The Kammu people are one of the ethnic groups in South-East Asia that traditionally live on mountain slopes, where they grow rice and other crops and hunt. They live in northern Laos and adjacent areas in Vietnam, Thailand, and China. The Yùan area, from which most of the material used here stems, is located in the Nalè area in the southern part of the Luang Namtha Province (Map 2). Kàm Ràw (also known as Damrong Tayanin, 1938–2011) was born and grew up in the village of Rmcùal (Figure 1). As a young man, he came to Thailand and eventually to Sweden, where he became a key informant and co-worker in the ‘Kammu Language and Folklore’ project at Lund University. Most of the vocal expressions studied in this chapter were performed by him, and he also wrote about his life in his home village. Like many other mountain villages, it has been abandoned and the villagers have moved to nearby communities.¹

Children didn’t sing together with the grown-ups at parties, but in the fields they could do that. At feasts boys at the age of six or seven could sit on their fathers’ knees. The elders used to say: ‘You will never get good relations with other people if you don’t sing. You will stay by yourself. You will never get people for parties and you can not go to another village or play with girls’. So the boys liked to sit on their fathers’ knees and to learn. The following morning my father used to ask: ‘Did you hear which song we used? Why do you think we used those?’ Boys also used to gather by themselves during parties and try to sing like the grown-ups. Women did not join the drinking men but used to sit by themselves. If somebody sings to you then, like a question, you should be able to answer correctly in order not to lose face. At feasting, a singer chooses who should answer by nipping at that person’s straw to the wine jar.²

Figure 1 Kàm Ràw (right) performing a perception test with another Kammu speaker in the village of Thapen, close to Luang Prabang in Laos, 1996.

In a Kammu village, there should be at least one shaman and one medicine man, since the Kammu people believe in spirits. In my home area, for instance, there were over 20 Kammu villages, but there was only my uncle Sɛ́ɛn who was a shaman in the entire area. Unfortunately he died in 1960. After he had died we had to find a shaman from a distant village when we needed him.

Anyone could become a shaman, depending on who is able to do the work. Both men and women can become shamans. In 1958, my uncle Sɛ́ɛn taught me and one of my cousins, Kàm Mán, how to become a shaman. In fact we knew most of the things already, because we had been together with him for years, but now he gave us extra, intensive training for a few months. To learn how to be a shaman is not easy at all, because one has to learn many different kinds of magic formulas and rites. There are hundreds of different spirits as I said above. There are specific magic formulas to call the various kinds of spirits. The suitable period for learning to be a shaman is when the sesame sets flowers, that is in September and October. This period is the best time for learning the magic formulae. Kammu people believe that if a person learns the magic formulae during this specific time, the spirits will both respect and fear him or her as a shaman.

Map 2 Map of Laos. The arrow shows the approximate location of the Yùan Kammu area.

When someone catches an illness caused by an evil spirit, that means that a spirit has entered his or her body. A shaman may then chase away the evil spirit from his body. We cannot heal illnesses caused by spirits by giving the patient medicine, the spirits have to be chased away. On the other hand, when someone has caught an ordinary illness or has wounded himself, the medicine man may use plant-medicine to heal the wound or cure the illness.³

Kammu language and vocal expressions

Previous studies of Kammu language and music had made it clear that the analysis of some vocal expressions from the perspective of a performance template is a rewarding pursuit.⁴ One main objective at the start of this study was to investigate more vocal expressions in a search for different performance templates, as well as for differences and similarities between them. There was also some prior knowledge about the interplay of melody and lexical tones.⁵ Another objective was to investigate how this interplay was practised in the various performance templates. Other factors, such as thematics in narratives, intonation, and reduplication of syllables, had not been studied before. Different methods of description or transcription of performances have been employed in this study, depending on the aims of the analysis in question, but also in order to communicate detailed information in a way that would be easily understandable for both linguists and musicologists.

The Kammu (Kmhmu’, Khmuʔ, etc.) language belongs to the Mon–Khmer branch of the Austroasiatic language family and is spoken by some 600,000 people in northern Laos and adjacent areas of Vietnam, Thailand, and China.⁶ There are three major dialects: Northern, Western, and Eastern Kammu. The Yùan variety of Kammu treated here is a sub-dialect of Northern Kammu, spoken in the Nalè area in the southern part of the Luang Namtha Province. Northern and Western Kammu have developed a system of two lexical tones, High and Low (denoted ´ and `, respectively), while Eastern Kammu, spoken further to the east and south in Laos, and also in Vietnam, retains the original state without distinctive lexical tones. These tones have developed from voiceless and voiced initial consonants, which gave rise to High and Low tone, respectively. For example, the Eastern Kammu minimal pair klaaŋ ‘eagle’ ~ glaaŋ ‘stone’ with a voiceless vs. voiced initial consonant corresponds to Northern Kammu kláaŋ ~ klàaŋ and Western Kammu kláaŋ ~ kʰlàaŋ with High vs. Low tone. Similarly, the Eastern Kammu pair hraaŋ ‘tooth’ ~ raaŋ ‘flower’, with a voiceless vs. voiced initial sonorant, corresponds to Northern and Western Kammu ráaŋ ~ ràaŋ.⁷ Our analysis here concerns the tonal Northern Kammu dialect; material from other dialects is shown only for comparison, in certain cases.

Northern Kammu words are written in the practical orthography used in other publications from the Kammu project at Lund University. In this orthography, ñ and y are written for IPA [ɲ] and [j]. Aspirated stops are written kh [kʰ], etc. Words with an initial vowel have a glottal stop as onset (e.g. àn [ʔàn]), and words ending in a short vowel have a glottal stop coda (pə̀ [pə̀ʔ]).

Syllables and tones in Kammu

Like many other Mon-Khmer languages, Kammu has two kinds of syllables, usually called major and minor syllables. Minor syllables are unstressed pre-syllables with no phonemic vowel. Most words are either monosyllabic, consisting of one major syllable (e.g. tís ‘mushroom’), or sesquisyllabic, consisting of one minor and one major syllable, e.g. km̀mú [kə̀m.múʔ] ‘human being, Kammu’ and kmúul [kə́.múul] ‘silver’. In careful pronunciation, minor syllables are often pronounced with a schwa vowel [ə], not indicated in the phonemic transcription.

Like major syllables, minor syllables also carry a tone in the northern tonal dialect. In most cases, the minor syllable tone can be predicted from the segmental composition of the word. There is a potential minor syllable tone contrast only when the minor syllable consists of two consonants and has an unaspirated stop as onset. Examples of minimal pairs are pŋ́kà ‘to wear by the ear’ vs. pŋ̀kà ‘shy’ and pŕnɔ̀ ‘broom’ vs. pr̀nɔ̀ ‘carrying-sling’. Only about 30 minimal pairs have been found, whereas there are more than 900 minimal pairs (such as ráaŋ ‘tooth’ vs. ràaŋ ‘flower’) for the major syllable tone. In the transcription used here, the minor syllable tone is usually not indicated.⁸

Intonation

The main function of intonation in Kammu is to signal prosodic phrasing.⁹ The words in an utterance are grouped into phrases, and the end of each phrase is marked by a high (rising–falling) boundary tone. Here, ‘phrase’ is defined in prosodic terms, not to be confused with musical phrase. New information is placed utterance-finally, coinciding with the place of the boundary tone. This intonation pattern is found in both the tonal and the non-tonal dialects.

The lexical tones in the tonal dialects may interfere with the realization of the intonation pattern, but they do so only when the identity of a lexical tone is jeopardized. Example 1 is a schematized illustration of the intonation pattern (smooth curve) common to all Kammu dialects. Short horizontal bars show how the utterance intonation adapts to lexical tones. The high (rising–falling) phrase-final boundary tone is changed to low only when the combination of lexical tones in the final words is High–Low, since the identity of the Low lexical tone would be blurred by a high boundary tone.

Example 1 Stylized intonation pattern in Kammu. The curved lines show the dominating intonation pattern in both dialects. The shorter horizontal lines show how this pattern may be perturbed owing to the influence of the lexical tones (marked as H(igh) and L(ow)); in particular, a phrase-final High–Low lexical tone pattern (as in the final phrase) may eliminate the final rise.

The realization of lexical tones is relational (linear): a High tone is higher than a preceding Low tone, and a Low tone is lower than a preceding High tone. Low and High tones tend to be realized in different pitch regions.¹⁰ Henceforth, the terms low and high pitch region denote the different frequency areas used for Low and High tones in speech. A third, neutral pitch region is recognized in singing.

In addition to differences in F0,¹¹ timing of tones expressed as synchronization of the F0 peak in the syllable is relevant for describing lexical tones in Kammu. The tone is usually regarded as being realized in the entire rhyme (vowel + coda), at least in ‘typical’ Asian tone languages, such as Chinese or Thai. In Kammu, with no contour tones, the vowel kernel is a sufficient domain for realizing the tones, and tone is synchronized near the vowel onset, probably reflecting the involvement of syllable-initial consonants in tonogenesis.

Syllable reduplication and prolongation

A special kind of syllable reduplication, where the vowel of the base syllable is repeated, often occurs in recitation and song but not normally in speech. The reduplicated vowel may be short or long, sometimes longer than the vowel of the base syllable. Prolongation of the vowel or of a final sonorant is also common. Occurrences of reduplication and prolongation are genre-dependent.

Syllable reduplication is illustrated in Example 2, where vowel length is not indicated in the reduplicated form, as the duration may vary with the vocal genre. A dot (.) represents the syllable boundary. The coda of the base syllable often becomes the onset of the reduplicant (a). The reduplicant can also acquire the onset h or ʔ, especially when the base syllable is open, lacking a coda (b), but in other cases as well (c). Prolongation of syllables takes two different forms, either as a lengthening of the vowel or as a lengthening of a final sonorant: nàaŋ >nàa.ŋŋ. In very few cases, a sonorant onset is prolonged. The patterns of reduplication and lengthening differ between the different vocal genres.

Phonetic measurements

The main principle in analysing vocal expressions in Kammu was to regard the tonal phrasing pattern in speech (Example 1) as the underlying melodic default pattern for both tonal and non-tonal dialects.¹² The description concentrates on the extent to which the vocal genres retain the intonation of ordinary speech, and also on how they differ from speech in the treatment of lexical tones. In this way we can identify more or less speech-like vocal expressions. Pitch was measured in Praat, noting the measurements in Microsoft Excel and constructing graphs in Microsoft PowerPoint. For each syllable, F0 was measured at the vowel onset and at the next two turning points of the pitch movement within the rhyme. These F0 values were measured both in the original syllable and in the reduplicant (the reduplicated syllable). In some genres, syllables can have more than three turning points of pitch, especially in prolonged and reduplicated syllables resulting from the singer’s vibrato that belongs to the musical performance. Three turning points were measured, irrespective of the number of turning points within the syllable. In cases with less than three turning points, two points were taken. All measurements were performed manually. Furthermore, to compare genres the following parameters were taken into account:

• the amount of reduplication;
• the syllabic location of prolongations (onset, nucleus, or coda);
• tonal movements and their syllabic location (in original syllables or reduplicants) and their alignment (early/late within the syllable);
• the realization of lexical tones: by using separate pitch regions or by local pitch changes (relational realization).

The melodic transcription for each genre was checked for its relation to lexical tones. Melodies were analysed as underlying patterns of performance templates, similar to intonation in speech, and lexical tones as a separate level. This approach was used in order to establish whether melody only consists of lexical tones; whether there is a melodic pattern, separate from lexical tones, to which lexical tones are modulated; or whether the melodic pattern itself is modulated when conflicting with tones, as was seen to happen with speech.

An example of vibrato and measurement points is given in Example 3. The word is síi-pàay ‘bean’. The first syllable síi has High tone, the second syllable has Low tone, and both syllables are reduplicated: síi-Ɂi-pàa-yaa. Vibrato occurs in the reduplicant yaa. f0₁, f0₂ and f0₃ indicate the points of measurement in each syllable: f0₁ is measured at the vowel onset, and f0₂ (and f0₃) are measured at the next F0 turning points.

Example 3 Illustration of vibrato and measurement points of F0. The word is síi-pàay ‘bean’ with reduplicated syllables: síi-Ɂi-pàa-yaa. Female singer. Genre: tə́əm.

The vocal genres

The borderland between song and speech that is the focus of our research can be described as a section of a continuum of vocal expressions spanning from speech to ‘song’. When the Kammu vocal expressions are ordered on such a continuum, the poetic and musical complexity appears to increase the closer we get to ‘song’. Naturally, there will be some overlapping and other inconsistencies; but the image provides a useful overview. In this chain, each poetic technique may be seen as a special and more complex case of the preceding one: Reduplication → Lexical parallelism → Pivot-rhyme → Chain-rhyme → Cross-rhyme. Musically, the metrical and rhythmic organization becomes more fixed the closer we come to ‘song’, and this is paralleled by increased range and an increase in stable pitches (Table 1).¹³

The Kammu source material

Except where otherwise noted, the material referred to in this study stems from the Kammu research project at Lund University, which dates back to 1972 and lives on in the form of individual research and various short-term projects.¹⁴ The vocal expressions discussed here are unique in the sense that they all come from one performer, Kàm Ràw, born in the village Rmcùal in northern Laos, and were recorded over a long period of time. This repertoire has been chosen because many genres are represented and because it permits study of variations within the repertoire without the uncertainty that may be caused by individual variations. However, this does not mean that the research has been limited to this repertoire. On the contrary, many performances by representatives of the same dialect area, and other areas as well, have been analysed in order to verify the reasonableness of our results.

Table 1 Vocal genres in the speech–song continuum with regard to the poetic techniques used in this study

SPEECH							SONG
Lɔ̀ɔŋ (narratives)	Kàm à-thí-tháan (prayers)	Ceremonial expressions (Ɔ̀ɔc, etc.)	Krùu (spells)	Hrlɨ̀ɨ	Hrwə̀, Húuwə̀	Yàam	Yùun tíiŋ	Tə́əm
Analysis 1	Analysis 2	Analyses 3, 8a	Analysis 10	Analysis 4	Analyses 5–6	Analysis 7	Analysis 8b	Analysis 9
Anaphora Parallelism Vowel or consonant rhymes
	Pivot rhymes
		Chain-rhymes
				Cross-rhymes (trnə̀əm)
								Rhymed words of address

Analysis 1 Lɔ̀ɔŋ, narrative style

Meaning approximately ‘chanting’, lɔ̀ɔŋ is, in a general sense, used for various vocal genres. Here it will be exemplified by a specific style of vocal expression that occurs in narratives about daily activities, and in one prayer with the aim of showing the soul of a deceased person the way to the village of the dead.

By analysing syntactic and tonal properties of everyday speech, some types of narratives were found to have similar information and prosodic structure not found in other genres, for instance folktales. These narratives typically list certain types of activities. Most of the recordings are spontaneous accounts of rice-growing, from which it was possible to establish features typical for this genre. Compared to other speech genres, a high degree of rhythmization is found in the narrative accounts. Rhythmization is attained by tonal, segmental, and syntactic means. The same features were found in Lɔ̀ɔŋ ŋɔ̀ɔr, a prayer to guide the soul of a deceased person to the spirit village of his/her ancestors. Though ‘rice narratives’ belong to everyday speech and Lɔ̀ɔŋ ŋɔ̀ɔr has a ceremonial context, both are lists of activities and share similar syntactic and tonal patterns. This leads to systematic long-term mutual relations of boundary tones marking information units and thematic shifts. Structural means of signalling information structure are uniform through this kind of narrative. In this case, we may speak about a fixed performance template used in narrative accounts.

Informational structuring of rice narratives

The speech material analysed here consists of ‘rice narratives’, spontaneous accounts of rice-growing, from the first period of work in the field until the rice is steamed and eaten. Rice narratives related by 14 tonal and 10 non-tonal speakers were recorded in Laos in 2007 and in Thailand in 2008.¹⁵

Very few disfluencies (interruptions, hesitations, etc.) are found, as speakers are describing the growing of rice, an everyday activity well known to most of them. Although spontaneous, this material is structured in the same way by all speakers: new information is given at the end of an utterance (underlined in Example 4) and is then repeated as old information in the immediately following utterance. Using square brackets to mark prosodic phrases, the informational structuring is: [anchor + new₁] [old new₁ + new₂] [old new₂ + new₃] … The new information becomes an anchor point (old information) in the next utterance. There are thus a considerable number of repeated words in the speakers’ monologues, while anaphoric reference is not used as in this example: Before there is rice we have to clear the field […] After clearing the field we burn the field […] After burning we sow. The verbal contents can thus be seen as a list of successive events. Some speakers use only one utterance per event, while some include much additional information.

