ἔνδον, εἴσω
(Phryn. Ecl. 99, Antiatt. ε 9)

D. General commentary

The entries in Phrynichus (A.1) and the Antiatticist (A.2) are concerned with the correct usage of the adverbs of place ἔνδον ‘within’ and εἴσω ‘into’. ἔνδον derives from the preposition ἐνἐν and originally signifies state; εἴσωεἴσω (with the variant ἔσω mostly exploited for metrical purposes, but becoming standard in later Greek) derives from the preposition εἰς/ἐςεἰς and originally signifies motion towards. Both these adverbs are also used instead of the corresponding prepositions already found in Homer (see C.1 and LSJ s.vv. εἴσω b. and ἔνδον 2.), a feature which is sometimes condemned in the scholia (e.g. schol. [Ariston.] Hom. Il. 1.71.a [A], schol. [Ariston.] Hom. Il. 11.44 [A]). In the following discussion we will mostly consider the purely adverbial usages, though the prepositional usages are frequent at all chronological and stylistic levels and must have played a role in the perception and evolution of the adverbs as well.

Phrynichus (A.1) states that ἔνδον cannot be used with verbs of motion, such as εἰσέρχομαι, and should only be used with verbs of rest, like εἰμί. Likewise, according to Phrynichus, the correct adverb with verbs of motion is ἔσω (or εἴσω), which should not be used with verbs of rest such as εἰμί and διατρίβω (it may be noted, incidentally, that in prescribing ἔνδον with διατρίβω, Phrynichus could have had Pl. Phd. 116b.6 in mind: χρόνον γὰρ πολὺν διέτριψεν ἔνδον ‘[he] spent a long time within’). The abbreviated lemma in the Antiatticist (A.2) overturns these prescriptions by recalling exactly the opposite usage in Demosthenes, where εἴσω occurs with ἔχω (C.2 and C.3; see too F.1), which is not a verb of motion. The preserved direct quotation from Od. 7.13 (C.1), with κοσμέω, reiterates the point. The same quotation appears in Apollonius Dyscolus (B.1, see here below), who defends εἴσω for ἔνδον, and in the treatise On solecism and barbarism by Polybius of Sardi (B.2), where, to the contrary, the usage is overtly identified as a solecism. The same rule as in Phrynichus is enunciated, with more explanations and examples, by [Ammonius] (B.3). The entry quotes three Attic authorities, though in the first two ἔσω governs a genitive and is thus used as a preposition; only in the Eubulus fragment does it have a purely adverbial function.

These sources do not fail to notice that the confusion stigmatised by Phrynichus is already found in Homer. Apart from Od. 7.13 (C.1), it also occurs in Od. 18.48–9Hom. Od. 18.48–9 ([…] οὐδέ τιν’ ἄλλον | πτωχὸν ἔσω μίσγεσθαι ἐάσομεν αἰτήσοντα ‘we shall not suffer another beggar to mingle with us inside and beg’), Od. 18.96–7Hom. Od. 18.96–7 ([…] ὁ δ’ αὐχέν’ ἔλασσεν ὑπ’ οὔατος, ὀστέα δ’ εἴσω | ἔθλασεν […] ‘but [Odysseus] smote him on the neck beneath the ear and crushed the bones inside’), and in h.Merc. 6h.Merc. 6 (ἄντρον ἔσω ναίουσα παλίσκιον […] ‘living within a thick-shaded cave […])’. Based on Od. 7.13, in his treatise On adverbs Apollonius Dyscolus (B.1) provides what may well be the earliest defence of locative εἴσω. Applying the rules of analogy, Apollonius states that εἴσω behaves like other adverbs of place in -ω, where the same form signifies both state and motion towards (see too schol. D.T. GG 1,3.99.5–9Schol. D.T. GG 1,3.99.5–9). Apollonius seems to imply that between εἴσω and ἔνδον, it is the latter that is the odd one out, because of its morphological structure.

