Language and Meter

[This essay is an online second edition of an original printed version that appeared as Chapter 25 in A Companion to the Ancient Greek Language (ed. E. J. Bakker; Blackwell Companions to the Ancient World 2010) 370-387. In this online second edition, the original page-numbers of the first edition will be indicated within braces (“{” and “}”). For example, “{371|372}” indicates where p. 371 of the first edition ends and p. 372 begins. These indications will be useful to readers who need to look up references made by others to the first edition of this essay.]

The term meter, as used in the study of literature, is ordinarily associated with rhythm in poetry. As such, this word is more specialized in its applications than the ancient Greek word metron from which it is derived, which means simply ‘measure’. In this work, the term meter is analyzed in its more basic sense of ‘measure’. As we will see from the ancient Greek evidence, meter was a ‘measure’ in the sense that it gave ‘measure’ to language, so as to create a special language that was differentiated from whatever was understood to be everyday language. Such special language was understood to be a form of art—a verbal art as distinct from visual arts like painting and sculpting. The most comprehensive term for such verbal art was mousikē.

Like the derivative term meter, which is more specialized than the ancient term metron, the derivative term music is more specialized than the ancient term mousikē, which referred not only to music in the sense of playing a musical instrument. Instead, mousikē referred to all forms of verbal art, not only to the art of playing musical instruments. In terms of mousikē, as we will see, the playing of musical instruments was in fact an aspect of verbal art.

The word mousikē derives from a combination of the feminine noun tekhnē, which means ‘art’ or ‘craft’, with the feminine adjective mousikē, which means ‘belonging to the Muses’. The expression hē mousikē, with or without the word tekhnē explicitly added to it, means the ‘art’ or ‘craft’ of the Muses.

Who, then, are the Muses, who preside over this tekhnē? The noun Mousa, from which the adjective mousikē derives, derives in turn from the root *men-, which conveys the basic idea of ‘have in mind’ (Nagy 1974:249-250, West 2007:34). In the plural, the noun Mousai refers to the Muses, goddesses who inspire the special state of mind required to create the special language that they control. And the creation of that special language is hē mousikē, the ‘art’ or ‘craft’ of the Muses.

The mythical dimensions of visualizing the acquisition and the creation of this special language can be seen most clearly in the Theogony, a composition attributed to a prehistoric figure named Hesiod. In this composition, the definitive form of which {370|371} dates back to the sixth century BCE, the figure of Hesiod describes the moment when he acquired the special language of the Muses: it happened when these goddesses appeared to him in an epiphany and gave him a special voice that enabled him to practice his verbal art (Theogony 21-34). The poet is literally ‘inspired’ by the voice that the Muses literally ‘breathed’ into him (Theogony 31).

In this chapter, the focus is not on the myths that tell about the creation of the ancient poetic language but rather on the realities of such a special language. These realities are expressed in the surviving evidence about the ‘art’ or ‘craft’ of the Muses, which is mousikē. And this concept of mousikē is essential for understanding the ancient Greek concept of what we call meter.

The most informative sources for understanding the realities of mousikē as the ‘art’ or ‘craft’ of the Muses are two major exponents of philosophical thinking in the fourth century BCE, Plato and Aristotle. I will mine these sources not for their philosophical agenda but for the linguistic information they provide about the usage of key words like mousikē.

Especially relevant is a passage in Plato’s Laws (2.668e-670b). In this passage, we see an anonymous Athenian speaker engaged in an ongoing conversation with a Cretan and a Spartan about the problem of making a distinction between the Muses as divine practitioners of mousikē and the human practitioners of mousikē, whom the speaker calls poiētai. Essential for understanding this distinction is the meaning of this word poiētai. The ancient term poiētēs (singular) / poiētai (plural), unlike the modern term poet that derives from it, refers not only to composers who compose in the medium of poetry as we know it. Like the Muses, who are the ideal practitioners of mousikē, the poiētai compose not only in the medium of poetry but also in the media of song and dance and instrumental music.

In a moment, we will look at the relevant wording used by the Athenian speaker in Plato’s Laws, which will show clearly that the Muses are in fact being imagined as the perfect model for the poiētai to follow as practitioners of these multiple media of mousikē. Before we do so, however, I must highlight a basic fact about this wording. The fact is, the ancient Greek words mousikē and poiētai are used by Plato’s Athenian speaker in a traditional rather than an innovative way. It is not a philosophical innovation for Plato to think of mousikē as the art of composing in the various media of poetry and song and dance and instrumental music. In the historical era of Plato, such a way of thinking about the art practiced by poiētai was perfectly traditional. What is innovative and untraditional about Plato’s way of thinking is merely his negative attitude toward this art as it existed in the real world of his era—which he holds up as a foil for the ideal and therefore perfect world of the Muses.

According to the Athenian speaker in Plato’s Laws, the Muses themselves are divine poiētai, to be contrasted with the ‘human’ poiētai (2.669d: ποιηταὶ … ἀνθρώπινοι). By contrast with the ideal that is represented by the divine Muses, whose compositions must be perfect, the compositions of the all-too-human poiētai are imperfect and therefore inferior. And this inferiority, according to the Athenian speaker, is the actual {371|372} cause of an all-important distinction between Muses and the poiētai as practitioners of mousikē, of the ‘art’ or ‘craft’ of the Muses (669c: διὰ τὸ τοὺς ποιητὰς φαυλοτέρους εἶναι ποιητὰς αὐτῶν τῶν Μουσῶν ‘because of the fact that the poiētai are inferior, as poiētai, to the Muses themselves’).

