course, the word ovpavog which could be written either with o- or with w-32. Thus, *Kaxopog would have been interpreted by the grammarian who came up with this parallel as a derivative of ovpog, the conjecture being further facilitated by the existence of the adjective snovpog33. Moreover, the same correspondence seems to underlie the variae lectiones preserved by the scholia to Il. 11, 296-298. The generally accepted text of this simile, likening Hector's descent into the battle to a stormy wind, is the following:
amdg S' ev npmToioi peya (povsmv ¿PsPiiKsi,
ev S' Snso' vopivr) vnspaei ioog aellr),
30 sXsyov Se aiyiSag Tag vvv KaTaiyiSag [...] 'AXKaiog Ss nov Kai Zanpm rdv
ToiovTov ävsßov Karmpn Xsyovaiv (Porhyr. Quaest. Hom. ad II. 2, 447 p. 47 Schräder); in Eustathius the gloss KaTapqg is not explicitly linked to KaTaiyig, as the etymology had been explained in an earlier note (Eustath. 193, 38-41 ad Il. 2, 148), and he does not choose to repeat it at this stage. For other references to the equation aiyiSsg ~ KaTaiyiSsg in the scholia, see n. 24 above.
32 See V. Langholf, s.v. ovpog in LfgrE (2004: III, 876-878). 32 opävw (Sappho fr. 52; fr. 54; Alcaeus fr. 338), but mpavm (Sappho fr. 1, 11; Alcaeus fr. 355). See Wackernagel (1970: 136, n. 1), Hooker (1977: 4647). In the Homeric scholia an important discussion of the correspondence between -o- and -ov- is preserved in the scholium A in Il. 2, 153 that 3p3articularly mentions Aristarchus.
E'iO' avsposaaa Tig / ysvoiT' snovpog ¿oriwTig avpa, / ffTig p' anoiKiasisv ek TÖnmv... (Soph. Trach. 953-955). See Kamerbeek (1970: _203); and Chantraine (1999: s.v. ovpog) for the list of words derived from ovpog.
tf ts KaOallo/svr iosiSsa novtov opivsi (Il. 11, 296-298)34.
The scholium of the Venetus A to v. 297 (remounting to Aristonicus and, consequently, to Aristarchus) records an alternative reading vnspovpsog found "in some copies":
vnspasi iaog asllfl: oti sv Tiai ypapsтal „vnspovpsog iaog asllr\'" td Ss avTd Kai sk tov vnspasi ar/aivstai, otav sk tov vnspOs mnov KaTa(sprTai nvorf. tavtag Ss ¿/sig KaxaiyiSag Kalov/sv and tov Katw ai'aasiv (schol. A in Il. 11, 297) .
K. Lehrs corrected the reading of the Venetus A, dividing the single epithet vnspovpsog into vnsp ovpsog (and H. Erbse in his edition followed his suggestion). However, the recognition that the original vKo/uvmua contained the analogy aiyig : caraiyig :: ovpog : KATOPOZ renders this correction not only superfluous, but also misleading. The adjective vnspovpsog is attested in Theocritus (Id. 24, 95) who uses it as a variant of vnspopsog36; however, the commentator writing on the aegis appears to have interpreted it as a prefixed adjective derived from ovpog and proposed it as synonym or substitute for vnsparfg (that he saw as a stray relative of aiyig and of aiaarn). This conjecture is in keeping with the analogy reflected in the Quaestiones homericae (aiyig : vnspasi :: ovpog : vnspovpsog), and the fact that Porphyry who is our best authority for the reconstruction of the vno/uvn/ua on the aegis cites Il. 11, 297-298 between the statement of correspondence between aiyig and caraiyig and the gloss caTrnpqg (Quaest. Hom. B 447, 7-8) does not seem accidental.
The next step in the transformation of KATOPOS would have arisen from its dissociation from ovpog through an alternative
34 "Proud Hector himself advanced among the first, and fell into the battle, similar to a storm rushing from on high that, leaping down, arouses the violet-coloured sea".
35 "Similar to a storm rushing from on high (vnspas'i): [diple], as in some copies it is written "wind rushing from above (vnspovpsog), similar to a storm". The same phenomenon, when the gust comes from above, is denoted by vnspaijg. We also call such winds "hurricanes" (KataiyiSsg), because they rush down".
