"Sagging transitions" between high pitch accents in English: experimental evidence
"Sagging transitions" between high pitch accents in English: experimental evidence
复制标题
DOI：
10.1016/s0095-4470(02)00073-6
复制
发表时间：
2003-01-01
影响因子：
1.9
通讯作者：
Schepman, A
中科院分区：
文献类型：
--
作者：
Ladd, DR;Schepman, A
英语语调的ToBI转录系统区分了两种涉及局部F0最大值的音高重音，即H* 和L + H*。在L + H* 中，上升到F0最大值始于一个实际的语音目标（L），但在H* 中，上升的开始（这里是“F0 min”）被认为没有语音地位，其语音特性由各种语境因素决定。这里报告的三个实验提供了反对后一种说法的证据。实验基于英语名+姓短语（例如，Norman Elson）。在实验1中，我们表明，F0分钟是可靠的对齐与姓氏的重音音节的开始，从而与词边界的区别在最小的对像诺曼·埃尔森/诺玛纳尔逊。在实验2中，我们表明，实验修改对齐的F0分钟，在这样的分段模糊的短语影响听众的判断，他们听到的名字。在实验3中，我们表明，F0分钟和第二H* 口音的F0水平的影响的两个重音音节之间的干预音节的数量，在某种程度上是不预测的Pierrehumbert的“下垂过渡”模型，这是中央H* 和L + H* 之间的区别。因此，我们认为，在H* 和L + H* 有不同的L和H的目标，这两个应该被视为属于一个单一的重音类别。这一分析使英语语调的描述与欧洲其他语言的语调描述在理论上更加一致。该分析还有助于解释ToBI转录员在区分H* 和L+H* 时所表现出的困难。(C)2003爱思唯尔科技有限公司版权所有。
The ToBI transcription system for English intonation draws a distinction between two kinds of pitch accents involving local F0 maxima, namely H* and L + H*. In the L + H*, the rise to the F0 maximum begins with an actual phonological target (L), but in the H* the beginning of the rise (here "F0 min") supposedly has no phonological status and its phonetic properties are determined by various contextual factors. The three experiments reported here provide evidence against this latter claim. The experiments are based on the phonetic properties of the medial F0min in H* H* sequences on English given name+surname phrases (e.g., Norman Elson). In Experiment 1, we show that the F0min is reliably aligned with the beginning of the accented syllable of the surname, thus correlating with the word boundary distinction in minimal pairs like Norman Elson/Norma Nelson. In Experiment 2, we show that experimentally modifying the alignment of the F0 min in such segmentally ambiguous phrases affects listeners' judgment of which name they are hearing. In Experiment 3, we show that the F0 level of the F0 min and of the second H* accent is affected by the number of syllables intervening between the two accented syllables, in a way that is not predicted by Pierrehumbert's "sagging transition" model, which is central to the distinction between H* and L + H*. We therefore argue that in both H* and L + H* there are distinct L and H targets, and that the two should be regarded as belonging to a single accent category. This analysis makes the description of English intonation more theoretically consistent with that of various other European languages. The analysis also helps explain ToBI transcribers' demonstrated difficulty in making the distinction between H* and L+H* reliably. (C) 2003 Elsevier Science Ltd. All rights reserved.