Structuring of the informational flow into [given + new] units is systematically cued by prosodic means. Thus, the tonal prosodic phrase boundaries in the [given + new] units are upstepped in relation to one another: each succeeding boundary is phonetically realized with a higher F0 value.¹⁶ The final word of the last prosodic phrase gets the highest F0 value, coinciding with the [new]. This is fully realized when no conflict with lexical tones occurs. The high boundary tones are on the right edge of prosodic phrases in Kammu, as illustrated in Example 1. This is also the case for rice narratives, as shown in Example 5 by the tonal course of the utterance in Example 4 spoken by a female tonal speaker.

Example 5 Tonal course of the utterance in Example 4. The division into three [given + new] units is shown by vertical lines. The high boundary tones of focused phrases, occurring on the underlined words in Example 4, are shown with arrows. Note that the final syllable starts rising, as expected for a boundary syllable. A Low lexical tone may suppress the rising movement. • 01 Rice narrative

The following part of a rice narrative stems from a male speaker of a non-tonal Kammu dialect. The basic structure is the same as in Example 4. Activities mentioned at the end of one phrase are immediately repeated at the beginning of the next phrase (Example 6).

In this case, too, the tonal course of the utterance shows the same characteristics with high boundary tones at the end of prosodic phrases (Example 7).

Example 7 Tonal course of the utterance in Example 6. The end of each [given + new] unit is indicated by an arrow.

In Melodyne a thin curve shows the detailed movement, and pitches are approximated as straight lines of varying width, so-called ‘blobs’, which serve as a base for the automatic notation. This process is similar to what happens in the manual transcription of music. When applied to rice narratives, the tonal course is clearly visible and the musical notation is fairly accurate. The automatic score notation gives a reasonable description of the melodic and rhythmic movement, though one should be aware that bars and barlines are not optimal (Example 8).¹⁷ The pitches are relatively fixed but not organized in a tonal system. The rhythmic pattern is notably regular, without having a regular beat. In these respects lɔ̀ɔŋ differs from ordinary daily language and also from vocal music.¹⁸

In the illustration of the non-tonal speaker in Example 9 the same characteristics are obvious, but there is a definite difference between the two performances in tonal range: 8 semitones in Example 8, compared to 13 in Example 9. There are no cases of syllable reduplication, but the number of syllables in the phrases differs. The corresponding numbers of syllables are 4–6–5 in three phrases and 6–6–6–8 in four. The manner of performance thus permits varying phrase lengths, which produces some rhythmic irregularities.

Example 8 The segment from the rice narrative in Examples 4–5 in a Melodyne graph. Arrows show boundary tones of focused phrases. The words and the lines in the notation that show downwards sliding motion have been added manually. • 01 Rice narrative

Example 9 The segment from a rice narrative in Examples 6–7 in a Melodyne graph. Arrows show boundary tones of focused phrases.

Division of narratives into topics

A thematic episode is seen as an informatively coherent part of discourse, with a clear beginning and end.¹⁹ Each episode has one central topic. As we are dealing with narratives about a traditional activity, the Kammu agricultural calendar as described by Kàm Ràw (Table 2, right column)²⁰ was used as a reference framework for division into thematic episodes. Most speakers use the topics listed in Table 2, left column. Some speakers have additional topics, such as making field-houses, protecting crops from animals, or different ways of cooking rice. All topics in the left column except (5) and (7) occur with all speakers.

Each thematic episode includes several [given + new] units. A schematic illustration of the prosodic organization of narratives is provided in Example 10.

Speakers structure thematic episodes in their narrative account as [description + name of activity that is described]. For example: We go to seek a field, seek in the forest, after finding the field we clear, the part before clear being the description. Clear is the name of the activity and coincides with the end of the thematic episode. Clear is thus the topic of the thematic episode, and all the previous descriptions (seek a field, seek in the forest) are about what is included in clearing. However, in some cases there is another type of structuring of topics, where the topic is introduced at the beginning and then described; here is an example: We seek a place we will clear, yes, search the forest, look for a place that will be good for the rice and we clear. In this case, clear is introduced in the beginning as a new topic, and its development comes afterwards. This kind of topic gets the highest F0 at the beginning of the topic instead of at the end. It should be observed that clear is repeated at the end, obviously indicating that the episode about clearing is finished here.

Example 10 A schematic illustration of the performance of lɔ̀ɔŋ. Right-edged boundary tones of prosodic phrases (p-phrase) are organized in a template that marks the informational structure of the narrative. Smooth lines illustrate the approximate tonal contour. The template is similar for tonal and non-tonal dialects, but is perturbed by lexical tones in the tonal dialect.

Tonal structuring of rice narratives by boundary tones is exemplified in Examples 11 and 12. As the narratives are long, only the last words in [given + new] units are included. A non-tonal speaker (Example 11) has 10 episodes in his narrative. He was living in a village in Laos and is familiar with work in the field.

Example 11 Rice narrative related by a non-tonal male speaker, Phet Smai from the Konkeo village, Vientiane Province. Top: Phrase final words coincide with new information. Square brackets show the division into thematic episodes. Bottom: Tonal movement with tonal boundaries marking ends of thematic episodes, indicated by arrows.

One tonal speaker had moved to a town in Thailand and only knew about rice-growing from his parents. His narrative is shorter with fewer details, consisting of 5 episodes (Example 12). Nonetheless, the melodic movement is similar to that of the non-tonal speaker in Example 11. The highest boundary tones (shown by arrows) mark the end of thematic episodes.

Example 12 Rice narrative related by a tonal male speaker, Somswàt Búnkə́ət from the Òm Kɔ́ɔ village, Bo Keo province, Kwὲɛn dialect. Tonal boundaries marking ends of thematic episodes are indicated by arrows.

[rɨ́am]	[pɔ́ɔk, réʔ, púur]	[cmɔ̀ɔl, hɛ́ɛl, hɔ́ɔt, híc]	[kùum, rùŋ]	[pə̀ʔ]
[clear]	[burn, field, burn]	[sow, weed, pound]	[winnow steam]	[eat]

Lɔ̀ɔŋ ŋɔ̀ɔr, ‘Showing the way’

A syntactic structure similar to that of rice narratives is found in the Kammu Lɔ̀ɔŋ ŋɔ̀ɔr, literally, ‘chanting the way’, which belongs to the funeral wakes where it was performed in order to guide the soul of the deceased to the spirit village where he/she would then live as an ancestral spirit. The extract in Example 13 consists of three lines from a performance of a total of 86 lines.²¹

The initial part is structurally similar to the rice narratives: new information is given at the end of prosodic phrases and repeated in the beginning of the following phrase. The phrases achieve a high tonal boundary (Example 14, bottom panel).

A similar pattern is found in a performance of Lɔ̀ɔŋ ŋɔ̀ɔr in the Rɔ̀ɔk dialect area, south of Yùan. This performance is realized as a call-and-response pattern, where the lead performer’s lines are immediately repeated by those present. As illustrated in the top panel of Example 14, the right phrase boundary is marked by a high boundary tone (suppressed by a Low lexical tone in the second phrase), and phrases comprising new information (the second and the last one) are realized in a lower pitch area. This is also the case in the Yùan-area example at the bottom panel of Example 14; in addition, it is clearly shown by the graph in Example 15. As was the case with the rice narratives, pitches appear to be quite stable, and the performance is rather rhythmic.

Example 14 Tonal course of two Lɔ̀ɔŋ ŋɔ̀ɔr performances showing alternating high and low pitch areas related to a call-and-response performance manner. The horizontal line indicates an approximate division into high and low pitch levels and arrows show ends of phrases. The performers are Tá Khám, Lwà village, Rɔ̀ɔk dialect area, ca. 1980 (top) and Sét Mán, Yùan area, ca. 1980 (bottom). Both are male tonal speakers. • 02 Lɔ̀ɔŋ ŋɔ̀ɔr

As mentioned above, everyday spoken narratives (rice monologues) are highly rhythmic, and their prosodic and syntactic features are very similar to those of Lɔ̀ɔŋ ŋɔ̀ɔr. Tonal phrasing and the use of lengthening are similar in the two genres. The main difference is the use of pitch levels. In Lɔ̀ɔŋ ŋɔ̀ɔr, the speaker shifts between higher and lower pitch levels (registers) by performing every second phrase at a lower level. This comes out quite naturally in the call-and-response pattern of the Rɔ̀ɔk-area performance (Example 14 top), where those present repeat each new sentence in unison at a lower pitch. Kàm Ràw would never perform Lɔ̀ɔŋ ŋɔ̀ɔr outside its actual function at a funeral, but he said that this alternation was the normal way of performing it in the Yùan area. This might explain why the Yùan-area performance (Examples 14 bottom and 15) also has lines alternately in a higher and lower pitch area, even though there is only one person who performs it. Both the narrative accounts and Lɔ̀ɔŋ ŋɔ̀ɔr may be rather long. Perhaps the alternation of phrases in a high and a low pitch area in Lɔ̀ɔŋ ŋɔ̀ɔr may be seen as an organizing principle.

Example 15 The melodic movement of the excerpt of Lɔ̀ɔŋ ŋɔ̀ɔr from the Yùan area in Example 13 showing alternation in pitch register where prosodic phrases in a high register (boxes with solid lines) are followed by phrases in a lower register (boxes with dashed lines). Notes within brackets are sobs or spoken words. • 02 Lɔ̀ɔŋ ŋɔ̀ɔr

The lɔ̀ɔŋ performance template

The difference between language and music – or spoken language and vocal music – is often expressed by the pulsating-rhythm and tonically ordered pitch parameters. While these are generally lacking in spoken language, they are present in vocal music. In these respects, lɔ̀ɔŋ is closer to spoken language than to vocal music; but it has a tendency towards repeated rhythm and pitch patterns, which are comparatively stable. Lɔ̀ɔŋ differs both from spoken language and from vocal music and fits into the borderland under study here.

Melody

• Melodic movement with relatively fixed pitches without a tonal centre (see Examples 8–9, 15).
• Range: 8–13 semitones.

Rhythm

• Rhythmic movement in rather uniform repeated patterns without a regular beat.

Form

• Litany: succession of prosodic phrase-pairs connected by repeated word(s).
• Call (higher pitch level) and response (lower pitch level) pattern occurs (Lɔ̀ɔŋ ŋɔ̀ɔr).

Phrasing

• Successive right-edged phrase boundary tones, with the highest phrase boundary tone coinciding with the end of each episode.
• New information is given at the end of an utterance and is then repeated in the immediately following utterance: [anchor + new₁] [old new₁ + new₂] [old new₂ + new₃] …

Word variations

• Schwa vowels (ə) are usually not audible.

Lexical tones

• Low lexical tone may make the boundary tone at the end of an episode low.

Analysis 2 Kàm à-thí-tháan, prayer

Prayers directed at a large number of spirits, róoy, accompany many activities in Kammu life.²² Numerous Kammu prayers have been published, but there are fewer recordings. Two prayers used in shaman ceremonies are discussed in this analysis.²³

Ì càk cə̀ən, ‘We called you’

These words may also be referred to as Krùu kʔə́əy mà krùu, ‘Spell for apologizing to the magic power’. They belong to the part of a shaman ceremony when the shaman spirits have arrived and sing through the shaman as a medium asking why the people called the spirits: Mɔ́ɔ rɔ̀ɔt màañ, ‘The shaman arrives and asks’. The people may reply with these words:

Prosodically Ì càk cə̀ən resembles ordinary speech, and prosodic phrasing is built on syntactic groups (Example 17). The syntactic groups consist of two parts linked by pivot rhymes (kháat/pyàat, cə̀ən/ə̀ən, pát/pát, tə̀lhát/pát).²⁴ Schwa or epenthetic vowels are pronounced ə (tə̀lhát and kə̀séh).

The performance starts high, and the first line falls (Example 18). The remainder is within the rather narrow range of approximately 6 semitones, and the intervals between syllables are small. The smaller prosodic groups are cued by a high boundary tone at their right edge. Lines 3 and 4 end on high boundary tones, The last of them is the highest, which is the normal way of marking the end of a continuous narration in Kammu. It should be noted, however, that line 2 ends on a Low lexical tone while the other lines end on a High. The small prosodic phrases are grouped into larger prosodic phrases²⁵ which, in addition to the final high boundary tone, are also marked by lengthening the final syllable. After the opening phrase, there is a fairly regular rhythm. The tempo is fast, with 33 syllables in 9 seconds – an average of nearly 4 per second, compared to an average of 2.3 in the Lɔ̀ɔŋ ŋɔ̀ɔr in Example 15. In lines 2–4, the average is 5 per second.

The overall shape of this vocal expression is that of a high and falling start, beginning with a possibly vocative əə. Then follows a swift, gradually rising delivery of words at fairly narrow intervals, where phrasing is marked by starting each prosodic phrase on a slightly higher F0 than the last word of the preceding phrase (Example 18). The prayer then ends with an upward motion. It may be noted, however, that there are several examples of movement contrary to lexical tones and that the tones and pitches tend to coincide on the final syllable of a syntactic group. For the most part, the tonal movements within syllables are flat or falling.

Example 18 Melodic movement of the prayer Ì càk cə̀ən, ‘We called you’. The arrow indicates gradually rising phrases. Performed by Kàm Ràw, ca. 1983/84.

Wéey mὲɛ krùu, ‘Worship one’s power’

The shaman speaks these words, also referred to as Krùu kʔə́əy mà krùu ‘Spell for apologizing to the magic power’, in the form of a prayer after having made a mistake in a spell, or before starting in order to prevent punishment if a mistake should occur:

The first and last lines are addressed to the shaman spirits. The underlined final word of the second line (krùu), rhymes with the first word of the third line (kúu). Lines 3 and 4 are linked by parallelism, as are lines 5 and 6. The introductory word sáa-thú is performed rather slowly as an opening, and with a falling interval. The speed is high, similar to the previous prayer.

As regards phrasing (Example 20), there is no resemblance to daily speech or rice narratives, so the ending of prosodic phrases have no high boundary tones. Syllable lengths within phrases are short or long, and the final syllable of a phrase is lengthened, marking the end of a phrase. In addition, the initial syllables consisting of the vocative sáa-thú are lengthened. Compared to other genres that are delivered at a slower tempo, as for instance in tə́əm (Analysis 9), the fast tempo does not allow room for syllable reduplication. Though lexical tones are sometimes realized, every syllable is performed with a falling pitch, which is unusual in Kammu genres (Example 21).

Example 21 Wéey mὲɛ krùu, ‘Worship one’s power’. Because of a high tempo, many words are performed in very short time spaces. Therefore, some parts of the transcription have been omitted. Performed by Kàm Ràw, ca. 1983/84.

If the introductory word is disregarded, the melody has less movement than the prayer Ì càk cə̀ən, ‘We called you’ (Example 18); and although the prosodic phrases produce a certain regularity, it is not as clearly organized into distinct rhythmic units as the genre lɔ̀ɔŋ (Analysis 1). Most of the time, it moves in the narrow space of approximately 4 semitones, and boundary tones are not particularly marked. Especially the end of the melody is falling, independently of the lexical tones. The feeling is that of a rather flat melody with very little room for intonation.

The kàm à-thí-tháan performance template

Melody

• Melodic movement is rather level but slightly rising (disregarding initial formula), with no tonal centre (see Examples 18, 21).
• A narrow range of 4–6 semitones (disregarding initial formula).

Rhythm

• Fast tempo.
• Rhythmic movement is basically created by the lengthening of final syllables in prosodic phrases.
• Rhythmic movement in rather uniform repeated patterns, with no regular beat.

Form

• Litany: consecutive prosodic phrasing built on syntactic groups with pivot rhymes.

Phrasing

• Smaller syntactic groups end on a high boundary tone, each slightly higher than the preceding group, with a lengthening of the final syllable.
• Larger prosodic phrases are marked by the lengthening of the final syllable.

Initial/final formulae

• The initial phrase is vocative, starting at a high pitch and falling with lengthened syllables.

Word variations

• Schwa vowels (ə) are very short and hardly audible.

