TL;DR
This paper investigates how vowel duration (quantity) varies across Japanese, Finnish, and Czech, focusing on how word structure and surrounding sounds influence these differences. The study reveals that word structure plays a more significant role than language itself in shaping vowel quantity, with Japanese exhibiting more consistent durational patterns.
What Problem Does This Paper Address?
Understanding how languages distinguish between short and long vowels (vowel quantity) is crucial for phonetics and speech processing. While many languages have this distinction, the acoustic realization and the factors influencing it can vary significantly. Japanese and Finnish have phonemic distinctions for both vowels and consonants, while Czech only distinguishes vowel quantity. This paper aims to explore how these linguistic differences, combined with varying word structures (like CVCV vs. CVVCVV) and surrounding consonants, affect the actual duration of vowels in Japanese, Finnish, and Czech. The goal is to pinpoint whether language-specific phonological rules or more universal acoustic principles govern these durational patterns.
Methods
The researchers conducted a cross-linguistic study using nonsense words to control for lexical meaning and focus purely on phonetic properties. They created four disyllabic word structures (CVCV, CVCVV, CVVCV, CVVCVV) using the consonant /m, p, s/ and the vowel /a/ (both short and long). Native speakers of Japanese, Finnish, and Czech read these words, and their speech was recorded. Spectrogram analysis was used to measure the segmental durations of vowels and consonants. Durational ratios between short and long vowels were calculated at both the segmental and lexical levels, and statistical analyses (ANOVA) were performed to assess the significance of language and word structure differences.
Datasets
The study utilized newly created nonsense words based on specific structural templates (CVCV, CVCVV, CVVCV, CVVCVV) and phoneme combinations (/m, p, s/ and /a/). These words were spoken by native speakers of Japanese (Tokyo dialect), Finnish (Helsinki dialect), and Czech (Bohemian dialect). The recordings were then analyzed using the Praat speech analysis program.
Key Insights and Ideas
- Word Structure Dominates: The most striking finding is that differences in word structure had a more significant impact on vowel quantity than the language itself. This suggests that the acoustic realization of vowel length is highly sensitive to the phonetic environment created by the word's structure.
- Japanese Isochrony: Japanese exhibited more "isochronal" increasing patterns in lexical durational ratios as the number of morae increased. This implies a more consistent and predictable temporal control within Japanese words.
- Czech Syllable Timing Clues: Czech showed a tendency for longer durational ratios in the second syllable, and its overall durational variation (range and standard deviation) was the smallest. This rigidity, along with its vowel quantity distinction, might support theories suggesting Czech is a syllable-timed language.
- Consonant Influence Varies: The surrounding consonants affected short vowels more in Finnish and long vowels more in Japanese. This highlights how different languages might prioritize different phonetic cues for vowel length distinction.
- Finnish and Czech Stress/Phonotactics: The paper speculates that Finnish and Czech vowel quantity might be more influenced by stress position and phonotactic constraints within the language.
Key Results
- Japanese segmental ratios between short and long vowels were longer than in Finnish and Czech.
- Czech durational ratios were longer in the second syllable compared to the first.
- Word structure differences significantly affected vowel quantity, particularly for vowels in the first syllable.
- Japanese showed more consistent increasing patterns in lexical durational ratios across word structures.
- The range and standard deviation of durational variations were generally smallest in Czech, indicating more rigid control.
Limitations and Future Work
The authors acknowledge that the study was limited to disyllabic words and the vowel /a/. They suggest that future research should incorporate a wider range of test words, including meaningful ones, and a broader set of vowels to expand the findings on vowel quantity. They also propose further investigation into the role of stress and phonotactics in Finnish and Czech.
Summary updated