arXiv:2606.23232eess.AS2026-06中稿 · Interspeech 2026 M…

研究发现说话末尾的词会明显变长,是语言交接的重要提示信号。

Word Lengthening as a Function of Utterance Position: A Multi-Corpus Study

论文配图:Word Lengthening as a Function of Utterance Position: A Multi-Corpus Study
图 1 · 摘自论文原文
  • 通过多语种大数据分析,发现句末词比中间词长约191毫秒。
  • 即使控制词汇和说话人差异,句末词仍平均长80毫秒且显著。
  • 变长主要集中在最后一个音节,是精准的语言交接线索。

高效对话需要对话者在几百毫秒内预测话轮结束。除了句法和语用完成外,韵律(尤其是边界前延长)有助于预测。我们检验了句末词是否比句中词更长,该现象是否反映韵律调整而非词汇选择,以及延长集中在词的哪个部分。分析涵盖四种语料库(英语、西班牙语):Switchboard、Columbia Games、BU Radio、Glissando,涉及超过500名说话者,共约39,500个话轮,包含39,470个句末词和206,268个句中词。结果表明,句末词更长(平均约191毫秒;效应量d=1.14)。在匹配词、同说话人比较中,该效应依然显著(80毫秒;p<0.001),且主要集中于最后一个音节(效应量d=0.89)。因此,句末延长是一种稳健且局部化的语用转移提示。

原文摘要 · Abstract (English)

Efficient turn-taking requires interlocutors to predict turn endings within a few hundred milliseconds. Beyond syntactic and pragmatic completion, prosody (especially pre-boundary lengthening) supports projection. We test whether turn-final words are longer than mid-sentence words, whether this reflects prosodic modification rather than lexical choice, and where within the word it concentrates. We analyze four corpora spanning styles and two languages (English, Spanish): Switchboard, Columbia Games, BU Radio, and Glissando, with >500 speakers, $39{,}470$ turn-final and $206{,}268$ mid-sentence tokens across $\sim39{,}500$ turns. Turn-final words are longer (mean ${\approx}191$\,ms; $d=1.14$). The effect persists in matched-word, within-speaker comparisons ($80$\,ms; $p<0.001$) and is localized mainly to the final syllable ($d=0.89$). Turn-final lengthening thus emerges as a robust, localized cue to floor transfer.

语音韵律对话分析语言交接

Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。