If Joan Miro Understood Dongba Scripts (2024)

Dongbaró is an interactive generative art system that reinterprets the endangered Dongba script—a pictographic writing system of China’s Naxi people—through Joan Miró’s surrealist abstraction. By integrating AI, the project bridges temporal and cultural divides between these two artistic traditions, transforming heritage into dynamic visual language.

Rooted in the concept of "cultural translation," Dongbaró goes beyond preservation to actively reinterpret traditional symbols within contemporary contexts. The Dongba script, with its pictorial grammar and shamanic roots, embodies cultural memory.  Yet static archiving struggles in the digital age. Dongbaró innovatively converts anthropological field materials into generative art, using AI to reinterpret heritage and offering an experimental pathway to observe cultural evolution through a techno-philosophical lens, transforming the dongba script into a living, evolving entity that interacts with contemporary practices.












The system analyzes user-input poetic texts, extracts themes and metaphors, and maps them to Dongba symbols via a dataset of 1,404 pictographs. This semantic matching extends beyond literality to conceptual dimensions, capturing both original context and poetic essence. The script’s structured forms are then fused with Miró’s fluid abstraction, preserving core traits while enabling creative reinterpretation.




Dongbaró redefines cultural preservation—activating the script as a dynamic medium rather than a static relic. This practice responds the discussion of postcolonial cultural production, exploring how to establish cultural identity within a globalized context through reconstructing traditional forms. Dongbaró serves both as an experiment in how technology mediates cultural heritage and as a reflection about the tension between preservation and transformation in digital cultural practices.


Interact with Dongbaró Generative System:
https://huggingface.co/spa ces/initialneil/DongbaDreamer








Jia-Qi Shi
&
Zhijing Shao