Published online by Cambridge University Press: 11 February 2009
In a recent article Margalit Finkelberg raises the question of whether or not the phrase κλοσ ἄπ;θιτον at Iliad 9.413 is indeed a Homeric formula:
λετο μν μοι νóατοσ, τρ κλοσ ἄπ;θιτον ἔσται
Her purpose is to ‘test the antiquity of κλοσ ἄπ;θιτον on the internal grounds of Homeric diction’ (p. I).1 Proposing to use specifically the analytic techniques of oral theory, she argues that this phrase does not represent a survival from an Indo-European heroic poetry, as has been suggested from the occurrence of its exact cognate, śápos;rdvas áksitam, in Vedic poetry. To this end Finkelberg presents a precise and carefully organized argument. I briefly summarize its two branches as follows:
(1) It is the formulae that comprise the oldest stratum of Homeric diction, and so it is here that one would find survivals of Indo-European poetic diction. κλοσ ἄπ;θιτον ἒσται (which Finkelberg correctly argues to be the complete phrase), however, cannot be judged a Homeric formula by the criterion of repetition since it is a unique phrase in Homer. Nor can it be judged a formula by the ‘functional’ criterion since the better attested κλἒοσ οὒ ποτ óλεῖται expresses the same essential idea in the same metrical shape.
(2) A unique phrase such as that in question might nonetheless be ancient. The development of the use of ἄπ;θιτοσ, first to modify concrete nouns, and only later with abstracts, however, would indicate that its use with κλοσ is late. The demonstration that κλοσ ἂθιτον ἔσται is a ‘formulaic expression’, moreover, argues that it was coined for this specific context in Iliad 9 by analogy with other Homeric formulae, and so does not preserve an Indo-European formula.
1 Finkelberg, M., ‘Is κΛΕΟΣ ΑΠ;Δitona Homeric Formula?’, CQ 36 (1986), 1–5CrossRefGoogle Scholar. For a complete bibliography of this much-discussed phrase see Finkelberg's article. Her argument is intended chiefly as a correction of Nagy, G., Comparative Studies in Greek and Indie Meter (Cambridge, MA, 1974CrossRefGoogle Scholar, and Schmitt, R., Dichtung und Dichtersprache in indogermanischer Zeit (Wiesbaden, 1967)Google Scholar.
2 As Finkelberg points out (p. 4, note 13), κλοσ ἄπ;θιτον is attested elsewhere in archaic Greek poetry, notably: Hesiod, fr. 70.5 (M.–W.); Sappho, fr. 44.4 (L.–P.), Ibycus, fr. 1.47 (P), and an inscription from Krisa, Schwytzer nr. 316. See Schmitt, op. cit., 62ff. Whether these are Homeric reminiscences or independent occurrences of a traditional phrase remains uncertain. In addition to κλοσ ἄπ;θιτον Schmitt (72–102) discusses among others the phrases, μγα κλοσ and εκλησ (cf.κλοσ ἄπ;θιτον) in connection with the corresponding Vedic phrases uniśravas, máhi śrávas, and vásuśravas. Since Schmitt's argument about κλοσ ἄπ;θιτον is cumulative, based upon these and other phrases, Finkelberg's discussion of that phrase in isolation is perhaps somewhat narrow in scope.
3 By ‘functional redundancy’ (p. 3) Finkelberg presumably refers to Parry's criteria of extension and simplicity (discussed below), which preclude the existence in Homer of synonymous, metrically equivalent formulae.
4 Lord, (The Singer of Tales [Cambridge, MA, 1960])Google Scholar defines the formulaic expression as ‘…a line or half line constructed on the pattern of the formulas’ (p. 4), and later, formulaic expressions ‘…follow the basic patterns of rhythm and syntax [sc. of formulas] and have at least one word in the same position in the line in common with other lines or half lines…’ (p. 47). The ‘formulaic expression’ had not emerged as an explicit concept in Parry's work, but see ‘Studies in the Technique of Oral Verse-Making I’, Making of Homeric Verse, p. 312. Finkelberg does not use this term either, but argues that is based upon κλοσ ἄπ;θιτον ἔπ;θιτον oAeu-ai (3 times); κλοσ οποτ’ óλεῖται (4 times); and ἄπ;θιτον αíεí (4 times); and ∪∪– κλοσ ε࿖ναι (4 times.
