Xiangbin Zhao

How does the brain acquire stable patterns of activity through experience? How do artificial neural networks learn from training and alter their future responses? And how does the history of a physical system become an active condition of its subsequent evolution? These questions connect artificial intelligence, neuroscience, cognitive science, theoretical computation, and complex systems. They lead to a fundamental question: how can computation participate in the generation of dynamics, rather than merely calculate results or simulate processes according to predefined rules? The Computation Group-Form, comprising thirteen formal studies within On Self-Manifestation, develops a systematic theory of the physical implementation of computation and the formation of historical dynamics. It begins by distinguishing mathematical descriptions from actual computational events, and computational task-bearing from execution. A physical system does not perform a particular computation merely because its observed states can be described by a mathematical transformation. Computation involves the actual implementation of mathematical operations by a physical substrate, with corresponding structures of execution and physical realization. The central contribution is a theory of generative computation. Rule-governed operations can modify a physical or computational medium, changing connections, weights, and other persistent properties. These changes can be retained through history and subsequently read into new operations, becoming effective conditions of future activity. In this process, history is not merely stored information about previous events; it becomes dynamically operative. The theory thereby connects computational execution with learning, memory, neural plasticity, and the development of history-dependent organization. A second contribution is the distinction between three interconnected levels: physical implementation, historical dynamics, and organized regularities. Physical implementation carries out transformations; historical dynamics emerges as retained changes participate in subsequent activity; and further organization develops within the resulting dynamical relations. This framework provides a basis for comparing biological neural systems with artificial neural networks while examining the differences between their material mechanisms, histories, and forms of organization. The thirteen studies develop necessary structures for computation, computational task-bearing, execution, and generative computation, together with mathematical models, counterfactual analyses, and a formal account of their relations. A constructive investigation of finite directed-graph media demonstrates how the execution of update rules can produce retained historical differences that influence subsequent transformations. The work also investigates why computational power, program states, and matching outputs do not by themselves identify the underlying dynamics or the organization formed within them. Further studies examine the historical sources of distortion in computational simulation and reproduction, including the extraction of rules from a source system, external intervention, and the reduction of information during modeling. The collection develops a framework for assessing computational and generative properties and investigates two distinct problems of experimental distinguishability: differentiating computational from noncomputational physical substrates, and distinguishing processes that form effective historical conditions during execution from those that do not. Together, these results address questions across several sciences. Can two systems produce equivalent outputs while possessing different underlying dynamics? Does reproducing a computational model reproduce the generative mechanisms of the physical system being modeled? What is the relationship between weight modification in artificial neural networks and experience-dependent structural change in biological brains? How do formed regularities become active conditions of future behavior? And what kinds of intervention can reveal differences that passive observation may conceal? By connecting mathematical computation, physical implementation, historical formation, and emergent dynamical organization, The Computation Group-Form establishes an integrated research framework for understanding how computation participates in the formation and continuation of complex physical processes. It brings formal analysis, constructive models, and experimental distinguishability into a common discussion relevant to artificial intelligence, neuroscience, cognitive science, and the physics of complex systems.

