Despite uncertainties of formal detail (v. Heubeck 1988:249), it is uncontroversial to assume that Καλυψώ derives from καλύπτω ‘cover’ (< *ḱel- ‘id.’), that this derivational relationship was apparent both to bard and audience, and that the name corresponds overtly to the character’s principal activity: the literal isolation and complete obscuration of Odysseus. Though the name alone is so transparent as to speak for itself (cf. Ἀντίνοος), it may be argued that this saliency, rather than precluding word-play, conditioned Καλυψώ’s participation in an etymologically allusive phrase: ἔρυκε Καλυψώ ‘Calypso detained’.
As Risch (1947:72–91) demonstrated, Homeric word-play is of an altogether greater subtlety than Hesiod’s more conspicuous aetiological etymologizing. For example, as I would argue, the echoic word-play with Ὀδυσσεύς by means of δύστηνον ὀδυρόμενον κατερύκει (1.55) and ὠδύσαο (1.62), though actual etymology may be unintended, still alludes variously to the hero’s unhappiness and subjection to superhuman ire. In contrast, the formula Ἅρπυιαι ἀνηρέψαντο ‘the Snatchers/whirlwinds snatched’ (1.55; 14:371; 20:77) is clearly etymological word-play effected by phonic resemblance founded on an underlying correspondence of meaning (cf. ἀναρπάξασα θύελλα 20:63); that is, the poetic intimation of lexical kinship is mediated by segmental and syllabic similarity and conditioned by the bard’s native feel for the intended semantic field. As regards ἔρυκε Καλυψώ, while the metrically principal internal echo of the liquid-initial syllables constitutes the medium (-ρῡ ˉ˘˘: -λυ- ˉ˘˘), a sense of the message emerges from the frequency with which (κατ)ερύκω is predicated of Calypso (1.14, 1.55, 9.29, 23.334). Indeed, the two are so bound together that the poet can play with the audience’s expectations by having Athena/Mentes mention the concealer’s location and activity (κατερύκεται…νήσῳ ἐν ἀμφιρύτῃ…ἐρυκανόωσι 1.197–9), while concealing her agency (χαλεποὶ δέ μιν ἄνδρες 1.198). Not only does this diction display and play on word-play, it may also betray in solely semantic terms an etymological basis recoverable through morphological analysis of ἐρύκω itself.
Despite Chantraine’s derivation from (ϝ)ἐρύω ‘drag’ (2009: 375–6), it is generally agreed that ἐρύκω, formed with the marginal, yet paralleled, -κω present suffix (e.g. ὄλε-σα- : ὀλέ-κε/ο- :: ἔρῡ-σα- : ἐρύ-κε/ο-), derives from ἔρυμαι ‘guard, protect, ward off, check’ (Schwyzer 1939:702; Frisk 1960–1973:568; Risch 1974:279), because of semantic overlap and lack of initial digamma (v. Chantraine 2013:136–7). Viewed comparatively, ἔρυμαι is formally and semantically problematic. In form and largely in meaning, the best etymon is *ser- ‘aufpassen auf, beschützen’ (LIV2:534; v. Chantraine 1968–1980:376), from which was formed the u-present reflected by psilotic ἔρυμαι (cf. YAv. ni-šhauruuaiti ‘watches’ Jasanoff 2003:142). Though it is possible for a preform *ser- of the sense ‘watch/guard’ to underlie all of ἔρυμαι’s above-mentioned meanings, there are two instances where ἔρυμαι simply means ‘cover’ (5.484; 6.130). The first, when Odysseus covers himself with leaves (φύλλων…ἔρυσθαι 5.483–4) is soon echoed by φύλλοισι καλύψατο (5.491); the second, when Odysseus uses leaves to conceal his genitals before meeting Nausicaa (ὡς ῥύσαιτο 6.130), is without any sense of physical protection. Since semantic change from ‘cover’ to ‘protect, shelter, etc.’ is straightforward (e.g. Ved. śárman- ‘protection’ < *ḱel-mn̥ ‘cover’), it is possible that ἔρυμαι, as the formal reflex of *ser-, has appropriated the meaning ‘cover’ from an originally distinct but formally and semantically overlapping root *h1u̯er- (e.g. Εὐρυσίλαος, Chantraine 2009:376; cf. conflation in Ved. vṛṇóti ‘cover, shut, protect, ward off’ LIV2:227–8). Thus, by virtue of ἔρυμαι’s ‘cover’ meaning, ἐρύκω inherited the sense ‘confine’ (cf. Lat. arceō ‘enclose, shut > confine, hinder, ward off’; *h2erk- > arx ‘Schloss’ and Hitt. ḫarzi ‘holds’ Watkins 1970:67).
In sum, I suggest that the poet chose ἐρύκω precisely because its meaning of confinement, along with derivationally inherited associations of covering and concealing, satisfied the semantic correspondence needed for etymological word-play. Accordingly, just as ἀνηρέψαντο pairs etymologically with Ἅρπυιαι, so too was ἔρυκε Καλυψώ intended as an etymological figure informed by a native feel for the type of “detention” involved.
Greek and Latin Linguistics