The Yezidi Religious Music
A First Step in the Analysis of the Acoustic Shape of Qewls
By Estelle Amy de la Bretèque, Khanna Omarkhali
Oral Tradition, 35/2 (2022):309-30
1. Introduction: Types of Performances and Learning Techniques
While the Yezidi religious textual tradition, including its collection, translation, and analysis, has already become a separate subject of investigation in the field of Yezidi Studies,1 its religious music remains largely unstudied. Based on the analysis of Yezidi religious vocal performances, this contribution aims at understanding the concept and uses of kubrî (“melody”), as well as the relation between text and music. The meaning of a particular oral “genre,” or a category, depends not merely on the text and its verbal performance, but also on rich nonverbal features, such as music and the special role of instruments, ritual ceremonies, the interaction between the reciter and his audience, and so forth.
This exploratory study is a first step towards an understanding of the acoustic shape of the Yezidi religious hymns, Qewls, based on two religious hymns. We will compare their words and acoustic shape in order to find out the parameters of changeability of the variants in performance. The soundtracks analyzed, which can be found in the eCompanion to this article, were recorded by the authors during fieldwork conducted in Armenia, Iraq, Russia, and Germany between 2006 and 2010. Five of the recordings are variants of one Qewl: two were performed in a ritual setting and three were recited at our request in the houses of the informants. In the same period, moreover, Khanna Omarkhali recorded two variants of another Qewl. This second Qewl is known to be uttered using two different “melodies” (kubrî/kubir). We will analyze the differences between the two melodies of the second Qewl and compare them with the five variants of the first Qewl.
A few aspects of music among Yezidis have already been studied. Estelle Amy de la Bretèque conducted research on melodized speech among the Yezidis of Armenia, where mourning, exile, and heroism are expressed through a specific system of melodized utterances. Combining ethnographic and linguistic approaches, as well as acoustic analysis, she showed in her book Paroles mélodisées (2013) that this use of the voice constructs a form of ideal life,
1 Most of Yezidi religious texts were published in the following collections: Celîl and Celîl 1978a and 1978b; Silêman and Cindî 1979; Hecî 2002. Some were translated into English, notably in: Kreyenbroek 1995; Kreyenbroek and Rashow 2005; Omarkhali 2017. The authors are grateful to Ourania Lampropoulou for her very insightful comments on the musical analysis.
309
linking the living to the absent and deceased. The book is accompanied by sixty-five audio and video recordings.2
In her monograph on the Yezidi religious textual tradition, Khanna Omarkhali (2017) recorded, listed, and (without musicological analysis) briefly described the extant kubrîs of the Yezidi religious texts. This is the first near-comprehensive collection and recording of the extant kubrîs of the Yezidi religious hymns. The accompanying CD contains recitations of the seventy-one Yezidi religious texts including variants—sixteen video files and fifty-five audio tracks (total time: 208:43).3
Nura Cewarî (2009:204-18) wrote an article in Kurdish about the melodies of the religious texts of the Yezidis in Armenia, providing a few musical transcriptions. Another relevant study is Scheherazade Q. Hassan’s (1976:53-72) article on the musical instruments of the Yezidis in Iraq, with the focus on the contexts in which these instruments are used, as well as the status of musicians.
Yezidi religious performances take place in different situations, from the highly planned (for example, big religious feasts) to the impromptu. The recitation is done only from memory; no notes are used during performances. Some modern experts, however, use notes and published collections for “refreshing” their memory before performances. Omarkhali (2017:161) proposes a general classification of performances of the Yezidi religious oral texts according to their function, that is, (1) as liturgical texts during religious rituals and gatherings; (2) for teaching and learning purposes,as well as for edification; and (3) as “rehearsals” to prevent the reciter from forgetting his (in very rare cases: her) own repertoire.4
The first text chosen here for analysis is Qewlê Şêxê Hesenî Siltan e (“The Hymn: Sheikh Hesen is the Sultan”), which in Armenia and Georgiais also called Qewlê Bore-borê (“The Hymn of Bore-borê”)because of the repeated refrain: Laliş ya bor, dîwan ya bor, merge ya bor, ya bor, ya bor.
In Armenia and Georgia, the hymn Qewlê Şêxê Hesenî Siltan e is recited at the grave, practically always by two or three “priests” together. The performance is similar to that during the funerals, with the audience standing quietly around the grave.
While some poetic compositions of the Yezidis are recited without any “melody,” the majority of them have their own distinctive “melody,” called kubrî in Iraq and kubir in Armenia and Georgia. The “melodies” are called the kubrî of such and such a Qewl, or the kubrî of such and such a Beyt, and so forth. The recitation of the Yezidi religious poetic texts may vary from speech-like or fast recitation without melody, to intoned recitation and melodic performances (Omarkhali 2017:92).
2 They are available at http://ethnomusicologie.fr/parolesmelodisees.
3 They comprise fifty-two Qewls (with variants), eight Beyts, and eleven recitations of other texts. Besides five audio recordings, another fifty audio files and all the video files were recorded by Omarkhali during her fieldwork in Iraq (Baʿdre, Beʿshiqe, Khanke, Lalish), Armenia (Jirarat, Qamishlo, Shamiram), Germany (Celle, Nienburg, Nienhagen, Oldenburg, Sehnde) and Russia (St. Petersburg). The CD also includes rare recitations of religious texts by women.
4 For more on the religious teaching and learning practices among Yezidis and the types of performances, see Omarkhali 2017, Chapter IV.
310
If a text has a kubrî, then experts would prefer to (and often do) recite the text in public (for example during funerals, feasts, open gatherings, and so on) with the melody, while during other types of performances, often private or in a close small gathering, teaching sermons, and so on, they might recite the text without a melody. The impact of the religious texts becomes stronger with the kubrî, and they are often performed with these melodies in public. If the audience is smaller, the experts might also recite the same Qewl without melody, which demands less energy. As far as the teaching and learning of the melodies is concerned, kubrî, as well as the texts, are to be recited regularly to prevent their forgetting.
