The Canela (Eastern Timbira), I: An Ethnographic Introduction.
By William H.Crocker
Smithsonian Contributions to Anthropology,
Number 33, 487 pages, 11 tables, 51 figures, 78 plates, 1990.

<<back   table of contents  next>>
<<return to literature


Appendix 1    Ten Field Trips to the Canela Over 22 Years

The first trip to Brazil took place in March 1957, facilitated by a University of Wisconsin James Campbell Goodwill traveling fellowship. Obtaining the permission for a scientific expedition alone, aside from the Indian Protection Service (Serviço de Proteçăo aos Índios: SPI) permission, took three months, most of which was spent in Rio de Janeiro. This time was devoted to learning Portuguese and paying visits to a number of Brazilian colleagues, including Drs. Darcy Ribeiro and Roberto Cardoso de Oliveira in Rio and Dr. Herbert Baldus and Sr. Harald Schultz in Săo Paulo. Time was also spent in Belém consulting with Dr. Eduardo Galvăo.

The first period among the Canela Indians extended from August into December. Then, upon return to Rio de Janeiro, my Indian service permission was rescinded and a wait of four months ensued, with Dona Heloísa Alberto Torres finally retrieving the permission. Subsequently, I spent seven months (April through October) with the Canela and Apanyekra tribes in 1958, after which I returned to Madison, Wisconsin, for much needed rest and consultations on the 11 months with the two tribes. The final predoctoral period of 13 months spent with the tribes was from June of 1959 through part of August of 1960. This time, Smith, Kline and French, a pharmaceutical company in collaboration with the American Anthropological Association, helped with a grant to study herbal medicines.

With the meeting of the Brazilian Anthropological Association in Săo Paulo in early July of 1963, came the opportunity for returning to the tribe for two weeks. In the halls of the meeting place one morning, Herbert Baldus told me that, according to the newspapers, the Canela had been machine-gunned by backland gunmen. My plan had been to visit the Canela directly after the meeting in Săo Paulo, but this incredible report raised a number of questions. Dona Heloísa Alberto Torres, then the head of the National Council for Protection of the Indians (Conselho Nacional de Proteçăo aos Índios), issued me a permission to enter the Canela area, and Sr. Olímpio Martins Cruz (an old SPI friend; Figure 7), who was in charge of the Indian Protection Service in Săo Luis at that time, encouraged my joining the Canela to discover what had really happened. This was probably the most fascinating time of my research career because it slowly became evident, through the casual information obtained from previous Canela research assistants, that they had just come through a full messianic movement.

Going back to the tribe in February 1964 was made possible by the joint support of the National Science Foundation and the Wenner-Gren Foundation for Anthropological Research. The specific purpose this time was to study a tribe that had been relocated from the cerrado to a dry forest environment (avarandados; Figures 5, 6) in order to determine the various factors contributing to their ecological maladjustment. Five months (not including emergency exits and returns) were spent on this ecological study, on the messianic movement, and on the morphology of the language.

The winter and spring months of 1966 were made notable by the invitation to attend the Pan-Gę seminars held at Harvard University by Professor David Maybury-Lewis and his student colleagues. Their emphasis on kinship and other relationship systems became the focus of my return visits to the field thereafter, except in 1976 and 1978–1979.

The return in 1966 was determined by two professional meetings: the Biota Amazônica (with the Brazilian Anthropological Association) at the Museu Goeldi in Belém, Brazil, in June, and the Congress of Americanists in Mar del Plata, Argentina, in September. Thus, six weeks of June, July, and August were spent in the two tribes.

The 1969 trip was a major one; consequently, its timing was not organized around professional meetings. I spent four months from August into December on census taking, kinship, and language, but only among the Canela. Then sabbatical time was taken to visit colleagues in southern South American countries, after which I spent five more months—with a long departure for the Congress of Americanists in Lima, Peru—in the two tribes in 1970. The fieldwork included major festivals, census taking, myth recording, kinship quantification in both tribes, and a diachronic marriage study done only among the Canela.

In 1971, the timing of the field visit to the two tribes was determined by the invitation of Professor Egon Schaden to give a paper at a meeting of the Brazilian Anthropological Association in Săo Paulo. Thus, three weeks were devoted to the Apanyekra in August and September and three weeks were spent with the Canela in October. I carried out kinship studies, census taking, and color perception (Munsell color chip) analyses in both tribes.

The next major trip to the field, to both tribes, came during the final two months of 1974 and during eight months (with some time out in Brazilian cities) of 1975, from January into October. The kinship and other relationship studies were completed (including the marriage study), and major festivals were studied and censuses were carried out in both tribes. The myth translations were nearly completed at this time.

I made one quick visit of four days to the Canela in February of 1976, just after the meeting of the Brazilian Anthropological Association in Salvador, Bahia. Besides routine checking, research assistants helped each day and evening on folk Catholic beliefs versus aboriginal Canela origin myths concerning the earth, the Canela, and the backland Brazilian.

With studies parallel to those of professional colleagues largely completed, it was possible to turn to my own interests in 1978 and 1979, but only among the Canela, not the Apanyekra. I studied topics such as festivals, religion, ethnobiology, and cognitive psychology during one of the longest field stays (14 months) from July into October of the following year, with visits to Săo Luis for family reasons in January, April, and July. The purpose was to complete the Canela field research in case I never returned. There were 12 separate research trips to the Canela undertaken between 1957 and 1979. I hope to return for a short trip in the 1990s.

[2003: Between 1991 and 2001, I returned to the field eight more times to total 74 months in the Canela and Apanyekra villages. During 1991, sponsored by the famous Brazilian anthropologist Berta Ribeiro, I returned with my new wife Jean to spend three weeks there to prepare for a major research trip in 1993. The four months there during 1993 were spent on recovering the history of the 1980s; on studying the language with new materials from the linguist-missionary Jack Popjes; on helping the Smithsonian cinematographer Carl Hanson shoot all the house and building structures, inside and outside, as well as two festivals (with my son Myles as his assistant); on carrying out a full census with demographer Margaret Greene, Ph.D., including many questions on culture change; on making and documenting a large collection of the still existing items of material culture for the Museu Nacional of Rio de Janeiro, for the Museu Goeldi of Belém, and for the Smithsonian Institution; on tracing the successions and changes in festival ownership of rites since the 1970s, and on learning and enhancing the various procedures of collective farming. During 1974 and 1975, I returned three times for short check up visits to study and enhance Canela collective farm methods.

In 1997, cinematographer Steven Schecter and I spent three weeks with the Canela filming them for footage to compare the 1990s with the 1970s, when Schecter had spent two months filming them for socialization practices. During the summer of 1999, I passed one month with them working closely with research assistants to obtain sensitive materials from old Canela women on the demise of the extramarital sex system as well as on its nature. From mid-September into early December of 2001, I led a team of five, including the Brazilian ethnologists Ana Carolina Pareschi and Hélder Sousa, the American ethnologist Priscilla Linn, Ph.D., and my photographer son Myles Crocker. We carried out a careful culture-change census of peoples, materials, and practices, using many of the same questions asked in 1993, 1979, 1975, and 1970. Linn collected materials on the various messianic movements, especially those of the 1980s and the one of 1999. During the month of August 2003, I hope to collect materials on revenge, on sources of outside income, and on the reasons why so many young Canela men and women want to get Brazilian educations in Barra do Corda.]


Appendix 2   Canela and Apanyekra Collections at the Smithsonian Institution

Besides the various kinds of collections that are at the Smithsonian Institution, as reported below, collections of almost all sorts (photographs, film, tapes, artifacts) have been given to the Indian Protection Service of Brazil (SPI) and the National Indian Foundation of Brazil (FUNAI) during a number of different years. The same general materials (including manuscripts as well) have been donated, but in far greater numbers and quality, to the Museu Paraense Emílio Goeldi of Belém, Brazil (de La Penha et al., 1986). Lesser material artifact collections have been sent to the Museu Nacional (Rio de Janeiro), Museu do Índio (Rio), and the Museu Paulista (Săo Paulo). Small artifact collections (about 120 items each) were given to the following institutions in Brazil: the Catholic Church and the Maranata School, both in Barra do Corda; and the Institute of History and Geography and the Indian Protection Service, both in Săo Luis. Similar small collections were sent in 1960 to the University of Wisconsin and Harvard University. Considerable numbers of black and white and colored prints (including Polaroids) have been left with both the Canela and the Apanyekra on every return visit.


Number of Items: 934

Years Collected: 1964, 1970, 1975, 1993, 1999

Comment: This Canela collection, maintained by the Department of Anthropology, National Museum of Natural History (NMNH), may be the largest held by any museum, but the one at the Goeldi Museum in Belém is far superior in quality.


The original negatives and prints are stored in my office in the NMNH, but representative collections have been given to the National Anthropological Archives, Smithsonian Institution.

