六鸟：涌现演算的基础
Six Birds: Foundations of Emergence Calculus
AI总结：
本文提出了一种跨学科的涌现演算，通过六个闭包变换原语统一序理论和动力学诱导的内射映射，并证明了在粗粒化下时间箭头函数的单调性及反饱和机制。
摘要：
我们开发了一种跨学科的涌现演算，将理论视为作用于描述上的幂等算子的固定点。我们证明，一旦过程是可组合的，但对底层系统的访问通过一个有界的观察接口进行中介，一组六个闭包变换原语（P1-P6）是不可避免的。该框架将序理论的闭包算子与由马尔可夫核、粗粒化透镜和时间尺度τ构建的动力学诱导的内射映射E_{τ,f}统一起来。我们引入了一个可计算的总变差幂等度缺陷用于E_{τ,f}；小保留误差意味着近似幂等性，并在选定的τ内产生稳定的“对象” within一个固定的透镜。为了解决方向性问题，我们定义了一个时间箭头函数作为正向和时间反转轨迹之间的路径空间KL散度，并证明其在粗粒化（数据处理）下是单调的；我们还正式化了一个协议陷阱审计，表明仅凭协议缠绕本身无法在没有提升动力学中的真实亲和力的情况下维持不对称性。最后，我们证明了一个有限强制式计数引理：相对于基于分区的理论，可定义的谓词扩展是指数稀有的，从而为严格梯子攀登提供了干净的反饱和机制。
英文摘要：
We develop a discipline-agnostic emergence calculus that treats theories as fixed points of idempotent operators acting on descriptions. We show that, once processes are composable but access to the underlying system is mediated by a bounded observational interface, a canonical toolkit of six closure-changing primitives (P1--P6) is unavoidable. The framework unifies order-theoretic closure operators with dynamics-induced endomaps $E_{τ,f}$ built from a Markov kernel, a coarse-graining lens, and a time scale $τ$. We introduce a computable total-variation idempotence defect for $E_{τ,f}$; small retention error implies approximate idempotence and yields stable "objects" packaged at the chosen $τ$ within a fixed lens. For directionality, we define an arrow-of-time functional as the path-space KL divergence between forward and time-reversed trajectories and prove it is monotone under coarse-graining (data processing); we also formalize a protocol-trap audit showing that protocol holonomy alone cannot sustain asymmetry without a genuine affinity in the lifted dynamics. Finally, we prove a finite forcing-style counting lemma: relative to a partition-based theory, definable predicate extensions are exponentially rare, giving a clean anti-saturation mechanism for strict ladder climbing.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