A novice learns the texts stanza by stanza, word for word. The master teaches him a part of the text every day, its length depending on the capacity of the pupil. Traditionally, there are two ways of learning. The first is to begin directly with the text and its kubrî, the second is divided into two steps: first learning the words of a religious composition, and later studying its melody. Sometimes a few months pass between these two steps. The melody may not be learned by a qewlbêj (“a reciter, expert on religious texts”) at all. The melody of the text is also studied stanza by stanza. Some informants mentioned that they were forced to learn the text by one preceptor and the kubrî by another (Omarkhali 2017:140).
Kubrîs are always linked with the words of specific texts; they do not exist without words. When people try to remember a Qewl, they recite the words without a kubrî. Omarkhali (2017) listed the melodies of about one hundred Qewls in “Part II, Survey of the Qewls”of her monograph. Omarkhali (2017:92) writes that Yezidi experts characterized various types of kubrî,for example as bilind, “high,” giran, “heavy,” or bi lez, “quick”; these terms may vary slightly or significantly in Iraq (also between the communities of Sinjar, Sheikhan, Beʿshiqe, and Behzane), Syria, Turkey, Armenia, and Georgia.
Unlike the Qewls, which may or may not have kubrî, all Beyts and Qesîdes have kubrî. The most complex forms of recitation are found in the melodies of the Qewls. They might include two or three different melodies in one text; or different melodies of the same text depending on the religious occasion on which they are performed. Based on extensive fieldwork, Omarkhali (2017:93) came to the following conclusions: there are about forty Qewls that do not have any melody. Around forty other Qewls have a kubrî that is unique to these religious texts. There are a few Qewls that are recited with the same kubrî. She writes (2017:93):
Usually, a Qewl has one melody for the whole text (for example Qewlê Axiretê, Qewlê Koçeka, Qewlê Meha, Qewlê Şêşims). In some cases, though, one hymn has more than one melody, for instance, the beginning of the Qewlhas one kind of melody, in the middle it changes, and the last part of the Qewlhas yet another melody; or the melodies might be different depending on the ritual (for example Qewlê Firwara Şêx Fexir, Qewlê Kofa, Qewlê Makê, Qewlê Mezin). Sometimes this is because a number of text units have been combined into one. There is a group of Qewls that have different melodies depending on when they are performed. Thus, it can be assumed that the melody in the Yezidi religious tradition helps to memorize the texts better and helps to distinguish them from others. The melody also provides the more sacral and solemn style of performance.
311
Description of the Recordings
The authors analyzed five variants of the “same” hymn: Qewlê Şêxê Hesenî Sultan e. One of them was recorded in Russia (consult the eCompanion, recording 3), two of them in Armenia (recordings 4 and 5), and two from Yezidis of Iraq (recordings 1 and 2). All of them are performed by members of endogamous religious groups (Pîrs and Sheikhs) with the exception of one performance (recording 1). One version is performed by a woman, the other four are performed by male experts. Recordings 1, 2, 3, 6, and 7 were collected and recorded by Khanna Omarkhali during her fieldwork in 2007 and 2010, not in a ritual context. Recordings 4 and 5 were recorded by Estelle Amy de la Bretèque during her fieldwork among Yezidis in Armenia in 2007 during funerals.
The first recording is the recitation performed by Merwanê Xelîl during an interview in Oldenburg (Germany) in July, 2007. Merwanê Xelîl is not a member of a religious endogamous group (Sheikhs and Pîrs), but he is known as a qewlbêj orexpert on religious textual tradition. The second recording was performed by a woman Sheikh, Dayê Xemê, the sister of a known expert on religious knowledge, Sheikh Deştî in Khanke (Iraq) in 2010.The third recording was collected in Saint Petersburg (Russia) in September, 2007; it was performed by Sheikh Şamilê Kereman at Khanna Omarkhali’s request.
The fourth recording was recorded in Arevik (Armenia) in September 2007 at the funerals of Yurîk. Two Sheikhs (Sheikh Agit and Sheikh Kiniazê Tayar) and one Pîr performed Qewlê Şêxê Hesenî Sultan e on the way to the cemetery located a few kilometers from the village. Following the huge crown of flowers, the men from the “priestly” groups, all in black, uttered the Qewl: the Sheikh leading the performance uttered the entire stanzas, the two other men joined for the refrain. Under the burning sun, they were followed by the coffin (closed because the death occurred days before as the body was brought back from Germany) and an endless crowd of silent men and weeping women.
The fifth recording was performed by Sheikh Kino and recorded in Alagyaz (Armenia) in February, 2007, at the funerals of Rexbet. Sheikh Kino performed Qewlê Şêxê Hesenî Sultan e on the way to the cemetery located about one kilometer north of the village on the top of hill. The procession included a man holding a big crown of white flowers, two men carrying the top of the coffin, Sheikh Kino performing Qewls, a female Pîr with a bowl of “bread of the deceased” (nanê miriya),5 the coffin with the deceased carried by four men (and still open), and a crowd of villagers, first men, mostly silent, and at last women, crying and weeping.
In this study, we compare the first hymn to two versions of another hymn, namely, Qewlê Kofa (“The Hymn of Headdress”). Both versions were recorded by Omarkhali during her fieldwork from an expert on religious knowledge, Merwanê Xelîl. Recording 6 is a recitation of Qewlê Kofa with a quick kubrî, while recording 7 is a recitation with a “heavy” or slow kubrî.
All recordings are subtitled by the authors.
5 Nanê miriya is the food distributed in the name of the deceased.
312
2. Description and Analysis of the Qewls
2.1 The Hymn: Sheikh Hesen is the Sultan
The Qewl we chose to analyze first is Qewlê Şêxê Hesenî Sultan e (recordings 1, 2, 3, 4, and 5 in the eCompanion). Qewls are religious hymns that occupy the highest position in the hierarchy of religious literature. They are traditionally believed to be ultimately of divine origin. Among them, Berane Qewls (“Rams’ Qewls”; also called Qewlê beranî or Qewlêd beran) are considered the most sacred and rich in content regarding the Yezidi religion. The interpretation of such Qewls may be complex because different layers of meanings exist there. In the Kurdistan region of Iraq, a Yezidi reciter of religious texts may memorize Berane Qewls and learn its interpretation only if he already possesses adequate knowledge as an expert in Yezidi religious texts. Qewlê Şêxê Hesenî Sultan e is considered a Berane Qewl.