[Ap.2.b.(1)] Black and White

Number of Items: ~960

Years Collected: 1957–1958, 1959–1960, 1964, 1966

Film Size: Both 56 and 35 mm in the late 1950s; only 56 mm in the 1960s

Comment: Leica and Rolleiflex cameras were used.

[Ap.2.b.(2)] Colored Slides

Number of Items: 20,000+

Years Collected: 1957–1958, 1959, 1963, 1964, 1966, 1969–1970,1971, 1974–1975, 1978–1979, 1993, 1999, 2001

Film Size: 56 and 35 mm in the late 1950s; only 35 mm thereafter

Comment: Informal training from National Geographic Society in 1969 led to a change from German to Japanese cameras (Nikon and Nikkormat) and taking more and better slides of festivals, rites, sports, houses, house interiors, people, etc.

[2003: Myles Crocker has contributed 5 thousand slides taken in connection with the 2001 census.]

[Ap.2.b.(3)] Color Prints

Number of Items: 1000+

Years of Collection: 1978–1979, 1993, 1999

Film Size: 35 mm

[Ap.2.b.(4)] Polaroid Prints

Number of Items: 425

Years of Collection: 1964, 1966, 1970, 1975, 1979

Comment: Used to supplement family census of Canela and Apanyekra tribes in 1970, 1975, 1979 (black and white; color, but only for the Canela in 1979). A vertical meter rod can be seen in almost all photographs and may be used to measure approximate height of individuals.

[2003: Myles Crocker took over 1,000 large size digital photos and William Crocker shot several hundred small size digital shots during the census taking in 2001.]


Originals of the Super-8 films (all in color) have been given to the National Anthropological Archives. This film is of the Canela only.

[Ap.2.c.(1)] Silent Super-8 Film

Size:    ~16,000 feet

Years of Collection: 1974–1975, 1978–1979

Comment: Used to supplement field notes of festivals, life cycle rites, sing-dancing, and any sports.

[Ap.2.c.(2)]  Sound Super-8 Film

Size:     ~2000 feet

Years of Collection: 1978–1979

Comment: Used to supplement field notes of festivals, life cycle rites, and sing-dancing.

[Ap.2.c.(3)] 16 mm Film

Size:  3 to 4 hours

Years Collected: 1957–1958, 1959, 1964, 1970

Comment: About 20 minutes of an edited film was produced on the Festival of Masks from the many hours of film taken of the Mask and Pepkahŕk festivals by Ray Roberts-Brown in 1970; and about one hour and a half was taken of four festivals in the late 1950s and 1964 by Crocker, with subtitles in total footage, and with notes on sound tracts (no sound-track notes for 1964).

[2003: Considerable video footage was taken by Carl Hansen of the Smithsonian during August, 1993, of Canela and FUNAI structures and of festivals (Pepkahàk and Closing Wè?tè). This Hi-8 video has been converted to DVD. Steven Schecter took footage in September 1997 to supplement his footage taken in 1975 and 1979, all toward the making of the video Mending Ways (1999).]


Originals are stored in my office in the NMNH; representative collections have been given to the National Anthropological Archives, Smithsonian Institution. All the 1978–1979 Nagra-recorded musical tapes were cleaned and copied by what is now the American Folklife Center, Library of Congress. A copy of these tapes was given to them.

[Ap.2.d.(1)] Music (Canela Only)

Size:     ~140 hours (total)

Years Collected: 1957–1958, 1959 (spring-wound Nagra) + 1966 (small Norelco) + 1969–1970 (Uher) (total-22 hours):

Fifteen Nagra 5 inch reels were sent to Alan Lomax’s Choreometrics Project as data for his 1968 book.

1975 (~12 hours): Cassettes of casual and ritual singing, not of the great festivals.

1978–1979 (~20 hours): Cassettes made by the younger Tep-hot of the younger Tŕŕmi (a great sing-dance master) singing alone were collected for the “song-saving” program.

1978–1979 (~84 hours): A modern Nagra with open-reel tape was used to record individual singing for the “song-saving” program, and for the great festival group chanting.

Comment: The whole musical collection constitutes primary data for potential publications (see item 16 in the planned Canela series as described in the Preface [Pr. l], and Appendix 3 [Ap.3.e.(2)].

[2003: The 1978–1979 musical recordings were made available to The Library of Congress for cleaning, copying, and storage there as well as at the Smithsonian. The entire Canela musical collection was made available to the ethnomusicologist, Anthony Seeger, for his copying, analysis and eventual publication. Dr. Seeger was with Smithsonian Folklife during the 1990s, but now is with the University of California at Los Angeles.]

[Ap.2.d.(2)]    Myths and Stories

Size:     ~210 hours (total)

Years Collected: 1957 (~1 hour): Canela myths by the older Mďďkhrô in Gę

1963, 1964 (~14 hours): Stories about the Canela messianic movement of 1963 (partly in Gę)

1970 (~14 hours): Myths by old narrators, the older Pů?tô (Canela) and Kupaakhŕ (Apanyekra)

1975 (128 hours): Canela and Apanyekra myths and war stories in Gę translated to backland Portuguese by the younger Kaapęltůk and summary of discussions in English

1978–1979 (64 hours): Canela myths and war stories processed as in 1975

Comment: It is expected that most of these narratives will be translated into English and published. In any case, they are primary data for potential publications (see item 13 in the Preface [Pr.1]; see also Appendix 3 [Ap.3.b]).

[Ap.2.d.(3)] Judicial and Political Meetings (Canela Only)

Years Collected: 1974–1975 (20 hours), 1978–1979 (68 hours): On cassettes in Gę of legal hearings between extended families in family houses and of political meetings in the plaza

Comment: This collection constitutes primary data for a later study of the judicial and political processes (see item 14 in the Preface [Pr.1] and Appendix 3 [Ap.3.a]. The recording conditions were poor.


Years Collected: 1970–1979 (708 hours): In Gę made continuously by the younger Kaapęltůk and Kaapręępręk

Comment: The recording was done in relatively slow, understandable Gę and is of good quality. These personal documents are sources for potential psychocultural and acculturative studies (see item 12 in the Preface [Pr.l] and Appendix 3 [Ap.3.a]).

[2003: This program was continued in 1991 in Canela. By 1995, the diarists were sufficiently proficient in Portuguese for them to speak only in this language for two hours a month. Six diarists contributed in this way during 2001, but by 2003, two were added. They are the folowing: Younger Kaapêltùk, Major Tep-hot, Hõõkô, Piyat, Yiilot, Kää-re, Kaarà-mpey, and Tê?tê.] 

[Ap.2.d.(5)] Sound Tracks of 16 mm Film (Canela Only)

Years Collected: 1975 and 1979 (73 hours): In collaboration with Smithsonian’s National Human Studies Film Center

Comment: The sound tracks were transferred to cassette tapes, which were transcribed (typed) onto paper in Gę and backland Portuguese by Kaapręępręk and Yaako and are mostly of children’s and their relatives’ conversations and of festivals. These materials constitute primary data for potential publications (see item 15 in the Preface [Pr. 1] and Appendix 3 [Ap.3.a]).

[Ap.2.e]   Autobiographical Diary Manuscripts

(Canela Only)

Years Collected: 1964 to 1970 (30,000 pages), 1970 to 1975 (18,000 pages), 1975 to 1977 (10,080 pages), 1977 to July 1978 (9,720 pages), July 1978 to January 1979 (3,960 pages), January 1979 to October 1979 (6,480 pages), October 1979 to November 1979 (180 pages): all equal a total of 78,420 pages. Written by the younger Kaapęltůk, the younger Pů?tô, Hŕwpůů, Kaapręępręk, the younger Tep-hot, Yaako, Krôôpey, Yôk, Hőőkô, Kôyapŕŕ, Yiirot, and the younger Krôôtô.

Comment: This extensive collection of personal documents constitutes a source of primary data for the study of psychocultural and acculturative trends and changes. It can also be the source of autobiographies and biographies. Some writers wrote in Canela and others in Portuguese, and still others wrote in Canela and translated their work the next day into backland Portuguese [I.F].

[2003: This program was continued in 1991, but by 1995 the writing of manuscripts was terminated in favor of speaking onto micro-cassette tapes in Portuguese for two hours a month. This procedure has been used to collect special data for publications such as materials on the demise of the extramarital sex system and on the subject of revenge.]


Appendix 3   Primary Materials For Future Studies

These associated publishable materials are primary field materials which constitute a source of data for potential publications in which collaboration with other specialists would be welcome under certain conditions and agreements.