36 ¿¡pi Ss avlls^aaa Koviv nvpdg a/plnolwv Tig / pipatw, sv /ala naaav vnsp nota/oio (spovaa / pwyaSag Eg nstpag, vnspovpiov, aiy Ss vssaOw / aatpsnrog, "And on the morrow let one of the maids, having gathered my ashes, take them to the rugged cliff and cast them out upon the river, so that they will no longer be in your boundaries, and let her return without looking back" (Theocr. Id. 24, 93-96).
analysis of its inner form, and from observation that neither -ород nor -юрод is a productive formation for adjectives (cf. Risch 1974: 69-73): thus, a form in -цд would have been proposed, just as was the case with уёюрод (Hdn. 3, 1, 200; Hsch. v 422; Phot. Lex. v 164) and У£<арцд (vide supra). It was thus that каттрцд joined the small group of adjectives in -трцд assembled by the grammarians (with the disadvantage of becoming virtually impervious to etymology). The form *каторцд, on the other hand, was reinterpreted as катарцд (cf. lesb. атротод - ion.-att. атратод)31, which afforded the possibility of linking the adjective to катараааю. This transformation will have taken place before Herodianus (2nd century A.D.).
Considerations on the dating of the йяо^пца on the aegis
Given the proliferation of explanations of каттрцд (катарцд) in ancient scholarship, it is likely that the reinterpretation of the group кат ород as an epithet goes back to an authoritative early scholar who worked not only on Homer, but also on lyric poets (in particular, on Sappho). However we might evaluate Porphyry's and Eustathius' SаípоvEg EÍSюХожоющEvоl ¡львштд, the second part of the tradition involving the reference to the Lesbian poets in the context of a linguistic analogy, should be taken more seriously than it has been hitherto.
The terminus ante quem seems to be Aristarchus as the correspondence агугд ~ ката1угд is well attested in Aristonicus scholia. Aristarchus seems to have written on the агугд himself (Lehrs 1882: 190) and might have come up with the equation; however, if our reconstruction of the train of thought in the note on Il. 11, 297 is correct, it implies a polemic reference to a prior edition (ev nai урафвтаг)38. Seen in this light, Eustathius' attribution of the treatise on the aegis to Aristophanes of Byzantium gains in reliability. We can deduce from other sources that Aristophanes' work as editor involved serious philological reflection on the text,
37 Cf. Hamm (1957: 28 § 57a); Bowie (1984: 98-99); O'Neil (1969: 26-28 et passim).
38 As a supplementary argument in favour of an early date for the discussion on aiyig and naTaiyig, the appearance of the rare word ko.t&i£ in Callimachus and Apollonius Rhodius may be cited (Call. Hec. 238, 29; Hymn. 3, 114; Apoll. Rhod. 1, 1203). Stephens (2015: 137, ad loc.) suggests that KaTait, was "probably invented by Callimachus and imitated by [Apollonius]". On Callimachus' practice of responding to contemporary philological exegesis of Homer through his poetry, see Rengakos (1992).
with careful consideration of parallels in other poets, amounting in some cases in separate treatises on a particular philological question, as was the case with Archilochus' expression axvv/usvn rncvxaXq that he connected with Alcaeus' fr. 359 Voigt39. Aristophanes' excellent knowledge of lyric poets could well have suggested to him that the reading *caropog in Sappho's fr. 47 could be used as an analogy to caraiyig. Finally, we know that Aristophanes applied the analogical method in morphology, although the extent of its theorization at this stage is disputed40. None of these considerations can be used as proof, but they speak in favour of Eustathius' attribution. On the other hand, Nauck's doubt that Aristophanes could resort to allegorical exegesis which Porphyry's and Eustathius' lists of deities and the expression siSmlonoiia Sai/uoviojStfg seem to imply, may perhaps be laid to rest, if we recall that Aristophanes would have had to explain the stock epithet Qnaaavosaaa "tasselled", used exclusively with the aegis41: it would have been natural to explain this association with animal fleece by suggesting that the poet attributed to Zeus' storm-raising aegis a number of commonplace properties, similar to the naturalistic features of Love, Strife, and other personifications.