5 In The Making of Homeric Verse. The Collected Papers of Milman Parry, Parry, Adam, ed. (Oxford, 1971)Google Scholar, the English translation of Parry's, dissertation V Épithet Traditionnel dans Homére (Paris, 1928)Google Scholar.
6 See Traditional Epithet, pp. 13–14.
7 For systems of simplicity and extension see Traditional Epithet, pp. 16–18, 37ff. Parry discusses Apollonius and Vergil at pp. 24–36 of the same study. Parry expresses the opinion, following Meillet, that all of Homeric diction is formulae at pp. 8–9, 20–1, 79–83.
8 Parry maintains that poets do not memorize songs (‘ Studies in the Technique of Oral Verse-Making II’, Making of Homeric Verse, pp. 330–7), but he clearly thinks of the formula as a fixed, memorized phrase. His comment at Traditional Epithet, p. 20, that ‘…the bard knew [this technique] without being aware that he knew it…’ should, however, make us wary of thinking of memory in too simple or too mechanical terms.
9 Hainsworth, J. B., The Flexibility of the Homeric Formula (Oxford, 1968), pp. 23–32, 72–3Google Scholar. Parry himself was well aware that breaches of ‘simplicity’ occur within formula systems (Traditional Epithet, pp. 175–7; ‘Studies I’, p. 272), and notes that the degree of economy and simplicity of the noun-epithet formula is extraordinary ('Studies I', pp. 277—8). Lord (op. cit., pp. 42–3) anticipates Hainsworth in holding that formulae exist only for the most frequently used and common ideas.
10 Parry, , Traditional Epithet, pp. 175–7Google Scholar, discusses the examples of νησ εεργσ and ποντóποντóποροσ νησ, concluding ‘…both formulae were in the common stock of epic diction, both kept there by the ever-present models of the formulae in the oblique cases, so that Homer learned them both from his predecessors’. See Hainsworth's brief discussion of other examples at op. cit., p. 23.
11 See Parry, , ‘Studies I’, pp. 274–9Google Scholar, and Lord, op. cit., pp. 4,30. Parry contradicts his position in Traditional Epithet (p. 16), but does not alter his definition of the formula in this later work since he believes that the number of examples within a formula type (e.g.ατρ πεí + finite verb) and the frequency with which individual examples are repeated imply respectively extension and simplicity. This change does mark, however, a move from a qualitative to a quantitative definition of the formula motivated by the desire for the means to compare Homer to other poets, by the impossible magnitude of analysing all of Homeric diction in terms of simplicity and extension, and by the recognition that the noun-epithet formula is something of a special case. Lord continues to be concerned with simplicity (‘thrift’) in particular as a sign of traditional, oral diction (see op. cit., pp. 50–3), but following ‘Studies I’ neither it nor extension occupies a central position in the development of oral theory. See Hainsworth's discussion of the move from a qualitative to a quantitative definition of the formula, op. cit., pp. 16–18.
12 This double aspect of the formula as stylometric test and as generative mechanism is overtly acknowledged at the outset of Lord's chapter on the formula (op. cit., p. 30) when he contrasts the formula as a ‘tool’ and as a ‘living phenomenon of metrical language’, and goes on to propose ‘…to look at the formula not only from outside in terms of textual analysis, but also from within, that is, from the point of view of the singer…’.
13 Lord argues that the ratio of formulae to formulaic expressions to unique expressions can distinguish the oral text from the literate text in a ‘conventional’ style from the straight literate text. See op. cit., pp. 130–1; and ‘Homer as an Oral Poet’, HSCP 72 (1967), 19–24Google Scholar.
14 The distinction between what the rate of repetitions in a text tells us about that text as a whole and what it tells us about an individual repetition in that text is a fine point, but one thrust to the forefront by Finkelberg's argument.
