ἱστών and other place-nouns in -(ε)ών
(Phryn. Ecl. 137, Phryn. Ecl. 222, Phryn. PS 52.16, Antiatt. α 96, Σ^b α 1933)

D. General commentary

Phrynichus (A.1, A.2, A.3), the Antiatticist (A.4), and an Atticist source used in the expanded Synagoge (which may well be identified with Orus, A.5) deal with several place-nouns in -(ε)ών-(ε)ών. Only in A.1 is the noun’s form directly addressed (with Phrynichus prescribing – as may be expected – the Attic ending -ών instead of Ionic -εών), while the other entries discuss the individual words’ admissibility (A.2, A.5) or simply recommend their use (A.3, A.4). Following a brief historical-linguistic exploration, the commentary will discuss the place-nouns in -(ε)ών and their treatment in Atticist lexicography in the following order: first, the nouns for which the Atticist norm (A.1) explicitly prescribes the ending -ών rather than -εών (these are ἱστ(ε)ών ‘weaving-shed’, καλαμ(ε)ών ‘reed-bed’, ἱππ(ε)ών ‘horse stable’, and ἀνδρ(ε)ών ‘men’s apartment’); second, the nouns that are dealt with by the Atticists (in this case A.2 and A.5) with respect to their admissibility (these are κοιτών ‘bed-chamber’, προκοιτών ‘ante-chamber’, κοπρών ‘latrine’, ἀφεδρών ‘privy’, and λυτρών ‘privy’); third, the word βοών ‘cow stable’, which Phrynichus (A.3) recommends, alongside ἱππών, without discussing it further. Pollux’s lexicon includes (and thus approves) several other place-nouns in -(ε)ών, but these will not be treated in detail here given their absence from the other main Atticist lexica, e.g. λουτρών ‘bath-house’ (in Poll. 7.167Poll. 7.167, cf. e.g. [X.] Ath. 2.10), περιστερεών ‘dovecote’ (in Poll. 9.16Poll. 9.16, cf. e.g. Pl. Tht. 197c.3), and ξενών ‘guest room’ (in Poll. 9.50Poll. 9.50, cf. e.g. Eur. Alc. 543, 547). Similarly, ἀκανθεών (‘bush of acanthus’) and αὐλών (‘channel, trench’) appear to have been included respectively in Aelius Dionysius’ (cf. Ael.Dion. α 60Ael.Dion. α 60 reconstructed from Eust. in Il. 3.870.1–3) and Pausanias’ lexica (cf. Paus.Gr. α 168Paus.Gr. α 168 reconstructed from Eust. in Od. 4.232.12–4 and Σ^b 2413 = Phot. α 3181, ex Σʹʹʹ) but do not appear in the Atticist lexica of the imperial period. Finally, βολεών (‘dunghill’) is also overlooked by the main Atticist lexicographers but will be briefly investigated in the commentary, given the interest it raised among Hellenistic scholars investigating the Attic language (B.1, B.2, F.1).

Ancient Greek has a group of ‘derived place-nouns’ (Barth 2016, 1) that consistently end in -ών in Attic and in -εών in Ionic: e.g. Attic ἀνδρών ‘men’s apartment’ (cf. C.1, see also A.4, B.3) vs. Ionic ἀνδρεών (cf. C.2, see also A.1). This group of substantives may be traced back to a complex suffixSuffixes -ĕi̯ōn, which – according to Barth’s convincing reconstruction (2016) – derives from the combination of two morphemes: on the one hand, the suffix *-ei̯o- of adjectives such as χρύσεος/χρύσειος ‘golden’ from χρυσός ‘gold’ (which are traditionally classified as ‘material’ adjectives despite actually encompassing a broader semantic spectrum, cf. Barth 2016, 9–10); on the other hand, the so-called ‘Strabon suffix’ -ōn-/-ŏn- ‘which substantivizes thematic adjectives’ (Barth 2016, 8). The resulting complex formant, which is productive throughout Ancient Greek (cf. e.g. Chantraine 1933, 164), allows to create derivative nouns denoting natural and cultivated areas (e.g. ἀμπελ(ε)ών ‘vineyard’ from ἄμπελος ‘grapevine’) as well as places of human construction (e.g. ἱππ(ε)ών ‘horse stable’ – cf. C.6, see also A.1, A.3, B.4 – from ἵππος ‘horse’); cf. Barth (2016, 3–5).

