Wany Bernardete de Araujo Sampaio Centro de Estudos da Linguagem Grupo de Estudos em Culturas, Educa

1
Wany Bernardete de Araujo SampaioCentro de
Estudos da Linguagem Grupo de Estudos em
Culturas, Educação e Linguagens.Universidade
Federal de Rondônia Brasilwsampaio_at_unir.brThe
Tupi-Kawahib Languages A Phylogenetic
Systematics based comparative study 1

• 1 Síntese de Tese de Doutorado em Lingüística
  (2001).
• Universidade Federal de Rondônia. Brasil.
• Orientador Prof. Dr. Mario Alberto Cozzuol

2
PRESENTATION

• Comparative study of Tupi-Kawahib languages,
  Tupi-Guarani Family.
• Tenharim, parintintin, juma, diahoi, karipuna,
  uru-eu-uau-uau e amondava Kayabi (like a
  hypothesis to be tested)
• Initial Studies
• Masters Degree - Lingüistics (Sampaio 1997),
• Comparative analysis phonologic end lexical
  levels (parintintin, tenharim, uru-eu-uau-uau e
  amondava)
• Phonetic similarities 80.875
• Phonemic similarities 86.375
• Conclusion these languages are varieties of a
  single language (dialects).

3
A Classification Problem
Lingüists Family Language Dialect
Rodrigues Tupi-Guarani Pano Parintintin Uruewauwáu Karipuna Diahói, Juma, Parintintin (Kagwahiv) e Tenharim
Mellati Tupi-Guarani Pano Kawahib Karipuna Parintintin, Diahói, Tenharim, Juma, Kayabi
Betts e Pease Tupi-Guarani Parintintin Uru-eu-uau-uau Parintintin e Tenharim. Uru-eu-uau-uau e Amondawa

4
OBJECTIVES

• To compare the percentage of similarity among
  tenharim, parintintin, uru-eu-uau-uau, amondava,
  diahói, juma, karipuna and kayabi
• To analyse the phylogenetic relationships among
  these laguages
• To discuss previous classifications of
  Tupi-Kawahib languages
• To propose a phylogenetic classification of
  these languages, based on lexical data
  comparison, using Phylogenetic Systematics
  methodology.
• To elaborate a hypothesis about the kinships of
  the Tupi-Kawahib languages, observing the
  distancing in the phylogenetic sequence
• To test the hypothesis the phylogenetic
  relationships point to the existence af a common
  ancestor to the members of one linguistic group.

5
METHODS

• a) Comparative Linguistics phonostatistics and
  lexicostatistics
• To measure the percentage of similarities among
  the languages. (intercomprehension)
• b) Phylogenitic Systematics (Comparative
  Biology)
• To measure the percentage of similarities among
  the languages.
• To define the sequence of differentiation among
  the probable Kawahib varieties, which allows us
  to propose a phylogenetic classification for such
  varieties
• To contribute to the discussion on the
  geographical dispersion of this linguistic group.

6
CONTRIBUTIONS

• Scientific
• a new discussion about the proposed
  classification for Tupi-Kawahib languages
• explore a different methodology, the Phylogenetic
  Systematics, which features a practice of
  interdisciplinarity
• Social
• indigenous school education
• more certainty in the proposition of learning
  resources, especially those for reading and
  writing in mother tongue.

7
ORIGIN OF DATA

• CATEGORIZATION INTO 2 GROUPS
• Languages of the internal group
• Parintintin, tenharim, uru-eu-uau-uau Betts
  Pease
• Uru-eu-uau-uau, amondava Sampaio da Silva Sinha
  et al.
• Diahoi and karipuna Sampaio and Bezerra
• Juma Betts Pease (1977) and Abrahamson (1963)
  
• Kayabi Dobson (1988).
• Languages of the external group
• Wayampi Jensen (1984)
• Tembé Sampaio (1995) (Pará - Emilio Goeldi
  Museum).
• Bibliography about languages in question.

