Court and religious music (2): music of gagaku and shōmyō

Authored by: Steven G. Nelson

The Ashgate Research Companion to Japanese Music

Print publication date:  October  2008
Online publication date:  February  2017

Print ISBN: 9780754656999
eBook ISBN: 9781315172354
Adobe ISBN: 9781351697613




The previous chapter introduced the common history and theoretical basis for the traditions of gagaku and shōmyō. In the present chapter we look separately at the musical details of these two genres: instrumentation, notation, performance practice and so forth.

 Add to shortlist  Cite

Court and religious music (2): music of gagaku and shōmyō

1  Introduction

The previous chapter introduced the common history and theoretical basis for the traditions of gagaku and shōmyō. In the present chapter we look separately at the musical details of these two genres: instrumentation, notation, performance practice and so forth.

2  Gagaku

2.1  The instruments and their notation

The complete instrumentarium of today’s gagaku is made up of wind, string and percussion instruments (details in garfias 1975, Nelson 2000b, Terauti 2002). These three classes are traditionally referred to as fukimono (blown things), hikimono (plucked things) and uchimono (struck things). Table 3.1 shows the use of instruments, song and dance in each of the sub-genres.

Winds: The kagurabue, ryūteki and komabue are transverse flutes, of lacquered bamboo, with six, seven and six finger-holes respectively. The hichiriki is a short double-reed pipe, of lacquered bamboo, with seven (front) and two (back) finger-holes and a comparatively large reed. The shō is a free-reed mouth-organ with 17 bamboo pipes (two reedless) inserted into a wind chamber, and is sounded by both exhaling and inhaling (Figure 3.1).

Strings: The wagon (or yamato-goto) is a six-stringed long zither, tuned in zigzag pentatonic tunings, often said to be indigenous to Japan, but likely based on an ancient Korean model. The biwa is a large four-stringed lute with four frets, played with a large hand-held plectrum. The koto (or ) is a 13-stringed long zither, tuned pentatonically, played with plectra placed on the thumb and two fingers of the right hand (Figure 3.2).

Table 3.1   Instruments, song and dance in the sub-genres of gagaku

A. Kuniburi-no-utamai (accompanied vocal music and dance of indigenous origin employed in imperial and Shinto ceremony)












B. Kangen and bugaku (instrumental music and accompanied dance deriving from the ancient performing arts of the Asian mainland)

















C. Saibara and roei (accompanied vocal music originating at the early Heian court of the ninth and tenth centuries)






















◊ The strings are used in exceptional performances of bugaku called kangen-bugaku, literally ‘pipe-and-string dance’.



Wind instruments, left, top to bottom:

Figure 3.1   Wind instruments, left, top to bottom: ryūteki, komabue, kagurabue, hichiriki; right: shō

String instruments:

Figure 3.2   String instruments: biwa and koto

Percussion: The shakubyōshi is a pair of clappers comprising two flat pieces of wood. The kakko is a small barrel drum struck with a separate stick on each of its two heads. The san-no-tsuzumi is a small double-headed hourglass drum struck on only one head with a single stick. The shōko is a small, suspended brass gong struck with two sticks. The taiko is a large shallow barrel-drum with ox-hide heads, struck on one side with two sticks. Large versions of the shōko and taiko (ōshōko and dadaiko), when available, accompany dance (Figure 3.3).

Notations for gagaku instruments apparently derive from ancient Chinese models. Substantial collections of notation for transverse flute, biwa and koto survive from the heian period, and music notation was clearly of great importance for the higher-ranking nobles in their performance and transmission of the art. In contrast, the tradition of the professional music families, upon which present-day palace practice is based, has for centuries put more emphasis on oral transmission, with notation viewed as a supplementary tool for facilitating memorization.

The notations for the wind and string instruments are essentially tablatures, that is, they show fingerings, strings or frets to be used, rather than the pitches to be sounded. For tōgaku and komagaku pieces, the modern notation for flute and hichiriki provides a mnemonic version of the melody, known as shōga. This expresses the melody in a systematic way, with consonant sounds indicating phrasing and certain performance techniques, and vowel sounds indicating pitch relations between neighbouring notes. Musicians learn their parts by singing the shōga, often for some time before the instrument is played.

Figure 3.4 shows the tablatures for wind and string instruments for the opening of the tōgaku piece Goshōraku no Kyū. The shōga for the ryūteki and hichiriki is written in the Japanese katakana syllabary as the central column in the notation for each instrument; tablature signs are given to its left. Shō notation gives single tablature signs that now indicate the names of the aitake chords. As shōga for the instrument, these names are sung to a melody most resembling that of the hichiriki. Notation for the biwa gives the names of the open strings and positions on the frets, while that for the koto gives string names. (Figure 3.4 is read right-to-left.)

2.2  Tōgaku in modern performance practice

Goshōraku no Kyū is one movement of the tōgaku suite Goshōraku, literally ‘Music of the Five virtues’. Said to have been composed in China in the Zhengguan era (627–49), it was first performed in Japan in 702. The title refers to the five moral principles of Confucianism (benevolence, justice, propriety, knowledge and honesty), one set within the Chinese theory of the Five elements. In its full form, Goshōraku is made up of four movements, demonstrative of the ‘ideal’ structure for a gagaku piece: Jo (introduction), an extensive, slow movement in free rhythm; Ei (song), originally with a sung Chinese text, now lost; Ha (breaking away), a long movement in a slightly faster, fixed rhythm; and Kyū (rushing), a shorter, faster final movement. At present, the whole can be performed in bugaku style as a dance suite for four dancers, while the final two movements can also be performed in kangen style. Today, however, it is rare for any but the last movement to be performed.