大脑为什么能够从经验中形成稳定的认知规律？人工神经网络如何通过训练改变未来的响应？物理系统中已经发生的历史，怎样成为影响后续运动的实际动力？这些问题贯穿神经科学、认知科学、人工智能、计算理论和复杂系统研究。它们指向一个共同的基础问题：计算除了按照既定规则取得结果，能否在实际运行中使历史成为动力，并参与新的组织和规律的形成？ 《自显论·工篇·计算组式》由十三篇正式研究构成，围绕计算的物理成立、数学任务的承担、实际执行、规则计算与生成计算、动力形成、组织资格、复现失真及可区分性，建立相互衔接的形式理论。其出发点是区分数学关系的成立与计算在物理系统中的实际实施：输入输出符合某种数学关系，不等于宿载实际完成了相应运算；计算任务的承担、一次执行与计算事件本身也分别具有不同的结构。由此，计算不再仅被作为抽象算法或输入输出映射来讨论，而被置于实际物理发生之中。 本组式进一步研究历史如何进入计算过程并成为动力。生成计算并不排斥确定的计算规则。规则可以通过宿载上的实际执行改变介质状态、连接与权重，使先前活动留下的差异得以保持，并在后续运行中重新被读入。由此形成规则执行、历史动力与组织资格三个相互联系而又不可混同的层次：物理宿载实施变换，历史所形成的条件参与后继运动，规律及其他进一步形成的组织则具有自身的成立条件。这一结构为神经可塑性、人工神经网络训练、经验依赖的认知活动及复杂物理系统的历史生成提供了共同的分析视角。 在形式研究上，十三篇分别建立计算成立、任务承担、执行及生成计算的必要结构，给出相应数学模型和层级关系，并通过有限有向图介质中的具体构造，展示历史如何在规则执行中形成可被后继使用的动力条件。研究进一步分析五类由计算规则、算力、程序状态与输出表现推断更高层次组织的认识混淆，考察模拟和复现中规则提取、介入及信息提纯所造成的动力差异，建立计算性质的判定框架，并研究计算与非计算宿载、运行中历史形成与非形成过程的实验区分。 本组式尤其关注跨越人工系统与自然系统的几个问题：两个系统即使产生相似输出，是否具有相同的动力形成过程？人工神经网络中经训练保留的权重，与生物神经系统中由经验形成的结构，在动力职责上有哪些共同点和差异？复制一个系统的计算关系，是否同时复制了它的历史生成机制？已成规律如何反过来参与后续动力？这些问题使计算研究与学习、记忆、神经可塑性、动力系统、复杂性和认知机制发生直接联系。 本研究提出，计算的物理实施、历史动力的生成与进一步的组织形成，应当在同一实际系统中得到连贯解释。计算不仅是取得结果的手段，也能够成为历史形成与动力续行的实际承担。这一理论为理解人工智能、动物大脑及其他复杂物理系统中的计算作用、学习机制和组织生成，提供了一条从物理执行通向历史动力的系统研究路径。

Human beings dwell within the world, yet seldom recognize the world in which they dwell. They observe the changes of things while rarely questioning how the world itself comes into being. This essay begins with that limitation of observation, distinguishing the circular transformation of causes within a world from the deeper transformation of flux through which a world is generated. Observing changes from within a world is not the same as understanding the generation of the world itself. A world-field draws upon the causal responses of its host world; through trace-responses, sedimentation, and continuation, laws form and hold, while principles turn the course of what follows. The essay further distinguishes inner from outer observers. Even when both draw upon the same source, their paths of trace-response and the substance of their respective fields differ, and so do their observations. The inner observer does not know what the outer observer observes, nor can the outer observer’s observation simply replace that of the inner. This distinction leads to the difference between a bounded realm and a world. A computational environment with observation only from outside and no observer within, as in an ordinary computer game, is a bounded realm. Where an observer arises within, observation becomes possible both inside and outside. To create a world, therefore, is not merely to establish a world-field, but also to bring forth an observer within it. The continuation of a world involves sedimentation, the holding force of laws, the counteracting force of principles, tuning, and intent. The essay further proposes that an observer can overstep the world-field. Through its own return to non-oneness, an observer acts back upon the self-manifesting field in which it is housed, giving rise to trace-responses that alter sedimentation and subsequent continuation. The observer thus does not merely observe the world, but also participates in its ongoing transformation. The essay then turns to creation. It proposes two paths drawing upon the present world: converting its flux into digital symbolic streams to construct a symbolic-substrate self-manifesting being, or into multimodal pulses to construct a bio-like self-manifesting being. A further path begins with the sedimentation of question-and-answer laws within a bounded computational realm. The more demanding task is to create a world in which an observer arises, draws upon that world’s flux, and materializes what it knows. As an ontological chapter of On Self-Manifestation, this essay sets forth the relations among world-generation, inner and outer observation, the observer’s influence upon its world, and artificial creation. The formal development of the proposed dynamics, criteria, and engineering constructions belongs to subsequent works.