[Ap.3.a] DIARIES ON PAPER (1964–1979) AND TAPE (1970–1979)

Collecting daily diaries was possible because the school teacher, Dona Nazaré (later the wife of Sr. Hugo Ferreira Lima) who taught among the Canela from 1944 to about 1949 was successful in teaching six young Canela boys how to read and write in Portuguese. Thus, in 1964, it was possible for three Canela men to begin writing daily manuscripts for pay. Then in 1970, two new manuscript writers were employed, and two of the five manuscript writers started to make monthly cassettes on tape recorders left with them for this purpose. This collection of manuscripts (~78,420 pages) and tapes (708 hours) about the daily activities and thoughts of these 5 (and in 1978–1979, of 12) native research assistants, constitute a large source of primary field materials. These data, and some of the individuals involved in creating them, are more fully described in Part One [I.F.].

[2003: This program was continued in 1991 and is maintained currently as micro-casstte tapes spoken by diarists in Portuguese.]


About 120 myths have been collected from the Canela and Apanyekra tribes. This tape cassette recording was completed in 1970 and 1971, and added to in 1975 and 1979. The tapes collected in 1970 and 1971, in both tribes, were made by very old individuals, both of whom were dead by 1974. In 1975 these myths and war stories, together with additional ones, were translated by the younger Kaapęltůk from these tape cassettes into Canela backland Portuguese onto other tape cassettes in my presence. His procedure was to listen to the one of the original cassettes for several minutes and then to speak his free translation (and admittedly, maybe his own interpretation)25 onto another cassette. This was then a re-recording of the old speaker’s narration with the segments of Kaapęl’s translation interspersed about every three to five minutes. In 1975 and 1979, we continued to carry out this same procedure.

Sometime in the future, these primary materials need to be translated line-by-line, instead of just in the free parallel manner already done by the younger Kaapęltůk, for publication in a volume of Canela and Apanyekra myths and war stories. Six of these parallel translations have already been published in English by Johannes Wilbert and Karmn Simoneau (1984).

One of the particularly interesting possibilities for research will be the close comparison of Canela and Apanyekra versions of the same myth. These two tribes differ in certain small but very significant ways which might become further clarified and better documented through comparative analysis.


These tapes are of tribal council meetings [III.D.2.a] held in the plaza, with some 10 to 40 men attending but with only about a quarter of them actually speaking (one at a time). Several dozen 120-minute cassettes of such meetings have been archived in my office. However, since they are totally in Canela and the recordings were not taken under ideal conditions, their translation and interpretation will require a great deal of work, which could be greatly facilitated by help from Canela research assistants.

Tape recordings of the judicial meetings [III.D.3.b] that are held between extended families in one of their houses, or in the house of the chief, were collected more extensively and under better conditions than it was possible to carry out for the political meetings in the plaza. (There are 88 tape-recorded hours of political and judicial meetings.) In most cases, written notes were taken of such judicial hearings, including the sequence of speakers. In the late 1950s and in 1964, such meetings were recorded in speedwriting, which was later transcribed onto McBee keysort cards. Thus, a sizable body of primary materials exists on judicial and tribal meetings (më aypën pa: audięncia: hearings).

Politics is one of the most difficult topics to study among the Canela, because of their emphasis against interpersonal aggression within the tribe and against anyone pretending superiority over anybody else. Consequently, these primary data on tape should yield information about Canela political attitudes and actions that have not been possible to record in the more traditional ways; however, future Canela research assistants may be necessary to develop such data.


During prolonged stays in the field with the Canela in 1975 and 1979, cinematographer Steven Schecter filmed about 36,000 meters (120,000 feet) of 16 mm film, most taken in a random sample. He focused on child training, festivals, and life-cycle rites. This filming was carried out under the direction of Dr. E. Richard Sorenson and myself, and took place under the auspices of the Smithsonian’s National Human Studies Film Center (NHSFC) and the Museu do Indio in Rio de Janeiro. These films are not part of my collections (i.e., they are not under my control), but they are now with the Smithsonian’s National Human Studies Film Archives, successor to the NHSFC. In 1978, however, the sound track of the films made in 1975 was put on tape cassettes by Smithsonian personnel in Washington, D.C., and delivered to me before my departure for the Canela in early July. The 1979 film sound tracks were transferred directly to cassettes in the field among the Canela. Thus, during the entire 1978 and 1979 field stays, two young Canela men (Kaapręępręk and Yaako), who were trained by Jack Popjes, touch-typed the entire sound track onto paper in Canela and also translated it into their own version of Portuguese.

These sound tracks are mostly conversations among children, or admonitions from parents or parent surrogates. They represent a medium that would be interesting to contrast with the judicial and political tape recordings and the diary manuscripts and tapes. Eventually, the studies of the political and psychological dimensions of the Canela sociocultural system will be considerably enhanced if thorough research is carried out on these primary data by some researcher fluent in the Canela language.

[2003: These materials were used in 1998 to provide quotations ("native voices') for the video, Mending Ways, but were found poor in quality and therefore are not currently recommendable.]


The Canela and Apanyekra tribes are remarkable for their choral singing, and for the emphasis and amount of time they devote to such secular and ceremonial performances [II.F.1]. An insignificant percentage of the tape recordings are of Apanyekra singing, so comparative ethnomusicological studies cannot be carried out between these two tribes, as can be done for myths and war stories. Canela singing in the late 1950s, however, can be compared with the singing of much of the same choral music in the late 1970s.

In the late 1950s, the principal singing performed during three Canela festivals was recorded on a hand-wound Nagra. In the 1960s and the early-to-mid-1970s, additional recordings, though inferior in quality, were made of a number of other festivals.

[Ap.3.e.(1)] Quality Recording with a Nagra in 1978–1979

The major quality collecting was carried out in 1978 and 1979 when the Smithsonian’s National Human Studies Film Center lent me a modern Nagra tape recorder. So much taping in 1978–1979 was undertaken that most of the recordings from the whole 22-year period were made then.

During those 14 months with the Canela, 25 percent of my evenings were spent on the extensive recording of Canela choral chanting and individual singing, and in playing back these recordings (as required by the Canela). The singing and chanting were taped either in the plaza or along the boulevard if the singing was part of their on-going activities.

[Ap.3.e.(2)] Song Conservation Program

The recording for the song conservation program took place in my family house, where individuals came to sing special songs. The words sung by individuals are more likely to be distinguishable on tape than words chanted by a group during an on-going activity. Thus the words and music from these artificial recording procedures are more likely to be preserved for the future. Again, help from future Canela research assistants will be needed to carefully transcribe and translate such songs.

The work undertaken in 1979 in the field has already demonstrated to me that extracting meaning from such words and phrases takes all the training and ingenuity that the younger Kaapęltůk has to offer. It is notable that while certain colleagues in ethnology do translate songs and narratives on their own, Jack Popjes of the Summer Institute of Linguistics, with professional linguistic training and more than 20 years of experience with the Canela language, does not attempt to translate Canela for professional purposes without the help of his trained research assistants.

It is expected that some day an ethnomusicologist will carry out the analysis and interpretation of this collection of Canela individual chanting and choral singing from the point of view of its music rather than just the words.


Photographs were taken on every visit to the Canela, and usually to the Apanyekra, from 1957 through 1979. From the late 1950s through 1964, the emphasis was on taking black-and-white photographs in the 56 mm size (Rolleiflex). In the 1970s, a shift was made to 35 mm photography (Nikon and Nikkormat) and the number of shots taken was increased ten times.

[Ap.3.f.(1)] 16 mm Research Films

In 1957 a second-hand 16 mm camera (Bell and Howell) was brought to the field, and a limited amount of footage of festivals was taken in 1957, 1958, and 1959. The film was edited in the mid-1960s, titles were made to describe changes in subject matter, and a sound track of my commentaries was added. Additional lengths of film, also just on festivals, were taken in 1964 with the same camera, but no sound track was recorded for this footage. All footage is retained in the sequence in which it was recorded.

[Ap.3.f.(2)] Super-8 Film of Rites, Festivals, and Athletics

In 1974 and 1975, and then again in 1978 and 1979, extensive Super-8 footage was taken of the significant festivals, life-cycle rites, log races, and other athletic sports. None of these films have been edited in any way, but all of them have been copied several times. The most interesting ones have been sent to various places including Brazilian museums.

Copies of the unedited original, whole lengths of Super-8 films have been retained as field notes, while the original footage remains untouched for posterity. Thus, in the later studies of festivals and of life-cycle rites [Pr. 1], this Super-8 film will furnish an additional dimension to the usual handwritten or tape-recorded field notes. Some of the 1979 Super-8 lengths are in sound, but none of these films of any year include the Apanyekra Indians.

[Ap.3.f.(3)] Study of Canela Houses

Care was taken to photograph the interior and exterior of every house around the village circle of Escalvado during every visit, and to some extent around the villages of the Apanyekra, so that studies can be made using the visual evidence of change and acculturation during the 1970s. While such photography (on stills and super-8 film) includes a good coverage of log racing and the various track sports of both tribes, these collections are considerably less than complete with respect to photographs on farm-oriented topics, hunting, and food collecting.