Percussion instruments, left:

Figure 3.3   Percussion instruments, left: kakko; right: shōko; below: taiko

Tablature for winds and strings for opening of

Figure 3.4   Tablature for winds and strings for opening of Goshōraku no Kyū

Tōgaku pieces are classified according to size into three classes: taikyoku (large pieces); chūkyoku (middle-sized pieces); and shōkyoku (small pieces). The taikyoku are generally extensive suites in several movements. They are only sometimes performed as bugaku, and rarely as kangen, due to their extreme length. The individual movements of Goshōraku are classified as chūkyoku, except for the final movement, which is a shōkyoku.

Example 3.1 shows the repeat of the opening section of Goshōraku no Kyū, that is, what can be labelled in terms of melodic structure as the second A in the AABB form of the piece (☉ track 1; Audio/videography Gagaku: an Important …, disc 5). (The first A opens with a ryūteki solo and thus does not give an accurate picture of the full texture.) For winds, both the performed part and the shōga (as sung by male performers) are given. Close examination shows that the instruments play in a heterophonic ensemble style, characterized by bi- or polymodal treatment of the same melody with differing degrees of ornamentation. In modern performance practice, the flute and hichiriki parts are regarded as melodic, and the other parts as structural. As will be shown below, however, the relationship among instruments can be understood more fully if historical factors are considered.

Goshōraku belongs to the hyōjō mode, which can be described theoretically in terms of the Western church modes as a dorian species on e (e–f#–g–a–b–c#–d–e). The strings and mouth-organ shō largely limit themselves to these tones, but the flute and especially hichiriki often lower the f#, g, c# and d by a semitone or more. The hichiriki part largely eliminates the sharp tones and avoids g, producing an essentially phrygian version of the melody, while the flute part fluctuates between the two. Dissonances are hence common on tones other than e, a and b – the first, fourth and fifth degrees of the mode – which are stable and uninflected. A similar degree of modal complexity can be seen in all tōgaku modes, perhaps most extremely in the mode ichikotsuchō as a result of multiple layers of historical change.

Although Goshōraku occurs only in hyōjō, the modern repertoire includes a number of pieces in two or three modes as the result of transposition within the boundary of modal class (see Chapter 1, §6.1 and Table 1.4). The piece Etenraku, for instance, exists in all three ritsu modes (hyōjō, ōshiki, banshiki). Marett (2001b: 856) shows that, unlike the simple transpositions of the shō part, the hichiriki and ryūteki melodies have evolved very differently, with each version displaying the distinctive melodic characteristics of its mode. Nelson (2000b: 39–44) examines the historical factors in the development of this practice. 1

The metrical structures of tōgaku pieces with fixed rhythms are built up of repetitions of rhythmic cycles. This is described according to the formula ‘X-Y-hyōshi

Opening section (repeat) of

Example 3.1   Opening section (repeat) of Goshōraku no Kyū

(-byōshi) hyōshi Z’. Hyōshi is a term with a broad range of meanings including ‘beat’, ‘metre’, and ‘rhythm’. ‘X’ refers to the number and length of beats in a bar, ‘Y’ to the number of bars in one rhythmic cycle, and ‘Z’ to the number of cycles that make up a piece. The hyōshi in the final expression refers to the strong beat of the taiko that occurs once in each cycle. In the final few cycles, a more complex pattern known as kuwae-byōshi (added beats) is generally used. For Goshōraku no Kyū, the metrical structure is described as haya-ya-hyōshi hyōshi-hachi, literally ‘quick eight beats, [taiko] beats eight’. This means that eight (ya) bars of 4/4 (haya) make up one rhythmic cycle, which is repeated eight (hachi) times. A kuwae-byōshi pattern is used in the last two cycles (sue ni-hyōshi kuwae).

Example 3.2 shows the percussion patterns of Goshōraku no Kyū. The first five bars (uchi-hajime) are followed by an eight-bar pattern repeated five times, and a kuwae-byōshi pattern repeated twice. The strong taiko beat at the end of each pattern corresponds to the first beat of the fifth bar in each eight-bar section. The rhythmic cycles appear to be out of alignment with the overall melodic structure, but this is simply the result of the rhythmic cycles being written traditionally with the strong taiko beat at their conclusion. Comparing examples 3.1 and 3.2 shows that the strong taiko beats fall on the first beats of odd-numbered bars, reinforcing the fundamentally two-bar melodic phrasing.

A performance of Goshōraku no Kyū in kangen style begins with solo ryūteki. The percussion begin from the third bar, and all other wind players join in at the first strong beat of the taiko, at the beginning of the fifth bar. The strings then enter in staggered order, two bars apart: lead biwa at bar seven, lead koto at bar nine, second biwa at bar 11, and second koto at bar 13. This texture is maintained until just before the final taiko beat. There, all except the lead wind and string players stop playing; the leaders then perform a standard free-rhythm cadence known as tomede, which ends with single notes on the string instruments.

In a good performance of kangen, a gradual increase in tempo occurs as the piece proceeds. A piece begins at a very leisurely pace and reaches its fastest tempo (still comparatively slow) somewhat before the end, usually close to the shift to the kuwae-byōshi percussion pattern. 2 In addition, certain beats, especially the final beats in even bars, tend to be drawn out, as the finger-changes of the shō are articulated carefully during breaks in the flute and hichiriki melodies, and the biwa plays the arpeggio which often precedes the first beat of the following bar.

