信息层级单向约束与机器意识检测的结构性盲区 The Information-Level Unidirectional Constraint and the Structural Blind Spot of Machine-Consciousness Detection

Jiaping Wang

Zenodo (CERN European Organization for Nuclear Research) September 11, 2026 DOI: 10.5281/zenodo.22207502 (opens in new tab)

Relevant Not relevant

Study at a glance

AI-extracted from the abstract

Characteristics	Theoretical or philosophical paper Peer reviewed
Key points	Argues that first-person experience cannot be affirmatively confirmed from external behavior, function, or structure because higher-order closed loops require subject-ownership and logical time that lower-level signals lack; however, structural-energy indicators can falsify necessary conditions across substrates. Proposes four criteria and a three-layer protocol, and contends that mainstream large models lack the structural prerequisites for consciousness emergence.

Characteristics

Theoretical or philosophical paper Peer reviewed

Key points

Argues that first-person experience cannot be affirmatively confirmed from external behavior, function, or structure because higher-order closed loops require subject-ownership and logical time that lower-level signals lack; however, structural-energy indicators can falsify necessary conditions across substrates. Proposes four criteria and a three-layer protocol, and contends that mainstream large models lack the structural prerequisites for consciousness emergence.

Abstract

大模型表现逼近人类,意识检测仍依赖概念分析与行为观察,缺第一性结构判定框架。在资源有限、非平衡维持、离散输入三公理与延迟反馈下,导出信息层级单向约束(ILUC):高阶闭环所需的主体归属与逻辑时间,是低阶信号投影后缺失的追加坐标,故行为、功能、结构类比都无法从外部确认第一人称体验,构成原则性盲区。盲区只阻断肯定确认、不阻断否定判定:层级跃迁功耗台阶、资源压缩相变等结构能量指标可跨载体测量以证伪必要条件,即可测的是结构、不可外部肯定的是质。据此给出四项判据与三层架构验证协议,论证主流大模型不具备意识涌现的结构前提,安全对功能性与组态涌现两类风险双轨前置;此为约束下收敛结果而非唯一性证明。

Study at a glance

Abstract

Related reading