Among Yezidis from Iraq Qewlê Şêxê Hesenî Sultan e is mainly recited in Sema’ ceremonies. It is one of the hymns that are recited at a grave(ser mezela). This Qewl is recited upon request of a member of the deceased’s family. The type of performance during the funerals is solemn, the audience stands or sits quietly, and two (in some cases three) people may recite the Qewlê Şêxê Hesenê Siltan e simultaneously. When Qewls are recited during funerals by Yezidis in Armenia, they are very often recited by two people, usually a Sheikh and a Pîr at the same time. It is not recited for the death of children or young people. In Iraq this Qewl is often recited together with Qewlê Kofa (Omarkhali 2017:541).
The hymn tells about grief and fate, and about the miraculous power of Sheikh Hesen. In the oral tradition, words may vary slightly from one performance to another.6 To give an idea of the textual content of this Qewl, here are, as an example, the words uttered at the funerals of Yurîk in 2007 in Armenia (recording 4).7
N | Original text | English translation |
1 | Şêxê Hesenî Siltan e, ey Siltan e ‘Ezîzê mala bavê mino Bavek ji bavê qelpê re’man e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the Sultan, hey, he is the Sultan O dear one of my father’s house Such a father whose heart is merciful The Prince of the qelenders (“wandering dervishes”), the prince, the eternal prince. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano ‘Ezîz Siltan Şîxadî hemû derdanî bi dermane. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor.[7] O healer of men O dear Sultan Sheikh Adi, a remedy for all pains. | |
2 | Şêxê Hesenî se’îd e, ey seîd e ‘Ezîzê mala bavê mino Bavek ji bavê qelpê mirîde Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the Sayyid, hey is the Sayyid O dear one of my father’s house Such a father whose heart [belongs] to the Mirîds (“laymen”) The Prince of the qelenders, the prince, the eternal prince. |
6 On the variability of religious texts read more in Omarkhali 2017:177-253.
7 Dîwan (lit. “assembly”); Mergeh is used for Lalish.
313
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano ‘Ezîz Siltan Şîxadî hemû derdanî bi xebîre. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. O healer of men O dear Sultan Sheikh Adi, [he] knows [how to cure] all pains. | |
3 | Şêxê Hesenî mezin e, ey mezin e ‘Ezîzê mala bavê mino Bavek ji bavê qelpê momin e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the great, hey is the great O dear one of my father’s house Such a father whose heart [belongs] to the believers The Prince of the qelenders, the prince, the eternal prince. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano Qapiya miraza, miraz-xasa qapiya Melek Şîxisin e, ciyê gilî gazinane. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. O healer of men The gate of wishes, those who wish, is the gate of the Angel Sheikhisin, the place of complaints. | |
4 | Şêxê Hesenî esed e, ey esed e ‘Ezîzê mala bavê mino Bavek ji bavê qelpê semed e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the lion, hey is the lion O dear one of my father’s house Such a father whose heart [belongs] to the Throne The Prince of the qelenders, the prince, the eternal prince. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano Hezar serî neyê malê. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. O healer of men A thousand people will not come home. | |
5 | Şêxê Hesenî feqîr e, ey feqîr e ‘Ezîzê mala bavê mino Bavek ji bavê qelpê mîr e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the feqîr, hey is the feqîr O dear one of my father’s house Such a father whose heart [belongs] to the prince The Prince of the qelenders, the prince, the eternal prince. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano ‘Ezîz Siltan Şîxadî hemû derdanî bi dermane. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. O healer of men O dear Sultan Sheikh Adi, remedy for all pains. | |
6 | Şêxê Hesenî ‘emer e, ey ‘emer e ‘Ezîzê mala bavê mino Xaliqekî minî her bi her e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the red, hey he is the red O dear one of my father’s house My eternal Creator The Prince of the qelenders, the prince, the eternal prince. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano Bi firwara Siltan Şîxadî ciset qalibî ruhê . . . dermanê. | Laliş ya bor, dîwan ya bor, . Mergeh ya bor, ya bor, ya bor. O healer of men At the behest of Sultan Sheikh Adi, the body, the form of the soul . . . remedy. | |
7 | Şêxê Hesenî dur cemal e, ey dur cemal e ‘Ezîzê mala bavê mino Borekî qudretê yî siyar e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the beautiful pearl, hey, is the beautiful pearl O dear one of my father’s house [He] is the rider of the mighty steed The Prince of the qelenders, the prince, the eternal prince. |
314
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano ‘Ezîz Siltan Şîxadî hemû derdanî bi dermane. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. O healer of men O dear Sultan Sheikh Adi, a remedy for all pains. | |
8 | Şêxê Hesenî dur cemal e, ey dur cemal e ‘Ezîzê mala bavê mino Borekî qudretê yî siyar e Mîrê qelendera, mîr herê mîrê. | Sheikh Hesen is the beautiful pearl, hey is the beautiful pearl O dear one of my father’s house [He] is the rider of the mighty steed The Prince of the qelenders, the prince, the eternal prince. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. Ya hekîmî mêrano ‘Ezîz Siltan Şîxadî hemû derdanî bi dermane. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. O healer of men O dear Sultan Sheikh Adi, a remedy for all pains. |
9 | Şêxekî mi hebû Hekar da, ey Hekar da, ‘Ezîzê mala bavê mino Zor kewanê qudretê palda Mîrê qelendera, mîr herê mîrê. |
My Sheikh was in Hakkari, hey in Hakkari O dear one of my father’s house [He] laid down the strong power (?) The Prince of the qelenders, the prince, the eternal prince. |
Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. | Laliş ya bor, dîwan ya bor, Mergeh ya bor, ya bor, ya bor. |
Qewlê Şêxê Hesenî Sultan e is known to have a characteristic kubrî.8 Our analysis shows that four versions are very similar, while the last one is quite different. In order to compare these recitations, we conducted a comparative analysis based on several parameters: words, melodic curves/cadence, rhythm, and breath groups (see transcriptions 1, 2, 3, 4, and 5, below). Afterwards, we compared this hymn with Qewlê Kofa (see transcriptions 6 and 7), which is known to have two different kubrîs.
