Pindar's Homer
The Lyric Possession of an Epic Past
a machine readable edition
Gregory Nagy

Copyright © 1980, 1997 The Johns Hopkins University Press. All rights reserved. This document may be used, with this notice included, for noncommercial purposes within a subscribed institution. No copies of this work may be distributed electronically outside of the subscribed institution, in whole or in part, without written permission from the JHU Press.

Go to Previous chapter; Next chapter; Table of Contents; Information

Chapter 3

The Panhellenization of Song

§1. The concept of Panhellenism helps explain not only how the multiple traditions of Archaic Greek oral poetry became a synthetic tradition but also how this tradition, as visualized in the hypothetical schema that has been offered, tended to counteract the emergence of historically verifiable authorship. Further, the concept of Panhellenism also helps explain why the oldest body of Greek literature to survive--the poetry of Homer and Hesiod--is representative of oral poetry, not song. I argue that the Panhellenization of poetry preceded the Panhellenization of song because the traditions of song were more diverse than the traditions of poetry. In the Archaic and even the Classical period of Greece, it appears that the greatest diversity in local oral traditions was on the level of song, with a wide variety of different melodic patterns native to different locales. 1 In their diversity, the local traditions of song were less adaptable to the evolving synthesis that I call Panhellenization.

§2. The earliest attestations of a critical mass of actual compositions in Greek song--which we may also call lyric poetry--are represented by the surviving texts attributed to Alcman, Stesichorus, Alcaeus, Sappho, Ibycus, Anacreon, Simonides, Bacchylides, and Pindar; in the era of Alexandrian scholarship, these nine names constituted the inherited canon of lyric poetry (cf. Palatine Anthology 9.184, where all nine are enumerated). 1 The relative dating of these poets covers a period stretching roughly from 650 to 450 B.C. 2 These attestations of song are the reflex of what we may call a second wave of Panhellenization, achieved through an ongoing synthesis as represented by the myths about the "inventions" of figures like Terpander and Olympus, to whom we turn presently. As with the first wave, namely, the Panhellenization of oral poetry, the relatively later Panhellenization of oral song would entail a progressively restricted series of recompositions, in ever-widening circles of diffusion, with the streamlining of convergent local traditions at the expense of divergent ones. 3 In this way, a preexisting multitude of local traditions in oral song could evolve into a finite synthetic tradition of fixed lyric compositions suited for all Hellenes and attributed by them all to a relatively small number of poets.

§3. Wilamowitz was struck by the limited number of poets just listed, nine, whose lyric compositions were being edited in the era of Alexandrian scholarship, 1 and he was ready to conclude that these Archaic poets were for all practical purposes the only ones whose texts of lyric poetry had survived from the Classical into the Hellenistic period. 2 This line of thinking fails to distinguish between the relatively small selective canon inherited by the Alexandrian scholars and the massive nonselective collection of works that was at their disposal in the Museum; 3 it is enough to say that the canon of nine lyric poets was inherited from the Classical period. 4 I argue that this limited number of nine lyric poets is due to the Panhellenization of preexisting traditions in oral song, just as the comparably small number of canonical Archaic poets who are credited with compositions in hexameters (predominantly two, Homer and Hesiod) 5 or in iambics (Archilochus, Hipponax, Semonides, Solon, and so on) and elegiacs (Archilochus, Callinus, Mimnermus, Tyrtaeus, Theognis, Solon, Xenophanes, and so on) is due to the Panhellenization of preexisting traditions in oral poetry. 6

§4. I have suggested that the very evolution of what we know as the Classics--as both a concept and a reality--was but an extension of the organic Panhellenization of oral traditions. 1 In line with this reasoning, the evolution of an ancient Greek canon in both poetry and song need not be attributed primarily to the factor of writing. 2 Granted, writing would have been essential for the ultimate preservation of any canon once the traditions of performance were becoming obsolete; still I argue that the key to the actual evolution of a canon must be sought in the social context of performance itself.

§5. In Archaic Greece the form of song or lyric poetry was functionally divided into two distinct media of performance. This division is most explicitly formulated in a hypothetical discussion, as framed by Plato Laws 764c-765b, of an idealized system of festivals in an idealized polis; the presuppositions of this discussion, however, are based on the institutional realities of the polis. On the one hand is monôidiâ 'monody' (764de, 765b), that is, performance by a single professional such as a rhapsôidos 'rhapsode' or kitharôidos 'lyre singer' (764de). On the other hand is khorôidiâ 'choral song' (764e), that is, performance by a nonprofessional khoros 'chorus', a singing and dancing ensemble of selected men, boys, or girls (764e), 1 who represent some aspect of the polis as a whole. 2 The essential context of public performance in both the monodic and the choral media is competition--among rhapsodes, among lyre singers, among choruses. This competition is called krisis (765b), and those who are officially chosen to select the winners are kritai 'judges' (ibid.). 3 Granted, there is no traditional distinction made between a monodic and a choral mode of composition. 4 Still the medium of both monodic and choral composition is public performance. Moreover, it is possible to explain the distinctiveness of solo lyric poetry, or monody, in terms of choral lyric poetry, rather than the other way around. 5 For now, however, I confine myself to observing the traditions of melody in singing and instrumental accompaniment and postpone till later my account of the complex patterns of differentiation in choral and monodic performance. 6

§6. Because of the diversity of localized traditions in melody, oral poetry, not oral song, was better suited for Panhellenic diffusion in that rhythmical (metrical) and phraseological systematization would not violate localized perceptions of what is correct as readily as the synthesis of diverse melodic patterns. Granted, the melody of song would have promoted diffusion from the standpoint of mnemonic utility. 1 Still, melody would also have impeded diffusion from the standpoint of contextual sensitivity. Thus the process of Panhellenization took effect relatively later for oral song than it did for oral poetry.

§7. The Panhellenization of song required an ongoing synthesis of patterns in vocal and instrumental traditions. Within the actual traditions of song, however, what I am calling a synthesis is treated as if it resulted from inventions by prototypical figures. The two names most commonly associated with these "inventions" are differentiated in terms of instrumental accompaniment: Terpander in the realm of string instruments, to which I refer with the general term kitharâ 'lyre', 1 and Olympus in the realm of wind instruments, that is, the aulos 'reed'. 2 The most comprehensive direct account of the relevant traditions is to be found in "Plutarch" On Music 1032cd-1034b. Even before we consider this account, however, I should stress that instrumental accompaniment does not require the point-for-point following of the vocal part by the instrumental part: in "Plutarch" On Music 1137b, with a description of an Archaic musical style associated with Terpander and Olympus, it is made clear that the instrumental part calls for complexities that do not match the vocal part. 3

§8. Let us begin with Terpander, that is, Terpandros 'he who gives pleasure to men.' 1 Tradition has it that he was a singer from Lesbos who moved to Sparta, where he was the first of all winners at the Spartan festival known as the Karneia (Hellanicus FGH 4 F 85 by way of Athenaeus 635e). The Feast of Karneia was reportedly founded in the twenty-sixth Olympiad, that is, between 676 and 672 B.C. (Athenaeus 635ef). 2 In other words the inception of the Karneia, an institution that was recognized by tradition as the oldest established festival of the Spartans, was reckoned in terms of Terpander's victory in a contest of singing to the accompaniment of the lyre (Athenaeus 635ef). That Terpander was eventually thought to be a solo singer is clear from his being regularly designated as a kitharôidos 'lyre [kitharâ] singer' ("Plutarch" On Music 1132d, 1133b-d). 3

§9. Tradition also has it that Terpander "invented" what are called the nomoi of kitharôidiâ 'lyre singing': Boeotian, Aeolian, Orthios, Trokhaios, Oxus, Kêpiôn, Terpandreios, Tetraoidios ("Plutarch" On Music 1132d; cf. 1132c; supplemented by Pollux 4.65). 1 For the moment the word nomoi has been left untranslated. It is enough to observe at this point that Terpander's "invention," as discussed in "Plutarch" On Music 1132cd and Pollux 4.65, is traditionally seen as the forerunner of a specific genre 2 known in the time of Plato as the kitharôidikos nomos 'citharodic nome' (Laws 700b). In this context the word nomos is specific to the given genre: thus it is reported in "Plutarch" On Music 1132c that Terpander was the "inventor" of the kitharôidikos nomos 'citharodic nome', which is to be compared with the tradition, as reported in Herodotus 1.23, that ascribes to Arion the "invention," in Corinth, of the dithurambos 'dithyramb'. 3 Arion is described in Herodotus (ibid.) as the premier kitharôidos 'lyre singer' of his era, who was the first to name, compose, and teach the dithyramb at Corinth in the era of the tyrant Periandros. In this early description, we see that the word kitharôidos originally did not exclude involvement in choral performance.

§10. That the strict distinction in genres between citharodic nome and dithyramb may be relatively late is suggested by a detail in the story of Arion: when he is abducted by pirates and performs for them a solo lyric composition that ultimately saves his life, this composition is described as an orthios nomos (Herodotus 1.24.5). 1 This naming of Arion's tune corresponds to one of Terpander's nomoi in the list of his citharodic nomes (Pollux 4.65). Still, for Plato, the kitharôidikos nomos 'citharodic nome' and the dithurambos 'dithyramb' are to be treated as parallel genres inherited from the Archaic period (again Laws 700b).

§11. Let us move beyond the use of the term nomos in the Terpander tradition, in the context of Terpander's having supposedly invented the various categories of nomoi just listed. In this context nomos is a matter of a specific genre and reflects a specialization of usage. In other contexts, however, the word nomos refers more generally to various types of local melodic patterns. In generalized references to song within song, nomos has the general sense of 'localized melodic idiom' (as in Aeschylus Suppliants 69); 1 such a usage meshes with the basic meaning of nomos, which is 'local custom'. 2 Just as nomos as 'local custom' refers to the hierarchical distribution or apportioning of value within a given society (root *nem-, as in nemô 'distribute'), so also nomos as 'localized melodic idiom' refers to the hierarchical distribution or apportioning of intervals within the melodic patterns of song. 3

§12. In line with the earlier argument concerning melodic traditions as extensions of patterns in pitch or intonation, 1 I suggest that the various Greek systems of nomoi evolved in symbiosis with the patterning of pitch accent in the phraseology of song. 2 As a melodic pattern that is characteristic of distinct speech in distinct habitats, nomos serves as the ideal metaphor for conveying the distinctiveness of bird song, as when the voice of Alcman declares that he knows the nomoi of all the different kinds of birds in the world (PMG 40). This theme can best be understood in the context of Alcman PMG 39, where the poet names himself as the "discoverer" of melody and words that put into human language the voices of partridges. In other words, the song of Alcman is being conceived as a mimesis of bird song, and the varieties of bird song resemble the varieties of nomoi. Thus song, as a mimesis of speech, can extend into a mimesis of the "speech" of birds. 3

§13. In the context of Terpander's "invention" of nomoi, however, nomos takes on more restricted meanings. From the standpoint of our main source, "Plutarch" On Music (1132de), the nomoi of Terpander are the result of a systematization of preexisting melodic patterns. 1 The figure of Terpander represents the common denominator of these patterns. The fitting of these patterns into a system attributed to Terpander is formalized in the tradition that attributes to this figure the "invention" of a seven-note scale that accommodates his nomoi ("Aristotle" Problems 19.32; "Plutarch" On Music 1140f, 1141c). 2

§14. Corresponding to Terpander's "invention" of this scale is his "invention" of the seven-string lyre, displacing the older four-string type (Strabo 13.2.4 C618, quoting Terpander PMG, p. 363; "Plutarch" 1141d). The iconographical evidence of the eighth and early seventh centuries B.C. corroborates this tradition: the norm during this period is a four-string instrument, which is replaced after this period by a seven-string instrument. 1 It has been said that the spread of the seven-string lyre in the seventh century "betokens a revolution in music." 2 In terms of the Panhellenic synthesis that I am proposing, the older four-string lyre would be adequate for any single local nomos, while the newer seven-string lyre, which represents the "revolution" of the seventh century, would fit a wide variety of nomoi, irrespective of local provenience, within a new interrelated system. 3 In the diction of Pindar, Apollo is represented as leading the choral performance of 'all sorts of nomoi' pantoiôn nomôn Nemean5.25) as he plays on the seven-string lyre, which is described as heptaglôssos 'having seven languages' (5.24).