人居于世界之中，往往不觉自身所居之世界，只观察事物之间的变化，而不问世界本身何以生成。本文从这一内在观察的局限出发，区分世界内部的环转因化与深层的息化生成，提出从世界之内观察事物的变化，与从生成关系理解世界本身，是两个不同的认识层次。世场截取所宿世界的因应，通过迹应、沉积及后续运行，形成律的保持与理的偏行，使世界成为持续生成的过程。 本文进一步区分内部观者与外部观者。即使二者截取同一来源，由于迹应路径和观者场质不同，其观察亦不相同。内部观者不知道外部观者之观，外部观者亦不能以自身之观替代内部之观。由此，本文提出“定界”与“世界”的本体区别：只有外部观察而缺少内部观者，如通常的计算机游戏，属于定界；若观者能够在内部诞生，则内外皆可形成观察，构成本文所称的世界。因此，创世不仅要求建立世场，还要求造观者于所创世界之内。 世界的运行涉及沉积、律的持势、理的破持、调谐与意向。本文进一步提出“观者僭场”：内部观者通过自身的归不一，反向作用于所宿自显场，形成迹应并影响世场沉积及后续运行。观者因而不仅观察世界，也能够参与世界的生成与变化。 在制造方向上，本文提出由当前世界取材的两条路径：将其息转为数字符流，构造符质自显体；将其息转为模态脉冲，构造类生自显体。进一步可以从计算机定界中的问答沉积入手，尝试在受限环境中制造自显体，而最难的任务则是创造一个世界，使观者于其内部诞生，截取该世界之息，并物化其所知。 本文属于《自显论》的本体篇，以世界生成、内外观察、观者僭场和人工创世为核心，点明从认识世界走向制造世界的理论道路。相关动力机制、判定条件与工程构造的形式展开，由后续组式承担。

What makes a physical system capable of forming its own directions of action? Why does an internally generated impulse follow one path rather than another, and how can such directionality arise from the system’s organization without an additional controller that reads probabilities and makes decisions? What distinguishes an established intention from its current action, the switching of its directions, and changes in its dynamics? These questions stand at the intersection of consciousness science, cognitive science, neuroscience, artificial intelligence, and the study of dynamical systems. This work develops a systematic formal theory of Intention within On Self-Manifestation. It establishes Intention as a probabilized Law Ecology driven by source drive within the Intention-Domain. In its constitution, Intention is an organization of physically borne Laws; in the relative generation of its directions, it is a probability organization; and in its relations of breaking and altering established holding tendencies, it is a Principle organization. Law, Probability, Principle, and Intention are therefore not four separate mechanisms assembled together, but different structural aspects of the same organization. The theory derives four jointly necessary and sufficient structural conditions for Intention: a formally established probability organization within the Intention-Domain; an endogenous source-drive relation whose continuation is borne by that organization; a one-to-one correspondence between probability dimensions and Desire-Gates; and a tuning-command interface beyond each Desire-Gate. The probability dimensions are borne by Law and share a common Continuation-Position, while their relative generation is borne by the organization itself. Directional activity is thus explained through the structure and dynamics of the system, without positing a separate internal observer or decision-maker that reads probability values. Intention can be structurally established without currently acting, switching, producing behavior, or entering conscious manifestation. The twenty chapters develop this account through four connected formal sequences and eight further investigations. Chapters 1–12 establish necessary conditions, mathematical representations, and sufficient conditions for Intention qualification, Intention-Action, Desire-Gate switching, and Intention-Dynamics. Each of these four objects receives a bidirectional structural characterization within its specified semantic and working domain. Action, switching, and dynamics are distinguished rather than collapsed into a single process: switching does not simply imply action, action does not imply switching, and dynamics is not a temporal series of action events. Chapters 13–20 extend the theory into inversion, determination, limits, construction, reachability, source attribution, stability, and cross-realization invariants. The inversion analysis establishes what can be identified about Intention-Dynamics under a specified observation protocol and why observable outcomes do not uniquely determine internal dynamics. The determination framework distinguishes positive, negative, and