Appendix. Colometry and Formulae

Oral-traditional tale-telling includes embedded cues to the larger back stories that begin at the level of cola and formulae. They are the first level of consideration in any search for characterization in Homer. [1] In colometry, “words,” to borrow Foley’s description, from particles to longer lexical (i.e. formulaic) components, are preserved whole. [2] What becomes clear from a survey of metrical studies of Homeric prosody since the initial work of Parry and Lord is “the intrinsically deep connection of many formulas and formulaic substitution systems to firm metrical structures (usually colometric units)” (Russo 1997:258). Cola and formulae are often coextensive, although formulae or formulaic systems can span multiple cola.
Cola are principal compositional units for an oral traditional composer. [3] It is to the individual and composite cola that we can often look for the building components of formulae and formulaic systems. In short, the colometrics of Homer’s lines act as a guide for comprehending the poet’s traditional idiom. The schema I adopt for the make-up of verse cola throughout this study is that of Peabody, [4] whose work is a variation of Fränkel. [5]
The Homeric line is made up of shorter cola, which can be divided under normal circumstances in any six places, as follows: [6]
– ⏔ | – | ⏔ – | ⏓ | ⏓ – | ⏔ | – ⏕ – x
      A1   A2      B1   B2    C1   C2
Peabody’s particular scheme, based upon his research of Hesiod’s Works and Days, was praised by Albert Lord as representative of oral traditional composition in Greece (and of the Balkans). [7] Since Lord’s observations reflect extensive fieldwork and a thorough knowledge of Homer, it seems reasonable to suggest, as a general rule of thumb, that colometric analysis is closely representative of the tectonic structure and compositional patterns of the original singers who gave us early Greek epic poetry. [8] Understanding Homer’s language, then, involves the task of recognizing formulaic elements with the assistance of the natural cola of a line. While this approach is not always adequate, as Haslam (2003) has noted, and as anyone who works with the cola of the Homeric line can attest, it does very often indicate Homeric compositional units.
The preceding system of colometry can be illustrated by an example from the first book of the Iliad, where Achilles welcomes Agamemnon’s heralds who come to take his girl, but adds damning aspersions on their despotic leader (1.333):
οὐδέ τι | οἶδε νοῆσαι | ἅμα | πρόσσω καὶ ὀπίσσω
           A1                        B2       C1

And not in any way | does he know how to consider |at the same time | before and after
As we saw in Chapter 4, the poet employed this final colon formula to suggest significant implications for Agamemnon’s characterization for his core audience. [9] Cola commencement, bridging, and termini, often mark the beginning and completion of formulae (or formulaic systems). [10] In Iliad 1.148, we first noted the formulaic line: [11]
τὸν δ’ ἄρ’ ὑπόδρα ἰδὼν προσέφη πόδας ὠκὺς Ἀχιλλεύς

(“At him darkly looking, spoke swift-footed Achilleus”)
Here we saw the presence of a bridged A colon, with the formula τὸν δ’ ἄρ’ ὑπόδρα ἰδών (“At him darkly looking), ending at B1. As we noted there, the formula has implications as an idiom for Homer’s audience.
Further, within and joining the cola of Homer’s poetic lines are formulae or formulaic elements that are sometimes difficult to demarcate absolutely or exclusively, as Russo’s review of studies since Parry’s initial formulation suggests. [12] One cannot limit formulaic diction as a whole, moreover, to formulae that only fit Parry’s early definition of “an expression regularly used, under the same metrical conditions, to express an essential idea,” [13] at least not for purposes of exploring what may in fact be the full extent of “formulaic” diction. [14] The more theoretical insights and developing approaches such as those of Nagler, Visser, and Bakker, however intangible, cannot be ignored. [15]
Yet, for practical purposes in seeing and comparing formulae, clear boundaries are necessary. I therefore limit my research to collocations of words that more obviously constitute formulae. I have employed the term formula to speak of recurring words or traditional idioms, employed, as Parry stated it, “under the same metrical conditions, to express an essential idea.” [16] Of course, what that “essential idea” is has been suggested by scholars working to comprehend the referentiality inherent in these units of meaning. The question of the “essential idea” has been the focus of my present study, as well. Formulae can also be part of formulaic systems, patterns of words with one or more important parts recurring as component(s), but with a varying amount of replacement of parts within the traditional idiom (and which can occasionally vary in length). Further, modifications and substitutions that affect metrical shape in a formulaic system have been noted throughout our present study.
Finally, since cola form the basic units that the Homeric poets employed for traditional language cues, my translations throughout the present work have been concerned first, not with the idiomatic rendering of Greek into English, but with maintaining, as far as possible, the cola of the original Greek line. A tendency toward formal equivalence, then, rather than dynamic equivalence, has been the goal in translation. [17] The style at times appear wooden, but also allows English readers to see many of Homer’s sense units more clearly, as they hear Homeric characterization through traditional language cues. [18]


