Phonetic manifestation of word accents in sentence perspetive -a comparison of Tokyo and Kochi Japanese

1
Phonetic manifestation of word accents in
sentence perspetive-a comparison of Tokyo and
Kochi Japanese

• Yasuko Nagano-Madsen
• Dept. Of Oriental African Languages,
• Göteborg University

2
Accent is weakened in the following cases(for
text reading and for learners of
Japanese)appears at the reduced pitch
register/downstep, catathesisfrom Kori
(1997189)

• When a noun is modified by a preceding ajective
  or with the genitive particle no
• aoi doresu a blue dress
• watashi no doresu my dress
• When a predicate verb is modified by a preceding
  adverb(ial) element or with the particle .
• kinou yonda I read (it)yesterday
• hitori de yonda I read (it) alone

3
3. The second word in a parellel expression
nihon to amerika Japan and USA aka ka
shiro red or white4. group of words after
a focused word
1-3 are syntactically conditioned 4 is
pragmatically conditioned
4
Tokyo vs. Osaka Japanese

• Kori (1989120)
• In Osaka Japanese, the magnitude of reduction is
  not as great as that in Tokyo Japanese
• Sugito(2001204-5)
• Tokyo Japanese -gt intonation lang.
• Osaka Japanese-gt(word)accent lang.

5
Purpose of the present study

• Pilot study to examine dialectal differences
  regarding accent manifestation in relation to
  syntax (modifier)
• Tokyo vs. Kochi Japanese
• Hypotheses
• 1. Syntax prosody mapping is manifested less in
  Kochi dialect (sf. Sugito 2001)
• The magnitude of accent reduction for the second
  accent is less in Kochi dialect
• (cf. Kori 1989)

6
Meterial

• NP consisting of a noun modified by another noun
  with the genetive particle no
• okaasan no presento mothers present
• okaasan no huku mothers dress
• neko no prezento present from a cat
• neo no huku a dress for cat
• When a predicate verb is modified by a preceding
  adverb(ial) element or with the particle .
• isoide aruita I walked fast
• dondon aruita I walked fast

7

• 3. NP(subject) ga V (predicate)
• Hanako ga aruita
• Hanako walked (it is Hanako who walked, not
  Taro)
• 1 and 2 appeared in isolation, as subject phrase,
  part of sentence modifier, and in focal
  position. 36 sentences.

8
Speakers and recording procedure

• Recordings were made in summer 2006
• Speakers were university teachers, administrative
  staffs, and students
• 20 65 years
• Both male and female speakers
• Speakers took a look of wordlist (5 minutes)
• Slow and fast speaking rates, sometimes speakers
  had to re-read the utterance

9
measurements

• PRAAT
• Accent peaks were measured both in Hz and
  semitone

10
Pitch range in semitone scale-Chinese (Lin 2004)
spontaneous speech-Japanese (Nagano-Madsen
Ayusawa 2005) simulated emotional speech by
voice actors
Male average Female average
Chinese 125.7 Hz 16.4 st (N37) 217.9Hz 16.2 st (N42)
Japanese 13.9 st (N3) 14.3 st (N3)
11
Results

• Tokyo 4 speakers,
• Kochi 5 speakers
• NP (N no N)
• Ad V
• Sentence modifier

12
1. Syntax prosody mappingN no N (98 vs. 94)Ad.
V (100 vs 90)
13
Great difference in syntax prosody mapping was
found for sentence modifier (relative clause)

• Example
• Okaasan no purezento ga todoita asa datta
• Mother-Gen present-Nom arrived morning
  it was
• It was the morning when mothers present
  arrived.

14
Tokyo (speaker S.B.)
slow
fast

• F0 contours (fast speech) for
• okaasan no purezento ga todoita asa datta
• It was the morning when mothers present arrived

15
Kochi
(speaker Y.K.)
slow
fast
F0 contours (fast speech) for okaasan no
purezento ga todoita asa yatta It was the
morning when mothers present arrived
16
Sentence modifier consisting of 3 or 4 accents N
no N V N or (N ga) Ad V N 1 unit no F0
reset 2 units1 F0 reset 3 units2 F0 resets
17
2. Magnitude of reduction in semitoneNP (N no N)
e.g. okaasan no purezento mothers present
18
Kochi Japanese
19
Magnitude of accent reduction (in semitone) N no
N (5.2 vs. 3.5) Ad V (6.4 vs. 4.2) NP (subject)
ga V (predicate)(2.9 vs. 2.5)
20
Conclusion

• 1. Syntax prosody mapping was manifested less in
  the Kochi dialect
• 2. The magnitude of accent reduction (downstep)
  was consistently less for the Kochi dialect.