What, then, is this distinction? As we are about to see, it is the fact that the all-too-human poiētai, unlike the divine Muses who practice their art in their ideal world of mousikē, are unable to produce compositions that integrate perfectly the media of poetry and song and dance and instrumental music.

The Athenian speaker in this passage from Plato’s Laws gives an example of the imperfect mousikē of the poiētai. These craftsmen, he says, make the mistake of mixing things up when they are composing song and dance. What happens in the imperfect world of mousikē as practiced by these poiētai would never happen in the perfect world of mousikē as practiced by the divine Muses themselves. These goddesses would never make the mistake of confusing what goes with what in the process of composing song and dance:

Clearly, the Athenian speaker is referring to the actual artistic practices of poiētai in the era of Plato and Aristotle. These poiētai, as we see from the examples cited here by the speaker—and as we can see also from the testimony of other sources surviving from that era and from even earlier eras—indulged themselves in the artistic bravura of mixing the existing forms of composition in the media of song and dance.

But Plato’s speaker does not think of such mixed forms positively as evidence for the artistic bravura of these poiētai. Rather, he thinks of them negatively as evidence for the imperfection of the art of mousikē as actually practiced by poiētai who composed in the media of song and dance.

Counting himself among the exceptional few who are perceptive enough to notice all such confusions, the Athenian speaker now goes on to give another example of outrages committed against the art of mousikē by the poiētai:

In terms of this formulation, the poiētai create the special media that we know as poetry and music by dismembering the components of the general medium of mousikē. In order to create the special medium of poetry, the poiētai separate—and exclude—the components that we know as instrumental music and dance, while they include only the component that we know as the words of poetry. Alternatively, in order to create the special medium of music, the poiētai separate—and include—only the components that we know as instrumental music and dance, while they exclude the component that we know as the words of poetry.

The word that the Athenian speaker uses here in referring to the “measures” of the words of poetry, stripped bare of instrumental music and dance, is metra (μέτρα). In general, the word metron means ‘measure’. In particular, as we are about to see, a metron is a way of measuring two irreducible elements that cannot be taken out of the words of the special language that is mousikē. These two irreducible elements are rhythm and melody.

What we mean by these two words rhythm and melody is different from the meanings of the ancient Greek words from which they are respectively derived, ῥυθμός and μέλος, as we can see from the usage of the Athenian speaker in Plato’s Laws.

Let us begin with the word rhuthmos. In the passage I have quoted from Plato’s Laws containing the formulation of the Athenian speaker, this word rhuthmos can be translated not only as ‘rhythm’ but also, more generally, as ‘dance movement’. That is because the noun rhuthmos derives from the verb rheîn, in the sense of ‘flow’ (Chantraine 2009: s.v.). The basic idea inherent in rhuthmos is that whatever bodily movement there is in dance has a “flow” to it. But the question remains, how does ‘flow’ become ‘rhythm’? The answer has to do with the fact that the flow of movement in dance is counterbalanced by a holding up of the flow, as expressed by way of the noun skhēma. This noun, which I have translated up to now as ‘dance pose’, derives from the verb ekhein in the sense of ‘hold’ or ‘hold up’. As we can see from the combination of the words rhuthmos and skhēma in the passage I have quoted from Plato’s Laws as also in other passages, dance is pictured as a counterbalancing of movement and non-movement. {373|374}

But how does this picture square with the fact that dance is basically a motor activity? How can dance be seen as a counterbalancing of movement and non-movement, of motor- and non-motor activity, of rhuthmos and skhēma? The answer is this: the actual counterbalancing of motion and non-motion can be seen overall as motion in its own right. That is why the “rhythm” of dance can be seen overall as rhuthmos. I now restate in terms of Prague School linguistics: in the opposition of rhuthmos and skhēma, rhuthmos is the unmarked and inclusive member of the opposition, while skhēma is the marked and exclusive member. (For a relevant commentary on the inclusiveness and exclusiveness of unmarked and marked members of oppositions, see Nagy 1990a:5-8.)

Besides the unmarked rhythm of dance, there is also a marked kind of rhythm: as we will see later, this marked kind of rhythm is the meter that frames the verses of poetry. Those aspects of rhythm that are needed only for song and dance can be taken out of the words that are the building blocks of poetry. But there are other aspects, as represented by what we call meter, that cannot be taken out of the words. These aspects are irreducible and inherent in the words of mousikē. And these irreducible aspects of rhythm, as we will also see later, correspond to a phenomenon that can best be described as stress accentuation in the ancient Greek language.

Earlier, I said that a metron ‘measure’ is a way of measuring two irreducible elements that cannot be taken out of the words of the special language that is mousikē. So far, we have examined the first of these two irreducible elements, rhythm, derived from the Greek word ῥυθμός. We turn now to the second of these two irreducible elements, melody, derived from the Greek word μέλος.

In the passage I quoted from Plato’s Laws (2.669d-e), melos actually refers only to an unmarked kind of melody, which is the melody that is sung and danced in song. But there is also a marked kind of melody: as we will see later, this marked kind of melody is the melodic contour that frames the verses of poetry. Those aspects of melody that are needed only for song and dance can be taken out of the words that are the building blocks of poetry. But there are other aspects, as represented by what I call the melodic contour, that cannot be taken out of the words. These aspects are irreducible and inherent in the words of mousikē. And these irreducible aspects of melody, as we will also see later, correspond to a phenomenon that can best be described as melodic accentuation in the ancient Greek language.