[2003: The recording of Canela and FUNAI structures was continued, especially, in 1993 and 2001. Carl Hansen of the Smithsonian video-recorded all structures in 1993 and Myles Crocker took digital stills of all structures, inside and outside, in 2001. During most other visits during the 1990s, I took color stills of most structures, including those of farm communities.The 1993 video is now on DVD.]


lthough collecting items of material culture was not a principal focus of any of the ten Canela field trips, extensive collections were made in 1960, 1964, 1970, 1975, and in 1979. (In the last case, the entire collection went to the Goeldi Museum in Belém). Field notes on the more valuable items in these collections were made at the time I purchased them. A special study on the relationships of items of material culture to the social and cultural life of the tribe was carried out in 1970. Only about 10 percent of the collections come from the Apanyekra.

There are 150 traditional categories of material artifacts, but items in only about 100 of these categories are still being made. An outstanding turning point in artifact fabrication occurred between 1964 and 1966 when traditional artifact sales increased tremendously in the commercial markets arranged by the Indian service. As expected, the quality of these traditional artifacts, when prepared for commercial outlets, was considerably lower and the materials used also varied from the traditional. Both tribes, however, still put high quality work into the traditional items that are used and distributed within the tribes, though this quality is not as high as it was during the time of Nimuendajú. Furthermore the supply of some raw materials, such as certain hardwoods, has been exhausted in the Canela area.

In 1975, but even more so in 1979, my orientation to collecting changed. Instead of trading for almost every category artifact to fill out the traditional list of categories, resources went for a small number of their traditional but personal items of great honor and prestige. Such items as necklaces for little girls and singing belts for men have a large number of ceramic beads and, therefore, were costly. In addition, these artifacts were their treasured honor-award items which young women and young men had won and stored carefully away over the years.


About 120,000 feet of film were taken among the Canela by Steven Schecter of the National Human Studies Film Center (now Archives) of the Smithsonian Institution in collaboration with the Museu do Índio in 1975 and 1979. He focused on socialization and festivals, particularly the Pepyê festival of 1975. This footage should be annotated by Schecter and myself before it can be considered "research footage" and ready for editing into "research films." On short research film of this sort was processed in 1975 and copies were delivered to the Museu Goeldi and to the FUNAI at the time of the meetings of the Associação Brasileira de Antropologia in Salvador in 1976, as agreed to before the filming tookplace in the field.


Appendix 4   Linguistic Notes

No attempt is made here to present a full linguistic description of any part of the Canela language. This has been largely accomplished already by Jack Popjes (Popjes and Popjes, 1972) of the Summer Institute of Linguistics (Wycliffe Bible Translators). However, a partial description is furnished to make certain limited aspects of the phonemics of the Canela-Krahó language easily available to anthropological readers. I am responsible for this partial description, and where Popjes has made a major contribution to my materials, this is acknowledged below.

Kenneth L. Pike’s (1947) approach to phonemics is followed in this monograph. The material below is an expansion of the material presented in a tabular form in the “Linguistic Key” [In.5].

[Ap.4.a]  PHONEMES

Canela is written phonemically in this monograph with only a few systematic exceptions, as explained below.

[Ap.4.a.(1)] Vowels

The Canela-Krahó language has 15 regular (frequent) vowels and two relatively rare ones (ä and û). Seven unnasalized vowels are similar to vowels in Portuguese (a coincidence) and are written in the same way. The orthography for these seven vowels is followed in this sentence by their approximate equivalents in English, vowel length not being considered: i (beet), ę (bit), e (bet), a (hurrah), u (boot), ô (boat), and o (bought). The vowels u, ô, and o are rounded but the others are not.

[2003:] The phonemic nasalization of o and ŕ is shown by the tilde (ő and ă), while the phonemic nasalizaion of i, e, a, and u is shown by the dierisis (ď, ë, ä, and ü). The nasalization of ù is shown by the circumflex (û). Some of these phonemic representations of nasals are followed in this sentence by approximations in Portuguese, in which a vowel is nasalized by being followed by m or n: ď (pinto), ë (pente), ã (lanche), ő (ponto), and ü (junto). The rare nasals ä and û cannot be appoximated in Portuguese. The nasal ä is really the nasaliztion of a, but ã was used for the nasalization of à because it is so frequent. This general confusion is forced on us because Latin only had five vowel letters while Canela has 17 vowel phonemes, and also because of the limitations of HTML as a system.]

There is also a nonphonemic partial nasalization of vowels when adjacent to m and n, depending on other factors of the environment.

There are three unnasalized, unrounded, back vowels: ů (high and closed); č (mid and closed); ŕ (mid and open) (Pike, 1947:5). These three back vowels are difficult to approximate in English, but ŕ is close to the u in “puddle” but is more open. The ù may be like u in "tu" and the è like eu in "peux" in French to some people.

[Ap.4.a.(2)]  Semivowels

Canela has two phonemic glides written as w and y, which are found both before and after certain vowels; when they appear after a vowel, they serve to complete the vowel’s length. Their qualities are usually similar to English as found in “west,” “pew,” “yes,” and “coy,” but the initial w can be sometimes a voiced bilabial or labiodental fricative. The y in Canela ranges from the y in “yes” (most frequent), through the n in “new,” to the unvoiced s in “sky,” depending on its environment, but a y is always palatalized. The y is the most variable phoneme in Canela, having the most seemingly different allophones.

[Ap.4.a.(3)] Consonants

The lateral r [r,l] (often flapped) in Canela is similar to what is found in Portuguese between vowels. A Canela syllable initial r (this information is provided especially for northern Brazilians) is never a palatal or velar fricative like the French "r" or the Spanish "j," and a syllable terminal Canela r resembles the "l" of the same position in Spanish rather than the "l" of the same position in Portuguese or English.

The affricative ts varies between the [ts] in “its” (most frequent) and the [ch] in “church,” and is never voiced.

The nasals are m, n, and g. The Canela m and n are similar to English, and the only rare phonemic consonant g appears mostly as an alternative second person indicator: gapal / ayapal (your grandson); or gő?camarad/ayő?camarad (your lover). It is pronounced more to the rear of the mouth (palatal and velar) than in English, overlapping the stop k.

The stops p [p,b] (bilabial), t [t,d] (labiodental), and k [k,g] (palatal and velar)—all unaspirated—are found both unvoiced and voiced depending on the environment, while kh (aspirated velar) only appears unvoiced. The glottal stop (?) is held by Popjes to be merely an allophone of h. These two sounds cannot be found in the same environment: h always being syllable initial and ? always syllable final; thus Popjes may be correct, but this distinction will be preserved in my orthography anyway for certain reasons, which are noted below [Ap.4.c.(2)].

The voiceless fricative h varies considerably in degree of fricativeness and placement in the back of the mouth and throat.

[Ap.4.a.(4)]  Vowel Length

There are no diphthongs in Canela, but vowel length (a single vowel contrasted with a double one) is phonemic: single vowel length (V) is approximately doubled by the same following vowel (VV). The glides (w and y) complete syllable length, standing in the place of the second vowel which doubles the length of the same initial vowel. Consonants also complete this same syllable length (double vowel length). In summary, in syllable medial position, a short vowel’s syllable length can be completed, and the syllable terminated, by the addition (1) of the same vowel, (2) of one of the two glides, or (3) of a consonant. In a word terminal position vowel length is not expressed in an isolated word. A short vowel’s syllable length in a word terminal position when the word is not isolated, is often completed by the consonant that begins the following word. It also can be completed by the consonant affix of the next word, which is expressed just when a short vowel terminates the previous word. After a word-terminating long vowel, the consonant affix of the following word (if it has one) is not expressed.

In other words, if a vowel is short and syllable medial, a consonant must terminate the syllable, and if a vowel is short and syllable final, a consonant has to be drawn from the following word (if it begins with one) to complete the syllable; or, the affix used by the following word for this purpose, and others, has to be utilized to complete this preceding syllable. Adjacent long vowels terminating and beginning two words may be contracted into only double vowel length in ordinary speech. (Examples: kaa-tswa (salt), katswa/kat-tswa (night), (garden staples), hőő (buttocks), hő m-pey, (garden-staples fine/good-looking), hőő pey (buttocks fine/good-looking). (For more on syllable length, see Popjes, 1982g:7–8.)

[Ap.4.a.(5)]  Word Stress

Stress almost always falls on the last syllable of isolated words but is not phonemic. Stress is altered and determined at the phrase level rather than the word level. (This observation was first made by Jack Popjes (Popjes and Popjes, 1972).) Consequently, stress is meaningless and inconsistent with respect to separate words, and so, will only rarely be utilized in this orthography. The acute accent (´) over a vowel will be used to indicate stress whenever necessary. It is not used for any other purpose. However, when a stress mark is over an a and indicates the superlative, the sound of the a is altered to ae, as in “cat.”

