κόκκοι, ῥόα
(Antiatt. κ 53, Poll. 6.80, Moer. ρ 9, Phot. ρ 145)

D. General commentary

An entry in the Antiatticist (A.1) reports that some scholars did not deem the word κόκκος (‘grain; seed’) acceptable for denoting pomegranate seeds and grapes, and instead recommended ῥάξ ‘grape’. The Atticists’ predilection for ῥάξ is confirmed by several entries recommending it against its variant ῥώξ, which is later and more widespread: besides B.1, see [Hdn.] Philet. 182[Hdn.] Philet. 182 and Phot. ρ 38 (the latter identified as Ael.Dion. ρ 6Ael.Dion. ρ 6; see entry ῥάξ, ῥώξ). By contrast, A.1 is the only source to inform us that some scholars, perhaps stricter Atticists, objected to the use of the word κόκκος. Since both in its earliest occurrences (C.1, C.2, C.3, C.4) and in the erudite sources discussing it (A.1, A.2) κόκκος mainly denotes the pomegranate seed and is thus associated with the word ῥόα/ῥοιά ‘pomegranate’ – itself a debated form in Atticist and Byzantine lexicography owing to its range of variants – the following analysis discusses the Atticist theories concerning both κόκκος (A.1, A.2) and ῥόα/ῥοιά: (A.3, A.4, B.6).

κόκκος is of unknown origin (see DELG and EDG s.v.) and is certainly very ancient, since it already occurs in the Homeric Hymn to Demeter (see h.Hom.Cer. 372 = C.1, 412). In Classical Greek, it is found in Herodotus (C.2) and in Hermippus (C.3), as well as in the compound σκληροκόκκος (‘with hard seeds’), a hapax in Antiphanes (C.5; see Olson 2023, 237–8). In all these early occurrences κόκκος always denotes the pomegranate seed; numerous later occurrences, however, show that it could also denote the seed or kernel of any fruit or plant, including the kermes oak, whose berry was used to produce the red dye, hence the meaning ‘scarlet’ (LSJ s.v. κόκκος, 2) and the diminutive κόκκινοςκόκκινος ‘red’, still used in Modern Greek. In the feminine, κόκκος denotes the kermes oak tree itself (see LSJ s.v. κόκκος, 2.2). In antiquity and the medieval period, the word is especially frequent in medical treatises describing the properties and uses of seeds.

In spite of the Antiatticist’s claim (A.1), ancient scholarship – as far as can be determined from extant sources – shows no reservations about κόκκος, nor in particular about the expression κόκκοι ῥόας. Pollux (A.2) tacitly accepts this phrase, and the same appears to be true of Aelius Dionysius, since, according to Eustathius’ testimony (B.5), he lists pomegranate seeds (κόκκοι ῥόας) among the ingredients of a sauce called ἀβυρτάκη. Nonetheless, given the distribution of the form, it is plausible that certain Atticists adopted a more rigorous stance on κόκκος than Pollux or the author of the Antiatticist, both of whom are known for their relatively permissive approach. Indeed, κόκκος is a rather common word across different registers of koine Greek. For instance, it occurs in well-known metaphors of the New Testament, including the parable in which the kingdom of God is compared to a mustard seed that grows into a tree high enough to offer shelter to birds (see C.6, Ev.Marc. 4.30–2, Ev.Matt. 13.31–2). Therefore, strict Atticists may have disapproved of the word solely because it remained in regular use in the lower registersRegister of the koine. Alternatively, κόκκος may have been perceived as insufficiently precise, given its broad applicability to a wide range of referents. Even the aforementioned classical occurrences may not have been regarded as a suitable precedent, either because they are markedly epic (C.1) or insufficiently canonical, as is the case of Herodotus (C.2) – on the controversial status of this author in the Attic canon, see Valente (2015, 54 n. 319); on ancient erudition on Herodotus’ language, and in particular on the notion that it was a mixed variety, see Tribulato (2022, 242–8). The same applies to Antiphanes (C.5), a playwright of Middle Comedy, whom stricter Atticists did not consider a suitable enough Attic model (see Tribulato 2024). Conversely, if the Antiatticist entry does not depend on a now lost locus classicus, both Herodotus’ passage and Hermippus’ line (C.3; see Comentale 2017, 150–2) may have constituted the classical precedent supporting the use of κόκκος (ῥόας) in the original version of the lexicon. Indeed, HerodotusHerodotus plays an important role in the Antiatticist’s canon, which is also more inclusive in its treatment of Attic comedy (compared, for instance, to stricter Atticists such as Phrynichus, who tend to rely on a narrow pool of authors; see Tribulato 2024).