So far, on the basis of the passage I quoted from Plato’s Laws (2.669d-e), we have seen that the ancient poiētēs can practice the art of mousikē by composing either in the medium of poetry or in the medium of music. But there is more to it. The ancient poiētēs can practice the art of mousikē by composing also in a medium that is more basic than either poetry or music. He can compose also in the medium of song, and he can combine the danceable rhythm and the danceable melody of his words with the danceable music played on musical instruments like the kithara ‘cithara’ and the aulos ‘reed’. This third and more basic medium is highlighted in another passage from Plato, where we see Plato’s Socrates eliciting a simple answer to a simple question:

The term for referring to the composer of such mousikē, as we have seen from Plato’s Laws, is poiētēs. Unlike the derivative term poet, which refers to the composer of only one medium, which is poetry, the ancient term poiētēs refers to the composer of multiple media—not only poetry but also song and dance and instrumental music.

But the problem is—and the Athenian speaker says so in the wording I already quoted from Plato’s Laws—the poiētai in the real world fail to integrate these multiple media of mousikē. Only the Muses in the ideal world of mousikē do not fail. By contrast with these goddesses, who are of course perfect, the imperfect poiētai do fail. One reason for their failure, as we have seen, is that they mix things up. Another reason, as we have also seen, is that they leave things out, as when they compose poetry by leaving out dance and instrumental music from song, or when they compose unsung instrumental music by leaving out of song the actual words of song.

The very idea that poiētai supposedly leave things out in the composition of mousikē is a key to understanding the media used by the poiētai. When the poiētai are using the medium of unsung instrumental music, they are not composing either poetry or song, since both poetry and song require words. Conversely, when the poiētai are using the medium of poetry, they are not composing instrumental music, since this kind of music has various kinds of rhythms and melodies that poetry does not have. More than that, the poiētai are not even composing song, since song too has various kinds of rhythms and melodies that poetry does not have.

I return to an essential point that is being made by the anonymous Athenian speaker in Plato’s Laws (2.668e-670b). The point is, only the divine Muses in the ideal world can succeed in integrating the media of poetry and song and dance and instrumental music in practicing their own divine tekhnē, the ‘art’ or ‘craft’ that is mousikē. By contrast, the all-too-human poiētai in the real world fail to integrate these media. Their tekhnē of mousikē has disintegrated.

This point about the art or craft that is mousikē, as spoken by the Athenian Plato in the words of his anonymous Athenian speaker, is a remarkably insightful appreciation of the actual state of the art in Athens in the era of Plato and Aristotle. In that era, mousikē as the tekhnē of the poiētai was not an integral whole, even if the meaning of the word mousikē as the tekhnē of the Muses idealizes the idea of an integral whole.

But the integral whole that is mousikē had not really disintegrated. To say that mousikē had disintegrated is to take literally the metaphorical world of Plato. Rather, mousikē had become differentiated. To say this much is simply to describe the state of the art in the era of Plato. The form of art known as the tekhnē of the Muses had become differentiated into multiple forms of art, multiple tekhnai, as it evolved over {375|376} time. In terms of the evolution of mousikē, the difference between the idealized tekhnē of the Muses and the actual tekhnē of the poiētai is a matter of differentiation, not disintegration.

There are indications of these multiple tekhnai, as practiced by poiētai, in the examples used by Plato’s Athenian in his efforts to show the disintegration of mousikē. I will focus on analyzing examples of poetry without music, song with music, and music without words. These examples have to do with not one but five distinct tekhnai, as we know from historical evidence. This evidence can be summarized in terms of two historical facts:

1. There were mousikoi agōnes ‘competitions in mousikē’ held at a festival in Athens known as the Panathenaia, which was one of the two major festivals of the Athenians in the era of Plato and Aristotle.
2. These competitions had to do with the actual performance of mousikē.

Up to now, we have been considering the art of mousikē only in terms of composition. From here on, however, we must consider the same art also in terms of performance. In the real world of the art of mousikē as practiced at the Panathenaia in the era of Plato and Aristotle, the performance of the four media of poetry and song and dance and instrumental accompaniment was differentiated from the composition of these four media.

In the era of Plato and Aristotle, there were basically five tekhnai or ‘arts’ of performance in play at the mousikoi agōnes ‘competitions in mousikē’ held in Athens on the seasonally recurring occasion of the festival known as the Panathenaia.

The first of these five tekhnai was the tekhnē of performing poetry. This tekhnē was known as rhapsōidikē, the ‘art’ or ‘craft’ of rhapsōidoi ‘rhapsodes’ (Plato Ion 538b, 538c, 538d, 539e, 540a, 540d, 541a). At the Panathenaia, this tekhnē was practiced by performers called rhapsōidoi ‘rhapsodes’, who competed with each other in performing poetry—without musical accompaniment (Nagy 2002:36, 41-42).

The second and the third of these five tekhnai were two tekhnai of singing to the musical accompaniment of the kithara and the aulos. These two different forms of tekhnē were known as kitharōidia and aulōidia (Plato Laws 3.700d). At the Panathenaia, these two different tekhnai were practiced separately by performers called kitharōidoi ‘kithara-singers, citharodes’ and aulōidoi ‘aulos-singers, aulodes’, who competed separately with each other in singing songs accompanied by the kithara and by the aulos respectively. In the case of the citharodes, the accompaniment was performed by the singers themselves. In the case of the aulodes, on the other hand, the singers and the accompanists were different competitors.