[Ap.4.a.(6)] Nonphonemic Orthography

Some words are written nonphonemically in order to indicate their most frequent pronunciation as found in certain environments. For instance, the word “buriti” (a palm), /krow/, is usually written krówa because this is the way it sounds, though the terminal a is not phonemic and is often not expressed in sentence sequences. The stress mark is added over the o (ó) to indicate that when found alone, this word is pronounced with the stress on the second to the last syllable instead of the last one, according to the usual practice for isolated words.

Examples of nonphonemic orthography and the use of the stress mark are the following.

Khęętúwayę for /Khęętuwyę/ (first initiation festival)

píyapit or /piypit/ (unmarried nonvirginal woman or man)

pŕlrŕ for /pŕl/ (non-buriti log). The acute accent over the first ŕ cannot be expressed because there is a grave accent there already.

Pró-khămmă for /Pro-khăm/ (principal age-set of the council of elders). Again, the acute accent cannot be expressed over the first ă.


Before Jack Popjes arrived among the Canela in the village of Sardinha in 1968, I sent him the phonemes, some grammar, and a long list of phrases (several hundred). He later confirmed the existence of the three rare phonemes I was already suspicious of, namely, /a/ nasalized and /ů/ nasalized, and /g/. Our phonemic writing of the language in 1969 and even now differs mostly because I hear (adjacent to /m/ or /n/) /a-nasalized/ as being /a/ while he hears it as /a-nasalized/. He is right, of course, that is, I do not usually “hear” the distinction between /a/ and /a-nasalized/, when adjacent to /m/ or /n/. Consequently, I have changed my earlier phonemic orthography in this respect only in a few cases (päm: father), because I have not personally mastered this difference. This is also the case with other vowels coming before /m/ or /n/ such as tum, which Jack would probably write tüm.

Later, Jack decided to merge one of the two rare phonemes, /a-nasalized/, with the frequent phoneme which is written as /ă/ in my current orthography; that is, /a-nasalized/ (a front, unrounded, low, open vowel) is merged with /ă/ (a back, unrounded, mid, open vowel) for the purposes of facilitating the Canela to write their language. However, this practice is not followed here.

When it is possible to identify this rare phoneme, /a-nasalized/, with certainty in my current orthography, instead of representing it as merged with the frequent phoneme mentioned above, /ă/, I use /ä/, as in päm (father). Û is used with the other rare vowel phoneme. Thus, /ů-nasalized/ becomes û.

Jack arrived to find several trained research assistants, especially the younger Kaapęltůk, whom as late as 1984 Jack used as the final judge when it came to clarifying shades of meanings (personal communication). When Jack said that my presence for 11 years before him in the tribe had advanced his work by two years, he was referring mostly to the ability of Kaapęl to help him. Kaapęl, who had been writing almost phonemically for me since 1964, could consciously and deliberately identify the phonemes in almost any word; and later, he learned to do this even for the rare phonemes.


Before 1968, I used symbols from Kenneth Pike’s Phonemics (1947:5,7) in my fieldwork orthography and published articles. The younger Kaapęltůk also used this source in his diary manuscript writing. In 1969, in cooperation with Jack Popjes, I adopted his orthography, both for my own field notes and for the writings of the younger Kaapęltůk, but not for my publications. Among the several changes, my writers and I used /z/ instead of /y/ and /c’/ instead of /kh/.

[Ap.4.c.(1)]  Modifications in My Orthographic System

For my publications after 1971, several of the symbols used in my earlier Pike-derived orthography were modified in order to eliminate the more irregular and unusual letters and their diacritics. This approach was designed to be easier for a publisher and for colleagues with typewriters.

Beginning in 1989, for this monograph and any later publications, additional modifications are being adopted to make the orthography still clearer and easier to use. The dieresis in my post-1971, but pre-1986, publications is being eliminated entirely as a difficult diacritic to write. Thus, in the new orthography, /ä/ becomes /ŕ/, /ë/ becomes /č/, and /ď/ becomes /ů/, with the i being changed to u because handwriting a superscript over an i creates difficulties. Moreover, to the Brazilian ear, this phoneme sounds more like a u than an i. Whereas the /č/ sounds more like an open o to the Brazilian, it is nevertheless being left as an e to remain more consistent with my earlier publications, with the publications of Nimuendajú, and with the publications of my Brazilian colleague Júlio Melatti, who is the expert on the closely related Krahó tribe. (Jack Popjes considers the Krahó dialect to be the “same language” as Canela and Apanyekra, whereas the other Timbira languages are not included in this category.)

The apostrophe, when it represents the Canela phonemic glottal stop, is likely to be overlooked and accidentally left out entirely by future publishers and colleagues. Moreover, it can be taken by mistake to be the symbol to indicate stress. Because the traditional symbol for the glottal stop is not available at the publishers, the question mark (?) is used herein. This symbol will never be confused with the question mark at the end of an interrogatory statement because the Canela glottal stop is never found in a word terminal position.

This recalls a difference between Jack Popjes and myself in the phonemic representation (although not in the hearing) of the boundaries between what is either one or two phonemes. One orthographic symbol can be utilized for the syllable initial h and the syllable terminal ? because they do not appear in contrast with each other in Canela; but whether they are two phonemes, or one, is still a question I am sufficiently unclear about. On the advice of the linguist Floyd Lounsbury, however, the syllable initial /h/ and the syllable terminal /?/ are being left as two phonemes here, just as I wrote them before Jack’s arrival in Brazil.

Just after my arrival in Brazil at Belém in late October 1974, both Jack and I were asked by the Indian service in Brasília to attend a meeting with certain Indian service and SIL personnel. The purpose of the meeting was to reach an agreement with respect to how the various Northern Gę languages should be written by Indian service and SIL personnel. It was agreed to use the same symbols for the quite consistent vowel qualities from the Canela in the east to the Kayapó in the west. The consonants are not consistent in these languages, so a variety of symbols were to be allowed for their representation. I was assured that anthropologists were specifically exempted from using this 1974 Indian service/SIL (i.e., FUNAI/SIL) orthography in their professional publications.

The letter y was adopted there as the sixth symbol for vowels in Indian service and SIL writing. Because Canela has 15 regular vowel phonenes and two rare ones, an extra symbol for use as a vowel considerably facilitates the writing of the language so that fewer diacritic marks have to be employed. The utilization of y as a vowel, however, is very unappealing to most Brazilians who do not use this letter in Portuguese and only employ it in the spelling of foreign names and words. There was great resistance to this usage among Indian service agents in the 1970s, though not among the Canela themselves. Consequently, y as a vowel has never been used in my publications out of respect for Portuguese speakers, but all my Canela manuscript writers have done so since 1974. Jack Popjes (1982a–f) has published myths using this orthography and Darrel Posey (1981, 1982, 1983) has used it in a number of publications about the Kayapó.

In the 1974 Indian service/SIL practice, /y/ is the current /ů/ of this monograph, /&/ is the /č/, and /tilde y/ the /mu/. Then, if y is used as an extra symbol for a vowel, it cannot be used as a syllable final glide for which it is well suited traditionally in English and Spanish. A j would have to be used instead, which is the SIL practice. This solution would be unappealing to English, French, Portuguese, and Spanish speakers because to these first three speakers j is a voiced affricative/fricative, and to the Spanish speaker it is an entirely different unvoiced velar fricative. In Canela, in sharp contrast, the /y/ sound is usually like the English y in “yes” and “pay” (and like the Mexican y in yo and mayo), though it varies in range more than any other phoneme in the language. This variation ranges from the English n in “new,” (Yő?hę: a male name) and the English y in “yes” (Pepyę) to the unvoiced, palatalized sibilant in “dossier” (i-?khyę: her/his-sibling). While they were in Washington, D.C., in 1987, I consulted extensively with both Júlio and Delvair Melatti on this point, and also on the rest of the orthography proposed in this monograph for the Canela/Krahó, and they concurred on every point.

My choice, keeping /h/ and /?/ separate, does constitute a distinct contrast with Jack’s solution and the 1974 Indian service/SlL’s subsequent decision. Moreover, even if these two phonetic sounds were allophones of one phoneme, it would be preferable to preserve (more conspicuously for the reader) the more obvious representation of the glottal stop, which an h as a symbol in a syllable terminal position does not do sufficiently.

