ποῖ, ποῦ
(Phryn. Ecl. 28, Moer. π 49, [Hdn.] Philet. 271, [Hdn.] Philet. 272)

D. General commentary

Phrynichus (A.1), Moeris (A.2), and the author of the Philetaerus (A.3, A.4), later followed by Thomas Magister (B.5), proscribe the use of the interrogative adverb ποῦ (‘where?’) instead of ποῖ (‘to which place?’), specifying that the former should be used only in reference to location in a place, whereas the latter indicates motion towards a place. The same doctrine is assisted in several ancient grammatical sources, of which a selection of which is offered in section B.

At first glance, the Atticists’ position on this issue does not appear entirely consistent with the data at our disposal, since the confusion between ποῦ and ποῖ is attested here and there even in Attic authors. As will be discussed below, this apparent discrepancy can be explained in several ways – not to mention that Atticist lexica, as we read them, often present black-and-white prescriptions that fail to account for the linguistic variety observable in the works of Attic authors. For the moment, it is important to stress that the number of ancient testimonies would increase significantly if other similar phenomena were taken into account (e.g. confusion between ὅπου ‘where’ and ὅποι ‘to which place’, ἐκεῖ ‘there’ and ἐκεῖσε ‘to that place’, ἔνδον ‘within’ and εἴσω ‘into’, etc.; on the latter, see the entry ἔνδον, εἴσω). However, in keeping with the limited scope of this entry, such cases will not be treated below (the same also applies to the references to the Homeric and Ionic πῇ ‘where?’ found in section B.).

The confusion between ποῖ and ποῦ (and the progressive disappearance of the former) is part of a wider phenomenon that may be described as a tendency towards simplification within the system of local adverbs – provided that the additional information once conveyed by these adverbs can be recovered in other ways (e.g. through verbs expressing motion: see C.3, where ποῦ ὑπάγει is perfectly intelligible, although in the koine ποῦ may denote both location in a place and motion towards a place, as we will see).

Atticist lexicographers were likely not the first to address this issue. Hellenistic scholars probably commented on it as well, since the distinction between ποῖ and ποῦ had already been lost in Post-classical Greek, with ποῦ taking on the function of ποῖ. The latter only survived in Atticising authors of the imperial period and later learned literature; unsurprisingly, it is absent from the New Testament (see Blass, Debrunner 1976, 82) – and the same holds true for documentary papyri. Already Eratosthenes may have been concerned with the double meaning of ποῖ as a local adverb: according to Hdn. Περὶ παθῶν GG 3,2.295.13–9 (= Eratosth. fr. 68 Strecker), he understood the colloquial expression ποῖ κῆχος as both ‘ποῦ ἐγγύς;’ ‘where nearby?’ and ‘εἰς τίνα τόπον;’ ‘to which place?’ (see AGP vol. 1, 424). Furthermore, B.1 shows that Demetrius Ixion (2nd century BCE?) regarded ποῖ as characteristic of the Attic dialect in contrast to the Homeric-Ionic πῇ (‘to which place?’). It is unclear whether his discussion was limited to a comparison between the Attic form and the Homeric one or whether it had a broader scope; in the latter case, his statement that ποῖ is characteristic of the Attic dialect could be taken as an implicit indication that the form was foreign to the koine. Moreover, Demetrius’ teaching may have served as the starting point for later scholarly discussions on the subject – perhaps even the Atticist ones. Caution is required, however, since we do not even know whether his treatment originally belonged to his Homeric studies or to the Ἀττικαὶ λέξεις (see Staesche 1883, 47).

In summary, the Atticists’ observations did not come out of nowhere. In this regard, one may wonder whether the entries collected in A. are reworkings of earlier scholarship with a broader aim than that usually pursued by Atticists – namely, to indicate the ‘correct’ use of local adverbs and to condemning what, in ancient grammatical thought, would have been considered a solecism (i.e., a syntactic error). A hint in this direction is the fact that Phrynichus and the Philetaerus use the verb ἁμαρτάνω ‘to be mistaken’ (accordingly, Moeris’ label κοινόν here may perhaps refer simply to a ‘mistaken use’, see entry Moeris, Ἀττικιστής; a different conclusion is reached by Monaco 2024, 55, according to whom ‘the label κοινόν here may refer either to a form of lower register Greek […] or to a fairly regular form’).