The precise mechanism that triggered the reinterpretation of cat opog as a single word is difficult to reconstruct, especially as it runs counter to a topos, already found in the Homeric similes, representing the hardships of mountain oaks42. It might have been due to the continuous writing in one of the copies the critic was using; it might have been the deliberate wish to correct the slight roughness of syntax in the phrase cat opog Spvaiv ¿/unsroiv43; it might
39 For the testimonies and commentary, see Slater (1986: 132-133); cf. also N auck (1848: 273-275) and Pfeiffer (1968: 181-182).
40 See Callanan (1987: 107-122), with discussion of examples; cf. Pfeiffer 41968: 202-203).
1 The formula aiyiSa dvaaavosaaav is found in II. 5, 738; 15, 229; 17, 593; 18, 204; 21, 400.
43 E.g., Il. 11, 492-495; Il. 12, 132-134; Il. 14, 398-399; Il. 14, 414-417.
43 This syntactic difficulty has been recognized by modern scholars. The editors of the 1937 Budé edition disagreed on the fragment's interpretation to the point of giving it a double translation: « Éros a secoué mon âme comme le vent, qui vient de la montagne, tombé dans les chênes » (Reinach) and « Le vent qui, sur la montagne, s'abat sur les chênes » (Puech). Tzamali (1996: 255) explains the syntax thus: "Kara + Akkusativ könnte hier die abwärtsgehende Richtung bezeichnen, wie es oft bei Verben der Bewegung der Fall ist. Mit dieser Bedeutung verträgt sich aber nicht gut der folgende Zieldativ Spvaiv, der gerade das ,hinein' ausdrückt. Daher liegt es
have been a detail in the context which is not preserved, or the simply the aesthetic preference for the epithet as more poetic. Whichever was the cause in this case, such editorial and interpretative practices have good parallels. The best known example is the suggestion of an unknown scholar that the expression Staarrfrnv épiaavrs (Il. 1, 6) in the proem of the Iliad should in fact have been written día ar^r^v épiaavrs, where arrfrn would be a dialectal gloss for "woman"44. More importantly, several corrections, based on the abolition or reinsertion of a word boundary, have been preserved specifically under Aristophanes' name45, so that it is not unlikely that in the case of kox opoq he was responsible for the gloss.
There is, however, a specific context that would have attracted the ancient scholar's attention to the resemblance between Sappho's and Homer's denomination of the winds - Homer's description of Zeus' intervention in the battle:
Kai tót' apa KpovíSr¡g sXst' aíyíSa Ovaaavósaaav
pappapér¡v, 'ISr\v Ss Kara vstpésaai KáXvpsv,
áaTpáiyaq Ss páXa psyáX' sktvks, Tf¡v Ss TÍva^s,
vÍK^v Ss Tpmsaai SíSov, é^ópr¡as S' A/aiovg (II. 17, 5 93-5 96)46.
This passage not only describes Zeus' wielding of the aiyig, but also uses the relatively rare verb xrvaaaa)47, the same that will later be used by Sappho in her comparison "Epog d' ériva^é poí / tppévaq. It is possible that the author of nspi aiyidog had recognized a Homeric
näher, anzunehmen, daß Kara hier im Sinne von ev zur Angabe des Ortes der Handlung verwendet wird; [...] KaT opog heißt also 'an vielen Punkten des Berges'".
44 f mg Ev t& Bopw AmciaSov (v. 1) ^ yvvtj sipr\Tai ct^t^, EnsiSij Tivsg tö nap' 'Opiipm "SiacT^T^v EpicavTs" ommg EtnyicavTo, Sia Tiva yvKaiKa (schol. D. Thr. [Zd] 11, 24; text cited from Erbse 1969: I, 4). Cf. Dyck 41987: 126).
5 Thus, schol. A in Il. 20, 30b preserves the correction of vnsp pöpov to vnspßiov, and the scholiast stresses that Aristophanes joined word group into one adjective: vnsp popov EtaXXanatni: 'ApicTopävqg mg "vnspßiov", sv pspog Xöyov noiwv (Eustathius 1193, 57 ad Il. 20, 30 attributes the reading to Aristarchus); similarly, schol. HMaO in Od. 3, 82b (Pontani) records Aristophanes' replacement of ov Sijpiog by EKSipiog; cf. also schol. in Od. 2, 338a (Pontani).