The fourth and the fifth of these five tekhnai were two tekhnai of performing music without words. These two different forms of tekhnē were known as kitharistikē and aulētikē (Plato Gorgias 501e) or, more simply, kitharisis and aulēsis (Plato Ion 533b). At the Panathenaia, these two different tekhnai were practiced separately by performers called kitharistai ‘kithara-players, citharists’ and aulētai ‘aulos-players, auletes’, who competed separately with each other in performing instrumental music—without words—on the kithara and on the aulos respectively.

It is essential to keep in mind that these five different tekhnai—that is, rhapsōidikē, kitharōidia, aulōidia, kitharistikē, aulētikē—were involved in five separate mousikoi agōnes ‘competitions in mousikē’ at the seasonally recurring festival of the Panathenaia in Athens. Corresponding to these five different tekhnai were five separate competitions at the Panathenaia—of rhapsōidoi ‘rhapsodes’, of {376|377} kitharōidoi ‘citharodes’ (= kithara-singers), of aulōidoi ‘aulodes’ (= aulos-singers), of kitharistai ‘citharists’ (= kithara-players), and of aulētai ‘auletes’ (aulos-players). We learn about these separate categories of competition from an Athenian inscription dated at around 380 BCE (IG II² 2311), which records the winners of Panathenaic prizes (Nagy 2002:38-39, 42n16, 51). And we learn about these separate categories of competition also from Plato’s Laws (6.764d-e), where we read about rhapsodes, citharodes, and auletes—and where the wording makes it clear that the point of reference is the Panathenaia (Nagy 2002:38, 40, 42).

These five separate tekhnai as practiced in five separate mousikoi agōnes ‘competitions in mousikē’ at the Panathenaia were all subcategories of one overall tekhnē—which was mousikē. Direct confirmation comes from the Aristotelian Constitution of the Athenians (60.1), where the author refers to the overall agōn ‘competition’ in mousikē at the Panathenaia (τὸν ἀγῶνα τῆς μουσικῆς).

Although the word mousikē was applicable as an overall term for the five separate and differentiated tekhnai as practiced at the festival of the Panathenaia, it was generally not applied to other tekhnai that were practiced at the festival of the City Dionysia, which was the other of the two major seasonally recurring festivals of the Athenians in the era of Plato and Aristotle. At this other major Athenian festival, there were four separate and differentiated tekhnai, which were tragedy and comedy and dithyramb and satyr drama.

Even though all four of these tekhnai having to do with the competitions at the City Dionysia accommodated all four of the media of mousikē as described by the Athenian speaker in Plato’s Laws—poetry and song and dance and music—we find a striking absence of any overall reference to these competitions in terms of mousikē. There is a simple explanation for this absence. It has to do with the dramatic frame for all the performances at the City Dionysia. The composers of the poetry and song and dance and music to be performed at this festival were composing for performers who represented characters inside the composition and who, as non-composers, had no claim to any direct inspiration by the Muses. Only the poiētēs, as the maker of the given composition, could have claimed to be inspired. So the concept of mousikē as the ‘art’ or ‘craft’ of the Muses could in principle apply only to the composition of the poetry and song and dance and music at the City Dionysia, not to the actual performance of these media.

By contrast, the performers at the Panathenaia represented the composers themselves, and they could thus claim, like the composers, to be inspired by the Muses. So the ancient concept of mousikē as the ‘art’ or ‘craft’ of the Muses could apply to performers as well as composers at the Panathenaia. That is one reason why the performers at the Panathenaia, unlike the performers of drama at the City Dionysia, did not wear masks.

We have seen, then, that the differentiation between composer and performer is linked with a pattern of differentiation in the actual use of the word mousikē as practiced by the poiētai. By the time of Plato and Aristotle, the word mousikē was being used not only in referring to the craft of composition as practiced by the poiētai but also in {377|378} referring to the craft of performance as practiced by craftsmen who directly represented the poiētai. These craftsmen were the rhapsodes, citharodes, aulodes, citharists, and auletes who competed at the Panathenaia.

The use of the word poiētai itself is a sign of this differentiation. Whereas mousikē as a tekhnē is practiced by the Muses themselves in the ideal world, there are no corresponding ideal practitioners in the real world of Athens in the era of Plato and Aristotle. In this real world, there are merely poiētai, who compose but do not necessarily perform what they compose. Unlike the divine Muses, who practice their integral art of mousikē by simultaneously composing and performing in their ideal world, the poiētai are primarily composers, not performers. And so the tekhnē of these poiētai is not so much mousikē as it is simply poiētikē.

This word poiētikē, derived from the word poiētai and meaning literally ‘the art of the poiētai’, is essential for understanding the differentiation between the art of the Muses as practiced by poiētai in the real world and the art of the Muses as practiced by the Muses in the ideal world. This differentiation is made explicit in the Ion of Plato, where poiētikē is used in the specialized sense of referring to the tekhnē of composing only, not performing. In this Platonic dialogue, such a specialization of the word poiētikē becomes overt when Socrates traps the Panathenaic rhapsode Ion into agreeing with the argument that rhapsodes, since they are performers and not composers, have no mastery of the art that is poiētikē.

