Some Questions on Greek Poetry and Music

Alan Shaw

Was ancient Greek a musical language?

Arguments for the intrinsic musicality of a language are apt to be rather circular. Why is Italian said to be musical? Because of the number of operas written in it. And why are so many operas written in Italian? Because of the musicality of the language. Specific qualities--such as the prevalence of open vowels--are often pointed to, but it seems dubious to call these qualities musical in and of themselves, apart from any particular style of singing. One could just as well say that English is more musical than Italian because it has a much greater variety of vowel sounds.

The musicality of ancient Greek, though, has long been an article of faith not only among casual students, but serious scholars as well. And there are reasons why Greek seems a special case: the intimate relation-- indeed the theoretical identity--between Greek music and poetry, and the fact that the two most basic elements of music--the duration of sounds and their pitch--form two clear and distinct systems in Greek, whereas in many languages, including English, they tend to get confused. Poetic meters were based on the relative durations of syllables, which permitted a fairly direct translation into musical terms. Word accent was based solely on pitch, and hence has often been called a "musical" accent.

It seems to me, though, that the notion that ordinary spoken Greek was naturally closer to music than other languages is a misleading one, and has done a lot of harm to our understanding of ancient poetry and its relation to music. It may be true that certain qualities of the language made it easier for the Greek poet-musician to set words to music. But the fact that something is easily done does not necessarily guarantee a superior artistic result. English, for instance, falls easily into verse measures of four beats, which is the "common time" of most Western music. But this has rarely been used as an argument for the inherent "musicality" of English, and for song composers it can be a hindrance as much as an aid, since they must constantly evade the obviousness of the four-beat pattern to achieve anything original.

Did the melodies of ancient Greek music follow the accents of the text?

The evidence is confusing. The few fragments of music from the classical era seem to indicate that they did not necessarily do so. Also, it has been argued, most lyrics were in strophic form, and a melody designed for one strophe would rarely fit the accentuation of the others. Implicit in this argument is that the same melody was used for every strophe, as in modern strophic songs. Some ancient sources seem to imply that this was indeed the normal practice, but we can't be sure. When they speak, for example, of certain verses being sung to a certain nomos, does this really mean they were all sung to the same tune? Perhaps nomos did not mean "tune", but a tune-making formula or family of tunes, much like the Indian rag (the more general meaning of nomos, which was "convention" or "law", would support this interpretation). And even if all the verses were theoretically set to the same tune, we don't know how strictly it would have to be followed to be considered "the same" in Greek musical practice. Jazz singing allows great freedom in this regard, mostly for purely musical reasons, but often to better express the words of different verses as well. But if there was no requirement at all that different strophes have the same melody, it may be that the metrical identity--the identical pattern of longs and shorts--between strophes was enough for the Greek ear to recognize them as the same, and indeed this could be regarded as a particularly subtle form of strophic song, of which modern examples could be found as well.

Would Greek words set to music be clearly understood if the accents were not observed? In Greek, just as in English, there are words that are identical except for their accent. But these are relatively few, and in context one would not be likely to confuse them (In Chinese, by contrast, the number of homophones distinguished only by their tones is very large). So the words would probably not be unintelligible. But would they sound wrong? This is much harder to answer. There is a famous anecdote about a Greek actor who broke up the audience by speaking the line "I spy a calm spot in the water" with the wrong accentuation, so that it came out as "I spy a weasel in the water." But this was in the spoken dialogue of the play; greater leeway could have been allowed in singing.

We know that the accents in ordinary speech were indicated by pitch, but exactly how we cannot say. The ancient grammarians distinguish several kinds of accents, and one states that the pitch of the voice normally rose by the interval of a fifth to indicate an acute accent. This statement has occasioned much absurdity --such as teachers attempting to recite passages in Greek on two notes to illustrate the accents-- and is largely responsible for the notion that the Greeks "sang" their language rather than speaking it. Even Chinese, which relies much more on distinct tones than Greek, never confines itself to precise pitches in this way. All we can say, with a fair degree of certainty, is that the accents must have been marked by some perceptible jump in pitch.