Words
A Qewl consists of different sebeqs, often translated as “stanzas” or “strophes.” The sebeqs are not metrical. The words and the order of stanzas/lines of the Qewl may vary slightly from one performance to another. The table below shows the differences for the first stanza within our five recordings.
Recording 1 | Recording 2 | Recording 3 | Recording 4 | Recording 5 |
Ê Siltan e, ê Siltan e | Şêxê Hesenî Siltan e, î Siltan e | Şêxê Hesenî Siltan e, oy Siltan e | Şêxê Hesenî Siltan e, ey Siltan e | Şêxê Hesenî Siltan e, ay Siltan e |
O ay ezîzekî mala bavê mino | Ey ezîzekî mala bavê mino | Ezîzekî mala bavê mino | Ezîzê mala bavê mino | Ezîzî mala bavê mino |
Bavê qelpê rehmen e | Bavê qelpê rehmen e | Bavê qelpê rehmen e | Bavek ji bavê qelpê re’man e | Bavê qelpekî momin e |
8 For more details see Omarkhali 2017:527-44.
315
Mîrê qelender e. | Mîrê qelender e. | Mîrê qelendera. | Mîrê qelendera. | Mîrê qelendera. |
Ay li min ayê ê yê yê ê yê ê | Ey li min ayê êy | Mîr herê mîrê. | Mîr herê mîrê. | |
Laliş yevol, oy dîwan yevol | Laliş yevol, dîwan yevol | Laliş yevol, dîwan yevol | Laliş ya bor, dîwan ya bor | Laliş ya bor, dîwan ya bor |
Merge yevol yevol yevol | Merge yevol yevol yevol | Merge yevol yevol yevol | Mergeh ya bor, ya bor, ya bor | Merge ya bor, ya bor, ya bor |
Ya birîndarê mêrano Siltan Şêx Adî hemû derd e yî derman e. | Ya birîndarê mêrano Siltan Şêx Adî hemû derd e yî derman e. | Ya ezîzê mêrano Siltan Şêx Adî hemû derd e yî derman e. | Ya hekîmî mêrano ‘Ezîz Siltan Şîxadî hemû derdanî bi dermane. | Ya hekîmê mêrano, Siltan Şêx Adî hemû derdanî bi xebîr e. |
Way derman e. | Yay ezîzo. | Ey li min. |
We can mainly observe two kinds of variations:9 (1) change in semantic content, (2) change in interjections or ornamental syllables/words. The differences in semantic content are often linked to repetitions or omissions of words. Sometimes there is also the use of different words. For example, towards the end of the first stanza we have three different versions: Ya birîndarê mêrano (recordings 1 and 2), Ya ezîzê mêrano (recording 3), Ya hekîmî/hekîmê mêrano (recording 4 and recording 5). In case 2, words/syllables such as yay, oy, ayê, yê, way, ey li min ê or ay li min ê. These syllables/words have no clear semantic content, but they have a strong emotional impact. As described in Amy de la Bretèque (2013),10 these syllables/words recall a whole set of emotions and feelings related to loss and pain. They are uttered at the beginning or at the end of breath groups, that is the vocalization produced between two inhalations. They may also constitute entire breath groups as in recording 1 (Ay li min ayê ê) and 2 (Ey li min ê yê êy). In this case they constitute entire autonomous melodic patterns.
Structure of the Stanza and Rhythm
The structure of the first stanza is similar from one utterance to another. Each stanza starts with a recitative pattern uttered alone (when several representatives of the “priestly” groups are present). Then a refrain (often, but not necessarily, similar within one performance) is uttered by one or more persons in a slower tempo. The stanza ends with a last recitative section, shorter than the first one, uttered alone.
9 On the question of variation, change, and stability in the process of transmission of the Yezidi religious texts, see Omarkhali 2017:177-253.
10 On the use of non-semantic syllables with a strong emotional content in Yezidi laments and epic songs, see Amy de la Bretèque (2013:103-04). Similar uses of non-semantic syllables in utterances linked with sadness, loss, and pain have been described in other traditions: eleleu, aiai, and ototoi in Ancient Greece (Loraux 1999; Svenbro 2004), or terirem in Byzantine music (Jefferey 1992:109; Conomos 1974:261-86). Some genres are even named after non-semantic syllables: lailailar in Azerbaijan (Amy de la Bretèque 2005) and amanedhes in Rebetiko (Holst-Warhaft 2003:172-74).
316
The length of the first stanza varies greatly from one utterance to another: from thirty-two seconds (recording 4) to one minute and twenty seconds (recording 1).11 This variety may be explained by different parameters. Qewls are uttered with a “free” rhythm (or non-isochronic one, that is, a rhythm without a regular pulsation, which means that it is not possible to clap hands12 during its performance). The number of breath groups (represented by measure bars in the musical transcriptions) and length of silences between breath groups may vary.13 Speech prosody may also vary from one performer to another: length of pauses on specific pitches within musical phrases may vary. Nevertheless, notions of time and tempo are not foreign to this repertoire: some syllables are stressed with longer pauses on one or several pitches, while others are uttered on a fast tempo (probably much faster than daily speech14). One pattern is a recitative one, with a large number of syllables uttered in a fast tempo and on few joint pitches. Another pattern used mainly for non-semantic syllables is more melismatic (each syllable is uttered on several pitches). A third one, used in the refrain, has a rhythm that is regular, but the four repetitions of melodic/rhythmic sequences are not aligned by a common pulse (there are small gaps between them). Finally, the last syllable of each breath group is longer, a particularly remarkable phenomenon.
Scale and Melodic Patterns
Qewlê Şêxê Hesenî Sultan e is uttered on stable pitches. Except for recording 5, the performances are based on the same scale and have very similar melodic patterns.
