PURA. Purism In Antiquity: Theories Of Language in Greek Atticist Lexica and their Legacy

Lexicographic entries

ἀνατοιχέω, διατοιχέω
(Phryn. PS 62.1–3, Phryn. Ecl. 132, Antiatt. δ 28, Poll. 1.114)

A. Main sources

(1) Phryn. PS 62.1–3: διατοιχεῖν· τὸ εἰς τὸν ἕτερον τοῖχον τῆς νεὼς διαβαίνειν ἐν τῷ πλῷ, ὅπερ οἱ ἰδιῶται ἀνατοιχεῖν λέγουσιν.

ἐν τῷ πλῷ de Borries : πλοίῳ cod. : ἐν τῷ πλοίῳ Lobeck | ἀνατοιχεῖν Lobeck : ἀντιτοιχεῖν cod.

διατοιχεῖν: To move to the other side of the ship while sailing; [an action] which laymen call ἀνατοιχεῖν.

(2) Phryn. Ecl. 132: ἀνατοιχεῖν μὴ λέγε, ἀλλὰ διατοιχεῖν.

Do not use διατοιχεῖν, but ἀνατοιχεῖν.

(3) Antiatt. δ 28: διατοιχεῖν· ἀντὶ τοῦ ἀνατοιχεῖν. Εὔβουλος Κατακολλωμένῳ.

διατοιχεῖν: Meaning ἀνατοιχεῖν. Eubulus in the Katakollοmenos (fr. 50 = 51 Hunter = C.1).

(4) Poll. 1.114: ἐρεῖς δὲ καὶ ‘διατοιχεῖν ἔδει’. τὸ γὰρ ἀνατοιχεῖν ἰδιωτικόν.

διατοιχεῖν most manuscripts : διατυχεῖν AVC | ἔδει MAV : omitted in FSBC | ἀνατοιχεῖν most manuscripts : ἀνατυχεῖν AV.

[For expressions concerning sailing] you will also use ‘it was necessary to move to the other side’; for ἀνατοιχεῖν is unsophisticated.

B. Other erudite sources

(1) Phot. δ 470 (= Su. δ 799, [Zonar.] δ 534): διατοιχεῖ· ἀντὶ τοῦ ποτὲ μὲν ἐπὶ τοῦ δεξιοῦ τοίχου τοῦ πλοίου, ποτὲ δὲ ἐπὶ τοῦ ἀριστεροῦ γίνεται.

διατοιχεῖ: Meaning ‘he is at one time on the right side of the ship, at another on the left’.

(2) Phot. ε 1344: ἐπαμφοτερίζειν· τὸ πρὸς ἑκατέραν τὴν στάσιν ἀποκλίνειν καὶ διατοιχεῖν.

ἐπαμφοτερίζειν: To incline towards both sides and to change sides.

C. Loci classici, other relevant texts

(1) Eub. fr. 50 (= fr. 51 Hunter) = Antiatt. δ 28 re. διατοιχεῖν (A.3) (see D.).

(2) Aristid. 14.22.8–10 Lenz–Behr (= 36.691.14–7 Dindorf): […] ἀλλὰ δεῖ τάξιν ἐξ ἀρχῆς ἑλομένους ἀντέχεσθαι τῆς ῥοπῆς, καὶ μὴ τὸ τῶν πλεόντων μεταστρέψαι πρὸς τὸν ἐλάττω διατοιχοῦντας ἀεί.

[…] but after choosing a position, we must hold fast to the shift in balance, and not in the manner of sailors, always be changing sides and shifting to the lighter one. (Transl. Behr 1986, 392).

(3) Arr. Epict.–2: ἑτεροκλινῶς ἔχω πρὸς ἡδονήν· ἀνατοιχήσω ἐπὶ τὸ ἐναντίον ὑπὲρ τὸ μέτρον τῆς ἀσκήσεως ἕνεκα.

ἀνατοιχήσω Schweighäuser : ἀνισοτοιχήσω Wolf : ἂν … ἀ….ήσω S : ἂν ἀτυχήσω s : ἀναχωρήσω S (correction by a more recent hand).

I am inclined to pleasure: I will throw myself beyond measure in the opposite direction, for the sake of training (Transl. Hard 2014, 472).

(4) Simp. in Epict. 38.710–2: τεκμήριον δὲ καὶ ἀποτέλεσμα τῆς ὄντως μεταμελείας τὸ μηκέτι τὰ αὐτὰ ἁμαρτεῖν μηδὲ τὰ ἐκείνων ἐλάττονα. Δεῖ γὰρ ἀνισοτοιχοῦντα, ὥσπερ οἱ πλέοντες, ἐπὶ τὸ ἐναντίον μεθίστασθαι.

