υἱεύς, υἱέως, υἱέα
(Phryn. Ecl. 45, Phryn. Ecl. 234, Phryn. PS 118.3–4, Phryn. PS 118.5–6)

D. General commentary

Phrynichus’ glosses are concerned with three forms of υἱός with which he finds fault: the nominative υἱεύς, the genitive υἱέως, and the accusative υἱέα. The origin of these forms is analogicalAnalogy. The existing forms υἱέος, υἱεῖ, and υἱεῖς were treated as though they belonged to a noun in ‑εύς, and so the new forms υἱεύς, υἱέως, and υἱέα were created, following the declension type of, e.g., βασιλεύς. Phrynichus too seems to recognise that υἱεύς, υἱέως, and υἱέα are analogical (see A.2: ‘For one would not automatically find υἱέα too [just] because υἱέος and υἱεῖ exist’). When, in his discussion of υἱέως (A.1), he ascribes the use of this form to the ψευδαττικοίοἱ ψευδαττικοί (‘pseudo-Attic speakers’), this must be taken as an indication, to use modern terminology, that such forms are pseudo-learned hypercorrectionsHypercorrection (on Phrynichus’ polemics with the sophist Alexander see F.1). In rejecting υἱέα, Phrynichus quotes the earlier authority of Philoxenus (fr. 216), who in one of his treatises on Homer ostensibly discussed the occurrence of this form in a passage of the Iliad (C.1) deeming it incorrect. Although Phrynichus quotes Philoxenus’ position to justify his own, it remains to be seen how faithfully he reports Philoxenus’ views and motivations (see F.2). However, we should remark here that the Iliadic occurrence of υἱέα (C.1), and the occurrences of υἱέα in later hexametrical poetry, will require a different explanation than the occurrences of υἱέα in imperial prose writers that Phrynichus (A.2) has in mind (see F.3).

To explain how the analogical forms υἱεύς, υἱέως, and υἱέα came about, we need to provide a brief sketch of the various stem formations and declensions of υἱός in the history of Greek.

In Greek, the inherited form of the word meaning ‘son’ is υἱύς, whereas the thematic form υἱός is secondary and likely developed by dissimilation (i.e. to avoid the repetition of [u]; see DELG s.v. and EDG s.v.). The form υἱύς, though unattested in literary sources, is well-documented in inscriptions (see Bile 1988, 191 and Threatte 1996, 221 for the evidence in Cretan and Attic). υἱύς is inflected as a u-stem, though with alternations between the full-grade (*υἱεϜ-) and the zero-grade (*υἱυ-) and consequently with the generalisation of one or the other apophonic grade to the rest of the declension. So, from the full-grade (*υἱεϜ-), one has the declension υἱύς, υἱέος, υἱέι (contracted into υἱεῖ from *υἱεϜ-ι), υἱέες (contracted into υἱεῖς), and υἱέων (contracted into υἱῶν), whereas from the zero-grade (*υἱυ-/*υἱϜ-), the declension is υἱύς, υἷ(Ϝ)ος, υἷ(Ϝ)ι, υἷ(Ϝ)ες. As regards the accusative singular and plural, the forms υἷα, υἱέας, υἷας, υἱεῖς are all innovations that replaced the inherited forms υἱύν and υἱύνς (both attested in Cretan, see Bile 1988, 192 and 195–6) through the addition of the consonantal stem ending. Homeric υἱέα belongs in this discussion and must be treated on an equal footing with the early forms υἷα, υἱέας, υἷας, and υἱεῖς (see F.3).

The alternation between the various stems and declensions of υἱός is well-documented in Homer (see Chantraine 1958–1963 vol. 1, 228) and post-Homeric sources, but the thematic form υἱός progressively takes over, replacing the u-stem type. In classical Attic, the two declensions alternate in literary and documentary sources. However, the thematic υἱός rapidly gains ground from the 4th century onwards. For example, in Xenophon, Demosthenes, and Plato, the two types of declension still alternate: the u-stem forms υἱέος, υἱεῖ, and υἱεῖς are more common than the corresponding forms of the thematic declension, whereas υἱός and υἱούς are more common than the u-stem forms. In Menander, however, υἱός is already inflected as a thematic stem in all cases of the declension. Similarly, in Attic inscriptions the u-stem type basically disappears, except in metrical texts, after the mid-4th century BCE (see Meisterhans, Schwyzer 1900, 144–5 and Threatte 1996, 220–2). The situation remains unchanged in koine Greek. Polybius only uses forms of υἱός (except four instances of υἱεῖς), and Diodorus Siculus does so too (except one instance of υἱεῖς). In more informal texts, υἱός likewise predominates: it is the only form attested in the New Testament (see Blass, Debrunner 1976, § 52.1), and the third-declension forms are extremely rare in papyri and inscriptions (see Mayser, Gramm. vol. 1,2, 20–1 and Gignac 1981, 101).