Recordings 1 to 4 follow melodic patterns well known in areas using makam.15 Although performers of Qewls are probably not aware of the makam theory, they most certainly are familiar with melodies following makam patterns. In recordings 1 to 4 the melodic patterns are very typical of the tetrachord uşşak (part of uşşak makam according to Ottoman/Turkish terminology, or bayyati maqam in Arabic terminology).16 This tetrachord is characterized by the following intervals: Re -Mi∂-Fa -Sol. The symbol ∂ means that the Mi is uttered slightly lower, by one comma, in ascending path, and slightly lower—until 2.5 commas in descending path.17
11 Recording 4: thirty-two seconds; recording 5: forty-six seconds; recording 3: forty-eight seconds; recording 2: one minute and one second; and recording 1: one minute and twenty seconds.
12 This parameter has also been noticed by Cewarî (2009:213).
13 Numbers of breath groups for the first stanza are the following: recording 1: ten breath groups; recording 2: thirteen breath groups; recording 3: ten breath groups; recording 4: eight breath groups; and recording 5: eight breath groups.
14 For an acoustic analysis of speech (axaftin) among the Yezidis of Armenia, see Amy de la Bretèque et al. 2017.
15 Makam (or maqâm, mougham, maqôm . . .) is a system of melody types used in musics of a vast area ranging from Maghreb to Western China. This system provides a complex set of rules for performance. Each makam has a unique intervalic structure and melodic development.
16 For more detail on uşşak makam (and on makams more widely), see for example Aydemir 2010:108-10.
17 A comma is a very small interval. One tone contains nine commas.
317
In recordings 1 to 4 the first part (up to the refrain) evolves around Fa, Sol, and La (in recording 4 up to Do) in a recitative fast tempo pattern. All of them end on the lowest note of the melodic range: Re (or Do in recording 3, a possible addition to the tetrachord uşşak at the end of melodic phrases). When non-semantic syllables have been added at the end of this section (recordings 1 and 2) the melody is more melismatic and with a slower tempo. The refrain starts with an ascending path (Re-Sol or Fa-Sol) and then follows a clear descending path up to Re or Do. The rhythm is close to a triplet (with no regular pulse in silences). Then the end of the stanza is again a fast tempo recitative pattern on Fa and Sol (Mi and Fa for recording 4) ending with a descending curve towards Re, except recording 2 which still remains on Fa. When non-semantic syllables have been added at the end of this sentence (recordings 1, 2, and 3), they also end on Re. In recording 2 it is in this last part, which is more developed and melismatic, that the performer utters a descending melismatic pattern towards Re.
318
319
In recording 5 the intervals are not the same and the melodic range is larger: more than one octave (Do-Re-Mi-Fa-Sol-La-Si-Do-Re). This scale gives the color rast. But in two melodic phrases the La becomes Lab, which gives a color similar to suznak (or suzinak) in the makam theory (although to fit with suznak, the Mi should be uttered slightly lower). As a consequence, the authors prefer not to link this utterance with any specific path of makam.
The general melodic shape is quite similar to the other recordings: the first part (up to the refrain) follows a recitative fast tempo pattern, the refrain starts with an ascending path (Fa-Sol) and then follows a clear descending path up to Do, and the end of the stanza is again a fast tempo recitative pattern ending with a descending curve. It seems that there is a similar global shape, but the scale used is different.
320
Conclusion
In the five versions of Qewlê Şêxê Hesenî Sultan e, some parameters remain identical: the structure of the stanza and the use of a non-isochronic rhythm. Other parameters are not similar in all versions. The scale and melodic path are shared by four of the five versions. It seems to indicate a preferred (or at least more common) way to utter this Qewl. The existence of the fifth version, which is different (recording 5), raises interesting questions. Can we consider that the melodic line in performance 5 is a variant of the melodic line in performances 1-4? Is it possible that the performer mistakenly sings another kubrî (because his knowledge of kubrîs and their differences is not very good)? This hypothesis would be in line with the fact, mentioned earlier, that this utterance does not behave as if it was sung with a clear makam structure in mind. However, performance 5 was recorded in a ritual setting: to judge by the sounds of this performance, the deceased was taken to the cemetery and buried surrounded by people, and nobody complained. In the ritual performance, this utterance was efficacious! Then, would any melodic line fit if uttered by the right person (a religious specialist)? To answer these pragmatic questions, more ethnography is needed.
On a theoretical level, the five versions of Qewlê Şêxê Hesenî Sultan e tends to indicate that a kubrî is not defined by stable intervals between pitches. Should it be analyzed as an indication of the flexibility of a kubrî? Or should we reconsider our definition of kubrî? The understanding/definition of kubrî is probably related to a general musical shape including a combination of distinct patterns (recitative or more melodic/rhythmic), general melodic curves (ascending, descending), and rhythmic accents on some syllables. The analysis of another hymn—Qewlê Kofa—will help us in defining the concept of kubrî further.
321
2.2 “The Hymn of the Headdress”
Qewlê Kofa, “The Hymn of the Headdress,” also called Qewlê Kofiya, Qewlê Pîrê Libnana (“The Hymn of Pîr Libnan”), or Qewlê Dilê Min î bi Kovan e (“The Hymn ‘My Heart is Sad’”), is an example of the hymn that has different melodies (kubrî) depending on the ceremony during which it is recited. Often it is recited together with Beyta Cindî (“The Beyt of the Commoner”), as well as with Sema’ya Şêxê Hesen (“Sema’ of Sheikh Hesen”) and Qewlê Babekrê Omera (“The Hymn of Babekirê Omer”). When this hymn is recited with the Sema’ya Şêxê Hesen and Qewlê Babekrê Omera, it is performed with a much slower “melody,” but when it is recited with Beyta Cindî, it has another kubrî, which is quicker (Omarkhali 2017:534-35). Qewlê Kofa is believed to have been composed by Dawidî bin Derman.
In variant A (recording 6 in the eCompanion) presented below, which is recited with a special recitation marked by the expert as “quick” (bi lez), the refrain is recited not after each stanza, but once after three stanzas. Variant B (recording 7) is recited with a so-called “heavy” (giran)18 kubrî, which is quite slow and is recited with a refrain after every stanza.
Below is the table with the wording of the analyzed excerpt from both recitations by the same expert ofthe Qewlê Kofa.
