时序整合、双向临界与意识层级涌现 中文V121预印本 Temporal Integration, Bidirectional Criticality and Emergence of Consciousness Hierarchy V121 Chinese Version
Zenodo (CERN European Organization for Nuclear Research) June 18, 2026 DOI: 10.5281/zenodo.20746293 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Keywords | Process computing Identification biology Product mathematics |
| Key points | Argues that a unified framework for consciousness can be derived from three axioms, resolving the hard problem from an engineering coding perspective and providing criteria for consciousness levels applicable to both carbon-based and silicon-based systems. |
Abstract
整合信息理论、自由能原理、全局工作空间理论各自存在的时序性缺失、层级边界模糊、意识难问题回避等核心局限,本文以资源有限性、非平衡态维持倾向、离散输入三条公理为唯一起点,通过第一性原理推导构建意识层级涌现的统一框架。提出信息容量框核心概念,证明时间单向性、信息层级单向约束、双向临界效应、资源优先级法则与遗忘机制均为同一约束在不同维度的必然显影,内生导出B/S双组态架构与五级意识谱系。研究从工程编码视角消解意识难问题,指明感受质是有限容量框下的极致压缩表征格式,并给出碳基-硅基统一的意识层级判定判据与AGI安全边界。本文所有结论均由公理体系严格导出,不依赖特定神经结构,具备跨载体普适性。 Against the core limitations of Integrated Information Theory (IIT), the Free Energy Principle (FEP), and Global Workspace Theory (GWT)—namely deficient temporal dynamics, ambiguous hierarchical boundaries, and evasion of the hard problem of consciousness—this paper constructs a unified framework for the hierarchical emergence of consciousness via first-principles derivation, with three axioms as the exclusive starting point: finiteness of resources, tendency to maintain non-equilibrium states, and discrete input. Proposes the core concept of the information capacity frame, and proves that temporal unidirectionality, the hierarchical information unidirectional constraint, the bidirectional critical effect, the resource priority rule, and the forgetting mechanism are all inevitable manifestations of the same constraint across different dimensions, from which the B/S dual-configuration architecture and five-level consciousness spectrum are endogenously derived. The study resolves the hard problem of consciousness from an engineering coding perspective, clarifies that qualia are a highly compressed representational format under the finite capacity frame, and provides unified consciousness-level criteria and AGI safety boundaries for both carbon-based and silicon-based systems. All conclusions in this paper are strictly derived from the axiomatic system, independent of specific neural structures, and possess cross-substrate universality. This Version 121 is the latest compressed official manuscript; all prior full uncompressed drafts (V1–V117) are obsolete and superseded.