§15. Parallel to the tradition about Terpander's "invention" of the seven-string lyre is the story in Homeric Hymn to Hermes51: when Hermes "invents" the lyre, it has seven strings. In the narrative, Hermes gives the kitharis 'lyre' to Apollo (Homeric Hymn to Hermes499, in the context of 475 and following), the figure conventionally associated with this instrument (e.g., Hymn to Apollo131, 188). We may note, in connection with the traditional provenience of the Terpander figure from Aeolic Lesbos, the actual form kitharis 'lyre' in the Hymn to Hermes (e.g., 499) and in Homeric diction in general (e.g., Odyssey i 153): the accentuation of this word follows a clearly Aeolic pattern. 1 In yet another version Amphion "invents" the seven-string lyre in Thebes (Pausanias 9.5.7). In this case two of the nomoi "invented" by Terpander for kitharôidiâ 'lyre singing' are precisely the Aeolian and the Boeotian (again "Plutarch" On Music 1132d). This tradition squares with the linguistic facts: the dialectal heritage of Thebes is Aeolic in general and Boeotian in particular.

§16. Before we pursue further the topic of an implicit Panhellenization as reflected by the "inventions" of Terpander in the realm of the lyre, let us consider the counterpart of Terpander, Olympus, and his corresponding "inventions" in the realm of the aulos. Just as Terpander is a prototype of the kitharôidos 'lyre [kitharâ] singer', so Olympus is presented as a parallel prototype of the aulôidos 'reed [aulos] singer' (cf. "Plutarch" On Music 1137b, 1133def; Aristotle Politics 1340a). 1 Like Terpander, Olympus is credited with the "invention" of specific nomoi ("Plutarch" 1133d). 2 Olympus is a decidedly mythical figure, a disciple of the mythical aulos player Marsyas of Phrygia ("Plutarch" 1133e). Olympus too is said to be a Phrygian (ibid.); 3 this detail becomes more significant as our discussion proceeds. 4 Besides Olympus, there are later figures who are less remotely mythical in appearance and who tend to be synchronized with Terpander. One such figure is Clonas, described as an inventor of nomoi for aulôidiâ 'reed [aulos] singing' ("Plutarch"1132c), who supposedly lived shortly after the time of Terpander (1133a). 5

§17. The systematizations attributed to figures like Terpander and Olympus are comprised of nomoi. Thus the nomoi, from the standpoint of these traditions about "inventors" and their "inventions," are no longer separate melodic idioms: they are patterns that are already integrated with each other into a larger system that controls its constituents. In other words we see here an early stage in the ongoing Panhellenization of local traditions in song.

§18. In considering the performance traditions of Archaic Greek song, I concentrate on the internal and external references to the performance of compositions attributed to the canonical nine poets of lyric. In these references we can find clear traces of Panhellenic systematization, as we have seen in the example from Pindar: the seven-string lyre, presented as a symbol of systematization, allows Apollo to lead the choral performance of 'all sorts of nomoi' pantoiôn nomôn Nemean5.25).

§19. The word nomos, however, is hardly adequate for designating the actual process of systematization since its basic meaning of local custom retains a built-in emphasis on the local origins of the constituents of the system. A more adequate word is harmoniâ, in the specific sense of a tuning or accordatura that fits a given melodic idiom, as attested in Aristophanes Knights994. In a more general sense harmoniâ can be understood as a 'system of intervals in pitch', as in Plato Republic 397d, where the point is that the traditional harmoniâ and rhythm of song is regulated by the words of song. Plato's usage in this passage, it has been observed, "points to the fact that the existence of melody depends on the prior existence of an organised scheme of pitches standing to one another in determinate relations, on the basis of whose relations the selection that generates a melody is made." 1 We must distinguish this notion of harmoniâ from the later notion of tonos, especially as developed by Aristoxenus, pupil of Aristotle, who himself was the son of a professional musician from Tarentum (Suda s.v. Aristoxenos. The difference has been formulated as follows: "each tonos had the same pattern of intervals: they differed one from another, as modern keys do, only in respect of pitch." 2 In contrast the harmoniai, as Plato understands them in the passage under consideration, "were distinguished from one another primarily by being constituted out of different sequences of intervals." 3 Correspondingly "rhythm" in this passage means "the element of rhythmic organisation that any composition must possess, an individual rhythm being the formal rhythmic structure underlying an individual piece or type of piece, its overall pattern of movement." 4

§20. Let us consider the six names of various harmoniai as discussed in Plato Republic 398e-399a: the Ionian [= "Iastian"], the Dorian, the Phrygian, the Lydian, the Mixolydian, and the Syntonolydian. From other testimonia, to be cited later, we see that such harmoniai were the basic melodic patterns of Archaic Greek lyric in its attested phase of development. Moreover, Plato's list of Archaic harmoniai corresponds to actual self-references found in Archaic Greek lyric poetry.

§21. On the basis of these self-references, it becomes clear that there were in fact other such harmoniai, besides the ones listed en passant by Plato in Republic 398e-399a. Most notably missing from Plato's list is the harmoniâ named as the Aeolian in the Archaic diction of lyric poetry. I cite the specific reference to a harmoniâ described as Aeolian in Lasus of Hermione PMG 702.3, where it is described as barubromos 'deep-roaring', suggesting that the tessitura of the Aeolian was marked by its frequency of lower notes. 1 Plato's omission of the Aeolian, in view of his inclusion of the Ionian and the Dorian, is asymmetrical from the standpoint of the Greek language, the major dialectal subdivisions of which are Aeolic, Ionic, and Doric. 2 But this very asymmetry reflects the transformations over time in the systematization of Greek modes. The Aeolian harmoniâ was already replaced by Plato's time with the so-called Hypodorian. The Archaic nature of the Aeolian harmoniâ within the newer systematization of the harmoniai is indicated by the older concept of an Aeolian nomos within the framework of an older systematization of nomoi ascribed to Terpander. 3

§22. Not only the names of the harmoniai listed by Plato but even the descriptive modifications associated with them, such as suntono- 'tense' in Syntonolydian, correspond to actual contrastive self-references found in Archaic Greek lyric poetry. For example, in the words of the Archaic poet Pratinas of Phleious, an older contemporary of Aeschylus, what are described in terms of lyre tuning as suntono- 'tense' and aneimenê- 'lax' melodic patterns or "Muses" are rejected for the moment as extremes in favor of a "moderate" Aeolian harmoniâ (PMG 712). Plato uses the notion of khalarâ- 'lax' in describing the Ionian harmoniâ (Republic 398e).

§23. In our present survey of Archaic references to harmoniai, I draw particular attention to a quotation from Terpander, where the frame of quotation specifies that the song is being composed in the Dorian harmoniâ (PMG 698, in Clement Stromateis 6.88). Thus Terpander is associated with not only the Aeolian melodic tradition, compatible with an older system of modes known as nomoi, but also the Dorian, compatible with a newer system of modes known as harmoniai, which accommodate both Dorian and Aeolian. This accretive combination of older Aeolian and newer Dorian tuning or accordatura dovetails with the myth that tells how Terpander came from Aeolic Lesbos to Doric Sparta. 1 It dovetails also with a linguistic given: that the dialectal texture of the medium that we know as choral lyric is dominantly Doric, with significant recessive elements of Aeolic. 2 This dialectal texture is most evident in the choral lyric poetry of Pindar. 3 In fact the medium of Pindar provides the following explicit self- reference:

Aioleus ebaine Dôrian keleuthon humnôn
Pindar F 191 SM

An *eolic man went along the Dorian path of songs. 4

Appropriately these words are framed in the metrical system known as dactylo-epitrite, which is the Doric counterpart to the other major metrical system used in Pindar's choral lyric compositions, the Aeolic. 5 All the attested lyric poetry of Pindar, with only a few exceptions, is composed in one or the other of these two kinds of meters. 6

§24. From such evidence I infer that the choral lyric traditions represented by Pindar resulted from an accretive blend of Aeolic and Doric poetic language, where all the various elements of mode, rhythm, and the words themselves can be seen participating in the synthesis. The model that I posit here for the choral lyric tradition of Pindar can be extended to the lyric tradition of Sappho and Alcaeus. In this case the available evidence points to a diachronic blend of Aeolic and Ionic poetic language. 1

§25. The synthesis of Aeolic and Doric traditions is expressed overtly in Pindar's Olympian1, where a Dorian lyre is playing (17) while the molpê 'singing and dancing' is described as Aeolian (102). In this particular case I infer that the actual mode is Aeolian since the meter of this composition is Aeolic. Elsewhere self-references to the Aeolian mode correspond to the actual composition of the given ode in Aeolic meter (Pindar Pythian2.69, Nemean3.79). 1 So also with the Dorian mode: we have already seen an example of a reference to this mode within a song composed in dactylo-epitrite, that is, Doric meter (Pindar F 191 SM), and there are other examples (e.g., Pindar Olympian3.5). 2 Such a neat pattern of convergence between the given meter and the self-reference to a mode seems to work only where the mode is Aeolian or Dorian. Other Pindaric compositions in the Aeolic meter contain self-references to the Lydian mode (Olympian14.17, Nemean4.45; cf. Olympian5.19). Also, one composition in dactylo-epitrite meter contains a possible reference to the Lydian mode (Nemean8.15). The Lydian mode may be adaptable to more than one native metrical tradition. After all, in Plato's list of six harmoniai in Republic 398e-399a, we hear of three different kinds of Lydian mode: Lydian, Mixolydian, and Syntonolydian. Perhaps these differentiations result from different adaptations to different native meters.

§26. What seems at first to interfere with a coherent picture of melodic patterns in the development of traditions in Archaic Greek song is the traditional naming of seemingly "foreign" modes or harmoniai, such as the Lydian, which are distinct from the three native names of Aeolian, Ionian, and Dorian. For example, there is a reference to the Phrygian melodic patterns in Alcman PMG 126, as also in a fragment from the Oresteia of Stesichorus, PMG 212. Elsewhere Alcman is said to have named some famous Phrygian aulos-players in his compositions (Athenaeus 624b). In Pindar's own words (F 125 SM) a string instrument called the barbiton was "discovered" by Terpander at the feasts of the Lydians, the sound of which would answer to that of the pêktis, another Lydian string instrument. Alternatively the "invention" of the barbiton is attributed to Anacreon (Athenaeus 175e). On the authority of Posidonius, Anacreon is said to have referred to three kinds of melody, which were supposedly the only three he used: the Dorian, Phrygian, and Lydian (Athenaeus 635cd). We may compare the repertoire of Pindar, as gleaned from the self-references surveyed above: Dorian, Aeolian, and Lydian. As for the Lydian pêktis, noted by the words of Pindar (F 125 SM) in the context of praising Terpander as a forerunner, we may compare the report that credits Sappho with being the first to use this instrument (Athenaeus 635e). 1 On the authority of Aristoxenus (F 81 Wehrli), Sappho is also credited with the "invention" of the harmoniâ called Mixolydian, which the composers of tragedies supposedly learned from her songs ("Plutarch" On Music 1136cd). 2

§27. The contrastive nature of these "foreign" harmoniai turn out to be a key to our understanding the systematization of Archaic Greek modes. I draw attention to the contrastive mentions of explicitly local harmoniai, in particular Pindar F 140b SM, with its mention of a local harmoniâ of the Locrians that is described as a rival of the Ionian "Muse." 1 In order to explore further the principle of contrastiveness as it operates within a system of harmoniai, let us return to Plato's list of harmoniai in Republic 398e-399a: Ionian [= "Iastian"], Dorian, Phrygian, Lydian, Mixolydian, and Syntonolydian. These six harmoniai are described in terms of fixed scales by Aristides Quintilianus, p. 19.3-10 (ed. Winnington-Ingram), whose testimony seems to be a genuine reflex of old traditions in the actual performance of Archaic Greek lyric poetry. 2 Notable for its omission in both Plato Republic 398e-399a and Aristides Quintilianus, p. 19.3-10 is the Aeolian mode, specifically designated as a harmoniâ in Pratinas of Phleious PMG 712 3 and Lasus of Hermione PMG 702. 4 Heraclides Ponticus (F 163 Wehrli, by way of Athenaeus 624e) equates the Aeolian of Lasus with a new replacement category, the Hypodorian. Other sources (e.g., Cleonides, p. 198.13 Jan) equate the Aeolian specifically with Locrian, as "invented" by Xenocritus of Locri (scholia to Pindar Olympian 11.17); clearly such an old category as Locrian would be out of step with any newer systematization. According to Athenaeus 625e the Locrian became obsolete after Pindar. Which brings us back to our point of departure, the reference in Pindar F 140b SM to a local harmoniâ of the Locrians that is a rival of the Ionian "Muse."