N | Variant A: quick kubrî | Variant B: heavy kubrî | English translation |
1 | Dilê minî bi kovan e Pîrê bi nav Libnan e Pîrê Libnano, giyano zeynê bi sura Şêxalê Şemsa ne. | Dilê minî bi kovan e Pîrê bi nav Libnan e Pîrê Libnano, giyano zeynê bi sura Şêxalê Şemsa ne. | My heart is sad The Pîr is named Libnan O Pîr Libnan, dear one, adornment of the Mystery of Sheikh Alê Shemsa. |
Ay Pîrê way ez bim xulamo. | Hey Pîr, let me be your slave. | ||
2 | Kofiya te ye bi qewî ye Lê cema bûn ewliya Pîrê Libnano giyano berê zeynê bi sura Şêxê Adî ye. | Kofiya te ye bi qewî ye Lê cema bûn ewliya Pîrê Libnano giyano berê zeynê bi sura Şêxê Adî ye. | Your headdress (kofî) is strong The saints (ewliya) gathered around it O Pîr Libnan, dear one, adornment of the Mystery of Sheikh Adi. |
Ay Pîrê ez bim xulamo. | Hey Pîr, let me be your slave. | ||
3 | Kofiya te ye cî da Lê cema dibûn mirîde Pîrê Libnano giyano zeynê bi sura Siltan Êzîd e. | Kofiya te ye cî da Lê cema dibûn mirîde Pîrê Libnano giyano zeynê bi sura Siltan Êzîd e. | Your headdress (kofî) is in its place The Mirîds were gathering around it O Pîr Libnan, dear one, adornment of the Mystery of Sultan Ezid. |
Ay Pîrê ez bim xulamo. | Ay Pîrê ez bim xulamo. | Hey Pîr, let me be your slave. | |
4 | Kofiya te ye mezin e Lê cema bûn momine Pîrê Libnano, giyano zeynê bi sura Melik Şêxisine. | Your headdress is great The believers gathered around it O Pîr Libnan, dear one, adornment of the Mystery of the angel Shekhisin. | |
5 | Kofiya te ye bû kêr Wa dinya bûye dukir Pîrê Libnano, giyano zeynê bi sura Şêxûbekir. | Your headdress is pure It appeared in the world O Pîr Libnan, dear one, adornment of the Mystery of Sheikhubekir. |
18 Giran literally means “heavy” but can also mean “slow” and “serious,” when it describes a context, a circumstance, or a form of speech. This use is not limited to Qewls. Some types of Kurdish dances, for example, govend, may also be defined as giran.
322
6 | Kofiya te bi nar e Lê cema bûn babzere Pîrê Libnano, giyano zeynê bi sura Şêşimsê Tetere. | Your headdress is luminous Good people gathered around it O Pîr Libnan, dear one, adornment of the Mystery of Sheikh Shems the Tatar. |
Ay Pîrê ez bim xulamo. | Hey Pîr, let me be your slave. |
Words
Qewlê Kofa also consists of different sebeqs (“stanzas”). As with Qewlê Şêxê Hesenî Sultan e the stanzas of this hymn are not metered. The textual content of the two variants of Qewlê Kofa is identical. This may be explained by the fact that the two recordings are uttered by the same person. Very few non-semantic syllables are used in these two utterances of Qewlê Kofa: ay (at the beginning of the refrain) in both, and ay within the refrain in the text recited with the “heavy” kubrî.
Structure of the Stanza and Rhythm
In these two utterances of Qewlê Kofa the structure varies. With the quick kubrî the structure is the following: first stanza—second stanza—third stanza—refrain, fourth stanza—fifth stanza—sixth stanza—refrain. Thus, the refrain (Ay Pîrê ez bim xulamo)comes after three stanzas. With the heavy kubrî the structure is as follows: first stanza—refrain—second stanza—refrain—third stanza—refrain, that is, the refrain comes between all stanzas.
The rhythm of this Qewl is here again “free” or non-isochronic. The length of each stanza varies greatly from one utterance to another: from nine seconds (quick kubrî, recording 6) to twenty-four seconds (heavy kubrî, recording 7).19 The quick kubrî is uttered fast with few and short breaks. The heavy one is uttered much slower with longer breaks. In both utterances the rhythm of the prosody mainly follows a recitative pattern uttered on few joined pitches, with the exception of some syllables which are more ornamented. Non-semantic syllables are all ornamented.
The breath groups (represented by measure bars in the musical transcriptions) also vary considerably. With the quick kubrî we count six breath groups up to the refrain (three stanzas plus the refrain). Breath groups do not follow the melodic/semantic lines. They do not cut melodic patterns or semantic lines, but they sometimes encompass more than one line/pattern. As a result, the breath does not always follow the structure of the stanza. For example, the last line of the first stanza (Pîrê Libnano, giyano zeynê bi sura Şêxalê Şemsa ne) and the beginning of the second stanza (Kofiya te ye bi qewî ye) are uttered within the same breath. With the heavy kubrî, breath groups are synchronized with melodic lines (and to a certain extent to semantic lines as well). We count four breath groups (for one stanza plus the refrain) uttered on the following words: (1) Dilê minî bi kovan e, Pîrê bi nav Libnan e;(2) Pîrê Libnano, giyano, zeynê bi sura Şêxalê Şemsa ne;(3) Ay Pîrê; (4) way ez bim xulamo.For the refrain, which is cut into two parts, the breath underlines the two different melodic models that constitute the refrain: a first one
19 Measured on the first stanza without the refrain.
323
which is melodic and a second entirely recited on one note (Re). Breath groups 3 and 4 are introduced by a non-semantic syllable.
Scale and Melodic Patterns
Qewlê Kofa is uttered on stable pitches. Recording 6 has a melodic range of 3.5 tones (from Do to Sol). Recording 7 has a melodic range of five tones (Do to Sib). But for both recordings Do, the lowest note of the ambitus, appears very rarely and mainly in the last notes of the melodic phrases. The last note is in most melodic lines a Re. The Do appears then as a possible “addition” to the scale (as described with recording 3). We can consider that for recording 6, the scale in which the melody of the Qewl is uttered is from Re to Sol and for recording 7 from Re to Tib.
