ὀστοῦν, ὀστέον
(Moer. ο 27)

D. General commentary

An entry in Moeris’ Atticist lexicon (A.1) prescribes the contracted form ὀστοῦν ‘bone’ against the uncontracted ὀστέον. While this is not the only Atticist prescription concerning thematic nominal forms whose stem ended in a vowel that was susceptible to contraction with the thematic vowel, this noun’s origin and distribution differ somewhat from those of contract adjectives of materialAdjectives of material in -οῦς < -εος (see entry χρύσεος, χάλκεος, φοινίκεος) and multiplicatives in -πλοῦς < -πλόος (see entry ἁπλᾶ, διπλᾶ, τριπλᾶ).

The noun ὀστέον is likely a thematic derivative *h₃esth₁-éi̯-o- of the i-stem found in Sanskrit ásthi, Hittite ḫaštāi (< *h₃esth₁-i-; on the accentuation see Dieu 2022, 280; the reconstruction of the initial laryngeal is disputed: *h₃e- is defended by Beekes, EDG s.v. and Kloekhorst 2008, 325 on the basis of Latin os, in which an e-grade is expected; *h₂o- by Mayrhofer 1992, 150‒1 and Kimball 1999, 392 on the basis of Welsh asgwrn). Unlike other contract nouns and adjectives, ὀστέον from the beginning included an accentAccent on the penultimate syllable (itself considered exceptional by Herodian, who compared other nouns in -εον that were proparoxytone before contraction, see B.1) and regularly ended up with perispomenon accentuation after the contractions took place. Thus, Atticists were concerned only with the contract endings and not with the accentuation.

The distribution of ὀστέον/ὀστοῦν in literary texts differs somewhat from that of contract adjectives (see, in detail, Gammage 2018, 134‒6). While this noun is customarily contracted in Plato, Aristotle, and the orators, its distribution in tragedyTragedy appears to be tied to specific case-forms: the three major tragedians all attest the uncontracted gen. plur. ὀστέων, while Euripides also has 7 instances of the contracted nom.-acc. plur. ὀστᾶ. In comedy, only contracted forms are attested (see, e.g., C.1 in Cratinus, quoted by Herodian), except for ὀστέων in Ar. Ach. 1226 (C.2), in a paratragic contextParody. Non-Attic literary texts from the classical period, from Homer to choral lyric and Herodotus, only attest uncontracted ὀστέον, whose frequency increases in the koine. In the Septuagint, this noun is mostly uncontracted, except for the nom.-acc. neut. plur. ὀστᾶ; the only occurrence of contraction in the New Testament is ὀστοῦν at Ev.Io. 19.36, in a quotation from the Old Testament (see Blass, Debrunner 1976, 37). In terms of non-literary sources, ὀστᾶ is the usual form in Attic inscriptions, but ὀστέα appears in late or metrical texts (see Threatte 1996, 37‒8). Both ὀστοῦν and ὀστέον occur in documentary papyri from the Roman period (Gignac 1981, 33‒7). In sum, the distribution of contracted and uncontracted forms in both Attic and Hellenistic sources is far from clear-cut (see Gammage 2018, 134‒6). Nevertheless, Moeris labelled only ὀστοῦν as the proper Attic form (see also B.2, while some later sources [B.3] simply present the two forms as equivalent). It is possible that such prescriptions were based not on imperfect knowledge of actual Attic usage but on the desire to provide those who wished to write in correct Attic Greek with a clear and straightforward rule. Given Attic’s overall strong tendency towards contraction, it is unsurprising that purists recommended the contracted form for this noun.

The practice of 2nd-century-CE writers, however, only partly squares with prescriptions such as that given by Moeris. If the manuscript evidence as represented in modern editions is to be trusted, the most rigorous authors are Aelius Aristides, who uses only the contracted forms (4x), and Lucian, with 22 occurrences of contraction against a single occurrence of ὀστέα at VH 1.31 (ὀστέ’ at Par. 47.8 is part of a quotation from Hom. Il. 23.83) and one of ὀστέων at Nec. 15.11 (but with the varia lectio ὀστῶν). Some uncontracted forms that occur alongside the contracted forms in Aelian and Philostratus the Elder are considered by Schmid to be literary Ionisms (Atticismus vol. 3, 19; vol. 4, 14), but this seems unlikely. Moreover, the actual figures for Aelian differ substantially: Schmid based his work on the text of the De natura animalium edited by Hercher (1864), who arbitrarily corrected several linguistic features to bring them in line with Atticist norms and introduced forms such as ὀστοῦν for ὀστέον (see Rodríguez-Noriega Guillén 2005, 457; García Valdés, Llera Fueyo, Rodríguez-Noriega Guillén 2009, XIX). In their new edition, García Valdés, Llera Fueyo, and Rodríguez-Noriega Guillén (2009) print 14 uncontracted forms and 3 contracted forms (always ὀστᾶ), reporting no manuscript variants. Three further uncontracted forms are found in other works by Aelian (ὀστᾶ 2x in VH 9.8.16 and fr. 297, ὀστῶν in VH 9.8.16), but these do not alter the impression that uncontracted forms were the norm for this author, except for an apparent preference for contraction in the nom.-acc. plur. As for Philostratus, at Im. 2.24.2 and 2.25.1 uncontracted ὀστέα and ὀστέων are transmitted, but Benndorf and Schenkl (1893) printed ὀστᾶ and ὀστῶν; at 1.24.3 ὀστῶν is transmitted without variants. On the sole occurrence of ὀστέον/ὀστοῦν in Achilles Tatius (C.4), the manuscript evidence is unfortunately inconsistent (see F.1). In Galen, who for reasons of content has numerous occurrences of the term ‘bone’, the ratio is 1975 contracted to 362 uncontracted forms, with the contracted variant outnumbering the uncontracted form in every paradigmatic form, albeit in varying proportions. Here, too, closer examination of the textual tradition would likely reveal some inconsistencies; it is interesting, for example, that the treatise De ossibus ad tirones, which includes 108 occurrences of the word, contains only one instance in which the editors print an uncontracted form (ὀστέων), which, moreover, is transmitted as contracted in another branch of the tradition (C.3).