Lexical tones

• Lexical tones are not very pronounced.
• Movement contrary to lexical tones is common.
• The melodic movement within syllables is flat or falling.

Analysis 3 Ɔ̀ɔc, ‘Begging’, at New Year

In late autumn, when the rice had been harvested and the master of the house had sacrificed taro, sweet potato, and elephant grass to the ancestor spirits in order to inform them that the old year was ended, young people used to gather in small groups and wander all over the village to collect gifts. This wassail went on for three consecutive evenings, and each household would have several visits.

The youths sang outside the houses of the village, expecting gifts consisting of taro, sweet potato, and peanuts which the house owner would first hold over the stove and over the table while praying to the spirits of waste to leave those places and to leave his house. The wassail was thus intended to drive out the spirits of waste, i.e. róoy yáap and róoy hʔéep, which are the spirits of stillborn babies and of those who died in accidents. It was believed that these spirits could not find repose in the village of the dead, but were doomed to roam around alone. Unlike other spirits, the spirits of waste did not receive blood sacrifices. Therefore, people believed that these spirits were constantly hungry, that they would be attracted to the fields by the growing rice and that they might follow the harvested rice into the village, houses, and barns. In such a case the crop would not last long, or it would rot.

On the evening of the third wassail day, the youths would gather in one of the common-houses of the village to have a feast on the food they had collected. A little of each offering was saved, however, and carried to the village boundary, where it was thrown away towards the west – that is, where the sun sets. This was how they disposed of the waste spirits.

This wassail was called Ɔ̀ɔc, ‘Begging’. The words are organized into units, within which one of the final words of a line rhymes with one of the first in the following line. Each line consists of 5 syllables followed by the non-lexical word ís, which makes 6 syllables. The only exception is the first line (A in Example 22), which consists of 3 + 1 syllables. The non-lexical word ís that occurs at the end of each line should probably be seen as a ‘song-word’. Since this is the only known occurrence of the syllable ís, it has been assigned a High lexical tone based on the way it is performed in Ɔ̀ɔc.

The melody used in the village of Rmcùal consists of a short motif that is repeated over and over with minor variations. The range is a sixth (9 semitones), and the rhythm is even and suitable for walking. The syllables 2, 4, 5, and 6 have fixed pitches throughout the performance. Since each of these (except syllable 6) is combined with High or Low lexical tones, agreement between pitch and lexical tone is accidental. Syllables 1 and 3 have two possible pitches, which permits adaptation to High or Low lexical tone as shown by the arrows in Example 22. Lexical tones are thus only partly realized in the performance.

Example 22 The three variants of the Ɔ̀ɔc melodic motif. A is the initial formula; B and C are repeated throughout. Arrows mark the only pitches that vary and can be adapted to High or Low lexical tone. • 03 Ɔ̀ɔc

As can be seen from the translation (Example 23), the verbal content of the song is an enumeration of details that explain why the singers are visiting a household. Except for the initial formula built on words without lexical meaning, the prosodic phrases are organized in pairs. These phrase-pairs are held together by parallelism and rhymes arranged so that one of the last syllables of a line rhymes with one of the first in the following line: for instance lines 6/7: tə̀-làa / pə̀-yàa or 10/11: lə̀m-táas / kráas. The vowel ə is a schwa vowel that is more or less inaudible in ordinary speech. The word tə̀-làa is normally written t̀làa, but since the schwa vowel is given the same length as any other vowel in the performance, it is written out in this transcription.

In Kammu practice, most melodic phrases end on a low pitch. The rising pitch at the end of the Ɔ̀ɔc melodic motif is hence unusual. The enumeration in the verbal contents and the paired phrases probably reflect the same prosodic practice as the lɔ̀ɔŋ narrative style (Analysis 1), but are here combined with a melodic motif with a regular rhythm and a tonal centre.

Number notation is used for the transcriptions, since it provides a description of melodic and rhythmic features that is sufficiently detailed for our purpose, which in this case is to determine whether lexical tones are realized in the performance or not (Table 3; for more details see Appendix 2). The aim has been to try out a simple transcription that is easily communicated and yields a clear picture of musical pitch vs. lexical tone, and also of metre and poetic form. The transcriptions are therefore also arranged so that parts that can be interpreted as initial and final formulae of a line are directly aligned below each other and can easily be compared.

Table 3 Number notation limited to the signs used in the following examples^*

Half note (minim):		1 –
Quarter note (crotchet):		1
Dotted quarter:	.	1·
Eighth note (quaver):		1
1 = tonic.
If 1 = c then 1–2–3–4–5–6–7 corresponds to c–d–e–f–g–a–b
Pause:	0
Lowered (by a semitone):	♭1, ♭2 etc.
Glissando (sliding between tones):	⇒

Note: ^* For a complete description, see Internet reference: Numbered musical notation in the References.

The Ɔ̀ɔc performance template

In our material, this melodic formula is only used with the words of the Ɔ̀ɔc wassail festivities. The short melodic formula is repeated for each line of the words which, in principle, are organized as phrase-pairs connected through rhymes. It might be discussed whether the concept ‘performance template’ is relevant for this melodic formula. On the other hand, it cannot be ruled out that this melodic formula could function as such, for it has some characteristics that generally occur in Kammu performance templates.

Melody

• A short melodic formula with a tonal centre (see Example 22).
• Melodic range: 9 semitones.

Rhythm

• Regular pulse.
• Syllabic.

Form

• Litany: prosodic phrase-pairs generally connected through rhymes.

Phrasing

• A high boundary tone coincides with the end of a phrase.
• Prosodic and musical phrases are aligned.

Initial/final formulae

• An initial formula at the very beginning
• A final song-word in each phrase (ís).
• The penultimate syllable of a line is lengthened.

Word variations

• A song-word without lexical meaning at the end of each line: ís, which is always performed high.
• The initial formula consists of song-words.
• Same duration for long and short vowel, minor and major syllable.
• Schwa vowels (normally ə) have the same duration as all other vowels that are performed as short.

Lexical tones

• Melody-centred.
• Lexical tones only partly realized.

Analysis 4 The vocal genre hrlɨ̀ɨ

In the Kammu village of Rmcùal in northern Laos, the predominant vocal expression in feasting situations was tə́əm, which could be used only in such situations. If the same kind of orally transmitted poems – called trnə̀əm – were to be performed at other times, the vocal genre hrlɨ̀ɨ would be used instead. Hrlɨ̀ɨ could also be used outside the village, not least in the fields where many young people spent much time watching over the growing rice. It also functioned as a mnemonic tool for learning and remembering the trnə̀əm poems. Generally speaking, hrlɨ̀ɨ is a rhythmicized way of performing a trnə̀əm freed from speech intonation, but with the lexical tones realized.

Example 24 shows the words of a typical trnə̀əm, which is the kind of orally transmitted poem used for hrlɨ̀ɨ and some other vocal genres. It is organized into two stanzas. Generally, the second stanza contains a more or less concrete message, whereas the first stanza is more metaphoric and serves as a ‘poetic parallel’ to the second. A high degree of parallelism is evident from the repetition of phrases in which only a few words are exchanged in the repetitions. The words that are exchanged are most often rhyme words, and the rhyme pattern stretches over both stanzas. This particular trnə̀əm is of the kind that would be used when visiting somebody in another village more or less uninvited.²⁶

The symbols after each line are used to identify lines of stanzas. The numerals stand for stanzas (1, 2, and so on), the letters for lines within stanzas (a, b, and so on), and the index for variations of lines (´). In the ‘basic’ form of the trnə̀əm in Example 24, the order of lines is: 1a–1b–1a–1b´, 2a–2b–2a–2b´. In Example 25, the same trnə̀əm has been reorganized and the lines of the poem have been regrouped into 1a–1a–2a–2a, 1b–1b´–2b–2b´. This happens in performance, but here it serves to show consistencies in the intonation of lexical tones and to bring out the pattern of parallelism and rhymes.

Hrlɨ̀ɨ is strictly syllabic, i.e. each tone corresponds to one syllable. For the most part, it employs one tone duration only and has a regular pulse – the exceptions are the penultimate syllable of a line, which is longer, and occasionally also the very last syllable of a stanza. With these exceptions, each syllable is given the same duration, regardless of vowel length. A minor syllable is, without exception, treated in the same way as a major syllable and is thus given a much longer relative duration than in speech.²⁷

Prolongation takes place in syllable-final sonorants (nɔ́ɔŋ ‘still’ > nɔɔŋŋ) and by inserting epenthetic vowels into minor syllables and prolonging them (rŋ.kɨ̀l ‘no more than’ > rəŋŋ.kɨl). Thus, prolongation is achieved using positions in which the length contrast does not occur. This differs from the situation in normal speech, where minor syllables are unstressed and have a very short epenthetic vowel that often disappears completely.

A hrlɨ̀ɨ performance has three pitch levels: high, low, and neutral, and a tonal centre. The main part of the performance is limited to two pitch levels, high and low – the only exception being a few syllables at the beginning of a poetic line after a pause, which are performed as neutral tones at lower pitch and may be considered as an initial formula. The size of the interval between the two dominating pitches varies from a second to a minor third (2–3 semitones), which means that the pitches can be easily recognized by ear. The high and low pitches are almost invariably used for High and Low lexical tone respectively. In Example 25, the repetitions of words in this performance are grouped in boxes to demonstrate that the same lexical tones are performed in the same way each time they occur.

Example 25 An approximate transcription of a hrlɨ́ɨ performance of the trnə̀əm in Example 24, in which the lines have been reorganized so that the consistent realization of L = Low and H = High lexical tones can be readily seen. After a performance by Kàm Ràw.

In the investigation, we used a material consisting of performances by one Kammu speaker, Kàm Ràw (Damrong Tayanin), speaking the Yùan sub-dialect of the Northern dialect.²⁸ A ‘studio sample’, which was performed at the informant’s own initiative, consisted of 12 performances and 909 syllables in total. A ‘laboratory sample’, which was performed at the request of the researcher (HL), consisted of 24 performances and 1,393 syllables. Finally, there was an ‘experiment sample’ consisting of another 12 performances that the informant had not performed in this style before. This was made in order to test predictions about rhythm and pitch relative to lexical tones. A fourth sample consisted of 2 performances by other Yùan informants, bringing the total up to 50 performances.

The results showed that the hrlɨ̀ɨ material followed the linguistic analysis of syllabification and of the individual tones, both on major and minor syllables. There were no mismatches at all in syllabification and the tonal mismatches were extremely few, less than 0.5%.

The trnə̀əm presented in Example 26 praises a woman’s beauty. This is made clear in the second stanza. The poetic form is similar to that of Example 24. This trnə̀əm will be used as an example of several vocal genres starting with hrlɨ̀ɨ in Example 27, since, by using lines from the same trnə̀əm, the differences between vocal genres will become clearer.

In Example 27, musical notation has been used for the transcription in order to indicate the sound of a hrlɨ̀ɨ performance. No bar-lines have been used as the musical phrases evidently depend on the poetic lines, in this case combinations of 5 and 7 syllables. For the same reasons, the staffs are aligned after the phrase-endings. When lexical tone and musical pitch coincide, this is indicated by capital letters, L and H, respectively. In this performance, they always coincide; but there are examples where High lexical tones are performed to the lowest, ‘neutral’, pitch which functions as an initial formula – in this case, the first notes of lines 1, 3, 5, and 8. Lines 3, 5, and 8 start with the short added word sáh, ‘I say’, at the lowest, ‘neutral’, pitch level. This is a common auxiliary word in hrlɨ̀ɨ and some other vocal genres. It has not been transcribed in the example. The last two tones of a line are performed in the same way, regardless of lexical tone, and hence function as a final formula.²⁹

Example 27 Hrlɨ́ɨ performance in notation. The word sáh, approximately ‘I say’, is an extra word often used at the beginning of lines. It is very short and does not affect the rhythm. Performed by Kàm Ràw, ca. 1979/80. Original pitch: c ≈ 130 Hz. • 04 Nàaŋ mὲɛn, hrlɨ́ɨ

H: High lexical tone and high pitch level coincide.

L: Low lexical tone and low pitch level coincide.

l, h: Lowest pitch level used in initial formula regardless of lexical tone (parentheses show the initial formulae).

This particular trnə̀əm will be often referred to in the following, as it will be used for analysing several Kammu genres. In Example 28, the first stanza is employed in order to demonstrate some basic characteristics of hrlɨ̀ɨ performance. It demonstrates that the majority of the words are performed at two pitches (around C3 and D3, respectively). The former are all syllables with Low lexical tone and the latter with High. It may also be noted that the initial 2–3 syllables of lines 1 and 3 are much lower – these sections serve as initial formulae at the beginning of lines, usually after a breathing pause.

Example 28 A Melodyne graph of the first four lines (= first stanza) of the performance in Example 27. High lexical tone is realized at a rather fixed high pitch relative to the Low. [ ] = the extra word sáh with High tone is normally performed low in the initial formulae. Vertical: pitches, horizontal: time (one column = 1 second).

The performer realizes High tones in the region above approximately 130 Hz, while Low tones are between 100 and 130 Hz. Lines start below 100 Hz and the separation between the tones is rather limited in this region. Examples 29–31 show the details of the lexical tone as well as the movement within them. It may be noted that there is very little movement in those words with Low lexical tones that are realized as low, while there is movement in words with High tone.

Example 29 Pitch measurements of lexical tones and movements within syllables in phrases 1–4 of the performance in Example 27. In this graph, tone length is disregarded. Vertical: Hz, horizontal: time.

In Examples 30 and 31, phrases 5–6 and 7–8, respectively, are illustrated in the same manner. Two measurement points per base syllable are shown: the initial F0 and the first subsequent F0 maximum within the same syllable. Dashed lines show approximate pitch registers for low and neutral (below 100 Hz) lexical tones. It can be noted that High lexical tones are performed with a sliding upward motion when approached from the lowest ‘neutral’ pitch register (Example 30: the fourth syllable kɔ́ɔn and Example 31: the tenth syllable kɔ́ɔn). Similarly, a High lexical tone followed by a Low can be performed with a sliding downward movement (Example 30, the eighth syllable ˀyúut).

Example 30 Pitch measurements of phrases 5–6 of the performance in Example 27.

Example 31 Pitch measurements of phrases 7–8 of the performance in Example 27. The auxiliary word sáh (syllable 6) with High lexical tone is performed in the higher area of the lowest pitch and glides downwards.

The hrlɨ́ɨ performance template

Melody

• Level melody without speech intonation with a tonal centre (see Example 27).
• Two pitch levels for the main part of the performance (high and low) with an interval of about 2–3 semitones.
• Melodic range: 9 semitones.

Rhythm

• Strictly syllabic, with a regular pulse.
• One tone duration dominates.

Form

• Strophic (trnə̀əm poem).

Phrasing

• Prosodic and musical phrases are aligned.
• Verbal phrasing dominates (a 7-syllable line is two tone durations longer than a 5-syllable line, etc.).

Initial/final formulae

• In an initial formula, syllables at the beginning of a poetic line are performed at an extra-low tone. Occurs at the very beginning and when a line starts after a pause.
• The penultimate syllable of a phrase and the very last syllable of a stanza are longer.

Word variations

• Same duration for long and short vowel, minor and major syllable.
• An auxiliary word sáh, ‘I say’, often occurs at the start of phrases, performed very short and at an extra-low level.
• Schwa vowels (normally ə) have the same duration as all other vowels that are performed as short.

Lexical tones

• Lexical tones are realized, except in the initial formula.
• High and low pitch levels correspond to lexical tones.
• Tones are realized within separate pitch regions (Low tones in the 100–130 Hz region, and High tones above 130 Hz).
• High lexical tones are performed with a sliding upward motion when approached from a lower pitch. A High lexical tone followed by a Low is performed with a sliding downward movement.
• There is very little movement in words with Low lexical tones that are realized as low, while there is movement in words with High tone.

Analysis 5 The vocal genre hrwə̀

Hrwə̀ is used particularly by young people, especially when away from the village, out in the fields, or in the forest. It was used for trnə̀əm with contents that fitted these contexts. The performance is built on a short melodic motif that is repeated with small variations (Example 32). It makes use of three basic pitch levels. In details, there is much variation in the execution of pitches and in the rhythmic delivery. The highest and the lowest pitch levels are approximately a fourth (5 semitones) apart. The middle pitch is normally a minor third (3 semitones) above the lowest. The last note has a downward slur with an indefinite final pitch approximately a fourth below.

