ἀρχῆθεν
(Phryn. Ecl. 66, Phryn. PS 9.9–11, Antiatt. α 138, Σ^b α 2201)

D. General commentary

Atticist lexicographers disputed the acceptability of ἀρχῆθεν. This form is very common in post-classical literature (see below). However, given that ἀρχῆθεν is unattested in classical Attic prose while occurring in tragedy (albeit rarely) and Herodotus, rigorous Atticists, such as Phrynichus (A.1, A.2) and possibly also Orus (A.4, though see F.1), proscribe such a form. By contrast, more open-minded lexicographers – typically represented by the Antiatticist (A.3) – are inclined toward the view that the form’s attestations in Herodotus are sufficient to support its approval.

Adverbs ending in -θεν are characteristically ablativeAblative in function (i.e., they indicate the origin of a movement), and so the adverbial suffix can be attached to nouns (e.g., οὐρανόθεν ‘from heaven’), pronouns (e.g., ἄλλοθεν ‘from another place’), and prepositions (e.g., πρόσθεν ‘before’) (the standard treatment of these formations is that by Lejeune 1939; on the accentuation see Ringe 1977). In archaic and classical texts, these adverbs are typically poetic and are amply employed in Homeric epic (see Chantraine 1958–1963 vol. 1, 241–3). In classical Attic, they enjoy more limited use but are by no means unattested. As Lejeune (1939, 149–51) demonstrated, they are largely confined to poetry, but some fixed forms also occur in prose. To give a concrete example, the adverb ἀγρόθενἀγρόθεν (which occurs twice in the Odyssey, at 13.268 and 15.428) occurs only twice in Attic texts – once in Euripides (Or. 866) and once in the comic poet Alcaeus (fr. 19.1)Alc.Com. fr. 19.1 – while the standard Attic expression is ἐξ ἀγροῦ. However, since Alcaeus is parodying Euripides’ line, the use of ἀγρόθεν functions as an element of parodicParody poetic vocabulary that is used to mark the difference with ordinary speech (see further Orth 2013, 92). In some cases, however, the situation is reversed. For example, while ἕωθενἕωθεν is unattested in tragedy, it occurs several times in comedy (18x) and more rarely in 4th-century prose (1x in Demosthenes and 1x in Aeschines but never in Lysias or Isocrates, for instance) (see Lejeune 1939, 172–3). In Attic inscriptions, the adverbial suffix -θεν is common (even standard, in some cases) in toponyms (Ἀθήνηθεν, Ἐλευσίνοθεν, etc.) and prepositions (ἔξωθεν, πρόσθεν, etc.), but no examples with common nouns are attested to date (see Threatte 1996, 397–404). This general situation changes in post-classical times, in that adverbs ending in -θεν that are derived from common nouns become increasingly ubiquitous. Lobeck (1820, 93–4) conveniently lists several such forms, focusing on those that occur in prose during the Hellenistic and – particularly – the imperial periods (see also the interesting observations by Lejeune 1939, 171 regarding παιδ(ι)όθεν in Septuagint Greek and documentary sources).

The case of ἀρχῆθεν wholly aligns with these tendencies. It is rendered less conspicuous by the relative rarity of temporal adverbs in -θεν (see Lejeune 1939, 170), and this is the only such form that is continuously attested from the late archaic to the Byzantine period (for an overview, see also Lejeune 1939, 170–1). In archaic and classical literature, ἀρχῆθεν is attested only in Pindar (ἀρχᾶθεν in O. 9.55, in I. 3/4.25), Aeschylus (C.1), and Sophocles (fr. 126.3), while in prose, it occurs only in Herodotus (C.2, plus six further instances) and once in the corpus Hippocraticum. In post-classical prose, ἀρχῆθεν is well-attested in the high koine (Polybius: 4x; LXX: 1x, 3Ma. 3.21 which is written in high koine; Dionysius of Halicarnassus: 4x; Josephus: 5x; Plutarch: 3x; Galen: 4x; Sextus Empiricus 23x) and in documentary sources: IG 12,9.4.6 [Euboea, 1st century BCE], IOSPE I².404.23 [Chersonesus, 185/186 CE], P.Oxy. 66.4526.6 (= TM 78598) [69/70 CE], P.Lund. 3.10.24 (= TM 63054) [provenance unknown, 98 CE], P.Oxy. 22.2341.10 (= TM 22213) [208 CE], SB 14.11899.12 (= TM 26548) [Oxyrhynchus, 2nd century CE]. Similarly, ἀρχῆθεν is common or very common in Christian writers (Theophilus, Tatianus, Clemens of Alexandria, Origen, Eusebius, Athanasius, Basilius, Procopius of Gaza, etc.). In contrast to this increasing diffusion, however, ἀρχῆθεν occurs only occasionally in Atticist writers (Appian 1x; Dio Chrysostom: C.3 plus two more instances; Libanius: 2x; see Schmid, Atticismus vol. 1, 149). The only known instance of ἀρχῆθεν in post-classical poetry occurs in the early Hellenistic writer Rhianus (CA 13.2). Two passages from Eusebius and one from Philo of Carpasia attest the reinforced syntagm ἐξ ἀρχῆθεν (Eus. Demonstr. Evang. 8.1.33 and Comm. in Psalm. MPG 23.905.2, Philo Carpasius Enarratio in Canticum Canticorum 117.8).