The precise rhythmic relationships between the instruments in kangen style give way to a rhythmically stronger ensemble style when the same piece is performed as bugaku, that is, as accompaniment to dance. 3 The emphasis here is on providing

Rhythm patterns used in

Example 3.2   Rhythm patterns used in Goshōraku no Kyū

a strong and steady beat at a quicker tempo for the dancer or dancers to move against. The shō players do not worry about articulation in the finger movements, but make them all at once with a great vitality, half a beat earlier than they do in kangen. In the case of Goshōraku no Kyū, the whole piece (AABB) must be repeated a total of five times for the dance to be completed. The relatively free 4/4 rhythm of kangen shifts to a stricter 2/2, although the principle of a gradual increase in tempo is still observed. 4 The faster tempo also means that melodic phrases that are split in kangen performance are played in a single breath.

2.3  The structural core of tōgaku and the ancient melodies of Tang China

The question of what forms the melody in modern tōgaku performance is commonly answered in two contrasting ways. Many Japanese performers and scholars point to the detailed melodic movement in the flute and hichiriki parts, while in the most widely read study of tōgaku in english, garfias bases his analyses on an abstracted melody ‘which follows the overall contour of the composition but which omits the individual stylistic characteristics of any one instrument’ (1975: 85). Research undertaken by the late Laurence Picken and his doctoral students (chiefly Rembrandt Wolpert, Allan Marett and elizabeth Markham) at Cambridge in the 1970s and ‘80s, however, has made it clear that the key to understanding the relationship between the melodic lines in modern performance practice lies in a structural core that derives from the ancient melodies of Tang China. 5

In short, this view holds that the ancient melodies are carried in the modern shō and biwa parts like a type of cantus firmus, slowed down by a factor of four or eight. In the case of Goshōraku no Kyū, for instance, this can be demonstrated as in example 3.3. Staff A shows the melody produced by reading the modern shō line with time values reduced by a factor of four, showing only the pitch indicated by the notated tablature sign (that is, ignoring the notes added in the aitake chords in modern performance practice). Staff B shows the melody produced by reading the modern biwa part in the same way, ignoring unnotated arpeggios. Staff C shows the biwa line as notated in the twelfth-century source Sango Yōroku. It can be seen that these three lines are largely identical (indeed, B and C are identical) and conform to the theoretical form of the hyōjō mode in original Chinese practice, namely dorian on E.

The implications of the view forwarded by Picken and his associates are manifold. First, if the present-day structural core is indeed the original melody, it follows that the performance tempo must originally have been faster, since it is impossible to perceive the shō or biwa lines as melody in modern performance. It also follows that the complex modern texture is an aggregate of layers of accretions to the original melody: unwritten notes added in the aitake chords on the shō; unwritten arpeggio notes on the biwa; figures once ornaments in the flute and hichiriki parts that now form the melodic surface, part notated, part orally transmitted; and corresponding modal changes in these parts conforming with the modal preferences of musicians of the later ages during which the ornaments were transformed into melody.

Proponents of the Picken view have asserted that it was proven conclusively in Markham’s doctoral work (later published as Markham 1983). As pointed out above (Chapter 2, §2.2), many saibara songs are traditionally said to have the ‘same melody’ (dōon; perhaps ‘same sound’) as certain tōgaku or komagaku pieces. On the face of it, this seems unlikely in view of the great dissimilarities between the paired pieces in modern performance. Markham, however, demonstrated that vocal melodies of saibara that she realized from biwa and koto parts of the twelfth century were closely related variants of the ancient melodies of their paired tōgaku and komagaku pieces. According to the Cambridge argument, the close identity of the two ‘could have been perceived only if what was recorded in the sources were indeed melodies’ (Marett 1986: 35; see also Marett 2001a: 860).

Example 3.3   Goshōraku no Kyū in reduced time values

2.4  Issues in research

The work of the Cambridge group has clearly been revolutionary, but it must be noted that both Japanese and non-Japanese scholars have expressed doubts. (Marett 2006 discusses and addresses these issues.) Criticism from the Japanese side has centred on philological and bibliographic issues. 6 Although much of this criticism is quite valid, Japanese scholars are often so disturbed by the shortcomings of foreign scholars in terms of scholarly literacy that they fail to appreciate the significance of the broader arguments presented. 7 Even so, there seems to be room for criticism of a more substantive nature.

Markham’s extensive comparisons of ancient and modern saibara, for instance, are made with barely a mention of the extinction and subsequent revival of the tradition. She extracts her early saibara melodies from biwa and koto parts, which scholars of another persuasion might view as accompaniment. Although her realizations of the vocal part are based on the reasonable assumption that the ensemble style was heterophonic in nature, they need to be tested against the vocal notations (that is, notations of the vocal melodies) that survive from the thirteenth to fifteenth centuries, and therefore pre-date the extinction of the genre.

Notwithstanding criticism of the Cambridge view, there is no room to doubt the importance, in analysis of musical structure, of the historical perspective they champion. The idea of change as a result of gradual accretions made to ancient melodies is far more persuasive than the opinion, sometimes still voiced by Japanese performers and scholars, that tōgaku has largely remained unchanged, and also explains the relationships between the instrumental lines in modern performance practice more plausibly than garfias’s idea of a hidden melodic abstraction treated differently by each of the instruments in the ensemble.

In coming years, research is needed on the sub-genres of gagaku that have so far attracted little musicological research: komagaku, rōei, and the various song and dance suites of the kuniburi no utamai repertoire, especially the major form mikagura. In addition, there is an extensive body of historical sources that would be worthwhile producing as facsimiles and/or typographical reproductions. Very few of these sources have been translated or even adequately annotated in Japanese. Acquiring the linguistic and other skills needed for these tasks is not an easy matter, however, and scholars interested in pursuing gagaku as a field of study should be ready to make a substantial investment of time and energy.

