上海视觉

上海视觉 ›› 2026, Vol. 0 ›› Issue (3): 29-35.

• 视觉与图像研究 • 上一篇    下一篇

“视觉无意识”及其在后光学时代的嬗变

陈希   

  1. 东华大学, 上海 201620
  • 出版日期:2026-09-20 发布日期:2026-10-26
  • 作者简介:陈希(1989— ),女。华东师范大学哲学博士,东华大学马克思主义学院讲师。研究方向为马克思主义技术哲学与传播政治经济学。
  • 基金资助:
    本文系2024年度教育部高校思政课教师研究专项“知识图谱赋能‘马克思主义基本原理’课程精准教学研究”(编号24JDSZK159)

The Optical Unconscious and Its Transformations in the Post-Optical Era

CHEN Xi   

  • Online:2026-09-20 Published:2026-10-26

摘要:

通过类比“驱力无意识”概念,本雅明提出了“视觉无意识”概念,用以描述光学成像技术所揭示的、由那些未曾被主动观看的视觉对象所构成的无意识视域。本雅明从技术、主体与政治三个维度剖析了“视觉无意识”的辩证结构,使其成为把握当代社会视觉经验转型的关键理论枢纽。当前,多模态大模型技术正推动图像生产迈入后光学时代,“视觉无意识”呈现新特征。人工智能图像生成技术揭示了隐藏在大数据中的新型“视觉无意识”,同时也以机器视域框定乃至取代着人类视域。主体被转变为“创众”与“分众”的复合体,在享受有限自由个性创作的同时,深陷“视觉信息茧房”之中。如何应对智能技术逻辑所引发的“视觉无意识”递归式自动再生产,正成为视觉内容生态治理的新挑战。

关键词: “视觉无意识”, 后光学时代, 多模态大模型, 人工智能图像生成技术

Abstract:

By analogy with the concept of the instinctual unconscious (Triebhaft-Unbewussten), Walter Benjamin proposed the notion of optical unconscious (Optisches-Unbewusstes) to describe an unconscious visual field revealed by optical imaging technologies, composed of visual objects that have never been actively perceived. Benjamin analyzed the dialectical structure of the optical unconscious across three dimensions — technology, subjectivity, and politics — establishing it as a pivotal theoretical nexus for understanding the transformation of visual experience in contemporary society. Today, multimodal large models are driving image production into a post-optical era, endowing the optical unconscious with new characteristics. Intelligent image-generation technologies unveil a new form of the optical unconscious latent within big data, while simultaneously framing and even replacing human vision with machine vision. The subject is transformed into a composite of prosumers of creation and segmented audiences: while enjoying limited freedom in personalized creation, they become trapped within a visual information cocoon. Addressing the recursive automatic reproduction of the optical unconscious induced by the logic of intelligent technology has emerged as a new challenge in the governance of the visual content ecosystem.

Key words: Optical Unconscious (Optisches-Unbewusstes), Postoptical Era, Multimodal Large Models, AI Image Generation