This theory (repeated in Apoll.Dysc. Adv. GG 2,1.204.26–205.4Apoll.Dysc. Adv. GG 2,1.204.26–205.4) is unattested elsewhere, but most erudite sources defend the admissibility of locative εἴσω (an exception would be the source – perhaps Aristarchus – on which the scholium to Od. 7.13 is based: see B.9 and Pontani ad loc.). Apart from the Antiatticist, see the synonymic-contrastive lemma in Hesychius (B.4, deriving from Cyril’s lexicon), the two lemmas in Photius (B.5, B.6, both of which refer to the same Demosthenes passage also quoted by the Antiatticist, which is the source of B.5), the Lexeis rhetorikai (B.7), and Priscian (B.8). Priscian identifies the interchangeable use of ἔνδον and εἴσω as an overtly Attic feature (he also mentions the locative use of ἔνδοθενἔνδοθεν, though this is not part of the general rule enunciated in the canon, which does not deal with adverbs of motion from). This passage comes from the last section of book 18 of the Ars, containing an alphabetical Greek-Latin lexicon where Greek syntactic peculiarities are described according to Attic models (the so-called Atticismi: see Rosellini 2015, XXVIII–XXIX, Spangenberg Yanes 2017, LIII–LVII). Although Priscian’s passage is partly reminiscent of Apollonius Dyscolus, whom Priscian used in the preceding parts of the Ars, this section likely depends on an anonymous Atticist syntactic lexicon, probably produced around the 1st–2nd centuries CE (Valente 2014, 76–7, Spangenberg Yanes 2017, XLIII–LIII). Finally, even Thomas Magister (B.10), who usually agrees with Phrynichus, displays an array of good authors (Demosthenes, Aelius Aristides, Libanius) in order to criticise his predecessor.

Was Phrynichus therefore mistaken in condemning locative εἴσω? Perhaps his stance can be explained by considering its occurrences in Greek literature more closely. Locative εἴσω is frequently attested in koine texts (e.g. C.6, C.7, C.8) and mirrored in post-classical inscriptions and papyri (e.g. C.5, though it should be noted in passing that in papyri the most common function of εἴσω followed by a noun is temporal, ‘within a certain time’, cf. LSJ s.v. εἴσω II). The increased frequency in post-classical texts provided a likely explanation for both Phrynichus’ interest in this usage and the great number of other erudite sources that reflect on it, often with different views to Phrynichus. The behaviour of Attic is of course crucial for this enquiry. In Attic texts of all genres, including inscriptions, εἴσω – both adverbial and prepositional – is overwhelmingly used in its original function (‘into’). Unambiguous locative usages of εἴσω with a verb of rest in prose are rare (see C.2, C.3). In other cases, εἴσω occurs as a modifier of nouns or as a preposition meaning ‘within’. Examples include Thuc. 3.21.3Thuc. 3.21.3 διήκοντες ἔς τε τὸ ἔσω μέτωπον αὐτοῦ ‘reaching to the inner front of it’, Pl. R. 407d.5Pl. R. 407d.5 τὰ δ’ εἴσω διὰ παντὸς νενοσηκότα σώματα ‘the bodies affected by generally diffused internal maladies’, D. 7.41D. 7.41 οἳ οἰκοῦσιν εἴσω Ἀγορᾶς ‘those who live within [i.e. on this side] of Agora’, X. An. 3.3.7X. An. 3.3.7 εἴσω τῶν ὅπλων κατεκέκλειντο ‘they were shut within the heavy-armed troops’ (here however the pluperfect κατεκέκλειντο may be thought to imply a motion before the state of being ‘shut in’). In classical inscriptions, too, locative εἴσω mostly occurs in the fixed phrase εἰς τὸ εἴσω ‘in the inner part’ (e.g. IG 2².463.61IG 2².463.61 [Attica, 307/6 BCE]) or as an adverbial modification of syntagms (e.g. IG 4²,1.109.3.136-7]IG 4²,1.109.3.136-7 [Epidaurus, early 3rd c. BCE]: τᾶς χρίσι[ος τ]ῶν εἴσω τοίχων ‘of the varnishing of internal walls’, I.Ephesos 3.5–6I.Ephesos 3.5–6 [Ephesus, ca. 290 BCE]: ἐκ τῆς μὲν ἔσω μοίρης ‘from the internal part’).

