Skip to main content

There are two statements often made, in the literature, about the following famous Iliadic line:

ὃν πότμον γοόωσα λιποῦσ' ἀνδροτῆτα καὶ ἥβην (Il. 16.858, 22.364)

The first statement is that this line contains a metrical irregularity (the scansion of ἀνδροτῆτα) which betrays great antiquity of the line itself (see West 1982:15, Watkins 1995:499 with references). The second is that this line specifically belongs to the death scene of Achilles (in a lost poem or tradition), which the poet of the Iliad is openly referencing (recently Barnes 2011).

The aim of this paper is to review and critique each of these statements, within a new approach to Homeric formularity, which uses the linguistic concept of Construction (Goldberg 2006) to describe the workings of Homeric technique both synchronically and diachronically (Bozzone 2010). In doing so, this paper will grapple with the fundamental questions of how can we establish the meaning of formulas and assess their antiquity.

Concern with the meaning of formulas has been acute since Parry’s determination that Homeric style was traditional. Scholars have long debated whether formulas can have a specific, contextual meaning beyond their traditional, generic one (what Foley 1991 calls traditional referentiality), as well as whether the usage of a formula can constitute a deliberate bond between two specific passages (what Burgess 2012 calls intertextuality, and I shall call textual referentiality).

This paper argues that while textual referentiality in Homer exists, it is predicated on the interaction between formulas and their context of usage, not on formulas alone. For instance, nobody would argue that, in using the formula

εἵλετο δ’ ἄλκιμον ἔγχος, ὅ οἱ παλάμηφιν ἀρήρει (Il. 3.338 of Menelaos, Od. 17.4 of Telemachos)

the poet of the Odyssey is trying to make an explicit reference to the arming of Menelaos in the Iliad, for no specific contextual connection holds between the two scenes.

In the death scenes of Patroklos and Hektor, it is the context of usage that sets up our formula as evocative of Achilles’ death, not the formula itself. Thus, one should not talk about ‘the formula for the death of Achilles’, but about how within a given passage a traditional, generic formula has the capacity of assuming specific referentiality. Likewise, one should not assume that if a formula has specific referentiality in one passage, it will do so everywhere else.

As for the antiquity of our formula, this paper will introduce criteria based on Grammaticalization Theory (Bybee 2010) for establishing the relative age of formulas within the poetic language. For a formula to preserve a prosodic feature of Mycenean age (as ἀνδροτῆτα has been claimed to do), one would have to show that it belongs to a fossilized, unproductive type in the language. Yet close constructional parallels for our line are found all the way through the Odyssey. Compare the following lines, all employing a coordinated object structure involving a trisyllabic –της abstract starting at the hephthemimeral caesura and a bisyllabic aorist in the middle of the line:

ἰδὼν ταχυτῆτα καὶ ἀλκήν (Od. 17.315)

μίγην φιλότητι καὶ εὐνῇ (Il. 3.455, 6.25, Od. 5.126, 23.219)

On the basis of these lines, one can write a construction:

3b[˘—]Verb 4a˘τητ˘ κα— ⩀] Dative/Accusative.NounPhrase

which readily generates our line above. This was a standard construction for trisyllabic –της abstracts in Homer, which all start with two light syllables; ἀνδροτῆτα (for reasons worth exploring) is the only exception. In following this construction, the poet ignored this discrepancy: he treated ἀνδροτῆτα as a regular trisyllabic –της abstract, thus violating the quantitative scheme (Chantraine 1958:107 could call this abrègement métrique, cf. δήϊον πῦρ).

This paper concludes that λιποῦσ' ἀνδροτῆτα καὶ ἥβην is a young, flexible formula, and that though the form ἀνδροτῆτα itself may be old (and deserving of individual treatment), its usage in our line is not. This conclusion directly undercuts attempts to use this line as evidence for the origin of the hexameter (Tichy 1981, Berg 1978).