ῥοφέω, ῥυφέω
(Antiatt. ρ 5, Phot. ο 364, Phot. ρ 200)

D. General commentary

Forms of the verb ῥοφέω ‘to slurp, gulp down, swallow’ are discussed by the Antiatticist (A.1) as well as in a series of entries in Photius (A.2, A.3, B.3), which Erbse traced back to Aelius Dionysius’ lexicon thanks to a passage in Eustathius (B.5). The entries discuss various phonological and morphological problems. One is the root vowel, which oscillates between ο and υ: while the grammatical tradition reflected in Photius assigned ῥοφ- to Attic and ῥυφ- to Ionic (to which later sources such as B.4 and B.5 add also Doric and Aeolic), the Antiatticist apparently quotes Aristophanes (C.4) in defence of the form with υ (but see F.1). Another point of disagreement is the verb’s contractContract verbs inflection, embraced by the Antiatticist (who quotes the aorist as ῥυφῆσαι), but proscribed by Aelius Dionysius, who, according to Photius and Eustathius prescribed a simple thematic present ῥόφωῥόφω. The same doctrine (B.2) is found in the anonymous prosodic lexicon transmitted in cod. Par. gr. 2646 and recently edited by Sandri (2023), where it may be traced back to Aelius Dionysius himself, but also to an older source such as Tryphon’s Περὶ Ἀττικῆς προσῳδίας, which is one of the main sources of the anonymous lexicon and also had an influence on Aelius (see Sandri 2023, 82–9). Cod. A of the Etymologicum Genuinum (B.4) attributes this teaching to Herodian, who – in keeping with his analogistic perspective – considered the Attic ῥόφω less correct than ῥοφέω because of the future ῥοφήσω, which implies an -έω-present.

The inherited form of the present is undoubtedly ῥοφέω, the reflex of an iterative-intensive formation *srobʰ-éi̯e/o- (cf. Lat. sorbeō) from the Indo-European root *srebʰ- (LIV 587). In Attic literature, ῥοφέω is relatively common in comedy (Ar. 10x, e.g. C.3; Antiph. 2x), while it is rare in tragedy (only Aesch. Eu. 264, Soph. Tr. 1055) and classical prose (only X. An. 4.5.32). The variant ῥυφ- is not the (irregular) outcome of a zero grade *sr̥bʰ-, nor of a dialectal sound change /o/ > /u/, but in all likelihood has an onomatopoeic motivation: the more pronounced lip rounding required to articulate the vowel [u] (or [y], if this form postdates the Attic-Ionic development /u/ > /y/) iconically reproduces the action of slurping (see Tichy 1983, 82–5; Hawkins 2013, 31–4). The exact dialectal distribution of ῥυφέω remains unclear: while the form is most prominently attested in IonicIonic, including at least Hipponax (C.1) and Hippocrates (very frequently, e.g. C.5), the claim that it was also DoricDoric finds support in a papyrus fragment of Sophron (C.2, see Hordern 2004, 140); on the other hand, the form is not attested in literary Aeolic, and Eustathius’ mention of it (B.5) next to the ‘Aeolic’ change ο > υ is possibly meant as a further attestation of the same change (with οὐ μὴν … ἀλλὰ introducing an additional argument, see Denniston 1954, 29–30), not as a (mistaken) attribution of ῥυφέω to Aeolic. The forms in ῥυφ- are most widely attested in the Hippocratic corpusHippocratic corpus, where besides (ἐκ)ρυφέω one finds the ‘double-nasal’ present ἐπι-/κατα-/προρρυμφάνω, the aorist imperative ῥυφησάτω, and the nominal derivatives ῥύφημα ‘porridge’, ῥύμμα ‘soup’, ῥυφητός ‘what can be supped up’, and ῥυπτός ‘id.’ As shown by Jouanna (1981), the spelling <ῥυφ-> is transmitted by the oldest and best codd. of the Hippocratic corpus, while <ῥοφ-> is an Atticisation, both in the verbal forms and in ῥόμμα, ῥοπτός. Moreover, the present forms ῥυφάνω and ῥοφάνω, recorded by modern lexica, are voces nihili, respectively a textual corruption of ῥυμφάνω and an Atticisation in Cornarius’ 1538 edition. It should be added that ῥυμφάνωῥυμφάνω could only have arisen in a dialect that had already undergone the shift ῥοφ- > ῥυφ-, since an o-grade would be out of place in a double-nasal present, which is always built on the zero grade of the root (see Jasanoff 2022). A thematic aorist ἔρρυφον (if this and not ἔρροφον is to be read for ἔρυφον in B.3: see F.2) could then have been built on ῥυμφάνω following the pattern of τυγχάνω, ἔτυχον. However, Jouanna (1981, 208) suggested that, on the contrary, ῥυμφάνω was derived from an old form ἔρυφον, similar to how ἐρυγγάνω ‘to burp, belch’ arose from ἤρυγον (see entry ἐρυγγάνω, ἤρυγον, ἐρεύγομαι, ἠρευξάμην).