In Old Comedy, too, εἴσω is routinely used with verbs of motion and the few possibly locative usages are ambiguous, since the meaning ‘within, inside’ could be the result of preceding motion. For instance, in Av. 435–6Ar. Av. 435–6 (ταύτην λαβόντε κρεμάσατον τύχἀγαθῇ | εἰς τὸν ἱπνὸν εἴσω πλησίον τοὐπιστάτου ‘you two there, take this and hang it up inside near the fire’), the use of εἴσω may have been influenced by the preceding accusative of motion (εἰς τὸν ἱπνὸν, governed by κρεμάσατον), while in Av. 1056–7Ar. Av. 1056–7 (ἀπίωμεν ἡμεῖς ὡς τάχιστ’ ἐντευθενὶ | θύσοντες εἴσω τοῖς θεοῖσι τὸν τράγον ‘Let us go there as quickly as possible, to sacrifice the goat inside’), εἴσω – though apparently modifying θύσοντες – may be required by the motion verb ἀπίωμεν (see too Hunter 1983, 130). A rare locative εἴσω in Aristophanes is found in Lys. 1052Ar. Lys. 1052, ἔσω ἐστίν: this locus alone provides a good Attic example with which to discredit Phrynichus’ statement that only ἔνδον ἐστι is correct.

Instead, the locative use of εἴσω (both adverb and preposition) is very frequently attested in tragedy (e.g. Aesch. Th. 232, C.4, etc.; Eur. Hipp. 2, Cycl. 679, Heracl. 584, etc.; Soph. Ai. 105, OT 58, etc.). Phrynichus rarely resorts to tragedy in the Eclogue and his treatment of tragic forms is often cautious (see entries on ἀλεκτορίς, ἀρχῆθεν, κυνηγός, λίβανον, Λάκαινα, πρόσφατος). Similarly, the line of Eubulus quoted by [Ammonius] (B.3), if known to Phrynichus, would hardly have represented a suitable example, given that Phrynichus rejects the models of Middle and New Comedy.

In conclusion, a possible explanation for Phrynichus’ proscription is that he did not find sufficient – and sufficiently representative – Attic models for locative εἴσω. [Ammonius] shows a similar attitude when he dismisses the examples in both Sophocles and Eubulus as mistakes. Unlike Phrynichus, the syntactic lexicon on which Priscian depends seems to have defended a more moderate form of Atticism (Valente 2014, 64–5; against the identification of this source with the Antiatticist, see Valente 2014, 69–70). For more tolerant Atticists, the use of locative εἴσω in tragedy and some other Attic authors may have been sufficient proof of its Attic character. For Phrynichus, on the other hand, the limited Attic attestations of locative εἴσω and its likely abuse in the lower koine may have been sufficient proof of its inadmissibility.

Turning now to ἔνδον with motion verbs, this usage is uniquely post-classical. It is open to speculation whether Priscian’s testimony (B.8) can be used to trace the origin of this use back to classical Attic (see Spangenberg Yanes 2017, 180–1). The two attestations in Dio Chrysostom (C.9) and Aelian (C.10), quoted in the LSJ, exemplify the use with a transitive verb (εἰσάγω) taking an object (ἡμᾶς) and with the intransitive verb of motion ὠθέομαι, respectively. In P.Tebt. 1.46.16 (C.11) the motion verb is εἰσβιάζομαι ‘force oneself in’. This use of ἔνδον, however, remains sporadic in papyri. The exact collocation criticised by Phrynichus, with εἰσέρχομαι, finds numerous parallels, starting with John Chrysostom (C.12; see also C.13, a text probably dating to the 6th century CE). There are also examples with ἔρχομαι alone, starting with the Visio Dorothei (C.14), a Christian poem in Homeric Greek transmitted by P.Bodm. 19 (see also C.15 and C.16). These usages continue into the Byzantine age, with a tendency to characterise prose in a lower register (see e.g. C.17, a hagiography which is likely to date from no later than the mid 10th century, cf. Rydén 1995, vol. 1, 41–56).