Example 32 Hrwə̀ musical phrases of a 5- and a 7-syllable line (for comparative reasons the syllables are numbered backwards). ‘V’ refers to vowel reduplication and ‘v’ to prolongation (with glissando). • 05 Nàaŋ mὲɛn, hrwə̀

Using two lines from the trnə̀əm in Example 26, a Melodyne graph is shown in Example 33. The melodic shape is clear and the graph also demonstrates the difference between 5- and 7-syllable lines: the 5-syllable line is prolonged by adding two more syllables. When a line is prolonged by added syllables, they are added in the middle of the line without affecting the initial or final formulae. On the other hand, the long final tone is shortened so that the length of each phrase is approximately 7 seconds. In that way the musical metre is preserved.

Example 33 Hrwə̀ performance of the trnə̀əm Example 26. Phrase 3: Nàaŋ mὲɛn trə̀ mà mɔ̀ɔn (5 syllables not counting reduplicants) and phrase 4: Ɔ̀ɔn-crɔ̀ɔn nàaŋ kɔ́ɔn mɨ́an ɔ̀ɔn yòl (7 syllables). Performed by Kàm Ràw, ca. 1979/80. Original pitch: final tone ≈ 130 Hz.

In some genres of vocal expressions in Kammu culture, there is vowel reduplication. This means that the vowel of a syllable is repeated, and the reduplicant is often much longer than the original vowel. In this particular case reduplication takes the following form (the reduplicant is underlined). For full text and translation, see Example 26:

In hrwə̀, every line is divided into two or three prosodic groups marked by a higher first lexical tone irrespective of L or H and by successive lowering (declination) of pitches within each prosodic group (Example 34). Lines visualize pitch declination. Each new prosodic group starts with a high pitch which is usually lower than in the previous group. Within a declining group, High tones are usually performed lower than preceding Low tones. If the first lexical tone in a phrase is High, it is boosted, as in the case of the word kɔ́ɔn in Example 34.

Example 34 Hrwə̀. First phrase of the trnə̀əm Example 26 showing melodic descent and a case where a high lexical tone (kɔ́ɔn) is boosted so that the second phrase starts higher than the preceding one.

Example 35 displays F0 measurement points for each syllable, including reduplicants, and also in sonorant codas (when prolonged). For example, in nàaŋ, six points were measured: start, middle, and last, both in the original syllable nàa and in the reduplicant -ha. Frequencies (in Hz) are shown on the vertical axis. Lines visualize pitch declination through each of the three prosodic groups. Creakiness often occurs at the end of a prosodic group and might signal a phrase boundary, as happens with regard to speech in several languages.

Example 35 Hrwə̀. First phrase of the trnə̀əm Example 26. • 05 Nàaŋ mὲɛn, hrwə̀

The hrwə̀ performance template

Melody

• Formulaic descending melody with a tonal centre (see Example 32).
• There is successive lowering or declination of pitches within each prosodic group.
• Syllables have a rising-falling pattern.
• Melodic range: 5 semitones.

Rhythm

• Regular pulse.
• One tone duration except at phrase endings.

Form

• Strophic (trnə̀əm poem).

Phrasing

• Prosodic and musical phrases are aligned.
• Verbal phrasing dominates (a 7-syllable line is proportionally longer than a 5-syllable line, etc.).
• Lines are divided into prosodic groups marked by a higher first lexical tone.
• The last word of a phrase is prolonged and lower, regardless of tone.
• Creakiness occurs at the end of a prosodic group and might signal a phrase boundary.

Word variations

• Same tone duration for long and short vowels, minor and major syllables.
• Coda prolongation is frequent.
• Syllabic reduplication is frequent. Reduplicants are generally lower than the original vowel.
• Schwa vowels (normally ə) have the same duration as all other vowels that are performed as short.

Lexical tones

• Lexical tones are often ignored. When realized, lexical tones are relative.
• Because of declination, a High tone can start lower than a preceding Low tone in the same prosodic group.

Analysis 6 The vocal genre húuwə̀

Húuwə̀ also belongs to the young people – especially young girls – and is used in the fields or the forest. It was used for trnə̀əm with contents that fitted these contexts. The melodic formula is similar to that of hrwə̀, but the phrases start low and have a recurring refrain (Example 36, for full text and translation see Example 26).

Example 36 Húuwə̀ performance of the trnə̀əm Example 26. Phrase 3: Nàaŋ mὲɛn trə̀ mà mɔ̀ɔn (5 syllables not counting reduplicated vowels) and phrase 4: Ɔ̀ɔn-crɔ̀ɔn khɔ́ɔŋ kɔ́ɔn mɨ́an ɔ̀ɔn yòl (7 syllables), thus demonstrating the difference between lines with 5 and 7 syllables. In contrast to hrwə̀ (Example 33) the first syllable(s) of a line is/are always performed low. Performed by Kàm Ràw, 1978. • 06 Nàaŋ mὲɛn, húuwə̀

The refrain is performed to the syllables húu wə̀ in the beginning and after each stanza (Examples 37–38). In the neighbouring Kwὲɛn area, the corresponding genre was called əə lə̀ə after its refrain in that area.

The refrain is very similar to the refrain of a lullaby, the words tuul luul being used for lulling a child.

Example 37 Húuwə̀ refrain. Performed by Kàm Ràw, 1978. Original pitch: c ≈ 145 Hz.

Example 38 Húuwə̀ refrain.

Húuwə̀ involves many reduplications and tonal movements in all syllables, both main and reduplicants. There is no use of separate pitch regions to realize tones. Instead, the tonal movement is rising-falling in base syllables, where the highest F0 point in High tone syllables is higher (starting from 200 Hz for our performer) than for words with Low tone. Reduplicants ignore lexical tones. The tonal contour is declining, building on the lowering of succeeding Low tones with High tones performed higher than this ‘melodic’ base line. The first High tone that occurs tends to be the highest within each line. These characteristics are easily observed in the graphs based on measuring height in the mid-point of the syllable (Example 39).

Example 39 Húuwə̀: the first two lines of the trnə̀əm Example 26, showing declining movement and realization of lexical tones.

Example 39 displays F0 measurement points for each syllable, both base and reduplicant, or within a sonorant coda when prolonged. For example, six points were measured in nàaŋ: start, middle, and last, both in the base syllable nàaŋ and in the reduplicant -ha. Frequencies (in Hz) are shown on the vertical axis. This performer realizes High tones from or above 200 Hz. Example 40 displays measurements of the maximum pitch for each base syllable.

Example 40 Húuwə̀: the sixth line of the trnə̀əm Example 26, showing realization in base syllables of 5 consecutive High lexical tones: St́ˀyúut khɔ́ɔŋ kɔ́ɔn mían performed: Sə̀-tə ˀyúu-tu khɔ́ɔŋ kɔ́ɔn mɨ́a-na ràa-ŋa rɔ̀ɔn (in this line the word khɔ́ɔŋ has replaced the synonymous word nàaŋ).

There are many reduplications and syllable prolongations; for example, the first phrase nàaŋ mὲɛn trə́ mà mɔ̀ɔn is realized as nàa-ŋa mὲɛ-nɛ trə́-ʔə màaa mɔ̀ɔnnn. Tonal movements occur in every syllable, both base and reduplicant. The tonal properties of base syllables are illustrated in Example 41.

Example 41 Illustration of tonal properties within each syllable of the two first phrases in a húuwə̀ performance of the trnə̀əm Example 26. Three F0 points in every base syllable are shown. Reduplicants are not shown.

The lexical tones are realized, but by a different technique from the one in hrlɨ̀ɨ. In hrlɨ̀ɨ tones are realized within separate pitch regions (Low tones in the 100–130 Hz region, and High tones above 130 Hz); but the realization of tones in húuwə̀ is relative: a High tone is higher than a preceding Low tone, and a Low tone is, with few exceptions, lower than a preceding High tone.

The húuwə̀ performance template

Melody

• Formulaic descending melody moving with a tonal centre.
• Successive lowering or declination of pitches within each prosodic group.
• The tonal movement in base syllables is rising-falling.
• Melodic range: 5 semitones.

Rhythm

• One tone duration, except at phrase endings.
• Regular pulse.

Form

• Strophic (trnə̀əm poem) with refrain (see Example 37).

Phrasing

• Prosodic and musical phrases are aligned.
• Verbal phrasing dominates (a 7-syllable line is two tone durations longer than a 5-syllable line, etc.).
• The last word in a phrase is prolonged and lower, regardless of lexical tone.
• Creakiness occurs at the end of a prosodic group and might signal a phrase boundary.

Initial/final formulae

• The first syllable(s) of a line is/are performed low.

Word variations

• Same tone duration for long and short vowels, minor and major syllable.
• Coda prolongation is frequent.
• Syllabic reduplication is frequent.
• Schwa vowels (normally ə) have the same duration as all other vowels that are performed as short.

Lexical tones

• Mainly tone-centred.
• The realization of tones is relative: a High tone is higher than a preceding Low tone, and a Low tone is, with few exceptions, lower than a preceding High tone.
• Reduplicants ignore lexical tones.

Analysis 7 The vocal genre yàam

Yàam, ‘weeping’, is used particularly by women: in the village it was used for soothing babies and called Táa yàam, ‘Don’t weep’; for dirges belonging to funeral wakes (Yàam róoy); and for guiding the soul of the deceased to the land of the dead (Lɔ̀ɔŋ ŋɔ̀ɔr) (Analysis 1). If somebody had died in a rice field, the clothes of the deceased would be laid in that place, and women would regularly go to yàam there. After the harvesting, a ceremony would be held in the field, and the women would perform dirges.

Even though yàam is particularly connected with female activities and funeral practices, this was not exclusively so, as is demonstrated by the existence of a Tə́əm yàam, ‘Weeping tə́əm’. Kàm Ràw performs this in the yàam vocal genre. In this case, the words are not arranged in stanzas, as is the case with trnə̀əm, but constitute about 20 lines which are tied together by chain-rhymes. It is more similar to certain ceremonial vocal expressions. Actually, the latter half is identical to the ‘Tiger dance’ (Yùun rwàay) that is associated with a ceremony for driving out the ‘Tiger spirit’ (róoy rwàay), a kind of vampire spirit. The Tə́əm yàam was performed with an unfaithful lover in mind. According to Kàm Ràw, it is very serious and was actually used with the intention of sending away the soul of the person in question to the land of the dead.

Outside the village yàam was also used for trnə̀əm, for instance when women were by themselves while washing in a river, fishing, or catching frogs. Young girls and boys would use it occasionally. In these cases, yàam was not used for praise or deprecation but rather for joking or for youth trnə̀əm. Boys and girls could also use yàam for a different kind of rhyme that might often be spoken.

Yàam is based on a short musical motif that may be prolonged by adding new word-pairs and repeating the penultimate ‘bar’ (Examples 42 and 43). The melody is arch-shaped: it starts low, rises and falls again. It ends a second (2 semitones) below (B♭) at what might be perceived as the tonic (C), and feels like a ‘blue ending’.

Example 42 Yàam melodic formula for a line of 5 syllables. ‘V’ stands for a reduplicated vowel, ‘v’ for a less frequently reduplicated vowel. • 07 Nàaŋ mὲɛn, yàam

Example 43 Yàam. First phrase of the trnə̀əm Example 26. Performed by Kàm Ràw, 1978. • 07 Nàaŋ mὲɛn, yàam

Lexical tones are realized at the vowel onset of the base syllables (Example 44). Tonal movement occurs in almost every base syllable, and this involves pitch going up or down, or up–down, down–up. High lexical tones start at a high pitch level or occasionally at a medium pitch level, in which case the high pitch occurs within the vowel. In initial formulae, all words are performed at low pitch level, regardless of lexical tone.

Example 44 Yàam. The second phrase of the trnə̀əm in Example 26, demonstrating tonal movement in base syllables. The tonal movement in some prolonged codas (ŋ, n) is shown separately.

The yàam performance template

Melody

• Formulaic arch-shaped melody with a tonal centre ending 2 semitones below the tonic (see Example 42).
• Melodic range: 9 semitones.

Rhythm

• Regular pulse.
• Iambic movement dominates.

Form

• Strophic (trnə̀əm poem) [litany also occurs].

Phrasing

• Prosodic and musical phrases are aligned.
• Verbal phrasing dominates (a 7-syllable line is two tone durations longer than a 5-syllable line, etc.).
• The last word in a phrase is prolonged and lower, regardless of lexical tone.

Word variations

• Same tone duration for long and short vowels, minor and major syllable.
• Coda prolongation is frequent.
• Syllabic reduplication is frequent.
• Schwa vowels (normally ə) have the same duration as all other vowels that are performed as short.

Lexical tones

• Mainly tone-centred.
• Low lexical tones start at a low or medium pitch level and continue to high.
• High lexical tones start at a high pitch level, or occasionally at a medium pitch level, in which case the high pitch occurs within the vowel.
• Lexical tones are realized at the onset of base syllables.
• Tonal movement in almost every base syllable (pitch goes up or down, or up–down, down–up).
• Reduplicants ignore lexical tones.

Analysis 8 The vocal genre yùun tìiŋ

8a: Yùun tìiŋ, ‘Water-tube dance’

In a Kammu village, the period that preceded the sowing season was the most common season for repairing houses and building new ones. When people worked on the houses, many water-containers, called tìiŋ, were constructed. Made from bamboo, they were about one or one and a half metres long. At feasts during the house-building, young boys and girls used to perform the words and movements of Yùun tìiŋ (yùun means ‘dance’), so it can be translated ‘Water-tube dance’. It was accompanied by the water-containers, which served as stamping tubes when struck against the ground. The dancers stood behind one another in a circle, each person holding a water-tube; and they walked around striking the tubes, making a curtsy rhythmically. The version in Example 45 stems from the Rmcùal village, but there are also a few recordings from neighbouring villages with variants of the words and melody.

Lexical tones might be expected not to be systematically performed in this melody type. Therefore, an approximate method of representation has been chosen that would make it easy to spot the relationship between lexical tones and pitches. This involves transcription – including reduplication of syllables – and parallel translation of the words. Initial and final formulae with regard to which it may be expected that lexical tones are not always reflected in performance are marked. The length of tones is not shown in the transcription but may be understood from Example 46. Pitches are given in the high (H), low (L), and neutral (N) categories. Neutral is in this case located between the high and the low pitches. A higher octave is marked with a + sign. When High and Low lexical tones coincide with high and low pitch, respectively, they are indicated in bold italics: H, L.

Even though many lexical tones coincide with musical pitch, one may interpret such occasions in the initial and final formulae as accidental. For the remaining portion of the lines, there are as many cases when they do not coincide. Often a High or Low slides to the neutral pitch. One may conclude that this melody has a more fixed melodic and rhythmic structure and is less flexible where the realization of lexical tones is concerned.

The schwa (ə) is short but clearly audible. Like all vowels, it can be reduplicated and performed to the long second tone of an iambic unit. Two examples of this occur in the first line of Example 45 (reduplicated schwa underlined; other vowels that fall on the long tone are also doubled): Kə̀n-əə-tíik-ii tìiŋ-hee rə̀-əə-háaŋ-aa […]

Example 45 Yùun tìiŋ, Water-tube dance, with realization of lexical tones. Performed by Kàm Ràw, ca. 1977. • 08 Yùun tìiŋ

H = high pitch.

N = neutral pitch.

L = low pitch.

L+, H+ = low, high in a higher pitch level (initial formula).

N–L = gliding from N to L.

L–N = gliding from L to N, etc.

H = High lexical tone performed high.

L = Low lexical tone performed low.