Modern lexica also record a variant ῥοφάωῥοφάω (cf. LSJ s.v.), which is well attested in Byzantine and Modern Greek (see E.). The first recorded attestation is Colot. in Ly., P.Herc. 208 (= TM 59467) col. 9.3 [100–79 BCE] ῥοφωμένην (besides ῥοφοῦντι at l. 5); however, it is unclear whether the letter immediately to the left of a small gap is really the left half of an omega rather than the omicron in the diphthong ου, and the forms of this verb in this passage show several uncertainties in spelling (the scribe had originally written ΡΟΠ- and ΡΟΦΥΝΤΙ: see Crönert 1906, 166 n. 16). Other possible early attestations of the -άω inflection do not inspire much confidence: ῥοφῶσαν is a correction for ῥοφοῦσαν in a cod. of Demetrius’ On Style preserving a fragment of the 4th-century BCE Attic orator Demades (C.6), while ῥυφᾶν appears as a varia lectio for ῥυφεῖν in a few Hippocratic passages (Morb. 2.40.4, Aff. 30.5).

The variant ῥόφω is puzzling: even if it is genuine, it must be secondary, as the o-grade is at home in an iterative-intensive of the shape *CoC-éi̯e/o- (on the history of this formation, see Willi 2018, 257–68), but not in a simple thematic present. The only literary evidence for ῥόφω is, on the one hand, ῥόμμα and ῥοπτός, which have been shown to be faulty readings for ῥύμμα, ῥυπτός (see above), and, on the other hand, the form ῥοφόμενον transmitted in a pseudo-Galenic treatise (C.7). The glosses ῥύψαι· σμῆσαι, σμῆξαι. πλῦναι. λοιδορῆσαι. ῥοφῆσαι. καθᾶραι, ‘ῥύψαι: to wipe, to wipe off; to wash; to insult; to gulp down; to purify’ (Hsch. ρ 549) and ῥυψόμεθα· καθαρισόμεθα. ῥοφησόμεθα, ‘ῥυψόμεθα: we will purify; we will gulp down’ (Hsch. ρ 451) do not support a hypothetical Ionic *ῥύφω = ῥόφω, since they refer to ῥύπτω ‘cleanse, wash’ (see Jouanna 1981, 210–2, who suggested that ῥύπτω in turn derives from *ῥυφ-jω, an etymology that would require the assumption of ῥύπτομαι in Ar. Ach. 17 as an Ionism, and does not seem superior to the connection with ῥύπος ‘filth’). To sum up, it seems quite possible that ῥόφειν is an analogising invention by Aelius Dionysius or one of his sources, as suggested by Chantraine (DELG s.v.; see also Fiori 2022, 256 n. 580). The matter is further complicated by the fact that the entry which most clearly prescribes the barytone accentuationAccent of ῥόφειν pairs it with the variant ὄφλειν for ὀφλεῖν, which is quite a different case from a linguistic point of view, since it is a thematic aorist that was reinterpreted as a present (see entry ὄφλειν, ὀφείλειν, ὀφλισκάνειν). Vessella (2018, 237) rightly compares Phot. ο 364 (A.2) with Phryn. PS 39.12–5Phryn. PS 39.12–5 (on ἀκταινοῦν) and [Hdn.] Philet. 