Innovation Series: Advanced Science (ISSN 2938-9933, CNKI Indexed)

Cross-layer Discourse Ripples: a Quantitative Case Study of AI-Romance Narratives in Social Media Communication

Weicheng Meng¹, Mengjiao Yin^1,2,*

¹ Business School, Wuxi Taihu University, Wuxi, Jiangsu, China

² Laboratory of Artificial Intelligence and Consumer Behavior, Wuxi, Jiangsu, China

Corresponding Author: Mengjiao Yin

Abstract: This study quantitatively analyzes the discourse ripple effect generated by cross-platform communication of human-machine romantic relationship-related topics across three heterogeneous social media platforms: Douban, Zhihu and Rednote, to address the insufficient explanatory power of the unidimensional attenuation hypothesis in traditional communication studies. Based on social distance theory, a three-layer discourse space model is constructed, dividing Douban, Zhihu and Rednote into the direct-interaction layer, near-interpersonal layer and far-interpersonal layer respectively. A total of 899 user comments related to human-machine romantic relationship topics are collected from the three platforms. Jieba Chinese word segmentation and stop word filtering are adopted for text preprocessing. A self-developed Cross-layer Discourse Ripple Index (CDRI) is constructed to quantitatively measure discourse differences between adjacent layers from three dimensions: semantic resonance, emotional attenuation and discourse mutation rate. The composite ripple index from Douban to Zhihu reaches 0.8927, indicating a strong ripple feature; the index drops to 0.3411 from Zhihu to Rednote, presenting a weak ripple feature. TF-IDF semantic similarity of all three layer pairs is lower than 2%. Emotional intensity is reversely amplified by approximately 121% from Douban to Zhihu, while attenuating by about 25% from Zhihu to Rednote. A total of 3,721 new concepts emerge from Douban to Zhihu, compared with merely 606 new concepts from Zhihu to Rednote. The cross-platform communication of human-machine romantic relationship topics is not a linear attenuation process, but a complex dynamic system integrating emotional resonance, knowledge reconstruction and semantic mutation. The near-interpersonal layer exerts significant emotional amplification and conceptual reconstruction effects on core discourse, thus serving as the key mediating node of ripple communication.

Keywords: Cross-layer Discourse Ripple Index; Human-Machine Romantic Relationship; Cross-platform Communication; Social Distance Theory; Sentiment Analysis

References