46 "And then the son of Kronos lifted his tasselled flashing aegis, and hid mount Ida under the clouds, and hurling his lightning thundered very loudly, and shook it (i.e. the aegis), and gave the victory to the Trojans, but on the Achaeans he sent terror".
47 See H. V. Nordheider, s.v. Tivaccm in LfgrE (2006: IV, 529-530)
allusion in Sappho of which we cannot be sure in the absence of broader context48.
Despite the elegant reasoning, later editions of Sappho49 seem to have reverted to the much more natural reading mg avsp.og cat opog Spvaiv ¿/unsrojv (Sappho, fr. 47 Voigt), preserved by Maximus of Tyre; but by then the varia lectio had already begun its independent existence as a gloss cited by lexicographers and Homeric commentators. As regards Alcaeus' appearance together with Sappho in Porphyry and Eustathius, this reference is doubtful (although we cannot completely exclude the possibility that the group cat opog appeared in his poetry as well): the amplified reference may be due to the fact that the scholars were no longer able to trace the quotation, as Porphyry's 'Alcaiog Ss nou cai haiupw does seem to indicate.
The results of this study add little to our knowledge of Sappho's and Alcaeus' poetry. We have tried to show that the adjective Kaxmpnq (Kazap^g) probably stems from a scholarly interpretation of the text of Sappho fr. 47 Voigt. This does not mean that it should be excluded from editions of the Lesbian poets50: however, in Sappho's case its closeness to fr. 47 Voigt should be emphatically stressed; as for Alcaeus fr. 412, it seems advisable to include it in the category of dubious fragments. On the other hand, while tracing of the gloss' transformations, we catch a glimpse of the sophisticated work of Alexandrian scholars and later grammarians where considerations of etymology, analogy, and dialectal traits were combined to attempt to solve the question of the Homeric aiyig. Moreover, the recognition of the four-term proportion aiyig : caraiyig :: ovpog : KATOPOE lends additional weight to Eustathius' attribution of the nucleus of the interpretation of the aiyig as the instrument that Zeus uses to raise storms to Aristophanes of Byzantium. As is often the case, the study of caTrnpqg (cazap^g) yields more to our knowledge of the ancient scholarship than to the object of their inquiry, and its results are in keeping with the principle formulated by Fr. Montanari (2011: 24): "everything that is of no aid in specifically interpreting Homer or
48 On Homeric allusions in Sappho, see in particular P. Rosenmeyer (1997).
49 On the question of reconstruction of Sappho's Alexandrian edition(s), see Yatromanolakis (1999).
50 Just as tpaivszai foi k^voç (Sappho fr. 165 Voigt) exists as a separate fragment, although it is surely a variant of the opening words of Sappho fr. 31 with the pronoun modified.
Pindar or Aristophanes from our point of view, is of the greatest aid in interpreting Zenodotus, Aristophanes of Byzantium, Aristarchus, and in understanding their cultural context and their intellectual milieu".
Bibliography
Bergk 1882 - Poetae lyrici graeci. Vol. III: Poetae melici / ed. Th. Bergk.
Lipsiae: B. G. Teubner, 1882. Bowie 1984 - Bowie, A. M. The Poetic Dialect of Sappho and Alcaeus.
Salem: Ayer Company, 1984. Calannan 1987 - Calannan C. K. Die Sprachbeschreibung bei Aristophanes
von Byzanz. Göttingen: Vandenhoeck & Ruprecht, 1987. Campbell 1982 - Greek Lyric. Vol. I: Sappho and Alcaeus / ed., transl. D. A. Campbell. Cambridge, Ma., London: Harvard University Press, 1982.
