The voice of the puppet

: The puppet as performing object invites vocalization that is different from the normal human voice. Choices to suit body size, stylized features, and the animated objects tendency toward extra-daily sound—song and music—reinforce the otherness of the performing object. Three major types of vocal concepts are presented: 1) the one person, many voices show, which can use different resonators, pitches, and speeds as exemplified in the wayang golek of West Java; 2) the two-voice show, where the puppeteer may speak with a voice modifier such as a swazzle and the dialogue interpreted by a more normative human voice, in traditional puppetry it will often be a musician or interlocutor/back chat who sits outside the booth; and 3) the multi-voice show where individual speakers are often used for each figure, while dialogue may pre-recorded, delivered from the side by appropriate voice actors, or by each of the individual manipulators who speak for their own figure.


Revista de Estudos sobre Teatro de Formas Animadas
Much of the interest of puppets is they are like us, but not. That "but not" also feeds into their vocal/sound stylization. The voice of the puppet typically is not a normative human voice (though, of course, there is a human behind). Since figures are generally larger or smaller than humans and their features, often exaggerated, stylization in vocalization is the norm. Small puppet are often squeaky, fast, and abrupt in vocal characteristic (think of Punch or his many cousins). Large figures for street parades or large stages may have deeper resonances and booming sounds if they talk at all. (To make them perhaps more awesome the figure may be silent with the music and/or percussion as a sound replacement-think of the dances of shi-shi (lions) in East Asian cultures or the Balinese barong whose manipulator may add rhythms of the figure's clacking the jaw to create interlocking percussion with the drums and gongs.
Vocal eccentricity is a plus in puppetry and performers capable of distinctive voices are usually valued. Consider contemporary animation films, where popular actors with unusual voices are routinely cast: puppets and animation visuals invite distinctive vocal choices. Puppets, due to otherness, also slip comfortably into singing-sometimes their talking is just a prelude or interlude in what is largely a sung/danced presentation with song text that may range from the doggerel of the clowns to the lofty poetry of the noble figures. Meanwhile contemplative songs may sound behind a screen ornament displayed in the Turkish karagoz. Sometimes a singer or chanter delivers the words as narrative, while a figure is just displayed. One can argue that the song/music/stylized poetry here is about distancing, since it may be in archaic language-as with the Java's Kawi (old Javanese/puppetry language)-a language only dimly grasped by audiences and sometimes obscure to the puppeteers themselves. Music, chant, poetry, distant language choices becomes a surrogate "voice" that at least in traditional theatres, helps establish puppetry as the link to ancestors and other worlds, where spirits or the dead are an implied audience. In sum, puppets are "other," and sound suits that difference.

MÓIN-MÓIN
The strategies for voicing the figures are, of course, multiple. Here I will outline some of the major variations: 1) the one manmultiple voices show; 2) the two-person show with the speaking puppet(s) and a backchat or interpreter; and 3) multi-person shows where the individual performers may voice the figures they manipulate, or speech is given by voice actors (prerecorded or live) while manipulators dance the figure.
The one man-multiple voices show Certain forms like wayang theatre of Indonesia, the bunraku of Japan, the tolpavakuttu and other narrative forms of India rely on a single voice, the dalang, the tayu, and the natuvar respectively. In each case we can see some relation to older traditions of narrative recitations (perhaps of story scrolls: wayang beber in Indonesia, precious scrolls in China/Japan, phad/pad in India). These have been expanded into theatre by having puppets act out stories that originally were just chanted or correlated to a painting of the scenes and using a pointer to direct the viewer to the place where the particular event is displayed. Hence, singing narrators would deliver tales, which were later reinterpreted in puppet show action. Though there is not space here to discuss these traditions fully, it is clear that scroll narration was important in many religious traditions of Asia and the Middle East. We often see them associated with wandering Buddhist monks, Hindu singers, Sufi or Shiite singers in the Middle East to South Asia Monk performers in China might use a "hell scroll" to scare listeners into good behavior, other scrolls might tell stories that release souls from hell as with the Bhima Suwaga (Bhima Goes to Heaven) of Bali where a Mahabharata hero saves the soul of his parents.
Characteristic of these storytelling genres is the demand for a multi-talented artist who can breathe emotion into a wide range of characters as well as present the narration only portions that tie the dialogue sections together and sung passages that ornament the form. Even in the secular world of today, such scroll narrators