Here is how Plato develops this argument. First, Plato’s Socrates induces the rhapsode Ion to agree that poiētikē is a tekhnē and that this tekhnē is a holon, an integral whole, just like other tekhnai (Ion 532c). Then he induces Ion to agree that the tekhnē of painters, graphikē, is likewise a holon ‘whole’; and the same goes for the tekhnē of sculptors (532e-533b). Comparable is the pairing of mousikē and graphikē in Aristotle’s Politics (8.1337b24-25). It follows, then, that the multiple tekhnai practiced by performers such as auletes and citharists and citharodes and rhapsodes are disintegrated aspects of a single tekhnē, to which Socrates had referred earlier as that integral whole, poiētikē (Ion 533b-c).

In terms of Plato’s argument, the word poiētikē has replaced the word mousikē in conveying the integrated totality of the tekhnē of the poiētai. And these poiētai are primarily composers, not performers. Accordingly, poiētikē is primarily the tekhnē of composing, not performing.

The same pattern of replacement is evident in the work of Aristotle on poiētikē, known to us as the Poetics, which is in Greek terms a discourse about poiētikē tekhnē. In the opening of the Poetics, we see a listing of the various forms of composition as practiced by poiētai (1447a13-15): these forms are epic, tragedy, comedy, dithyramb, and the various forms of composition for the aulos and the kithara. All these forms, as listed by Aristotle in the opening of the Poetics, correspond to the forms of composition that were actually performed at the two major festivals of the Athenians:

• At the Panathenaia, there were five separate forms of composition, corresponding to five separate tekhnai: (1) epic accompanied by no instrument, (2) song accompanied by the aulos, (3) song accompanied by the kithara, (4) instrumental music played on the aulos, and (5) instrumental music played on the kithara. I have listed these {378|379} five forms here in the order indicated by Aristotle’s own listing.
• At the City Dionysia, there were four separate forms of composition, corresponding to four separate tekhnai: (1) tragedy, (2) comedy, (3) dithyramb, and (4) satyr drama (Nagy 1996:81-82; 1999:27; Rotstein 2004).

Now that we have seen the state of the art of mousikē in the historical context of Athens in the era of Plato and Aristotle, the time has come to assess how the realities of this historical context, as captured by the alternative term poiētikē, relate to the idealization of mousikē as the ‘art’ or ‘craft’ of the Muses.

From a synchronic point of view, the realities of poiētikē in the Athens of Plato and Aristotle during the fourth century BCE indicate the existence of multiple differentiated systems of verbal art; from a diachronic point of view, on the other hand, the idealization of mousikē as pictured by the Athenian speaker in Plato’s Laws indicates the pre-existence of a single and undifferentiated system. (For a relevant commentary on the terms synchronic and diachronic, see Nagy 1990a:4-5.)

Although the Athenian speaker is using a mythological instead of a scientific model in explaining the idea of mousikē as a single undifferentiated system of verbal art, picturing this system as an integrated and perfect whole that becomes disintegrated into imperfect parts in the real world of the poiētai, this mythological explanation is nevertheless most intuitive. That is because it approximates an explanatory model that is truly scientific, not mythological.

Such a scientific explanatory model is supported by the empirical and comparative perspectives of not one but two sciences, linguistics and musicology. In terms of these two sciences, as we will now see, the multiple differentiated systems of ancient Greek poetry and song and dance and even instrumental accompaniment can all be explained on the basis of one single undifferentiated system, which is the ancient Greek language.

Applying such an explanatory model, I start by reassessing the formulation of the Athenian speaker:

In attempting to describe here the differentiation of (1) poetry and (2) music in terms of (1) words without song and (2) song without words, the Athenian speaker is forced to use terms that are imprecise in expressing that differentiation. {379|380}

A case in point is the term metra, which the Athenian speaker applies to the “measures” or “measuring units” of elements that are built into the words of poetry. As I have argued, these measurable elements are rhythm and melody. The problem is, the term metra is imprecise for such an application. As we will see, metra in the sense of ‘measures’ or ‘measuring units’ can apply to the elements of rhythm and melody not only in poetry but also in song. And we already know that song is distinct from poetry. Although song is like poetry in having rhythm and melody, it is unlike poetry in having patterns of rhythm and melody that are more varied than the reduced patterns of rhythm and melody that we find in poetry. Also, song is unlike poetry in having dance and instrumental music as optional features, whereas poetry has no such options.

A related case in point is the pair of terms that the Athenian speaker applies to the elements of rhythm and melody. The two terms are rhuthmos and melos. I have already argued that these terms are likewise imprecise.

First, let us review the case of rhuthmos. This term can apply not only to rhythm as sung in song (or as recited in poetry) but also to rhythm as produced by instrumental music, with or without song. Further, as we have already seen, this term can apply not only to rhythm but also to the motion or motor activity of dance.

Second, let us review the case of melos. This term can apply not only to melody as sung in song but also to melody as produced by instrumental music, with or without song.

These cases of imprecision in the references made by the Athenian speaker to the uses of rhythm and melody in the four media of poetry and song and dance and instrumental music provide valuable evidence for reconstructing an earlier phase in the evolution of ancient Greek verbal art when these four media were as yet undifferentiated from each other.

Such an earlier phase is evident in the usage of the Athenian speaker when he speaks about rhuthmos in poetry. This linking of rhuthmos and poetry points to an earlier phase when poetry was as yet undifferentiated from song. To be contrasted is the later phase—as reflected in what the Athenian speaker says about the state of the art that is mousikē. In this later phase, song and poetry were already differentiated from each other. In this later phase, the differentiated medium of song could still be coordinated with the medium of dance, but such coordination had already broken down in the differentiated medium of poetry.