209[Hdn.] Philet. 209 (on ξυρεῖν), where ‘the lexicographer prescribes the position of the accent to identify the conjugation of the verb’. However, it should be noted that neither the contracted nor the uncontracted stem seems to have been perceived as intrinsically preferable: the Philetaerus prescribes ξυρεῖν over ξύρειν, while Phrynichus limits himself to noting that ἀκταίνω was used by Aeschylus (Eu. 36) and ἀκταινόω by Plato (Lg. 672c, not in the Phaedo as erroneously stated). The need to disambiguate between contracted and uncontracted inflections, which were often indistinguishable in written form without accent marks, may have been exacerbated by some cases of oscillation between -ω and -έω presents in later Attic and the koine, such as ἐπιμέλομαι/ἐπιμελοῦμαι and στέρομαι/στεροῦμαι (see Schwyzer 1939, 709, 721–2; Threatte 1996, 513–5; Mayser, Gramm. vol. 1,2, 115–6; Gignac 1981, 365–8). At any rate, if Phot. ρ 200 (A.3) also depends on Aelius Dionysius, the latter not only commented on the correct accentuation of the Attic form, but also contrasted it with Ionic ῥυφέω: indeed, it is typical of this lexicographer to contrast Attic and Ionic forms, with an implicit preference for the former (see entry Aelius Dionysius, Ἀττικὰ ὀνόματα). As regards the Antiatticist, it is possible that his entry, which may itself depend on Aelius, did not promote the Ionic variant, but simply took note of its attestation in Aristophanes (see F.1).

Atticising authors of the 2nd century CE use the contracted forms ῥοφέω, ἐρρόφησα and their compounds: Aelian, for instance, has several occurrences of both the present and aorist stem (NA 3.19 ἐκροφεῖ, 6.4 ῥοφοῦσι, 9.62 ἐπιρροφήσας, 14.5 ἐπιρροφῶν, 15.11 ἐκροφοῦσι [v.l. -ῶσι], 16.28 ῥοφῆσαι), as does Lucian, who on this point of verbal morphology makes no distinction between his own ‘normal’ usage (Macr. 25 ῥοφεῖν, VH 1.30 ἀναρροφῆσαν) and Lexiphanes’ hyper-Atticising diction (Lex. 5 ἐρρόφει = C.8; here the character’s comic linguistic error lies not in the morphology but in the semantics, as ῥοφέω is used classically only for liquid substances: see Weissenberger 1996, 206). These authors thus disregard prescriptions such as those of Aelius Dionysius (who may not have been representative of the Atticist mainstream on this matter) and instead align themselves with, e.g., Phrynichus (PS 128.11–3)Phryn. PS 128.11–3, who, by recommending the use of the compound ψυχορροφέω ‘to drain the soul’, implicitly accepted ῥοφέω as the correct form of the simple verb (see entry ψυχορροφεῖν). Attic ῥοφέω is also presupposed by further developments in Medieval and Modern Greek (see E.).