In A.1, Phrynichus prescribes the contract ending -ών for the nouns ἱστ(ε)ών ‘weaving-shed’, καλαμ(ε)ών ‘reed-bed’, ἱππ(ε)ών ‘horse stable’, and ἀνδρ(ε)ών ‘men’s apartment’. Given that -ών is the regular Attic ending, this prescription is unsurprising. However, the potential reasons that led Phrynichus to specifically choose these four terms to illustrate the norm are worth examining.

The lemma of A.1 is the form ἱστεών (‘weaving-shed’, from ἱστός ‘loom’), which occurs once in Menander’s Samia (C.9) and is otherwise attested only in A.1, B.6 and Hsch. ι 1011 (ἱστεῶνα· ἱστῶνα). The form ἱστών (which was perceived to be the correct Attic form), instead, has no classical attestation and occurs almost exclusively in grammatical treatises or lexica (A.1, B.6, B.11). PapyriPapyri have both ἱστεών and ἱστών (the count includes only the occurrences of the word for which the ending is preserved in its entirety and excludes those cases in which it is supplied by the editors): the Ionic form in -εών occurs only in the Hellenistic period (5x), while the Attic form in -ών is found exclusively from the 1st century CE onwards (8x; see also the noun ἱστωνάρχης ‘controller of weaving’ and the derivatives ἱστωναρχία and ἱστωναρχικός, all first appearing in the 1st century CE). The epigraphic evidence is limited to the genitive ἱστῶνος, which occurs in a metrical inscription from Asia Minor (I.Mus. Iznik 103.2 [1st century CE (?)]; Şahin 1979 actually prints ἱστεῶνος, but on the stone – of which only a transcription survives – the ending clearly does not have an ε, cf. Peek 1980, 37). Overall, judging from the literary and documentary evidence, it appears that ἱστεών was the more ancient form while ἱστών appeared only later. However, Phrynichus likely prescribed ἱστών in accordance with the other nouns for which the opposition between the Attic ending -ών and the Ionic ending -εών was evident and well attested in classical authors (e.g. ἀνδρών vs. ἀνδρεών, cf. C.1, C.2). Moreover, given its scanty occurrences in literature, it seems reasonable to assume that Menander’s use of ἱστεών in C.9 was Phrynichus’ implicit target in A.1. A less straightforward picture emerges from Pollux’s passage on ἱστ(ε)ών (B.6), in which the lexicographer begins by stating that the ἱστών may also be called ἐριουργεῖον and that the noun ἐριουργία and the verb ἐριουργέω are attested in ancient authors. He then goes on to say βέλτιον δ’ ὁ ἱστεών (ὁϊστεύων BC), ‘it is better [to use] ἱστεών’, thus adopting an oppositional stance to that recommended by Phrynichus in A.1. The reading ὁϊστεύων of the MSS clearly points towards ἱστεών, and it appears reasonable that Pollux would recommend this form given that it is – to the best of our knowledge – the word’s only literary occurrence and that Pollux tends to accept Menander as an authoritative source of Attic language (albeit with varying degrees of caution, see Tribulato 2014, 204–7; Favi 2022, 311–2). Meanwhile, the previous section concerning ἐριουργεῖον and its acceptability based on the ancient attestations of ἐριουργία and ἐριουργέω suggests that βέλτιον should be interpreted as the expression of Pollux’s preference for ἱστ(ε)ών over ἐριουργεῖον, rather than his preference for one of two possible spellings of the ending. From this perspective, however, an explanation of the transmitted ὁϊστεύων as originating from ὁ ἱστών would be considerably less straightforward.

