1、 Introduction

The current situation and importance of constructing Chinese interlanguage corpora were discussed, and the shortcomings and challenges of multimodal corpus research were pointed out, including the difficulty of collection, transcription, and annotation. At the same time, the collection and transcription of spoken and video language materials, as well as the neglect of spoken language in Chinese language teaching research, were analyzed. Finally, the necessity of strengthening the research and construction of oral and multimodal databases was proposed, and the research gap in body language annotation was pointed out. The aim is to explore the body language issues in video corpus annotation and promote related research and corpus construction.

2、 The value of body language annotation

In depth exploration of the value of body language in the annotation of multimodal corpus of Chinese interlanguage. Firstly, body language is defined as the facial expressions and body movements that accompany speech, similar to non acoustic features in paralanguage but with a greater emphasis on motor phenomena. Secondly, the importance of comprehensiveness and segmented annotation was emphasized, and it was pointed out that this has been effectively applied in the construction of the global Chinese interlanguage corpus. Then, the value of body language annotation was elaborated from four aspects: 1) revealing the speaker's true intention and helping listeners understand the implied meaning; 2) Reflecting learners' Chinese proficiency and guiding teachers in organizing teaching; 3) Promote the in-depth analysis of multimodal discourse and provide rich linguistic resources; 4) Understand the cultural differences between China and foreign countries, examine body language errors and negative transfer phenomena. These values reflect the important role of body language in language communication, teaching, research, and cross-cultural communication.

3、 Annotated content

Elaborated in detail the principles and content of body language annotation in the multimodal corpus of Chinese interlanguage. Firstly, it has been clarified that the annotated object should be a paragraph rather than a single sentence, as a paragraph can more fully express the central meaning in communication. A paragraph is composed of at least two sentences that are coherent and jointly express a central idea, which is crucial for cultivating the ability to express oneself in paragraphs. Secondly, it is pointed out that the annotation of body language should focus on body language that has auxiliary effects on meaning expression, such as smiling, excited expressions, and gestures such as waving and waving. Body language that is not helpful for meaning expression, such as blinking and other physiological phenomena, does not need to be annotated. Finally, 10 aspects of body language annotation were detailed, including expressions, facial movements, head movements, etc., providing specific guidance for body language annotation in video corpora.

4、 Annotation method

Elaborated in detail the annotation methods for body language annotation in the multimodal corpus of Chinese interlanguage. Firstly, the model of "error labeling+basic labeling" was introduced, aiming to comprehensively analyze the correct and incorrect application of body language. Secondly, the annotation method using human labeling and machine assistance was discussed, emphasizing the combination of manual analysis and computer-aided methods to improve annotation efficiency and accuracy. Next, the multimodal corpus transcription or annotation software, especially the characteristics and advantages of ELAN software, were introduced, and it was pointed out that it is suitable for the construction of multimodal Chinese interlanguage corpora. In addition, the presentation of annotated content has been specified, including the use of solid square brackets and curly brackets, as well as incorrect and correct body language marker codes. Finally, the annotation rules were clarified, including body language annotation for speakers and listeners, annotation for multiple body languages, overlapping of annotation codes, and the placement of annotation codes. These rules help to standardize and systematize annotation.

5、 Conclusion

Video corpus plays a core role in the construction of multimodal corpora and is crucial for Chinese language teaching and research. The current construction of multimodal corpus for Chinese interlanguage is insufficient and needs to be strengthened. Body language annotation is a crucial step that requires in-depth research. This article provides preliminary conclusions, but they need to be validated and deepened in broader practice.