8
GEOGRAPHICAL LOCALIZATION

• People of internal group
• a) Tapajós-Madeira parintintin, tenharim,
  diahoi, kaiabi
• b) Juruá, Jutaí, Purus juma
• c) Rondônia karipuna, uru-eu-uau-uau and
  amondava
• People of external group
• Area Southeastern of Pará tembé
• Area Amapa - North Pará wayampi

9
CULTURAL COMMONALITIES

• Social Organization based on exogamous
  marriage and division into two clans (moitiés)
  called by name of birds.
• Uru-eu-uau-uau and amondava
• Mutum (Mutua) and Arara (Kanindea)
• Tenharim, parintintin and diahoi
• Gavião (Kwandu) and Mutum (Mytu)
• Karipuna
• Tucano (Tukanahua) and Mutum (Mytu)
• Juma
• Mutum (Mutua) Arara (Kanindea)

10
Social organization (by a child amondava)
11

• patrilineal descent
• matrilateral complementary filiation
• levirate - inheritance of the wife by the brother
  of her deceased husband
• fraternal polyandry husband can loan his
  wife(s) to his brother.
• patrilocal residence
• they dont use tobacco at the time of the first
  contacts
• none of them grows tobacco currently.
• facial tatoo
• kinship structure determines the onomastic of the
  group.
• burial of the dead
• a good spirit Tupanangá and an opponent
  Mbahira.
• use of flutes made with taboca
• Yrerua ritual dance
• food and beverages, rituals (mbotawa and chicha)

12
Yrerua
13
METHODS OF COMPARATIVE LINGUISTICS

• CORPUS
• 10 lists with 200 words (10 languages) 2.000
  words
• Inspection Methodology
• Phonetic similarities among the compared
  languages.
• Sincronic point of view
• degree of intercomprehention between languages
• the more these languages are similar to each
  other, the more probable it is that they
  originated from the same language.
• Methods used in computing the similarities
• Deibler e Trefry (1963, apud Sanders, 1986 35)
• Carrol Dyen (1962 id. 36).

14
Deibler and Trefry method

• Traditional comparative analyses
• Compare lists of words in pairs, one word at a
  time, a phoneme at a time.
• Principle of analysis
• association of phonostatistics with
  lexicostatístics.
• Computing data and calculating the percentage of
  similarities
• equal words (4)
• words with only one different phoneme(3)
• words with two different phonemes (2)
• words with tree or more different phonemes, but
  considered cognates(1)
• not cognate words (0)
• not available data (-) (not considered in
  determining the comparison).

15
Percentage of phonetic similarities and degree
of intercomprehension
PAR URU AMO KAR DIA JUM KAY TEM WAY
TEN 85.8 67.1 65.7 71.0 57.3 59.1 45.2 43.0 45.1
PAR 69.8 67.0 68.0 54.6 59.3 42.7 41.2 45.3
URU 86.1 70.4 52.2 60.0 41.0 37.1 36.6
AMO 85.1 54.1 60.4 40.2 37.6 36.8
KAR 34.5 62.1 44.9 40.8 45.4
DIA 47.3 36.7 34.7 35.2
JUM 43.9 40.1 41.5
KAY 52.7 51.8
TEM 52.5
16
Phonostatistical Results

• Criterion to determine the intercomprehension
  50
• Tenharim, parintintin, uru-eu-uau-uau, amondava,
  karipuna, diahoi e juma belong to a single
  linguistic group.
• Distance of 50
• Karipuna e diahoi (34.5)
• Juma e diahoi (47. 3).
• Are there 2 groups into the internal group?
• Kayabi, tembé and wayampi appear as belonging to
  another group of interrelated languages.

17
Carrol Dyen method

• Consider the global similarities.
• Principle of analysis
• lexicostatistics.
• Comparative basis
• The language which has the highest number of
  data. (Tenharim)
• Criteria for determining the cognates formal and
  observable similarities between the words of
  each set of words.
• Each set of cognates was taken separately and
  were assigned code numbers equal for all cognates
  of the basis language.
• After the encryption of data, we had a count of
  codes, to know the total amount allocated for
  each code, considering the sum of the values in
  each language.