The second term that Phrynichus (A.1) cites as an example of the wrong (i.e. Ionic) ending of the place-nouns in -(ε)ών is καλαμεών (‘reed-bed’, instead of καλαμών, from κάλαμος ‘reed’). This form is found only in A.1 and then in late antique and Byzantine works (see E.) and is clearly less common than καλαμών, which is first attested in Strabo (C.11) and remains in use up until the Byzantine period (cf. E.). The documentary evidence is either rather scanty (with only two occurrences of καλαμών in two Carian inscriptionsInscriptions, i.e. Ι.Μylasa 803.10 and Ι.Mylasa 814.9 [2nd–1st century BCE]) or dubious (two instances of the genitive singular but with abbreviated ending καλαμ( ) in P.Ross.Georg. 2.28.26, 29 (= TΜ 12891) [Arsinoites, after 164 CE]).

The third form listed in A.1 as an evident blunder is ἱππεών instead of ἱππών. A.1 is the only attestation of ἱππεών, which is otherwise unattested both in literary and in documentary texts (perhaps this was a nonce-formation by Phrynichus?), while ἱππών is found already in Xenophon (3x, cf. C.6), then in historians from the post-classical period (e.g. Polybius 1x, Diodorus 1x, Strabo 2x, Flavius Josephus 1x), 1x in Lucian, and later mostly in grammatical texts, lexica, and Byzantine authors (cf. E.; as regards documentary texts, ἱππών is attested in inscriptions and in papyri dating to between the 3rd century BCE and the 4th century CE).

The last word mentioned in A.1 is ἀνδρ(ε)ών (‘men’s apartment’). In this case, there is a clear ancient opposition between the Attic form, which occurs already in Aeschylus (2x, cf. C.1), and the Ionic form, which was used by Herodotus (6x, cf. C.2). The Attic ἀνδρών is also found in Aristophanes (2x) and Xenophon (2x), in post-classical authors such as Philo (8x) and Plutarch (17x) as well as in Atticising writers such as Aelian (3x) and Philostratus (5x). By contrast, the Ionic form virtually disappears after Herodotus (with the exception of an epigram by Leonidas of Tarentum, AP 9.322 and a single occurrence in a late-Byzantine historian, cf. E.). Epigraphic evidence for ἀνδρών is abundant (it is found in more than 30 inscriptions, most of which are 3rd–2nd-century BCE inventories of offers redacted by the Delian ἱεροποιοί, cf. e.g. IG 11,2.156.35–9, and I.Délos 290.25–6; cf. Vial 1984, 351–3), while ἀνδρεών is found only in two Sicilian inscriptions, also from the 3rd century BCE (in which ἀνδρεών seems to indicate a room in the gymnasium used only by men, see Nenci 1991, 923–4). The distribution of ἀνδρών and ἀνδρεών in papyri stands in contrast to that of ἱστών and ἱστεών: the Attic form alone is attested in papyri from the Hellenistic period (5x), while ἀνδρεών first appears in the 4th century CE (5x total occurrences). The Antiatticist (A.4) dedicates an entry to this word, stating that it applies ‘to the big room’ (ἐπὶ τοῦ μεγάλου οἴκου) and referring to Herodotus’ first book (C.2; but the historian uses the term also in Hdt. 3.77.3, 3.78.3, 3.121.1, 3.123.1, and 4.95.3). The entry signals the semantic nuance ‘big room’, i.e. ‘banqueting hall’ (cf. LSJ s.v.): this is surely the meaning implied in Aeschylus (C.1) but not necessarily also in Herodotus (C.2), given that a banqueting hall does not appear to be a suitable location in which to store weaponry, while the men’s apartments would be suitable for such a purpose. Nonetheless, the Antiatticist’s entry might have served a further purpose – namely, that of promoting the Ionic form ἀνδρεών (i.e. the form that was likely spreading in non-literary Greek in the imperial period, as the papyri evidence would indicate; cf. above). This is suggested not only by the fact that the author quoted in the entry is Herodotus, who obviously uses the Ionic variant, but also, by contrast, by the fact that Phrynichus prescribed ἀνδρών. In sum, it is highly likely that the lemma of A.4 originally read ἀνδρεών (as already suspected by Valente 2015, 108).

