Cross-System Evolutionary Dynamics: The FIT Framework (Force – Information – Time) — Origin and Design Goals

We present the Force–Information–Time (FIT) Framework, a minimal meta-theory for reasoning about evolutionary dynamics across physical, biological, cognitive, social, and artificial systems. Starting from a fundamental question—why do hierarchical structures emerge, and why do systems fail not from lack of power but from wrong tempo?—we compress "evolution" into three primitive variables: Force (F) as directed influence, Information (I) as persistent causal structure, and Time (T) as emergent characteristic scales. Extended with Constraint (C) and State (S), these five primitives generate a level-aware framework that addresses the fragmentation problem in cross-domain evolutionary science. Unlike domain-specific theories (thermodynamics, evolutionary biology, machine learning optimization), FIT provides a meta-language—a common syntax for expressing diverse evolutionary phenomena without claiming to replace specialized frameworks. The framework is explicitly designed to be: (1) minimal—using the fewest primitives without assuming specific mechanisms; (2) operational—every theoretical claim translatable to measurement protocols; (3) falsifiable—generating specific, testable propositions; and (4) level-aware—binding all statements to explicit estimator tuples. We articulate six principles governing system evolution, distinguishing near-tautological core principles (ℒ1–ℒ3) from empirical working hypotheses (ℋ4–ℋ6), and derive 18 falsifiable propositions. The Estimator Selection Theory (EST) introduced in v2.4 provides formal axioms (A1–A8) preventing "estimator-hacking" critiques while acknowledging observer-dependence. Initial Tier-1 validation on cellular automata (Conway's Game of Life, Langton's Ant) demonstrates both predictive power and sensitivity to measurement choices. This paper serves as the foundational reference for FIT, establishing the philosophical motivations, formal structure, and epistemological commitments of the framework. It is intended for researchers in philosophy of science, complex systems, and cross-disciplinary evolutionary theory. [中文/Chinese] FIT(力–信息–时间)动力学:起源与设计目标 本文提出力–信息–时间(FIT)框架,这是一个用于推理物理、生物、认知、社会和人工系统演化动力学的最小化元理论。从一个根本问题出发——为什么层级结构会涌现?为什么系统的失败往往不是因为力量不足,而是因为节拍错误?——我们将"演化"压缩为三个基元变量:力(F) 作为有向影响,信息(I) 作为持久的因果结构,时间(T) 作为涌现的特征尺度。扩展后的 约束(C) 和 状态(S) 共同构成五个基元,生成一个层级感知的框架,解决跨领域演化科学中的碎片化问题。 不同于领域特定理论(热力学、演化生物学、机器学习优化),FIT 提供的是元语言——一种表达多样演化现象的通用语法,而非声称要取代专业框架。该框架明确设计为:(1)最小化——使用最少的基元,不预设特定机制;(2)可操作——每个理论断言都可转化为测量协议;(3)可证伪——生成具体的、可检验的命题;(4)层级感知——所有陈述都绑定到显式的估计量元组。 我们阐述了支配系统演化的六个原则,区分近乎重言式的核心原则(ℒ1–ℒ3)和经验性工作假说(ℋ4–ℋ6),并推导出18个可证伪命题。v2.4中引入的估计量选择理论(EST)提供了形式化公理(A1–A8),防止"估计量黑客"批评,同时承认观察者依赖性。在元胞自动机(康威生命游戏、兰顿蚂蚁)上的初步Tier-1验证展示了预测能力和对测量选择的敏感性。 本文作为FIT的基础参考文献,确立了框架的哲学动机、形式结构和认识论承诺。适用于科学哲学、复杂系统和跨学科演化理论的研究者。