Here we must reckon with a salient fact about the state of the art that was poetry in the fourth century BCE, the time of Plato and Aristotle. At that time, the basic metrical feature of poetry was rhythm as embedded in the words of poetry. This fact is the single most noticeable and most obvious reality we find in the surviving texts of ancient Greek poetry. The study of ancient Greek meter as we know it is basically a study of rhythm. This kind of rhythm in poetry was differentiated from rhythm as sung and danced in song and as played on a musical instrument.

What is a less noticeable and less obvious reality about Greek poetry at that time is the fact that melody, like rhythm, was also embedded in the words of poetry. This kind {380|381} of melody in poetry was differentiated from melody as sung and danced in song and as played on a musical instrument.

These twin facts about rhythm and meter as embedded in the words of Greek poetry are exemplified in the use of the word metron ‘measure’ by Herodotus, who flourished in the fifth century BCE. There are two cases to consider.

The first case is straightforward, involving a reference to the measuring of rhythm in terms of metron. Herodotus (1.12) uses the expression ἐν ἰάμβῳ τριμέτρῳ ‘in an iambic poem that has three measures’ with reference to verses of Archilochus, dated to the seventh century BCE. Herodotus is referring here to a form of invective poetry (iambos) composed of verses containing three metra or ‘measures’. The meter of these verses is known to us as the iambic trimeter. In terms of the embedded rhythms of this meter, as produced by the alternation of long and short syllables, the iambic trimeter can be described as three consecutive units shaped ⏓ – ⏑ – (– = long syllable; ⏑ = short syllable; ⏓ = short or long syllable). As we see from the internal evidence of all attested verses composed in iambic trimeter, this meter is in fact a “trimeter” in terms of its rhythmical structure of three consecutive units measured as ⏓ – ⏑ –.

The second case is less straightforward and more complicated, involving a reference to the measuring of both rhythm and melody in terms of metron. Herodotus (1.47) uses the expression ἐν ἑξαμέτρῳ τόνῳ ‘in a tune [tonos] that has six measures [metra]’ with reference to verses uttered by the Delphic Oracle, which are actually quoted in this context. Herodotus is referring here to a form of epic or oracular poetry composed of verses containing six metra or ‘measures’. The meter of these verses is known to us as the dactylic hexameter. In terms of the embedded rhythms of this meter, as produced by the alternation of long and short syllables, the dactylic hexameter can be described as six consecutive units shaped – ⏑ ⏑ or – –, with the sixth unit truncated from – ⏑ ⏑ to – ⏓ (again, – = long syllable; ⏑ = short syllable; ⏓ = short or long syllable). As we see from the internal evidence of all attested poetry composed in dactylic hexameter, this meter is in fact a “hexameter” in terms of its rhythmical structure of six consecutive units measured as – ⏑ ⏑ or as – – or, at the end of the line, as – ⏓. (In Aristophanes Frogs 650-651, there is talk of two kinds of rhuthmoi ‘rhythms’, and one of these is described as κατὰ δάκτυλον ‘dactyl by dactyl’, which is evidently a reference to the dactylic hexameter.) But now we run into a complication. Although the six metra or ‘measures’ of the dactylic hexameter, as we see from the internal evidence of the alternating long and short syllables in this meter, are basically six units of rhythm, these same six units, as Herodotus describes them, are being measured in terms of melody rather than rhythm, that is, in terms of the ‘tuning’ or tonos of a string instrument, which is the kithara by default. Evidently, Herodotus is thinking of this meter in terms of the “measures” of singing it to the musical accompaniment of a kithara.

So it appears that in this case Herodotus is thinking of poetry primarily in terms of melody and only secondarily in terms of rhythm as a working component of the metron that is the ‘measure’ of the dactylic hexameter. We can see other cases of this way of thinking. For example, in terms of what the Athenian speaker says in Plato’s Laws, we have already seen that poetry contains melos ‘melody’ as well as rhuthmos ‘rhythm’. Or, to say it in Greek terms, the metron that measures the basic measurable units of poetry is measuring melos as well as rhuthmos. {381|382}

A rare and most precious example of the use of the term melos with reference to melody as embedded in poetry composed in dactylic hexameter is a passage in Plato’s Ion (536b-c) where we see a mention of the melody inherent in the dactylic hexameters of epic poetry attributed to Homer. In this passage, Socrates finds fault with the rhapsode Ion for being a specialist in the poetry of Homer as poiētēs or ‘poet’—to the exclusion of all other poiētai. Socrates playfully describes how this Panathenaic rhapsode is inattentive and dozes off whenever he has to hear the poetry of poiētai other than Homer, but he wakes up whenever he hears a performer sing a melos or ‘melody’ that is typical of the verses of Homer himself:

Plato’s Socrates goes on to compare the behavior of the Corybantes, who are figured as mystical Phrygian dancers: those dancers, he says, are attentive to one single melody that inspires them to dance and to sing the words that go with the dance. One single melos or ‘melody’ can activate for those dancers the skhēmata ‘dance poses’ and the rhēmata ‘words’ that go with that one single melody:

Although there is an element of metaphorical play here in what Plato’s Socrates says about the rhapsodic soul that dances to the distinctive melos ‘melody’ of Homeric verses, the actual presence of melody in Homeric verses is not a metaphor but a reality. The meter known as the dactylic hexameter, which was the one single rhythmical frame for the composition of epic verses attributed to Homer, was simultaneously a melodic frame for these verses. To say it more technically, each hexameter had its own distinctive melodic contour.