§28. There is much confusion arising from the fact that such terms as Dorian and Phrygian earlier used to designate the harmoniai, were later adapted to the newer theoretical notion of tonos, as used by Aristoxenus. 1 In the so-called Greater Complete System of Aristoxenus, some earlier terms reflecting regional differentiation were replaced by newer terms designed to connect even more explicitly the various systems that had at earlier stages been separate from one another: thus the categories of Aeolian and Ionian were replaced respectively by Hypodorian and Hypophrygian. 2 In the case of Hypodorian we have noted the explicit testimony of Heraclides Ponticus (F 163 Wehrli, by way of Athenaeus 624e), who equates the Aeolian of Lasus of Hermione PMG 702 with Hypodorian, a component of a newer system of interrelations as recognized by Aristoxenus. Such names of harmoniai, as reused in the later classification- systems of Aristoxenus and the like, 3 reflect not the local melodic idioms themselves but the eventual Panhellenic systematization of categories that resulted from the ongoing juxtaposition of the local melodic idioms. 4

§29. Such categories as Phrygian and Lydian are particularly indicative: they convey identification of the locale indirectly, not by way of naming what is native but by way of representing what is alien to the native--in this case not just alien but "barbarian" as well. It has been pointed out that the harmoniai with barbarian names have no common feature that distinguishes them from the ones with Greek names, or the other way around. 1 The same sort of mentality is at work in the distinctions among harmoniai with Greek names. In the case of Ionian, for example, this harmoniâ would represent a synthesis of what is alien to, say, native Locrian traditions (again I cite Pindar F 140b SM). Still the synthesis implied by a concept such as the Ionian harmoniâ could take place in terms of local traditions. In other words the local traditions are the frame of reference. 2 In this way the alien is appropriated because it can be performed. In other words alien becomes native, in a broader and inclusive sense of native. 3 Such thought patterns of inclusion reflect the real beginnings of Panhellenism.

§30. It has been calculated that changing from any one harmoniâ to another meant re-tuning at least five strings on a seven-string instrument. 1 Which leads to the following inference: "To minimize the inconvenience which such re-tunings involved, musicians must have striven to find as much common ground between different modes [= harmoniai] as they could, and to identify certain notes in one where possible with notes in another." 2

§31. Still the various harmoniai were distinct enough to require considerable effort in the development of a performer's repertoire. In the comedies of Aristophanes we see the ridiculing of a character on the grounds that he could learn only the Dorian harmoniâ when he had been a boy in school (Knights 985- 995). 1 The context for such learning of the harmoniai can best be observed in the Clouds of Aristophanes, with its informative description of old-fashioned Athenian paideiâ 'education' (tên archaian paideian 961), the kind that purportedly produced the men who fought at the Battle of Marathon (985-986). 2 Boys learn selected compositions of old lyric masters in the house of the kitharistês 'master of the kitharâ' (964), who teaches them to learn by heart (promathein: 966) the performance of famous lyric compositions (967) 3 and who insists on their adherence to performing these compositions in the proper harmoniâ 'mode' that had been 'inherited from their forefathers' (968; cf. 969-972). 4

§32. In the same context, Aristophanes Clouds969-972, there is mention of a composer, called Phrynis, who is ridiculed for modernizing the old conventions of harmoniâ. This Phrynis belongs by our standards still to the Archaic period: he was a kitharôidos 'lyre [kitharâ] singer' from Mytilene who won first prize in a contest at the Panathenaia of 456 (or possibly 446). 1 He was primarily known for his virtuosity in the genre of the kitharôidikos nomos 'citharodic nome' (Athenaeus 638c). According to "Plutarch" On Music 1133b, he was the first to introduce the practice of modulating between harmoniai within a single composition (apparently a citharodic nome: ibid.; see also Pherecrates F 145.14- 18 Kock). Such testimony implies that the contemporaries of Phrynis, such as Pindar, did not yet compose songs that modulated between harmoniai. We have already seen a parallel on the level of meter: Pindar generally avoids modulating between the Aeolic and the Doric or dactylo-epitrite meters. Moreover, we have seen that the Doric and Aeolic meters of Pindar seem to be correlated respectively with self-references to the Dorian mode on one hand and to the Aeolian or the various "foreign" modes on the other. It appears that, to this extent, Pindar keeps his meters and his harmoniai distinct in any given composition, and that only on the level of dialect are the Aeolic and Doric elements accretively merged. Still the report that the Locrian mode, to be equated with the Aeolian (Cleonides, p. 198.13 Jan), had become obsolete after Pindar (Athenaeus 625e), combined with the fact that the concept of Hypodorian eventually replaced that of Aeolian (Athenaeus 624e, citing Heraclides Ponticus F 163 Wehrli), suggests that the distinction between the Aeolian and the Dorian modes had become obsolete after Pindar. The new nomenclature of Hypodorian suggests that the old Aeolian melodic tradition had become at best a residual subcategory of Dorian. 2

§33. Such shifts in classification might help explain what seem at first to be contradictions in the later testimony about the melodic pattern of various genres. 1 For example, there is a report in "Plutarch" On Music 1136f that the partheneia 'maiden-songs' of Alcman, Simonides, Bacchylides, and Pindar were composed in the melodic pattern of Dorian. 2 Yet, in the few fragments of attested partheneia collected as Pindar F 94a-104b SM, the metrical pattern is Aeolic, which suggests an Aeolian melodic pattern. If indeed such a pattern is perceived as Hypodorian, the term Dorian here may be viewed as inclusive of Aeolian. 3 An alternative explanation is that the partheneia of Pindar, like his epinîkia 'victory odes', may have been composed in either the Dorian or the Aeolian melodic patterns, and that the formulation offered in "Plutarch" On Music 1136f is overly restrictive. Still there seem to be clear signs of correlation between given genres and given melodic patterns, parallel to the well-known correlation between given genres and given metrical patterns. Thus, for example, Aristotle says explicitly that the Phrygian harmoniâ is natural to the dithurambos 'dithyramb' (Politics 1342b); he notes an anecdote about Philoxenus, who tried to compose a dithyramb in the Dorian harmoniâ and who could not help but fall back into the Phrygian (ibid.). 4 Further, we hear that the Hypodorian and the Hypophrygian, described as having relatively less melody than other patterns, are inappropriate for the choruses of tragedy ("Aristotle" Problems 19.48), and that the Hypodorian is the most suited of all harmoniai for solo kitharôidiâ by principal actors (ibid.). We may note the remark in Plato Republic 398e that thrênoi 'laments' are associated with such harmoniai as the Mixolydian and the Syntonolydian.

§34. Before we end this discussion of relationships between melodic systematization and the content of songs, it is important to consider yet another set of categories for melodic traditions, organized under the heading of genos 'genus'. Classification by way of genus was systematized by Aristoxenus. According to this system (Aristoxenus Harmonics 21.31- 27.14, 46.19-52.34) there is a fundamental unit of composition, the tetrachord, which consists of four notes and which is divided into three genera: (1) enharmonic, (2) chromatic, and (3) diatonic. The two outer notes of the tetrachord are constant for all three of the genera while the two inner ones are variable; the different locations of the inner notes within the tetrachord of four notes constitute the differences in genera. A tetrachord has different interval patterns according to its genus: in the enharmonic it is quarter-tone, quarter-tone, ditone (that is, 1/4, 1/4, 2); in the chromatic it is predominantly semitone, semitone, tone-and-a-half (1/2, 1/2, 1 1/2); in the diatonic it is semitone, tone, tone (1/2, 1, 1). 1 Although these categories themselves are late, 2 they clearly contain old patterns: it has been plausibly argued that the tetrachord, which is the basis of the three genera, is related to the system of tuning for the older four-string lyre. 3 Within the later system of tuning for the newer seven-string lyre, as reflected by the harmoniai, it is possible to find embedded the earlier system of tuning for the older four-string lyre, as reflected by the tetrachord, with its three categories of genus. 4 This diachronic hierarchy seems to be supported by the synchronic fact that "the tetrachord, whether it came high or low in the scale [of the harmoniâ], had primary status in the melody, it was a nucleus, while the notes outside the tetrachord derived their significance from their relation to it." 5

§35. We might say that the containers known as the three genera are new, but they contain three redistributed sets of old patterns. 1 Moreover, the actual distribution into containers reflects a relative ranking of features that are at least perceived as ranging from old-fashioned to innovative. The most old-fashioned features, clearly, fall into the category of the enharmonic genus. There are reports about later musicians who imitated Pindar and Simonides by deliberately avoiding the chromatic genus, thereby "sounding" enharmonic ("Plutarch" On Music 1137f; cf. 1145a). The enharmonic genus was considered typical of tragedy ("Plutarch" 1137de). 2 In the fourth century B.C., as we know from the testimony of Aristoxenus, the chromatic genus was the prevalent mode (Harmonics 23.9-22); 3 moreover, Aristoxenus specifies that it tended to displace the enharmonic genus, which was becoming obsolete (ibid.). 4

§36. When Aristoxenus discusses the order in which the genera came into being (Harmonics 19.17-29), he places the diatonic as the first and oldest on the grounds that it sounds more universal, as if the diatonic genus were the first system that would occur to human nature (prôton gar autou tou anthrôpou phusis prostunchanei ibid.); then comes the chromatic, and finally the enharmonic, on the grounds that human perception has the hardest time in getting accustomed to the enharmonic (ibid.). In "Plutarch" On Music 1134f-1135 there is a convergent formulation, attributed to Aristoxenus, according to which the mythical master of the aulos, Olympus the Phrygian, supposedly "invented" the enharmonic genus by experimenting with the diatonic genus and adjusting it; all music before Olympus was supposedly diatonic or chromatic (ibid.). According to Aristoxenus this "experimentation" of Olympus, "transforming" the diatonic genus into the enharmonic, led to a system of composition in the Dorian tonos ("Plutarch" 1135a). 1 The account goes on to admit that such a system, as attested in the compositions attributed to Olympus, cannot readily be classified under any one single genus, whether it be diatonic, chromatic, or even enharmonic, because the interval patterns reveal areas of nondifferentiation that do not correspond even to current enharmonic standards (On Music 1135ab). Thus the Archaic musical style of Olympus is to be considered a sort of early enharmonic (ibid.). This admission makes it clear that the enharmonic was in fact the basis for differentiation, and that the hierarchy in terms of myth has to be reversed in terms of the actual development of patterns. 2

§37. The hierarchy of myth, which is based on contemporary musical perceptions of what comes naturally, must be juxtaposed with actual contemporary musical trends. Despite the thought-pattern of myth, which insists on the invention of the chromatic out of the diatonic, it is generally agreed by present-day musicologists, on the basis of other indications in the ancient sources, that the diatonic genus superseded the chromatic as the prevailing musical style in post-Hellenistic times. 1 One such indication is the explicit testimony of Aristides Quintilianus, p. 16.10-15 (ed. Winnington-Ingram). Moreover, we have also seen the testimony of Aristoxenus to the effect that the chromatic genus tends to displace the enharmonic in his own time (Harmonics 23.9-22). Thus it would appear that the enharmonic is more Archaic in its interval patterns than the chromatic. Similarly, if indeed composers avoid the chromatic in order to imitate the older masters--and we have seen that Aristoxenus verifies this practice as a contemporary one--then the chromatic genus must surely be less Archaic than the enharmonic. We may infer then that the Archaic masters like Pindar and Simonides "sounded" more enharmonic than chromatic or anything else.