For his part, Pollux, besides (implicitly) accepting κόκκος, does not hesitate to promote the use of its derivative κοκκίζω ‘to remove the seeds’, a hapaxHapax occurring in an Aristophanic line said to be composed ‘as Aeschylus’ (C.4), most likely meaning that it quotes Aeschylus’ text and thus perhaps is part of a paratragic passage. The interpretation of the line remains uncertain due to the lack of further context: see Bagordo (2016, 139–42) for various possibilities, including an erotic understanding of the pomegranate’s ripeness. The prefixed ἐκκοκκίζωἐκκοκκίζω is better attested than the simplex form: it occurs three times in Aristophanes (Ach. 1179; Pax 63; Lys. 364), always with figurative meanings – respectively ‘to break (a bone)’, ‘to empty (the city)’, and ‘to tear out (hair)’ (see LSJ s.v.) – which are amply discussed by ancient scholiasts (see also Su. ε 524).

As for the word ῥόα ‘pomegranate’ (denoting both the tree and the fruit, see below), it displays a series of variants. Besides ῥόα, there are also the forms ῥοιά and ῥοιή, the latter typical of epic diction (the form occurring in Homer; see C.1, Od. 7.115 and 11.589) and of Ionic (as in Herodotus; see C.2). As for the accentuationAccent, the word is usually paroxytone, but the oxytone form ῥοά, is promoted by [Arcad.] 114.9 (cf. [Hdn.] Περὶ καθολικῆς προσῳδίας GG 3,1.301.27) and occasionally occurs in the manuscripts (see the apparatus of B.2). The series of variants ῥόα/ῥοιά/ῥοιή is not unparalleled, but appears analogous to that of χρόα/χροιά/χροιή ‘skin, appearance’, as noted by both ancient (B.4) and modern scholarship (see DELG s.v. and EDG s.v.). The origin of this word has been debated at length; in addition to DELG s.v. and EDG s.v., see Witczak, Zadka (2014, 65–7) for an overview of the various interpretations proposed in modern scholarship. They argue that ῥόα/ῥοιά represents the Greek development of the IE *rosi̯ā́ (or *rosā́) ‘juicy fruit’, referring to the pomegranate’s juiciness and thus originally denoting the fruit. The form may be attested as early as Mycenaean Greek: tablet KN Xd 148 attests a form ro-a, which may correspond to the plural ῥόαι ‘pomegranates’, while tablet PY Va 482 attests a form ro-i-ko, possibly reflected in the Greek diminutive ῥοΐσκος ‘small pomegranate’ (see Witczak, Zadka 2014, 62–3; note, however, that ῥοΐσκος is not attested before the Septuagint).