ἀνισοτοιχοῦντα BCG (the latter from conjecture) : ἀνισοτειχοῦντα A : ἀνισοτειχοῦντες HΣ : ἀνισοτoιχοῦντας J : the end of the verb is abbreviated in DEF.

The proof and fruit of true repentance lies in no longer making the same mistakes, even the less serious ones. For one must change position to the opposite side, as sailors do, when there is an imbalance.

D. General commentary

διατοιχέω ‘to move to the other side of the ship’ is a very rare verb in Greek: apart from lexicographical sources, it is only securely attested in Aelius Aristides (C.2). Phrynichus (A.1, A.2) and Pollux (A.4) are adamant in favouring this form against its synonym ἀνατοιχέω, which was considered to be typical of the language of the ἰδιῶταιοἱ ἰδιῶται (see Matthaios 2013, 108); the Antiatticist (A.3), in contrast, probably defended it (see here, below). ἀνατοιχέω is never attested outside Atticist lexica: its only possible literary occurrence – in Arrianus, quoting Epictetus (C.3) – is a modern conjecture (for a corrupt reading in the main manuscript of the Dissertations: see F.3). There is a further ghost-attestation in Ar. fr. 605 Kock, quoted by Poll. 4.114; the correct reading of this form is ἀνατειχίσας, reported by Pollux’ cod. F, and as printed by Kassel and Austin in Ar. fr. 657. That διατοιχέω and ἀνατοιχέω were rarely used is confirmed by the frequent corruptions in all the testimonia of the lexicographical passages, which usually confuse these words with the aorist of compound forms of τυγχάνω (see apparatuses of Phryn. Ecl. 132, A.2, Antiatt. δ 20, A.3; Poll. 1.114, A.4; see also C.3; at Phryn. PS 62, A.1, ἀνατοιχέω is transmitted as ἀντιτοιχέω, not attested elsewhere: see Lobeck 1820, 161).

διατοιχέω and ἀνατοιχέω are denominative verbsDenominative verbs derived from prepositional governing compounds (on this category, see Risch 1974, 187). The prepositional first element governs the nominal second element, in this case the noun τοῖχος, in its specialised meaning ‘side of a ship’ (LSJ9 s.v., 2). The adjectival compound διάτοιχος, formally the base of διατοιχέω, is only attested in the meaning ‘extending through the width of the wall’, as in IG 22.463.57, 4th c. BCEIG 22.463.57]. There is also the substantivised noun διάτοιχοςδιάτοιχος ‘bounding stone’ (IG 11,2.144a.85IG 11,2.144a.85, 4th c. BCE). Both meanings are confirmed by Hsch. δ 1388Hsch. δ 1388: διάτοιχοι· οἱ ἀπὸ τοίχου ἐπὶ τοῖχον ‘those [stones?] which extend from [one] wall to [another] wall’. Another compound verb in which the second member ‑τοιχέω has a similar meaning is ἀνισοτοιχέω ‘to be out of balance’, a hapax attested in Simplicius’ commentary on Epictetus’ manual (C.4; see further below and F.4). Adjectival compounds in ἀνισο‑ are common in Post-classical Greek, but the denominative verbs derived from them amount to only ἀνισοπαχέω ‘to be of unequal thickness’ (Hero, 2nd–1st c. BCE), ἀνισοκρατέω ‘to be unequal in strength’ (Sextus Empiricus, 2nd c. CE), and ἀνισοτοιχέω.

The formation pattern of διατοιχέω deserves further scrutiny. Semantically, at least judging from the attested meanings of the verb, the preverb δια‑διά expresses the notion of a movement through the ship, which functions as the ‘landmark’ (cf. LSJ⁹ s.v. διά, 1 and 2). This is what Luraghi (2012, 384) defines ‘a unidirectional trajectory through the interior of a landmark’ (see too Luraghi 2012, 367–8 and, for its expression in prefixed verbs, Schwyzer, Debrunner 1950, 450). Given this meaning, one would expect the compound to have a word like ναῦς as the second member. Instead, at least in theory, the second member τοῖχοςτοῖχος implies something that stretches across a wall (as in διάτοιχος discussed above). The same semantic problem concerns ἀνατοιχέω. Here the preverb ἀνα‑ emphasises either an up and down motion (cf. LSJ⁹ s.v. ἀνά I.2), or an upward movement tout court. Again, one would have expected the second member of the compound to be ναῦς, not τοῖχος. Be that as it may, the synonym ἀνατοιχέω was probably created to signify that, when the ship is unbalanced, one moves to the side that is higher up, which is also the lighter one (for this image, cf. C.2: πρὸς τὸν ἐλάττω διατοιχοῦντας).

