SIBE - Sociedad de Etnomusicología
Explorar TRANS:
Por Número >
Por Artículo >
Por Autor >


Share |
Suscribir RSS Suscribir RSS Feed

Blog Observatorio de Prácticas Musicales Emergentes
ETNO Revista de música y cultura
IASPM - Espana
Musicalogía feminista
IASPM - International Association for the Study of Popular Music

< Volver

Language and Music in fusion: The Drum language of the Banda linda (Central African Republic)

Simha Arom


Key words: Banda-Linda (Central Africa Republic); lenga (wooden drums); tone language; drum messages.

El presente artículo trata de definir el uso de los tambores para la transmisión de mensajes lingüísticos entre los Banda-Linda (República Centroafricana), demostrando así la fusión e interacción que se da entre el lenguaje, que en el caso de los Banda-Linda es tonal, y la música llevada a cabo por los tambores.

El lenguaje de los tambores de los Banda-Linda se basa en el mismo esquema de reproducción de tres alturas tonales, así como en los mismos patrones rítmicos que tiene el lenguaje.

La novedad en la aproximación al fenómeno de Arom es que, a diferencia de sus predecesores, éste toma el punto de vista del receptor y no sólo el del emisor. Así realiza un análisis del fenómeno que le permite descubrir una anatomía de los mensajes.

Palabras clave: Banda-Linda (República Centroafricana); lenga (tambores de madera); lenguaje tonal; mensajes de tambor.

With the exception of vocal music —i.e. when music becomes the vehicle for words— the use of drums for the transmission of linguistic messages appears to be the only situation in which the fusion of language and music takes place. In Subsaharan Africa drum languages rely “on the emission of sound signals which are neither words nor substitutes for a graphic code” (Cloarc-Heiss 1997: 136). This mode of communication is widely promoted by the fact that a large number of languages happen to be tonal in nature. In a “tone language” each vowel is necessarily affected by a pitch : this means that any change in a vowel’s tone may lead to a modification in the meaning of the word containing that vowel. This in turn means that in such languages, tones —as well as vowels and consonants— play a distinctive function.

Drum languages in black Africa constitute semiotic systems whose rules generally concern both language and music. In fact, musical instruments are used in order to convey linguistic messages, and so are two major musical parameters : pitch and duration, which moreover operate at a pace that is identical with that of speech. This is a mode of communication which corresponds to a soliloquy, in the sense that the only aim of information thus disseminated is to move the addressee/s to engage in some form of action and never to elicit a response through the same medium.

The Banda-Linda

In terms of demography, the Banda are the main ethnic group in the Central African Republic. They live in the Eastern half of the C.A.R. and have a population of about 400,000 individuals divided into many sub-groups, each with its own dialect and its own music.

The present paper focuses exclusively on the Linda dialect, one of the most important linguistic varieties used by this group and one which is spoken by some 27,000 people living in the very heart of the Banda area.

The Linda language consists of 32 consonants and 8 vowels. It presents no phonological opposition between short and long vowels, nor does it use accents. But each vowel is affected by a tone. The language has three registers —High, Medium, Low— and two “modulated” tones realized through a quick gliding from the low tone into the high tone, and vice versa . [2]

The “logistics” of drumming

As happens with the other drum languages in the region, the drum language of the Banda-Linda is based on the schematic reproduction of the pertinent tonal pitches affecting the vowels in the language. These pitches are produced at a pace that is identical with that of human speech. Observation shows that there is a term-to-term correspondence between the three pitches produced by the drums and those of vowels in the same utterances in the natural language. Likewise, the comparison between a series of drummed-out messages and their decoding into the Banda-Linda language reveals a perfect match between their rhythmic articulation and that of the corresponding utterances in the spoken language. Such a close connection between speech and drumming bears proof that the drummer conceives his messages in the natural language while he is transmitting them. Therefore, the encoding of messages relies on the natural language which the drum language is but a calque of from the melodic and rhythmical standpoint.