Chantraine 1999 - Chantraine P. Dictionnaire étymologique de la langue
grecque. 2e éd. Paris : Klinksieck, 1999. Cohn 1881 - Cohn L. De Aristophane Byzantio et Suetonio Tranquillo Eustathi Auctoribus // Jahrbücher der classischen Philologie, Supplementum 12. Lipsiae: B. G. Teubner, 1881. P. 285-374. Deacy, Villing 2009 - Deacy S., Villing A. What was the Coulour of Athena's Aegis? // Journal of Hellenic Studies. 2009. Vol. 129. P. 111129.
Dyck 1987 - Dyck A. R. The Glossographers // Harvard Studies in
Classical Philology. 1987. Vol. 91. P. 119-160. Erbse 1969-1977 - Scholia Graeca in Homeri Iliadem / ed. H. Erbse. Vol.
1-7. Berlin: De Gruyter, 1969-1977. Fowler 1988 - Fowler R. L. Air- in Early Greek Language and Myth //
Phoenix. 1988. Vol. 42. P. 95-113. Funaioli 1973/1974 - Funaioli M. P. Sapph. fr. 183 V. // Museum criticum.
1973/1974. Vol. 8/9. P. 129-130. Hamm 1957 - Hamm E. M. Grammatik zu Sappho und Alkaios. Berlin: Akademie-Verlag, 1957 (Abhandlungen der Deutschen Akademie der Wissenschaften zu Berlin, Klasse für Sprachen, Literatur und Kunst 1951 nr. 2).
Hobein 1910 - Maximus Tyrius. Philosophumena / ed. H. Hobein. Lipsiae:
B. G. Teubner, 1910.
Hooker 1977 - Hooker J. T. The Language and Text of the Lesbian Poets. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 1977.
Janda 2014 - Janda M. Purpurnes Meer: Sprache und Kultur der homerischen Welt. Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck, 2014. Kamerbeek 1970 - The Plays of Sophocles: Commentaries / ed., comm. J.
C. Kamerbeek. Leiden: J. C. Brill, 1970.
Lehrs 1882 - Lehrs K. De Aristarchi Studiis Homericis. 3. Aufl. Leipzig: S. Hirzelius, 1882.
Lenz 1979 - Herodiani Technici Reliquiae / ed., expl. A. Lenz. T. 1 : Praefationem et Herodiani Prosodiam Catholicam continens. Hildesheim; New York: Georg Olms Verlag, 1979.
LfgrE - Lexikon des frühgriechischen Epos. Bd. 1-4. / begründet von B. Snell, hrsg. von Thesaurus linguae graecae. Göttingen: Vandenhoeck & Ruprecht, 1955-2006.
Libermann 1999 - Alcée: Fragments / éd. trad. comm. G. Libermann. Vol. 1-2. Paris: Les Belles Lettres, 1999.
Lobel 1925 - ZAn^OYZ MEAH: The Fragments of the Lyrical Poems of Sappho / ed. E. Lobel. Oxford: Clarendon Press, 1925.
Lobel 1927 - AAKAIOY MEAH: The Fragments of the Lyrical Poems of Alcaeus / ed. E. Lobel. Oxford: Clarendon Press, 1927.
Lobel, Page 1968 - Poetarum Lesbiorum Fragmenta / ed. E. Lobel et D. Page. Oxford: Clarendon Press, 1968.
MacPhail 2010 - Porphyry's Homeric Questions on the Iliad: Text, Translation, Commentary / ed. J. A. MacPhail Jr. Berlin; New York: De Gruyter, 2010.
Montanari 2011 - Montanari F. Ancient Scholarship and Classical Studies // Ancient Scholarship and Grammar: Archetypes, Concepts and Contexts / ed. by S. Matthaios, F. Montanari, A. Rengakos. Berlin; New York: De Gruyter, 2011. P. 11-24.
Nauck 1848 - Aristophanis Byzantii Grammatici Alexandrini Fragmenta / ed. Aug. Nauck, accedit R. Schmidtii comm. de Callistrato Aristophaneo. Halis, 1848.
O'Neil 1969 - O'Neil J. L. The Treatment of Vocalic R and L in Greek // Glotta. 1969. Vol. 47. P. 8-46.
Pfeiffer 1968 - Pfeiffer R. History of Classical Scholarship from the Beginnings to the End of the Hellenistic Age. Oxford: Clarendon Press, 1968.