Ancient lexicographers, including the Atticists, frequently debated the variants ῥόα, ῥοιά, and ῥοιή. In classical Attic, the preferred form is ῥόα, which is used by Plato (Lg. 845b.7) and by the comic playwrights: see the occurrences in Hermippus (C.3), Aristophanes (cf. C.4 and V. 1269; the form ῥοιά in Pax 1001 is due to metrical constraint), Epilycus (fr. 2, see A.4), Antiphanes (see C.5 and fr. 66, although Athenaeus transmits the form with ι in both cases), and Ephippus (fr. 24). As Olson (2023, 237) argues, the Ionic form tends instead to prevail from the 4th century onwards. The existence of multiple variants appears to have caused some confusion among ancient scholars, so much so that some resorted to an ill-founded distinction to explain the variation between them: according to this theory, to which Pollux himself (A.2) appears to lend credence, ῥοιά denotes the tree, while ῥόα the fruit (cf. [Ptol.Ascal.] Diff. 135 Palmieri, [Ammon.] 430). Both Eustathius (B.4) and Thomas Magister (B.6) rightly deny the validity of this distinction – a distinction in which they may have taken particular interest, since in the spoken language of their time the tree and its fruit were designated by two different words (see E.). Nor was there consensus on which form was preferable. Indeed, ῥοιά is labelled as IonicIonic in an entry in Photius’ lexicon which has been conjecturally attributed to Aelius Dionysius (see A.4, with apparatus). Although this doctrine cannot be securely attributed to Aelius, it is very likely to originate in Atticist erudition, given the mention of the Old Comedy playwright Epilycus with regard to the use of ῥόα, which the entry implicitly presents as the Attic form. Conversely, Moeris (A.3) – later followed by Thomas Magister (B.6) – states that ῥοιά is the Attic, and therefore preferable, form for ‘pomegranate’. It is unclear why Moeris recommends ῥοιά over ῥόα, and it cannot be ruled out that the present form of the entry may result from the reduction of a fuller discussion into the binary format in which Moeris’ lexicon circulated (on which see entry Moeris, Ἀττικιστής).

E. Byzantine and Modern Greek commentary

Both κόκκος and ῥόα/ῥοιά remain in use throughout the late antique and Byzantine periods. In the medieval period, κόκκος was widely used across registers with very different levels of literary ambitions. As a common term in Medieval Greek (see Kriaras, LME s.v., LBG s.v.), it is employed in lower-registerRegister texts such as the War of Troy and the Digenis Akritis; nonetheless, it is also found in classicising literature, in both prose and verse, and is regularly used, among others, by Photius, Michael and Nicetas Choniates, Manuel Philes, and Maximus Planudes. The presence of seed metaphors in the New Testament – and in particular the image of the mustard seed (see D.), which enjoyed a great popularity – certainly contributed to the word’s widespread use in Byzantine texts at all levels.

In Modern Greek, κόκκος remains standard and primarily denotes cereal grains, but also coffee beans, peppercorns, and even grains of sand. It is also used metaphorically to indicate a minimal quantity, even in a moral sense, e.g. of common sense (compare the English expression a grain of…, and the Italian un briciolo di…). Although κόκκος can be used, the more common word for the seeds of fruits and plants in Modern Greek is the already ancient σπόρος. The derivative κόκκινος ‘red’ is still standard in Modern Greek.

ῥόα/ῥοιά is likewise extensively used by Byzantine authors. The distribution of these variants is difficult to evaluate with certainty, both because these forms are likely to have been altered in transmission and because of editorial practices. In general, ῥόα and ῥοιά appear to be interchangeable, but more learned authors were arguably aware that ῥόα was the Attic form: for instance, it is likely no coincidence that Photius, who records this information in his lexicon (A.4), transmits Antiphanes’ fr. 66 with the form ῥόας, whereas Athenaeus quotes the verse with the form with ι (as does Hesychius in C.3; see apparatus). In the medieval period, both ῥόα and ῥοιά belong to literary registers, since the everyday forms for the fruit are ροίδι(ν) and, with dissimilation of the diphthong, ρόδι(ν), deriving from the ancient ῥοΐδιον. The latter is the diminutive of ῥόα/ῥοιά and itself attracted the attention of Atticist scholars (see AGP vol. 2, Word formation, forthcoming). The name of the tree, by contrast, alternates between ῥοϊδέα/ῥοϊδία/ῥοϊδιά. ῥοϊδέα is offered by [Zonar.] 1619.4 as the interpretamentum of ῥόα/ῥοιά (ῥόα καὶ ῥοιά· ἡ ῥοϊδέα). Further evidence, especially regarding dialectal forms (including those of southern Italy), is collected in Witczak, Zadka (2014, 65), with bibliography.

The medieval forms continue in Modern Greek, which distinguishes between ροδιά ‘pomegranate tree’ and ρόδι ‘pomegranate’. An older form ρόιδο, now obsolete, survives in the saying τα έκανες ρόιδο, meaning ‘you’ve made a mess!’.

F. Commentary on individual texts and occurrences

N/A