3  Shōmyō

3.1  Shōmyō in its ritual context

A comprehensive grasp of shōmyō must begin with an understanding of its place in Buddhist ritual. First, a distinction is to be made between two Japanese terms, hōe and hōyō. The first refers to Buddhist ceremonies, and the second to particular rites (or liturgies) celebrated within them. Although some ceremonies are celebrated as one rite lasting only a short time, up to several hours, other ceremonies are held over the interval of several days, usually an odd number. Long ceremonies last for seven or even multiples of seven days. Each day can accommodate a plural number of rites.

The difference between ceremony and rite can also be understood as a distinction between purpose and content. For example, the Shingon hōe Kōbō daishi Shō-Mieku is a ceremony held on a fixed date (21 March) as a memorial service for Kūkai, the sect’s founder. The hōyō rite often performed on this occasion is the Kongōkai-date Nika hōyō-tsuki Chūkyoku Rishu Zanmai. 8 The elements of this name designate quite specifically its content: it is an esoteric rite focusing on the diamond Realm Maṇ ḍ ala (Kongōkai-date), with the standard shōmyō pieces Bai and Sange (Nika hōyō), 9 centred around a reading of the esoteric Rishu Sutra (Rishu Zanmai) recited in chūkyoku modal style (Chūkyoku; Table 1.4, types 2 and 3). This rite can be performed with an altered structure for a different purpose at a different ceremony.

As can be inferred from this description, Buddhist rites are classified in a number of different ways. An important distinction is that between esoteric (mikkyō-date) and exoteric (kengyō-date) rites. The former include an esoteric ritual celebrated at the altar by the master, or dōshi. The latter lack such a ritual and are regarded as more ‘open’. Esoteric rites are further classified according to which of the two maṇ ḍ alas central to Japanese esoteric ritual practice is focused on: kongōkai-date rites focus on the diamond Realm Maṇ ḍ ala, while taizōkai-date rites focus on the Womb Realm Maṇ ḍ ala. 10 In addition to this difference in ritual practice, the distinction between esoteric and exoteric also has a great bearing on the choice of texts. The standard piece Bai, for instance, exists in a number of forms with different texts, which are used in either esoteric or exoteric rites.

Another classification is made according to the use of the four standard pieces, Bai, Sange, Bonnon and Shakujō, in the introductory section of the rite. When only Bai is used, a rite is said to have an ikka hōyō (one-part ritual); this form is used in both esoteric and exoteric rites. A rite with Bai and Sange (with the piece Taiyō appended as a rule) has a nika hōyō (two-part ritual); this form is generally used in esoteric rites. When all four are used, it has a shika hōyō (four-part ritual); this form is used only in exoteric rites. A third classification characterizes rites according to their central element. Kyō-date rites centre on the reading of a sutra (as is the case with the Rishu Zanmai), kōshiki-date rites are built around a kōshiki, while shu-date rites centre on the chanting of mantras.

In general terms, the programme (shidai) of a rite is decided according to the purpose, character, scale, and time of its ceremony, and can take the form of any of a number of variations on several standard structures. Rites are celebrated by a group of priests, shikishū, led by the master, dōshi, and helped by one or more assistants. The dōshi sits at the altar, generally on a worship dais, situated at the foot of a painted or sculpted image. The shikishū take their places to the left and right of the open area in front of the altar. Many of the shikishū have particular roles in a rite, generally involving solo singing at some point.

Throughout a rite, the dōshi provides signals in accordance with the programme using music instruments known as narashimono or hōgu, and the shikishū sing shōmyō following these signals. Other narashimono are played by the shikishū and attendants. The narashimono used in the Kongōkai-date Nika hōyō-tsuki Chūkyoku Rishu Zanmai, for instance, are as follows:

  • hanshō: a medium-sized suspended bell struck to call the priests together
  • inkin: a small resting bell struck during processions to and from the place of worship and at certain points during the rite
  • hachi and nyō: cymbals and a gong played together at the conclusion of Sanskrit and Chinese san (hymns)
  • kei: a suspended metal plate struck by the dōshi at important points during the rite
  • gokorei: a handbell with a five-pronged vajra (diamond pounder) head, rung by the dōshi during the esoteric Rishu Kyōbō ritual, which he performs at the centre of the rite
These are only six of the more than 20 narashimono used in the rites and daily ritual practices of the Shingon sect.

Japanese Buddhist rites are generally constructed in a symmetrical or cyclical fashion around their central element. Individual pieces of shōmyō are set out in this framework and perform specific functions within the ritual context. In the case of the esoteric Nika hōyō-tsuki Chūkyoku Rishu Zanmai rite, there are two central elements performed simultaneously: a reading of the Rishu Sutra by the shikishū, and the esoteric Rishu Kyōbō ritual celebrated by the dōshi at the altar. Table 3.2 shows the structure of the Kongōkai-date Nika hōyō-tsuki Chūkyoku Rishu Zanmai as it might well be performed at a memorial ceremony for Kūkai (held annually on 21 March). As well as pieces in Sanskrit or Chinese with substantial sections sung in unison by the shikishū that are used in this rite irrespective of ceremonial context, there are pieces (mainly in Japanese) sung solo by either the dōshi or one of the shikishū that are unique to this ceremonial context and are replaced when the rite is used for a different ceremony. Note the mirror symmetry in the relationship of the zen-shōrei and go-shōrei sections (each made up of a group of several pieces) and the zen-san and go-san categories of hymns of laudation, as well as the five-section structure that can be seen working in the background. 11

