Question prosody : an African perspective

1
Question prosody an African perspective
Conference ESF Tone and Intonation, Santorini,
Greece, 9-11 September 2004

• Annie Rialland
• Laboratoire de phonétique et phonologie,
• UMR 7018, CNRS/Sorbonne-Nouvelle, Paris

2
Question Prosody typology Background

• A widespread  universalist view 
• High-pitched question prosody is a (near-)
  universal.
•  it is almost invariably the case that high or
  rising pitch signals the former question
  whereas low or falling pitch, the latter
  statement 
• Ohala (1983)  Cross-linguistic Study of Pitch
  an Ethological View  Phonetica 40

3

• High-pitched question intonation is
  viewed as a grammaticalization of a natural
  tendency according to the  Frequency code 
•   the pattern is too widespread to be
  explained by borrowing, descent from a common
  linguistic source, or chance. It follows that
  there is something common to all human speakers,
  at all stages in history, which creates this
  phenomenon .
• Ohala (1984),   An ethological perspective of
  common cross-language utilisation of F0 of
  voice , Phonetica 41

4
What are the databases on Question prosody?

• Hermann E. 1942.  Problem der Frage. Nachrichten
  von der Akademie der Wissenschaften in
  Göttingen . Philologisch-Historische Klasse, Nr
  3-4
• 175 languages
•  found without exception a tendency to
  higher pitch somewhere in the
• utterance  Bolinger (1978)
• Ultan R. 1969.  Some General Characteristics on
  Interrogative Systems , Working Papers in
  Language Universals 1.
• 53 languages
•  same results high pitch without exception 
  , 71 rising intonation
• (quoted by Bolinger 1978)
• Bolinger D. 1978  Intonation across languages 
  in Universals of
• Human languages , J. Greenberg (ed.)
• 41 languages
• (all with h or rise, except Papago,Itonama,
  Quechua for which  no h or rise has been
  reported )

5

• Some languages have been added by various authors
• Chickasaw and Roermond Dutch
• by Gussenhoven (2002)  Intonation and
  Interpretation Phonetics and phonology , in
  Speech Prosody 2002
• Currently
• The most important databases (Hermann, Ultan) are
  almost unavailable. Their conclusions are known
  through Bolinger (1978).
• We dont know which languages are included or,
  consequently, how many African languages have
  been taken into account

6
Our in-progress African Database
7
Languages currently in our database 74

• NIGER-CONGO (60)
• 16 Gur languages
• 5 Mande
• 2 Atlantic
• 6 Benue-Congo
• 7 Kwa
• 3 Ijoid
• 2 Kru
• 1 Adamawa-Ubangi
• 1 Bantoid
• 17 Bantu
• KHOISAN (1)
• AFRO-ASIATIC (8)
• 6 Chadic
• 2 Cushitic
• NILO-SAHARAN (5)

8
Prosodic markers

• HIGH-PITCHED MARKERS
• cancellation/reduction of downdrift, register
  expansion
• raising of last H(s) (not necessarily
  sentence-final)
• cancellation/reduction of final lowering
• final H tone or rising intonation (final H)
• final HL melody
• final L tone or falling intonation (final L)
• final polar tone or M tone
• length (VV or V)
• breathy termination
• cancellation of penultimate lengthening
• open vowel

9
Languages without any type of high-pitched
melodiesin RED on the hand-out

• with L or L
• with length
• with breathy termination
• with open vowel
• in any combination
• 27 languages
• 15 Gur, 2 Mande, 4 Kwa, 2 Benue-Congo, 1 Ijoid,
• 1 Kru, 2 Bantu
• There is no  near totality  of languages with
  high-pitched melodies.