One might speculate that διατοιχέω originated as a verb for masonry (as its connection with διάτοιχος would suggest) and was later extended to nautical contexts. In its metaphorical application, διατοιχέω refers to changing one’s mind according to convenience, and specifically to a change in political allegiances: see the occurrence in Aelius Aristides (C.2), where διατοιχέω is used to signify that the Athenians should not betray the Thebans, and Phot. ε 1344 (B.2), where it is employed as an explanation of ἐπαμφοτερίζω ‘to play a double game’. These meanings stem from the metaphorical use of τοῖχος to refer to a political position, illustrated e.g. in Ar. Ra. 533–8:Ar. Ra. 533–8 ταῦτα μὲν πρὸς ἀνδρός ἐστι | νοῦν ἔχοντος καὶ φρένας καὶ | πολλὰ περιπεπλευκότος, | μετακυλίνδειν αὑτὸν ἀεὶ | πρὸς τὸν εὖ πράττοντα τοῖχον | μᾶλλον ἢ γεγραμμένην | εἰκόν᾽ ἑστάναι, λαβόνθ᾽ ἓν | σχῆμα: ‘The mark of a man with brains and sense, one who’s voyaged far and wide, is ever to shift to the safe side of the ship and not just stand fast in one position like a painted picture’ (Transl. adapted from Henderson 2002, 97). As for ἀνατοιχέω, its only likely occurrence in a literary text – if the conjecture proposed for Arrian’s account of Epictetus’ lectures (C.3) is correct – confirms that the verb implied counterbalancing an unsafe situation (in this context, the tendency to pleasure) by shifting to a safer side.

Phrynichus’ and Pollux’ dislike of ἀνατοιχέω and approval of διατοιχέω are not substantiated by quotations from classical or contemporary authors (see further F.2 on Pollux). Their reason for preferring διατοιχέω and classifying ἀνατοιχέω as ‘unsophisticated’ (ἰδιωτικόνἰδιωτικός) can only be guessed. The likely occurrence of ἀνατοιχέω in Arrian/Epictetus may confirm its lower register: the style of Epictetus’ lectures was apparently characterised by colloquial expressions, appropriate for the communicative purposes of his teaching (see Zangrando 1998). The Atticist pedigree of διατοιχέω is confirmed by Aristides’Aelius Aristides usage (C.2). Aristides’ example may have influenced the lexicographers’ approval, since he was one of the few contemporary sophists whom Phrynichus admired, as he apparently told his dedicatee Menodorus in the prefatory letter of the eleventh book of the Praeparatio sophistica (see Phot. Bibl. 158). Recent scholarship has made a strong case for their encounter at Pergamum, perhaps at Aristocles’ lectures (Jones 2008, 254–61; Berardi 2016, 249–51). Aristides could perhaps have derived διατοιχέω from a lost classical text; the only preserved reference in lexicography, in the Antiatticist (A.3), brings the playwright Eubulus (Middle Comedy) into the picture, but it is impossible to ascertain in what context he employed διατοιχέω.

The entry in the Antiatticist (A.3) is ambiguous and deserves further scrutiny. At face value, it may seem that the lemma is devoted to confirming the Attic pedigree of διατοιχέω through Eubulus. The Antiatticist quotes Eubulus no less than nine times to defend Hellenistic usages, but the comic playwright was not among Phrynichus’ approved authors (cf. Ecl. 315). Hunter (1983, 138–9) deems this contradictory, expecting the Antiatticist ‘to adduce classical evidence for the form ἀνατοιχεῖν, and it is the latter form which Eubulus is likely to have used’. He therefore proposes to reverse the order of the two verbs in the Antiatticist entry (cf. also PCG ad Eubul. fr. 50 and Matthaios 2013, 108 n. 188). Hunter’s interpretation is probably correct, but there is no need to change the transmitted text of the Antiatticist. In this lexicon the expression ἀντὶ (τοῦ)ἀντὶ τοῦ usually introduces the Attic form, or the form that other sources may consider more correct and for which the compiler of the Antiatticist provides a synonym which he considers to be equally acceptable. One example is Antiatt. α 26Antiatt. α 26: ἀγοράσαι· ἀντὶ τοῦ ὠνήσασθαι, οὐκ ἐπὶ τοῦ ἐν ἀγορᾷ διατρίβειν, where the more recent meaning of ἀγοράζω ‘to buy’ (= ὠνήσασθαι) is further contrasted with its classical meaning ‘to frequent the ἀγορά’. Following this example, Hunter would be right in expecting ἀνατοιχεῖν to be the lemma and διατοιχεῖν to be its classical counterpart. However, the preserved order may still be defended by assuming that διατοιχεῖν· ἀντὶ τοῦ ἀνατοιχεῖν is a drastically abridged version of a longer sentence, something which could have sounded like διατοιχεῖν ἀξιοῦσι λέγειν ἀντὶ τοῦ ἀνατοιχεῖν· ἀλλ’ Εὔβουλoς ἀνατοιχεῖν λέγει ἐν τῷ Κατακολλωμένῳ (‘[the Atticists] recommend using διατοιχεῖν instead of ἀνατοιχεῖν, but the latter is used by Eubulus in the Katakollomenos’; cf. Valente 2015, 47 for entries with this structure and Antiatt. γ 6). Though not conclusive, this argument warns against amending the order of the transmitted Antiatticist text.