3.2  Pitch, mode and vocal range; melody and ornamentation

By early medieval times, both Shingon and Tendai shōmyō had developed a complex pitch and mode theory, largely shared with gagaku. Although there was an awareness of the manifold possibilities of combining the 12 tones of the chromatic scale with the various modal types (ryo, chūkyoku and ritsu 12 ), both sects limited themselves to a much smaller range of modes. Shingon shōmyō has traditionally used only five: ichikotsuchō (ryo on D), hyōjō (ritsu on E), sōjō (ryo on G), ōshikichō (chūkyoku on A) and banshikichō (ritsu on B). Tendai shōmyō is more complex, employing a total of six or seven finals, the five of Shingon shōmyō in addition to shimomu

Table 3.2   Structure of Kongokai-date Nika Hoyo-tsuki Chukyoku Rishu Zanmai as for a memorial ceremony for Kukai



Type of piece

Concrete realization (name of piece or group of pieces)

Language of text (see §3.3)




tengu-san [‘hymn of offering’]

Shichi no Bongo


shikishū: solo + group

address to Kūkai

saimon [‘festival texf]

Kōbō Daishi Shō-Mieku Saimon

‘hard’ Japanese

shikishū solo

purification and adornment of place of worship

nika hōyō [‘ritual in two parts’; Bai and Sange with Taiyō]



shikishū solo



shikishū: 1 or 2 soloists + group

+ Taiyō


shikishū: 1 or 2

soloists + group


explanation of purpose of ceremony

hyōbyaku [‘expression’]

Mieku Hyōbyaku

‘hard’ Japanese

dōshi solo

invocation and intercession

jinbun [‘gods’ parf]

Mieku Jinbun

‘hard’ Japanese

dōshi solo


butsumyō [‘names of the Buddhas’]



dōshi solo


kyōke [‘song of praise’]

Mieku Kyōke

‘soft Japanese

dōshi solo


zen-shōrei [‘opening invocation’]

Kongōkai Shōrei

Chinese with some Skt.

dōshi solo + shikishū group


zen-san [‘opening hymn’]

Shinryaku no Bongo


shikishū: solo + group



kyō [‘sūtra’]

Rishu-kyō (in chūkyoku modal style)


shikishū: solo + group


shuhō [esoteric ritual]

Rishu Kyōbō




go-san [‘closing hymn’]

Fudō no San


shikishū: solo + group


go-shōrei [‘dosing invocation’]

Kongōkai Shōrei (continued)


dōshi solo + shikishū group



ekō [‘transfer of merit’]

Shishin Ekō


shikishū: solo + group

(F#) and (exceptionally) kamimu (C#). Modes of all three types are said to exist on all finals.

The phenomenon of modulation from one mode to another, either within a piece or between pieces as a consequence of structuring in a particular rite, was explained theoretically from the early thirteenth century. Tendai shōmyō stipulates modulations of three types, with the final shifting a minor third (for example ryo ichikotsuchōritsu banshikichō), a major second (for example ryo ichikotsuchōritsu hyōjō) or a perfect fourth (for example ryo ichikotsuchōritsu ōshikichō). Shingon practice makes use of only the first of these, that is, a relationship that can be understood in Western terms as a modulation between a major key and its relative minor.

The range of one octave is referred to with the term (layer), and three of these, shojū, nijū and sanjū (‘first, second and third layers’ from lowest to highest), are joined together in most shōmyō pieces to make a theoretical range of three octaves. In practice, however, most pieces limit themselves to two octaves or less, around the central nijū octave. In the narrative genre kōshiki, however, the three-layer system was originally built up of overlapping rather than conjunct octave ranges, so that the shojū and sanjū layers were an octave apart, with the intermediate nijū layer overlapping with both. Changes in performance practice obscure this system today.

The greater part of the shōmyō repertoire is made up of pieces with Chinese texts originating in China, in addition to pieces with Sanskrit texts that were transmitted via China, and hence passed through the filter of the Chinese language. As a consequence, it seems likely that the melodies of imported pieces (at least) would reflect the Chinese tone-accent of their texts. While this assumption is often repeated in Japanese scholarship on shōmyō, there is little research that goes towards demonstrating it in concrete terms. In terms of melodic ornamentation, the tones that make up each mode are thought to have distinctive melodic characteristics, so that certain figures occur only on certain degrees. For example, it is only kyū, the final of a mode, and chi, the fifth, that are ornamented with the common figure yuri – essentially movement away from a tone and back to it, either upper or lower, sometimes fast and sometimes slow.

The pitch, mode and melodic theory of shōmyō are complex matters. Things are complicated even further by the fact that changes in performance practice during centuries of largely oral transmission have produced an enormous gap between theory and practice. In terms of its concern for correctly expressing pitch, Tendai shōmyō displays a vivid contrast with that of Shingon, where the great majority of pieces are sung in a way that has little to do with traditional stipulations regarding pitch and mode. Conversely, in the performance of the narrative genre kōshiki, Shingon Buzan-branch practice preserves a version much closer to the old form (as preserved in medieval notations) than the apparently reconstructed Tendai practice.

3.3  Language of the texts and musical styles of shōmyō

The long history of Buddhist chant is reflected in the variety of language to be found in the texts of shōmyō, and even within single rites. Four major language types can be identified.