324
In both versions of Qewlê Kofa the first part of the stanza (Dilê minî bi kovan e / Pîrê bi nav Libnan e), uttered within one breath group, is composed by two short melodic lines nearly identical on the lower part of the ambitus. It is followed by a recitative part up to the refrain (for the heavy kubrî) or up to the next stanza (for the quick kubrî). Both versions end this second part with a descending curve: on Do for the heavy kubrî, and on Re or Do for the quick kubrî. With the quick kubrî these two parts are repeated three times before the refrain, while with the heavy one, the refrain is added each time. The refrain can be divided in two parts in both versions: a first melodic part (on Ay pîrê) with melismas on few syllables, and a second part (on ez bim xulamo), which is recitative on few notes.
Recordings 6 and 7 share a common general melodic shape, but they are not uttered on the same melodic scale. In recording 7 the melodic patterns are very typical of the tetrachord uşşak (as described for recordings 1, 2, 3, 4 of Qewlê Şêxê Hesenî Sultan e). The Qewl is uttered within this tetrachord (Re-Mi∂-Fa-Sol). Two notes are added at the two ends of this tetrachord: low Do for the beginning and the end of melodic phrases, and high Sib as an ornament with La in a recitative pattern (line 2 in transcription 7). In recording 6 the intervals used are close to
325
recording 7, but the ambitus is shorter, and the Mi is always flat. We have, then, the following intervals (Do-)Re-Mib-Fa-Sol. This scale is characteristic of tetrachord kurdî. The note Sol however is quite rare and uttered systematically in descending path, which makes the scale in use close in perception to the trichord Re-Mib-Fa. Uşşak (recording 7) and Kurdî (recording 6) are modes considered within the same family of makam: their mood and behavior are very similar. Thus, the scale of recordings 6 and 7, although not similar, can be understood as close in their mood and behavior. This is an interesting result. One might have thought that the kubrîs are different because their “mood” is different. In fact, however, their melodic modes are all within the same “mood” (as in the makam theory).
Conclusion
The heavy kubrî and quick kubrî differ in many ways, namely, in their scale, speed, and breath groups. But they share a common melodic shape: first stanza uttered on two short melodic lines and on the lower part of the ambitus followed by a recitative part ending with a descending curve. In the local typology of vocal performances, quick and slow kubrî are considered as different: they constitute two separated kubrîs. Which parameter/s then differentiate two kubrîs? According to our analysis, it could be either the scale, the speed, or the repartition of breath groups, or all of these aspects. The fact that the two versions share a general melodic shape does not seem to be enough to make it the same kubrî.
3. Conclusion: Towards a More Accurate Understanding of a Kubrî Concept
Among the seven recordings analyzed, five followed the path of the tetrachord uşşak (recordings 1, 2, 3, 4, and 7), and one followed the path of kurdî, which is a related scale (recording 6). Recording 5 is the only one with a totally different behavior, close to rast–suznak. The analysis of Qewlê Şêxê Hesenî Sultan e shows that a kubrî is not defined by a stable succession of pitch intervals (as long as different melodies are considered as one kubrî). Our first hypothesis is that the definition of kubrî is probably more related to a general musical shape including a combination of distinct patterns (recitative or more melodic/rhythmic), general melodic curves (ascending, descending), and rhythmic accents on some syllables. Theanalysis of Qewlê Kofa shows that, on the other hand, a common general musical shape is not enough to “make” one kubrî. Indeed, the two kubrîs (“quick” and “heavy”) of Qewlê Kofa share a general musical shape but are considered by Yezidis to be two different kubrîs.
Nûra Cewarî (2009) defined Qewls as sitran, which can be translated into English as a “song.” The word s(i)tran, however, is not used by Yezidis in the religious discourse. In the Caucasus the word s(i)tran is used exclusively for songs with a regular pulsation (often accompanied with dance). In Iraq s(i)tran is used also for epic/historic/heroic songs, a repertoire composed of exclusively recitative patterns (Allison 2001). The terminology may vary from one region to another, but the vernacular typology of vocal performance is quite explicit. As described in Amy de la Bretèque’s work, for nonreligious repertoires the performances are divided into two groups: repertoires to accompany dancing and repertoires to be listened to
326
quietly, such as epic/heroic songs or laments (Amy de la Bretèque 2012; Amy de la Bretèque et al. 2017). According to Cewarî (2009:208), the majority of the melodies of Qewls are close to those of the old traditional songs (sitranên gelêrî yên kevin). The acoustic shape of Qewls indeed shares common points with both nonreligious repertoires mentioned above: for example, they include recitative patterns (as in epic songs), as well as melodic parts (as in dance songs). The melodic parts, however, are not uttered in a regular tempo, which makes Qewls profoundly different from dance songs. Moreover, the recitative parts are more fixed than in epic/heroic narratives: the words vary less than in epic songs, and variations are added mainly on non-semantic syllables. Additionally, in the vernacular typology, Qewls are not compared or defined on the basis of this typology (dance songs vs. epic songs). In Yezidi discourse, religious repertoires constitute a separate category.
Yezidi religious music has hardly been explored until now. The present paper offers some initial conclusions and paves the way for further research. In order to understand the concept and uses of kubrî better, more ethnography is needed. More comparisons such as those in the present study may provide further insights. But this study already raises questions that can only be answered on the basis of more ethnography. For instance: when the performers of recordings 1-4 listen to performance 5, do they actually hear the same kubrî? How much of the kubrî typology is actually shared by the Yezidis in general (Sheikh, Pîr, Mirîd, in the Middle East, in the Caucasus, in the European diaspora, and so on)? Or how much of it is shared by the people whose performances are analyzed in the text? Also, ethnographic enquiry would seem to provide a direct way to understand what happened in “curious” performances like recording 5, or why the kubrîs of recordings 6 and 7 are said to differ. Τo answer these questions, ethnography is a key factor. Apart from this, a more exhaustive analysis of different kubrîs of each Qewl would give a better understanding of the relation between a kubrî and the text in which it is uttered, a more precise definition of kubrî, and would suggest new approaches for researching this repertoire, such as studying the emotional content of kubrî or the modes of memorization of melodies. Also, a study of the acoustic shape of a wider range of religious texts, including for example Beyts, Qesides,and so on, would help to gain a deeper understanding of the religious repertoires of the Yezidis and the musical characteristics of each category of religious repertoire.