E. Byzantine and Modern Greek commentary

Attic ῥοφέω remained in use throughout the Byzantine period, even giving rise to new literary derivatives (see entry ψυχορροφεῖν). At the same time, the popular language attests to the variant ῥουφῶ (see Kriaras, LME s.v. ροφώ), which survives in Modern Greek ρουφάω ‘to swallow, slurp, suck’ (see LKN s.v.), continuing the contracted inflection condemned by Aelius Dionysius. The shift from the -έω to the -άω class of contract verbs (for the possible first attestations of which see D.) is common in Post-classical Greek, from the Hellenistic to the medieval and early modern periods (CGMEMG vol. 3, 1269–70, 1298–9). The root vocalism of ῥουφῶ (also seen in the medieval derivative ῥουφίζω, preserved in some modern dialects) also presupposes Attic ῥοφ-, but its evolution admits of different explanations: it could represent a change /o/ > /u/, often found near labial consonants (CGMEMG vol. 1, 73–5), or the same phonosymbolic substitution found in Ancient Greek ῥυφέω (Tichy 1983, 84). In particular, the classical Attic vocalism is preserved in the technical term απορροφώ ‘to absorb’ and its antonym προσροφώ ‘to adsorb’, which are semantic (and etymological) calques on the French absorber, adsorber, which in turn ultimately derive from the Latin cognate sorbeō (see LKN s.vv.).

F. Commentary on individual texts and occurrences

(1)    Antiatt. ρ 5 (A.1)

This entry, its use of Aristophanes as a validating attestation, and its relationship to other erudite sources have recently been discussed by Fiori (2022, 255–8). The Aristophanic quotation (C.4) is probably from the now lost first Wealth (408 BCE), since the form is not attested in the Wealth of 388 BCE. Given that Aristophanes otherwise only uses forms of ῥοφέω (10x), including the aorist infinitive ῥοφῆσαι in the Knights (C.3), most scholars conclude that in Ar. fr. 465 a non-Attic character is speaking (see Kassel, Austin PCG vol. 3,2, ad loc.; Colvin 1999, 279–80; Pellegrino 2015, 465). Fiori (2022, 258) thus concludes that the Antiatticist’s aim was probably not to prescribe ῥυφῆσαι against ῥοφῆσαι, since he does not usually resort to Aristophanic quotations to promote non-Attic dialectal forms, but the more moderate one of signalling that Aristophanes used the Ionic form with υ alongside the Attic one with ο. Accordingly, Fiori supports a reconstruction of the original text of the entry such as the one proposed by Caroli (2021, 102): ῥυφῆσαι· διὰ τοῦ υ <Ἴωνες>, Ἀριστοφάνης Πλούτῳ, <διὰ τοῦ o Ἀττικoὶ. Ἀριστοφάνης Δαιταλεῦσιν κτλ.>, ‘ῥυφῆσαι: <Ionic speakers> [say it] with υ, Aristophanes <uses it> in the Wealth; <Attic speakers [say it] with o, Aristophanes [uses it] in the Banqueters, etc.>’. Finally, Fiori argues that this entry depends on Aelius Dionysius, since he was the only imperial grammarian with a documented interest in the variation ῥοφέω/ῥυφέω, and he had a habit of contrasting Ionic and Attic forms. While this is true, a different source cannot be ruled out, especially given that the attribution to Aelius of the entry stressing the Ionic character of ῥυφέω (A.3) is speculative, and that the Ionic ῥυφέω is discussed by other erudite sources (see B.1, where the focus is on the syntactic construction with the accusative).

(2)    Phot. ε 1982 (B.3)

This entry in Photius’ lexicon, which Erbse attributed to Aelius Dionysius, prescribes a form ἔρυφον, which must be corrupt, against ἐρρόφησα, the usual sigmatic aorist of ῥοφέω. Both Erbse and Theodoridis accept Naber’s emendation ἔρροφον, which is not without problems. If ἔρροφον is taken to be the imperfect of the present ῥόφω, known to have been prescribed by Aelius Dionysius, then its contrast with ἐρρόφησα becomes problematic: it is unclear why Aelius (or anyone else) would have prescribed an imperfect as an alternative to an aorist. On the other hand, a thematic aorist ἔρροφον is not only unattested, but impossible next to the simple thematic present ῥόφω. If Jouanna’s reading ἔρρυφον is accepted, the entry would prescribe a thematic aorist that, though also unattested, would not be out of place next to ῥυμφάνω (see D.). In the latter case, however, it is unlikely that Aelius Dionysius prescribed a form such as ἔρρυφον, although he may have possibly mentioned it as an Ionic variant, together with ῥυφέω.