¹ Ebeling, H., Lexicon Homericum (Leipzig, 1885)Google Scholar, 2.330: ‘dubiae originis et significationis vocabulum’; Debrunner, A., ‘Zu den konsonantischen i̯o-Präsentien im Griechischen’, IF 21 (1907), 201–76Google Scholar, at 252: ‘ganz unerklärt’; H. Frisk, Griechisches etymologisches Wörterbuch (Heidelberg, 1960–72), 896: ‘schon wegen der schwer bestimmbaren Bed. etymologisch dunkel’; Skoda, F., Le redoublement expressif: un universal linguistique (Paris, 1982)Google Scholar, 214: ‘obscur’; Hoekstra, A., A Commentary on Homer's Odyssey (Oxford, 1989)Google Scholar, 2.171: ‘exact sense and etym. unknown’; Rengakos, A., ‘Lykophron als Homererklärer’, ZPE 102 (1994), 111–30Google Scholar, at 120: ‘das immer noch unerklärte Hapax’; P. Chantraine, Dictionnaire étymologique de la langue grecque (Paris, 2009²), 1077: ‘expressif et obscur’; Beekes, R.S.P., Etymological Dictionary of Greek (Leiden, 2010), 1482Google Scholar: ‘origin ?’; Bowie, A.M., Homer Odyssey Books XIII and XIV (Cambridge, 2013), 116Google Scholar: ‘a very rare word, of unknown meaning and etymology’.

² The reference is to Leumann, M., Homerische Wörter (Basel, 1950)Google Scholar, whose approach is discussed below.

³ Similar translations have been preserved in other grammatical literature, e.g. the Homeric lexicon by Apollonius Sophista (152.33 Bekker: τὴν τροφὴν ἀποθησαυρίζουσι ‘they lay food aside’) or the Etymologicum Magnum (758.16 Gaisford).

⁴ Hsch. τ 862 Hansen–Cunningham: τιθαιβώσσουσιν· ἐν ἀποτίθενται, ἀποθησαυρίζουσι τὴν τροφὴν αἱ μέλιτται, τὸν λεγόμενον μελίκηρον (‘the bees put away inside, preserve the provisions—namely, the honeycomb’).

⁵ For the sake of space, I am not citing all modern works of reference in which τιθαιβώσσω has been thus translated. One voice of dissent is by T.V. Gamkrelidze and V.V. Ivanov, Indo-European and the Indo-Europeans (Berlin and New York, 1995), 519, who render the verb as ‘they nest’, arguing that the stone vessels metaphorically represent beehives; in fact, clefts in rock form the natural abode of bees.

⁶ For the Koine the beginning of this sound change can be dated to the third century b.c.e.

⁷ This poem (Πινδάρου γένος δι’ ἐπῶν) is transmitted in several Pindaric manuscripts such as Laurentianus 32.37, Laurentianus 32.35 and Parisinus 2403.

⁸ For further discussion of the metaphor of a honeycomb in the mouth, see M. Lefkowitz, The Lives of the Greek Poets (Baltimore, 2012²), 62, 176 n. 5; and for general association between bees and poets, see M. Davies and J. Kathirithamby, Greek Insects (Oxford, 1986), 70–2.

⁹ The same meaning must have been the one known to Porphyry, who explains τιθαιβώσσειν as τὸ τιθέναι τὴν βόσιν ‘putting away food’ (De antr. nymph. 18).

¹⁰ It is unclear what inferences can be drawn from Antimachus’ choice of an Aeolic form of the participle.

¹¹ V.J. Matthews, Antimachus of Colophon (Leiden, 1996), 442.

¹² Cazzaniga, I., ‘Osservazioni critiche intorno allo hypomnema antimacheo di Pap. Mil. Vogl. I 17, 33–6 (= fr. 182 Antimachi W.)’, PP 22 (1967), 63–74Google Scholar, at 72 n. 15.

¹³ Matthews (n. 11), 283.

¹⁴ Elderkin, G.W., ‘The bees of Artemis’, AJPh 60 (1939), 203–13Google Scholar remains magisterial; see also R.D. Carlson, ‘The honey bee and apian imagery in classical literature’ (Diss., University of Washington, 2015).