One interesting suggestion, first made, as far as I know, by the poet Gerard Manley Hopkins (and rarely mentioned since) is that the same accent may sometimes have been indicated by a downward rather than an upward jump in pitch. Though this would seem to go against the statement of the grammarian (who lived, let us remember, several centuries after the classical era, at a time when the old pitch accent was already being lost), it seems quite possible, though of course there is no way of proving it. But if this happened, however occasionally, in ordinary speech, it would make a "correct" observance of accents much easier in singing, especially if different verses were really required to follow the same melody. It would be intriguing, with this in mind, to look again at the extant musical fragments (dubious though they are) and see how often accented syllables are set to notes that represent a marked jump in pitch, whether upward or downward, with respect to the unaccented syllables around them.

But even if we had many more authenticated fragments of Greek music, and none of them showed any clear correlation between melody and accent, it would be hard to say whether this was because the accents were simply ignored in singing, or because we don't know how the Greeks heard the accents, or indeed how they heard music. Pitch in a musical context is a very different thing from pitch in ordinary speech, and strange things, akin to optical illusions in painting, can happen. As song composers know, the same sequence of pitches can accentuate a syllable in one context, and leave it unaccented in another. So even if we could see no clear correlation between melody and accent, we would be wrong to conclude that the Greeks heard none.

Did ancient Greek poetry have a beat?

The simple answer would be that some did and some didn't. But we need first to clarify what we mean by a "beat."

In English prosody "beat" is sometimes used as a synonym for "stress" or "accent." Thus the pentameter will be spoken of as a five-beat line. More careful authors confine the notion of "beat" to stresses that occur in a regular metrical position, as opposed to "extra-metrical" stresses, and still others identify it with beat in the musical sense, that is, with the perception of a regular pulse in time, created by but somewhat independent of any regularly occurring sounds. Since the first two usages depend on the notion of "stress," which has no part in classical Greek prosody, it is only this last, the notion of beat in the musical sense, that will concern us here.

The ancient term most often translated as "beat" is ictus, about which much was written in the nineteenth century. The testimony of the ancients was that ictus clearly existed, at least in poetry associated with the dance, but there was controversy about its nature. On one side were those who, unable to conceive of beat in poetry except as stress, denied that it could have existed at all, since classical Greek had no stress accent. Therefore, they argued, ictus must mean something other than "beat." At the opposite extreme, others, like the classical scholar Jebb, tried to revise the understanding of Greek metrics to bring it in line not only with musical rhythm in general, but with the regular duple or triple musical rhythms characteristic of the modern European tradition.

Of course the musical notion of "beat" is itself a relative thing. In one sense all music, or at least any music that involves more than one performer, has a beat; otherwise the players or singers couldn't stay in time. But we also distinguish, quite often, between music that "has a beat," such as rock n' roll, and music that doesn't, of which the clearest example is Gregorian chant. The ethereal rhythms of chant have attracted many as a model for what Greek choral music must have been like. Like Greek music, chant was monodic, and drew its rhythms directly from the text. But there are other things about chant that make it an unlikely model. First, it was not danced to, as Greek choral music was, and indeed its originators, who lived among the lingering remains of the high pagan tradition, took care that it couldn't be danced to, eliminating anything that would appeal, as they saw it, to the pagan body rather than the Christian soul. Second, the words were in Latin, not the native vernacular of the singers, and they were delivered at a pace much slower than normal speech, with many melismas on single syllables, whereas the normal Greek practice was one note per syllable.

Tempo is an important factor in the perception of beat: if you slow down a lively dance tune to half its normal tempo, the beat, in the more limited sense of a strong kinetic pulse, begins to be lost. This certainly must be considered when we hear classical scholars, for whom Greek is nevertheless still a foreign language, recite Greek verse at a plodding pace and then conclude that the verse has no beat. Also to be considered is the lingering prejudice, inherited from the early church and its music, against any music with a strong beat, the association of such music with "low" as opposed to "spiritual" traditions, a prejudice and association that the body-loving Greeks are not likely to have shared.

But those willing to concede that Greek verse had a beat have themselves been often misled as to its nature. When Jebb tried to determine the beat of ancient verse, he was hampered by the idea that a strong beat can arise only in a rhythm that is invariably duple or triple. Since Greek verse, scanning by the rule that one long syllable equals two short ones, is often neither clearly in one nor the other, he was forced to abandon the classical rule for longs and shorts, and assume that other quantities were used as well. Nowadays we are somewhat more familiar with other musical traditions, such as the African, in which a strong beat can be combined with rhythms that are not simply duple or triple.

