Introduction: apart from objective nature contact, people's subjective view are a crucial individual factor for the psychological effects of nature to take hold. While most studies are based on Western cultural contexts, the ancient and contemporary content of nature-related views among the Chinese people has yet to be deeply explored.
Purpose and methods: this study aims to (1) conduct Chinese BERT-based machine learning model and Latent Dirichlet Allocation (LDA) on 68,461 nature-related posts (year range: 2009-2025) from popular contemporary Chinese social media platforms (Sina Weibo, RedNote, Zhihu, and Bilibili); and (2) use Large Language Model (LLM) to extract core themes of views of nature from a random selection of 242 ancient Chinese text (year range: 1st BCE - 18th CE).
Results: (1)the machine learning pre-trained model was well performed, and the LDA extracted four themes from contemporary Chinese views of nature: Adapting to Nature (e.g., acceptance, mind-body health), Protecting Nature, Affection for Nature (e.g., strength, joy, freedom, gratitude), and Nature-based Healing (e.g., meaning, tranquility, mindfulness, stress coping, sleep aid). (2) Snow NLP(Natural Language Processing) showed the general affection towards nature was positive (score = 0. 94). (3) LLM extracted four themes of ancient Chinese views of nature: Conforming to Nature (unity of heaven and humanity), Caring for Nature (concern and affection), Peace of Mind (emotion regulation), and Self-Transcendence (meaning making). (4) Despite China's rapid urbanization and industrialization, the unique views of nature nurtured by Chinese civilization over millennia are similar.
Conclusions and implications: This study provides evidence from ancient and contemporary China to enrich the biophilia hypothesis and stress reduction theory, highlighting the potential benefits of Chinese views of nature for mental health. These findings offer new directions for developing nature-based psychological interventions that utilize views and emotions related to nature, as well as inform the conceptual measurement of the dimensions.