¹⁵ For the sake of simplicity, Pseudo-Lycophron is referred to as Lycophron below.

¹⁶ See Hornblower, S., Lykophron: Alexandra (Oxford, 2015), 266Google Scholar.

¹⁷ See Richardson, N.J., The Homeric Hymn to Demeter (Oxford, 1974), 167Google Scholar.

¹⁸ Cf. Pind. fr. 158 S.–M. ταῖς ἱεραῖσ<ι> μελίσσαις τέρπεται ‘(Demeter) delights in her priestesses’ (this fragment, cited in schol. Pind. Pyth. 4.106a, is followed by a clarification that μελίσσαις δὲ τὰς ἱερείας, κυρίως μὲν τὰς τῆς Δήμητρος). The Hellenistic poet of a hymn to Demeter addresses her priestesses (or initiates?) as μέλισσαι (SH 990.2). Apollodorus of Athens reports that the women participating in the Thesmophoria were called μέλισσαι (FGrHist 244 F 89). The scholiast on Theoc. Id. 15.94 explains Persephone's epithet Mελιτώδης by saying τὰς ἱερείας αὐτῆς καὶ τῆς Δήμητρος μελίσσας λέγεσθαι; similarly, Porphyry states τὰς Δήμητρος ἱερείας ὡς τῆς χθονίας θεᾶς μύστιδας Μελίσσας οἱ παλαιοὶ ἐκάλουν (De antr. nymph. 18). Finally, note Hsch. μ 719 (Latte–Cunningham) μέλισσαι· αἱ τῆς Δήμητρος μύστιδες (‘bees the initiatresses of Demeter’).

¹⁹ See Crane, G., ‘Bees without honey, and Callimachean taste’, AJPh 108 (1987), 399–403Google Scholar, who plausibly argues that μέλισσαι here should be understood as actual bees, not as priestesses of Demeter.

²⁰ J.-M. Jacques (ed.), Nicandre: Œuvres. Tome 2: Les Thériaques (Paris, 2002); transl. A.S.F. Gow and A.F. Schofield, Nicander (Cambridge, 1953).

²¹ It is possible that νεοττοτροφοῦσιν ‘rear young birds’ listed in Schol. V as a paraphrase of τιθαιβώσσουσι at Od. 13.106 refers to the passage in Nicander.

²² From τιθήνη (‘nurse’), ultimately from the root of θῆσθαι.

²³ The question why there should be bees in the Cave of the Nymphs lies outside the scope of this paper. It is possible that an analogy was sought with the bees who helped feed the infant Zeus in the Dictaean cave (Epimenides, fr. 4.70 EGM = BNJ 457 F 17), but the insects’ general association with purity, chastity and holiness is more likely to have played a role.

²⁴ See G.S. Kirk, The Iliad: A Commentary. Volume 1: Books 1–4 (Cambridge, 1985), 378 and C. Brügger, M. Stoevesandt and E. Visser, Homers Ilias: Gesamtkommentar, vol. 2, fasc. 2 (Berlin, 2010²), 86.

²⁵ Leumann (n. 2), 211.

²⁶ See Nikolaev, A., ‘Homeric ἀάατος: etymology and poetics’, Sprache 50 (2012–13), 182–23Google Scholar9, at 197–8.

²⁷ ‘Leumannian misunderstanding’; so dubbed by Burkert, W., ‘ΘΕΩΝ ΟΠΙΝ ΟΥΚ ΑΛΕΓΟΝΤΕΣ’, MH 38 (1981)Google Scholar, 195–204 = Kleine Schriften (Göttingen, 2001), 1.95–104. While ‘Leumannsches Missverständnis’ is most frequently used to refer to the process that S. Reece, Homer's Winged Words (Leiden, 2009) has termed ‘junctural metanalysis’ (e.g. pre-Homeric [πολέμου] ἐπιδημίοο κρυόεντος resegmented as [πολέμου] ἐπιδημίο’ ὀκρυόεντος, hence Il. 9.64 ἐπιδημίου ὀκρυόεντος), Leumann's magisterial book contains discussions of many other processes that contributed to the creation of ‘Homeric words’.