The transmission of messages is done by means of two lenga of different sizes. The lenga are wooden drums. Carved on a single hard-wood block, the instrument consists of a barrel-shaped empty body that rests on four thick legs and has two protruding handles. The drum’s convex sides are unequal in thickness —which allows the production of two sounds with different pitches— and are separated on the upper side by a 5-centimeter long slit. The bigger of the two is called eyi-lenga —“mother-lenga”— and is about 1,20 metres long, while the smaller one is called ako-lenga (“husbandlenga”). Traditionally, each village owns a pair of drums and it is the chief of the village who keeps them.

In order to produce a message the two drums are placed in parallel. The performer may be sitting or standing. The low and medium tones are respectively played on either side of the eyi-lenga, while the high tone is played on one of the sides of the ako-lenga. Drumming is done by means of two sticks, the ends of which are covered with strips of wild rubber.

Master-drummer Michel Wanga, Ippy, June 1973


The length of the drummed-out messages is of approximately five minutes. Depending on the time of the day and the atmospheric conditions, these messages can reach a distance between 2 and 12 kilometres; in order to extend this range, messages are transmitted during the night, right before sunrise (the time which provides the most favourable conditions). The messages can be further relayed within an infinitely broader range through the use of “retransmitters”, but quite evidently the limits here are set by dialectal boundaries, which in this case allow for a maximum radius of 100 kilometres (Bambary-Ippy-Bria). As a general rule, messages are intended for local dissemination, so that the normal reach of plain, straightforward transmission is more than enough.

Functions of the drum messages

In theory, the messages that can be conveyed by drumming make up an open inventory. Félix Éboué (1933 : 80) is right when he points out that through the drum language the Banda may “express absolutely every idea that can be expressed through speech”. But it is not clear whether it is also possible to understand all those messages…
In fact, the use of the drum language —which above all fulfils a social function— has been limited to a set of situations or circumstances which always require that one or several people should move to the source of emission, and which have changed in the course of time. This is why it seems useful to draw a distinction between the following two categories: the purely traditional messages that were already used before colonization; and those others that have been gradually incorporated after the arrival of the Europeans.

To the first category belong the following events: births, deaths, the organization of assistance after a hunting accident; the investiture of a chief; the ending of a period of mourning; the consecration of new drums; or the transmission of instructions so that someone who has got lost may safely return to the village.

In the second category are included the various calls issued by the komanda (the administrative authority) to the village chiefs (màkonji) in order to arrange for the collection of the tax money; summon them to an official meeting; or inform the population of the arrival of some official person, the presence of a sanitary team, or of the setting up of the annual cotton market.

Beginning in the late nineteenth century, a great number of studies have been devoted to African drum languages (Cfr. References). What these studies have in common is that their authors explored these communication systems purely from the point of view of encoding —i.e., emission— while neglecting the none-the-less important aspect of decoding or reception. Thence many misinterpretations regarding the modes of these messages’ perception. Indeed, while it is true that drum languages based on pertinent pitches allow one to produce all possible utterances or successions of utterances in the language ─as is rightly claimed by several authors (among others Éboué: 1933)─, it is also worth pointing out that, because of the significant rate of tonal homophony shared by many words, the likelihood of a real understanding would be extremely low.

Research and Analysis

Our aim was not just to unearth the system underlying the drum language of the Banda-Linda, but also to ascertain the extent to which the addressees manage to make out the messages’ content. In this sense, our research had to involve, as well as a drummer-transmitter, a local person who would act as a decoder.

Each one of the messages making up our corpus has been collected in two successive versions: one, performed in the traditional —i.e., continuous— way; the other in discontinuous or alternate fashion, that is to say, where each drummed utterance was followed by a pause, used by the decoder –—ignorant of the content of the message— to “translate” it on the fly into Banda-Linda. The aim of this procedure was, on the one hand, to confirm the parallelism in melodic profiles between the tonal patterns of the drummed-out messages and those of the spoken language; and, on the other, to double-check the extent and pace of understanding by the potential addressee.