18
Percentage of cognates by global similarity
Code Quantity Percentage
1 1735 86,96
2 163 8,17
3 68 3,40
4 19 0,95
5 8 0,40
6 2 0,10
Total Geral 1995 100,00
19
Lexicoestatistical Results

• Considering the global similarities
• Code of higher occurrence 1
• similarity among the linguistic data compared
  86,10.
• This suggests that all these languages have high
  degree of intercomprehension, belonging,
  therefore, to a single group.

20
Phonostatistics X Lexicostatistics

• Each one points in a different direction
• Phonostatistics suggests the existence of at
  least two distinct groups of languages in the
  internal group
• Lexicostatistics suggests that all the languages
  (internal and external group) belong to a single
  group.
• Original doubt are there or not a Tupi-kawahib
  linguistic group?
• the percentage of lexical or phonetic
  similarities between languages is not enough to
  measure the degree of intercomprehension between
  them
• throughout the history of Linguistics, these
  criteria have been used without major challenges,
  both for the classification of varieties
  language, and for the development of studies
  aimed to the classification and historic
  reconstruction of languages.

21
PHYLOGENETIC SYSTEMATICS

• Compare the similarities of the characters of
  languages among themselves.
• Coding number of words assignment of a numerical
  symbol for each state of a different character in
  the series of transformation the codification
  can inform about the natural ordination between
  the states of a character
• The coding was optimized by a computer program of
  cladistic analysis.
• The numerical matrix was then submitted to the
  ordination of the characters, weigh them and
  choose a type of parsimony in the optimization of
  the characters shared.
• For the parsimony analysis of the data, we used
  the program Hennig 86 (James Farris, 1988) and
  the program Tree Gardner 2.2 (Tiago Courrol
  Ramos, 1997).
• The analysis provided the search for a tree
  representation, which reflects the optimization
  of the matrix number
• The tree obtained was adopted as the best chance
  for the interpretation of the phylogenetic
  relationships among the languages studied.

22
Cladogram
23
PHONOSTATISTICAL PHENOGRAM The top scale
indicates that the closer to 100, the greater the
degree of similarity between the languages. .
24
LEXICOSTATISTICAL PHENOGRAM The top scale
indicates that the closer to 0.00, the greater
the degree of similarity between the languages
25
Analysing the diagrams

• the results represented by the phylogenetic tree
  are compatible with the both results of
  phonostatistic and lexicostatistic phenograms,
  when submitted to computer analysis of clusters.
• phonostatistics and lexicostatistics analysis,
  using traditional methods of Comparative
  Linguistics, have brought us different results
• Phonostatistics two different language groups
• Lexicostatistics only one linguistic group.
• Phylogenetic analysis shows that all languages of
  the internal group belong to one lineage.

26
FINAL CONSIDERATIONS

• The languages Tupi-Kawahib - among those studied
  - are juma, tenharim, parintintin, diahoi,
  karipuna, uru-eu-uau-uau and amondava
• Considering the evolutionary sequence, juma was
  the first language to separate from the others
  within the group Tupi-Kawahib, followed by
  tenharim and parintintin uru-eu-uau-uau and
  amondava karipuna and diahoi
• All these languages, belonging to the group
  Tupi-Kawahib, originates from a single ancestral
  language
• All ethnic groups, users of languages here
  postulated as Tupi-Kawahib, come from a single
  ancestral ethnic group that has been split
  successively in the course of time
• The divisions in the group contributed to the
  divisions in the language.
• The synapomorphies between the languages of the
  group Tupi-Kawahib allow us to refine the
  hypothesis about their relations of kinship,
  classifying them as languages closely linked
  evolutionarily.

27
THANKS

• University of Portsmouth
• Dr. Chris Sinha
• Lund University
• Dr. Alf Hornborg
• University of Rondônia