There is documentary confirmation in papyrus texts of the Homeric Iliad and Odyssey: in some of these texts, most of which can be dated to the second and the third centuries CE, we can see markings of accents that correspond to the melodic contour of the hexameter (Nagy 2000). The markings are placed over vowels of selected syllables within the wording framed by each hexameter. These markings correspond to {382|383} pitch accents that are built into those same syllables—accents that exist within the individual words of the overall wording. Also, these markings correspond to melodic peaks that are built into the intonation of the overall wording. (The importance of the overall wording is indicated by the fact that ancient Greek was normally written in scriptio continua: that is, the overall wording was written out without indication of word-breaks, that is, without leaving spaces between words.)

In using the term intonation here, I am referring to the patterns of melodic accentuation in ancient Greek, which are not only word-bound but also phrase-bound. The term melodic accentuation derives from the research of W. S. Allen (1987:116-131), who shows that the ancient Greek accents that we know as acute (´) and grave (`) and circumflex (ˆ) are reflexes of a system of melodic accentuation that operates on the level of phrase-units as well as word-units. He describes these patterns of ancient Greek accentuation not only in terms of intonation or contonation but also in terms of melodic accent (Allen 1987:131). The term melodic, as used by Allen, is most fitting for describing the phonetics of ancient Greek accentuation. It points to an understanding of melody as a metrical feature that derives from the ancient Greek language itself (Nagy 1990:34-35, 39-41; for more on accent and melody, see Probert 2006:47-48; also her pp. 45-47 on evidence from the papyri).

When we examine the evidence of Homeric texts as transcribed in papyri dated mostly to the second and third centuries CE, we find that there are generally no more than two or at the most three melodic peaks indicated for each hexameter, and the selective marking of these peaks in the scriptio continua of these poetic texts is a way of recording the traditional patterns of intonation embedded in the poetry itself (Nagy 2000). These patterns of intonation are embedded in the traditional phrases contained by the metrical framework of Homeric verses, and these patterns, which are traditional in their own right, combine to form the melodic contour of these verses (Nagy 2000:17).

In what follows, I show an example of the marking of melodic contour in the Homeric text of the so-called Bankes Papyrus (= Papyrus 14 in the Oxford edition of the Iliad (Monro and Allen 1920), which is a fragment from a papyrus manuscript of the Iliad produced in the second century CE. The Homeric verse I have chosen as an example from the Bankes Papyrus corresponds to line 796 of Iliad 24. I first give the wording of the verse as written in the scriptio continua of the papyrus, and I then give the same wording as it is written in the Byzantine spelling system:

It would be insufficient to say that the pitch accents we see built into the words καλύψαντες and μαλακοῖσιν in this verse actually determined the melodic contour of {383|384} the overall wording contained within the frame of the hexameter. Rather, the melodic contour was determined by the intonation of the overall wording, within the overall syntax of the Homeric verse. And it was this melodic contour that ultimately preserved the older phrase-by-phrase pattern of pitch accentuation (Nagy 2008).

Such markings of pitch accentuation can also be found in texts containing other forms of poetry and song, as in the case of papyri featuring the songs of the poet Bacchylides, who flourished in the fifth century BCE (Nagy 2000).

In the case of song, it must be added that its melodic patterns are more stylized than the melodic patterns of poetry—and far more stylized than the melodic patterns of everyday speech. That is because melodic stylization in song is measured in terms of melodic intervals—corresponding to melodic intervals produced by musical instruments of accompaniment. The testimony of the ancients highlights the distinctions between the melodic intervals of song, which is diastēmatikē ‘marked off by intervals’, and the melodic contours of everyday speech, which is sunekhēs ‘continuous’ (Aristoxenus Elementa Harmonica 1.8-9 ed. Da Rios 1954; also Dionysius of Halicarnassus De compositione verborum 11.13-18 ed. Aujac and Lebel 1981; see Probert 2003:4-7).

By now we have seen that the irreducible element of melody as embedded in ancient Greek poetry, as also of course in ancient Greek song, corresponds to the phenomenon that Allen describes as melodic accentuation in the ancient Greek language.

This formulation about melody can be matched with a formulation about rhythm—or, to say it more specifically, about rhythm in song and about meter in poetry. That is, the irreducible element of rhythm / meter as embedded in ancient Greek song / poetry corresponds to a system of stress accentuation in the ancient Greek language.

Although the system of stress accentuation was not indicated in traditional ways of writing Greek, the existence of such a system has been argued persuasively by W. S. Allen (1966, 1973; a similar but in many ways different argument is offered in the work of Devine and Stephens 1984, 1994; Allen 1987:139 comments on their work). In following Allen’s argument, I distance myself from the position taken by Meillet (1923:11), who argued that ancient Greek rhythm / meter was determined exclusively by quantitative or durational differences between syllables (for a critique of this position, see especially Allen 1973:98).

The basic rules of stress accentuation in ancient Greek can be summarized as follows (Nagy 1972: 26-27):

• a) Words were primarily stressed on their last heavy syllable. (On the concepts of “heavy” and “light” syllables, see Probert 2003:2.) Words containing only one syllable could have either stress or no stress on that syllable.
• b) A secondary stress fell on preceding heavy syllables if separated from the primary stress by at least one mora of quantity. (On the concept of a “mora” of quantity, see Probert 2003:16.)