One needs to bear in mind that in the traditional environment, the decoding of drummed-out messages is not a job for the specialist: any Banda- Linda speaker is prepared to decipher them. In fact, the speed and accuracy with which the successive utterances have been decoded dramatically proves the effectiveness of this type of communication.

The structure of the messages is not rigid. The comparison, for each one of them, between the continuous and the alternate version shows the existence of variants that essentially concern the sequential order of utterances. However, and with rare exceptions, their linguistic content derives from verbal language itself : messages rely on the transmission of full utterances whose syntax corresponds to that of the spoken language.

Anatomy of a message

At first hearing, a drummed message sounds as follows: a long succession of quickly drummed beats on three fixed pitches that constitute melorhyhtmic segments separated by pauses, some of which are immediately identifiable as such because of their sporadic recurrence, which indicates a high rate of redundancy. The presence of a series of pauses along the syntagmatic chain allowed us to carry out a preliminary segmentation, which was further completed by a series of more detailed analyses. Just as the pauses enabled us to parse the message into units, so the phenomenon of recurrence brought to the surface the existence of several paradigmatic classes. By paradigm we understand here a series of terms —i.e., utterances or parts of utterances— whose central notion, as defined by the meaning —that we shall call keysemanteme—, constitutes a recurrent feature.

Such a definition allows us to distribute all the utterances in the corpus into paradigmatic classes. The analysis thus obtained shows the existence of two major categories :

- One includes the utterances that are common to all messages: formulae for summoning, inviting people to listen and authentication.

- The other brings together the several elements making up each message’s specific information.

Each one of these utterance types, regardless of whether it belongs to the first or the second category, constitutes a paradigm, or more accurately, a “paradigmatic theme”. For the sake of convenience, we shall call the several types included in the first category general formulae (GFs), while the term specific formulae (SFs) will be used for types belonging to the second category. To these two categories we should add fixed formulae used as terms of address —except when the message is addressed to a specific individual, for whom a codified formula is used (Cfr. infra)— and by way of conclusion.

The message announcing the investiture of a traditional chief will provide an example of what I have just described. Here we have two versions. The first one corresponds to the decoding of the transmitted message performed in normal conditions (continuous version). The second one corresponds to the transcription obtained on the basis of the alternate version.

The text’s segmentation, such as it appears here, follows the very segmentation of the message at the sound level. The units thus defined are numbered in order of appearance . [3]

Continuous version [4]

1. Hey ! Guests !

2. I am speaking to you about the feast.

3. I am speaking about a feast.

4. And I say :

5. The investiture will happen.

6. And I say :

7. Get together !

8. I am speaking to you about the drum feast.

9. And I say :

10. Don’t be fools.

11. Come to me quickly.

12. I am speaking to you about a feast.

13. I am speaking to you about Wayewo’s feast.

14. Come to me for Wayewo’s feast that I am speaking about.

15. Get together and come to me quickly !

16. Get together and come to me quickly !

17. I am speaking to you about the investiture feast.

18. I am speaking to you about the investiture feast.

19. Come to me for the rejoicing of the feast.

20. I am speaking to you about Wayewo’s investiture.

21. I am not mistaken.

22. Get together !

23. And come to me, quick, quick, quick !

24. I am speaking to you about the feast.

25. Get together and come to me quickly !

26. I am speaking to you about the investiture.

27. Get together and come to me quickly !

28. Come to me quickly. Come to me !

29. Get together you all !

30. And come to me for the feast of the investiture of Wayewo, from the Ngora lineage.

31. Get together and come to me quickly !

32. Get together and come to me quickly !

33. I am speaking the truth.

34. You are hesitant.

35. You are hesitant.

36. I am not mistaken.

37. Get together and come to me quickly.

38. Come to me quickly !

39. Come to me quickly !

40. I t is the feast of the investiture of Wayewo, from the Ngora lineage.

41. I am speaking to you about the feast of the investiture of Wayewo, from the Ngora lineage.

42. D id you understand?

43. I am speaking to you !

44. Get together and come to me quickly.

45. I am speaking to you about the rejoicing of the feast.

46. I t is the feast of the investiture of Wayewo, from the Ngora lineage.

47. I am speaking to you about the feast of the investiture of Wayewo, from the Ngora lineage.

48. Get together and come to me quickly, come to me.

49. I am not mistaken.

50. Come to me quickly !

51. Come to me quickly !

52. I am speaking to you about the feast of the investiture of Wayewo, from the Ngora lineage.

53. Get together and come quickly, quick, quick, quick !

Alternate version [5]