A.2 and A.5 discuss other place-nouns but only with respect to their admissibility. The text of A.2, however, poses several interpretive problems, as far as the word κοιτών (‘bed-chamber’ from κοῖτος ‘bed’) is concerned. Indeed, in its transmitted form, the entry reads τὸ μὲν κοιτὼν ἀδόκιμον, τὸ δὲ προκοιτὼν οὐ δόκιμον, i.e. ‘κοιτών is unapproved, while προκοιτών is not approved’, which seems rather abstruse wording for simply proscribing two terms (unless one translates ‘κοιτών is unapproved, προκοιτών is [also] not approved’ or ‘κοιτών is unapproved, and [therefore] προκοιτών is not approved’). Still, had Phrynichus’ intent been to reject both κοιτών and προκοιτών, one might expect a more straightforward formulation, such as κοιτὼν καὶ προκοιτὼν ἀδόκιμα (‘κοιτών and προκοιτών [are] unapproved’, which, incidentally, is the wording found in fam. q, cf. below). Lobeck (cf. the apparatus of A.2) proposed deleting the ἀ- in ἀδόκιμον, while Fischer suggested integrating <οὐκ> ἀδόκιμον (cf. Phryn. Ecl. (fam. q) 174, on which, see entry οἰκόσιτος, αὐτότροφος, οἰκότριψ, οἰκογενής). Both solutions move towards restoring an opposition between the disapproval of προκοιτών and the approval of κοιτών: a parallel in this respect would be the rejection of the derivative προβασκάνιον and the acceptance of simple βασκάνιον in Phryn. Ecl. 60Phryn. Ecl. 60, although, in this case, both forms are synonyms for ‘amulet’, while κοιτών and προκοιτών appear to denote two different rooms. Nonetheless, if Phrynichus actually accepted κοιτών and rejected only προκοιτών, the wording in fam. q may be explained as a rephrasing of the already corrupt wording of the main redaction; MS T, meanwhile, omitted the first portion of the text, which appears to further confirm that τὸ μὲν κοιτὼν ἀδόκιμον, τὸ δὲ προκοιτὼν οὐ δόκιμον in the main redaction was perceived as problematic. Fischer’s proposal <οὐκ> ἀδόκιμον seems preferable to Lobeck’s {ἀ}δόκιμον, in that it gives a better sense of Phrynichus’ unenthusiastic acceptance of κοιτών, a word that – aside from an isolated attestation in Aristophanes (C.4) – was found almost exclusively in koine Greek: it is amply attested in the Septuagint (15x) and other religious and hagiographical texts but also in pagan authors of the imperial period (Galen 6x, Athenaeus 6x, Cassius Dio 9x; for a similar case in which Phrynichus appears to accept a form contrary to his own volition and judgement on the sole grounds that it is scantily attested in Attic authors, see Phryn. Ecl. 334Phryn. Ecl. 334 and entry βόλβιτον, βόλιτον). The diffusion of κοιτών in koine texts is confirmed by its numerous documentary attestations: inscriptions attest in particular to the phrase ἐπὶ τοῦ κοιτῶνος (11x) indicating the Roman cubicularius (i.e. the slave in charge of the master’s bed-chamber) and, later, the imperial chamberlain (cf. e.g. I.Nysa 23.4–5 [Caria, 117–138 CE]). Moreover, the word occurs in more than 30 papyri dating between the 3rd century BCE and the 7th century CE. The determining factor in Phrynichus’ cautious acceptance of κοιτών was arguably that the word was attested (though only once) in Aristophanes (C.4), and could thus be considered properly Attic, despite its large use in koine texts. In his discussion of κοιτών, Pollux (B.3) reports that the word was in fact used not only by Aristophanes but also by Menander (C.10), in a manner that implied that the playwright considered the word ‘barbaric’ (on the possible context in which κοιτών occurred in Menander, cf. AGP vol. 1, 190–1, with further bibliography). In sum, Pollux’s treatment of κοιτών (B.3) seems to confirm that the word was perceived by the Atticists as a ‘borderline case’ that required argumentation to justify its status as an approved term (see, in particular, Pollux’s remark that Aristophanes is πιστότερος ‘more reliable’ than Menander).