In deference to Jebb, though, we should recognize that there are many aspects of Greek musical rhythm that can't be gleaned from the verse. The assumption that the basic rhythm was given by the scansion of the verse seems well-founded: Greek musical notation, after all, consisted only of marks to indicate pitch; time values, being given by the verse itself, were not needed. But how were pauses treated? Were they counted as beats or parts of beats, like most "rests" in our own music, or not counted at all, like breathing marks or rests under "fermatas" ? Were they sometimes counted and sometimes not? According to what rules? All this we can only conjecture.

Likewise we cannot rule out the use of time values other than "long" and "short" in actual music as opposed to theory, though it seems safe to say that if such were consciously used, it was not to achieve the regular duple or triple rhythm that Jebb desired to find. We know from modern researches in baroque and other early music that even a notation capable of minute distinctions in time values does not necessarily specify the ones the performers actually use, and the notation of jazz melodies shows the same thing: in both cases a "swung" rhythm, in which adjacent notes notated with identical time values are made unequal, is common practice. In such cases the notation gives the rhythm only in a schematic way, not in detail, and the same could be said, a fortiori, of any system of verse scansion.

So in sum we can say that Greek poetry, when it was sung, probably did have a beat, and when it was danced as well, the beat could have been fairly kinetic. In some types of modern dance the relation between movement and music is not at all obvious, but the identification of certain dance figures with certain rhythms, for which we have good ancient evidence, would suggest that Greek dance was of the more common variety in which the main steps are performed in time with the music.

But not all Greek poetry was sung. The dialogue in plays, for instance, was written in iambic trimeter, which was considered the type of verse closest to ordinary speech. As it appears in tragedy, iambic trimeter represents a poetically heightened form of speech, but the meter seems to have had its origins in a sort of anti-poetic impulse among some earlier poets, like Archilochus, who grew weary of the melodic mythologizing of their colleagues, and wanted something more down to earth. For this they devised a meter that, apart from being regular, had little in it that was suggestive of song. The succession of quick iambs, for which slower spondees are unpredictably substituted, seems to guarantee that this meter will have no musical beat, or at most a very faint one. Of all Greek meters this is the only one that has found a successful equivalent in English; it was in fact one of the models for the blank verse of the Elizabethan dramatists.

Was Greek music an amateur or a professional art?

Clearly both: poet-composers like Pindar were commissioned to write odes to Olympic victors, pipers were hired to play at feasts, and so on; tragic choruses, on the other hand, consisted of ordinary citizens recruited for the occasion, and any educated Athenian was expected to be reasonably proficient at singing and playing the lyre. The place of music in Greek culture seems, in this way, not terribly different from what it was in Elizabethan times, when most households owned a lute and sales of songbooks by the leading composers were brisk, or during the Regency, when an evening's entertainment was to assemble around the pianoforte and hear the young lady of the house accompany herself singing.

Of course the question has implications beyond payment or nonpayment for services rendered, and in the Greek case these implications are of special importance, having to do with the central role of the poet in Greek musical culture. The young lady in a Jane Austen novel would in all probability be singing something in Italian, by a minor poet whose name she wouldn't even care to know. The song was the composer's, not the poet's. The Elizabethans, on the other hand, would have preferred to sing in English (though Italian influence was already rife); the lyrics, even when anonymous, would have been of far higher quality, and in at least one case--the songs of Thomas Campion--poet and composer would have been the same. But the view of Campion, certainly later and probably even in his own time, was that he was--simply by virtue of being a poet --not really a professional composer. The Greeks, too, distinguished between the professional musician exclusively devoted to the art of sound, and the poet-composer who put noble words to music, only in their culture it was the latter who had far greater prestige: mere pipers and such might be virtuosos, but knew nothing of "rational" music, which always begins with words.

So ancient and modern times seem to agree that the poet was, in some sense, an "amateur" in music, and this has colored much of the scholarship on Greek music. It has led many to suppose that the music of Greek tragedy, for instance, was a pretty simple affair. And in a certain sense this must be true: it is hard to imagine that Aeschylyus or Sophocles, who wrote a hundred plays each, fought as soldiers, tended to their estates, and were active in their city's affairs, had much time or inclination to become virtuosos (though Sophocles was noted for his skill on the lyre) or explore different modes and scales in a purely abstract way, like modern musicians studying harmony. We are reasonably certain, too, that the music of Greek tragedy was 1) monodic, 2) accompanied by only one or two instruments, and 3) sung by an amateur chorus with only a few weeks rehearsal time.