10
Presentation of the prosodic markers
11
Register expansionCancellation/ reduction of
downdrift
1

• Wolof examples

Statement
H-
H
Peer and Samba came yesterday.
L-
L-
L-
L
H-
L-
L-
H-
Peer and Samba came yesterday?
L-
L
Question
Rialland A. Robert S, 2001,  The intonational
system of Wolof  Linguistics 39-5

• Number of languages with this marker in our
  database 21

12
Raising of last H(s)
2

• Domains
• Last H tone (not necessarily sentence-final)
• Last sequence of H tones
• H tones of the last phrase
• Languages with raising of last H(s) in our
  database
• 2 Bantu languages
• 4 Chadic languages

13
Cancellation/reduction of final lowering
3

• Domains
• last syllable
• last word
• Languages with cancellation of final lowering in
  our database
• 1 Bantu language
• 1 Kwa language
• 1 Cushitic language

14
Final tones

• 4 H or H in 19
  languages
• 5 L or L in 31 languages
• 6 HL in 3 languages
• 7 polar tone or M tone in 4
  languages

15
Differences between T and T
Gulmancema examples
 an arrow   an arrow? 
L H L H H
statement question
H - higher than H - sentence-final
target L - at the bottom of the pitch range
But differences between T and T are unclear in
many descriptions.
16
Non-melodic markers

• 8 Length VV, CC or V 22
  languages
• 9 Breathy termination 5 languages
• 10 Cancellation of penultimate lengthening 2
  languages
• 11 open vowels 18
  languages

17
Lengthening and breathy termination

• Moba examples (Togo)

Statement
 Stones. 
Question
 Stones? 

• Lengthening
• Prolongation of F0
• Intensity decrease

18
Statement
airflow decreases
Question
airflow increases opening of the glottis
From Rialland A. , 1984, "Le fini/l'infini ou
l'affirmation/l'interrogation en moba (langue
voltaïque parlée au Nord-Togo), Studies in
African Linguistics, supp. 9
19
The melodic contour stretches out the tone
realizations
 Beans.   Beans? 

• In Moba, question prosody
• no specific tone or melodic contour.
• lengthening
• breathy termination
• base form of words (without truncation or
  metathesis occurring elsewhere)

20
Ncam (Togo) examples
 It is equal.   Is it equal? 
 a slave   a slave? 

• Falling intonation.
• lengthening
• intensity decrease

21
In some contexts, an open vowel a is added or
replaces the last vowel
 the child   the child? 

• In Ncam, the yes-no question markers are
• falling intonation
• lengthening
• breathy termination
• a vowel a in some contexts

A  lax  prosody
22

• Distribution of  lax prosody  characteristics
• Length and breathy termination
• Gur languages Moba, Ncam, Akaselem, Moore,
  Gulmancema
• Hausa
• but significant lack of data
• Low tone or falling intonation
• 29/60 Niger-Congo languages (few Bantu 2/17)
• 2/24 other families
• open vowels
• in all families
• -à is widespread.

23
Proposal a  lax  question intonation

• Besides  high-pitched intonation   with tension
  and raised larynx ,
• I propose a  lax  question intonation,
• with various facets
• - low pitch
• - breathy termination
• - open vowel

24
Cancellation of penultimate lengthening

• Zulu
•  The interrogative clause loses its length on
  the penultimate syllable  P.C. Talgaard and
  Bosch (1988)
• ukalile  he/she cried 
• ukalile  did he/she cry? 
•  The end of an interrogative sentence is
  generally pronounced with a higher tone or
  voice. 

25
Conclusions

• Question prosodies without any type of
  high-pitched correlates are not just exceptions
• They are widespread in Africa
• almost all of the Gur languages
• many Kwa, Mande, Kru languages
• some Benue-Congo, Ijo languages
• a small number of Bantu languages
• Our proposal
• besides the well-known set of high-pitched
  question markers  with tension and raised
  larynx , there is an opposite set of question
  markers with  lax  features
• low pitch (L or L)
• breathy termination
• open vowel

26
Further research

• A study of the relationship with statement
  intonation
• one of the difficulties in many descriptions,
  it is difficult to know whether statement
  intonation has a L or not.
• Note that in our database, there are only two
  examples of a reversal of the  expected 
  relationship between statements and questions
• Nembe (Ijoid), Williamson (1979)
• statement final low tone becomes high
• question final high tone becomes low
• Isoko (Ijoid), Williamson (1979) from Elugbe
  (1977)
• statement final series of lows raised to mid
• question additional final low added
• Improving the database
• adding languages
• improving the data on many languages