Even though the u-stem type had generally disappeared, some of the third-declension forms came back into use in imperial prose. The most common were υἱέος, υἱεῖ, υἱεῖς, and υἱέων. In all likelihood, this was an archaising imitation of classical Attic prose writers like Thucydides, Plato, and Demosthenes, who mostly used υἱέος, υἱεῖ, υἱεῖς, and υἱέων over the corresponding forms of the thematic declension. The use of third-declension forms interspersed with second-declension ones is common among several Atticist prose writers (see Schmid, Atticismus vol. 4, 584), but it is also documented in authors who are not strict Atticisers. In Dionysius of Halicarnassus, there are almost as many third-declension forms as there are second-declension forms (υἱεῖ occurs 2x, υἱεῖς 4x, and υἷας 12x). In the writings of Plutarch and Josephus, the forms of the u-stem type are not nearly as common as those of the thematic declension, but are still attested in sizeable quantities (Plutarch: υἱέος 2x, υἱεῖ 1x, υἱεῖς 8x, υἱέων 6x, υἱέσι(ν) 4x, υἱέας 1x, and υἷας 1x; Josephus: υἱέος 6x, υἱεῖ 5x, υἱεῖς 21x, υἱέων 4x, υἱέσι(ν) 7x, υἱέας 1x, and υἷας 1x.

The development of the analogical forms υἱεύς, υἱέως, and υἱέα originates in this context. In trying to recover old Attic forms that had fallen out of use, the forms of the u-stem type υἱέος, υἱεῖ, υἱεῖς, and υἱέων were re-interpreted as belonging to a noun in ‑εύς like βασιλεύς. Thus, the new forms υἱεύς, υἱέως, and υἱέα were created in accordance with this declension type (for the parallel development of υἱῆος, υἱῆι, etc. see below). υἱεύς, υἱέως, and υἱέα are therefore pseudo-learned hyper-corrections, as Phrynichus rightly recognised (A.1, A.2). This status also explains why none of these forms is ever attested in documentary papyri and inscriptions. Interestingly, these forms are also occasionally attested as variant readings in the manuscripts of classical authors (see Lobeck 1820, 68–9 and Rutherford 1881, 142).

The distribution of υἱεύς, υἱέως, and υἱέα in ancient texts is uneven.

The analogical nominative υἱεύς is only attested in grammatical writings (B.1, see also Et.Gud. 540.57–9Et.Gud. 540.57–9, Eust. in Il. 2.124.4–7Eust. in Il. 2.124.4–7 ~ in Il. 2.263.21–264.3Eust. in Il. 2.263.21–264.3, Eust. in Il. 3.22.12–4Eust. in Il. 3.22.12–4).

Though regarded as the prototypical form and the original nominative of all the third-declension forms of υἱός with the full-grade in the stem (i.e. υἱέος, υἱεῖ, etc.), it was also deemed to have always remained unused (see B.2). It is likely that this form was created merely to provide a nominative for the declension to which these third-declension forms belong. This was due to the fact that υἱύς, which we know is the original nominative of the u-stem type, is unattested in literary sources and only occurs in inscriptions (see above).

The evidence for υἱέως in pre-Byzantine writers is limited to one occurrence in Cassius Dio (C.5) and three in Galen. The genitive υἱέως was perhaps a less obvious development, not least because there is substantial classical evidence for the genitive υἱέος. In fact, grammatical sources claim that υἱέος is exceptional in the declension of υἱεύς (see B.1, Choerob. in Theodos. GG 4,1.215.28–30Choerob. in Theodos. GG 4,1.215.28–30, EM 775.17–8EM 775.17–8), which is reminiscent of the doctrine according to which the dative plural should have been υἱεῦσι (cf. βασιλεῦσι). Surprisingly, this form too is replaced by υἱέσι (B.1, see also Choerob. in Theodos. GG 4,1.139.4–6, Eust. in Il. 2.124.4–7 ~ in Il. 2.263.21–264.3).

Finally, the accusative υἱέα is the best attested of these three analogical forms. It occurs in Arrian (C.2), Dio Chrysostom (C.3, C.4), Galen, in one of the Socratic epistles (C.6), Julian (2x), and Themistius (C.7 and in Ὑπατικὸς εἰς τὸν αὐτοκράτορα Ἰοβιανόν 70d.3). Especially noteworthy are the occurrences of υἱέα in Atticising writers like Dio Chrysostom and Themistius.