For illustration, I show two sample verses, one composed in dactylic hexameter and the other in iambic trimeter. The highlighting of vowels indicates the placement of stress on the syllable occupied by those vowels: {384|385}

(The line-final syllable ⏓ counts as latent ⏑ or – when the preceding verse-rhythm is … – ⏑ … or … ⏑ – … respectively; Allen 1987:134 explains this “law of indifference.”)

As we see from these illustrations, the patterns of stress accentuation were independent of the patterns of pitch accentuation in ancient Greek. (In later phases of the ancient language, however, the old patterns of pitch accentuation were replaced by corresponding new patterns of stress accentuation, which persist into Modern Greek, while the old patterns of independent stress accentuation were lost; see Horrocks 1997:67.)

The model built by Allen for describing the ancient Greek system of stress accentuation “gives an immediate and simple explanation of a number of the ‘laws’, ‘canons’, ‘bridges’, etc., regarding the positions at which heavy word-finals may or may not occur; all reduce simply to the avoidance of word-division where this would produce conflict between stress and ictus—more particularly in the coda section of a metrical structure” (Allen 1966:146; on ictus, see Allen 1973:276-279). Especially productive is the application of this model to “Porson’s Law” in iambic trimeter (Allen 1966:129-135).

Allen’s use of the term “metrical structure” highlights a differentiation in terminology. The concept of rhythm as a general term may be contrasted with meter as a specific term referring primarily to a stylization of rhythm in poetry. Applying a combination of synchronic and diachronic perspectives, I have built a model for explaining such a stylization:

This model also helps explain the relationship of meter and formula in the making of Homeric verse (on the concept of formula: Nagy 1990b:29). An alternative model is the formulation of Hermann Fränkel (1960) concerning what he sees as four “cola” contained by the dactylic hexameter of Homeric verse. (On the concept of the “colon,” see West 1982:5-6.) Such a model cannot account for the full range of formulaic variation in the making of Homeric verse (Nagy 1990b:29-35; see also Clark 1994, 1997).

In terms of this argumentation, then, rhythm in ancient Greek poetry and song was a function of stress accentuation, just as melody was a function of melodic accentuation. {385|386}

In closing, I stress that the patterns of rhythm and melody in song may diverge as well as converge with the patterns of stress and intonation in speech, and that such divergence is actually an aspect of the overall metrics of song, despite the ultimate derivation of song from speech. More than that, there is an actual esthetic at work in such divergence. A striking example is the relatively greater degree of divergence between stress and rhythm in the opening as opposed to the closing of the iambic trimeter (Allen 1966:125; Nagy 1972:27-28). Another example is the interplay of partial convergences with partial divergences between rhythm and stress or between melody and intonation in the dynamics of responsion, that is, in the matching of strophe and antistrophe, stanza and counter-stanza. Studies of this phenomenon of responsion show that the convergences are more pronounced in the rhythm than in the melody.

A pioneering example of studies in melodic responsion is the work of Wahlström (1970). For a critique, which is overly negative in my opinion, I cite Devine and Stephens (1994:169). For another critique, I cite Probert (2006:48n88): “apparent accentual responsion in poetry (significantly greater than that found in prose) can result from the fact that sequences of words with similar distributions of word boundaries and heavy syllables are more likely to have similarly-placed accents than sequences of words with no such constraints on the location of word boundaries or heavy syllables.”

I also stress that the elements of dance and instrumental music are both relevant to the linguistic basis of rhythm and melody in the overall metrics of song (Nagy 1990a:33-42).

1. In the case of dance, which is basically a stylization of movement as produced by any part of the human body, I quote a formulation by Allen (1973:100): “Implicitly or explicitly underlying [the] identification of stress as the basis of rhythm is the conception of rhythm as movement, and of stress, in the production of audible linguistic phenomena, as the motor activity par excellence.” (See also Nagy 1990a:38.)
2. In the case of instrumental music, which is basically a stylization of rhythm and melody as produced by the human voice, I refer to a generalized formulation by the musicologist Bruno Nettl (1965:41), who points out that the limitations of the human voice (not to mention the limitations of the human ear), as contrasted with the relatively greater freedom of sound-range in musical instruments, may lead to differences in the patterns of evolution for vocal and instrumental music. Instrumental music may not only diverge from the human voice: such patterns of divergence may become part of an esthetic of interplay between the human voice and its instrumental accompaniment. (See also Nagy 1990a:34.)

From a diachronic point of view, dance and instrumental music may be seen as differentiated elements that derive from song—elements that can become further differentiated as either contrasting with song or parting with song altogether. (Examples of such a parting are forms of dance or instrumental music that exist independently of song.) {386|387}

From a synchronic point of view, on the other hand, any such contrast may be seen as the basic state of affairs. That is, song and dance and instrumental music may be seen as separate elements that happen to come together in the art of mousikē.

Either way, separate or unified, song and dance and instrumental music are regulated by the measures of their rhythm and their melody. And such measures are based on language. That is the essence of meter.

The basic facts about ancient Greek rhythm and melody as linguistic phenomena can be found in the introductory books of Allen 1987 and Probert 2003. There are specialized accounts in Allen 1973 and Nagy 1990a:33-42. Another specialized account is the book of Devine and Stephens 1994, which offers alternative views. A standard work on Greek meter is West 1982. The work of Blanc 2008 is a model of sound methodology in studying the interactions of meter and poetic language.