  • Sanskrit: Texts with Sanskrit origins are pronounced in a form approximating the original pronunciation, passed through several filters, that is, with the Japanese pronunciations of Chinese characters that were once phonetic equivalents of the syllables of the original Sanskrit text. Many of the esoteric Sanskrit san (hymns) make use of the verse form śloka, the most important epic metre used in classic Sanskrit works, comprising two lines, each of two parts (pādas) with eight syllables each. Other Sanskrit texts are in prose, many in the form of mantras. In either case, the texts retain a certain symbolic power in Sanskrit that would be lost if translated into the Japanese vernacular. Communication of verbal meaning is not important, and there is hence a strong tendency for their vocal melodies (especially those of the san) to be melismatic.
  • Chinese: Overwhelmingly numerous, this class includes texts from both China and Japan. Again, two forms can be observed: texts in verse, often in rhyming couplets, and texts in prose. The Chinese equivalent of the Sanskrit śloka is a pair of couplets; typical examples are the esoteric Chinese san. Prose texts include the greater part of the texts of the sutras. The melodic realization of texts in Chinese covers the full spectrum from highly melismatic pieces in free rhythm to syllabic chanting on a single tone. The usually important distinction between long and short vowels is blurred in the reading of the sutras, since each character (rather than each syllable) is given the value of one beat.
  • ‘hard’ Japanese: This refers to Japanese in kanbun kundoku style, that is, texts either originally Chinese or drafted in Japan in Chinese (that is, in Chinese word order and made up exclusively of Chinese characters) but performed in Japanese word order with the necessary grammatical inflections added. While this technique can be used to read almost anything written in Chinese, even the sutras, it is generally limited to classes of explanatory or narrative prose texts sung solo. Major genres of shōmyō that employ this highly literary style include: hyōbyaku (also hyōhyaku/hyōhaku), texts explaining the purpose of a ceremony to the principal image of worship; jinbun, texts inviting the celestial multitude be present; saimon, texts addressed more clearly to those attending a ceremony; rongi, extensive catechism-like texts in question and answer form dealing with doctrinal matters; and kōshiki, narrative sermon-lectures in multiple dan (sections) that praise various figures in the Buddhist cosmos. The need for these texts to communicate their contents means that they are performed in syllabic settings with unobtrusive melodic ornamentation, often at the ends of phrases.
  • ‘Soft’ Japanese: Texts were apparently composed in the Japanese vernacular from the early years of shōmyō in Japan. Most are in the classical Japanese of the heian period, often in the alternating seven- and five-syllable lines of Japanese poetry. Genres employing this (originally readily comprehensible) language include: sandan, short songs of praise used in exoteric rites; wasan, more extensive songs, often epic in nature; and kyōke, short songs of praise in fixed textual and musical structures of four lines, often used in esoteric rites. The relatively simple poetic content of most of these texts means that their musical settings are often melismatic in part or throughout.
The following musical examples juxtapose two contrasting language types and musical styles: two excerpts from a melismatic free-rhythm piece in Sanskrit, and an excerpt from a syllabic narrative piece in ‘hard’ Japanese. Example 3.4a contrasts the first half of the Sanskrit san (hymn) Shichi no Bongo as performed in the Tendai Ōhara branch and the Shingon Buzan branch (☉ tracks 2 and 3). This is a san dedicated to Mahāvairocana of the diamond Realm, the cosmic Buddha of esoteric Buddhism and embodiment of the truth of the universe. The shichi or ‘four wisdoms’ of its title are those represented by the four directional Buddhas who manifest central aspects of Mahāvairocana of the diamond Realm, and who are given positions around Mahāvairocana in the diamond Realm Maṇ ḍ ala. The text of the piece praises these four wisdoms in a symbolic way by including the Sanskrit names of the first and foremost of the four Bodhisattvas who are expressions of the basic functions of each of the directional Buddhas. 13 The text takes the form of the mystic syllable oṃ followed by a śloka verse.

  Text of Shichi no Bongo

sanskrit original:

oṃ vajrasattva samgrahād vajraratnam anuttaraṃ vajradharma gāyanair vajrakarma karo bhava.

Tendai pronunciation:

om basara-satanba shigyaraka basara-ratanna madotaran basara-daruma kyayanai basara-geruma gyaro haba.

shingon pronunciation:

om banzara-satoba sogyaraka banzara-ratannou madotaran banzara-tarama kyayatai banzara-kyarama kyaro hanba.


Oṃ. Receiving the blessing of the wisdom of Vajrasattva, we obtain the ultimate vajraratna [diamond treasure]. By chanting the vajradharma [diamond law], we may attain the vajrakarma [diamond deed] of enlightenment.

Both versions begin with a solo singer, who is joined at some later point by other priests singing in unison. The Tendai Ōhara version given here is classed as an otsuyō ryo piece in ōshikichō, that is in a ryo mode on A, beginning from the fifth of the mode (otsu referring to the fifth, as opposed to the term , which refers to the final). The Shingon Buzan version is traditionally classed as a kōyō ryo piece in ichikotsuchō, that is in a ryo mode on d beginning from the final of the mode. Since it includes modulation to banshikichō (ritsu on B), however, its traditional notation

First half of

Example 3.4a   First half of Shichi no Bongo as performed in the Tendai Ōhara branch (adapted from Nakayama 1963)

First half of

Example 3.4b   First half of Shichi no Bongo as performed in the Shingon Buzan branch (transcribed by Kondō Shizuno and Uchida Atsushi)

has, since the medieval period, been written throughout in ritsu banshikichō, so that the notation of the piece commences not on the final but on the second (actually augmented second, namely d) of banshikichō. In performance, however, the pitch level is approximately a fifth higher than this. It can be seen from both examples that many short melodic fragments and combinations of them (senritsukei, ‘melodic patterns’) are given specific names in modern performance practice. These names are nouns or verbs reflecting either the movement of the vocal melody or the appearance of the traditional notation, and appear to be a comparatively recent development.