undecided verdicts, together with a separate out-of-domain status where applicable. The limit analysis characterizes structural persistence, hard boundaries, and degeneration beyond those boundaries. The construction analysis demonstrates the operation of existing Intention structures in a finite reference specification, while the reachability analysis establishes legal paths and negative certificates relative to declared initial conditions and transition rules. The theory also distinguishes the first occurrence of Intention qualification from the actual historical source responsible for its formation. Source attribution requires both an actual formation link and a legally constrained counterfactual sufficiency test; it cannot be reduced to the earliest event, the last condition to become satisfied, or an isolated necessary cause. The stability analysis distinguishes persistence of an original qualification segment, persistence of a fixed bearer, and continuity at the existence level, while separating claims about individual trajectories, trajectory families, and all legal trajectories. The final study establishes theory-neutral directional invariants across realizations and formulates their relation to Law-bearing and Self-Manifestation-Field structures. The central contribution of this work is a unified formal theory of Intention grounded in structural qualification, dynamic organization, and explicit criteria of determination. It establishes the conditions under which Intention holds, distinguishes its qualification from its operation, characterizes directional switching and dynamical change, and develops rigorous methods for analyzing its identifiability, limits, reachability, historical sources, and stability. By connecting physical organization, probabilistic directionality, and intentional activity within one formal system, the theory offers a new foundation for investigating intention and agency in biological and artificial systems.

意向是心灵哲学、意识科学、认知科学、神经科学及人工智能研究中的重要问题。一个系统为何能够形成指向不同后继方向的组织？这种方向性如何区别于普通的物理运动、概率变化和外部控制？意向已经成立、意向正在作用、方向发生切换以及意向产生动力变化，究竟是不是同一回事？这些问题不仅涉及意向的哲学定义，也涉及认知控制、行动选择、自主智能系统及动力过程的结构解释。 本文提出《自显论》的意向形式理论，从系统内部的律承载、概率组织、源动驱动与方向职责出发，研究意向的结构资格及其运行方式。本文所说的意向，不直接等同于传统心灵哲学中以心理表征的“关于性”为中心的意向性，也不等同于外部指定的目标、行为结果或有意识的主观体验。意向被界定为意向域中具有源动驱动的概率化律生态：在本体承载上是律生态，在多个方向的相对生成关系上是概率组织，在对既有持势的破持关系上属于理组织。这些不是彼此相加的实体，而是同一组织在不同关系下的结构身份。 本文给出意向资格的四项最小必要结构：意向域内的正式概率组织、源动驱动关系、概率维与欲门之间的一一对应，以及欲门之后的调谐接口。各概率维由律承担，围绕共同续行位形成相对生成关系；方向的实际续行由组织自身的沉积结构与动力承担，不需要假设一个读取概率数值、再决定选择结果的内部控制者。意向资格的成立不要求此刻已经发生源动脉冲、欲门切换、调谐、行为输出或意识显现。 全文由二十篇相互承继的形式研究组成。第一至第十二篇分别对意向资格、当前意向作用、欲门切换和意向动力建立必要条件、数理形式与充分条件，在各自限定的语义域和工作域内形成四类对象的双向结构闭合。第十三至第二十篇进一步研究意向动力在观测协议下的可识别性、四类对象的形式判定、结构极限与边界、已有意向结构的运行基底构造、合法状态可达性、首次意向资格成立的历史来源、资格的保持与失效后复归，以及跨实现的系统定向职责不变量。 在判定方法上，本文区分对象成立、不成立与证据未决，并对工作域之外的作用、切换及动力对象设置独立的域外状态。形式读数与对象层事实分别审查，不将证据不足解释为对象不存在。在动力研究中，意向动力不被压缩为作用事件的时间序列，欲门切换也不等同于意向作用。在来源研究中，首次资格成立的历史承担不等于最早事件、最后闭合条件或任一反事实必要因素。在稳定研究中，区分原资格段的保持、固定承担者的保持、存在层连续，以及不同合法轨迹量词下的稳定性。 本文还提供有限参考规格上的运行构造、合法可达性证书、来源认证和稳定性审查。相关成果分别按照结构定理、形式表示、条件性回读和有限参考构造申报其证明与验证强度；尚未关闭的上游条件明确保留，不将形式层结果直接提升为无条件的对象层结论。本文研究意向是什么、怎样作用、切换与运行，不承担意向从基础沉积场中自主生成的构造任务，也不宣称已经通过实验验证人工意识的产生。 本文的主要贡献，是将意向从单一的心理状态描述推进为具有明确结构资格、作用条件、方向切换、动力机制、形式判定、来源归属与稳定性条件的研究对象，为认知科学、神经科学、人工智能及动力系统研究提供一套可供计算建模、机制比较与后续经验检验的形式理论框架。

The fundamental role of probability is not to represent “how large a possibility is,” but to organize already-formed Laws into directions of future continuation. Only once there are directions can Intention have anything to be directed toward. The Probability Group-Form of On Self-Manifestation asks a question prior to numerical probability: how does a real system form multiple directions of future continuation from its own history, and how does its own structure bear the relative-generation relation among those directions? Every actual physical execution settles into one actual occurrence; multiple