[L+	H+	L+–N]	N	L	N	H–N		[N–L–N	H–N]
kǹ-ə-ə	tíik-i	tìiŋ-he-e	r̀-ə-	háaŋ-a	tə́ən-ə-	trə́-ə		tə́ən-ə-	trə́-
strike	tube		bamboo		bounce			bounce
down
[L+	H+	L+–N]	N	N–L	L–N	H		[L–N	H–N]
kǹ-ə-	tíik-i	tìiŋ-hə-ə	r̀ə-ə-	yə̀-ə	hḿ-ə-	pɨ́al-ɨ		hḿ-m-	pɨ́al-ɨ
strike	down	tube	tree		rebound			rebound
[L+	H+	L+–N]	L	N	H–N	L	N	[L–N	H–N]
yùun-u,	tìiŋ-i,	yùun-hu-u	tìiŋ-i	pə́ə	yùun-u	tìiŋ-i	tə́-ə	tìiŋ-i	tə́-ə
dance,	tube,	dance	tube	not	dance	tube	broken	tube	broken
[L+	H+	L+–N]	L	N	H–N	L	N	[N–H	H–N]
yùun-u,	tìiŋ-i,	yùun-hu-u	tìiŋ-i	pə́-ə	yùun-u	tìiŋ-i	páak-a	tìiŋ-i	páak-a
dance,	tube,	dance	tube	not	dance	tube	split	tube	split
[L+	H+	L+–N]	N–H	H–N	N–L	N		[N–L–N	H–N]
kǹ-ə-	tíik-i	tìiŋ-he-e	r̀-ə-	háaŋ-a	tə́ən-ə-	trə́-ə		mm	nàaŋ
strike	down		tube	bamboo		bounce			dear
[L	H–N	H–L]	N–L	L	N	H–N		[N–L–N	H–N]
kǹ-ə-	tíik-i	tìiŋ-hə-ə	(t)ə-ə-	yə̀-ə	hḿ-ə-	pɨ́al-ɨ		hḿ-m-	pɨ́al-ɨ
strike	down	tube	tree		rebound			rebound
[L+	H+	L+–N]	L	N	H–N	L	L	[N–H	N]
yùun-u,	tìiŋ-i,	yùun-hu-u	tìiŋ-i	pə́ə	yùun-u	tìiŋ-i	tə́-ə	tìiŋ-i	tə́-ə
dance,	tube,	dance	tube	not	dance	tube	broken	tube	broken
[L+	H+	L+–N]	L	N	H–N	L	N	[N–H	H–N]
yùun-u,	tìiŋ-i,	yùun-hu-u	tìiŋ-i	pə́-ə	yùun-u	tìiŋ	páak-a	tìiŋ-i	páak-a
dance	tube	dance	tube	not	dance	tube	split	tube	split

Interpretation:

Strike down bamboo tubes, bounce

Strike down wooden tubes, rebound

The one that doesn’t dance is broken, is broken

The one that doesn’t dance is split, is split [etc.]

8b: The vocal genre yùun tìiŋ

The melody used for yùun tìiŋ also commonly served as a basic melody for a style of performing non-ritual trnə̀əm (Example 46). This particular performance style differs from the majority of basic melodies used for trnə̀əm in most Kammu dialect areas in northern Laos. According to Kàm Ràw, it was originally used in the so-called Cwàa area to the north-east of the Yùan area on the east side of the Nam Tha river and then spread into the Yùan area, probably in the 1960s. The melody may be historically related to the cèem ə̀əy style, commonly used among the Kammu who live around Luang Prabang.³⁰

Example 46 The basic melody of the vocal genre yùun tìiŋ. The example shows two kinds of musical phrases: a and b, and three phrase lengths: 11 (or 10), 9, and 7 syllables, respectively, numbered backwards in order to show how syllables are added in the middle of lines while the beginning and ending are unaffected.

The longer the phrase to be performed, the more tones are put into the middle of the template, while its beginning and end function as formulae and are stable in either of two variants (a and b in Example 46). In the vast majority of cases, High lexical tones are not realized by higher pitch (Example 47). Schwa vowels (ə) are clearly audible and normally short, but they are long when they fall on the second syllable of an iambic unit (underlined). This is the case in the first line of Example 47: pə̀ kə́ə-tóŋ-hoo […]

Example 47 Yùun tìiŋ genre with lexical tones. A suite of three trnə̀əm: Yèm àay yèt tàa lès pŋkà (When close to you I’m shy) > Mà àay àn àay tèn wáay wèc (Mother asked me to go and to hurry home) > Prɔ́ àay yèt rùam kúŋ (I wish we could live together in the village). Performed by Kàm Ràw, ca. 1977.

H = high pitch level.

h = medium high pitch level.

l = medium low pitch level.

L = low pitch level

L+, H+ = low, high in the higher octave (initial formula).

H–L etc. = gliding from H to L, etc.

H, L, h, l = lexical tone and tone-level coincide.

[l+	h+	l+–l]	l	L	L–l	h–L			L–L–l	h
pə̀	kə́-	tóŋ-ho	lìit-e	théeŋ-e	làm-a	ŋà			làm-a	ŋà
eat	egg		beetle		tasty	itching			tasty	itching
[l+	h+	l+–l]	l	L	L	l			L–L–l	h
pə̀	kə́-	tóŋ-ho	sén-	wà	làm-a	tées-e			làm-a	èes
eat	egg		beetle		tasty	itching			tasty	itching
[l+	h+	l+–l]	L	L	L–L–l	l	h-L	L-h	H-L	h–l
yèm-e	àay-a	yèt-e	tàa	lès-e	kàay-a-a	sáh	pŋ̀-ə-	kà	pŋ̀-ə-	kà
when	I	stay	at	close	then	say	embarrassed		embarrassed
[l+	h+	l+–l]	l	L	L	l	H–l	h–L	L–L–l	h–l
yèm-e	àay-a	yèt-e	tàa	cà-a	kàay-a	sáh	sŕ-e	èeŋ-e	sŕ-	èeŋ-e
when	I	stay	at	far	then	say	yearn		yearn
[l+	h+	l+–l]	l	h–L					L–L–l	h–l
mà	àay-e	àn-ha	àay-a	téc-e					pháay-a	lɔ̀ɔ
mother	my	let	me	sell					cotton	bobbin
[L	h	H–h]	h–l	l–h					H–l	h–l
mà	àay-e	àn-a	àay-a	téc-e					pháay-a	làn
mother	my	let	me	sell					cotton	reel
[l+	h+	l+–]	l	L					l–h	H–I
mà	àay-e	àn-ha	àay-a	tèn-e					wáay-a	wèc-e
mother	my	let	me	stay					hurry	return
[L	h	H–h]	h–l	L					h	L
mà	àay-e	àn-a	àay-a	tèn-e					wáay-a	kàay-a
mother	my	let	me	stay					hurry	return
[l+	h+	l+–l]	l	L	l–h	H–l			L–L–I	h–l
ò	lə̀ə	pàt-ha	pɔ̀ɔc-ɔ	tàa	sń-ə-	trùh-u			lɔ̀ɔc-ɔ	kɔ̀ɔŋ-ɔ
I	then	cut	bamboo	at	downstream				forget	cook
[L	h	H–h]	h–L	L	l–h	H–l			h–L–l	h–l
ò	lə̀ə	pàt-ha	pɔ̀ɔc-ɔ	tàa	sń-ə-	trùh-u			lɔ̀ɔc-ɔ	cáa
I	then	cut	bamboo	at	downstream				forget	all
[l+	h+	l+–l]	l	L	h	H–l			l–L–l	h–l
àay	lə̀ə	tḿ-hə-	pùh-u	kɔ́ɔn-ɔ	pr̀-ə-	yɔ̀ɔŋ-ɔ			lɔ̀ɔc-e-e	wèc
I	then	run	across	child	dragon				forget	return
[L	h	l+	h+	l+–l]	l	h–l			L–l	h
àay-a	lə̀ə	tḿ-hə-	pùh-u	kɔ́ɔn-ɔ	pə̀-	yàa			lɔ̀ɔc-e	kàay-a
I	then	run	across	child	title				forget	return
[l+	h+	l+–l]	l	h–L	L–l	h–L			L–l	h
prɔ́	àay-a	plúŋ-hu	rùam-m	cét-e	dὲɛ	nàaŋ-a			dὲɛ	nàaŋ
wish	we	sprout	together	sour	oh	dear			oh	dear
[l+	h+	l+–l]	l	L	L–l	h–L			L–l	h
prɔ́	àay-a	plúŋ-hu	rùam-m	èn-e	dὲɛ	khɔ́ɔŋ-ɔ			dὲɛ	khɔ́ɔŋ
wish	we	sprout	together	0	oh	dear			oh	dear
[l+	h+	l+–l]	l	h–L					L–l	h–L
prɔ́	àay-a	yèt-he	rùam	kùŋ-u					dὲɛ	nàaŋ
wish	we	stay	together	village					oh	dear
[L	h	H–l]	h–L	L					L–l	h–L
prɔ́	àay-a	yèt-he	rùam	kàaŋ-a					dὲɛ	cùu
wish	we	sit	together	house					oh	dear

Interpretation:

Beetle eggs taste sour, sour, oh, dear

Spider eggs taste bitter, bitter

When close to you I’m shy, I’m shy

When far from you I yearn, I yearn

Mother asked me to go and sell bobbins

Mother asked me to go and sell reels

Mother asked me to go and hurry back

Mother asked me to go and hurry home

I cut bamboo downstream and forgot to cook it

I cut bamboo downstream and forgot everything

I met a mighty Dragon and forgot to go back

I met a wealthy Lord and forgot to go home

I wish we could sprout together bitterly, oh dear, oh dear

I wish we could sprout together happily, oh dear, oh dear

I wish we could live together in the village, oh dear, oh dear

I wish we could sit together in the house, oh dear, oh dear

The yùun tìiŋ performance template

Melody

• A fairly long melody with a descending contour, tonal centre, and triadic melodic movement (see Example 46).
• Melodic range: 15 semitones.

Rhythm

• Regular pulse.
• Iambic delivery of the words.

Form

• Strophic (trnə̀əm poem) [litany also occurs].

Phrasing

• Prosodic and musical phrases are aligned.
• Verbal phrasing dominates (a 7-syllable line is two tone durations longer than a 5-syllable line, etc.).
• The poetry may be varied by the addition of polite words or phrases.

Initial/final formulae

• Initial and final formulae.

Word variations

• Same tone duration for long and short vowel, minor and major syllable.
• Song-words with lexical meaning: dὲɛ nàaŋ, dὲɛ khɔ́ɔŋ (approximately ‘oh, dear’), each pair of words corresponding to an iambic unit.
• Reduplication is frequent: in principle one word with a long vowel or with reduplication per iambic unit.
• All vowels, including schwa (ə), have approximately the same duration. In addition, schwa vowels are long when falling on the second long tone of an iambic unit.

Lexical tones

• Predominantly melody-centred.
• Most of the phrases circle around high, ‘neutral’, and low pitch levels. Lexical tones are often realized in these phrases, but contrary motion occurs.
• A ‘neutral’ pitch (n) is often used both for High and Low lexical tones, but Low, and more often High, lexical tones can be realized by sliding from the initial pitch: Low may be performed l–n or n–h, High may be performed h–n.
• A lexical tone is often realized at the beginning of a syllable.
• Lexical tones are not realized in initial formulae that usually comprise the first three words of a line.

Analysis 9 The vocal genre tə́əm

Tə́əm is an elaborate form of performing the orally transmitted trnə̀əm poems. The performer may also elaborate the words by prolonging or contracting them, or by adding sets of words that may be traditional or made up on the spot. A tə́əm melody basically consists of a melodic formula that varies between the Kammu dialect areas, but it also varies in details between villages and even between individuals. There are usually rather distinct initial and final formulae, performed for various words that may be translated ‘hey’ or ‘oh’. A common final formula in the Yùan area is kàay sáh, ‘this I say’, and the word sáh, ‘I say’, which is often squeezed in at the start of lines within the trnə̀əm. Apart from these portions, the melody moves forward in rather narrow intervals that are more or less compulsively determined by the lexical tones.

In the research on related vocal expressions in South-East Asia, words or syllables that seem to be added in performance are normally treated as ‘fillers’ with the function of making the words of a poem fit a melody. Without denying the fact that fillers do exist, this view is based on an assumption that an existing poem with a stable form is fitted to a stable melody when performed. The process of performance is, however, a more complex creative activity which involves linguistic, poetical, and musical aspects that do not have stable forms, but are fluid. When looking at the graph in Example 48 as a formula used for a performance template, it is clear that some such words may actually be integrated and even necessary parts of the performance template, with the function of making people recognize that the performance is special for a particular area, village, or individual. The words həəy, eee and kàay sáh thus define tə́əm Yùan. Other tonal dialects in the area, like Kwὲɛn, have similar patterns using their specific words.³¹ The primary conclusion is that these are not random words put in to fill blank spaces in the melody, but necessary constituents for performance.

In major aspects of the performances, the words are grouped two by two and performed in an iambic pattern (short + long). In the case of parts where lexical tones and musical pitch coincide, we use the term tone-centred, whereas the other parts are referred to as melody-centred.³² After subtracting initial and final formulae (i.e. the parts where melody centration dominates), the remainder of the performance is more dominated by lexical tone centration.

Example 48 shows the main outline of a performance with the initial and final formulae. In its basic form, a trnə̀əm consists of two stanzas, the first one starting with a high-pitched Həəy and the second with Eee in the lower region. Each stanza normally consists of four musical phrases marked α, β, γ, δ. The α and δ musical phrases are initial formulae and the γ-phrase includes final formulae. These sections are mainly melody-centred. The tone-centred β-phrases are the major part of the performance. In reality there will be much variation that includes the length and order of the musical phrases.

Example 49 is an approximate graph of one stanza of a tə́əm Yùan performance (for full text and translation, see Example 26). The shaded parts show the areas in which words can be predicted to be performed if the lexical tones were different from this particular performance. This type of graphic representation was used in the initial phase of our research. It was devised in order to communicate musical characteristics in a manner that did not presume knowledge of musical notation and therefore served as a good starting point for discussions, while being easy to carry out with the use of an ordinary computer, basically using Microsoft Word for tables.

It should be noted that the first and third lines of the trnə̀əm are 5-syllable lines that are made into 7-syllable lines through prolongation: the added ‘padding words’ dὲɛ nàaŋ (dashed underlining), meaning approximately ‘oh dear’. Conversely, the last 9-syllable line, with the added ‘song-words’ kàay sáh, is turned into a 7-syllable line by means of contraction: four syllables are squeezed into a 2-syllable iambic unit (mɨ́an ɔ̀ɔn yòl kàay). These are common devices in tə́əm performance.

Example 49 One stanza of a tə̀əm Yùan performance of the trnə̀əm Example 26 in a simplified graphic transcription. H = high lexical tone, L = low. Vertical: pitch, horizontal: beats (1 square ≈ 1 eighth note). Performed by Kàm Ràw, ca. 1979/80. Original pitch: c ≈ 130 Hz. • 09 Nàaŋ mὲɛn, tə́əm

When performed in the vocal tə́əm genre, the words are fitted into the model in Example 48. Reduplication of vowels and coda prolongation are frequent. The musical transcription in Example 50 is based on the same performance as Example 49, but it also includes the second stanza. The order of the musical phrases follows the basic pattern: α–β–β–γ for the first stanza and δ–β–γ for the second. The initial long-drawn-out həəəy and the final kàay sáh are also in agreement with the model in Example 48. Most lines are preceded by the word sáh, but it serves as a very short upbeat to the next syllable and steals time from the preceding syllable. There are breathing pauses in two of the β-phrases, and the first three 5-syllable lines are made into 7-syllable ones by the addition of 2 syllables (underlined). It should be observed that the two final lines of 5 and 7 syllables plus the 2-syllable kàay sáh, 14 syllables altogether, are compressed into a δ-phrase of 6 iambic units, which would otherwise contain 7 syllables or less. It is a common stylistic device to shorten the stanzas in this manner, particularly the final stanza of a performance.

Example 50 Tə̀əm Yùan performance of the trnə̀əm Example 26. Dotted barlines mark divisions betweeen iambic units. Musical phrases are marked α, β, γ, δ (cf. Example 48). Added words are underlined. The word sáh is notated with an ‘x’. Performed by Kàm Ràw, ca. 1979/80. Original pitch: c ≈ 130 Hz. • 09 Nàaŋ mὲɛn, tə́əm

1 Pàh mòŋ cŋkwá ŋɔ̀ɔr, ‘Bright moon, widen my path’

A trnə̀əm is an orally transmitted poem that may be performed as tə́əm or in many other genres. It can often easily be divided into two stanzas, the first usually being an associative parallel to the second one. This means that the meaning can be expected to be clearer in the second stanza. In Example 51, it is a person walking on a forest path in the dark, wishing for the moonlight to light it up. But a trnə̀əm like this one can be performed on other occasions as well.