[ back ] 1. On the colometry of the epic hexameter line, see Fränkel 1955:104, Halporn et al. 1963:11-12, Jones and Gray 1972, Nagy 1974, Peabody 1975:66-70, Edwards 1987:4-54, Foley 1990:80-82, Sale 1993, Nagy 2000, Edwards 2002, Garner 2011:3-17, and Porter 2011.
[ back ] 2. I mean here by “words,” both the Greek words of a traditional lexicon and, as Foley (1991:26n57) argues, composite words such as formulae, which should not be divided. The difference between cola and metra in this singular respect suggests the superior internal logic of colometrics for a consideration of the structuring of the Greek hexameter.
[ back ] 3. Peabody 1975:68.
[ back ] 4. 1975:66-70. See Foley (1990:73-80) for a favorable overview of Peabody and statistics on cola placement.
[ back ] 5. 1955:104. The response of Kirk (1985:18-23; 1990: passim) to Fränkel and his predecessors was to question the developing principles of colometrics, including the four part line, essentially because of the number of bridges that occur and the location of the caesura in a certain percentage of Homeric lines. Kirk’s stance is ameliorated somewhat by his own observation that Fränkel’s (and other scholars’) “analysis has been a productive one, since many Homeric verses do naturally fall into those [four] cola” (1985:20). While it is necessary to adopt a working principle for considering the cola of any line, it is unwise to speak in absolutes.
[ back ] 6. The placement of the caesurae in no way negates the normal patterns suggested by Meyer’s or Hermann’s Bridge. In the case of Meyer’s Bridge (affecting where the break occurs for the first cola), the rule is that if the second foot is a dactyl (–⏑⏑), then the two short syllables must be part of the same word-unit. This means for colometric analysis, which does not divide words in any case, that the break (unless bridged) would come at A1. In the case of Hermann’s Bridge (affecting where the fourth cola begins, at C1 or C2), which observes that if the fourth foot is a dactyl, then the two short syllables must also be part of the same word-unit, the break would come at C2 (the adonean clausula), unless the second hemistich is bridged.
[ back ] 7. Peabody 1975:xi-xiv. Cf. the comments of Austin 2009:95-96.
[ back ] 8. It needs to be said that I have based most of my conclusions in Chapters 2-5 (even when I am considering the findings of others) on firsthand data gleaned from innumerable searches of Homer using the TLG database and incessant analysis of the poetic verse using the system of colometry I outline here. Consequently, any statistical mistakes are almost always my own.
[ back ] 9. See 4.2.1 Agamemnon’s Dishonoring and Hubristic Actions: 1.6–344.
[ back ] 10. The most significant colon break is that of the mid line, followed by C2 respectively, but of course, formulae and formulaic systems can stretch to whole lines and beyond.
[ back ] 11. 2.2.4 Menestheus and Odysseus: 4.327–364.
[ back ] 12. Russo 1997: 2011.
[ back ] 13. Parry 1971:13.
[ back ] 14. Anyone who has spent time trying to find and analyze formulae has necessarily noted the flexibility of Homer’s use of formulaic elements, including the substitutions, expansions, parallelisms, and differing degrees of fixity for particular parts of Homeric verse.
[ back ] 15. Nagler 1967, Visser 1988, Bakker 1997. Russo’s (1997, 2011) and Edward’s (1997) overviews of these approaches are balanced. The sorts of observations that Hainsworth (1968) made about the greater flexibility of Homeric prosody are real and cannot be overlooked; cf. Higbie 1990:152-198. While I do not try to set rules as to the number of times a formula must occur to be named as such, I accept the idea of formula recurrence as one sign that an expression is formulaic. Sale (1993:101) calls a formula exactly repeating fewer than six times an “infrequent formula.” I do not, as Peabody (1975:97) or Parry (1971:275n.), attempt to establish the minimum length for formulae.
[ back ] 16. Parry 1971:13; cf. 272: “a given essential idea.”
[ back ] 17. As a starting point for formal versus dynamic equivalence, see the bibliography listed in Pedro 2000:415.
[ back ] 18. Kelly (2007) perhaps wisely eschews traditional terminology opting instead for a “unit” of meaning in his referential commentary.