Pontani 2007-2010 - Scholia Graeca in Odysseam. Vol. 1-2 / ed. F. Pontani. Roma: Edizioni di storia e letteratura, 2007-2010.
Reinach, Puech 1937 - Alcée, Sappho / éd., trad. Th. Reinach, A. Puech. Paris: Les Belles Lettres, 1937.
Rengakos 1992 - Rengakos A. Homerische Wörter bei Kallimachos // Zeitschrift für Papyrologie und Epigraphik. 1992. Bd. 94. P. 21-47.
Risch 1974 - Risch E. Wortbildung der homerischen Sprache. 2. Aufl. Berlin; New York: De Gruyter, 1974.
Rodriguez Somolinos 1994 - Rodriguez Somolinos, H. Notas Léxicas a Safo y Alceo // Emerita. 1994. Vol. 62. P. 109-123.
Rose 1971 - Rose V. Aristoteles Pseudepigraphus. Hildesheim, New York: Georg Olms Verlag, 1971 [repr. 1863].
Rosenmeyer 1997 - Rosenmeyer P. A. Her Master's Voice: Sappho's Dialogue with Homer // Materiali e discussioni per l'analisi dei testi clas-sici. 1997. Vol. 39: Memoria, arte allusiva, intertestualità. P. 123-149.
Schrader 1880 - Porphyrii Quaestionum Homericarum ad Iliadem Pertinentium Reliquiae. Bd. 1. / ed. H. Schrader. Lipsiae: B. G. Teubner, 1880.
Schwartz 1881 - Schwartz E. S. De scholiis Homericis ad Historiam fabularum pertinentibus // Jahrbücher der classischen Philologie, Supplementum 12. Lipsiae: B. G. Teubner, 1881. P. 405-463.
Schwyzer 1990 - Schwyzer E. Griechische Grammatik auf der Grundlage von Karl Brugmanns griechischer Grammatik. Bd. 1: Allgemeiner Teil, Lautlehre, Wortbildung, Flexion. 6. Aufl. München: C. H. Beck, 1990. Slater 1986 - Aristophanis Byzantii Fragmenta / ed., comm. W. J. Slater.
Berlin; New York: De Gruyter, 1986. Stephens 2015 - Callimachus: The Hymns / ed., comm. S. A. Stephens.
Oxford: Oxford University Press, 2015. Tzamali 1996 - Tzamali E. Syntax und Stil bei Sappho. Dettelbach: J. H. Röll, 1996.
van der Valk 1971-1987 - Eustathii Archiepiscopi Thessalonicensis Commentarii ad Homeri Iliadem Pertinentes. Vol. 1-4 / ed. M. van der Valk. Lugduni Batavorum: E. J. Brill, 1971-1987. Voigt 1971 - Sappho et Alcaeus: Fragmenta / ed. E.-M. Voigt. Amsterdam:
Athenaeum - Polak & van Gennep, 1971. Wackernagel 1889 - Wackernagel J. Das Dehnungsgesetz der griechischen Composita (dem Basler Gymnasium Feier seines dreihundertjährigen Bestehens gewidmet). Basel: L. Reinhardt, 1889. Yatromanolakis 1999 - Yatromanolakis D. Alexandrian Sappho Revisited // Harvard Studies in Classical Philology. 1999. Vol. 99. P. 179-195.
M. N. Kazanskaya. Sappho fr. 183 Voigt (= Alcaeus fr. 412 Voigt) and Ancient Scholarship on the aíyí<;
The article proposes an overview of the ancient testimonies on the word каттрцд (катарцд) said to have been used by Sappho and Alcaeus of a specific kind of strong wind, descending from above. The study of the testimonies shows that the reference to каттрцд (катарцд) stems from the Alexandrian exegesis on the Homeric aegis, where it appeared as one of the terms of the analogical proportion, while the gloss itself originated in the reinterpretation of the expression avspog кат ород in Sappho fr. 47 Voigt. Finally, arguments in favour of Eustathius' much contested attribution of the essay on the aegis to Aristophanes of Byzantium are cited.
Keywords: каттрцд (катарцд), Sappho, Alcaeus, aegis, Homeric exegesis, Porphyry, Eustathius, Aristophanes of Byzantium.