Within each stanza, there is much repetition or parallelism, which leads to considerable play on consonants and vowels (Example 51). In this case, the first stanza is dominated by c and a ~ ɔ and the second stanza by p and o ~ ɔɔ. There are also end rhymes that tie the two stanzas together, since the first word of a rhyme-pair comes in the first and the second rhyme-word in the second stanza (the syllables kɔ̀ɔr/ŋɔ̀ɔr, kɔ̀/kwá and kὲɛr/klɛ́ɛr).³³

The following examples illustrate the realization of the lexical tones in certain melodic positions. Example 52 demonstrates that High lexical tones are realized higher than Low lexical tones, with only one exception: the fourth syllable in line δ, which is the first line after a breathing pause, usually the third line of a stanza (see Example 48), which is always performed low even if the lexical tone is High.³⁴ This is a case of exceptional melodic dominance that sometimes occurs close to a phrase ending. Note that the graph (Example 52) has four high pitches: H, H2, H3, and H4 and three low pitches: L, L1, and L2. This is designed to cover the melodic contour which descends more than one octave.

Example 52 Four lines from the first tə̀əm stanza of Pàh mòŋ cŋkwá ŋɔ̀ɔr in Example 51, schematically showing how lexical tones are realized (the lines in this performance are line 1 of stanza 1, line 1 of stanza 2, line 4 of stanza 1, and line 4 of stanza 2, with variations). H = high lexical tone, L = low. Performed by Kàm Ràw, 1981. Vertical: pitch levels, horizontal: beats (1 square ≈ 1 eighth note). Original pitch: L ≈ 110 Hz. • 10 Pàh mòŋ, B1a

This particular trnə́əm was chosen as it has very few words with High lexical tones, so it is fairly easy to spot how they are realized. There is actually one line that consists only of words with Low lexical tone. In a β-phrase, these are evidently performed either at low or medium pitch (Example 53). The β-phrases are tone-centred parts of the performance and therefore dominated by the lexical tones, while the melody plays a smaller role.

Example 53 The realization of a β-phrase of Pàh mòŋ cŋkwá ŋɔ̀ɔr, Example 51, made up of words of only Low lexical tones. For the β-phrase, see Example 48. Vertical: pitch levels, horizontal: beats (1 square ≈ 1 eighth note).

The different iambic units that occur in β-phrases in the same trnə̀əm are listed in Example 54. It shows that High lexical tone can be performed high, or high sliding to low. In certain positions both High and Low can be realized at medium pitch.

Example 54 The different iambic units that occur in β-phrases in the same trnə̀əm as the previous examples: Pàh mòŋ cŋkwá ŋɔ̀ɔr, Example 51. H = High lexical tone, L = Low. Vertical: pitch levels, horizontal: beats (1 square ≈ 1 eighth note).

Musical transcription – whether in musical notation or graphs – tends to approximate pitches musically, i.e. the main part of a pitch that corresponds to a linguistic unit is transcribed. Initial or final movement within a sound is normally considered to be just that: a tone is reached or left by a sliding motion. The result will be a musically relevant transcription that communicates the musical sound as experienced by the person who is transcribing. An analysis based on phonetic expertise will approach this differently, and this is one of the advantages of cross-disciplinary work.

In Example 55, a section of the α-phrase in Example 52 has been measured with regard to the realization of lexical tones. This measurement adds the information that, in tə́əm, the lexical tones are realized at the vowel onset, while all other pitch movements are in the coda or in added syllables.

Example 55 An α-phrase of Pàh mòŋ cŋkwá ŋɔ̀ɔr, Example 51. Arrows demonstrate lexical tones that are realized at the vowel onset in tə́əm.

2 Táa píc àay yʌ̀ʌ prìaŋ, ‘Don’t abandon me to stay with others’

This is a widely spread trnə̀əm that belongs to the social situation when the performer asks people to let him/her be included, or stay included, in their friendship group (Example 56). It basically says: don’t force me to stay among strangers. It shares the general poetic characteristics of trnə̀əm.

The performance includes frequent reduplications and coda prolongations. The lexical tones are often realized, and it is possible to distinguish three pitch areas: ≤130 Hz, 130–180 Hz and ≥180 Hz. High lexical tones lie in the highest pitch area and Low tones in the middle pitch area. Then there are portions with even lower pitch. The initial formula that starts at a high pitch covers the first 3–4 words (Examples 57–58). In this case, all the words have High lexical tone; but in the case of Low lexical tones, or a mixture, the pitch will also zigzag down to the low pitch area, and the realization of the lexical tones will be relational. It then stabilizes in the middle area where the Low lexical tones are realized.

Example 57 Realization of lexical tones in the first phrase, the α-phrase, of Táa píc àay yʌ̀ʌ prìaŋ, Example 56.

Example 58 Realization of lexical tones in the same first phrase, the α-phrase, of Táa píc àay yʌ̀ʌ prìaŋ as in Example 57. The tonal movement within each word is shown. Vertical lines show word boundaries. The initial həəəy (not shown in Example 57), with a short sáh squeezed in before the first word of the trnə̀əm (táa).

Examples 57–58 show the realization of lexical tones in the first phrase of a Yùan tə́əm performance. This is a melody-centred section, but lexical tones can still be seen to be realized in the performance. In this case, High lexical tones are performed higher than 180 Hz and Low tones mainly in the middle area. As shown in Example 59 this is even more evident in the second phrase, which is a tone-centred section. As these examples show, High syllables have falling tone, whereas Low syllables have rising tone in most cases.³⁵

Example 59 Realization of lexical tones in the second phrase, β-phrase, of Táa píc àay yʌ̀ʌ prìaŋ, Example 56.

Example 60 shows the final phrase. Like all tə́əm performances of the Yùan style, it finishes with a falling movement in the lowest pitch area with the words kàay sáh, meaning ‘this I say’. The whole phrase is performed low. Even the initial plɔ́ɔy with a High lexical tone is performed in the middle pitch area but still high, relatively speaking. Here the performance moves downwards, and 5 (or 6) syllables are performed in the lowest pitch area. They all happen to have Low lexical tones, but it can also be seen that there is no movement within the tones. This technique of flattening out the melody – and the tones – appears particularly when the tempo increases and the performer is exhausting the last remaining air in his lungs.

Example 60 Realization of lexical tones in the final phrase, γ-phrase, of Táa píc àay yʌ̀ʌ prìaŋ, Example 56. • 11 Táa píc, Last phrase

The tə́əm performance template

Melody

• A fairly long descending melody with a tonal centre (see Example 48).
• Melodic range: 16 semitones.

Rhythm

• Regular pulse.
• An iambic pattern dominates: normally one word-pair or one word with reduplicant per iambic unit.

Form

• Strophic (trnə̀əm poem).

Phrasing

• Prosodic and musical phrases are aligned.
• Verbal phrasing (syllables added through prolongation or contraction) and musical phrasing of approximately equal weight.
• The poetry can be varied by the addition of lines or stanzas.

Initial/final formulae

• An initial formula consists of an initial, long high-pitched tone (Həəəy) and falls with the very first syllables.
• An initial formula of phrases inside a stanza (generally the third line) consists of an initial, long low-pitched tone (Eee) and rises with the very first syllables.
• There is a final formula with the words kàay sáh (approximately ‘this I say’), ending approximately 2 semitones below the tonic.

Word variations

• Song-words with or without lexical meaning.
• An auxiliary word sáh (‘I say’) often occurs at the start of initial phrases, performed at approximately the same pitch as the first syllable of the phrase.
• Same tone duration for long and short vowels, minor and major syllables.
• Coda prolongation is frequent.
• Syllabic reduplication dominates.
• All vowels, including schwa (ə), are of approximately the same duration. In addition, schwa vowels are long when falling on the second long tone of an iambic unit.

Lexical tones

• Contains both melody-centred and tone-centred features.
• Lexical tones are realized at three pitch levels.
• A medium pitch (M) is often used both for High and Low lexical tones.
• A lexical tone is often realized at the beginning of a syllable.
• Low and more often High lexical tones can be realized so that they slide from the initial pitch: Low may be realized as L–M or M–H, High may be realized as H–M.
• Lexical tones are sometimes ignored in the middle of the second part of a stanza when all syllables are performed L. There is no contrary movement in these passages.

Analysis 10 Krùu, ‘spells’

The Kammu shaman is called mɔ́ɔ. This term can also be used for a physician. More specifically, the Kammu shaman is called mɔ́ɔ róoy, ‘spirit doctor’. The shaman is a specialist in communicating with the spirit world. At the beginning of a seance, he/she calls for the assistance of the shaman spirits or kɔ́ɔn róoy, ‘spirit children’, the spirits of previous shamans. Apart from the sick person, a traditional shaman ceremony included the shaman, one or more helpers, and an ‘audience’. The aim of the seance was for the shaman to diagnose and cure illness by functioning as a medium for certain spirits, by visiting the land of the dead and/or driving out bad spirits.

The contextual information has been obtained from interviews with Kàm Ràw, and the krùu were not recorded in an actual seance. He had substantial knowledge of Kammu shamanistic rituals, knowledge that appears to be rapidly disappearing among the Kammu people in his home area. In northern Kammu tradition, the musical instruments used were the knobbed gong, mòoŋ, in combination with the cymbals, crɛ́ɛŋ. Apart from these instruments the shaman, and others present at a seance, perform various vocal expressions. The shaman may also whistle, either in communication with certain spirits or when certain spirits are considered to have entered his body and are talking through him. In Kammu tradition, the róoy hʔéep (i.e. the spirits of those who died in accidents) are particularly believed to whistle in a special manner.

The krùu are known only to those who (like Kàm Ràw) were taught to be shamans. Almost all words in the krùu are loans from Old Lao or Proto-Southwestern Tai, the ancestor language of modern Lao, Thai, and Lü.³⁶ Although some of these words occur in ordinary Kammu, most of them cannot be understood by ordinary Kammu speakers. Some of the words are ultimately of Indic origin, including the word krùu itself, which is etymologically related to Sanskrit guru, ‘teacher’.

One focus in this section is on the relationship between prosodic and musical phrasing in the performances. For this reason, rather long performances need to be examined. In order to facilitate this, number notation has been used for the transcriptions (see Appendix 2). This is also useful for the other focus, which is on the realization of lexical tones in this genre.

Vocal expressions in a Kammu shaman seance

There are different forms of vocal expressions during a shaman seance.

Calling shaman spirits

• Kʔə́əy kɔ́ɔn mɔ́ɔ ‘Call the shaman spirit(s)’. The shaman calls the spirits of dead shamans through these vocal expressions that start in a somewhat similar manner to the vocal genre tə́əm (used in social situations), but continue more like a polite krùu (see below). (Figure 2)
• Tə́əm. Those present may perform ordinary trnə̀əm of the type used at parties in order to encourage the shaman spirits to come (Analysis 9).
• Mɔ́ɔ rɔ̀ɔt màañ ‘Shaman arrives and asks’. When the shaman spirits have arrived, i.e. entered the body of the living shaman, they ask why people called using the same vocal genre as Kʔə́əy kɔ́ɔn mɔ́ɔ.
• Kʔə́əy hrmàal ‘Calling a soul’. The shaman may call a sick person’s lost or wandering souls back. The same vocal genre as Mɔ́ɔ rɔ̀ɔt màañ.
• Mɔ́ɔ krɔ́ɔ ùun róoy cɔ̀ɔy tèe ‘Shaman asks spirits for help’. A shaman may address the shaman spirit(s) for support. The same vocal genre as Mɔ́ɔ rɔ̀ɔt màañ.

Kàm à-thí-tháan, prayers

• Ì càk cə̀ən ‘We called you’. The people may reply Mɔ́ɔ rɔ̀ɔt màañ ‘Shaman arrives and asks’. These words are spoken quickly, like a prayer, without much intonation or marked pronunciation of lexical tones (Analysis 2).
• Krùu kʔə́əy mà krùu ‘Spell for apologizing to the magic power’ or Wéey mὲɛ krùu ‘Worship one’s power’. The shaman speaks these words in the form of a prayer after having made a mistake in a spell, or before starting in order to avoid punishment, should a mistake occur (Analysis 2).
• Mɨ̀a tàaŋ ròoŋ cét khát mɨ̀ŋ nə́ə ‘Go back to your seven-roof building’. These are words for seeing the shaman spirit(s) off. It is not among the recordings. This may indicate that it would be spoken as a prayer.

Krùu, spells

There are a great number of krùu, ‘spells’, for certain situations. They are mainly performed in a rhythmic manner, with a degree of melodic movement generally relating to the lexical tones. There are two major types of krùu:

1. Polite krùu are performed rather slowly in order to ɔ̀ɔy, ‘lure out’ or ‘call out’, a spirit that the shaman thinks is the reason for a problem without angering it. They include:
   • Krùu ɔ̀ɔy róoy, Calling out spirits
   • Krùu ɔ̀ɔy róoy phíi pàa, Calling out the sky spirit
   • Krùu plɔ́ɔŋ rúu, Opening your mind [= remember]
   • Krùu ràaŋ húum, Making someone appreciate a person.
2. Impolite krùu are performed in a very rapid manner in order to ptú, ‘drive out’, a dangerous spirit by scaring it. These include:
   • Krùu ptú róoy, Driving out a spirit
   • Krùu ptú róoy rwàay ɔ̀ɔk, Driving out the tiger spirit
   • Krùu ptú róoy cntràas, Driving out the lightning spirit

Figure 2 Kàm Ràw demonstrating the simultaneous playing of gong and cymb, crɛ́ɛŋ, when singing the song Kʔə́əy kɔ́ɔn mɔ́ɔ, ‘Calling the shaman spirits’, of the shaman seance. The gong is hung on the right arm and struck with a drumstick in the right hand, while one cymbal is held between the big toe and the next toe of the left foot and struck with the other cymbal held in the left hand. The photograph was taken in Lund, Sweden, in 1999.

Krùu performances

The performer, Kàm Ràw, said that krùu relates to à-thí-tháan (‘pray’) as tə́əm (Analysis 9) relates to hrlɨ̀ɨ (Analysis 4), which means that they are two ways of addressing an entity: krùu and à-thí-tháan address spirits, whereas tə́əm and hrlɨ̀ɨ address people. It also means that à-thí-tháan and hrlɨ̀ɨ are close to speech, while krùu and tə́əm are musically – and poetically – more complex. The relationship between lexical tones and musical pitch in krùu is rather complex, and closely related to the poetical form and poetical variation. The realization of the lexical tones in krùu is relative rather than expressed by exact pitch. In this respect, krùu has parallels with performances in the tə́əm genre, but also with hrlɨ̀ɨ. The performer, Kàm Ràw, uses the same frequency area (F0) that he uses in speech, which differs from those he uses in performing tə́əm (Analysis 9), for example.

Many krùu start and end with a vocative phrase. Apart from this, prosodic phrases are often paired in a manner similar to the lɔ̀ɔŋ narratives (Analysis 1). The second phrase of a pair partly parallels the first phrase, either antithetically, as question and response, or by repetition, with one or two words exchanged. These pairs are indicated in the translations in Examples 62 and 63, in which the second phrase of a pair is indented. Longer krùu have two or more sections (A, B, C, etc.). These sections function as thematic episodes, which means that several prosodic phrases are tied together into one thematic episode, closer to speech.

• Some krùu are very short. Most are long, however, and the long krùu are divided into more than one section.
• The end of a section is usually marked by breathing and by the performer clearing his throat (transcribed ‘hrm’). This is called km̀rὲh and is intended to scare spirits away.
• Each section is performed in one breath.
• The tempo usually increases with each section.
• Some krùu start with lines of 3 syllables that are performed with long note values, i.e. more slowly than the rest.
• Basically, the krùu consists of lines of 5, 7, or 9 syllables, musically marked by a long final tone.
• Musical phrases generally end with ♭3⇒1 or with 1, which often, but not always, reflect lexical tone.
• The first two lines of a section are often tied together: For instance two lines of 5 syllables may be performed as 10 syllables, before finishing with a long final tone.
• In long krùu, it is becoming more common for lines to be tied together, in some cases into very long units.

Krùu ɔ̀ɔy róoy (Examples 61–62) is a polite type of krùu, which is used when the shaman is asking a spirit to leave a person’s body or when something is gently requested. This type is more melodious, long-drawn-out and slower compared to the krùu for driving a spirit out by scaring it. It has four sections of increasing speed. Section A is the most melodious one and has a tempo of 75 beats per minute, only slightly faster than a resting heartbeat. In comparison, the tempo of ‘standard’ krùu is 150–190, more than twice as fast. The tempo increases in the later sections: B has 96, C 115, and D 164. Musical phrases centre on the tonic (1 in number notation).

