The musical symbol of Central Asia is a unique polyphonic style, known by the names “overtone singing” or “throat singing”. A solo performer produces two different sounds simultaneously. This singing style has been found on a wide territory comprising Tuva, western Mongolia and Altay and Sayan mountain ranges. According to the information received from Yuri Sheikin from Yakutsk, there are few different distinct styles of overtone singing in this region:
(1) Northwestern Tuva is the most important region for this style, with 14 styles of singing (term – khoomei).
(2) Unlike western Tuva, eastern Tuva is much less known for overtone singing traditions, and singers here are far apart. Three styles have been documented here (the term khoomei is used here as well).
(3) Five styles of overtone singing have been recorded in Mountain Altay. Overtone singing in known here under the term kai.
(4) Three styles have been recorded among the Mountain Shoria. They are also known under the term kai.
(5) Khakassia has two styles of overtone singing, known as khai.
(6) One style of overtone singing has been documented in Yakutia as well, under the name khabarga.
(7) And finally, outside the Russian Federation, six styles of overtone singing have been documented in western Mongolia (under the term khoomei)
Overtone singing is positioned ambiguously between the polyphonic and monophonic singing traditions. From a musical point of view overtone singing is polyphony, as two functionally completely different parts (drone and a pentatonic melody) are heard simultaneously. From a social point of view, overtone singing is not polyphony, as this singing style lacks crucial social element of the polyphonic singing tradition – active musical interaction between several (at least more than one) performers.
The singer produces a sustained pitch using a specific tense sound, then, using this sustained sound as a drone, he changes the mouth cavity shape with his tongue, lips and some other parts of the mouth to produce different harmonics (overtones) and construct melodies from these overtones. Interestingly, singers use only the selected set of overtones, carefully avoiding two overtones that do not fit the pentatonic scale (Levin, 2002:982). So, if the drone is on “C”, the singer will be using the overtones from “G” to the next “G”. This part of the overtones contain G, Bflat, C, D, E, F# and G. Out of these overtones “B flat” and “F#” are carefully avoided. This fact suggests that singers do not follow the naturally existing sound material. Instead they follow their aesthetic and cultural preferences. Overtone singing was traditionally performed by men only, but now there are women singers as well.
There has been a diversity of opinion about the origins of overtone singing, ranging from the most archaic periods (30-40 000 years ago), the “pre-speech articulation” epoch of human history (see, for example, Ikhtisamov, 1984:180-181) to the first millennia A.D. (Vainstain, 1980). I will discuss this interesting question in a separate section (or “Case Study”) in the second part of this book.
There has been a great deal of interest in the overtone singing style, and many singing-loving westerners have learned this unique sound-production technique. I remember myself sitting together with my Georgian friend and colleague, Edisher Garakanidze, under a big tree in Switzerland, near Geneva on September 26th, 1991, trying to learn the basics of overtone singing under the guidance of the brilliant performer of this style (and not only!) and ethnomusicologist Tran Quang Hai. After some time I came to the conclusion that, although I managed to produce few audible overtones, singing Georgian polyphony was still much easier for me.

Although Central Asia is justifiably known as a monophonic region, elements of “usual” polyphony (when polyphony is created by the group of singers) are present in some genres. I well remember the excitement among Soviet ethnomusicologists when an example of vocal polyphonic music was found in Central Asia. This example of rare polyphonic singing came from Tajikistan. I am very grateful to Zoia Tajikova who provided this information together with her own transcription of this unique version:
Z. Rabiev recorded in 1964 this polyphonic version of the traditional naksh in Ura-Tiube, in Fergana Vale, northern Tajikistan. The song is performed as an alternation of the soloist (sarnaksh) and the mixed choir of 10-12 singers. The soloist singing sections are in free rhythm, and the choir singing sections are metred. It is difficult to say whether all these spectacularly clashing dissonant chords were intended as they were performed, but it is clearly not a case of performers “not achieving the unison”, because the song actually starts with a long section where the choir does sings in unison (and octaves).

As I know, although the tradition of vocal polyphony has not been documented, strong vocal-instrumental polyphonic tradition is present in solo epic genre (zhyrau) in Kazakhstan. In this genre solo singing is supported by dombra playing, which accompanies the vocal melody in parallel fourths and fifths. Quite interestingly, in virtually all the earlier transcriptions of Kazakh epic tradition scholars were transcribing only the monophonic version (melody only), without the two-part dombra accompaniment. Singing in groups and antiphon between the groups of singers is another interesting tradition. “The most popular wedding song, “Zhar-Zhar,” which is found in different versions throughout Central Asia, constitutes a genre on its own, defined by an antiphonal performance style in which two choirs – one male and one female – sing responsorially” (Kunanbaeva, 2002:951).