To conclude, it is worth observing that the re-interpretation of υἱέος, υἱεῖ, υἱεῖς, and υἱέων as forms belonging to a noun in ‑εύς (and consequently the analogical development of υἱεύς, υἱέως, and υἱέα) is not at all an isolated development in the history of Greek. Indeed, the very same situation is paralleled in the formation of the pseudo-epic forms υἱῆος, υἱῆι, υἱῆα, υἱῆες, υἱήων, υἱήεσσι(ν), and υἱήας, which are continuously attested in hexametrical poetry, from Apollonius Rhodius to the writers of late antiquity. These forms too are created in imitation of the declension of the nouns in ‑εύς, which in epic poetry do not undergo quantitative metathesis (one may compare Ἀχιλῆος, βασιλῆος, etc.).

E. Byzantine and Modern Greek commentary

Unlike the examples found in classical sources, the nominative υἱεύς is occasionally attested in late Byzantine poetry – once in Manuel Philes (Carmina 5.7.81) and three times by Theodorus Metochites (1.109, 1.238, 1.260). In such cases, υἱεύς is simply a metrically useful alternative to υἱός, a feat made possible by the continuing notion that υἱεύς and υἱός are equally correct (see Gennadius Scholarius Grammatica 2.444.27Gennadius Scholarius Grammatica 2.444.27 υἱεὺς δὲ καὶ υἱός, τὸ αὐτό ‘υἱεύς and υἱός: [they are] the same’).

The genitive υἱέως is continuously attested in late antique (1x in Basilius of Caesarea, 5x in Theodoretus), middle Byzantine (1x in Constantinus VII), and late Byzantine writers (2x in Nicetas Choniates, 1x in Georgius Acropolites, 15x in Nicephorus Gregoras, 1x in Macarius Macres, and 3x in the letters of Michael Apostolius).

The accusative υἱέα is well-attested in Byzantine prose. The authors who use it more frequently include Sozomenus (x2), Theodorus Studites (x4), Tzetzes (3x in his prose writings), John Cinnamus (x4), Nicetas Choniates (x4), and Nicephorus Gregoras (x4). The innumerable occurrences of υἱέα in Byzantine poetry should be treated alongside those in Hellenistic and imperial poetry (see F.3).

F. Commentary on individual texts and occurrences

(1)    Phryn. Ecl. 234 (= Philox.Gramm. fr. 216) (A.2)

The sophist Alexander who, according to Phrynichus, used υἱέα in one of his letters, is Alexander of SeleuciaAlexander of Seleucia. Phrynichus also criticises him in Ecl. 324Phryn. Ecl. 324 for using the optative δῴη and διδῴη in place of δοίη and διδοίη. Philostratus devotes a detailed biography to Alexander (VS 2.5), which provides us with a certain amount of information for illuminating the nature (and possibly the aims) of Phrynichus’ hostility towards him. It seems that in condemning Alexander’s linguistic mistakes, Phrynichus aims to attack more deeply the cultural milieu and the sophistic circles to which Alexander belonged.

Alexander was initially a pupil of Aelius Dionysius, but he was forced to leave his school for family reasons and later became a close disciple of FavorinusFavorinus of Arles (VS 2.5.576.5–10). Favorinus is notoriously one of the constant polemical targets in the Eclogue. Phrynichus condemns him as a non-Greek who, though lacking the necessary linguistic competence, is unreasonably regarded as a model of Greek speaking. Thus, Alexander’s education in Favorinus’ circle makes him highly suspect. In addition, Philostratus says that Alexander spent most of his life in Antiochia, Rome, Tarsus, and Egypt, and only spent a little time in Athens (see VS 2.5.571.9–13); he died while living in Gaul or Italy (VS 2.5.576.8–10). Regarding his weak connection with the Greek world, Philostratus openly comments that Alexander was not popular in Greece (see VS 2.5.574.16–7). As in Favorinus’ case, Alexander’s stronger affiliation with the cultural circles of the Western part of the empire rather than with Greece may have been a further motivation for Phrynichus’ hostility.

Philostratus’ extended quotations from Alexander’s speeches demonstrate sufficiently the sophist’s markedly Asianic styleAsianic style (see VS 2.5.574.27–33; short cola, as in VS 2.5.574.22–7). Philostratus also repeatedly describes Alexander’s diction and style as pathetic and solemn (see VS 2.5.572.17–9, VS 2.5.574.19–21). This was mildly parodied by Herodes AtticusHerodes Atticus (see VS 2.5.573.25–30 and 2.5.574.3–6, see Russell 1983, 86: ‘Presumably he [i.e. Herodes] wanted to show that this more exuberant style [i.e. Alexander’s] was really quite easy’) and openly derided by Herodes’ disciples (see VS 2.5.573.17–9). Moreover, Alexander delights in embellishing his speech with tragic quotations and tragic allusions (see VS 2.5.571.17 and [Aesch.] PV 32; in VS 2.5.574.6-8 he alludes to Aeschylus’ description of his tragedies compared to the Homeric poems as reported by Ath. 8.347e, see further Civiletti 2002, 550 n. 28). All these features of Alexander’s style are clearly incompatible with Phrynichus’ preferences.