1. Hey ! Guests !

2. I am speaking to you because of the feast of the investiture.

3. I am speaking the truth.

4. I am speaking to you about the feast of the investiture of Wayewo, from the Ngora lineage.

5. A nd I say:

6. Get together and come to me for the feast of the investiture of Wayewo, from the Ngora lineage, that I am

speaking to you about.

7. I am not mistaken.

8. Did you understand well ?

9. Get together and come to me you all !

10. Come to me !

11. I am summoning you to the feast of the investiture of Wayewo, from the Ngora lineage.

12. Get together and come to me !

13. Get together and come to me !

14. I am speaking the truth.

15. I am not mistaken.

16. Get together and come to me quickly !

17. Hey ! Fathers !

18. Hey ! Mothers !

19. Get together and come to me !

20. I am summoning you to the feast of the investiture of Wayewo, from the Ngora lineage.

21. Get together and come to me for joy rejoicing !

22. I am speaking to you about the feast of the investiture.

23. Did you understand ?

24. Get together and come to me for the rejoicing !

25. I am speaking to you because of the feast of the investiture of Wayewo.

26. Get together and come to me !

27. Run to me !

28. Come to me !

29. Quick ! Quick ! Quick !

30. Come to me !

31. You all !

32. I am speaking the truth.

33. Did you understand?

34. Did you understand?

35. Don’t be fools !

36. Come to me !

37. I am speaking to you about the rejoicing of the feast.

38. I am summoning you to the investiture of Wayewo.

39. Come to me !

40. It is the feast of investiture.

41. Did you understand?

42. Get all together !

43. Get all together !

44. Come to me in a big crowd !

45. Come to me in a big crowd !

46. I am summoning you to the feast of investiture of Wayewo, from the Ngora lineage.

47. Don’t be fools !

48. Final formula.


A close look at the first few units in both versions ─continuous and alternate─ reveals the coexistence of the two above-mentioned categories:

- a more general one that may occur in any message, such as: “I am speaking the truth”

- and another one which carries the message’s more specific information. For instance : “I am speaking to you because of the feast of the investiture of Wayewo.”

In the continuous version we see 33 occurrences of GFs (general formulae) by contrast with 20 occurrences of SFs (specific formulae) ; whereas in the alternate version GFs occur 33 times and SFs only 14. However, while recurring utterances provide a significant feature, the number of distinct utterances is relatively low (about a dozen for the former type, and two or three for the latter). We realize that the specific information is always distributed along —somehow embedded— within the GFs, whose role is to draw and keep the attention of the addressees and, when necessary, signal the message’s end. Therefore, the sequence of occurrence of the several formulae —whether general or specific— is not codified. Rather it is left to the drummer’s choice.


Altogether the general formulae, the specific formulae, the terms of address and the conclusion formula provide the necessary elements for the production of messages. When the message is addressed to a specific individual, the latter’s identification is guaranteed by overdetermination. The sender drums out : “x-x, the son of y-y, from the village of the children of the z-z lineage”.

As we have seen, the GFs frame the information contained in the SFs and work as markers which allow the identification of the specific information in each individual message. This information revolves around one or several nouns, those with the heaviest semantic load. The occurrence of both types of formulae does not follow any regular pattern : while the messages are tightly structured, their form can vary.