In Brasília in 1974, we also accepted an orthographic and phonemic distinction between the unaspirated and aspirated palatal/velar stops, the former as /c/ and the latter as /k/. Exceptions were to occur, however, if /c/ appeared before i and e; and in these cases qu was to be utilized instead of c to maintain the practice used in Portuguese. I have used this clumsy aspect of the Indian service/SIL orthography since 1974 but am returning now in 1989 to my pre-1968 easier utilization of /k/ and /kh/ which requires no exceptions. The use of qu before the various phonemic forms of i and e, involving their several combinations with diacritics to create different phonemes, is quite confusing. Moreover, the h in the kh visually implies and accentuates the contrasting aspiration between the two phonemes, whereas the relationship between /c/ and /k/ does not.

I also use ts for the 1974 Indian service/SlL x, Pike’s [˘,]. Although ts consists of two symbols and therefore is more clumsy, it is at least close in quality to the Canela sound in the pertinent major Western languages.

The above are the changes in my 1989 orthography and the reasons for the modifications. Obviously, any symbol can be used for any phoneme and the purely professional linguist often does not care about such matters as the symbols being generally appealing—as already representing certain sounds to certain speaking groups. But this choice of letters is an attempt to furnish a phonemic orthography that Portuguese, Spanish, and English speakers will find easy to use without requiring much adaptation on their part.

[Ap.4.c.(2)]  Comments on Brazilian Orthography

A principal difference between my orthography and a usual Brazilian’s use of letters in writing Canela exists in my use of y instead of i as a glide following a vowel to complete syllable medial length and syllable final length (itswčyyę instead of itswčiyę: out-sister-in-law, and impey instead of impei: it is good).

Jack Popjes first made the point that an i or u following a vowel is really a glide (i.e., y or w) rather than a second vowel forming a diphthong.

This y as a glide instead of i is consistent with English (say) and Spanish (soy: I am), but the Portuguese writer uses i instead (boi: bull). Similarly, following a single vowel, a syllable terminal w completes the required syllable length. Thus, more correctly, it should be katsaw (same day), not katsau; i-yapal (my-nephew), not iapal; and wapo (machete), not uapo.

This use of y and w is carried out to be phonemically correct. The Canela do not have diphthongs to lengthen the single vowel to occupy the length of time that would ordinarily be spent on double vowel length; instead they have only double vowel length, or single vowel length doubled by a vocalic glide, which terminates the syllable. Then, if y and w are used syllable terminal in this manner, they must be used syllable initial in a similar manner, though this syllable initial usage is not related to vowel length. Thus, the use of a syllable beginning or syllable terminal i and u, preceding or following a vowel, provides a wrong impression of the nature of the language. Moreover, such a syllable initial or syllable terminal i and u, as found in Portuguese, is not an allophone of /i/ and /u/ in Canela, but is rather an allophone of the separate phonemes /y/ and /w/, respectively. This practice is consistent with the 1974 Indian service/SIL-accepted orthography except that where they use j, I use y, but we all use w in the same way.

[Ap.4.c.(3)] Suggestions for Brazilians Writing Canela

In 1970 and 1971, with the help of Olímpio Martins Cruz and meteorologist António Gomes Cordeiro, I developed a way for Brazilians of the interior to write Canela: an orthography suited to their ear and their preferences in symbols. Such materials were developed because of local resistance to the particular use of y, w, j, x, h, and other letters, found in the 1974 Indian service/SIL authorized orthography.

The exposition of this Portuguese-oriented orthography will facilitate the understanding of certain phonetic sounds of the Canela language as they have to be included in their phonemic categories to write the language accurately. These clarifications are especially useful for northern Brazilians, who strikingly modify the alveolar stops t and d when found in certain environments, making them affricatives. The problem in northern Brazilian is that a t or d before an i or e is pronounced as an affricative, more like ts (tio [tsiu]: uncle) or dz (dia [dzia]: day). Consequently, /ts/ in Canela is confused with the Portuguese t before i or e so that one Canela phoneme (/ts/) is sometimes heard as one phoneme (/t/: tio) in Portuguese and then a different Canela phoneme (/t/) is heard as the same phoneme (/t/: tanto: so much) in Portuguese at another time.

A similar problem for Brazilians in general is that they tend to use i for both /i/ and /y/ and u for both /u/ and /w/ in Canela. Thus, in this orthography created especially for local Brazilians, the “one phoneme in one language for one phoneme in the other language” principle had to be violated to some extent. Our solution is to not make such a violation with respect to the Portuguese t and the Canela /ts/ and /t/ but to turn around and give way to the Brazilian ear and its need to merge the Canela /y/ and /w/ with the Canela /i/ and /u/, respectively, in all environments. At least this is a simple merging of two phonemes into one, twice, and not a phoneme-dividing violation of the “one phoneme for one phoneme” principle, which otherwise is maintained throughout the recommended orthography. A Canela /iy/ becomes i (iyapal: iapal) in a reduction of the number of phonemes represented, but a /wu/ sequence does not occur in Canela except in chanting.

[Ap.4.c.(4)]  Distinctions Not Heard by Nimuendajú and Others

The grave accent (`) is used regularly in Portuguese, so /ů/, /č/, and /ŕ/ could be accepted quite easily as ů, č, and ŕ by the northern Brazilian, but we (Cordeiro, Crocker, Cruz) wrote / as ň because it sounds more like an open o to the local ear. The problem here was that the inexperienced ear does not distinguish between /ů/ and /č/, /o/ and /, and even /o/ and /č/. When SIL specialist Sarah Gudschinsky found the younger Kaapęltůk in Brasília in 1964, even she did not distinguish /ů/ from /č/ in her entire Museu Nacional Questionário (Gudschinsky, 1960) analysis, a copy of which she later sent me. Nimuendajú often did not distinguish properly between /ů/ and /č/, and between /o/ and /ŕ/. For instance, his spelling of the Wů?tč girl should be Wč?tč and his representation of the Cakamekra tribe should be Tsookhămmë?khra, phonemically, with his hacek-cŕ correctly being tsoo, but still meaning “fox,” just as he maintained (Nimuendajú, 1946:34).

The two distinctions I often could not be sure of making by ear in words that were new to me were between /ů/ (: urucu) and /č/ (wakhč: metal, wire, machine) and between /?/ and /k/ when the latter, an unvoiced, unaspirated, velar stop, was occluded because of being followed by a stop beginning the next syllable (pa- ?prő: our-wife; pak-ti: scorpion augmentative). By the 1970s, however, the younger Kaapęltůk could make these distinctions for me, after noting the positions of his tongue and glottis.


Appendix 5 Concept of “Today”

After carrying out considerable research on the characteristics of the Canela unit “today,” and after relating Canela terms and expressions with solar events, I was able to reconstruct precise periods of time, around the clock and for several days ahead and behind, which follow their view of past and future Canela “todays.”


The word for “today” is ita-khăm (this-in: this one [day] we are in) phonemically and itá-khămmă phonetically [Ap.4.a.(6)]. It is, however, simply written as itakhăm.

If the speaker is talking during the daytime (i.e., between sunrise and sunset), the itakhăm she or he is referring to includes three periods: (1) the night before (sunset to sunrise), (2) the daytime she or he is in (sunrise to sunset), and (3) the next night (sunset to sunrise). Thus, itakhăm (today) for the Canela person speaking in the daytime (sunrise to sunset) refers to a period of 36 hours and its time markers are sunset (půt tsŕl: sun enters) and sunrise (půt katol: sun comes-out) (Table 6).

If the speaker wants to talk about the adjacent nights, she or he says itakhăm katswa ri (today’s night there), letting the context indicate which one is meant (Figure 12). If she or he wants to specify which night is meant more precisely, itakhăm katswari amu të (today’s night away moves: last night) indicates the past one, while itakhăm katswari aypęn të (today’s night this-way moves: tonight) indicates the approaching one. (Katswa and katswa ri have essentially the same meaning: night.)

The sequence of “todays” is seen as being linear, the past ones going away from the speaker, and the future ones coming toward her or him—to eventually pass by her or him and then go away in turn [V.A.6].

The term for tomorrow is apë-?nă (daylight-on: tomorrow), which is the next day, from sunrise to sunset, when the speaker is talking in the daytime. “Tomorrow night” is apë?nă katswa ri (tomorrow’s night there), from sunset to sunrise. Apë?nă hakpûû-mă (tomorrow’s backed-onto: tomorrow’s beyond day: the day after tomorrow) means the day after tomorrow, and apë?nă hakpûû-mă katswali (tomorrow’s backed-onto’s night: tomorrow’s beyond day’s beyond night: the night before the day after tomorrow) is the night just before the day after tomorrow. It is interesting that while the Canela unit of today that the speaker is in lasts 36 hours, the Canela tomorrow (one daytime and one nighttime) and the Canela day after tomorrow (one daytime and one nighttime) last only 24 hours each.