²⁸ See the references in n. 1 above. Few daring solutions that have been advanced do not stand scrutiny from the position of modern historical linguistics. For instance, C.A. Lobeck, Ῥηματικόν, siue uerborum graecorum et nominum uerbalium technologia (Königsberg, 1846), 248 assumed that ‘nourish’ was the original meaning and analysed τιθαιβώσσω as a ‘uerbum intensiuum’ made from the same root as τιθήνη ‘wet nurse’; this is impossible, since Latin fēmina ‘woman’ and other cognates of the Proto-Indo-European root ‘to suck’ make it clear that the -η- in the root of τιθήνη, θῆσθαι goes back to Proto-Greek *ē (and not *ā) and is therefore incompatible with the -αι- of τιθαιβώσσω. L. von Döderlein, Homerisches glossarium (Erlangen, 1850–8), 3.359 thought that the description of the cave in the Odyssey referred specifically to domesticated honey bees and proposed that τιθαιβώσσω with the alleged meaning ‘settle’ was etymologically related to τιθασός ‘cultivated’, which is improbable semantically, morphologically and phonetically (for Döderlein, the -β- in τιθαιβώσσω was a ‘hardened digamma’ used as a hiatus-filler). As to τιθασός, it is probably a foreign word; see Beekes (n. 1), 1482.

²⁹ G. Klingenschmitt, Das altarmenische Verbum (Wiesbaden, 1982), 70 derives πτώσσω from a primary stem *ptoh₂k-i̯e/o- and compared Armenian t‛ak‛č‛i- ‘to hide’, but the Armenian verb may also go back to the root of Latin tacēre, while πτώσσω can be analysed as a denominative verb derived from πτώξ ‘timid, hare’ (O. Hackstein, ‘Eine weitere griechisch-tocharische Gleichung: Griechisch πτῆξαι und tocharisch B pyāktsi’, Glotta 70 [1992], 136–65, at 137).

³⁰ Homeric κνώσσω ‘to be asleep’ and Hsch. θ 812 Latte–Cunningham θρώσσει⋅ γεννᾷ, φοβεῖται both have uncertain etymologies and are excluded from the present consideration.

³¹ See E. Risch, Wortbildung der homerischen Sprache (Berlin and New York, 1974²), 284. For an exhaustive discussion of secondary verbs in -ώσσω, see Debrunner (n. 1), 248–53.

³² For the agent noun suffix -της, see P. Chantraine, La formation des noms en grec ancien (Paris, 1933), 316 and the monographic treatment by A. Leukart, Die frühgriechischen Nomina auf -tās und -ās (Vienna, 1994).

³³ While -της/-τᾱς had become the most productive of the agent noun forming suffixes very early in Attic-Ionic and by the fifth century in all other dialects, in Homeric Greek we still find what may have been the original distribution of the suffixes: -της is used with compounds, while -τήρ and -τωρ are used with simplex nouns (ἡγήτωρ ‘leader’ vs κυνηγέτης ‘hunter’, βοτήρ ‘herdsman’ vs συβώτης ‘swineherd’; Risch [n. 31], 28–9). However, the -της suffix does not necessarily indicate that *τιθαιβώτης should be analysed as a compound: the reason such a form would be acceptable even in the oldest Greek is simply that it looked like a compound because of its length; it is also possible for an original *τιθαιβώτηρ to have been remade as *τιθαιβώτης on the analogy to compounds with -βώτης. In sum, the agent noun *τιθαιβώτης does not violate morphological rules of Homeric Greek.

³⁴ See E.F. Tucker, The Creation of Morphological Regularity: Early Greek Verbs in -éō, -áō, -óō, -úō, and -íō (Göttingen, 1990), 283–92. The formal difference between the two types would not have been perspicuous to a Greek, since factitives in general may have exactly the same meaning as instrumentals, e.g. αἰσχύνω ‘to furnish with αἶσχος’ (instr.) is equivalent to ‘to make αἰσχρόν’ (fact.), as noted by W.S. Barrett, Greek Lyric, Tragedy, and Textual Criticism: Collected Papers (Oxford, 2007), 344 n. 67.