Revista de Estudos sobre Teatro de Formas Animadas
may have an aura that comes via their command of important narratives with spiritual content of the past (i.e. wayang beber in Java). In the Indian tolpavakuttu and Japanese bunraku, the narrator does not manipulate the figures, which allows them to focus on the vocal demands and singing. Knowledge of the text sometimes in a semi-archaic language (17-18th century plays of bunraku or the 13-century text of the Tamil poet-saint Kamban) of tolpavakuttu sung for a Malayali speaking audience) is important.
By contrast in wayang the puppeteer manipulates all the figures as well as doing all the vocal work, and, probably as a result, we find that the dialogue passages are relatively static, with the dalang mostly moving the hand of the puppet for which he currently speaks. Dance and fighting manipulation comes when the singers in the gamelan orchestra supply the sound. By contrast bunraku with its three non-speaking manipulators per figure or tolpavakuttu, which may have one manipulator to a figure have more room to move while the narrator continues his vocal presentation (often sung or chanted). Clowning in both Indian and Indonesian arts may be a space where more normative contemporary language and slang enters, but the clown vocals are often distorted: they stutter, growl, and bluster as they speak. Clowns are less tied to the elaborate language for the high characters intoned by the dalang, tayu, or natuvar.
Here I will discuss the wayang golek of West Java as a specific example of a one-man show.
The voices, in my analysis, are generally four and are conceptualized as located in different parts of the body. The voice of the refined character is located in the back of the mouth with the jaw stretched a bit to give a feeling of roundness and greater resonance. This applies to both male and females of the refined type (white face, downward glance, small eyes and bodies), but the female voice may move more frequently to a higher note (1) on the five tone gamelan scale from the customary deep tone of (4) with this type's slow and melodic diction. Sometimes the puppeteer uses

MÓIN-MÓIN
the tuning phrase masehman, which has no actual meaning but allows for quick vocal placement before launching into a speech for the refined (lenyapan) characters. The measured pace and low pitches indicate a person who thinks before he/she acts and avoids conflict until it is necessary for defense.
The second type is the semi-refined (lenyap or ladak) character again male or female (with white or pink faces, small bodies and eyes, but a higher gaze). The voice hoovers with a musical lilt around note 1. The dental resonator at the top of the mouth is the vocal placement for this character and the tuning phrase is ke-ke-ke. These characters speak higher and faster than the refined figures, giving the feeling of rush. They are less considered in their actions and often put themselves at risk by hurrying into a dangerous situation. The voice moves as swiftly as the character, whose actions often require a refined character to save them when they are kidnapped or take on opponents that are beyond their strength.
The third voice is that of a punggawa (warrior) with dark colored skin, a larger body and staring eyes, and forthright gaze. The pitch that is home base for this character is, as with the refined character, the low note 4. Tensing the vocal cords and using a chest resonator produce a deep gutteral tone. The tuning sound is gr-grgrum, a growl deep in the throat. These characters are the workerbees of the wayang world who carry out the missions assigned by the refined king. An example is burly Bhima, the second and strongest of the five Pandawa heroes in the Mahabharata or Hanuman, the monkey general, who assists the noble Rama in the Ramayana. For Hanuman as a monkey the dalang sometimes adds the squeaks and blusters created by placing a piece of grass in the mouth to use as a temporary reed, or he may puff out lips and make animal-like sounds by sharp exhalation of breath.
The final major voice is of the big, red -faced, bug-eyed, upward glancing, emotionally uncontrolled king that attacks without thought and seizes with greed. This voice slides up and down between the high note 1 and low note 4, with a tendency to