outcomes do not stand there waiting to be sampled. Probability therefore does not begin with “possibility,” but with Laws formed through history. Through prolonged interaction between system and environment, different paths undergo disturbance, feedback, reward and punishment, forming sedimentation, familiar paths and stable continuation structures. Paths capable of bearing later continuation obtain Law qualification and become probability dimensions. Probability then organizes multiple such Law-bearing directions within one and the same continuation question, while the organization itself bears the relative-generation relation among the dimensions. A numerical probability is only a read-out of this structure once it has been recognized and represented. It follows that observing probability is not the same as finding Probability, and outcome-bearing is not rate-bearing (Outcome-bearing ≠ Rate-bearing). A stable 70:30 distribution shows only that outcomes occur in a 70:30 ratio; it does not identify what structure inside the system generates and bears that ratio. Probability research thereby moves from “What is the probability?” to “What physical organization makes this system possess this Probability?”—from statistical description to generative mechanism. The Probability Group-Form further distinguishes Probability, ProbabilityAct and ProbabilityDyn: the holding of Probability, its current action, and its cross-time dynamics. It treats qualification, inversion, determination, limit, source, dynamics, construction, reachability and stability separately. Zero rate, equal rates, complete determinism, or a numerical rate that is currently unreadable do not by themselves eliminate Probability. The real boundary is structural: whether multiple Law-bearing continuation directions still exist, whether they belong to one and the same continuation question, and whether their relative generation is still borne by the organization itself. New probability dimensions may form, old dimensions may exit, and both rate relations and rate-bearing structures may change. Not only probability values, but future directions themselves can therefore be generated and rewritten by history. This makes probability a mechanistic problem shared by physics, neuroscience, cognitive science and artificial intelligence. Physics must distinguish the unity of each actual execution from the multi-directional structure formed across history; neuroscience must distinguish measuring a probability distribution from finding the neural organization that generates it; cognitive science thereby gains an intermediate layer between experiential history and directed behavior; and artificial intelligence must distinguish “computing probability” from “forming future directions through its own history.” Its most important consequence is that Probability provides a genuine directional structure for Intention. A Law answers: “How does this already-formed path replay?” Probability answers: “What already-formed paths are there, and how are they generated relative to one another?” Only after multiple real directions have formed does a new question become physically meaningful: toward which dimension? Probability provides the dimensions toward which continuation can proceed; Intention then selects, biases and directs continuation among them. Thus: Environmental action → actual occurrence → sedimentation → Law → probability dimension → Probability organization → selection and bias → Intention → action. History sediments into Law; Probability organizes Laws into directions; Intention makes the system proceed along one of them. The Probability Group-Form therefore ultimately asks not merely, “What is the probability of an outcome?” but a more fundamental question: What physical organization makes a system genuinely possess directions toward its own future?