³⁵ The position of the accent in this hypothetical form cannot be ascertained. Another formation from which the verb *τιθαιβόω ‘to deposit’ is just as likely to have been derived would be ā-stem *τιθαίβη, cf. κορυφόομαι ‘rise up’, ‘be provided with a crest’ from κορυφή ‘top, peak, crest’. No decision can be made between *τίθαιβος and *τιθαίβη as putative derivational bases of *τιθαιβόω, and the choice is ultimately immaterial for the solution pursued in this paper. The presentation in the main text implicitly assumes *τιθαίβη as a viable alternative to *τίθαιβος.

³⁶ In many languages verbal abstract nouns (nomina actionis) in addition to denoting an actual action may also be used to denote either concrete objects or results of said action (nomina rei actae): cf. English construction ‘the process of building something’ but also ‘a building’, forgery ‘the process of faking something’ but also ‘a fake’, Italian discendenza ‘the process of descending’ but also ‘offspring’, Greek ἀοιδή ‘the act of singing’ but also ‘a song’.

³⁷ Across languages, chains of morphological derivation often become opaque to speakers and are abbreviated thanks to the workings of analogy. The verb ὑπνώσσω ‘be sleepy’ mentioned above may in fact have been derived directly from ὕπνος ‘sleep’ on the model of ἀγρώσσω ‘catch by hunting’ next to ἄγρᾱ ‘hunt’, ‘skipping’ the putative intermediate stage *ὑπνώτης (the absence of which therefore does not have to be viewed as an attestation gap after all); in other words, the speakers may have reanalysed the synchronic morphological relationship between ἀγρώσσω and ἄγρᾱ as direct derivation whereby the nominative singular ending was replaced by -ώσσω and left ἀγρώτης out of the derivation, even though historically the double -σσ- of ἀγρώσσω certainly goes back to the τ of ἀγρώτης followed by the suffix *-i̯e/o-. Similarly, τιθαιβώσσω may in theory have been formed directly from *τίθαιβος on the model of ὑπνώσσω : ὕπνος, etc.

³⁸ See Risch (n. 31), 8–14 and for details A.J. Nussbaum, ‘Agentive and other derivatives of “τόμος-type” nouns’, in C. Le Feuvre et al. (edd.), Verbal Adjectives and Participles in Indo-European Languages (Bremen, 2017), 233–66.

³⁹ See Chantraine (n. 32), 13.

⁴⁰ See for the last example F. Solmsen, Untersuchungen zur griechischen Laut- und Verslehre (Strassburg, 1901), 256 and for additional examples E. Schwyzer, Griechische Grammatik (Munich, 1939), 1.423.

⁴¹ Grassmann's Law: A. Sihler, New Comparative Grammar of Greek and Latin (Oxford, 1995), 142–4.

⁴² Cf. βέλτερος ‘better, stronger’ < *bel-, Vedic balín- ‘strong’, Latin dē-bilis ‘weak’, Russian bol’šoi ‘big’.

⁴³ See J. Clackson, Indo-European Linguistics (Cambridge, 2007), 46; M. Weiss, Outline of the Historical and Comparative Grammar of Latin (Ann Arbor, 2020²), 37.

⁴⁴ Cf. βαίνω ‘come’ < *g^wm̥-i̯e/o-, Latin ueniō or βοῦς ‘cow’ < *g^wóu̯s, Vedic gáu-, Latin bōs, Old English cū. See Sihler (n. 41), 161–2.

⁴⁵ See Sihler (n. 41), 45; Clackson (n. 43), 36; Weiss (n. 43), 45.

⁴⁶ For laryngeal metathesis, see M. Mayrhofer, Indogermanische Grammatik, vol. 1/2: Lautlehre (Heidelberg, 1986), 175. For instance, Greek πρᾱΰς ‘mild, gentle’ < *prāi̯u- goes back to the root *preh₂i- the zero-grade of which appears as *prī- < *prih₂- (instead of *prh₂i-) in preconsonantal position (cf. Vedic prīṇā́ti ‘gratifies’).