Language and music in fusion

The lenga are above all musical instruments used to keep the metric basis for the dance and to ensure its rhythmic foundation. Used in pairs, they require the presence of two drummers, each one being in charge of a drum. The high-pitched drum —the “husband”— ensures the synchronicity of the dancers’ steps : its role is limited to the uninterrupted performance, in the regular manner of an ostinato, of the specific rhythmic framework of each of the traditional dances. The low-pitched instrument —the “mother”— , on the other hand, becomes the attribute of the master drummer, generally a virtuoso. This drum is also assigned a specific rhythmic pattern for each dance, but there is room for greater freedom from the pattern in order to perform unconstrained improvisations. Meanwhile, the high-pitched drum guarantees the dance’s basic pattern. The tradition is that the master drummer alternates musical improvisations and drummed utterances addressed to dancers and viewers and containing invectives, jokes — generally obscene — and insults that elicit the audience’s laughter. It is then, when linguistic utterances whose rhythm is not subject to metrical constraints become embedded in a strictly measured musical organization that music and language truly merge.


The drum language of the Banda Linda is defined by the following characteristics :

- I t is based on the reproduction of the tonal pitches and on the rhythmic patterns of normal speech. The corpus of messages, however open, reflects the existence of a limited body of information items.

- Such items can be classified as belonging to two types : those which can appear in any message (GFs) and, in opposition to them, those others which belong to each particular case (SFs). The former simply provide support or a frame for the latter.

- All information items, notwithstanding their character (general or specific) are embedded into paradigms defined by a central idea in turn expressed by means of its own tonal scheme. Therefore, for each information paradigm there takes place a phenomenon of stereotypy.

- For the SF’s the tonal pattern is the sound expression of one or several key-semantemes.

- Each message shows a considerable level of redundancy.

- Perception is global. In view of the countless risks of tonal homophony, understanding cannot operate on the level of the isolated word or on that of the syntagm, but rather on the level of the drummed-out utterance ―a tonal succession between two pauses―, since the latter always matches the realization of one or more paradigmatic themes.


The Banda-Linda drum language appears as a functional system, synchronically closed but diachronically open. It resorts to a limited number of paradigms each of which may give rise to diverse realizations. Accordingly, this drum language can be considered as equivalent to a phonological system. It is therefore likely that the drum languages of most tone languages must rely on similar principles.


  • [1] The present text revisits several points of Arom & Cloarec-Heiss 1976.
  • [2] High tone, Medium tone, Low tone, High-low tone, Low-high tone.
  • [3] Note that the sequential order of GFs and SFs varies from one version to the other : since a message’s syntax is not rigid, its elements are alternated freely.
  • [4] Cfr. CD .II .1
  • [5] It is worth explaining that the alternate version does not stand on the same level as the preceding one, since the conditions in which it was collected were of an experimental kind. We realize, however, that the two versions do not differ in their form, content and length. Cfr. CD .II .2.