Example 61 The 2nd and 3rd prosodic phrases of Krùu ɔ̀ɔy róoy, Spell for calling out spirits (cf. Example 62). A vertical line indicates the break between the prosodic phrases. High and Low lexical tones are indicated below the words. Arrows show the realization of Low and High lexical tone at the phrase endings. The corresponding number notation is given at the bottom of the graph. Performed by Kàm Ràw, ca. 1983/84. Original pitch: 1 ≈ 139 Hz. • 12 Krùu ɔ̀ɔy róoy

As illustrated in Example 61, High lexical tones are normally performed at a higher pitch (approximately ♭3 or 4). A musical phrase generally ends on the tonic, which is indicated by arrows. If the last syllable is Low the musical phrase will end on a low pitch (1), and if the last syllable is High it will usually end high and falling (♭3⇒1). In the graph, a vertical line shows the division between the two prosodic phrases. In this case the prosodic and musical phrases coincide. Each phrase consists of 5 syllables (6 or 7 syllables if reduplicants are counted). In performances, there may be a lower or a higher number of syllables in a phrase.

Example 62 Krùu ɔ̀ɔy róoy, Spell for calling out spirits, Number notation. Performed by Kàm Ràw, ca. 1983/84. Original pitch: 1 ≈ 139 Hz.

See Appendix 2 for number notation signs.

Coincidence of lexical tone and musical pitch is marked in bold italics.

Glissando is marked with ⇒. For example: ♭3⇒1 denotes glissando from a pitch where tone and musical pitch coincide (♭3) to a pitch where they do not coincide (1).

[ ] = uncertain word or lexical tone.

( ) = uncertain pitch.

A Tempo: 1 ≈ 75/minute
Prosodic
phrase
No.:
	1	1	♭3⇒ 1 –
1	àk-	àk-	ʌ̀ʌy

	♭3⇒4	2 1	♭3⇒1	1	1	–
2	kúu	mὲɛn-ɛ	mɔ́ɔ	càaŋ	cɔ̀ɔy

	♭3⇒4	2 1	1 1	1	♭3⇒1	–
3	kúu	cə̀ə	ɔ̀ɔy-	ɔ	phíi	hàay

	1	1	1 1	♭3⇒1		(♭3)⇒1
4	kúu	cə̀ə	pàay-a	phíi		sɨ́a	Hrm

B Tempo: 1 ≈ 95/minute
	4	4	5⇒4	♭3⇒1	1 –
5	mɨ̀a	yə̀ə,	phíi	mɨ̀a	yə̀ə

	1	1	♭3⇒1	1	♭3⇒1 –
6	mɨ̀a	yàam	kúu	wàa	dáay

	1	1	♭3⇒1	(♭3)⇒1	1	–
7	pàay	yə̀ə,	phíi		pàay	yə̀ə,

	4	1	♭3⇒1	1		3⇒1
8	pàay	yàam	kúu	wàa		díi	Hrm

C Tempo: 1 ≈ 115/minute
	1	1	1	♭3⇒1	♭3⇒1	–
9	krɨ̀aŋ		díi,	pə̀ən	pɛ́ɛŋ	lὲɛw

	1	♭3	♭3⇒1	1	♭3⇒1 –
10	Nàaŋ	Kɛ́ɛw	ɔ̀ɔk-ɔ	màa	kín

	1	1	1	♭3⇒1	♭3⇒1	1
11	níi	yə̀ə,	Àay	Khóŋ	Khóot-	ɔ

	1	1	1	♭3	1
12	níi	yə̀ə,	Àay	Khóot [Khóŋ?]	Mɨ̀aŋ	Hrm

D Tempo: 1 ≈ 165/minute
	4	5⇒4	5⇒4	♭3⇒1	1 –
13	mɨ̀ŋ	bɔ́ɔ	ˀyáan	kúu	rɨ̀ɨ

	4	4	4	♭3⇒1	1	–
14	tíin	kúu	tém	cə́-	khép

	1	1	1	♭3	4	♭3⇒1	♭3⇒1	1
15	lèp	mɨ̀ɨ	kúu	lə̀ə	tém	mὲɛŋ	pɔ́ɔŋ-	ɔ

	1	1	♭3	1			♭3⇒1	–
16	tíin	kúu	tém	cə́-			khép,

	1	1	1	1	♭3	1	1	–
17	lèp	mɨ̀ɨ	kúu	lə̀ə	tém[kúu?]	ŋùu	lɨ̀am

	1	1	♭3	1	♭3⇒1	1	1	–
18	òom	rə̀ -	sáp,	Prà	báŋ	khán	kúu		Hrm

Interpretation:

[A] Àk-àk-ʌ̀ʌy,

I am a shaman who can help

I shall call out the evil spirits

I shall drive away the tiger spirit

Hrm

[B] Go back, spirits, go back

Leave while I speak politely

Go back, spirits, go back

Go back while I speak nicely

Hrm

[C] People have made good food for you

Miss Kɛ́ɛw come out and eat

Go away Master Khóng Khóot

Go away Master Khóot Mɨ̀aŋ

Hrm

[D] Do you not fear me?

My feet are full of centipedes

My nails are full of scorpions

My feet are full of centipedes

My nails are full of pythons

[Go away while I speak politely

Go away while I speak nicely]

om close, Buddha protect my tray

Characteristics of Krùu ɔ̀ɔy róoy

• The dominating pitch is 1. Range: 7 semitones.
• Syllabic and the last syllable of a line is long and generally glides upwards or downwards.
• Most lines start at a low pitch 1, independent of lexical tone. There may be 2–4 consecutive syllables like this. This is interpreted as an initial formula similar to the practice in hrlɨ̀ɨ (Analysis 4).
• Final syllables are performed 3⇒1 or 1 regardless of lexical tone. This can be considered a final formula in which lexical tone does not matter, similar to tə́əm (Analysis 9) and several other genres.
• High lexical tones are realized as ♭3, 4 or 5, often gliding to or from these pitches upwards: ♭3⇒4, 5⇒4, or downwards: ♭3⇒1.
• Low lexical tones are generally 1, but may be higher.
• The first lines of sections B, C, and D begin without an initial formula, so that the lexical tones are realized in musical pitches from the very start. The pitches are higher here. The patterns are 4–5–4–3–1, 4–4–5–3–1, 4–4–4–3–1. The range is one factor that gives Krùu ɔ̀ɔy róoy a more melodic touch, in combination with a rather low speed. These may be the features that make it ‘polite’. In these passages, words with Low lexical tone may be realized as high as 4. The principle seems to be that it is not higher than the adjacent High lexical tone.

Krùu ptú róoy rwàay ɔ̀ɔk (Example 63) is intended to drive out the tiger spirit, róoy rwàay, which is a vampire spirit considered to be very evil and dangerous. In contrast to the previous krùu, which talks fairly pleasantly to a spirit, this krùu is not at all polite but very direct or even rude, using strong words to scare out the spirit. It is longer and consists of 8 sections (6 of which were transcribed) and about 70 lines, compared to less than 20 for the previous one. The tempo is faster, equalling approximately 232 beats per second from the start. Section B has 257, compared with an average of 95 for the polite krùu (Example 62). The remainder averages around 260, which means that the tempo does not increase much but is rather fast from the start. The tonal range is smaller, encompassing 3 semitones (pitches 1–3), whereas the polite style has 7 semitones (1–5). This means that the High lexical tones, though clearly audible, are not very salient. The intervals between the pitches also tend to be narrower. Reduplication of vowels occurs only in the very first line. The musical phrases are constructed as in the previous krùu and as described in Example 61. One exception is that the final phrase of a section generally ends with an upward sliding motion (1⇒♭3 or 2⇒♭3).

In Example 63, vertical lines are added in the margins. A line in the left margin marks a prosodic phrase-pair. A line in the right margin marks a musical phrase that is prolonged to encompass two or more prosodic phrases. For practical reasons, solid and dashed lines are alternated. In this way, it is possible to obtain an overview of how prosodic and melodic phrases are combined.

The first 8 prosodic phrases coincide with the musical phrases in the same manner as in Example 62. The prosodic phrase No. 9, however, ends on a High syllable (ˀyóoŋ) that is performed low (1), while the musical phrase ends in No. 10, the last syllable of which is High and performed high–low (♭3⇒1). This results in one melodic phrase being prolonged to encompass two prosodic phrases (9–10). There are several cases of a musical phrase being prolonged to encompass 2–5 prosodic phrases. In prolonged musical phrases, long sequences of syllables will often be performed low (pitch 1), regardless of lexical tone. This practice can partly be explained by the fact that each section (A, B, C, etc.) is performed in one breath. When a section gets ‘wordier’, more words need to be squeezed in, since the performer needs to save the air in his lungs, and prosodic phrases are then tied together into prolonged musical phrases.

There is only one case where a prolonged musical phrase corresponds exactly to one prosodic phrase-pair (39–40), but there are more cases when a prolonged musical phrase includes a prosodic phrase-pair (16–17, 18–19, 30–31, 35–36). It may be concluded that prosodic and musical phrases often coincide; but when the tempo is fast and the number of words in a section increases, musical phrasing tends to dominate.

This krùu shares many of the characteristics of the former one (Example 62), but some differences may be noted:

• Range: 3 semitones.
• In most cases the final syllable of a section is performed with an upward glissando: 1⇒3 or 2⇒3.
• High lexical tones are realized at 3 or 2.
• Several sections start without an initial formula (sections B, D, E, H).
• In several cases whole lines are performed at one pitch (1) regardless of lexical tones, particularly in the latter part of sections.
• In several cases three lines are grouped together so that lines 2 and 3 begin without an initial formula.

Example 63 Krùu ptú róoy rwàay ɔ̀ɔk, Spell for driving out the tiger spirit. Number notation. Performed by Kàm Ràw, ca. 1983/84. Original pitch: 1 ≈ 139 Hz.

For number notation signs, see Appendix 2.

Vertical line left margin = prosodic phrase pairs.

Vertical line right margin = musical phrase longer than a prosodic phrase.

Coincidence of lexical tone and musical pitch is marked in bold italics.

Glissando is marked with ⇒. For example: ♭3⇒1 denotes glissando from a pitch where tone and musical pitch coincide (♭3) to a pitch where they do not coincide (1).

A Tempo: 1 ≈ 232/minute
	1 –	1	1	♭3	2	1	1	♭3⇒1 –
1	òom	cúk-	u	cík-	i,	Prà	cə́-	kháan

	♭3	1	1	1				♭3⇒1 –
2	sáaŋ	wáan,	Prà	cə́-				khóo

	♭3	♭3	1	1				♭3⇒1 –
3	phák	khóo-	mòo	àa-				cáan

	1	1	1	1	1	1	2⇒♭3 –
4	wáan	àn-	nì	pén	wáan	àa-	sáŋ	Hrm

B Tempo: 1 ≈ 257/minute
	♭3	2	1	1	1	1	1	1	♭3⇒1 –
5	wáan	àn-	nì	pén	wáan	ìi-	lì	ìi-	lì

	♭3	♭3	2	2					2⇒1 –
6	híin	khóp-	khép	khɔ́ɔ-					khɛ́ɛ,

	1	1	1	1	♭3⇒1 –
7	nàm	cók-	cék	cɔ́ɔ-	cɛ́ɛ

	♭3	2	1	1	♭3⇒1 –
8	kúu	cə̀ə	sá	phíi	pɔ́ɔp

	1	1	1	1	1 –
9	phíi	pɔ́ɔp	màa	ˀyóoŋ-	ˀyóoŋ

	1	1	♭3	2	♭3⇒1 –
10	kúu	cə̀ə	sá	phíi	phóoŋ

	♭3	2	1	1			1 –
11	phíi	phóoŋ	màa	ˀyáay-			ˀyáay

	1	1	1	1	2⇒♭3 –
12	kúu	cə̀ə	sá	phíi	[pháay|haay] Hrm

C Tempo: 1 ≈ 240/minute • 13 Krùu ptú róoy rwàay ɔ̀ɔk, C–D
	1	1	1	1			1 –
13	phíi	pháay*	màa	ˀyá-			ˀyá	[* haay?]

	1	1	♭3	♭3	♭3⇒1 –
14	kúu	cə̀ə	sá	phíi	ká

	♭3	2	1	1	1 –
15	phíi	ká	màa	ˀyɨ́aʔ-	ˀyɨ́aʔ

	1	1	♭3	♭3	2⇒1	1	1 –
16	phíi	lə̀ə	kép	sḿ-	pɔ́ɔy	tàaŋ	dèey

	1	1	1	1	2	1	1 –
17	phíi	lə̀ə	kép	sḿ-	pɔ́ɔy	tàaŋ	kók

	1	1	1	1	1	1	1⇒♭3 –
18	kìi	lə̀ə	yòk	sḿ-	pɔ́ɔy	tàaŋ	dèey

	♭3	2	1	2	1	1	♭3⇒1 –
19	phíi	lə̀ə	yòk	sḿ-	pɔ́ɔy	tàaŋ	háak

	♭3	2	♭3	1	1 –
20	máak	màn	tók	tàaŋ	dèey

	1	1	1	1	1	1	1⇒♭3 –
21	máak	màn	tók	sáam	sú	sáam	sáay

	♭3	♭3	♭3	2	1 –
22	Mɨ́ɨn	kɔ́ɔn	kép	bɔ́ɔ	dèey

	♭3	♭3	♭3	2	1 –
23	Sɛ́ɛn	kɔ́ɔn	kép	bɔ́ɔ	dèey

	1	1	1	1			1⇒♭3 –
24	Kúu	kép,	kúu	yàŋ			dèey	Hrm

D Tempo: 1 ≈ 282/minute • 13 Krùu ptú róoy rwàay ɔ̀ɔk, C–D
	♭3	2	1	1				♭3 –
25	Phíi	lə̀ə	wàa	ṕ-				sáŋ

	♭3	2	1	1	♭3⇒1 –
26	phíi	lə̀ə	wàa	sḿ-	pɔ́ɔy,

	1	1	1 –	1	♭3	♭3⇒1 –
27	kúu	[bɔ́ɔ]	wàa	lὲ	sḿ-	pɔ́ɔy

	1	1	1	1	♭3 –
28	kúu	cə̀ə	wàa	sóm-	pá

	2	2	♭3	2	1	1	1	1 –
29	kúu	cə̀ə	sá	nàm	tóo-	róo	tée-	rée

	♭3	2	♭3	2	3	1	1 –
30	phíi	pɔ́ɔp	ˀyúu	nèey	púum	kɔ̀ɔ	lèey

	1	1	1	1	⇒♭3	1		⇒♭3 –
31	phíi	pɔ́ɔp	ˀyúu	nèey	séey	kɔ̀ɔ		ɔ̀ɔk	Hrm

E Tempo: 1 ≈ 260/minute
	♭3⇒ –	[2⇒] –	2⇒1 –
32	Kín	kɛ́ɛŋ	khɛ́ɛ

	♭3	1	1	1	♭3⇒1 –			♭3⇒1 –
33	Kúu	cə̀ə	[tɔ̀ɔ]*	táa	phíi			[pɔ́ɔp|bɔɔt?]	[* tɔm?]