⁴⁷ Laryngeal consonants disappeared from most branches of Indo-European, but left important traces in the vowel system; in particular, any laryngeal lengthened the immediately preceding vowel before it was lost: cf. Greek δαίομαι ‘distribute’, Cretan δαῖσις ‘apportioning’ < *deh₂i- vs Vedic dī- ‘divide, share’ < *dih₂- (< *dh₂i- with metathesis). Another example of alternation between -ai- (< *-eh₂i-) and -ī- (< *-ih₂- < *-h₂i-) may be provided by Greek λιλαίομαι ‘desire’ compared by Solmsen, F., ‘Zur Geschichte des Dativs in den indogermanischen Sprachen’, ZVS 44 (1911), 161–223Google Scholar, at 171 to λαιδρός ‘bold, impudent’, to which we may add adj. λαιμός ‘wanton’ (Men. fr. 102 K.–A.) and reconstruct the root as *leh₂i-, the zero grade of which would be found in λῑρός ‘shameless’ < *lih₂-ro- (one wonders if Hittite laḫlaḫḫiya- ‘to be in [emotional] turmoil’ may belong to this PIE root). A similar alternation may be found in Greek αἱμύλος ‘seductive, binding’ (of words) vs ἱμάς ‘leather strap’, ῑ̔μονιά ‘rope’ going back to an n-stem *sīmon- < *sih₂mon- (Vedic sīmán- ‘boundary’, Old English sīma ‘rope’) and further to PIE root *seh₂i- ‘to bind’ (see M. Weiss, ‘On the prehistory of Greek desire’, HSPh 98 [1998], 51–6; M. Janda, Elysion: Entstehung und Entwicklung der griechischen Religion [Innsbruck, 2005], 46–7).

⁴⁸ Contrast EXDEICATIS in the same inscription, line 22, from dīcere < PIE *dei̯k̑-.

⁴⁹ Note also the gloss offīuēbant: claudebant seris ‘shut with bars’ (Abolita Glossary 132.1).

⁵⁰ Word-medial -g- in Classical Latin present fīgere was introduced by analogy to the perfect. The most recent discussions of the Latin verb are B. Bock, Die einfach thematischen Präsentien in der dritten Konjugation des Lateinischen (Graz, 2008), 239–40; R. Garnier, Sur le vocalisme du verbe latin: étude synchronique et diachronique (Innsbruck, 2010), 398.

⁵¹ In theory, *dʰīgʷ- attested in Latin, Germanic and Baltic may go back to *dʰih₁gʷ-, *dʰih₂gʷ- or *dʰih₃gʷ-, but Tocharian B tsāk^ā- to be mentioned momentarily rules out *h₁.

⁵² The Academic Dictionary of Lithuanian (Lietuvių kalbos žodynas) illustrates this meaning with a proverb that amusingly leads us back to Greek: bėk nuo grieko kaip nuo žalčio, nes, jei tu prisiartinsi, įdiegs tave ‘run away from a Greek as if from a snake: if they get closer, they will sting you’.

⁵³ See W. Smoczyński, Słownik etymologiczny języka litewskiego (Vilnius, 2007), 109.

⁵⁴ See A.L. Lloyd, O. Springer and R. Lühr, Etymologisches Wörterbuch des Althochdeutschen. Vol. 2: bî – ezzo (Göttingen, 1998), 630–4.

⁵⁵ See D. Ringe, ‘Evidence for the position of Tocharian in the Indo-European family?’, Die Sprache 34 (1988–90), 59–123, at 71; D.Q. Adams, A Dictionary of Tocharian B: Revised and Greatly Enlarged (Amsterdam, 2013), 800.

⁵⁶ See J.H. Jasanoff, ‘The impact of Hittite and Tocharian: rethinking Indo-European in the 20th century and beyond’, in J. Klein et al. (edd.), Handbook of Comparative and Historical Indo-European Linguistics (Berlin, 2017), 1.220–38.

⁵⁷ It is possible that the meaning ‘to sting’ attested in Lithuanian is an archaism. Note also Latvian daîga ₂ ‘kind of fly’ (K. Mühlenbach and J. Endzelin, Lettisch-deutsches Wörterbuch [Riga, 1923–36], 2.430).