§38. From the standpoint of the Archaic masters of lyric, it may even be enough to describe their enharmonic melodic patterns as simply harmonic. 1 In setting up a chronological hierarchy of (en)harmonic as the oldest, followed by chromatic and then by diatonic, we may perhaps draw some inferences from the presence of the so-called puknon in the enharmonic and chromatic genera, as opposed to its absence in the diatonic. The word puknon refers to a pattern where the two lowest intervals of the tetrachord, when added together, are less than the remaining interval of the tetrachord. Thus in the chromatic genus 1/2 plus 1/2 is less than 1 1/2, and in the enharmonic genus 1/4 plus 1/4 is less than 2. By contrast in the diatonic genus 1/2 plus 1 is not less than 1. In other words the distinctive feature of enharmonic and chromatic genera is the consecutive sequence of two small intervals, the puknon, while the rest of the tetrachord is occupied by one single large interval. In addition there is an auxiliary rule according to which the puknon cannot be followed in traditional melody by an interval that is shorter than one tone. 2 The smaller intervals in the enharmonic and the chromatic, which are relatively older genera, seem to resemble more closely the "speech melody" of ancient Greek accentuation than the larger intervals of the diatonic, which is relatively younger. 3 The smaller intervals in the enharmonic and chromatic are the very factors that would make ancient Greek music sound foreign from the standpoint of the tempered tones and semitones of the modern Western musical traditions. 4

§39. If indeed the enharmonic tetrachord were once simply the harmonic, preceding any differentiations leading to the chromatic and thereafter to the diatonic, then we may look for the clearest traces of preserved affinities with the pitch accent system of the Greek language precisely in the enharmonic genus. On the authority of "Plutarch" On Music 1135ab and 1137ab (especially 1137b), we know that some of the oldest melodic patterns, attributed to Olympus and Terpander, were trichords, not tetrachords. 1 It can even be argued, on the basis of such testimony, that certain types of tetrachords developed out of trichords. 2 The affinity of the (en)harmonic tradition with trichords, that is, three-note systems, suggests even closer links between the interval patterns of the enharmonic genus and the interval patterns of pitch accentuation in the Greek language. 3

§40. Rounding out this survey of the chronologically overlapping melodic systems of nomoi, harmoniai, tonoi, and genera, we may say that they all reflect in various degrees an ongoing process of mutual assimilation and systematization, to which I have applied the overall concept of Panhellenization. The present list is hardly exhaustive. For example, another important factor that contributed to the systematization of melodic traditions was the innovative interaction of conventions in accompaniment by lyre and reed (aulos), 1 and the actual conflation of distinct melodic patterns associated with the lyre, the aulos, and the voice, as pioneered by such antecedents of Pindar as Lasus of Hermione (Theon of Smyrna, p. 59.4 Hiller, Suda s.v.; "Plutarch" On Music 1141c). 2 This particular figure is reputed to have been a teacher of Pindar (scholia to Pindar Olympian1, p. 4.13-15 Drachmann) and a rival of Simonides (cf. Aristophanes Wasps1409-1411). Given that the provenience of Lasus is the Dorian city of Hermione, we may note again the specific reference to a harmoniâ described as Aeolian in Lasus PMG 702.3; 3 there is a comparable self-reference to an Aeolian song and dance in Pindar Olympian1.102 (aiolêidi molpai), the same composition in which the voice of the poet, at the beginning, asks that the Dorian lyre be handed over to him dôrian...phorminga 17). 4

§41. With this survey of melodic traditions in Greek song serving as background, we may pursue further the notion of a canon in our ongoing discussion of Panhellenization in poetry and song. What I have proposed is that the formation of a canon in song--which we can also call lyric poetry--started relatively later than the formation of a canon in nonlyric poetry proper. Once the Panhellenic breakthrough of song did happen, however, its transmission would have been facilitated to rival that of poetry not only because of the mnemonic utility of melody but also because of the relative brevity of song as opposed to the potentially open-ended length of poetry. In any inherited distinction between SONG and SPEECH, we would expect that the pressures of regularization in SONG would tend to delimit the length of production in contrast with the potentially open-ended length of speaking everyday SPEECH. So also in any differentiation of SONG into song vs. poetry, we would expect that song would be more clearly delimited in length of production by contrast with the potentially open-ended length of poetry in its imitation of SPEECH. 1

§42. With these considerations, let us examine the social context of performance. If indeed the transmission of Panhellenized song coexists with that of poetry, it stands to reason that the professional performer of such song, the kitharôidos 'lyre [kitharâ] singer' and the aulôidos 'reed [aulos] singer', would be valued on a scale comparable to that of the professional performer of poetry, the rhapsôidos 'rhapsode'. In fact the epigraphical evidence shows that rhapsôidoi, kitharôidoi, and aulôidoi, as they perform in competition at festivals, are awarded comparable sums for their prizes. 1

§43. In the following chapters, we have occasion to observe a recurrent pattern where the composer/performer of song or lyric poetry is eventually differentiated into a mythical protocomposer on the one hand and a contemporary professional performer, the kitharôidos or aulôidos on the other. But this pattern is just one of many other possible patterns of evolution. The category of lyric poetry includes performance not only by a single professional or nonprofessional performer but also by a nonprofessional group of specially selected natives of the polis, the khoros 'chorus', who both sang and danced the song. 1 In Pindar's time, as we see later, the nonprofessional chorus would be performing, on commissioned occasions, songs composed by professional poets of Panhellenic prestige, such as Pindar. 2 As we also see later, the institution of the chorus plays its own role in the emerging concept of authorship in Archaic and Classical Greece. 3

§44. Of all the composers of song, or lyric poets, I single out Pindar as the focus of our attention. Along with his near-contemporary, Bacchylides, Pindar is the latest and the last in the canon of lyric poets inherited by the Alexandrian editors. The last securely datable poem of Pindar, Pythian8, was composed for performance in 446 B.C. (for Bacchylides, the last datable compositions are Odes 6 and 7, performed in 452 B.C.). With this date of 446 we have an imprecise but revealing terminus in the history of ancient Greek poetry.

§45. A striking feature of this terminus is the fact that the canon of lyric poetry excludes poets who flourished in the second half of the fifth century or thereafter. This fact will in due course be linked with the argument that the canon of lyric poetry results from patterns of Panhellenization in oral traditions of song. Although there is ample evidence for the existence of poets who composed song in the second half of the fifth century and thereafter--the most prominent of whom are Timotheus of Miletus and Philoxenus of Cythera-- 1 there is also evidence that their song was a medium that had evolved beyond the lyric poetry represented by Pindar and the other canonical lyric poets. In particular the differences can be seen in the genres known as the citharodic nome and the dithyramb. 2 I quote from a brief summary of the situation: 3

One point about the development of lyric poetry in the latter part of the fifth both clear and relevant to the question [of why this poetry was excluded from the canon]: two kinds of lyric poetry, the [citharodic] nome and the dithyramb, began to dominate nondramatic poetic composition. Both were different from all the earlier lyric types, including the earlier nome and dithyramb, 4 in signficiant respects: they were nonstanzaic, the relative importance of music to words suddenly and greatly increased, and their affinity to drama was recognized; Aristotle groups them with tragedy and comedy in his classification of the mimetic arts at the beginning of the Poetics. It may be that the Alexandrian critics did not consider this new poetry, which continued dominant in the fourth century, to be of the same genre as lyric poetry (nearly all of which was stanzaic), 5 and for this reason excluded Timotheus, Philoxenus of Cythera, Cinesias, and the other writers of dithyrambs and nomes.

§46. I propose a different reason for the exclusion of these poets in the canon of the Alexandrians: the likes of Timotheus and Philoxenus, unlike the earlier masters of lyric, were already being excluded from the canon of traditional Athenian education in the "Classics," mainly on the grounds that the innovative virtuosity characteristic of such poets, and of the new genres that they represented, tended to restrict their oeuvre to performance by professionals and to defy the traditions of liberal education for nonprofessionals, that is, for the future citizens of the polis. 1

§47. With regard to this crucial era of the second half of the fifth century and thereafter, it is important to note that, alongside the emergence of new media of song as represented by the "new" nomos 'nome' and the "new" dithurambos 'dithyramb', 1 there is a concurrent obsolescence of the old media of song as represented by lyric poetry proper. In fact the traditions of composition in lyric poetry, as once practiced by Pindar, seem to be becoming extinct in this era of the "new" nome and the "new" dithyramb. Such a trend of extinction is most evident from the standpoint of traditions in performance. For example, in Eupolis F 366 Kock (paraphrased by Athenaeus 3a) the complaint is made that the songs of Pindar have for some time been covered over in silence, ignored by the audiences of the day (Eupolis was a contemporary of Aristophanes). In Eupolis F 139 Kock (quoted by Athenaeus 638e) we see a parallel theme: the speaker is complaining that the songs of Stesichorus, Alcman, and Simonides are considered out of date by contemporary audiences, who prefer the "modern" poetry of the likes of Gnesippus. The latter poet is ridiculed for his modernisms also by Cratinus (F 97 Kock), a pioneer of Old Comedy, who was an older contemporary of Eupolis and Aristophanes (Athenaeus 638ef).

§48. In Aristophanes Clouds1353-1358, the figure of Strepsiades is taking an old- fashioned stance in berating his son Pheidippides, a new convert to the school of Socrates, representing modernist trends of education that have eroded the traditions of old-fashioned liberal education in the "Classics." 1 At a symposium Pheidippides refuses a request to take up the lyre and sing a famous lyric composition by Simonides (Aristophanes Clouds1355-1356). The composition was an epinîkion, that is, a victory ode (Simonides PMG 507). Technically Pheidippides is refusing here to perform a skolion. This word skolion, as used in the time of Aristophanes, is an appropriate general designation for the performance, self-accompanied on the lyre, of compositions by the great lyric masters. 2 A notable example of this usage of skolion is Aristophanes F 223 Kock, with reference to the performing of compositions by Alcaeus and Anacreon, which are here called skolia. 3 Such a general sense of skolion is lost later as the word becomes progressively restricted in meaning (as we see from Athenaeus 694f-695f). 4 The performances at symposia of the great lyric masters correspond to the monodic medium of the kitharôidos. 5 To engage in these performances was an old-fashioned convention at symposia, as we can see from such references as the present passage from the Clouds of Aristophanes, the scholia to Aristophanes Wasps1222, and Eupolis F 139 Kock (by way of Athenaeus 638e). 6 According to "Plutarch" On Music 1140f, Pindar attributed the "invention" of the skolion to Terpander, who as we have seen is also the traditional "inventor" of the system of melodies used in kitharôidiâ 'lyre singing' (On Music 1132d). 7

§49. As we see from the passage in the Clouds, Pheidippides ridicules the singing of Simonides' lyric poetry specifically because he considers it something that is passé (Clouds 1357-1358). 1 There is still, however, an important last stand of old-fashioned lyric poetry in the second half of the fifth century and thereafter, and it is to be found in the choral traditions of Athenian tragedy and comedy. The poets of Old Comedy, as we have seen, even ridicule the new poetry that purports to displace the old poetry. On another level Old Comedy could also ridicule the old poetry of lyric traditions, as in the parody of Pindar F 105 SM in Aristophanes Birds926-930, 941-945(note too the adjacent reference to Simonides in Birds 917-919).

§50. The point remains that the old traditions of lyric are obsolescent by the time of Aristophanes, and in fact the Birds, presented in 414 B.C., is the last attested comedy of Aristophanes that mentions or parodies the compositions of Pindar. 1 We may note that Pindar F 105 SM is taken from a composition, known as a huporkhêma, intended for the tyrant Hieron of Syracuse. 2 The only other Aristophanic reference to a Pindaric composition where we know the identity of Pindar's intended audience is in Acharnians 637/639 and in Knights 1329/1323, both referring to Pindar F 76 SM, a passage from a famous dithurambos 'dithyramb' composed expressly for the glorification of Athens. Both these Aristophanic references to recognizable Pindaric passages, one the huporkhêma for Hieron and the other the dithurambos for the Athenians, focus on the beginning of a Pindaric composition. It seems that the allusion is being made to the most famous parts of famous compositions. Besides these two cases there seems to be only one more where we can be reasonably certain that the reference is to a well-known passage of Pindar: Knights 1264-1266, alluding to a prosodion 'processional song' (F 89a SM). 3 It has been observed that these three Aristophanic references to three passages apparently familiar to an Athenian audience can give us an indication of the kind of repertoire that was being taught to young Athenians in the years roughly between 450 and 420 B.C. 4 This repertoire is decidedly limited in scope, which converges with what we have observed about the canon of nine "Classics" of lyric poetry: given the vast variety of traditions that they represent, this canon is a relatively small corpus.