E. Byzantine and Modern Greek commentary

Throughout the Byzantine age, the adverb ἀρχῆθεν is well attested across all genres and in various registers (for an early instance in high-level prose, see, e.g. Procop. Vand. 3.16). Compared to the hundreds of occurrences of ἀρχῆθεν known from the 11th to the 13th centuries, during this phase the form’s occurrence is not particularly common (although not wholly unattested) in the writers who use the high language (Photius: 3x; Arethas: 5x; Michael Psellus: 14x; Anna Comnene: C.4 plus 3 further instances; Nicephorus Basilaces: 8x; Michael Choniates: 1x; Nicephorus Blemmydes: 4x; Nicetas Choniates: 6x). It becomes more ubiquitous in Palaeologan writers (including George Pachymeres: 63x; Theodore Metochites: 11x; Nicephorus Gregoras: 24x). ἀρχῆθεν does not survive in Modern Greek with the exception of certain dialects (see ILNE s.v.).

F. Commentary on individual texts and occurrences

(1)    Σ^b α 2201 (= Phot. α 2928, ex Σʹʹʹ = Orus fr. B 46 = Phryn. PS fr. 264) (A.4)

De Borries previously considered this lemma to be a fragment of Phrynichus’ PS, which he printed as PS fr. 264. However, Alpers (1981, 67 n. 64) refuted this interpretation based on the following argumentation. In the entry the PS (A.2), Phrynichus does not proscribe the use of ἀρχῆθεν, which is a tragic word used by Aeschylus and Sophocles (see D.), both of whom Phrynichus treats in the PS as approved models of Attic language. However, the lemma in the Synagoge is openly critical of Aeschylus’ use of this form, making its derivation from the PS unlikely. It is generally assumed, however, that Orus did not admit Aeschylus and Sophocles in his canon of approved Attic authors, since these authors are never mentioned in the fragments which are normally ascribed to his lexicon. Accordingly, Alpers infers that the entry in the Synagoge is more likely to derive from Orus’ lexicon than from Phrynichus’ PS. Alpers’ argumentation is sound, though we should at least admit the possibility that the lemma does in fact derive from one of Phrynichus’ lexica. First, although Phrynichus includes Aeschylus and Sophocles in the canon of approved Attic authors (see, e.g., PS 125.12–4Phryn. PS 125.12–4), it does not necessarily follow that he will also admit those forms that he regards as foreign to Attic. In this regard, ἀρχῆθεν may be particularly suspect because adverbs ending in -θεν as a whole are largely foreign to Attic (see D.), making it difficult to approve of such a form. Moreover, it is difficult to agree with Alpers’ assertion that the formulation παρὰ μὲν ταῖς ἄλλαις διαλέκτοις εὑρίσκεται, Ἀττικοῖς δὲ οὐ φίλον (‘[This form] is found in the other dialects, but [it is] not loved by Attic [writers]’) in the entry of the PS (A.2) should not be regarded as a proscription, particularly in the light of the subsequent remark about the absence of ἀρχῆθεν in Plato and Thucydides. Thus, Ἀττικοῖς δὲ οὐ φίλον may well indicate that ἀρχῆθεν cannot be approved based solely on its (admittedly rare) occurrence in tragedy. Finally, Alpers assumes that if Phrynichus is the author behind the entry in the Synagoge, then the specific source can only be the PS (see Fischer 1974, 48 and Alpers 1981, 64). However, we cannot firmly exclude the possibility that the entry in the Synagoge derives from the Eclogue, in which Phrynichus openly rejects ἀρχῆθεν (A.1). In considering this option, one must be open to the possibility that the surviving plenior redaction of this lexicon is actually an epitomeEpitome rather than the complete text (see Tribulato 2022) and that the Eclogue did not entirely disappear in the pre-Palaeologan age (on its circulation in pre-Palaeologan Byzantium, see also Alpers 2013, 146–8).