赵晨扬

（女，2019级博士，专业：语言学及应用语言学，研究方向：语音与言语科学，导师：李爱军）

　　摘 要

　　本文在自主音段-节律音系学理论（The?Autosegmental-Metrical?theory of Intonational Phonology, AM）的基础上，用实验语音学的方法，深入对比分析了粤、吴方言区中等水平英语学习者与英语本族语者语调的产出和感知特征、学习者母语方言语调对其二语语调的影响，并开展了面向粤、吴方言区英语学习者的可视化语调教学研究。

　　本研究的主要工作和创新点如下：

　　（1）考察粤、吴方言区学习者英语语调产出的音系和语音特征。重点分析了两个方言区学习者陈述句和一般疑问句的音系表征和语音实现。音系方面，粤方言区学习者陈述句和一般疑问句的语调模式分别为H* L-L%和L* H-H%，与英语本族语者相同，而吴方言区学习者则常采用H* L-L%和H* L-H%来实现以上两种句型，和本族语者存在差异；两方言区学习者句首和句中焦点正确实现为核心重音的比例高于句末焦点和宽焦点。语音方面，粤方言区学习者存在焦点音节凸显程度不足的现象，而吴方言区学习者的焦点凸显程度充分，甚至表现出了相对于本族语者而言更为夸张的凸显形式。

　　（2）探索学习者语调产出和感知间的关系。从位置方面来看，学习者更容易感知到位于句首和句中的核心重音，而这两个位置下的焦点恰好也更容易被两方言区学习者实现为核心重音，这一现象支持了SLM-r中“感知与产出同步发展的假设”。然而，从音高类型上看，吴方言区学习者在产出两种句型时常用的音高模式，并非最容易被感知到的类型，其可以感知到的特征并不能够被正确地产出，这一表现又与SLM中“产出发展滞后于感知”的情况相吻合。

　　（3）分析学习者的母语语调对其英语语调产出的影响。由于汉语的音高曲线同时编码了声调和语调，且分属词汇和后词汇两个不同的结构层次，因此方言区英语学习者语调的成分，即边界调和音高重音的产出，必然受到母语方言的影响。首先看边界调，疑问句中，粤方言区学习者由于其母语方言语调疑问句句末边界调H%实现为后续叠加边界调，边界音节声调后存在一个固定的高点目标值，因此其产出的英语一般疑问句边界调与本族语者一致，可以实现为升调，而吴方言中由于语调疑问句的边界调H%实现为同时叠加边界调，导致学习者的英语一般疑问句边界调实现为升调时存在困难；两方言区学习者母语方言边界调L%均实现为同时叠加低调特征，因此其陈述句的边界调实现都表现较好。粤方言和吴方言的语调表现存在一定差异，前者属于无焦点后压缩（Post-Focus Compression, PFC）的方言，焦点实现主要依靠时长手段，与学习者的目标语英语存在差异，而后者则为有PFC的方言，主要依靠音高手段来进行焦点的凸显。正是这一差异导致了两方言区学习者在产出英语语调时分别存在焦点凸显不足和焦点凸显充分的现象，从超音段层面上支持了迁移理论。同时，本研究结果也进一步印证了Mennen二语语调教学理论中的观点，即二语语调的感知与产出与音段不同，在预测二语语调特征产出的困难时，不仅只局限与一个维度，需要综合语调的音系表征、语音实现、语调的功能与形式的对应关系等。

　　（4）探索多模态、监督反馈等多种语调教学方式的有效性。通过设计可视化语调教学实验，用五种不同的教学方法（无监督教学、音频辅助教学、音频+可视化语图辅助教学、音频辅助教学+反馈、音频+可视化语图辅助教学+反馈）对江淮官话区、北方方言区、北京方言区等3个地区的英语学习者进行教学实验，证明音频+可视化语图辅助教学+反馈的方法最为行之有效。

　　（5）采用可视化反馈语调教学，探索粤、吴方言区学习者的语调教学效果，从音系和语音两个方面考察了学习者训练前后的语调表现差异，证明了可视化结合监督和反馈的教学方法的有效性。采用以上的语调教学方法，不仅可以解决音系上存在的偏误，提高边界调和核心重音类型实现的正确率，同时还可以使得学习者所产出的焦点音节在语音上更靠近本族语者，一定程度上解决了学习者在焦点处音高和时长范围扩张不充分，以及焦点后音高压缩程度不足的问题。