§51. It seems likely that the evolving predominance of Athenian theater as a poetic medium played a major role in the obsolescence of lyric poetry in other media and by extension in other genres. From Plato's writings we hear of complaints about theatrokratiâ (Laws 701a) 1 and about the intoxication of pleasure in the poetry of theater (Laws 700d), leading to 'transgressions' of genre (paranomia: Laws 700e). 2 To be contrasted are the good old days, as in the era of the Persian Wars (Laws 698b-700a), 3 when there were still distinct eidê 'types' and skhêmata 'figures' of song and dance (Laws 700a), 4 five of which are specified as examples: humnos 'hymn', thrênos 'lament', paiân 'paean', dithurambos 'dithyramb', and kitharôidikos nomos 'citharodic nome' (Laws 700b). These genres, 5 as well as other genres left unspecified (ibid.), 6 are the structurally distinct aspects of mousikê 'music' (that is, for all practical purposes, lyric poetry), parallel to the structurally distinct aspects of aristokratiâ in Plato's good old Athenian society (Laws 701a). 7 In contrast the progressive leveling by Athenian theater of generic distinctions in lyric poetry is for Plato parallel to the leveling by Athenian democracy of class distinctions in society. 8 What I have been describing as an infusion of lyric genres into theater, and their concomitant atrophy elsewhere, is seen by Plato as an illegitimate mixing of genres (Laws 700d), a degeneration into a superseding genre of lyric traditions in Athenian drama (ibid.). In contrast, as Svenbro points out, 9 Athenian drama is seen by Aristotle not as the product of degeneration but rather as a teleological organic development in the evolution of poetic traditions (Poetics 1449a14-15). 10

§52. From either point of view, the lyric poetry of Athenian theater would be considered the final productive phase of a medium that had otherwise become unproductive already by the second half of the fifth century. In Aristophanes Clouds1361-1376, as we have seen, Pheidippides refuses to take up the lyre and sing a lyric composition by Simonides, on the grounds that the singing of lyric compositions at symposia is passé. 1 It is clear that Pheidippides is not well-versed in the art form of this kind of performance. 2 Then he is asked at least to perform something from the compositions of Aeschylus, while holding a branch of myrtle (Clouds 1365). 3 The aulos, not the lyre, serves as the medium of accompaniment for the lyric compositions of Athenian Drama. 4 Singing to the lyre implies potential self-accompaniment, whereas singing to the aulos does not. Thus a lower degree of education is required for performing in the chorus of an Aeschylean tragedy or for reperforming at a symposium selections from the choral songs of such a tragedy. 5 Even this kind of performance is refused by Pheidippides, who elects to recite a passage from a speech in Euripides (Clouds 1371). The word rhêsis 'speech' (ibid.) makes it clear that the modern Pheidippides opts for a medium that is devoid of the lyric element. 6

§53. Plato's portrait of nostalgia for those earlier days when lyric poetry had not yet been absorbed and ultimately usurped by Athenian theater returns us to the era of Pindar, last in the canon of lyric poets. Let us consider the genres in which Pindar composed. It is best to begin with the inventory of an Alexandrian edition of Pindar as reported in the Vita Ambrosiana, according to which Pindar's compositions are subdivided into seventeen books corresponding to specific genres of lyric poetry. 1 We are struck by the fact that of the distinct genres of lyric poetry mentioned in Plato's partial list, namely the humnos, the thrênos, the paiân, the dithurambos, and the citharodic nomos (again Laws 700b), all but the last one are also represented in the Vita Ambrosiana inventory of Pindar's poems. The inventory of seventeen books attributed to Pindar contains ten distinct genres (an asterisk marks those genres that correspond to Plato's list):

5 It is difficult to be certain whether such an editorial organization of Pindar's poems goes further in time than the Alexandrian era--back to the time of Plato, for example. 6 But we do know for certain that Plato was familiar enough with Pindar's poems to refer to them at least sixteen times in the attested Platonic corpus. 7

§54. It remains to ask how exactly these references came about. One readily available explanation is that Plato was citing from a hypothetical edition of Pindar that was circulating in Athens. 1 Most likely such an edition would have been a school text going back to an earlier time when youths had still been well-educated in the actual performance of old-fashioned lyric compositions. 2 By Plato's time, however, it was becoming less and less likely that performers of kitharôidiâ or aulôidiâ, especially amateurs, could still have had in their repertoires selections from Pindar and other grand masters of lyric poetry. I quote from the summary of a musicologist: 3

The classic Athenian comedy had been made for a society which talked music as it talked politics or war. But in Aristophanes' post-war plays, a shrunken chorus gives us only a last flash or two of his musical parody; and his successors substituted entr'actes by variety artists. The Alexandrian era still has excellent stage gossip on performers, but a first-hand judgement on the style or quality of music is hardly to be found after the fourth century. Aristotle already prefers received opinions. His master Plato and his pupil Aristoxenus are the last who speak to us with the authority of musical understanding.

§55. It must be kept in mind that the contemporary Athenian traditions of composing and performing lyric poetry had already outgrown, well before Plato's time, the traditions represented by Pindar. 1 I do not rule out the possibility that some of the better schools, even in Plato's time, insisted on extensive memorization of the libretti of the lyric masters, but it is clear that most schools limited their requirements to a small repertoire of selected passages to be memorized for recitation (cf. Plato Laws 810e). 2 As for professional musicians, there is evidence that they still had access, even at as late as the era of Aristotle, to the melodic traditions of Archaic masters like Pindar. I refer to a revealing report in "Plutarch" On Music 1142b attributed to the music theorist Aristoxenus, pupil of Aristotle: according to Aristoxenus, the son of a professional musician from Tarentum, 3 there was a composer, one of his own contemporaries, who in the course of his career reverted from the musical idiom of Timotheus and Philoxenus to that of the Archaic poets, among whom Pindar is mentioned first. Clearly the musical tradition of Pindar had survived until then. Still there would have been no chance for any major ongoing recomposition of Pindar's lyric poetry through performance in that contemporary traditions of choral lyric composition would have been sufficiently differentiated from Pindar's old-fashioned traditions. In fact the very traditions of performing Pindar's compositions had become obsolete.

§56. In Pindar's own time, by contrast, his compositions could still be readily reperformed, ordinarily not by a chorus 1 but by individuals at symposia, simply as "Classics." As we have seen, Old Comedy represents the mode of reperforming an epinician composition of Simonides, at the symposium, in the format of a solo rendition with self-accompaniment to the lyre (Aristophanes Clouds 1355-1358). By implication there was a time when a choral composition, such as we see in the songs of Simonides or his contemporary Pindar, could be actively converted into a solo performance. Such interchangeability between choral and solo contexts is a clear indication, however indirect, of the flexibility of the choral lyric form as a still-living tradition in the era of Pindar. 2 This flexibility presupposes a solid education of nonprofessionals in the art of professionals known as kitharôidoi 'lyre singers'. 3 In the later era of Timotheus and Philoxenus, as we have also seen, such flexibility had broken down.

§57. As "Classics" the compositions of a Pindar could be reperformed at will in the "good old days," but they would have to be grounded in an awareness of the situations and ideologies in which Pindar was commissioned to give public poetic testimony. These situations and ideologies, or occasions, may strike us at first as potential obstacles to the Panhellenization of such compositions. The very occasions of this lyric poetry, however, were of Panhellenic importance, with an impact lasting in prestige. Each of Pindar's compositions was originally commissioned for a specific occasion, to be performed ostensibly by a chorus assembled and trained for that one original occasion. But the prestige of such an occasion was meant to reverberate indefinitely in time and space.

§58. This transcendent occasionality of Pindar's lyric poetry is most evident in the only genre of his compositions to survive almost intact: the four books of epinîkia 'victory odes'. These victory odes or epinicians were commissioned as ad hoc choral performances in celebration of victories won by athletes at the Panhellenic Games, and the four books of Pindar's epinicians match the four most prestigious Panhellenic athletic festivals: the Olympian, Pythian, Isthmian, and Nemean Games. Though each of Pindar's victory odes was an occasional composition, centering on a single performance, each containing details grounded in the historical realities of the time and place of performance, still each of these victory odes aimed at translating its occasion into a Panhellenic event, a thing of beauty that could be replayed by and for all Hellenes for all time to come.

§59. The occasion of the poet's victory ode not only conferred Panhellenic prestige: it also received it since the athlete's victory itself was a Panhellenic event, the prestige of which depended on the relative Panhellenic prestige of the given games in which the victory occurred. Just as the concept of Panhellenism is relative in terms of poetry and song, 1 so also in terms of athletics: the older the athletic festival, the more Panhellenic was its prestige. Thus the Olympics of Elis (traditional founding date of 776 B.C.) were more Panhellenic than the Pythians of Delphi (582 B.C.), the Pythians more than the Isthmians near Corinth (581 B.C), the Isthmians more than the Nemeans near Argos (573 B.C.). The list could be extended, with such runner-up festivals as the Panathenaia of Athens (with athletic competitions instituted at 566/5 B.C.). 2

§60. It is clear that there would have been no rationale for recommissioning a chorus to reperform such a composition since the original occasion would have been archetypal from the standpoint of the lyric poetry. To put it another way, the original occasion would have been gone forever from the standpoint of us outsiders who are critics of this poetry. For us, any reperformance of such a composition in, say, Aristophanes' time seems at first sight to be just a performance of a canonical poet. From the internal point of view represented by the poetic tradition, however, a reperformance in an old-fashioned symposium is a remaking of an original poetic event. There is no chorus, no chorus-leader present; instead a soloist performer must reconstitute their roles, while accompanying himself on the lyre. 1 The example of the young Pheidippides in the Clouds of Aristophanes, however, suggests an incipient failure of liberal education, even by the time of Old Comedy, to produce anyone to take up such a challenge. 2


§1n1. There is a useful survey by Comotti 1979.15-25 (see especially p. 18). For an insight into the character of local melodic patterns, consider the expression dêmôdê mousikên 'songmaking of the locale [dêmos]' at Plato Phaedo 61a, in the context of the discussion by Brisson 1982.55-56. In the discussion that follows, I use the same notions of pitch and melody that I have set up in the working definitions at Ch.1§33n1.

§2n1. See Pfeiffer 1968.205. On the concept of canon as used here, see Ch.2§25.

§2n2. Cf. Kirkwood 1974.3. The dating of Corinna as roughly contemporaneous has been a matter of controversy, discussed impartially by Page 1953.68-84. If indeed this lyric poetry is Archaic (cf. Gerber 1970.394-395), like the canon of the nine lyric poets, the question remains: why was Corinna ignored in the canon? I agree with the reason offered by Page: by the time that we reach the era of Alexandrian scholarship, Corinna would have "long ceased to rank among hoi prattomenoi, the poets whose works survived in universal and unbroken circulation (p. 69)." If then we suppose that Corinna is an Archaic poet, I would further suppose that the transmission of her lyric poetry happened not on a Panhellenic but on a more localized Boeotian level. (On the localized nature of the compositions attributed to Corinna, I cite the useful discussion of Davison 1968.300-302.) The absence of an ultimately Panhellenic transmission could perhaps be connected with the possible absence of an Athenian transmission. We would not expect Corinna, as a local poet of Boeotia, to be a "Classic" in the paideiâ 'education' of Athenian youths, on which see further at Ch.3§31. On the likelihood that the canon inherited by the Alexandrian scholars reflects primarily the traditions of Athenian paideiâ, see in general Ch.13. In any case it is important to distinguish between the selective canon inherited by the Alexandrian scholars and the nonselective repertory of works housed in their Museum: see Ch.2§25n6. For the Alexandrian scholars, exclusion of an author from the canon does not preclude an active interest in that author, even as a model for imitation. As Zetzel 1983.99 points out, "one of the most striking characteristics of Alexandrian poetry was its tendency to avoid the major classical genres and even to elevate to literary status forms probably not recognized previously as literature at all." This way the classifier sets himself apart as one who is beyond classification (Zetzel ibid. cf. also Rossi 1971.83- 86).

§2n3. This is not to say that we should expect the patterns of recomposition in poetry and song to be neatly parallel in every way.