In vivid contrast is the syllabic narrative piece in ‘hard’ Japanese, Nehan Kōshiki. This is the first of the four kōshiki that make up the Shiza Kōshiki completed by Kōben (Myōe Shōnin) in 1215 (see Chapter 2, §2.3) (☉ track 4). A full transcription is unnecessary, since the musical realization of the text is built up from comparatively simple melodic formulas known as kyokusetsu, with structures shown in Example 3.5. Narration in each kyokusetsu is carried out using two central tones separated by either a tone or a fourth; this reflects the high-low pitch accent of the Japanese language. In addition, each kyokusetsu has a falling cadential pattern that facilitates the link to the kyokusetsu that follows. In the modern performance practice of the Buzan branch, the kyokusetsu move in a circular fashion, from shojū to kō-nijū and through the subsidiary kyokusetsu chūon either down to shojū or up to sanjū. From sanjū, the only possible movement is down to kō-nijū.

  Text and translation of excerpt from Nehan Kōshiki





nyorai mata moromoro no daishū ni tsugete notamawaku

The Buddha announced further to the gathered masses,

‘ware ima henshin hiiraki itamu

‘Now my body is racked with pain.

nehan toki itareri’

the time of nirvāṇa has come’.

kono kotoba o nashi owatte

As he finished these words,

jungyaku chōesshite

he entered the various states of meditation

moromoro no zenjō ni iru

in a way that differed from normal men.

zenjō yori tachi owatte

awakening from these states,

daishū no tame ni myōhō o toku iwayuru

he pronounced a great teaching for the assembled masses.


‘mumyō honzai

‘although we may be unaware of the true nature of existence,

shōhon gedatn gakon anjū

we all possess the seed of enlightenment.

Now I attain great peace,

jōjaku mekkō myō dainehan’ to

eternal tranquillity and the extinction of light. This is called mahā-varinirvāṇa, the great extinction.’

daishū ni shimeshi owatte

Finishing his words to the assembled masses

henjin yōyaku katabuki

he leaned over,

ukyō ni shite fusu

and lay on his right side,

kōbe hoppō o makura to shi

his head to the north,

miashi wa nanbō o sasu

his feet to the south,

omote o saihō ni mukae

facing the west,

ushiro tōbō o somukeri

with his back to the east.


sunawachi daishi zenjō ni itte

he entered the fourth stage of meditation,

dainehan ni kishi tamainu

and finally the state of great extinction.


shōren no manako tojite

he closed his eyes of lotus blue,

nagō jihi no mishō o yame

and lost forever his smile of compassion.

tanka no kuchibiru

his lips, red like the fruit of


the bimbā tree, were shut,

tsui ni daibon no

and his beautiful voice was

aisei o tachiki

never to be heard again.

3.4  Shōmyō notation

The vocal melodies of shōmyō have long been written down using various systems of neumatic notation, generally known as hakase (-bakase). Straight and curved lines, and combinations of them, are used to indicate pitch according to a number of different principles. An enormous difficulty faced by anyone wishing to survey these systems comprehensively is that the traditional names tend to be valid only in sectarian terms, and that the same name can imply quite different things when used by scholars affiliated with different sectarian groups. Trans-sectarian studies have been undertaken in recent years by the scholar and Shingon-sect priest Arai Kōjun, 14 but it must be noted that his ideas have not yet been accepted by scholars with other affiliations.

Rather than attempting to deal with the complexities of historical and current notational practice, here I will simply propose in outline a systematic method for classifying traditional notation according to the way in which pitch is represented, providing new english labels but also mentioning the traditional nomenclature.

  • ‘Approximate’ notation: This is notation that lacks any systematic way of indicating pitch. The earliest hakase, which date from the early eleventh century, are short lines that emanate from dots added at points around the characters of the text to indicate tone-accent. The lines appear to represent the up-and-down movement of the vocal melody in a simple visual manner.

    Example 3.5   Kyokusetsu pitch structure and melodic movement according to Shingon Buzan branch practice. Pitches shown are relative

    Similar notations can be found in almost all subsequent ages. Relatively modern developments include styles of notation that represent the movement of short melodic patterns in a similar visual style. These became necessary with the increasing degree of ornamentation that came to be added as the various traditions were passed down orally. A common name for this ‘approximate’ type of notation is tada-bakase (simple hakase), while the antiquity of this style results in it being called ko-hakase (old hakase) in some schools. The supplementary nature of more modern developments in a similar vein is reflected in such names as kari-bakase (provisional hakase) and meyasu-bakase (tentative hakase). It should also be noted that most vocal notation for the vocal genres of gagaku belongs to this ‘approximate’ type.
  • ‘graph’ notation: In this style, the opening pitch for each character of the text is specified by the position and the angle from which the notational line leaves the character to which it applies. A model hakase graph is often given as the key to reading the notation, and in some schools multiple graphs, each suitable for the melodic movement of a particular piece or group of pieces, were used. This style appeared in both Tendai and Shingon shōmyō from the late twelfth century, and has traditionally been called zu-hakase (graph hakase). Although the initial pitch for each character can be indicated logically this way, melodic movement in melismatic phrases can only be indicated in the ‘approximate’ way. It comes as no surprise that both ‘approximate’ and ‘graph’ notations have often used supplementary signs to indicate pitch, such as the names of the five tones, or the tablature signs of the gagaku flute or koto.
  • ‘Five-tone’ notation: ‘graph’ notation was eventually improved in both Tendai and Shingon shōmyō, by means of linking pitch (as expressed by the names of the five tones) directly to the direction and angle of the notational line, so that the pitch of even long melismatic passages could be read automatically. The most successful notation in this style, called goin-bakase (five-tone hakase), was that developed by Kakui of the Nanzan-shin branch of Shingon shōmyō in the late thirteenth century. In this system, the lines indicating the five tones are arranged in clockwise order from low to high pitch around the model character, with the directions of the lines of adjacent tones differing from each other by 45-degree angles (see Malm and Hughes 2001b). It is used today in shōmyō branches of the Nanzan-shin lineage (including the Buzan branch). There it is usually supplemented with some sort of ‘approximate’ notation that better indicates the melodic ornamentation of the later tradition.
Figures 3.5 and 3.6 show the opening of traditional notation for the two versions of Shichi no Bongo given in example 3.4a. The Tendai Ōhara notation is of the ‘approximate’ type (meyasu-bakase), with supplementary signs indicating pitch (shorthand signs for the names of the five tones), the abbreviated names of melodic patterns, and pronunciation. Note that the hakase lines for the fourth and fifth characters depart from different positions, despite the fact that the pitch is the same. The Shingon Buzan-branch notation is made up of two columns of hakase, the older ‘five-tone’ notation developed by Kakui in thicker lines and newer ‘approximate’ notation (kari-bakase) in thinner lines. The latter is annotated generously with notes