Afghanistan (Nuristan)

One of the world’s most isolated, rich and interesting traditions of vocal polyphony exists in high mountains of East Afghanistan, among Nuristanians.
Hidden from the expansionist politics of Arabs, Mongols, and Persians by the high impenetrable mountains of southern slopes of the Hindukush mountains in East Afghanistan, about 150 000 Nuristanians maintained their independence until the end of the 19th century. In 1896 Afghani Amir Abdur Rahman Khan finally incorporated Nuristanians into Afghanistan, turned them into Moslems and gave their region a new name “Nuristan” (“enlightened”). Prior to this event this region was known in Afghanistan as “Kafiristan” (“Land of Infidels”). As you would expect from the people who live in such isolation, Nuristanians maintain elements of pre-Moslem practices, most notably music with the unique polyphony and dances. About half of the Nuristanians also live in neighbouring Pakistan. Nuristanians speak Kafir and Dardic languages (belong to Indo-Iranian languages), are agriculturalists and physically differ from the rest of the Afghanistan population.
Singing traditions of Nuristanians represent an extremely interesting and isolated case of vocal polyphony. I was privileged to become acquainted with over forty polyphonic songs from Nuristan, recorded by Herman M. Pressl in 1968 and 1969 (I am grateful to Viennese Phonorgammarchiv for making these recordings available to me). Together with the recordings made by Lennart Edelberg and Klaus Ferdinand in 1953-54 (and by Lennart and Margot Edelberg in 1964 and 1970), these recordings represent variety of examples of Nuristan traditional polyphony.
According to these recordings, Nuristan polyphony was (at least in the 1960s and the 1970s) a rich and live tradition, fully functioning in Nuristan society. Both men and women sing traditional polyphonic style. They mostly sing separately, but sometimes they do sing in mixed groups. Polyphony is mostly three-part. Rhythm is always precise. Metre is mostly triple (most popular being 6/8, although 4/4 and the peculiar 5/8 is also used). Lydian scale is one of the main scale systems in Nuristan polyphonic songs. Melodies have a small range (mostly up to fourth, or augmented fourth in Lydian scale). Polyphony is based mostly on the use of two principles: ostinato and drone. Drone is mostly rhythmic. Three-bar repetitive structures dominate. Songs are often accompanied by drums, string instrument wadzh (bowed harp), clapping and dancing.
Maybe the most important and salient feature of Nuristan polyphony is the amazing richness of dissonant chords and intervals. Nuristan vocal (and instrumental) music is a true kingdom of secondal dissonances. This abundance of seconds is often derived from the leading melodies “jumping” over the drone. If you think of three independent parts being squeezed in the space of the fourth (or augmented fourth), it is not surprising that in some songs seconds are virtually the only interval you hear during the whole song.
As transcribed examples of Nuristan polyphony are very scarce, I am including here few examples, which I transcribed from the recordings made in Nuristan by Herman M. Pressl in 1968 and 1969. Not available on multi-microphone recordings, these recordings are not easy to transcribe.
The “typical” Nuristan polyphonic song is a three-part composition, with two lead melodies (mil-alol and at-alol, “lead” and “support” parts) and mostly movable drone (asamchilog, choir). 6/8 metre is the most widely spread. Very peculiar string instrument – wadzh (four string bowed harp, tuned usually into Lydian tetrachord with the augmented fourth and very characteristic chords, full of seconds) is also often used for the accompaniment of singing and drumming. The following excerpt represents these features:
I am presenting here only six bars from each song as three-bar structures dominate Nuristan polyphonic songs. The songs mostly consist of the ostinato-like repetition of these six-bar structures (3+3 bars). One of the voices (at the end of the repetitive three-bar phrase) is improvising the short motif that leads to the next repetition of the three-bar structure.
The next example is in 5/8 metre with the typical secondal clashes between two parts. Song is accompanied by drumming (on the first and fourth beats of the five beat cycle):
These were so far examples of male polyphony. Women’s singing has the same features, and three-part singing is dominating. Here is the example of typical three-part polyphony from Nuristan women:
One of the songs from this collection is particularly interesting, as two versions of the same song (“Senkivar kasek”)` – were recorded in the same village on the same day. The first version is performed by women and the second is by mixed group. Most of the songs are sung by men and women separately. This kind of mixed performance might not be usual for Nuristanians, because this particular performance shows a bit of uncharacteristic features, and possibly even a “confusion” of the singers. It starts with the male lead and in the first phrase both man and women are singing the second “supporting” part. This is not traditional (to have more that one person singing any of the lead parts), and possibly both male and female singers are giving each other space, because for a while none of the “second soloists” is supporting the lead singer. Only about a minute later the female singer starts singing again the supporting lead part. Group of women also sing a drone in this version, and as women sing an octave higher, this version has uncharacteristically wide range of more than an octave. Second lead part, sung by a woman to the male lead part is very interesting. Instead of singing, as usual, a second lower from the lead melody, she follows the lead melody a seventh higher (singing the same notes, but an octave higher).
Nuristan polyphony is definitely one of the most isolated and interesting traditions of polyphony on our planet. We will discuss the unique features of Nuristan polyphony and its possible links to other polyphonic traditions in the second part of the book.