§3n1. Wilamowitz-Moellendorff 1900.63-71, especially p. 65.

§3n2. Ibid. See also Page 1953.68, who doubts that "any ancient lyrical poet whose works were in circulation up to the Alexandrian era was omitted by the Alexandrian editors from their collection." For bibliography on reactions to this view, see Pfeiffer 1968.205n4.

§3n3. See Ch.2§25n6.

§3n4. What Zetzel 1983.98 writes about the three canonical iambic poets (Archilochus, Hipponax, and Semonides of Amorgos: cf. Quintilian Institutio oratoria 10.1.59) can be applied also to the nine canonical lyric poets: "it was a selection and not, as has sometimes been thought, a complete list of all early authors--it was explicitly a list of earlier authors, stopping with the fourth century B.C." (emphasis mine). I would also compare the canonical notion of Seven Sages, datable all the way back to the Archaic period, where the number seven is a constant while the actual membership is variable, in that we witness attestations of different "members" at different times and different places (see Ch.8§44n2).

§3n5. As we have already observed in the case of Homer, the patterns of attribution to such a given "author" become progressively more exclusive as we move forward in time: see Ch.1§7n1. Thus the Epic Cycle, for example, becomes reassigned to distinct poets, whose canonical status is considered inferior to that of Homer: see Ch.2§38 and following. On Orpheus and Musaeus, presented as if they were earlier than Hesiod and Homer, cf. Aristophanes Frogs1032-1035 (further references at Ch.8§2n8). By contrast Herodotus 2.53.3 argues against the notion that there might be such poets older than Homer and Hesiod (cf. Ch.8§2n8). It is clear from the context that Herodotus places Homer and Hesiod as the earliest because he deems their canonical status the very highest.

§3n6. Ch.2 above. By the time of the second century A.D., the entire surviving corpus of Greek lyric in the broader sense, including iambic and elegiac poetry, was ranked as a nonpractical aspect of education in the Classics (cf. Dio of Prusa Orations 18.8, p. 478 R).

§4n1. Cf. Ch.2§25.

§4n2. See Ch.2§13, Ch.2§24n6.

§5n1. The element of dancing is made explicit in this context: orchêsesi (Laws 764e). In Ch.12, we shall consider various patterns of differentiation between singers and dancers in the context of the khoros.

§5n2. The valuable testimony in Plato Laws 764c-765b on the given of institutional differentiation, in terms of public performance, between the monodic and choral media is too readily dismissed by commentators. I stress also the crucial testimony in Aristophanes Frogs1329-1364, with parodies of the monodic technique (note monôidiâ at 1330) as distinct from the choral lyric technique, parodied in 1248-1329.

§5n3. Cf. also Plato Laws 659ab, as cited at Ch.2§25n9.

§5n4. This point is stressed by Kirkwood 1974.10, 212n16 (with specific reference to Plato Laws 764c-765b); cf. also Harvey 1955.159n3 and Färber 1936.16. Davies 1988.61 observes that "it is dangerously misleading to talk of choral or monodic poets."

§5n5. Cf. Ch.12§2 and following.

§5n6. Ibid.

§6n1. See Ch.1§63n1.

§7n1. For a useful survey, see Barker 1984.14. References to string instruments in Archaic Greek poetry tend not to differentiate kitharâ from, for example, phorminx: see Barker, p. 25n19, who also takes note of the later taxonomy of differentiation, as in Aristotle Politics 1341a. Cf. also Maas and Snyder 1989.5, 202.

§7n2. Survey in Barker, p. 15.

§7n3. Also I should caution against any general assumptions of parity between instrumental and vocal intervals: for a cross-cultural survey, see Nettl 1956.50.

§8n1. As the discussion that follows makes clear, I interpret this name as generic, in line with the programmatic use of the verb terpô 'give pleasure' in poetry to describe the effects of poetry, as in the case of Phemios at Odyssey i 347. The name Terpandros is analogous to the expressive patronymic Terpiadês, derived from verb terpô 'give pleasure', as applied to the singer Phemios at Odyssey xxii 376. More details in N 1979.17 §4n1.

§8n2. Athenaeus also reports a variant tradition according to which Terpander was a contemporary of Lycurgus (635f, on the authority of Hieronymus On Kitharôidoi). Lycurgus is credited with being one of the founders of the first numbered instance of the Olympic Games, that is, at 776 B.C. (Athenaeus ibid.).

§8n3. In this connection, "Plutarch" On Music 1133d mentions one Periclitus, who was at some undetermined later point likewise a winner at the Spartan Feast of Karneia in the contest of singing to the accompaniment of the kitharâ: like Terpander, he was a kitharôidos 'lyre singer' from Lesbos, and after his death the continuous tradition at Lesbos of singing to the kitharâ supposedly came to an end (ibid.). On the preeminence of kitharôidoi from Lesbos, see also Aristotle F 545 Rose and Hesychius s.v. meta lesbion ôidon.

§9n1. See the commentary of Barker 1984.96n16; also p. 251. The list of seven nomoi in "Plutarch" On Music 1132d omits the Orthios, present in the list of eight given by Pollux 4.65. In Suda s.v. orthios nomos, it is specified that there are seven nomoi; the account goes on to mention the Orthios along with the Trokhaios. In Suda s.v. nomos, the Tetraoidios and the Oxus are mentioned. Perhaps the number seven may have been a constant within some traditions, whereas the constituents were variable over the course of time. In Timotheus Persians (PMG 791) 225, there is a reference to ten ôidai 'melodies' of Terpander.

§9n2. On the appropriateness of the term genre here, see Svenbro 1984.225 and n135. See also Pfeiffer 1968.184 on eidos.

§9n3. On the "invention" of the dithyramb in Corinth, see also Pindar Olympian13.17-19.

§10n1. This reference in Herodotus 1.24.5 to the lyric performance of Arion as a nomos suggests that the earlier meaning of this word was broad enough in scope to designate simply a lyric composition that followed a set mode or melodic pattern, in this case specified as orthios 'shrill'. There is another reference to orthios nomos in Aristophanes Knights 1279. As for the story about the attempt by greedy sailors to rob Arion of his great wealth (Herodotus 1.24.1-2), we may compare the Homeric Hymn [7] to Dionysus, where Dionysus is abducted by pirates. This thematic connection between Arion and Dionysus is parallel to the connection between Arion and the dithyramb (again Herodotus 1.23), which is associated with the cult of Dionysus (e.g., Archilochus F 120 W).

§11n1. Cf. Comotti ibid. Cf. also Aristophanes Birds745 and the commentary of Fraenkel 1962.209-212.

§11n2. Chantraine DELG 742-743.

§11n3. On the specific parallelism between the hierarchical distribution of intervals within a melody and the hierarchical distribution of sacrificial meat within a community, see Svenbro 1984 and N 1985b.

§12n1. Cf. Ch.1§40n1.

§12n2. The phraseology of song, from the synchronic standpoint of Greek metrics, is organized along the lines of a stylized syntactical unit, the colon; cf. Appendix §1.

§12n3. On the basis of the self-references, I infer that this extension is not just a matter of metaphor: the mimesis of bird song seems to be part of an actual musical tradition. Cf. Nettl 1964.284 on a Shawnee song tradition imitating the call of the turkey, where the sound patterns "still fit into the musical structure."

§13n1. In this source, the nomoi of Terpander are understood anachronistically as equivalent to the nomoi of Timotheus of Miletus, a virtuoso composer of the late fifth century, who is said to have composed his earliest nomoi in dactylic hexameters: "Plutarch" On Music 1132e. At 1132de (see Barker 1984.209n25) the source infers that Terpander too composed primarily hexameters (though it would be more accurate to say, on the basis of Terpander PMG 697, that Terpander composed in meters related to the hexameter: Gentili and Giannini 1977.35-36). Timotheus was famed for a performance of a composition of his that happens to be a kitharôidikos nomos 'citharodic nome', known as the Persians, at the Feast of the Panathenaia at Athens, around 408 B.C. The composition is attested as Timotheus PMG 788-791, and we do indeed find a prominent deployment of dactylic hexameters alongside various lyric meters, even at the beginning of the song (PMG 788). In his Persians Timotheus overtly refers to Terpander as his predecessor (PMG 791.225-236). For more on the Persians of Timotheus, see in general the interesting discussion of Herington, pp. 151-160, who has enough information to calculate even the running time for the performance of the complete composition (thirty-five or forty minutes: p. 275n25).

§13n2. I should stress that the values within any scale at this particular stage of development in Archaic Greek song have to be considered in terms of relative rather than absolute pitch.

§14n1. There is a summary of the evidence in Wegner 1968.16, with helpful observations about the nonsignificance of occasional deviations from the number of four. I find it unnecessary to adopt, however, Wegner's restricted nomenclature of phorminx for the four-string and kitharâ for the seven-string instrument. On the still earlier use of seven- and eight-string lyres in the Minoan and Mycenaean periods, see Wegner, pp. 26-27. Also, I disagree with the theory of West 1981 that Homeric poetry was sung to a four-string instrument: the reasons for my disagreement have to do with the diachronic skewing of Homeric self-references, as discussed in Ch.1§9.

§14n2. West 1981.120. For a particularly early representation of a seven-string lyre, on a seventh-century Greek sherd found at Smyrna, see Boardman 1980.97-98, with illustration. For a comprehensive survey of evidence concerning ancient Greek stringed instruments, see now Maas and Snyder 1989; at pp. 27-28 and 203, they argue that the post-Mycenaean iconographic attestations of four-string instruments may be a matter of iconographic convention rather than reality, and that seven-string instruments may have been the norm even in the post- Mycenaean period. From the standpoint of comparative ethnomusicology, however, it is logical to expect the diachronic sequence of morphological development in stringed instruments to proceed from the four- to the seven-string configuration, though there is no reason to rule out even the coexistence of four- and seven-string instruments in any given era.

§14n3. Cf. Barker 1984.49: "A four-stringed instrument, perhaps tuned to a pair of fourths separated by a whole tone, does little to fix a clear-cut form of scale. A seven-stringed instrument must do so, though later writers were in doubt as to exactly which notes Terpander's scale incorporated." The interval of a fourth, known as a tetrachord, is the basis of formulating the three categories of genus, to be discussed below.

§15n1. Schwyzer I 385. The variant of this Aeolic form kithariW is kitharê = kitharâ (e.g., Herodotus 1.24.5, referring to the lyre of Arion), which is not even attested in the Iliad and Odyssey. Cf. Shelmerdine 1981.41n73.

§16n1. See also the commentary on "Plutarch" On Music 1133e in Barker 1984.212n51.

§16n2. "Plutarch" On Music 1133d marks a transition from the discussion of nomoi as sung to the accompaniment of lyre or aulos: now the subject shifts to nomoi for solo aulos. It is made clear that Olympus is credited with inventions of both kinds of nomoi, those for solo aulos and those for voice accompanied by the aulos (1133e). The distinction between an earlier and a later Olympus (1133de) reflects an attempt to resolve conflicting relative chronologies. On the reference to a nomos of Olympus in Aristophanes Knights8-10, see Bowie 1986.24.

§16n3. On the myth of Marsyas, cf. Plato Symposium 215bc.

§16n4. For more on the Phrygian connection of Olympus and Marsyas, see Barker 1984.210n32.

§16n5. The text does not say that Clonas actually originated aulôidiâ, so that the emendation of aulêtikôn for aulôidikôn at the end of "Plutarch" On Music 1132f, entertained as a possibility by Barker 1984.210n35, seems to me unnecessary. If we keep the text as it is, the source is giving the following relative chronology: Olympus/Terpander/Clonas. Our source claims that Clonas composed elegiacs and hexameters ("Plutarch" On Music1132c). We have already seen a parallel claim, that Terpander composed hexameters (Ch.3§13n1). Also, our source gives two variant traditions about the provenience of Clonas: he is a Tegean according to the Arcadians, a Theban according to the Boeotians ("Plutarch" 1133a). Also named in this same context is one Polymnestus of Colophon (ibid.). We know that Polymnestus was mentioned in the songs of Alcman and Pindar ("Plutarch" 1133a). A Pindaric mention of Polymnestus is quoted by Strabo (14.1.28 C643: Pindar F 188 SM).