Despite his (unenthusiastic) acceptance of κοιτών, Phrynichus (A.2) proscribed the derivative προκοιτών. This must have been in use to some extent in the imperial koine, given that Pollux mentions it (B.8), but, aside from the two lexicographers, the word is found only in grammarians (Hdn. Περὶ καθολικῆς προσῳδίας GG 3,1.36.24, GG 3,1.40.25, Hdn. Περὶ κλίσεως ὀνομάτων GG 3,2.727.31, Theodos. Περὶ κλίσεως τῶν εἰς ων βαρυτόνων 18.3, [Arcad.] 17.6) and is wholly absent from documentary texts. Again, Phrynichus’ choice is understandable in the absence of any authoritative Attic attestation of προκοιτών. Interestingly, however, the word he prescribes instead of προκοιτών, i.e. προδωμάτιον, is equally unattested in classical authors and is found, besides A.2, only in an entry in Hesychius (π 3390), with no documentary attestations. The simple δωμάτιον, on the contrary, is unquestionably Attic (Aristophanes 5x, Plato 1x, Lysias 5x) and Phrynichus’ preference for προδωμάτιον is clearly based on this.

A.5 (an Atticistic entry in the expanded Synagoge, likely going back to Orus) also addresses place-nouns in -ών, distinguishing between admissible and inadmissible forms (βάρβαρα). As may be expected, the terms rejected as ‘barbaric’ by A.5, i.e. ἀφεδρών and λυτρών (both meaning ‘privy’), first appear in koine Greek: ἀφεδρών is found twice in the New Testament and later only in religious texts. The only documentary attestation is in OGIS 483.233, 234–5 = SEG 13.521, an inscription from Pergamon that is generally dated to the 2nd century BCE, although the case has also been made for a later dating based, among other pieces of evidence, on the presence of ἀφεδρών in the text (cf. Zuchtriegel 2008). λυτρών, meanwhile, occurs once in the Septuagint and then only in lexica (cf. B.10) and has no occurrence in documentary texts. Another factor at play in the rejection of ἀφεδρών and λυτρών (besides the lack of classical attestations) may be the two nouns’ rather obscure derivation. Indeed, neither ἀφεδρών nor λυτρών has a noun from which it clearly derives and to which it is semantically connected (as happens for all the other place-nouns examined so far, cf. above). In the case of ἀφεδρών, the noun closest to it is the feminine substantivised ἡ ἄφεδρος, which generally means ‘menstrual period’ (cf. LSJ s.v.) but arguably originated as a translation of the Hebrew niddâ (‘menstrual impurity’ but also ‘lustration’), in turn, deriving from the root ndd (‘to chase away’, ‘to expel’) in relation to the Jewish custom of separating menstruating women from the rest of the family (cf. Milgrom 1991, 744–5): the literal meaning of ἀφεδρών, however, is ‘place of the external seat’ i.e. ‘outhouse’, ‘latrine’. λυτρών, meanwhile, is clearly linked to λῦμα ‘water used in (or dirt removed by) washing’ (cf. LSJ s.v.), but the two words are not as formally close as e.g. καλαμών and κάλαμος or ἱππών and ἵππος, and this may have posed a further problem as far as the Atticists were concerned.

κοπρών (‘latrine’ from κόπρος ‘excrement’) is accepted by A.5 on the grounds of its clear Attic pedigree (Aristophanes 2x, cf. C.3, Demosthenes 1x, Eubulus 2x; it also occurs in Attic inscriptions dating from the 4th century BCE onwards, i.e. IG 2².380.38–9, IG 2².2496.11, IG 2².2742.1–2, and Agora XIX H.110.1, but is absent from documentary papyri). This entry is not isolated in recommending κοπρών, for Pollux includes the term twice in his lexicon (B.5, B.7). Remarkably, the term is found in a 2nd-century CE papyrus commentary on an unidentified text – i.e. P.Oxy. 35.2744 (B.1) – with the Ionic ending -εών (acc. plur. κοπρεῶνα̣ς̣), which is otherwise found only in a few Byzantine works (mostly by John Tzetzes, cf. E.). In B.1, κοπρ(ε)ών glosses another place-noun with the same ending, i.e. βολεών (βο̣[λε]ῶ̣[ν]α̣ς δὲ καὶ τοὺς κοπρεῶνα̣ς̣ καλεῖσθαί φησιν̣ Ἀμερ[ί]α̣ς, ‘Amerias says that the latrines are also called ‘dunghills’’; for an analysis of the archaeological evidence concerning the terms κοπρών, βολεών, and ἀφεδρών, see Zuchtriegel 2006). Given that B.1 provides the only ancient attestation of the Ionic variant κοπρεών, one wonders whether this might have been a mistake influenced by the preceding βολεών (which, instead, only exists with this ending, cf. F.1). Ηowever, the fact that κοπρεών resurfaces during the Byzantine period (cf. E.) further complicates the picture.