As a parallel for the inflection of ὀστέον/ὀστοῦν, Herodian (B.1) mentioned κάνεον/κανοῦν ‘basket’, whose behaviour is, however, only partially similar. This noun is a substantivised material adjective derived from κάννακάννα ‘reed’ (a Semitic loan, cf. Akkadian qanû and see Rosół 2013, 44‒5; Mycenaean ka-ne-ja ‘of reed’ still has adjectival function); like other material adjectives, in Attic, it underwent contraction and shifted to a perispomenon accent (see entry χρύσεος, χάλκεος, φοινίκεος). The distribution, unlike that of ὀστέον/ὀστοῦν, is wholly unambiguous: this noun is regularly uncontracted in Homer (κάνεον 13x, plus κάνεια at Od. 10.355), Herodotus (1.119.16), and Hippocrates (in the variant κάνειον), while it is invariably contracted not only in Attic authors but also in prose writers of the Hellenistic and imperial periods. Its absence from the extant Atticist lexica, then, may simply be due to chance but is perhaps attributable to the fact that contemporary speakers would have had few doubts as to which was the correct Attic form, as it coincided with that still in use in the koine.

E. Byzantine and Modern Greek commentary

During the medieval and early modern periods, both the contracted and uncontracted forms of ὀστοῦν/ὀστέον may still be found (see Kriaras, LME s.v.); Modern Greek has a learned form οστό that, similarly to the adjectives, became a normal o-stem, in addition to the innovative κόκαλο, which had already developed in the Byzantine era (see LBG s.v. κόκκαλον and Kriaras, LME s.v. κόκκαλον) from the classical κόκκαλοςκόκκαλος ‘kernel’; the change in gender is likely due to the influence of ὀστοῦν. As for κάνεον/κανοῦν, it was gradually replaced by its diminutives κανίσκος, κανίσκιον (already in Ar. fr. 173.1), and κάναστρον/κάνιστρον (borrowed as Latin canistrum, whence English canister, Italian canestro ‘basket’). The lexica gloss κανοῦν with κανίσκιονκανίσκιον (Hsch. κ 626, Hsch. κ 679, Phot. κ 161, Su. κ 368); see also Poll. 6.86Poll. 6.86 τὸ δὲ νῦν κανίσκιον κανήτιον ἐκάλουν, implying that κανίσκιον was the current word in his time. During the medieval period, both κάνιστρο(ν) and κανίσκι(ν) continue to be attested, the latter also in the innovative meaning ‘gift basket’, or simply ‘gift’: Modern Greek retains κάνιστρο and the more popular variant κανίστρι in the sense ‘basket, bucket’ and κανίσκι in the sense ‘gift, present (especially wedding present)’.

F. Commentary on individual texts and occurrences

(1)    Ach.Tat. 4.19.3 (C.4)

This passage, which describes the anatomical peculiarities of the crocodile’s tail, contains the only occurrence of ὀστέον/ὀστοῦν in Achilles Tatius’ novel Leucippe and Clitophon. Unfortunately, the paradosis does not permit us to determine whether the author chose the contracted or the uncontracted form: the readings of the codd. are evenly divided between ἓν ὀστοῦν (VGE) and ἓν ὀστέον (WMD). Vilborg (1955, 86) adopted the reading ἑνὸς ὀστοῦ, found in a quotation of this passage by Eustathius of Antioch (Hex. 725.37), based on the assumption that the nominative ἓν arose by haplology of ἑνὸς ὀσ-: he then suggests, ‘but it constitutes the end of the only bone of the spine’ (Vilborg 1962, 91). Garnaud (1991, 130), meanwhile, adopts the reading ἓν ὀστέον and translates ‘mais c’est, en un seul os, l’extrémité de son épine dorsale’ (‘but it is, in a single bone, the extremity of its spine’); cf. ‘it is […] a single bone continuing and completing the vertebra’ (Whitmarsh 2001, 76, based on Garnaud’s text); ‘the continuation of its single spinebone’ (Winkler 1989, 232, based on Vilborg’s text). The various textual and exegetical difficulties of this passage are remarked and discussed by Vilborg (1962, 91‒3). Since, as noted above (D.), ὀστέον/ὀστοῦν is a word that could already appear uncontracted in Attic literature, even if Achilles Tatius had written ὀστέον, this would not contradict his general choice of a moderate Atticism (see Vilborg 1962, 12; Gammage 2019, 59).