The Atticists’ criticism of ἀνατοιχέω may not have stemmed only from its absence from canonical texts. The formation pattern of the verb itself and its semantics might also have been felt to be odd. In ἀνατοιχέω, the preverb not only has a prepositional governing function, but it also retains its original spatial function (‘up’: see above), which ἀνάἀνά, an infrequent preposition in Attic, had already lost in Hellenistic Greek (an exception being the expression ἀνὰ μέσον; see Bortone 2010, 185, 227 on the evolution of this preposition in Post-classical Greek). Such semantic evolution is mirrored in compounds. In verbs first attested in Hellenistic Greek, the prefix ἀνα‑ἀνά has a distributive meaning (see Mayser, Gramm. vol. 1,3, 207; 238–9); their number, however, is low compared to other prefixed verbs, confirming that ἀνάἀνά underwent an increasing loss of productivity both as a preposition and as a preverb. One could therefore speculate that ἀνατοιχέω, seemingly a rare verb only used in nautical contexts, was felt to be less semantically transparent for speakers of Post-classical Greek, where διάδιά and the verbs prefixed with it instead remained in common usage (see Bortone 2010, 158 with n. 78; 190; 228; Mayser, Gramm. vol. 1,3, 211–4; 238–9).

E. Byzantine and Modern Greek commentary


F. Commentary on individual texts and occurrences

(1)    Phot. δ 470 = Su. δ 799 (B.1)

The agreement between Photius and the Suda makes it likely that the two lexica inherit the lemma from a lost version of the Synagoge (Σ´´ in Cunningham’s terminology: cf. Cunningham 2003, 49), though the extant testimonies of the Synagoge do not preserve an entry on διατοιχέω.

(2)    Eubul. fr. 50 (= fr. 51 Hunter) (C.1)

The verb which Eubulus really used (διατοιχεῖν or ἀνατοιχεῖν) is destined to remain unknown (see above). Hunter (1983, 52), followed by Kassel and Austin (PCG vol. 5, 217), quotes the Antiatticist without attempting to restore the original text. He also chooses to reject Jacobi’s proposal (apud Meineke, FCG vol. 5,1, 184) to consider ἔδει of Poll. 1.114 (A.4) as part of Eubulus’ original line (Hunter 1983, 139; Kassel and Austin agree). However, this does not solve the problem that διατοιχεῖν ἔδει makes little sense within the syntax of Pollux’s sentence, which has ἐρεῖς as the governing verb. There are thus two possibilities: either διατοιχεῖν ἔδει is a quotation (as Bethe thought), but the preceding text has been made irrecoverable by epitomisation; or the text, as edited by Bethe, is faulty. ἔδει is omitted in codd. FSBC, the main testimonies for two of the four branches of the Onomasticon tradition: perhaps it was a gloss meant to explain the prescriptive ἐρεῖς.

(3)    Arr. Epict.–2 (C.3)

The main manuscript of the Dissertations, cod. Bodl. misc. gr. 251 (S), presents the corrupt sequence ἂν … ἀ….ήσω. This is corrected into ἂν ἀτυχήσω in the manuscripts copied from S (s); in S itself, a second hand has corrected the sequence into ἀναχωρήσω. Modern corrections, reported by Schenkl (1916, 268), include ἀνατειχίσω, ἀνακυκλήσω, ἀνισοτοιχήσω, and ἀνατοιχήσω, the last of which was chosen by Schenkl.