A broader concern with ‘correct Greek’, rather than with Attic usage alone, may perhaps explain an apparently problematic issue raised by the entries in A. As mentioned above, ποῦ instead of ποῖ occurs sporadically in Attic authors (along with other similar phenomena: see Lobeck 1820, 43–5; K–G vol. 1, 545; Radermacher 1927–1928, 133; Fraenkel 1959, 16 n. 2; Kannicht 1969 vol. 2, 210; Dunbar 1995, 135–6), so that its proscription does not seem entirely consistent with the evidence at our disposal. However, this discrepancy can be explained in several ways. To begin with, some cases are not easily assessed: for example, in C.2, despite the presence of the verb χορεύω, the idea of motion towards a place seems to be implied (see Dodds 1944, 88). Remarkably, such usage did not go unnoticed by Atticising authors, who elsewhere are rigorous in their distinction between ποῖ and ποῦ: see Luc. DMort. 10.2: νῦν οὖν ποῦ σου τὸ θεῶν ἐκεῖνο ἡμίτομον ἀπελήλυθεν; (‘Well then, where has your divine half gone at present?’; transl. MacLeod 1961, 53); Aristid. 50.75 Keil (= 1.524.4 Dindorf): ποῖ οὖν τοῦτο ἐτελεύτησεν; (‘How was this fulfilled?’; transl. Behr 1981, 333). Additionally, scribal errors may have slightly distorted the picture, as the corruption from ποῖ to ποῦ and vice versa is easy (see the discussion of Ar. Nu. 9 in Dunbar 1995, 135–6). Therefore, it is difficult to determine the true extent of the phenomenon in Attic authors: several modern scholars are even sceptical about the possibility of genuine confusion between ποῖ and ποῦ in tragic and comic passages and are instead inclined to regard the few occurrences as scribal errors (see Diggle 1994, 186 n. 23; Dunbar 1995, 135–6; Allan 2008, 230; contra, Radermacher 1927–1928, 133; Fraenkel 1959, 16 n. 2; Kannicht 1969 vol. 2, 210). Also relevant to our discussion is the fact that several alleged cases of confusion between ποῖ and ποῦ occur in the tragedians and in Xenophon – that is, in authors not included in Phrynichus’ and Moeris’ canons (see Favi 2022, 321; entry Phrynichus Atticista, Σοφιστικὴ προπαρασκευή (Praeparatio sophistica); entry Moeris, Ἀττικιστής). Sporadic occurrences in Aristophanes or other canonical authors may not have been sufficient to make the use of ποῦ instead of ποῖ acceptable. On the other hand, there may have existed a ‘milder’ Atticist position, allowing the use of ποῦ instead of ποῖ precisely by virtue of its (few) occurrences in Attic authors, as suggested by Priscian (B.3), whose supposed Atticist source perhaps admitted both ποῖ and ποῦ with reference to motion (see Valente 2014, 63–4; Spangenberg Yanes 2017, 385–6).

In conclusion, ποῦ instead of ποῖ was likely rejected by stricter Atticists as a solecism (i.e. a syntactic error) poorly attested in canonical authors but common in Post-classical Greek. Conversely, milder Atticists may have attempted to justify the confusion between ποῖ and ποῦ precisely on the base of its occurrence in classical authors.

E. Byzantine and Modern Greek commentary

As mentioned above, ποῖ disappeared in the koine; in Byzantine times it survives only in cultivated prose and poetry (to cite just one example, it occurs 8x in Symeon Metaphrastes). By contrast, ποῦ, denoting both location and motion, is firmly established in Medieval Greek and continues into Modern Greek, where the interrogative πού retains its local sense while also developing temporal and modal function.

F. Commentary on individual texts and occurrences

N/A