⁵⁸ Or ‘to bite’: even though Aristotle knew that female bees have a sting (e.g. Hist. an. 626a18), it is entirely possible that at the much earlier time when the Odyssey was composed Greeks still thought that bees and wasps bite and not sting. This hypothesis is borne out by the use of δάκνω of insects at Il. 17.572.

⁵⁹ The disyllabic structure of *d^heh₂ig^w- is awkward, but such an extended root would not be unparalleled: beside *g^weh₂id- mentioned in the main text above cf. *seh₁id^h- (> Greek εἶθαρ ‘immediately, right on’) next to *sh₁id^h- > *sih₁d^h- (> Greek ῑ̓θύς ‘straight’: M. Peters, Untersuchungen zur Vertretung der indogermanischen Laryngale im Griechischen [Vienna, 1980], 86) or *u̯reh₂ig^w- (> Greek ῥαιβός ‘bent inward’, Gothic wraiqs ‘crooked’: Frisk [n. 1], 639). In theory, *d^heh₂ig^w- may be viewed as a secondary root, viz. *d^heh₂(i)- with a Wurzelerweiterung *g^w, cf. *leh₁id- ‘let’ for which G. Klingenschmitt, Das altarmenische Verbum (Wiesbaden, 1982), 213 n. 69 proposed a connection with *leh₁(i)- (cf. Alb. la ‘s/he let’); another example may be found in Latin saepēs ‘hedge’ and Greek αἰπύς ‘steep’ (< *‘mit einer Befestigung versehen’), on the basis of which M. Janda, ‘Etymologie von altgriechisch αἰπύς’, SPFB(klas) 6–7 (2001–2), 123–34 reconstructed *seh₂ip- ‘bind’, clearly relatable to *seh₂i- ‘bind’ discussed above, n. 47. (It is quite possible that root extensions like *-d^h, *-g^h, or *-g^w represent fossilized second members of compounds: I. Balles, ‘Lang, rund und krumm: zu einigen indogermanischen Zusammenbildungen’, Die Sprache 48 [2009], 20–6.) The putative non-extended root *d^heh₂(i)- cannot be identified with certainty at present.

⁶⁰ In H. Rix and M. Kümmel, Lexikon der indogermanischen Verben (Wiesbaden, 2001²), 142 the root is reconstructed as *d^hei̯Hg^w- ‘to stick in’, but this is done solely on the evidence of Lithuanian díegti, Latvian diêgt and other forms in Baltic languages that point to Proto-Baltic *dei̯g-/*dai̯g- (similar reconstructions have been adopted in all standard etymological dictionaries of Baltic). If the reconstruction *d^hei̯Hg^w- is correct, Proto-Greek *t^(h)it^hai̯g^wo- with its -ai̯- cannot be related to this root. But this reconstruction is phonotactically suspect. More importantly, as is acknowledged in the first footnote to the LIV lemma, the Proto-Baltic full grade *dei̯g- may be secondary; there are many examples of such secondary full grades in Baltic languages. For instance, Lithuanian seĩlas ‘noose, rope’ (< *sei̯-) can only be explained as a derivative of the root *seh₂i- ‘bind’ discussed above in n. 47 under the assumption that a new full grade *sei̯- was created in Proto-Baltic to match the zero grade *sī- < *sih₂- < *sh₂i- ~ *seh₂i-. Similarly, the only way to align Lithuanian ríeti ‘to scold’, Latvian riêt ‘to bark’ (< *rei̯-) with Russian rajat’ ‘to make a noise’ and Latvian rãt ‘to rebuke, scold’ (< PIE *reh₂i-) is to posit a secondary full grade *rei̯-. Klingenschmitt (n. 59), 213 n. 69 plausibly analysed Lithuanian síekti ‘reach out’ (quasi *sei̯Hk̑-) and léisti ‘let down, let go’ (quasi *lei̯Hd-) as new full grades back-formed to zero-grade allomorphs *sih₁k̑- and *lih₁d- made from *seh₁i-k̑- and *leh₁i-d-, extended versions of *seh₁i- and *leh₁i-. Examples could be multiplied; the point is that Baltic evidence for *dei̯Hg- (to which one should add some Slavic forms overlooked in etymological dictionaries—namely, Slovenian and Croatian degáti se ‘to argue’ < Proto-Slavic *děgati, on which see O.N. Trubachev, ‘Ėtimologičeskij slovar’ G.A. Il'inskogo’, Voprosy jazykoznanija 1957/6, 91–6, at 95) is not incompatible with the reconstruction *d^heh₂ig^w- marshalled in the main text above.