Traditional notation for opening of Tendai Ōhara branch

Figure 3.5   Traditional notation for opening of Tendai Ōhara branch Shichi no Bongo (from Nakayama 1963)

on interpretation. Supplementary signs provide some details of pronunciation and the abbreviated names of melodic patterns.

3.5  Issues in shōmyō research

The greatest issue facing the future of research on shōmyō is how to overcome sectarian boundaries. It is unlikely that a truly comprehensive study of the subject can be written until something is done about the contradictions in terminology that plague most research published by performer-scholars on their own traditions. Moreover, there is a great unevenness in research on ceremonies and ritual forms. Some ceremonies, such as the Shunie of Tōdaiji, have been studied exhaustively, while others remain largely unexplored. Another facet of research still in the rudimentary stage is the musical analysis of shōmyō in current performance practice. Consideration of its historical change should lead in the same direction as research on gagaku has in recent years: an increasing degree of ornamentation over the centuries with the concurrent development of short melodic patterns, the slowing of tempo, and changes in modal form with the flattening of tones other than the final, fourth and fifth.

Traditional notation for opening of Shingon Buzan branch

Figure 3.6   Traditional notation for opening of Shingon Buzan branch Shichi no Bongo (from Shingon-shū 1998)

As is the case with gagaku, there is a great abundance of primary source material waiting to be studied. Collections of typographical reprints of theoretical sources have been published in recent years, but there is still much to be done. Very few facsimile reproductions of music notation have been published, and those that have often reduce the photographs so much as to frustrate attempts to examine the notation in any detail. One imagines that the abundance of old examples of notation would facilitate historical and comparative studies of musical content, but little has been attempted, perhaps because so few examples are readily available for examination, especially to ‘outsiders’ who are not affiliated with the particular sect or school in question. In many cases, scholarly access is not viewed as a priority, and it is not a simple task to open new doors in what is undoubtedly a conservative segment of modern Japanese society.


For Japanese-language commentary on watashimono, see Masumoto 1968: 248–57, Masumoto 2000: 208–13.

In the performance of Goshōraku no Kyū recorded on Gagaku: Kunaichō Gakubu/Gagaku: Music Department; The Imperial Housed (sic; mistake for household), an opening tempo of ♩ = 29 reaches = 45 by the end of the piece.

Not all kangen pieces ♩ are performed as bugaku: many pieces exist only in the kangen repertoire, some only in bugaku, and some in both. Also, the metre of some pieces changes with the change in repertoire. A small number of pieces are performed in tadabyōshi rhythm in kangen, with alternating bars of 2/4 and 4/4, and also in this metre, but more quickly, if danced as ‘Left’ (tōgaku) dances. When they are performed exceptionally as ‘Right’ (komagaku) dances, however, the final beat of the 4/4 bar is omitted, to produce the metre known as yatara-byōshi (alternating bars of 2/4 and 3/4).

In a 1999 Tokyo performance of Goshōraku in bugaku form by the group Jūnion-kai, the tempo increased from = 20 to final movement. ♩ = 56 in the space of the 12-minute performance of the final movement.

Clear expositions of the argument can be found in Marett 1985, 1986, 2001a: 859–61, and 2006, while the results of the research have been published in the series Music from the Tang Court.

See, for instance, Fukushima Kazuo’s review (1984).

The convoluted style of Picken’s prose, difficult even for native speakers, may be another reason why his ideas have met with less appreciation in Japan than they might.

This particular rite has been chosen whenever concrete examples are needed here for two reasons: 1) An abbreviated version is available on video/laser disc, as vol. 4 (VTMV-104) in the series Nihon Koten Geinō Taikei; 2) there is substantial english commentary on it in Nelson 1998.

Note that this hōyō is often written 法用, to distinguish it from the more general word for ‘rite’, which is written 法要.

The Maṇ ḍ alas are, respectively, the Kongōkai mandara (Skt. Vajra-dhātu Maṇ ḍ ala) and Taizōkai mandara (Skt. Garbha-dhātu Maṇ ḍ ala). They give concrete representations of the concepts expounded in the Mahāvairocana Sutra (Jap. Dainichi-kyō) and the Vajra-śekhara Sutra (Jap. Kongōchō-gyō).

The structure shown in Table 3.2 is an abbreviated version of the ceremony; Nelson 1998 gives details of contents and music of the full version.

See Table 1.4, for details of these types.

The conceptual background of the piece requires an understanding of the structure of the diamond Realm Maṇ ḍ ala; see Nelson 1998.

See Arai 1995, 1996 and 1999 for english translations of his historical studies on the development of hakase notation.

Search for more...
Back to top

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.