Revista de Estudos sobre Teatro de Formas Animadas
keep on the higher note but conclude on the lower. These kings also are distinctive in that they burst into a chant-like delivery in a pelog scale (which contrasts with all the other character who speak in normative selendro tuning). This gives the feeling that this is a character truly from another world. The normative actions of this type -kidnapping women and laying waste to kingdoms -confirm the alien status. In wayang cepak the puppet is often seen is called Kelana (literally "Wanderer), a misguided king from an overseas kingdom. Characters like Rawana the king of Langka who kidnaps Lady Sita in the Ramayana are of this category.
There are some characters that are felt to have voices specific to them, but even these figures often are covertly linked to the four types mentioned above. However though the vocal placement and note are the same, timing and additional vocal distortions are added. For example Semar the wise god-clown, who accompanies the white-faced refined hero, actually speaks in a voice, which is a variation of the refined character, but he draws out his words much more in a way that adds humor as he complains about his misguided sons or the sorry state of the universe. His third son Dawala does a permutation of the rather nasal sounding semi-refined character, while his eldest son Cepot does a even more macho version of the warrior voice. Lower class ogres often use a version of the warriors voice but with more vocal exaggeration, stuttering, mumbling, and shouts. A few of the comic ogres may evoke hilarity with their high-pitched squealing that borrows from the semi-refined voice.
The only character who does not use the notes 1 or 4 as his homing note is the god Narada whose dialogue is chanted and whose base pitch is 3. He is constantly being sent to the world by the refined high god of the universe, Batara Guru (Siwa/Shiva), to carry out some mission. His eccentric tonal choice may indicate his otherness in the narrative, which is a world where heroic humans and specious demons contend.
The dalang has a narrative voice that is generally hovering around tone 4 of the gamelan scale and that he uses for description and omniscient views regarding the action. He also sings mood songs to express the emotion of a scene (sadness, love, transforming from one body to another, etc.). For clowns and ogres he may also sing in that character's eccentric voice.
The evolution of the system of placements, tuning notes, and use of different resonators (mouth for refined, nose for semi refined, throat and chest for minister, top of the head for the emotionally uncontrolled king) surely comes from the fact that this is a solo performance genre. With additional slight modification of speed or nuance, the dalang can differentiate two characters of the same type that appear in the same scene. This system of placement, note, and speed allows the dalang to distinguish a wealth of characters (25-40 characters might appear in a single play).
Minor warriors or ogre types may make up a third of the voices and they will have little to say. The refined hero and the emotionally uncontrolled antagonist, a hard-fighting warrior, and a semi-refined heroine may have the bulk of the dialogue. This allows the audience to be clear about who is speaking, even when they are hearing to a show in a radio broadcast.
The choices of wayang do not fully apply to Japanese gidayu naration by the tayu whose vocal techniques may be related to early Buddhist chant traditions or the full-voiced singing of the Ramayana of Kamban by the Indian natuvar. Both the Japanese and Indian works are more through-composed and text-based narrative performances than wayang, where most of the ideas are delivered via the dialogue of the figures, improvised in performance. Still when these techniques delivered by one artist use dialogue they make use of the same tools-the voice tapping potential resonators and stylization by gender and character type are evident. The unitary body becomes an instrument the puppet mater plays to represent a cosmic whole: Gods, demons, spirits, animals, men and women (good, bad, and ugly) all come from one speaker who represents in himself the universe.

The two-voice show (puppeteer and interpreter)
The two-voice show generally works with the interpreter/back chat on the outside and the puppeteer inside the booth. Even in early television of the USA this very old model of puppetry was manifest in shows like Howdy Doody (with Buffalo Bill as the interpreter) and Kukla, Fran and Ollie (with Fran Brill as singerinterlocutor).
This is a pairing of puppet voice and a natural human voice is of course found in many traditional puppet theatres. The puppet speaks with all his/her/its eccentricity and the live person interprets, serving as a bridge between the two worlds. The similarity of this pattern to shamanic complexes, where medium while in trance channels the voice of the other (the dead, a heavenly goddess, demon, etc.) signaled by a change of voice, and the shaman sits outside and controls the event, interpreting it for spectators is sometimes seen as a potential antecedent to the puppet with his musician/interpreter/back chat. The dummy-manipulator model of ventriloquism, while united in one performer, still maintains the two-voices model (see Stockman, 2012). While puppetry usually requires two different performers (often hiding the puppeteer) and vent is only one person, these forms are related in that one voice is normative (the "real" or everyday voice of a performer) while the second is the altered voice (the other). This puppet-interpreter model was historically a part of low class, itinerant performances in many world regions.
Of course, we do not strictly need a shaman-medium model to explain the division of labor. The backchat outside the booth makes practical sense -for gathering an audience, providing accompaniment, passing the hat mid-show, and adding commentary to boost the performance. The interpreter is especially important in the two-voice show that uses a swazzle, making sure than none of the audience would miss any important content. Reed instruments (of which the swazzle is one) were and remain important in outside and parade presentations of puppetry and other forms. They catch the ear from afar as the puppeteers arrive. The sound (along with drum or other musical instruments often played by the interpreter) draws viewers to the performance. The reed's squeal seems to emanate from another dimension-it is human breath and word transformed: we can even hear the phrases, but they are only partial clear and seem as emanating from another dimension. The dialogue will often be full of jokes and misunderstandings, but the wordless screams as figures fight or die are only too clear.
In many instances, the backchat spends much of the time simply restating the puppets' dialogue (though sometimes in more polite language). In other instances the puppet's distortions might make the interpreter mishear language, allowing the puppet puns and inappropriate language, for which the backchat will reprimand the figure, highlighting the comedy. The swazzle may also be to signal music cues that the audience will not catch, providing a sort of secret communication between puppeteer and backchat. The interpreter often plays his/her own musical instrument, adding sound support for action or creating the ambience of a scene. The puppets' squeals draw attention, attracting passing crowds, but the more natural voice of the interpreter/musician who sits outside the booth helps keep the crowd attached and will end a scene when flagging interest in the audience is apparent.
The spread of the swazzle is wide. This reed instrument that normally sits on the performers tongue is largely used by low class performers who share what I have called "big bang" arts (done outside often with processions that lead to the booth, using loud music/sound with minimal players to gather audience, comprised of relatively direct short skits that the audience can understand at almost any moment they arrive, etc.). These shows contrast with what I have termed the "courtly consort" (presented in a pavilion or on specially built stage where people come specifically to see the show, often having multiple musician for accompaniment, using relatively complex plotline with viewers often expected to stay for longer periods to follow or even all night). Courtly consort genres