15 See Traditional Epithet, pp. 68–74; ‘Studies I’, pp. 267–79, 307, 319–23. In so far as Finkelberg relies upon Parry's work, it is fair to point out that he would first argue thatκλοσ ἄπ;θιτον ἓσται is simply a poorly attested formula (see ‘Studies I’, pp. 312–14), and further that he thought it impossible to infer a phrase's age from its status as a formula or lack thereof (see Traditional Epithet, pp. 68–9, and ‘Studies I,’ pp. 319–21).
16 Hainsworth, op. cit., pp. 33–5, objects to Parry's definition of the formula precisely on the grounds that it is unable to accommodate such a generative mechanism.
17 See Lord, op. cit., pp. 30–65. See also Kirk, G. S., The Songs of Homer (Cambridge, 1962), pp. 95–8Google Scholar regarding improvisation versus memorization and oral composition.
18 In his insistence upon the formula as a generative, dynamic element, Lord is in agreement with, among others, Kirk, The Songs of Homer; Hainsworth, op. cit.; Russo; Nagler, , Spontaneity and Tradition (Berkeley and Los Angeles, 1974)Google Scholar; and Peabody, Berkeley, The Winged Word (Albany, 1975)Google Scholar. Passages in Parry's work do indeed anticipate the direction taken by Lord; see Traditional Epithet, pp. 71–4 (regarding μπ;λνθεν ῂδσ and μπ;λνθε θλνσ τ), pp. 79–83; and ‘Studies I’, pp. 276–9, 312–14, 319–21. Although this distinction I draw between Parry and Lord is a fine one, it is of great importance for how we view the Homeric formula, and absolutely crucial for Finkelberg's argument.
19 Though Finkelberg's argument that the use of ἄπ;θιτοσ with an abstract concept such as κλοσ must be late is strictly outside the concerns of oral theory, I note here briefly my objection to this argument as well. This formulation of the abstract/concrete opposition is certainly anachronistic for Homer, whose heroes in any case view their κλοσ, like their τιμ, as something quite palpable. κλοσ ἄπ;θιτον, moreover, is no more an oxymoron of abstract and concrete than are κλοσ, κλοσ… ἄριτο, κλοσ ορανóν ἴκει, or κλοσ ερ
20 Peabody, pp.1–4, 30–1, 188–95; see also Lord, pp. 50–4. While Finkelberg ma y find monotonou s the line she claims Home r avoids, others might find it a n attractive an d elegant chiasmus, on e of Homer's characteristic figures (’ring composition’). Although Finkelberg is aware, moreover, that beyond its denotative meaning o f ‘unquenchable’ ἄπ;θιτοσ bears a connotation of ‘divine ‘(p. 4 with n.15), she does no t consider the possibility that the phrase κλοσ ἄπ;θιτον ἔσται might, therefore, express a different ‘essential idea’ than does κλἔοσ οποτ' óλεῖται I have mad e this suggestion elsewhere: see my Achilles in the Odyssey, Beiträge zur klassischen Philologie 171 (Königstein/Ts, 1985), pp. 75–9Google Scholar.
21 Tha κλοσ ἄπ;θιτον does represent a survival from a pre-Greek Indo-European heroic poetry seems to me likely to be the case, but I leave it to experts in this area to argue the point.
22 See Hoekstra, A., Homeric Modifications of Formulaic Prototypes (Amsterdam and London, 1964)Google Scholar, and Gray, D. F. H., ‘Homeric Epithets for Things CQ 61 (1947), 109–21Google Scholar. It is noteworthy that the evidence used by Hoekstra to identify late formulae, such as the role of nu-moveable or quantitative metathesis in certain formulae, is strictly external to oral theory itself, as is the archaeological evidence exploited by Gray. Of course, there are other methods than that of oral theory for studying such problems in Homeric diction. One might note among others the excellent work of Shipp, G. P., Studies in the Language of Homer 2 (Cambridge, 1972)CrossRefGoogle Scholar or Leumann, Manu, Homerische Wörter (Basel, 1950)Google Scholar, who bring to bear, however, techniques of analysis specifically designed to distinguish late and early diction in Homer.