  • Alexandre, P. 1969. “Langages tambourinés, une écriture sonore?” Semiotica 1(3): 272-281.
  • Ame s, D. et al. 1971. “Taaken sàmàarii : A Drum Language of Hausa Youth”. Africa, 41(1) : 12-31.
  • Armstrong, R.G. 1955. “Ta1king Instruments in West Africa”. Explorations 4 : 140-153.
  • Arom, S. and Cloarec-Heiss, F. 1976. “Le langage tambouriné des Banda-Linda (R.C.A.)”. I n: Théories et méthodes en linguistique africaine ed. by L. Bouquiaux. Paris : SELA F, 113-169.
  • Beart, C. 1953. “Contribution à l’étude des langages tambourinés, sifflés, musicaux”. Notes africaines de l’IFAN 57 : 11-14.
  • Beier, U. 1954. “The Talking Drums of Yoruba”. African Music 1 : 29-31.
  • Betz, R. 1898. “Die Trommelsprache der Duala” Mitteilungen von Forschungsreisenden und Gelehrten aus den deutschen Schutzgebieten 11 :1-86.
  • Burssens, A. 1939. “Le Luba, langue à intonation, et le tambour-signal”. Proceedings of the Third International Congress of Phonetic Sciences : 503-507.
  • Carrington, J.F. 1949. Talking Drums of Africa. London : The Carey Kingsgate Press.
  • Clarke, R.T. 1934. “The Drum Language of the Tumba People”. American Journal of Sociology 40 : 34-48.
  • Cloarec-Heiss, F. 1997. “Langage naturel, langage tambouriné : un encodage économique (Banda-Linda de Centrafrique)”. In Diversité des langues et représentations cognitives ed. by C. Fuchs et S. Robert. Paris-Gap : Ophyris, 136-149.
  • Denett, R.E. 1909. At the Back of the Black Man’s Mind. London : Macmillan.
  • Éboué, F. 1933. Les peuples de L’Oubangui-Chari. Essai d’ethnographie, de linguistique et d’économie sociale. Paris : Comité de l’Afrique Centrale.
  • Éboué, F. 1935. “La clef musicale des langages tambourinés et sifflés”. Comité de L’Afrique Occidentale Française 18 : 353-360.
  • Guillemin, L. 1948. “Le tambour d’appel des Ewondo”. Études Camerounaises (IFAN) : 21-22.
  • Heepe, M. 1920. “Die Trommelsprache der Jaunde in Kamerun”. Zeitschrift für Kolonialsprachen 10 : 43-60.
  • Heinitz, W. 1943. “Probleme der afrikanischen Trommelsprache”. Beiträge zur Kolonialforschung 4 : 69-100.
  • Herzog, G. 1945. “Drum-signalling in a West African Tribe”. Word 1 : 217-238.
  • Hulstaert, G. 1935. “De telefoon der Nkundo (Belgische Kongo)”. Anthropos 30 : 655-668.
  • Junod, H.A. 1927. The Life of a South African Tribe. London : Macmillan.
  • Labouret, L. 1923. “Langage tambouriné et sifflé”. Bulletin du Comité de l’Afrique Occidentale Française 6 : 120-158.
  • Meinhof, C. 1894. “Die Geheimsprachen Afrikas”. Globus 66 : 117-119.
  • Nekes, H. 1912.”Trommelsprache und Fernruf bei den Jaunde und Duala in Südkamerun”. Mitteilungen des Seminar für orientalische Sprachen 15(3) : 69-83.
  • Nketia, J.H.K. 1963. Drumming in the Akan Communities of Ghana. Edinburgh : Thomas Nelson and Sons.
  • Peters, C. 1891. New Light on Dark Africa. London : Ward Lock.
  • Rattray, R.S. 1923. Ashanti. Oxford : Clarendon Press.
  • Rialland, A. 1974.”Les langages instrumentaux sifflés ou criés en Afrique”. La Linguistique 10(2) : 105-121.
  • Rouget, G. 1964. “Tons de la langue en Gun (Dahomey) et tons du tambour” Revue de musicologie 10(l) : 3-29.
  • Schaeffner, A. 1951. Les Kissi, une société noire et ses instruments de musique. Paris : Hermann.
  • Schneider, M. 1952. “Zur Trommelsprache der Duala”. Anthropos 47 : 235-243.
  • Van Avermaet, J. 1945. “Les tons en kiluba samba et le tambour-téléphone”. Aequatoria VIII ,1 : 1-12.
  • Verbeken, A. 1920. “Le tambour-téléphone chez les indigènes de l’Afrique centrale”. Congo I, 3/4 : 253-284.
  • Westermann, D.H. 1907. “Zeichensprache der Ewevolkes in Deutsch-Togo”. Mitteilungen des Seminar für orientalische Sprachen 10(3) : 1-14.
  • Wilson, W.A.A. 1963. “Talking Drums in Guinea”. Estudios, Ensaios e Documentos 3 : 201-219.
  • Witte, P.A. 1910. “Zur Trommelsprache bei den Ewe Leuten”. Anthropos 5 : 50-53.

Subir >

TRANS - Revista Transcultural de Música