Finally, in A.3, Phrynichus recommends βοών (‘cow stable’, from βοῦς ‘bull, cow’) alongside the already mentioned ἱππών. βοών occurs only here and in a few other grammatical and lexicographical works (among which see B.9) but has some relevant documentary attestations: it is found in two inscriptions (i.e. IG 14.645.139 and 143 [Heraclea, Magna Graecia, 4th–3rd century BCE] and I.Mylasa 206.11 [2nd–1st century BCE]), as well as in five 3rd-century BCE papyri from the Arsinoites, i.e. P.Lond. 7.1995.206 (= TM 1557), P.Cair.Zen. 2.59292.136 (= TM 936), P.Cair.Zen. 2.59235.2 (= TM 880), PSI 5.497.4 (= TM 2124), and PSI 4.422.38 (= TM 2105; the papyrus reads βοουνων, but this is a misspelling of βοώνων, cf. Messeri Savorelli, Pintaudi 1995, 115; on scribal errors of this nature, see Mayser, Gramm. vol. 1.1, 76–7). In other words, despite the lack of literary attestations, βοών appears to have been a rather ancient term that was maintained in the early koine but – as far as the extant evidence allows us to reconstruct – fell out of use after the 3rd century BCE (interestingly, Philoponus in the 6th century CE described βοών as typical of the Greek of Alexandria, cf. B.9; on the Atticists’ perception of Alexandrian Greek, cf. entries ἐλέγοσαν, ἐγράφοσαν, ἐσχάζοσαν, ἐξαλλάσσω, ἐξάλλαγμα, and χειμάζω). This, along with the parallel with ἱππών (which instead is well attested in literature, cf. above), likely made βοών an acceptable word for Phrynichus (A.3).

In sum, given the productivity of the formant -(ε)ών for place-nouns in Classical and Post-classical Greek, the Atticists were faced with an array of substantives belonging to this morphological category, which had different features and posed different problems. Among those for which an actual alternation between the endings -ών and -εών is attested, ἀνδρ(ε)ών presented a truly ‘dialectal’ alternation between Attic ἀνδρών and Ionic ἀνδρεών, thus requiring Phrynichus to prescribe the former rather than the latter (cf. A.1 and above), while the Antiatticist probably allowed for ἀνδρεών (cf. A.4 and above). In the case of ἱστ(ε)ών, the alternation was not dialectal but rather was diachronic, with ἱστεών being the ancient form and ἱστών the more recent (nonetheless, Phrynichus preferred ἱστών in analogy with words such as ἀνδρών, cf. A.1). For καλαμ(ε)ών (and perhaps also κοπρ(ε)ών) the alternation was synchronic, with the form in -ών being the most widespread and that in -εών being a minor variant (which Phrynichus still made sure to proscribe, cf. A.1). Other place-nouns of this category, however, only ever existed with the ending -ών and therefore troubled the Atticists not in terms of their morphological appearance but rather in terms of their semantics, formation, and admissibility. In these cases, as may be expected, it was primarily the Attic attestations (or lack thereof) that settled the matter: ἱππών, κοιτών, and κοπρών were accepted (A.1, A.2, A.3, A.5, B.3, B.4, B.5, and B.7) on the basis of their occurrences in Xenophon (C.6) and Aristophanes (C.3, C.4), respectively (βοών, having no Attic literary attestation of its own, was likely recommended in analogy to ἱππών, cf. A.3 and above). Meanwhile, προκοιτών, ἀφεδρών, and λυτρών lacked reputable literary attestations and were rejected accordingly (A.2, A.5).