Revista de Estudos sobre Teatro de Formas Animadas
do not use a swazzle, but the big bang arts often do.
Probably because of the itinerant nature, big bang troupes travel light-the one man booth, a single musical instrument for accompaniment, and simple and light glove or rod puppets or 1-2 string marionettes are the norm. Forms like the Korean gogdu gaksi had a musician on the hourglass drum interpreting with the main puppet Pak Cheomji in a sequence of short skits and historically used a reed for the puppet. Rajasthani kathpuli had the male puppeteer with his booli (swazzle) for the puppets' voices while his wife played the dhol drum and interpreted during a series of short acts (drummer, snake charmer, dancer, etc.) (see Jairazbhoy, 2007). Persian kemeh shab bazi had the puppeteer with his safir (swazzle) in the booth performing the simple string figures with the blackfaced clown Mobarak featured: the morshed backchat outside the booth played his spike fiddle (sometimes with a drummer as well).
In Egyptian Kargoz glove puppetry the swazzle was called amana; the Italian Pulchinella had his pivetta, the Russian Petrushka, his pishtchik; the Turkish karagoz, his nareke, Punch who arrvied in England in the 17th Century often speaks through the swazzle to the present. Mali puppets use a kind of kazoo and Cameroon, Gabon, Equatorial Guinea, Niger, Nigeria, and the Democratic Republic of Congo all have voice modifiers (Prochan, 2012). In Bandung Indonesia on July 26, 2018, I saw a simulation of kuda kupang (horse trance dance) that began with a dukun (shaman) putting a dancer into real or simulated trance with drumming accompaniement-as soon as the dancer was in an altered state the body puppet of the lion he lay below reared up and the familiar sound of a reed voice modifier signaled his trasnsformation as he danced in agitated fashion, clapping his jaw. Marcel Violett (2012) notes voice accessories/modifiers are used whenever a voice needs to be deformed to fit a character's physique, personality or make him unique. In European traditions we know the swazzle was present in the 17th century and probably before and endures in a number of performances to the present.

MÓIN-MÓIN
Of course there are other methods of altering the voice and we find these more the norm in contemporary puppetry, where changing the register, employing a dialect or accent, mangling syntax for comic characters, and other means are used. So the disparity between normal language and that of the puppet can be manifest with or without the swazzle. Yet even when the puppeteer has no swazzle, the altered voice for the puppets and the puppeteer's or an interpreter's more natural voice for narration or comment provides the needed contrast between the figure's and the normative human plane. Whether it is a swazzle, an odd accent, a lowered pitch or heightened speed, the contrast of this "two-voices, one show" (puppeteer and interpretor/interlocutor) is likely to remain.

The multi-person show
Where companies support a large number of performers they may split the voice performers from the manipulators. Voice actors whose vocal quality fits the figure portrayed may in this iteration be used (or they may manipulate the puppet they voice). Women get the female voices (except for the shrew who will often be a man), grown men become villains or guides, youth get the lovers, and those with vocal quirks inhabit the clowns. This trend that approaches a kind of realism is found in marionette performances (often with pre-recorded sound tracks) where the manipulators of may just be the movers. In puppet operas, the soprano, tenor, and basso will normally sing for the heroine, hero, and villain, respectively. In contemporary claymation and animation, as mentioned previously, the voice actors are chosen for their ability to give the visual character the nuance desired. Note that each of these object-related genres (puppet opera, animation) work best with extra daily sounds. However these multi-person shows require considerable human resources and coordination. Rehearsal time multiplies and costs rises with many performers, hence a prerecorded sound track is often chosen in professional shows. Sound systems replace the vocal production expertise of the past. The squeal of the swazzle is no