Finally, we know from Philostratus that Alexander was appointed secretary ab epistulis Graecis (see VS 2.5.571.15–6). Bowersock (1969, 53–4) dates the beginning of his time in office to between the years 169 and 172. Since the dedicatee of Phrynichus’ Eclogue, Cornelianus, was also appointed secretary ab epistulis Graecis (see Ecl. 357Phryn. Ecl. 357), it may be that Phrynichus, in criticising Alexander’s linguistic ineptitude and his lack of the necessary skills, indirectly praises Cornelianus as being more qualified to hold the office of imperial secretary.

(2)    Philox.Gramm. fr. 216 (= Phryn. Ecl. 234) (A.2)

According to Phrynichus, Philoxenus discussed the form υἱέα ἐν τοῖς εʹ Περὶ τῆς Ἰλιάδος συγγράμμασι (‘In the five treatises On the Iliad’). This formulation is evidently corrupt. One possibility is that the text here should be ἐν τῷ εʹ Περὶ τῆς Ἰλιάδος συγγράμματι (‘in the fifth book of the treatise On the Iliad’), but certainty is unattainable. At any rate, we have no other evidence for the existence of a work by Philoxenus with this title. Thus, earlier scholarship (the bibliography is collected by Theodoridis 1976, 9 and n. 9) suggests that this is either a reference to Philoxenus’ treatise On the signs in the Iliad (see also Theodoridis 1976, 9 and 187–8, who agrees with this interpretation) or to On the glosses in Homer (see Theodoridis 1976, 11).

Considering Phrynichus’ passage, it seems that Philoxenus condemned the Homeric occurrence of υἱέα as morphologically incorrect. The formulation τοῦτο δὲ καὶ Φιλόξενος […] δαψιλέστατα ἀπέφηνεν, ἀδόκιμον μὲν εἶναι τὸν υἱέα, δόκιμον δὲ τὸν υἱόν (‘Philoxenus too showed this extensively […], [explaining] that υἱέα is unapproved, whereas υἱόν is’) would seem like a report of Philoxenus’ main conclusion. However, we have no confirmation that this inference is correct, and it might also be the case that Philoxenus simply remarked that υἱέα in Il. 13.350 is unparalleled in archaic poetry, without necessarily questioning its correctness or even the authenticity of the line. In such a case, the use of the adjective δόκιμον, the typical vocabulary of Atticist lexicography, might be Phrynichus’ way of adjusting Philoxenus’ original wording to his own ends.

The Homeric line in which υἱέα occurs is the object of criticism in the scholia (B.3, see also schol. [ex.] Hom. Il. 13.348–50 [T])Schol. (ex.) Hom. Il. 13.348–50. The alleged reason is the repetition of the same concept expressed two lines above (i.e. Zeus giving honour to Thetis). No concrete evidence can be adduced that the accusative υἱέα was also the object of criticism in Homeric scholarship, though this possibility should remain open (see Erbse ad loc.: ‘nescio, num sch. de forma singulari υἱέα umquam fuerit (fort. etiam ab Aristonico commemorata?)’).

(3)    Hom. Il. 13.347–50 (C.1)

This passage preserves the only occurrence of υἱέα in Homeric poetry (and in the whole of pre-Hellenistic poetry). The Homeric υἱέα should not be mistaken for the homophonous form υἱέα, which appears in Imperial prose. The Homeric υἱέα is created from the full-grade of the root (*υἱεϜ-) with the addition of the consonantal stem ending. In this way, υἱέα functions like υἱέας (which occurs in Hom. Il. 2.693, 5.149, 11.123, 23.175, 23.181, 24.205, 24.521, Od. 4.211, Hes. frr. 35.6 and 49 M.–W.). The rarity of υἱέα in Homeric poetry (and in epic poetry in general) is noteworthy, but not entirely surprising. As is rightly stressed by Schwyzer (1939, 573–4) and Chantraine (1958–1963 vol. 1, 228), the forms created on the full-grade (*υἱεϜ-) are rarer than those built on the zero-grade (and much rarer than those of the thematic declension). Regarding the accusatives, in Homeric poetry the ratio υἱέα : υἷα and υἱέας : υἷας is 1 : 7 and 8 : 36. In imitation of the Homeric passage, υἱέα began to be used more often in Hellenistic, Imperial, and Byzantine hexametrical poetry (Callimachus, Apollonius Rhodius, Moschus, Quintus Smyrnaeus, Gregory of Nazianzus, Nonnus, etc.).