The term for yesterday is i?-nő khăm (its-other in: in today’s other day), and yesterday’s night is i?nőkhăm katswari. The day before yesterday is i?nőkhăm hakpûû-mă (yesterday’s backed-onto: yesterday’s beyond day), and the night just before the day before yesterday is i?nőkhăm hakpûûmă katswari (yesterday’s backed-on-to’s night; yesterday’s beyond day’s beyond night). The Canela “yesterday” and their “day before yesterday” are each 24 hours long. Using the Canela concepts of the future coming to the speaker (with the speaker remaining stationary) and the past going away from the speaker, yesterday’s night is just beyond yesterday, and the day before yesterday’s night is just beyond the day before yesterday. Similarly, tomorrow’s night is just beyond tomorrow and the day after tomorrow’s night is just beyond the day after tomorrow.


If the speaker is talking during the night, “tonight” (that is, her or his “today” for that time of day) is katswa itakhăm (katswa ita-khăm: night this-in: during night’s today or during the nocturnal today). This unit of time, katswa itakhăm, covers a different block of hours than is covered by a speaker talking during the daylight, but it also is 36 hours long.

[Ap.5.b.(1)] When Talking before Midnight

If the speaker is talking before what is approximately midnight (katswa pikapôn: nighttime split-in-half), the unit of this “nocturnal today” includes (1) the whole night (from sunset to sunrise) in which she or he is speaking, (2) the preceding daytime (from sunrise to sunset), and (3) the nighttime before (i.e., beyond) this preceding daytime (from sunset to sunrise). But if the speaker is talking after what is approximately midnight, the unit for this other, later-in-time nocturnal today includes (1) the whole night (from sunset to sunrise) in which she or he is speaking, (2) the following daytime (from sunrise to sunset), and (3) the night after (beyond) this following daytime (sunset to sunrise). Before midnight the speaker looks backward into the past to visualize the main portion of her or his nocturnal today, while after midnight she or he looks forward into the future. (It is notable that the three relatively precise Canela time markers are sunrise, sunset, and midnight. Midday is not held as a time marker.)

If it is seven to ten o’clock in the evening, the Canela will say katswa itakhăm půt-khăm (night’s today’s sun-in-time: this nocturnal today’s late afternoon) for the late afternoon, or more simply itakhăm půtkhăm or půtkhăm. From seven to ten in the evening, they also will say itakhăm půt-te-?kapaa-khăm (today’s sun-has-blocked/pended-time: this nocturnal today’s sun-pending [from noon] time) for the (later in time, now going away) early afternoon. Again, from seven to ten in the evening, itakhăm meio-dia-khăm-půt yû (today’s mid-day-in-sun-sits: this nocturnal today’s noon) is the expression for the preceding noon period. Similarly, from seven to ten in the evening, the speaker says itakhăm irŕŕrŕn (today’s morning: this nocturnal today’s morning) for the preceding morning 10 to 12 hours earlier, but now going away. Then, still from seven to ten in the evening and talking about still earlier time (now going still farther away) in the same nocturnal today, the speaker can say katswa amu të itakhăm (night away moving, today: the night in the front, or the early part, of this nocturnal today; or, this nocturnal today’s moving away night).

During the evening period of a night, to refer to the day before the preceding (the moving away) night, the speaker can say i?nőkhăm. Then this preceding day’s preceding night is i?nőkhăm katswari, and its still earlier (moving away) day is i?nőkhăm hakpûûmă.

[Ap.5.b.(2)] When Talking after Midnight

If it is during the early dawn period of a night, the awkati-tsŕ?wčl (dawn-period toward), the speaker will say katswa ita-khăm (night this-in: in this nocturnal today’s night) for the preceding evening and for the whole of the night (sunset to sunrise) that she or he is in. Then the earlier periods of the arriving day are itakhăm půt katol tsŕ (today’s sun come-out time: sunrise) (or just půt katol) and then itakhăm irŕŕrŕn (or just iráŕrŕn), meio dia khăm půt yû, půt-te-?kapaa-khăm, půtkhăm, and půt tsŕl (sun enters: sunset), each further away upstream in the arriving future. Then a?pre?prel (dusk) precedes. Then, the coming night, which is still part of this nocturnal today, is katswa aypęn të, itakhăm (night this-way moving of-today: the coming [arriving] night of this nocturnal today).

During the dawn period of a night, to refer to the day before the coming night, the speaker can say apë ita (tomorrow this-one: tomorrow), and to the night immediately preceding it, the speaker can say apë-?na katswari (tomorrow-on its-night: tomorrow night). As for the day beyond the day beyond the coming night (coming from upstream), he can say apë?nă hakpûûmă, meaning day after tomorrow, and this day after tomorrow’s night is apë?nă hakpûûmă katswari.

[Ap.5.b.(3)] When Talking about the “Short” Side of the Nocturnal Today

If the individual speaks during the night, but addresses what might be called the “short” side of the nocturnal today (the “long” side extends two 12 hour periods beyond the night either into the future or into the past), she or he may talk in the evening (before midnight) about an event about to take place during the next arriving day just beyond sunrise. This is also the next day for the Canela, being beyond the nocturnal night of the speaker. Thus, she or he has to say apë?nă irŕŕrŕn (tomorrow morning) when referring to occurrences about to take place the next morning, only about seven to ten hours in the future, from the individual speaking in the evening. In contrast, for the individual talking in the short side of the nocturnal today but still speaking during the evening, the simple irŕŕrŕn, or the more precise but unnecessary itakhăm irŕŕrŕn, refers to the preceding morning 12 to 18 hours earlier, which is part of the speaker’s nocturnal today.

Similarly, if the speaker is in the dawn period of the night and refers back to the late afternoon sing-dance about 10 hours in the past, she or he must use i?nőkhăm půtkhăm (yesterday afternoon) or the listener would think the speaker was referring to the coming late afternoon sing-dance of the next late afternoon about 14 hours in the future. Terms for the extension of references to further-away days on the short side of the nocturnal today are similar to those found on the long side.


For expressing the number of days ahead or behind, the speaker can double and triple, etc., the expression hakpûûmă (i.e., apë?nă hakpûûmă hakpûûmă: the day after the day after tomorrow), but in modern times the days of the week in Portuguese are used. Moving on to further lengths of time, two Sundays ahead is Domďk píyakrut (Domingos [Sundays] two), three Sundays is Domďk kręę (Sundays three), and four Sundays is Domďk te quat (Sundays have quatro [four]), which is also putwrč putsęt (moons one: one month).


Appendix 6   Sources of Data

Material for Part II was collected in a number of ways. Information on the various tribes that speak Gę and on the population numbers of these tribes came largely from the literature. Canela definitions of their language context, however, found in the meanings of the terms and mëhïï (both meaning “Eastern Timbira”), came from fieldwork with experienced research assistants. Canela and Apanyekra population numbers come from my own census taking, conducted in 1960, 1964, 1966, 1969, 1970, 1975, and 1979 for the Canela and in 1959, 1970, 1971, and 1975 for the Apanyekra.

Geographical materials were gathered partly through a literature search and partly through observations made while traveling by horse in the backland region around and between the two reservations. Culture contact materials were collected while visiting the backland communities of Bacabal, Sítio dos Arrudas, Papagáio, Serrinha, Barreira, Ribeirăo, Jenipapo do Resplandes, Leandro, Curicaca, Ourives, Mucună and others, largely during the late 1950s [II.d] (Map 3). I also had long talks with key backlanders about their communities when they visited the Canela villages of Ponto, Baixăo Pręto, and Escalvado.

With the exception of “The Eastern Timbira” (Nimuendajú, 1946), materials for the historical chapter come only to a small extent from historical documents. Material in “Indigenous Accounts” [II.B.1] comes mostly from special work with my research assistant council [Pr.2], and material in “Acculturation” [II.B.2] comes from both my research assistant council and my daily notes taken over the 22-year period of my research. “The Eastern Timbira” was crucial here. Certain Indian service agents, such as Antônio (Gato) Ferreira do Nascimento, Virgílio Galvăo Sobrinho, and Júlio Tavares, also provided crucial information. Data on the history of Barra do Corda was largely furnished by its mayor, Lourival Pacheco, and the city’s meteorologist, Antônio Gomes Cordeiro. I was allowed to tape an interview with them, and the meteorologist provided a written document (Table 1). Jaldo Pereira Santos [I.H] corrected and added information to a long report I had produced using these materials. Some data (on Maranhăo and Piaui) also came from the Encilopédia dos Municípios do Brasil (Volume 15:70).

Some climatic cycle data are based on my 1972 article, but I recorded temperature and humidity from day to day during certain periods, especially May through August, using a maximum-minimum thermometer and a wet-bulb dry-bulb relative humidity instrument (a psychrometer). The meteorologist furnished the best data in the form of a table (Table 1). The weather shifts which were obvious and notable were recorded in my daily notes.