	1	1	1 –
34	Méey	wàn	ñɔ̀ɔt

	1	1	1	1	1⇒2 –	[2⇒♭3] –
35	Kúu	cə̀ə	tɔ̀ɔt	táa	[phíi|hɨ́ɨ ]	paaŋ

	♭3	♭3	[1	1]	1 –	[1]	1	2⇒♭3 –
36	phíi	pɔ́ɔp	sə́-	wáay	lɔ̀ɔŋ,	laaŋ	kɔ̀ɔ	hán

	♭3	2	1	1	♭3	1	2⇒1 –
37	Phíi	táa	dám	táa	khám	màa	khɨ́ɨn

	[1	1]	1 –	[♭3]	1			1 –
38	Ra	raam	ràay,	seey	yàam			ràay	Hrm

H Tempo: 1 ≈ 260/minute
	♭3	♭3	♭3	2	1	1	♭3⇒1 –
39	sɔ́ɔŋ	wák	hɛ́ɛ	kúu	mìi	mὲɛŋ	pɔ́ɔŋ

	♭3	♭3	2	1	1	1	♭3⇒1 –
40	sɔ́ɔŋ	tɔ̀ɔŋ	tɔ́ɔm	kúu	mìi	khwán	pháa

	1	1	1	[1]	1	[1]	[1⇒2] –
41	pə̀ən	lɨ̀ɨ-	sáa	kə́-	tháa	[mòn]	[díi]

	♭3	2	1 –
42	kúu	tə̀ŋ	mɨ̀aŋ

	♭3	2	1	[1]	1	1	1 –
43	phíi	táa	lɨ́aŋ,	tə̀-	nàŋ	khíi-	mìn

	1	1	1	1	1⇒♭3 –
44	sá	phíi	níi	tə̀ŋ	khám

	1	1	♭3⇒1 –
45	cám	nɔ̀ɔk	pháa

	1	[1]	1	1	1	1	1⇒♭3 –
46	òom	[rə̀-	tàa	Prà	mòn]	khàn	kúu	Hrm

Interpretation:

[A] Buddha, ckháan pepper

Arrange medicine, Buddha ckhóo

Herb teacher

This medicinal plant, what medicinal plant is it? Hrm

[B] This medicinal plant is a real, real medicinal plant

Khóp-khép khɔ́ɔ-khɛ́ɛ stone

Cók-cék cɔ́ɔ-cɛ́ɛ water

I shall drive away the tiger spirit

The tiger spirit returns again and again

I shall drive away the spirits of waste

The spirits of waste return again and again

I shall drive away the miscarriage spirit. Hrm

[C] The miscarriage spirit returns again and again

I shall drive away the tiger spirit

The tiger spirit returns again and again

Where do the spirits collect holy fruits?

The spirits collect holy fruits at the tree trunk

Where do the spirits collect holy fruits?

I shall collect holy fruits at the root

Where do the betel nuts fall?

The betel nuts fall everywhere

Ten thousand people cannot pick them up

A hundred thousand people cannot pick them up

Only I can pick them up. Hrm

[D] Some call it holy fruits,

Some call it holy water

I do not call it holy water

I shall call it sóm-pá

I shall pour out the tóo-róo tée-rée water

The tiger spirit in the stomach will be driven out

The tiger spirit in the guts will come out. Hrm

[E] Eating khɛ́ɛ soup

I will hit the tiger spirit’s eye

A wicker twined around the tree’s top

I will hit the tiger spirit’s eye blind

I indeed see the tiger spirit deep down in the stomach

The black spirit eyes, the golden spirit eyes are coming up

This is a hard time starting for the spirit. Hrm

[H] There are scorpions on my casting-net hook

There are heavenly souls on my tɔ̀ɔŋ tɔ́ɔm hook

All around the country my magic is praised

Yellow-eyed spirit,

Shaman’s helper khíi-mìn

Drive out the spirit so that it runs all night

Goes behind the sky

Òom rtàa, Buddha mantra, protect my tray

The lexical tones in krùu

Measurements were made and represented in graphs in order to see how the linguistic phrases occur in this krùu and how the lexical tones are handled. The number notation above the text is transcribed after what can be heard musically, that is the tonal ‘centre’ or ‘musical approximation’ of a performed tone, while the exact measuring focuses on the maximum points. These two do not always coincide. In many krùu, the separate sections are marked by a higher and louder ending, much like the rice narratives. In these cases, the melodic contour is hence dominated by the intonation at the endings.

In the following graphic representations of phrases from Krùu mòn òm smpɔ́ɔy, Spell for making holy water (Example 64), the maximum pitch of each syllable is shown. The prosodic phrases are grouped in phrase-pairs and the axis on which two phrases are mirrored, the mirror break, is marked with a vertical line. Example 64A shows the first two lines that constitute one thematic episode mirrored in meaning, as the vertical line shows. In this case, the thematic episode coincides with a musical phrase. The main melodic outline is often mirrored, too: the first part goes from high to low and the second part from low to high.

Example 64B shows a different melodic contour. The first part goes from low to high and the second part from high to low. In this part of the performance, the tempo has increased and the performer has started to combine more than two prosodic phrases into one musical phrase. The first part in this figure is actually the final part of a single musical phrase (ending high), and the second part is the beginning of the next musical phrase (starting high and ending low). It can also be noted that in the first part all syllables except the last one are performed low, regardless of tone. This occurs in krùu – as in tə́əm – in final musical phrases when the performer is running out of breath. This is also obvious in the final phrase of the performance, shown in Example 64C.

Example 64 Prosodic phrase pairs and musical phrases from Krùu mòn òm smpɔ́ɔy, Spell for making holy water. Performed by Kàm Ràw, ca. 1983/84. Original pitch: 1 ≈ 140 Hz.

Graphs: Vertical: Hz, horizontal: time (unspecified). The mirror break is marked by a vertical line.

Number notation: For number notation signs see Appendix 2.

A Speed: 1 ≈ 192/minute

B Speed: 1 ≈ 228/minute

C Speed: 1 ≈ 240/minute

This genre is melody-centred, and lexical tones accommodate to the rising or falling general tonal contour. The situation is similar to that of speech, with a fixed melodic formula and lexical tones accommodating to it. The main difference is that tonal identities are neutralized in krùu by levelling the transitions between tones when in direct conflict with the melody. The realization of the lexical tones in krùu is thus relative, rather than an exact pitch. Using our approach with a performance template in which melodic movement and tones can be studied as two separate phonetic levels, differences in the realization of lexical tones can be accounted for systematically.

Example 65 Illustration of two phrases without conflicts between lexical tones or between lexical tones and melody. The melodic outline is indicated by a line. Arrows point to transitions discussed in the running text. Only the F0 values at the beginning of the vowels in each syllable are shown, as this is the point at which the identity of lexical tones is phonetically anchored.

In Example 65, two phrases without conflicts between tones or between tones and melody are illustrated. The melody falls towards the end. Tonal combinations are transitions from sequences of High to sequences of Low tones. Hence, there are no tonal conflicts. Transitions between tones similar to the melodic direction are enhanced (indicated by arrows). There is reason to distinguish between an early and a late downslope or upslope in the melody (upslope is dealt with below). In the upper plot, there is a late downslope, and the melody is kept high with no enhanced transitions until the H + L tone sequence. Melodic downslope is aligned with this sequence and enhances the contrast between the two tones. In the bottom plot, there is an early downslope. Lexical tones are realized in such non-conflicting combinations of tones and melody.

Example 66 Illustration of phrases without conflicts between tones, but with conflict between tones and melody.

There are many tonal movements within syllables. Lexical tones are preserved by different techniques. For example, a rising movement may occur in the sonorant coda of a Low tone word while the Low lexical tone is preserved on the vowel. The relation between melody and lexical tones is therefore very similar to that found in speech. In principle, melody/intonation is leading unless it conflicts with lexical tones. In krùu, the High tone in the combination L+H+L in the rising contour boosts the High tone so that it dominates over the melodic contour. Furthermore, a final Low tone is realized higher when the line is the last one in a thematic episode. A thematic episode is normally marked by a high boundary tone that influences the realization of a final Low lexical tone by raising it. The same pattern is found in speech: intonational marking of boundaries (by high phrase boundary tones) influences the realization of lexical tones.

In the upper plot of Example 66, the melody is rising and conflicts with the final Low tones. The transition from High to Low in kə́ət tìi is not enhanced. This transition conflicts with the rising contour of the melody. Transitions between tones are suppressed and the melody does not rise until at the end, somewhat jeopardizing the identity of the last Low tone. Tones are neutralized and melody is prioritized. In this case, the phrase is tied together with the following phrase (not in the illustration), so there is no actual ending of the musical phrase in the example. The melody is rising on the last Low syllable because the first syllable of the following phrase is High and will be intoned high. In the bottom plot, the melody falls and conflicts with the final High tones. The transition from Low to High in pə̀ən tháak is suppressed as it directly conflicts with the falling contour, and the fall of the melody is aligned with the High–High sequence tháak bɔ́ɔ where it does not conflict so much with tonal identities.

Example 67 Illustrations of phrases with conflicts between tones (i.e. sequences with intervening opposite tones) and no conflicts between tones and melody. The dotted arrows show High tones for final syllables intoned with a downward sliding motion.

In the plots in Example 67, the conflict with the final High tones is of a different kind. This rising pattern is realized by suppressing High tones at the beginning of the phrases, and the first combination of tones Low–High (which goes in the same direction as the rising melody) is enhanced. Transitions of High–Low combinations are suppressed. At a faster tempo with too many conflicts between adjacent tones, transitions between tones are levelled, and there are no enhancements. The High tone of the last syllable is realized with a falling movement to the lowest pitch. The most common ending of a musical phrase is low, and this pattern dominates over the lexical tone. In this genre, however, the last musical phrase of a section ends with a high pitch (Example 68).

Example 68 Illustration of a phrase with many conflicts between lexical tones, performed at a rapid tempo.

It seems to be a common technique for Kammu genres that transitions between lexical tones are levelled in conflicting situations, so that High and Low are realized on the same level (usually within a somewhat lower pitch range). We do not find ‘wrong’ tonal transitions (e.g. falling pitch) between Low and High tones. The only exceptions are phrases in a very rapid tempo.

The krùu performance template

Melody

• Short melodic phrases with a tonal centre (see Example 61).
• Melodic range: 3–7 semitones.

Rhythm

• Regular pulse.
• Syllabic.

Form

• Litany: the same musical phrase repeated with variations.
• There are one or more sections performed in one breath and finished with ‘clearing the throat’ (‘hrm’).
• The tempo often increases with each new section.

Phrasing

• The prosodic phrases are organized as phrase-pairs: as question and answer, or antithetically.
• A number of prosodic phrases form a thematic episode.
• Verbal phrasing dominates (a 7-syllable line is two tone durations longer than a 5-syllable line, etc.)
• For most of a performance, prosodic and musical phrases are aligned.
• Musical phrases may be prolonged and incorporate two or more prosodic phrases.

Initial/final formulae

• Many krùu have an initial formula, generally with non-lexical words, performed to a variant of the basic musical phrase.
• There may be two or more tones at a low pitch in the beginning of a phrase. This is interpreted as an initial formula.
• Final syllables are performed 3⇒1 or 1. This may be considered a final formula in which lexical tone does not matter.

Word variations

• Same tone duration for long and short vowels, minor and major syllables.
• Reduplication occurs, but is rather infrequent.
• Schwa vowels (normally ə) have the same duration as all other vowels that are performed as short.

Lexical tones

• Basically melody-centred.
• Low lexical tones are generally 1 (but may be higher) or falling (3⇒1).
• High lexical tones are realized at 2, 3, or 4, often gliding to or from these pitches upwards: 1⇒3, 2⇒3, 2⇒3, 3⇒4, or downwards: 2⇒1, 3⇒1.
• The 4th and following syllables of the 2nd phrase of a phrase-pair are generally low, independently of lexical tones.
• In longer combinations, all or most syllables of the second line may be low (1) with no contrary motion in such a passage.

Kammu summary

In the investigation of ten genres of Kammu vocal expressions, musicological approaches and linguistic approaches were combined. We investigated how phonological features are phonetically implemented in performance. The main questions were if, how, and to what degree those phonological features exist in these vocal genres, and whether we would find features that do not occur in speech. We also studied how melody is built up in different genres, that is, whether the melody is primarily built up of lexical tones or if there is a tonal contour to which tones are accommodated.

The syllable was chosen as the domain for comparison. The following phonetic features were compared between genres: lengthening within syllables (whether onset, nucleus, or coda is lengthened in some positions), reduplication of syllables, tonal movements within syllables in terms of the number of tonal turning points and the direction of movement (e.g. tonal movement up–down, up–down–up, etc.), and the alignment of tonal movements within syllables (if the tonal movement starts at the onset or later).

This chapter has shown how performance templates are constructed and how one particular person realizes lexical tones in different genres while also performing vocal expressions in which lexical tones are realized differently. The main characteristics of these genres are listed in Table 4.

Three types of genre-dependent relationships between tones and melody occur in Kammu vocal expressions: tone-centred, melody-centred, and tone- and melody-centred genres. In tone-centred genres, the melody is built up by the lexical tones. In melody-centred genres, lexical tones adjust to a fixed melodic contour. Performance templates may have initial and final formulae with rather fixed pitches, sometimes also fixed words (melody-centred sections), and variation mainly occurs in the middle part (tone-centred sections). Those genres are tone- and melody-centred. In Table 4, the vocal expressions have been grouped in accordance with the existence of melody-centred, tone-centred, and mixed melody and tone-centred sections.

• Tone-centred: These genres are performed at pitches that are close to speech intonation (Analyses 1–2, 4). They do not have the character of fixed melodies, and lexical tones are realized. They can be adapted to poetical lines of varying lengths.
• Melody-centred: These genres may be experienced as having fixed melodies, and in two cases this really applies, since they afford little scope for realizing lexical tones (Analyses 3, 8), while lexical tones are realized in other cases (Analyses 5–7). On the other hand, they can all be adapted to poetical lines of varying lengths, which makes them function as performance templates.
• Melody-centred and tone-centred parts: These genres have shorter or longer initial and final formulae with rather fixed pitches (the melody-centred parts), and most of the remainder of the performance is dominated by lexical tones (the tone-centred parts) (Analyses 9–10). They also have their respective rhythmical metres and permit the lengthening or shortening of poetic lines.

Typically, vocal expressions that can be explained by performance templates are open to variation and improvisation. The templates include practices which enable the performance of longer or shorter poetical lines. In the majority of cases, prolongation occurs in the middle of the vocal expression, while initial and final formulae remain relatively unchanged. Existing lines can also be prolonged by putting in additional words at the end of a line by prolongation (the musical phrase is simply prolonged) or contraction (several syllables are squeezed into the musical phrase without changing its length). The rhythm patterns and poetic metre of the templates vary between different genres. Some genres make use of non-lexical words or ‘song-words’. These are integrated parts of the performance templates, and one of their functions is to mark the genre.

Some words are pronounced differently as compared with ordinary speech. In most of the genres no distinction is made between long and short vowels; the schwa vowels in minor syllables are pronounced with the same duration as phonemic vowels (vowels occurring in major syllables), and sometimes they are even prolonged. Otherwise, prolongation of syllables generally occurs in the coda. Reduplication of syllables is common. Embellishments such as vibrato and other fast tonal movements are normally realized on the final part of the base syllable, or on the reduplicant.

The two lexical tones are realized in most genres, but in different ways. In melody-centred sections, lexical tones are not necessarily realized in the initial or final formulae. Shorter or longer passages occur where syllables are performed low, regardless of lexical tones. In tone-centred parts, the Low and High lexical tones are realized in performance, at either relational or fixed pitch levels. Lexical tone is realized in the initial part of the vowel, and tonal movements in the reduplicant and coda are independent of the lexical tones. Some genres have complex tonal contours in the base syllable, for instance húuwə̀ (Analysis 6), while in other genres, such as tə́əm (Analysis 9), complex tonal movements occur only in reduplicants.

Prosodic phrases are often paired and linked to each other by parallelism and rhyme. In the case of the orally transmitted trnə̀əm poem, parallelism and rhyme are extended to form pairs of stanzas that are performed in various vocal genres depending on time, location, age, and gender (Analysis 4–9). Prosodic and musical phrases generally coincide. The musical metre tends to depend on the prosodic phrasing in terms of length and verbal metre, but there are also cases where a musical phrase includes more than one prosodic phrase.

In many cases, the melodic contour is descending, starting at a high pitch and ending lower. The final tone is low and sometimes ends by sliding further downwards. One exception is hrlɨ̀ɨ (Analysis 4), which apart from an initial formula has no melodic movement or speech intonation at all, only a high and a low pitch relating to lexical tones. The influence of speech intonation is noted in prayers (Analysis 2) and in vocal genres where the final tone is high, which is similar to speech intonation (Analyses 1, 3, 10). The majority of the vocal genres have a tonal centre and a regular rhythm (Analyses 3–11).

Most of the Kammu vocal repertoire can thus be understood as based on performance templates. This means that it is possible to analyse a quantitatively extensive material by combining the use of performance templates with graphic transcriptions, and then to actually analyse large parts only by hearing. Musical and linguistic methods of analysis complement each other by producing overall interpretations on the one hand and exact measurements on the other.