⁶¹ In PIE, one would expect a present stem with i-reduplication to show a zero grade root; however, in the absence of direct comparanda for the reduplicated stem, it is entirely possible that *d^hi-d^heh₂ig^w- (> Proto-Greek *t^(h)it^hai̯b-) is a product of remodelling of some sort: either athematic *d^he-d^heh₂ig^w- was remade as an i-reduplicated stem *d^hi-d^heh₂ig^w- (compare *d^he-d^heh₁- >> *d^hi-d^heh₁- > τίθημι) or thematic i-reduplicated *d^hi-d^hih₂g^w-e/o- was remade as an athematic stem *d^hi-d^heh₂ig^w- (compare *s(t)i-sth₂-e/o- >> *s(t)i-steh₂- > ἵστημι). I thank M. Weiss for pointing this out to me.

⁶² As A. Nussbaum reminds me, there is more than one way of arriving at *τιθαιβόω, and while it is possible that the hypothetical nominal stem *τίθαιβος/*τιθαίβη ‘biting’ was concretized to ‘a bite/sting’ and served as a derivational basis for a factitive verb *τιθαιβόω ‘to make a bite/sting’, a different approach is just as possible: one could posit a nominal stem *τιθαιβός with a passive meaning ‘bitten, stung’ that would make an -όω present stem meaning ‘to render bitten/stung’, hence ‘to bite, to sting’. Compare e.g. θοός ‘sharp, sharpened, whetted’ vs θοόω ‘to make sharp’, on which see A. Nikolaev, ‘Greek θοός “sharp”, Hittite tuḫš- “to cut”’, in D. Gunkel et al. (edd.), Vina diem celebrent: Studies in Linguistics and Philology in Honor of Brent Vine (Ann Arbor, 2018), 267–75.

⁶³ The natural quantity of the vowel -ι- in θιβ'ρός cannot be determined.

⁶⁴ On Θίβρος, Θίβρων, Θίβραχος, see F. Bechtel, Die historischen Personennamen des Griechischen bis zur Kaiserzeit (Halle, 1917), 508 and Arena, R., ‘Considerationi intorno agli aggettivi θιβρός e θεμερός’, RIL 104 (1970), 307–14Google Scholar, at 309 n. 10. The earliest onomastic attestation is Σέβρον, viz. Θέβρον, in Alcm. fr. 1.3 PMGF.

⁶⁵ Frisk (n. 1), 674: ‘wegen der unsicheren Bedeutung etymologisch mehrdeutig’; Chantraine (n. 1), 420 is similarly agnostic. Arena (n. 64), 314 derived θιβρός from PIE *d^heg^wh- ‘to be hot’, but this is linguistically impossible since Grassmann's Law (n. 41 above) affects the first of two aspirated stops: compare the regular development in *d^heg^wh-reh₂ > Proto-Greek *t^hek^whrā > *tek^whrā > τέφρᾱ ‘ashes’. An etymological connection with Hsch. θ 233 Latte–Cunningham θεμερόν· σεμνόν and θεμερῶπις proposed by K. Tsantsanoglou, Of Golden Manes and Silvery Faces (Berlin, 2012), 11 likewise lacks conviction: there is no ‘normal syncope + assimilation process’ in Greek that would convert θεμερός to θιμβρός and then to θιβρός.

⁶⁶ Transl. Gow and Schofield (n. 20), except that I hold back their translation of the word under discussion.

⁶⁷ θιβρὴν δὲ τὴν θερμὴν καὶ ὀξεῖαν διὰ τὰς ἐξ αὐτῆς γινομένας φλεγμονάς (‘θιβρήν means hot and sharp, on account of the inflammation that results from it’).