The environmental cycle material was developed largely with my research assistant council in 1979. Its members supplied the key concept of regeneration and the many descriptive phrases in Table 2. They also supplied the principal materials for the economic cycle (Table 3), debating among themselves about the monthly ranges in planting and harvesting their various crops. Although I visited Canela and Apanyekra farms (Plate 12c,d) a number of times, I studied their techniques for planting and harvesting only once, and that was among the Apanyekra (with Khrúwapu) where such techniques are more traditional. The macro elements of the economic cycle (Table 3), however, were familiar to me just from having spent many years with the Canela and from taking daily notes.

Material for the annual festival cycle (Table 4) was so familiar that I needed to consult my field notes only to check an occasional point or the spelling of a certain word. I had seen the festivals many times and specialized in recording them. Over two months were spent with my research assistant council in the fall of 1978 studying all the festivals act by act with xeroxed notes brought from the Smithsonian of all earlier festivals.

Similarly, the material for life cycle rites was very familiar, but additional data came from the notes of many rites observed over the years and some data came from research assistant council sessions. Crucial points like the purpose of the girl associate in men’s societies [II.D.2.e.(3)] and the nature of a woman’s “free” period in life [II.D.2.g] emerged during these sessions.

Description of the daily cycle, which I experienced almost two thousand times, came largely from my memory, but the phrases for describing the times of the day (Table 6) were provided by the research assistant council in 1975, as was the material on the Canela concept of “today” (Figure 12).

In “Recreation” [II.F] again, the familiar material came largely from memory, but the terms in Canela had to be checked. Data on body decoration was supplied by individual research assistants during the winter of 1960.

Almost all the information on the artifacts came from a special study of the associated social practices (by whom they were made, for whom they were made, how they were used and painted, how they were disposed of, and what they were meant to signify) carried out with my research assistant council for well over a month during the fall of 1969.

Turning to Part III, data for the chapter on socialization [III.A] were collected sporadically through daily observations in 1957, 1958, 1959, and 1960. Socialization was my major focus, besides festivals, of the late 1950s. In the winter and spring of 1960 I worked intensively with a number of research assistants (not quite yet my research assistant council) to systematize my understanding of the socialization process.

Material for the psychological chapter [III.B] was collected sporadically through the years in daily notes. Since anything psychological is a special interest of mine, these notes are extensive. The main body of material on polarity, however, was collected in the summer of 1979 in collaboration with Steven Schecter, Director E. Richard Sorenson’s cinematographer from the Smithsonian’s National Human Studies Film Center [Ap.3.d]. We had edited a film from his 1975 footage called “Individuality in Solidarity,” based on these Canela dualistic pairings, and we were searching for any number of other pairings (or polarities in behavior) that we could identify. A number of them are in this chapter.

Material for the chapter on socioceremonial units [III.C] came from various places but mostly from my research assistant council meetings in the fall of 1978, when we worked on festivals for two months. Material for the articles on this topic (W. Crocker 1977, 1979) came from work done on festival groups, especially the “river”-oriented ones [III.C.8.a], during the winter of 1975. Much of the analysis of these materials was carried out during a two-week rest stay that winter in the house of Jack and Josephine Popjes on the SIL base in Belém.

Material for the chapter on political matters [III.D] came largely from daily notes, since no research assistant council meetings were held on this topic [III.D.1.g.(1).(c)]. Some information also came from the work on festivals with my research assistant council in the fall of 1978.

The basic materials for kinship and other relationship systems [III.E] came from work with research assistants in the late 1950s and the summer of 1966. The quantification on a larger scale of what individuals call each other, however, was obtained while taking censuses in 1969, 1970, and 1971 when the major effort in the field was on gathering this kind of material. Solutions to the problems between the data collected and ethno-kinship theory were worked out in the fall of 1974 and the spring of 1975 with my research assistant council.

Material for the marriage chapter was collected in 1970 when I invited about eight research assistants into my field office each day for several weeks, each individual representing different parts of the village circle, to talk about the sexual and marital histories of 120 married or formerly married persons. Much of this material was presented in a paper read before the American Anthropological Association in New Orleans in 1973. This and the other papers in the New Orleans symposium on marriage in lowland South America were taken by Kenneth Kensinger, as editor, to form the basis for a book on this topic. Publication of this book was delayed, so in 1975 I went through the same process in the field, this time with about 18 research assistants to represent more thoroughly the different parts of the village circle (W. Crocker, 1984a).

The 1970 material on marriage was considerably modified and improved through this procedure. I also worked on critical points, such as refining the information on the stages of becoming married and on clarifying the circumstances of unencumbered women and men [III.F.4.b.(1),(2)], with my research assistant council in 1970, 1975, and 1979.

Materials for the chapters on tribal festivals and individual rites [IV.A,B] were recorded in the field while seeing them many times, and they were also collected from my research assistant council in the fall of 1978. For obtaining data for the chapter on oral history and cosmology, I worked with the younger Kaapęltůk and Rőő-re-?hô in the evenings for some months during the winters and springs of 1975 and 1979 to translate myths into Portuguese [Ap.3.b]. With these transla­tions in mind, I presented numerous questions about the mythology and war stories of the Canela to my research assistant council for them to debate and resolve. Some of the results were published in Portuguese (W. Crocker, 1978) but are presented here in English with additional materials.

Data for “Shamanism,...” [IV.D] comes from many sources and periods of my research. General material on shamanism was collected from older Kô?kanăl, a shaman (kay), in Baixăo Pręto in the late 1950s. Specific information (all taped) was obtained in a series of private interviews in Canela with the younger Mďďkhrô (kay) in 1975 (Plate 70c). Material in the form of questions and answers was obtained from Tsůůkhč (Plate 68c) during the summer of 1979 while he attended my research assistant council meetings. [2003: Special material—the only data on shamanic journeying—was obtained in 2001 from the kay Iromkuukhre (Pedro in the video Mending Ways).]

Material on pollution and medicine [IV.D.3,4] was obtained informally in the late 1950s through discussions with research assistants about restrictions against eating certain foods and having sex at certain times. I worked extensively on these topics with my research assistant council during the winter of 1964 and published some of these data in papers presented at the Congress of Americanists in Stuttgart in 1968 (W. Crocker, 1971a, 1971b).

Information on “transformations” of persons into one form or another [IV.D.6] comes only from the study of myths and war stories and consequent research assistant council discussions, but data about “affirmations” (positive chanting) [IV.D.5] comes solely from individual work with the younger Kaapęltůk (Figure 51) in the fall of 1978.

Data for Part V were collected during the winter, spring, and early fall months of 1979 through intensive work with my research assistant council members but only during the mornings. Usually, three or four research assistants helped on the weekdays [I.G.11,13,14, and others] and two on Saturdays [I.G.2 and Tel-khwčy] as well as the younger Kaapęltůk at both times. For other topics, I worked with my council in the afternoons as well as the mornings, but for cognitive studies such as these, I spent the afternoons reviewing notes, rethinking issues, and preparing for the next day. Moreover, besides taping group opinions and decisions in the mornings, I also wrote down and diagramed much of what was said. Then during the afternoons, I went over these written materials, taped them as well as additional thoughts on current issues, and red underlined matters to be brought up or retested the next morning. I did not work on these cognitive matters in the evenings being aware that clearer vision and greater awareness was necessary. (These evenings were devoted to recording myths and songs and to family activities.)

When I read other ethnographies, I note the exegesis reported with interest and envy, particularly for ceremonial materials. Canela research assistants were rarely able to provide exegesis. They simply said of a ceremonial activity when questioned, mam më nkętyę nkaakaa tsŕ ?khôt (early Timbira ancestor’s breath thing following: it is traditional), or é dos antigos (it-is of-the ancients). Only by 1978, and after my considerable encouragement, did the younger Kaapęltůk begin to “loosen up” and apply his imagination to occasional meanings behind words in songs or, for example, to why the Little Fox was so furtive in the contributing-father rite [IV.B.2.c.(5)]. But even Kaapęl was not very expansive in this case and in most others. Traditional interpretations rarely existed and individual ones were amyiá-?khôt (self-following) [III.B.1.k.(3)] and therefore of questionable worth and considerable risk, constituting a kind of disobedience [III.D.1.a.(2)].This is why the research assistant council was so necessary. Working together in groups, they remembered much more about almost any traditional matter and had far more courage to go i?-khay nă (it-off condition: “wrong”) about matters for which little traditionally expressed thinking existed [V.A.1].

[2003: By 2001, the younger Canela research assistants, such as Piyat and Kää-re, were far better at coming up with exegesis. Either they had more imagination—maybe due to their better educations—or they were less afraid to be individualistic. The younger Kaapêltùk had difficulty in providing the interpretations obtained from the materials in Part V. "The ancestors did not say this," he would say. However, once he became used to thinking in terms of "underlying" emic structures, he became freed from his ancestors and enjoyed the new "game."]

<<back   table of contents  next>>
[top of page]