Prosodic realization of focus in statement and question in tibetan (lhasa dialect)

Xiaxia Zhang, Bei Wang, Qian Wu, Yi XuInterspeech 2012

TL;DR

This paper investigates how speakers of Lhasa Tibetan use pitch and duration to highlight important words (focus) in both statements and questions. They found that focused words are spoken with higher pitch and longer duration, and importantly, that this focus marking is similar to other languages like Mandarin and English, including a phenomenon called "post-focus compression."

What Problem Does This Paper Address?

When we speak, we don't just convey information; we also signal what information is most important. This is known as "focus," and it's often marked prosodically – through changes in pitch, loudness, and duration. While this is well-studied in languages like English and Mandarin, understanding how focus is realized across diverse languages is crucial for building a comprehensive theory of human communication. Specifically, researchers are interested in a phenomenon called "post-focus compression" (PFC), where the pitch range of words after a focused word is reduced. The presence and absence of PFC across languages, even within the same language family, raises questions about its origin: is it learned locally, tied to linguistic features like tone, spread through contact, or inherited from ancient proto-languages? This paper tackles this by examining focus realization in Lhasa Tibetan, a tonal language, and comparing it to its non-tonal relative, Ando Tibetan, and other languages. They also explore how focus interacts with the prosody of questions.

Methods

The researchers recorded eight native Lhasa Tibetan speakers producing two target sentences. These sentences were uttered as both statements and yes/no questions, with specific words highlighted as the focus (initial, medial, final, or neutral focus). This resulted in a total of 384 utterances. Acoustic analysis was performed using the ProsodyPro software to extract fundamental frequency (Fo, or pitch) and duration of the target words. Fo values were converted to semitones for easier comparison. Statistical analysis, including repeated measures ANOVAs, was used to determine the significance of various factors (sentence type, word position, focus condition) on prosodic features.

Datasets

The study used newly recorded data from 8 native Lhasa Tibetan speakers. No pre-existing datasets were used for training or evaluation.

Key Insights and Ideas

  • Universal Focus Marking: Lhasa Tibetan speakers, like those of English and Mandarin, use a "tri-zone" prosodic strategy for focus: raising pitch and expanding pitch range on the focused word, keeping pre-focus words relatively stable, and significantly lowering and compressing pitch in post-focus words (PFC).
  • PFC is Robust in Lhasa Tibetan: The presence of PFC in Lhasa Tibetan, a tonal language, further challenges the idea that PFC is directly dependent on the presence or absence of lexical tone. This strengthens previous findings that PFC is not simply a feature of tonal languages.
  • Geographic Proximity vs. Spread: The finding that Lhasa Tibetan (with PFC) is geographically close to Yi (without PFC) raises questions for the "spreading hypothesis" of PFC, suggesting that simple geographic proximity might not be the sole driver of its distribution.
  • Focus and Questions Interact: While focus marking is similar in statements and questions, questions generally exhibit higher overall pitch, particularly in unfocused words. This suggests that communicative functions like focus and interrogativity are encoded in parallel but can also interact.
  • Lhasa Tibetan vs. Mandarin: While both are tonal and show similar focus marking, Lhasa Tibetan shows a difference in where the pitch rise for interrogative intonation occurs compared to Mandarin, suggesting subtle cross-linguistic variations even among related tonal languages.

Key Results

  • On-focus words: Show significant Fo rising, pitch range expansion, and duration lengthening.
  • Pre-focus words: Pitch range remains largely intact.
  • Post-focus words: Exhibit lowered and compressed pitch range (PFC). This pattern holds for both statements and questions.
  • Questions vs. Statements: Interrogative intonation generally has higher Fo than statements in unfocused words, but not in focused words.
  • Duration: Focused words are significantly longer than unfocused words in both statements and questions.

Limitations and Future Work

  • Limited Corpus: The study used only two target sentences, which might not fully capture the complexity of focus realization in all contexts. Future work could expand the sentence set and include more naturalistic speech.
  • PFC Origin: While the study provides evidence against the linguistic-dependence hypothesis and raises questions for the spreading hypothesis, the exact origin of PFC remains an open question. Further cross-linguistic research, particularly in areas where Tibetan dialects and other languages with differing PFC patterns interact, is needed.
  • Interaction of Functions: The paper notes that focus and interrogative meaning are encoded in parallel but interact. More research is needed to fully understand the nuances of this interaction across different languages and contexts.
  • Mandarin Comparison: While comparisons are made to Mandarin, a more direct, controlled comparison of focus and question prosody between Lhasa Tibetan and Beijing Mandarin would be valuable.

Summary updated

BibTeX
@inproceedings{zhang2012prosodic,
  author = {Xiaxia Zhang and Bei Wang and Qian Wu and Yi Xu},
  title = {{Prosodic realization of focus in statement and question in tibetan (lhasa dialect)}},
  booktitle = {Interspeech},
  year = {2012},
}
Figure