E. Byzantine and Modern Greek commentary

Among the place-nouns in -(ε)ών analysed here, κοιτών (‘bed-chamber’) is by far the most widely attested in the Byzantine period, both in high-register works (e.g. Constantinus VII 98x, Michael Psellus 5x, John Zonaras 12x) and in less linguistically refined texts, such as hagiographies and other religious works. ἀνδρών (‘men’s apartment’, ‘banqueting hall’) is also widely attested in high-register texts (e.g. Theodorus Studites 2x, Photius 3x, Michael Psellus 5x, Nicetas Choniates 7x), as are ἱππών (‘horse stable’, e.g. Photius 1x, Michael Psellus 2x, Tzetzes 1x) and ἀφεδρών (‘privy’, e.g. Photius 3x, Georgius Cedrenus 2x, Nicephorus Blemmydes 1x). Similarly, καλαμών (‘reed-bed’) is present in high-register works (e.g. Anna Comnene 6x), while the variant καλαμεών is considerably less frequent and is mostly confined to monastic documents. κοπρών (‘latrine’) also has a few attestations in high-register texts (cf. e.g. Leo VI 1x, John Tzetzes 1x, Michael Choniates 1x) and, interestingly, the variant κοπρεών occurs only three times, always in works by John Tzetzes (H. 6.59.515, 12.399.242, Prolegomena de comoedia 2.34). Byzantine attestations of προκοιτών (‘ante-chamber’), λυτρών (‘privy’), βοών (‘cow stable’), and ἱστ(ε)ών (‘weaving-shed’) are limited to lexicography and grammar (the only exception being one occurrence of ἱστών in Paul of Monemvasia, Narrationes 16.80).

In Medieval Greek, place-nouns in -(ε)ών become paroxytone nouns in -ώνας, which follow the declension pattern of substantives such as ἀστέρας (parisyllabic plural ἀστέρες; cf. the examples in CGMEMG vol. 2, 304). Several of the nouns examined here survive in this shape – namely, κοιτώνας (from ancient κοιτών), ἀφεδρώνας (from ancient ἀφεδρών), and κοπρώνας (from ancient κοπρών), see Kriaras, LME s.vv. Meanwhile, καλαμιώνας is a derivative in -ώνας from the Byzantine καλάμι (‘reed’). κοιτώνας, ἀφεδρώνας, and καλαμιώνας remain in use in Modern Greek (cf. LKN s.vv.).

F. Commentary on individual texts and occurrences

(1)    P.Oxy. 35.2744.col. ii.9–14 (B.1), Harp. β 16 (B.2)

Although it was apparently overlooked by the Atticists of the imperial period, βολεών (‘dunghill’, lit. ‘place of the throw’, from βόλος/βολή ‘throw’) appears to have attracted the attention of Hellenistic grammarians interested in the Attic dialect, as demonstrated by two testimonies from the 2nd century CE. One is P.Oxy. 35.2744 (B.1), a papyrus commentary on an unidentified work in which the commentator mentions the term βολεῶνες and states that, according to the Hellenistic glossographer Amerias, the βολεῶνες were also called κοπρεῶνες (see Valente 2005, 284–5; on Amerias in general, see Pagani 2021 with further bibliography). This information is then followed by a quotation from a speech by Dinarchus that contains βολεῶνες (C.7). The other testimony of a Hellenistic interest in βολεών as a typically Attic word is an entry in Harpocration (B.2) that attests that the Hellenistic erudite Nicander of Thyatira (FGrHist 343 F 3) defined the βολεῶνες as ‘[the places] in the fields, in which the dung is thrown’. The fact that both the papyrus commentary (B.1) and Harpocration (B.2) could refer to ancient scholars such as Amerias and Nicander of Thyatira in their discussions of βολεών suggests a common source that dealt with the word and cited multiple Hellenistic authorities. This common source may well be DidymusDidymus, as already proposed by Luppe (1970, 40–1).