E. Byzantine and Modern Greek commentary

As shown in D., ἔνδον with motion verbs is attested in Byzantine texts, mostly those of a lower register (see C.17). In classicising prose, the adverb is normally employed with verbs of rest (e.g. Anna Comnene Alexiad 3.10.7Anna Comnene Alexiad 3.10.7: θήκη διάχρυσος ἔχουσα ἔνδον τμήματα διαφόρων ἁγίων ‘a gold pyx having relics of several saints inside’), or in adverbial expressions which imply εἰμί and indicate ‘that which is inside’ (e.g. Anna Comnene Alexiad 13.3.3Anna Comnene Alexiad 13.3.3: ἀλλ’ οἱ ἔνδον καταγελῶντες τῆς τραγικῆς ταυτησὶ τειχομαχίας ‘but those who were inside, making fun of this tragic assault […]’).

The situation is different for εἴσω, which continues its double life as an adverb/preposition of both rest and motion, with a prevalence of prepositional usages. Two examples from the normally classicising prose of Anna Comnene will suffice to illustrate this point. In Alexiad 3.1.5Anna Comnene Alexiad 3.1.5, εἴσω accompanies γενόμενος, a verb of rest (ὁ δέ γε τέως τῶν σκήπτρων δραξάμενος Ἀλέξιος ὁ ἐμὸς πατὴρ εἴσω τῶν βασιλείων γενόμενος ‘as soon as my father Alexius grasped the sceptre and was inside the palace […]’). In Alexiad 6.2.3Anna Comnene Alexiad 6.2.3, it occurs with the motion verb εἰσέρχομαι (κᾷθ’ οὕτως εἴσω τῶν πυλῶν ἀπογραφομένους εἰσέρχεσθαι ‘[Alexius promised] that when their names had been inscribed [they] would enter the gates in that order’). The attestations of εἴσω in Nicetas Choniates’ Chronicle are overwhelmingly with verbs of motion, but expressions of rest occur as well (e.g. 384.24–5 van Dieten: στὰς τοίνυν θωρακοφόρος εἴσω τοῦ περιβόλου τῆς πόλεως ‘standing, in full armour, inside the walls of the city’Nicetas Choniates Historia 384.23–4 van Dieten). In texts in Medieval and Early Modern Greek, έσω (this is the standard form of the adverb in Medieval Greek) sometimes keeps its original motion function (see Kriaras, LME s.v. A.1), but it is more frequently used in expressions of rest, as in ο Ρίτζος είναι έσω μας χωσμένος ‘Rizzo is hidden in our home’ in the early 16th-century Narrative of The Chronicle of Cyprus by George Boustronios (170 Kechagioglou = p. 500 Sathas, quoted in Kriaras, LME s.v. έσω A.2). In Standard Modern Greek, έσω is a learned re-introduction, typical of a very high register, like ένδον and εντός (the normal adverb of place being μέσα). έσω however survives directly in some Modern Greek dialects (e.g. Southern Italian, Cypriot, Dodecanesian; I thank Ioanna Manolessou for discussing this matter with me).

F. Commentary on individual texts and occurrences

(1)    D. 19.251, 19.255 (C.2, C.3)

In both Demosthenic passages, the act of keeping one’s hand inside the cloak is a symbol of self-restraint in public speaking, of which Solon is adduced as a model (see MacDowell 2000, 309). In D. 19.255, Aeschines uses this image as a symbol of moderation, but Demosthenes retorts that the appropriate occasion on which one should keep one’s hand hidden is during an embassy, i.e. so as not to receive bribes (see MacDowell 2000, 313). This Demosthenic image enjoyed some popularity in rhetorical treatises (see [D.]. 227, on exhalation, and Hermog. 1.10.60, on the power of apostrophes), which may perhaps explain why the Antiatticist selected this locus classicus instead of another to illustrate the use of εἴσω with verbs of rest.