Appendix: Musical Transcriptions
Our transcriptions are an attempt to write on a musical staff the first stanza of each recording that is analyzed in the article. We used the following conventions:
Durations. Since there is no stable pulse in these Qewls, in these transcriptions the durations are only approximate. We chose to use only three values: half note, quarter note, and eighth note. Half notes are used to transcribe notes with a really long length and stable pitch. Eighth notes are used to transcribe recitative fast speed syllables. Quarter notes are for the in-between length. Bars do not indicate measures but breath groups.
Pitch. We transposed the transcriptions in order to facilitate their comparison. We transcribed recording 1 in real pitch, and the following recordings were transposed according to the same tonal center as the first one. (Real pitch of recording 2 is 1.5 tones higher than
327
transcription, real pitch of recording 3 is 1.5 tones higher than transcription, recording 4 is uttered 2.5 tones higher than transcription, and recording 5 is voiced 0.5 tone higher than transcription). For Qewlê Kofa we transcribed recording 7 in real pitch, and the transcription of recording 6 is 2.5 tones lower than the real utterance.
French National Centre for Scientific Research, LESC UMR 7186
Freie Universität Berlin
eCompanion: Recordings
References
Allison 2001
Christine Allison. The Yezidi Oral Tradition in Iraqi Kurdistan. Richmond, UK: Curzon.
Amy de la Bretèque 2005
Estelle Amy de la Bretèque. “Femmes mollah et cérémonies féminines de deuil en Azerbaïdjan.” Cahiers d’ethnomusicologie,18:51-66.
Amy de la Bretèque 2013
______. Paroles mélodisées: Récits épiques et lamentations chez les Yézidis d’Arménie. Paris: Classiques Garnier.
Amy de la Bretèque et al. 2017
Estelle Amy de la Bretèque, Boris Doval, Lionel Feugère, and Louis Moreau-Gaudry. “Liminal Utterances and Shapes of Sadness: Local and Acoustic Perspectives on Vocal Production among the Yezidis of Armenia.” Yearbook for Traditional Music, 49:129-48.
Aydemir 2010
Murat Aydemir. Turkish Music Makam Guide. Istanbul: Pan Yayıncılık.
Celîl and Celîl 1978a
Ordîxanê Celîl and Celîlê Celîl. Zargotina Kurda. (Kurdish Folklore.) Moscow: Nauka.
Celîl and Celîl 1978b
______. Zargotina Kurda. (Kurdish Folklore.) Yerevan: Sovetakan groch.
328
Cewarî 2009
Nûra Cewarî. “Sitranên Şêx û Pîrên Êzdiyên Ermenistanê.” (“The Songs of the Sheikhs and Pîrs from Armenia.”) In Şêx Fexrê Adiyan: Fîlosof û xasê ola Êzdiyatiyê. Ed. by Eskerê Boyîk, Bedel Feqîr Heci, Chaukeddin Issa, Siddik Düzgün, Ali Tuku, Ilyas Yanc, Kazim Oba, and Telim Tolan. Oldenburg: Dengê Êzîdiyan. pp. 204-18.
Conomos 1974
Dimitri E. Conomos. Byzantine Trisagia and Cheroubika of the Fourteenth and Fifteenth Centuries: A Study of Late Byzantine Liturgical Chant. Thessaloniki: Patriarchal Institute for Patristic Studies.
Hassan 1976
Scheherazade Q. Hassan. “Les instruments de musique chez les Yezidis de l’Irak.” Yearbook of the international Folk Music Council, 8:53-72.
Hecî 2002
Bedelê Feqîr Hecî. Bawerî û Mîtologiya Êzidîyan: Çendeha Têkist û Vekolîn. (Belief and Mythology of Yezidis: Some Texts and Their Investigation.) Duhok: Hawar.
Holst-Warhaft 2003
Gail Holst-Warhaft. “The Female Dervish and Other Shady Ladies of the Rembetika.” In Music and Gender: Perspectives from the Mediterranean. Ed. by Tullia Magrini. Chicago: University of Chicago Press. pp. 169-94.
Jefferey 1992
Peter Jefferey. Re-Envisioning Past Musical Cultures: Ethnomusicology in the Study of Gregorian Chant. Chicago: The University of Chicago Press.
Kreyenbroek 1995
Philip Kreyenbroek. Yezidism: Its Background, Observances, and Textual Tradition. Lewiston, NY: Edwin Mellen.
Kreyenbroek and Rashow 2005
Philip Kreyenbroek and Khalil Jindy Rashow. God and Sheikh Adi Are Perfect: Sacred Poems and Religious Narratives from the Yezidi Tradition. Wiesbaden: Harrassowitz.
Loraux 1999
Nicole Loraux. La voix endeuillée: Essai sur la tragédie Grecque. Paris: Gallimard.
Omarkhali 2017
Khanna Omarkhali. The Yezidi Religious Textual Tradition: From Oral to Written. Categories, Transmission, Scripturalisation and Canonisation of the Yezidi Oral Religious Texts. Wiesbaden: Harrassowitz.
Reşo 2004
Xelîl Cindî Reşo (Khalil Jindy Rashow). Pern ji Edebê Dînê Êzdiyan. (A Part of the Yezidi Religious Literature.) Volumes 1-2. Duhok: Spirez.
Silêman and Cindî 1979
Xidir Silêman and Xelîl Cindî (Khalil Jindy Rashow). Êzdiyatî: Liber Roşnaya
329
Hindek Têkstêd Aîniyî Êzdiyan. (Yezidism: In the Light of Some Religious Texts of the Yezidis.) Baghdad: Korî Zanyarî Kurd.
Svenbro 2004
Jesper Svenbro. “Le mythe d’Ajax: Entre Aietos et Aiai.” Mythe et mythologie de l’antiquité Gréco-Romaine, 82:154-73.