§19n1. Barker 1984.130n18.

§19n2. Barker, p. 164.

§19n3. Barker ibid., with a survey of references. This working definition corresponds to my understanding of mode, as discussed in Ch.1§63n5. In modern Greek folk music there is an analogy in the notion of dromos 'road, mode', which is parallel to Arabic maqâm 'mode' and Turkish makam 'mode': see Beaton 1980b, especially p. 8, emphasizing the independence of the systems of dromoi from the Byzantine Oktoechos. On the difference between harmoniai and tonoi, Solomon 1984.249 notes: "Aristoxenus no doubt had to squeeze some of the intervallic leaps used originally in the native, tribal harmoniai and then in the tonoi into or out of the great system, but such is the universally compromising force of standardization. The difference between our own tempered and nontempered systems provides somewhat of a parallel."

§19n4. Barker ibid.

§21n1. Cf. West 1981.126.

§21n2. The dialectal subdivisions correspond to political subdivisions as well: cf. the narrative master plan of subdividing the notion of "Hellenes" into Ionians, Aeolians, and Dorians in Herodotus 1.6.2 et passim.

§21n3. See Ch.3§9n1. In this connection, R. Hamilton points out to me the fact that we know of no nomos that is Dorian. Again I invoke Kurylowicz's fourth law of analogy," discussed at Intro. §13.

§23n1. Cf. Ch.3§8.

§23n2. Survey in Palmer 1980.119-130; a basic work in this regard is Forssman 1966. Further discussion at Ch.14§9.

§23n3. Palmer, pp. 123-127.

§23n4. The 'Aeolic man' need not be Pindar (as claimed in the apparatus of Snell in SM, p. 130); it could be Terpander himself, as a reputed founder of the choral lyric medium.

§23n5. The Doric affinities of the dactylo-epitrite are best illustrated by the metrical heritage of Stesichorus, as discussed at Appendix §21.

§23n6. Cf. Ch.1§56.

§24n1. On the evidence for Ionic in Lesbian poetic diction, see the summary in Bowie 1981.136. In the corpus of Sappho and Alcaeus there is probably more of an Ionic element than meets the eye; such an element can easily become blurred by the efforts of Alexandrian scholars in re-aeolicizing the transmitted text (on which topic see also Palmer 1980.115-116).

§25n1. Cf. Most 1985.100n26.

§25n2. The latter example is cited by Most ibid.

§26n1. Note that Sappho's own words (F 106 V) acknowledge the primacy of Terpander, the 'singer from Lesbos'.

§26n2. For a remarkable anecdote that pictures Euripides singing mixoludisti 'in the Mixolydian manner' to the members of his chorus in preparing them for performance of one of his compositions, see Plutarch On Listening to Lectures 46b. See also West 1981.125n73 on the reference by the Phrygian slave to his song as a harmateios nomos in Euripides Orestes 1384; according to Glaucus of Rhegium ("Plutarch" On Music 1133f), Stesichorus is credited with compositions in the harmateios nomos, derived from Olympus of Phrygia, and we have already seen a reference to a Phrygian tune in Stesichorus PMG 212.

§27n1. Commentary by Barker 1984.60-61.

§27n2. See especially West 1981.117-118. Cf. also Barker 1984.165-168. From the standpoint of someone like Aristides Quintilianus, the notion of scales is no longer anachronistic. The progressive systematization of relations between distinct harmoniai leads to a paradigm-shift from modes to scales.

§27n3. Cf. Ch.3§21.

§27n4. Cf. ibid. Cf. also West, p. 126.

§28n1. On which see Ch.3§1n1.

§28n2. For more on the Aeolian harmoniâ, see Ch.3§21 and Ch.3§27n2. Cf. Anderson 1966.48; also West 1981.128-129. Still, the nomenclature of the tonoi retains the notion of regional differentiation. Thus Aristoxenus Harmonics 46.20-47.1 compares the confusion of relationships between tonoi with the confusion of relationships between various days of the month in various regional calendars: what counts as the tenth day of the month for the Corinthians is the fifth for the Athenians and the eighth for yet others.

§28n3. See Ch.3§1n1.

§28n4. For possible traces of various stages in this process of systematization, see the interesting discussion of West 1981.127.

§29n1. West, p. 126.

§29n2. The synthesis may require sub-categories: there were "more or less co-ordinated 'families' of harmoniai, grouped together under headings such as 'Lydian', etc." (Barker, p. 167).

§29n3. For more on this type of thought pattern, see Ch.10.

§30n1. West 1981.127.

§30n2. West ibid. I would have preferred the words pitch or tone over note here.

§31n1. Anderson 1966.234n65 argues that it was singing different harmoniai that was difficult, not learning to play them. On the appropriateness of the Dorian harmoniâ in teaching the young, see Aristotle Politics 1342a-b.

§31n2. The "old" paideusis 'education' (Clouds 986) is associated with the era that produced the fighters at Marathon (985-986).

§31n3. The composition to which reference is made here in Clouds 967 is apparently that of Stesichorus (see Sommerstein 1982.207).

§31n4. Each harmoniâ, as Sommerstein (p. 207) emphasizes, "required a different tuning (the literal sense of harmoniâ) of the instrument."

§32n1. Pickard-Cambridge 1962.43-44n4; on the date 446 see Davison 1968 [1958] 61-64.

§32n2. Again we may apply Kurylowicz's fourth law of analogy, as discussed at Intro. §13.

§33n1. The genres about to be named are discussed in further detail at a later point, Ch.3§51.

§33n2. Cf. also Aristoxenus F 82 Wehrli on the practice of Simonides, who reputedly used the Dorian pattern for partheneia 'maiden-songs', prosodia 'processional songs', and paiânes 'paeans'. In Etymologicum Magnum 295.53 and following, there is a report that Apollonius the so-called Eidographos classified the compositions of the lyric masters by distinguishing the following categories: Dorian, Phrygian, Lydian, Mixolydian, Ionian, and so on (cf. Rossi 1971.92n63).

§33n3. The various existing patterns of inclusion led to various doctrines reducing the "original" number of harmoniai to three. According to Heraclides Ponticus (F 103 Wehrli, by way of Athenaeus 624c), the "original" three harmoniai were the Dorian, Aeolian, and Ionian; Heraclides makes the explicit equation of Aeolian with Hypodorian (F 163 Wehrli, by way of Athenaeus 624e). By contrast, according to the more prevalent doctrine, apparently espoused by Aristoxenus (cf. Athenaeus 635e, 637d), the "original" three harmoniai were rather the Dorian, Phrygian, and Lydian. Further references and commentary in Barker, p. 213n62. According to the doctrine of Aristoxenus, I would infer, Hypodorian was treated as not only a subcategory but also a derivative of Dorian. According to the doctrine of Heraclides, on the other hand, it seems to have been derived from the Aeolian, in defiance of the nomenclature. The latter doctrine comes closer to a true diachronic scheme.

§33n4. Cf. Barker, pp. 95, 181. On the genre of dithurambos 'dithyramb', see Ch.3§9.

§34n1. Cf. Barker, p. 216n77.

§34n2. See Barker, p. 184n8, for a collection of references.

§34n3. West 1981.115-121.

§34n4. Ibid.

§34n5. West, p. 118.

§35n1. Cf. Kurylowicz's fourth law of analogy, as discussed at Intro. §13.

§35n2. See also "Aristotle" Problems 19.15, Hibeh Papyrus 13 (see Barker, p. 184), and other sources listed by West, p. 117n30. According to Plutarch (Sympotic Questions 645de) the chromatic genus was first used in tragedy by Agathon. On the interpretation of khrôma 'coloration' as a "deviation from the standard," see West, p. 117.

§35n3. Cf. Barker, p. 225n132.

§35n4. Cf. Barker, p. 183.

§36n1. Clement Stromateis 6.88 attributes to Aristoxenus the observation that the enharmonic genus suits the Dorian harmoniâ and the diatonic genus, the Phrygian (cf. West 1981.128). He then proceeds to cite a Dorian harmoniâ in a hymn to Zeus attributed to Terpander, the beginning of which he quotes (PMG 698).

§36n2. Note that the reason given by Barker, p. 165, for his choice of the enharmonic genus in his own description of harmoniai is that "the members of the 'unsystematic' group are more nearly enharmonic than they are anything else."

§37n1. Henderson and Wulstan1973.30; cf. West 1986.44.

§38n1. Cf. Henderson 1957.389.

§38n2. See Henderson, pp. 364-366.

§38n3. Cf. Ch.1§40n1; cf. also Allen 1987.123 on the speech melody of ancient Greek accentuation, "which gradually rises towards the high pitch, whether by steps or glide, and then returns to the low." If there is more than one low pitch in a Greek word, that is, in polysyllables, there will be one lowest pitch while the other lows will be intermediate (ibid.). The smaller intervals in the enharmonic and chromatic may be a reflex of the distinction between lowest pitch and intermediate low pitch.

§38n4. Mountford and Winnington-Ingram 1970.707.

§39n1. West 1981.117.

§39n2. West ibid.

§39n3. See Ch.3§38n3.

§40n1. For a reference to patterns of tuning the lyre that corresponded to tunes played on the aulos, see Xenophon Symposium 3.1 and the commentary of Barker 1984.120n13.

§40n2. Cf. Comotti 1979.27-28 and Seaford 1984.15.

§40n3. Cf. Ch.3§21.

§40n4. Cf. Ch.3§25.

§41n1. A similar point is made, with illustrations from Arabic traditions, by Monroe 1972.40-41; cf. Zwettler 1978.212-213, 217.

§42n1. Cf., for example, IG II2 2311, an inscription concerning prizes at the Panathenaia (first half of the fourth century B.C.); also IG XII ix 189 (Eretria, ca. 340 B.C.). The most marked difference is that the kitharôidoi outrank the aulôidoi in the value of the prizes. To be compared is the hierarchy of listing in Plato Laws 658b: rhapsôidiâ kitharôidiâ tragôidiâ kômôidiâ. Note too the use here of the verb epi-deiknunai in the sense of public performance. Homer is specified as the exponent of rhapsôidiâ; exponents of the other three media are left unspecified. See also Laws 834e-835a. There is a kitharôidos featured in the representation of the Panathenaic procession on the Parthenon Frieze (Shelmerdine 1981.80).

§43n1. Cf. Ch.3§5.

§43n2. More on the professionalism of Pindar at Ch.12§5 and following.

§43n3. Cf. Ch.12§2 and following.

§45n1. See Kirkwood 1974.3-4.

§45n2. Cf. Ch.3§9.

§45n3. Kirkwood ibid. Cf. Herington 1985.228n39: "By the time that Aristotle was composing his Poetics, about 330 B.C., the dithyramb seems to have been the only kind of choral lyric that was still alive enough to deserve his notice in that work." On the performance of a citharodic nome, the Persians of Timotheus (PMG 788-791), at Athens around 408 B.C., see the informative discussion of Herington, pp. 151-160.

§45n4. I take note of the explicit contrast in "Plutarch" On Music 1142bc between the dithyrambs of Philoxenus and those of Pindar as representatives of the new and old styles, respectively. Kirkwood 1974.4 argues that the dithyrambs of Bacchylides represent a transitional phase, being more innovative than his victory odes in revealing a greater proximity to the new styles associated with drama. On the appropriateness of dithyramb for expressing the author as speaking in his own person, cf. Plato Republic 394c. On evidence for Pindar's entry in the dithyrambic competitions at the City Dionysia in Athens at the beginning of the fifth century B.C., see Oxyrhynchus Papyri xxvi no. 2438, column ii, lines 9-10. On the total of fifty-six victories claimed for Simonides in various dithyrambic contests presumably held at various places, see Palatine Anthology 6.213 = EG 27.

§45n5. This is not to say that there are no existing nonstanzaic types of Archaic lyric: see the general discussion at Ch.1§58 and following. The application of the word "genre" here seems to me too wide-ranging to be useful.