(4)    Simp. in Epict. 38.710–12 (C.4)

The lectio difficilior ἀνισοτοιχοῦντα is transmitted by the Vat. gr. 326 (B, second half of the 12th c.), the oldest testimony of the commentary, and by the Vat. gr. 327 (C), which derives from B; it is further attested in the Marc. gr. 261 (G) as a conjecture (by Bessarion). The Par. gr. 1960 (J) has the syntactic lectio facilior ἀνισοτoιχοῦντας, while three further manuscripts (D: Par. Mazarin. 4460; E: Par. gr. 2072; F: Par. suppl. gr. 1023) have the abbreviated form ἀνισοτοιχοῦντ-. Only the Bonon. 2359 (H) and the Venetian editio princeps (1528: Σ) present the banalising reading ἀνισοτειχοῦντες. For this passage, see Hadot (1996, 390); on the manuscript transmission of Simplicius’ commentary, see Hadot (1996, 163–72). The sentence equating repentance to the sailors’ change of position in the ship does not have a parallel in Epictetus’ Manual, but it most probably depends on the passage of the Dissertations in which Arrian/Epictetus uses the same image: a hint in this direction is Simplicius’ ἐπὶ τὸ ἐναντίον μεθίστασθαι (‘to change to the opposite side’), which is reminiscent of ἀνατοιχήσω ἐπὶ τὸ ἐναντίον in Arrian/Epictetus.


Cunningham, I. C. (2003). Synagoge. Συναγωγὴ λέξεων χρησίμων. Texts of the Original Version and of MS. B. Berlin, New York.

Behr, C. A. (1986). P. Aelius Aristides. The Complete Works. Vol. 1: Orations I–XVI. Leiden.

Berardi, E. (2016). ‘Maestri di atticismo. La lingua e i suoi modelli in tre testi di ambiente scolastico di Elio Aristide (or. 30-32)’. Pernot, L., Abbamonte, G., Lamagna, M. (eds.), Ælius Aristide écrivain. Turnhout, 245–62.

Bortone, P. (2010). Greek Prepositions. From Antiquity to the Present. Oxford.

Hadot, I. (1996). Commentaire sur le Manuel d’Epictète. Introduction et édition critique du texte grec. Leiden.

Hard, R. (2014). Epictetus. Discourses, Fragments, Handbook. Oxford.

Henderson, J. (2002). Aristophanes. Vol. 2: Frogs, Assemblywomen, Wealth. Edited and translated by J. Henderson. Cambridge (MA).

Hunter, R. L. (1983). Eubulus. The Fragments. Cambridge.

Jones, C. (2008). ‘Aristides’ First Admirer’. Harris, W. V.; Holmes, B. (eds.), Aelius Aristides between Greece, Rome and the Gods. Leiden, Boston, 253–62.

Lobeck, C. A. (1820). Phrynichi Eclogae nominum et verborum Atticorum. Leipzig.

Luraghi, S. (2012). ‘The Spatial Meaning of διά with the Accusative in Homeric Greek’. Mnemosyne 65, 357–86.

Matthaios, S. (2013). ‘Pollux’ Onomastikon im Kontext der attizistischen Lexikographie. Gruppen «anonymer Sprecher» und ihre Stellung in der Sprachgeschichte und Stilistik’. Mauduit, C. (ed.), L’Onomasticon de Pollux. Aspects culturels, rhétoriques et lexicographiques. Paris, 67–140.

Risch, E. (1974). Wortbildung der homerischen Sprache. 2nd edition. Berlin, New York.

Schenkl, H. (1916). Epictetus. Dissertationes ab Arriano digestae. Leipzig.

Schwyzer, E.; Debrunner, A. (1950). Griechische Grammatik. Syntax und syntaktische Stilistik. Munich.

Zangrando, V. (1998). ‘L’espressione colloquiale nelle “Diatribe” di Epitteto. Contatti con lo stile della predicazione diatribica’. QUCC 59, 81–108.


Olga Tribulato, 'ἀνατοιχέω, διατοιχέω (Phryn. PS 62.1–3, Phryn. Ecl. 132, Antiatt. δ 28, Poll. 1.114)', in Olga Tribulato (ed.), Digital Encyclopedia of Atticism. With the assistance of E. N. Merisio.
DOI: https://doi.org/10.30687/DEA/2021/01/000

This article deals with the verbs ἀνατοιχέω and διατοιχέω, discussed in the Atticist lexica Phryn. PS 62.1–3, Phryn. Ecl. 132, Antiatt. δ 28, and Poll. 1.114.

CompoundsDenominative verbsPrefixesPrepositionsSemanticsἰδιωτικός