　　（6）编制用于可视化英语语调教学平台的教学材料。学习课程共有7个章节，每个章节均包含语调特征讲解与感知产出练习两个部分，按照语调功能分类开展语调学习。通过可视化的语调训练方法(visualized intonation learning and training)，让学习者在明确不同语调模式的具体功能之后展开训练，帮助学习者建立语调意识（intonation awareness）。

　　关键词：英语作为二语的语调产出与感知；粤方言区英语学习者；吴方言区英语学习者；可视化英语语调教学

　　Abstract

　　Based on Autosegmental-Metrical (AM) theory, an in-depth investigation on the intonation production and perception of L2 learners with Cantonese and Wu dialecal background will be conducted by using the experimental phonetics method. The relation between learners English production and perception, learners’ mother tongue influence imposed on learners’ production and perception will be discussed as well. On this basis, a visualized intonation teaching will be given to these learners for the purpose of improving their English intonation production.

　　The main work and contributions of this research are listed below:

　　(1)Learners of the two dialectal areas are studied in terms of their English production and perception. In their phonological presentation, learners with Cantonese background (CS) tend to realize English statements by H* L-L% and YN-questions by L* H-H%, similar to the English native speakers’（NS） representation. While learners with Wu background (WS) usually adopt H* L-L% and H* L-H% in the above two sentence types respectively; in addition, both learner groups are likely to realize the sentence-initial and sentence-medial focus correctly, whereas find it difficult to produce sentence-final and broad focus correctly. In phonetic realization, CS focus realization are mainly represented by a less prominent nuclear accent, while some of the WS have their focus realization properly highlighted, and some of them even do this in a much exaggerated manner than NS.

　　(2)Relation between perception and production are explored. It is noticed that nuclear accents in sentence-initial and medial position are easier for learners to perceive, and focus in these two positions are exactly those easy to be realized as nuclear accent. This finding is in line with SLM-r, which claims that perception and production develop hand in hand. While WS’s inconsistency on the production and perception of nuclear accent type in YN-questions goes the opposite of SLM-r, but to some extent provide evidence for SLM, which indicates that the ability of production is acquired behind perception.

　　(3)Learners’ mother tongue influence imposed on their English production is analyzed. It is known that Chinese pitch contour encodes tone and intonation simultaneous, which basically belongs to lexical and post-lexical level respectively. For this reason, it is no doubt that learners’ English will be influence by their mother tongue dialects. For the production of boundary tone, CS are easier to produce a H% in YN-question, as there is a fixed high target at the end of Cantonese intonation questions. No matter how their intonation declines in the sentence main part, it will meet a rising at the end. But this phenomenon is absent in Wu dialect, for which, WS for it hard to realize the boundary tone in YN-questions as H%. L% in English statements can be well produced by both learner groups, as this feature is a superposition in the statements of the two dialects. Comparing the phonetic realization of English intonation between the two learner groups, it is found that some of the WS do it well, as focus in their dialect is featured by PFC and pitch range expansion, while their CS counterparts fail to realize the focus in a prominent manner, as Cantonese mainly use duration lengthening in doing focus realization, leaving learners’ pitch level of English focus syllable remain unchanged compared with those in broad focus.

　　(4)A pilot study of teaching experiment was designed, in which five different teaching methods were used to teach learners from Jianghuai dialectal areas, northern mandarin dialectal areas and Beijing dialectal areas. By the observation of learners’ pre and posttest data, a conclusion was drawn that the fifth teaching method, teaching with visualized intonation contour and after-training feedback, is the most effective one.

　　(5)The fifth teaching method was then used in Cantonese dialectal area and Wu dialect area. The data elicited was annotated and analyzed. By the comparison of production between learners in control and experiment group, it was found that there was an interactive relationship between production and perception of L2 learners. By this method, the fossilized deviant features in either phonological or phonetic aspects can be removed to a large extent.

　　(6)A learning and teaching intonation website was designed with 7 chapters. It was expected that by taking courses designed in this website, learners were able to establish intonation awareness, and able to be aware of the relationship between the intonation form and function.

　　Key words: Teaching English as a second language; English learners with Cantonese background; English learners with Wu dialect background; visualized English intonation teaching