§46n1. From here on, I use the expression liberal education in the sense expressed here. For more on Athenian paideiâ 'education', see in general Ch.3§31. Any eventual patterns of exclusion in fifth-century Athens, however, need not have affected the adoption of these "new poets" as "Classics" at a later time, in the context of revivals of "old masters." For example, there is the report of Polybius 4.20.8 (quoted by Athenaeus 626b) concerning the choral education, in his own time, of Arcadian youths who were being brought up on the compositions of Timotheus and Philoxenus: for these Arcadians, at least, these poets represent the "Classics." The phenomenon of shifting perceptions, where a given style is perceived as "modernistic" by one generation and "Classical" by another, is illustrated by the attitudes dramatized in Plato Laws 802cd. As for Plato, when he rejects the modernisms of Timotheus and the like, he is rejecting trends that were by then some eighty or more years old (as Barker 1984.128n13 points out). Plato's tastes are a matter of nostalgic retrojection into the Classical period. Such prescriptions as three years of liberal education, starting with the age of thirteen, in the art of the lyre (Laws 810a) are surely a mere exercise in idealization from the standpoint of Plato's own era. In the fourth century, even in Athens, the rapidly increasing specialization of "music" had increasingly restricted it to the professionals: I cite the revealing discussion in Aristotle Politics 1341a9-36, 1341b8- 18. The specialization is more pronounced in a place like Sparta: Politics 1339b1-4.

§47n1. . The new-style nomos and dithurambos are to be contrasted with earlier old-style attestations of these genres.

§48n1. For more on Athenian paideiâ 'education', see also Ch.3§31.

§48n2. Harvey 1955.162-163. That the skolion is not intrinsically monodic is made explicit in Athenaeus 694b, where we hear that older types of skolia could be choral. In this connection, we may note a report that Sophocles in his youth performed a dance, naked and anointed with oil, to the accompaniment of his lyre, around the trophy erected after the battle of Salamis (Athenaeus 20e-f); whatever we may think about the historicity of this account, its details point to a public choral setting. The "crookedness" implied by the word skolion is explained by Athenaeus (again 694b) as a metaphorical veering, by way of an individual's performance, from the "correctness" or "straightness" of collective performance in the singing and dancing of the chorus.

§48n3. For the attestation of an actual skolion (PMG 891), composed in the Attic dialect, that closely corresponds in phraseology to a stanza in a larger poem composed in the Aeolic dialect and attributed to Alcaeus (F 249.6-9 V), see Nicosia 1976.73-74.

§48n4. Harvey ibid.

§48n5. Cf. Ch.3§31.

§48n6. Ibid.

§48n7. Cf. Ch.3§9.

§49n1. On old-fashioned paideiâ 'education' in lyric, see also Ch.13§52 and following.

§50n1. Irigoin takes note of this fact and adds (1952.15n3): "La lyrique chorale, qui passait déjà pour surannée, est désormais laissée de côté; c'est un genre littéraire dont la vogue est terminée, à Athènes tout au moins. Rares sont ceux qui en garderont le souvenir."

§50n2. On the form of the huporkhêma, see Ch.12§30.

§50n3. Irigoin 1954.14.

§50n4. Irigoin, p. 16.

§51n1. Cf. Plato Laws 658a-659c, 669b-670b and the comments of Svenbro 1984.231n133.

§51n2. Cf. Svenbro, p. 232n136.

§51n3. Cf. Svenbro, p. 231n131. See at Ch.3§31 and following for a parallel theme in Aristophanes, where the nostalgia for old-fashioned traditions in song making is linked with the good old days of Athenian paideiâ 'education', the kind that purportedly produced the men who not only knew how to perform the traditional songs but also fought at the Battle of Marathon (Clouds 961).

§51n4. For the notion of skhêmata as 'dance figures', postures and gestures that represent, see Barker 1984.119n10 on Xenophon Symposium 2.15. Cf. Ch.1§39n3.

§51n5. On the appropriateness of the term genre here, see again Svenbro 1984.225 and n135.

§51n6. Cf. Plato Ion 534c, as discussed at Ch.12§30, where the dithurambos 'dithyramb' is treated as parallel to the huporkhêma (for the meaning, see Ch.12§30), the enkômion 'encomium', the epos, and the iambos (all in the plural).

§51n7. Cf. Svenbro 1984.225.

§51n8. Ibid.

§51n9. Svenbro, pp. 225-226.

§51n10. Cf. Svenbro, p. 226 and n140 (where 1449b should be corrected to 1449a). As Lucas 1968.82 points out, this would imply "that there was no important change later than the early plays of Sophocles, and that no further development is to be looked for." This same attitude is already alive and well in Aristophanes, for whom anything after Sophocles is decadent (cf. Frogs 76-82, 787-794, 1515-1519).

§52n1. See Ch.3§48. See further Reitzenstein 1893.34 for a survey of illustrative passages.

§52n2. Perhaps the most celebrated example of this theme is in Cicero Tusculan Disputations 1.4 on the embarrassment of Themistokles in being unable, at a symposium, to sing and accompany himself on the lyre; the traditions underlying this theme are examined by Reitzenstein 1893.33-34.

§52n3. On the custom of substituting a myrtle branch for a lyre in performances at symposia: Dicaearchus F 89 Wehrli. Cf. the useful discussion of Barker 1984.103n16. On the earlier applications of skolion in a broader sense of a lyric song sung at a symposium (cf. Aristophanes F 223 Kock) and on the eventual semantic narrowing of this word to designate only such special Attic forms as the Harmodios-song, see Harvey 1955.162-163.

§52n4. As Henderson 1957.339 puts it, a single aulos supports the tragic chorus.

§52n5. Cf. the scholia to Aristophanes Wasps1239. Cf. also Reitzenstein 1893.34. For further observations on Athenian paideiâ 'education', see Ch.13§52 and following.

§52n6. The supporting evidence is conscientiously surveyed and clearly discussed by Reitzensten 1893.32-39; cf. Harvey 1955.162. On the interchangeability of aeidô/âidô 'sing' (as in Clouds 1371) and legô 'say' in referring to nonlyric as well as lyric delivery, see the discussion of Herington 1985.13 and 224- 225n15; my interpretation of this interchangeability, however, differs from his.

§53n1. Pindar scholia I, p. 3 Drachmann. I say an Alexandrian edition rather than the Alexandrian edition in light of observations made by Race 1987 concerning Oxyrhynchus Papyri xxvi 2438, c. 200 A.D., a papyrus life of Pindar: this document shows that the sequence of categories in the Vita Ambrosiana was hardly the only one.

§53n2. Mentioned by Plato in Ion 534c; see Ch.3§51n6.

§53n3. Mentioned by Plato in Ion 534c; see Ch.3§51n6 above.

§53n4. The notion of epinicianf will be discussed in detail at Ch.4.

§53n5. This order of the epinician or victory odes seems to have been standard until the period when the text was transferred from papyrus roll to codex. When this transfer happened, the old sequence of Isthmians-Nemeans was reversed, perhaps accidentally, to become Nemeans-Isthmians: see Irigoin 1952.100. The fact that some of the genre categories of the Pindaric corpus take up more than one book has to do, of course, with the conventional space capacity of papyrus rolls. The expected maximum was roughly 2,400 lines, in that 2,400 is denoted by the terminal letter omega in the numbering system used by Alexandrian editors in counting the lines contained in a given roll (where each successive letter of the Ionic alphabet is 100 more than the previous one): see Irigoin 1952.40. The typical length of an Alexandrian roll is represented by the number of lines specified for Book 1 of Sappho in Oxyrhynchus Papyri x 1231, which is 1,320 lines. Irigoin, p. 41, compares the length of the Alexandra of Lycophron (1,474 lines), the Phaenomena and Prognostica of Aratus (1,154 lines in all), the Hymns of Callimachus (1,083 lines), and the four books of the Argonautica of Apollonius of Rhodes (ranging from 1,288 lines for Book 2 to 1,779 lines for Book 4). Such a length for a roll would be just right for an Athenian tragedy or comedy (see Irigoin, p. 41n3). On the editorial arrangement of the epinicians of Simonides, with different rolls assigned to different athletic events, see Irigoin, p. 38; as for the epinicians of Bacchylides, they all could fit within one roll (ibid.).

§53n6. For bits of evidence that may suggest Athenian editorial conventions predating the Alexandrian ones, see Irigoin 1952.39-40. Cf. also Young 1983.47-48.

§53n7. See Irigoin 1952.16-18. No. 1: F 105a.1 SM, a passage from a huporkhêma (composed for Hieron of Syracuse), in Meno 76d. We have already noted ( Ch.3§50) that this passage was parodied in the Birds of Aristophanes. No. 2: same passage, in Phaedrus 236d. No. 3: F 169a.1-5 SM, a passage also cited in Herodotus 3.38.4, from a poem of undetermined genre, in Gorgias 484b. Nos. 4-6: same passage, in Laws 714e-715a as also in 690bc and 10.890a. No. 7: same passage, in Protagoras 337d. No. 8: F 133 SM, a passage apparently (but not necessarily) from a thrênos, in Meno 81abc. Nos. 9-11: F 214, 213.1-2, 209 SM, passages from poems of undetermined genre, in Republic 331a, 365b, 457b. No. 12: F 292 SM, a passage from a poem of undetermined genre, in Theaetetus 173e. The rest are passages from the victory odes. No. 13: from Olympian2.16 (composed for Theron of Akragas), in Protagoras 324b. No. 14: from Isthmian1.2 (composed for Herodotus of Thebes), in Phaedrus 227b. No. 15: from Olympian1.1 (composed for Hieron), in Euthydemus 304bc. No. 16: from Pythian 3.54-57 (composed for Hieron), in Republic 408b.

§54n1. So Irigoin, pp. 19-20. Here is his summary of the major references: F 76 SM, a dithurambos for the Athenians, referred to by Aristophanes (Knights 1329 / 1323) and by Isocrates (Antidosis 15.166); F 89a SM, a prosodion 'processional song', apparently imitated by Aristophanes (Knights 1264-1266); F 105 SM, the huporkhêma for Hieron, mentioned in Aristophanes (Birds 926-930, 941- 945) and two times in Plato (Ch.3§53n4); F 169a.1-5 SM, cited in Herodotus 3.38.4, in Plato (five times, see Ch.3§53n4), and apparently in Aristotle (Rhetoric 1406a). Aristotle also refers to Pindar F 96 SM (Rhetoric 1401a) and to Olympian1.1 (Rhetoric 1364a).

§54n2. Cf. Ch.3§48, Ch.3§52. Cf. Ch.13§54n2.

§54n3. Henderson 1957.340.

§55n1. See Ch.3§48, Ch.3§52, Ch.3§54.

§55n2. Cf. also the discussion of Reitzenstein 1893.39n1. On the obsolescence of actual performance in song, see Reitzenstein, p. 42n1.

§55n3. See again Henderson 1957.343

§56n1. There are exceptional cases where the Pindaric composition was apparently meant to be performed on more than one occasion. In Pindar Olympian6.98-102, for example, the composition calls for its performance at two related but distinct occasions: both at Stymphalos in Arcadia, the native place of Hagesias, the immediate subject of praise, and at the court of the tyrant Hieron of Syracuse in Sicily. Cf. Mullen 1982.26.

§56n2. The nature of choral lyric will be examined in detail at Ch.12§2 and following.

§56n3. Cf. Ch.3§31.

§59n1. Cf. Ch.2§4.

§59n2. The status of the Nemeans as originally the last book of epinician odes (see Ch.3§53n2) helps explain why the last three of the odes in that book, Nemean9, 10, and 11, celebrate victories won not at the Nemean Games. One of these, Nemean10, celebrates a victory at the Heraia of Argos. Nemean9 celebrates a victory in chariot-racing at Sikyon. Cf. Irigoin 1952.42.

§60n1. On the reperformance at symposia of choral passages from Old Comedy itself, see Aristophanes Knights529-530, with references to the songs of Cratinus (F 69, 70 Kock).

§60n2. Reitzenstein, p. 42n1, observes that by the time of New Comedy even the convention of referring to the performance of skolion, not just in the broader sense of lyric types associated with the likes of Pindar but also in the narrower sense of the lyric type known as the Harmodios-song (e.g., PMG 894), had become obsolete.

Go to Previous chapter; Next chapter; Table of Contents; Information

Copyright © 1980, 1997 The Johns Hopkins University Press. All rights reserved. This document may be used, with this notice included, for noncommercial purposes within a subscribed institution. No copies of this work may be distributed electronically outside of the subscribed institution, in whole or in part, without written permission from the JHU Press.