⁶⁸ Followed by J.-M. Jacques, Nicandre: Œuvres. Tome 3: Les Alexipharmaques (Paris, 2007), 51 ‘œufs délicats’.

⁶⁹ A different way of understanding this passage is reflected in the scholium θιβρά⋅ θερμά. Indeed, νέρθε πυρός (questioned by Gow and Schofield [n. 20], 199) may refer to baking under the coals; cf. Schol. Alex. 555 Ábel and V'ari: ἑψηθέντα ἐπ’ ἀνθράκων. But this does not make the translation ‘hot eggs’ particularly plausible for the Alexipharmaka passage in which not just the eggs but all of the ingredients are said to be mixed on a fire.

⁷⁰ An excellent parallel in English, suggested to me by B. Maslov, may be seen in the word tart, whose meanings, at least through its history, have ranged from ‘sharp, severe, painful’ to ‘sharp to the sense of taste, pungent’ to ‘acrimonious’ (of a person).

⁷¹ Or Ἁρμονίης; cf. Plut. Mor. 769A τὴν Ἀφροδίτην Ἄρμα καλοῦσιν.

⁷² Following the scholia, students of these fragments have translated θιβρή with ‘hot’, ‘burning’, or ‘sultry’, e.g. G. D'Alessio, Callimaco (Milan, 1996), 2.761: ‘della bruciante Cipride’; J.L. Lightfoot, Hellenistic Collection (Cambridge, MA, 2010), 357: ‘sultry Semiramis’; B. Acosta-Hughes and C. Cusset, Euphorion. Œuvre poétique et autres fragments (Paris, 2012), 170: ‘l'ardente Sémiramis’.

⁷³ For Semiramis’ excessive lust, see e.g. Diod. Sic. 2.13.4.

⁷⁴ On this epithet, see Horn, F., ‘“Bitter-sweet love”: a cognitive linguistic view of Sappho's Ἔρος γλυκύπικρος (frg. 130 Voigt)’, Poetica 48 (2016), 1–21Google Scholar, who critically reviews C. Calame's suggestion (The Poetics of Eros in Ancient Greece [Princeton, 1992], 16) that γλυκύπικρος should be understood as ‘sweet-stinging’.

⁷⁵ δάκνω ‘bite’ is used both of insects (e.g. Il. 17.572) and of love, as in Aeschylus’ δηξίθυμον or in Soph. fr. 841 Radt τῳ δ᾿ ἔρωτος δῆγμα παιδικὸν προσῇ ‘but for him who has been stung by love for a boy’; I thank both B. Maslov and the anonymous reviewer for drawing my attention to this.

⁷⁶ But the interpretation of Alcm. fr. 58 PMGF (Ἔρως … ἄκρ᾿ ἐπ᾿ ἄνθη καβαίνων ‘Eros coming down over the flower-tips’) is debated; see Calame, C., Alcman (Rome, 1983), 555–6Google Scholar.

⁷⁷ E.g. Mel. Anth. Pal. 5.163 (= 4248–51 Gow–Page, HE) Ἀνθοδίαιτε μέλισσα … | ἦ σύ γε μηνύεις ὅτι καὶ γλυκὺ καὶ δυσύποιστον, | πικρὸν ἀεὶ κραδίᾳ, κέντρον Ἔρωτος ἔχει ‘O flower-nurtured honeybee … Is your message that she has Love's sting, both sweet and hard to bear, ever bitter to the heart?’; Strato, Anth. Pal. 12.249.1, 6 Βουποίητε μέλισσα … | κἠγὼ κέντρον ἔρωτος ἔχω ‘Ox-born bee … I, too, have a sting, even love's’; Anacreontea 35.15–16 πόσον δοκεῖς πονοῦσιν, | Ἔρως, ὅσους σὺ βάλλεις; ‘If the bee-sting is painful, what pain, Love, do you suppose all your victims suffer?’ Euripides’ comparison of Aphrodite/love to a bee flitting around (Hipp. 563–4) probably does not allude to stinging: Barrett, W.S., Euripides Hippolytos (Oxford, 1964), 266Google Scholar.