Now, apart from the fact that we are not quite so apt as the nineteenth century to equate musical sophistication with elaborateness, virtuosity, and large instrumental forces, there are reasons why we might want to qualify this view of the poets' music as an essentially simple art. The amateur status of the chorus, in the first place, tells us nothing about the simplicity or difficulty of the music they were required to sing: modern choral societies often consist of amateurs, too, in contrast to the professional orchestras that accompany them. Nor can we assume anything about the instrumentalists, who were in any case presumably paid. And monodic music, as we know from Gregorian chant and any number of non-Western examples, is not necessarily simple or unsophisticated.

What is perhaps hardest for us to determine is what role the different participants, whether amateur or professional, had in the actual making of music. Were the composers the sole creators, and the rest mere interpreters, as in the recent classical tradition? Or were the poets simply songwriters, like those of the thirties in America, surrounded by a crowd of creative performers who knew how to flesh out their tunes? Certainly by the classical era the poet's words, at any rate, were sacrosanct; no one would have thought of changing those. But were the poet's tunes treated with the same reverence? The invention of musical notation at about this time would seem to argue that they were, while its relative crudity, and the rarity with which it has been preserved, might lead us to think that the reverence was no greater than, say, a jazzman's reverence for a Cole Porter tune. We know that musical traditions can be transmitted fairly intact even without notation. But I think it can be said that we know of no other musical culture that has separated the roles of creator and interpreter as absolutely as Western classical music has.

So the trend now is to look for parallels to Greek music in non-Western societies, in Africa or Indonesia, for example, where music-making seems much more communal than in our own. And this is probably a good direction to take. Still, the Greeks were hero-worshippers; it is from them that we inherit that sometimes unfortunate trait, and to the Greeks of classical and post-classical times, the poets were undoubtedly heroes. They were to the average Greek what Bach or Beethoven is to your average music lover. If someone who had heard the original production of the Agamemnon were asked to sing the Hymn to Zeus, he would no doubt try to give what he thought of as a faithful rendition. Of course we have no idea what that might have meant in Greek musical practice; it could be that what they heard as a faithful rendition would sound to our ears like a wild improvisation.

Easily forgotten in all this is how much of the poet's composition, even in a musical sense, was already contained in the verse. Greek melodies, based as they were on an elaborate system of scales and modes, some of which engendered our own, some containing "microtonal" intervals that would have sounded very exotic to us, must have added many subtle inflections to the vocal delivery of the lines, and the traditional associations of different modes would make a wealth of allusions possible. But the main evidence for the compositional power of Greek music, so far as we can still see it, is in its rhythm. And for all but purely instrumental music, that was given - schematically of course, but with far greater precision than in any English song--by the scansion of the verse. When we read that "the rhythms of the epode, as Wilamowitz pointed out, seem to reflect traditional cult hymns; note especially the concluding lines, 3 pher. + glyc. + pher., the same arrangement as the 'rhythmic refrain' which runs through the cult hymn in the Heracles..." we are reading something closer to a musicological analysis than anything we would be likely to find in English literary criticism. The nearest parallel might be an analysis of The Tempest, say, that pointed out Shakespeare's use of verse forms characteristic of the masque, where the critic, however, would probably be noticing diction more than scansion, and it is significant that in this play, too, actual music was almost certainly involved.

In terms of rhythms, Greek verse had indeed a heritage almost comparable to the heritage of melodic and harmonic motifs in Western classical music. And in a large composition like the first chorus of the Agamemnon, this heritage is played on with the same kind of allusiveness, subtle modulations and unexpected transitions that we find in Beethoven's Ninth Symphony. It is tantalizing to glimpse this structure, and not have the melodic element that would have made the whole thing so much clearer and more vivid, but there is no doubt at all that the structure is there.

And it was the poet's creation, not a creation of "pure" musicians or an automatic result of the Greek language's musical qualities. The Greek poets were indeed "professionals", in any meaningful sense of the word, and their profession embraced music. If modern poets are intrigued by the results, they have only to look to the neglected music of their own languages, and see what might be done.

Sites of related interest:

Perseus Project Homepage

DIDASKALIA: Ancient Theater Today



Alan Shaw is a poet, playwright, composer and translator. His verse has appeared in Grand Street and Partisan Review, and on his Website. His verse translation of the classic Russian play "The Woes of Wit" (by Griboyedov) was published in 1993. He lives in New York City.




Copyright 1997 ebr and the author. All rights reserved.

riPOSTe:ebr@uic.edu