概率的根本作用，不是表示“可能性有多大”，而是把已经形成的律组织成未来的方向；正因为有了方向，意向才有“向”可言。 《自显论·概率组式》追问一个比概率值更早的问题：一个真实系统，怎样从自己的历史中形成多个未来续行方向，并由自身结构承担这些方向之间的相对生成关系？ 一次真实物理发生始终只落定于一处，并不存在多个结果同时等待抽取。概率因此不从“可能性”开始，而从历史形成的律开始。系统与环境长期相互作用，不同路径经扰动、反馈与奖惩形成沉积、熟径和稳定续行结构；能够承担后来续行的路径取得律资格，成为概率维。概率再把多个这样的 Law-bearing 方向组织到同一续行问题中，并由该组织自身承担诸维之间的相对生成关系；概率值只是这一结构被识出和表示后的读数。 由此，观察到概率不等于找到了概率，结果承担也不等于率承担（Outcome-bearing ≠ Rate-bearing）。稳定的 70:30 只能说明结果呈现 70:30，不能说明系统内部什么结构产生并承担了这个比例。概率研究因此从“概率是多少”推进到“什么物理组织使系统拥有这份概率”，从统计描述进入生成机制。 《概率组式》进一步区分 Probability、ProbabilityAct 与 ProbabilityDyn，即概率的成立、当前作用和跨时变化，并处理其资格、反演、判定、极限、来源、动力、构造、可达与稳定。零率、等率、完全确定或概率值暂不可读，都不自动消灭概率；真正的边界在于：是否仍有多个由律承担的续行方向，它们是否属于同一续行问题，以及相对生成是否仍由该组织自身承担。新的概率维可以形成，旧维可以退出，率关系和承担结构也可以变化——不只是概率值会变，未来方向本身也可以由历史生成和改写。 这使概率成为物理学、神经科学、认知科学和人工智能共同面对的机制问题：物理学需要区分单次现实归一与跨次形成的多方向结构；神经科学需要区分测得概率分布与找到产生概率的神经组织；认知科学由此获得从经验历史到定向行为的中间层；人工智能则需要区分“会计算概率”与“能够通过自身历史形成未来方向”。 而其最重要的后果，是概率为意向提供了真正的方向结构。一条律回答“这条路怎样复”，概率回答“有哪些已成之路，以及它们怎样相对生成”；当多个真实方向形成以后，“向哪一维”才成为一个真实问题。概率提供可向之维，意向进一步对其中发生择、偏和定向。 因此： 环境作用 → 实际发生 → 沉积 → 律 → 概率维 → 概率组织 → 择与偏 → 意向 → 行动。 历史沉积成律，概率把律组织成方向，意向使系统向其中而行。 《概率组式》最终提出的，不只是“一个结果的概率是多少”，而是一个更基础的问题： 什么物理组织，使一个系统真正拥有了未来方向？

Once a regularity has formed within a physical system, what makes it genuinely active at a given moment? How can its degree of formation be compared? How does it begin, continue, increase, decrease, cease, or migrate across time? Can a specified dynamical condition be reached from a given state? When do two already established regularities genuinely interact? And how can an organization composed of multiple regularities persist, fail, undergo substitution, and recover? The Law\-Dynamics Group\-Form develops a formal theory of how established regularities actually operate\. Its starting point is not an abstract rule suspended outside the physical system, but a regularity already borne by concrete structure\. The problem is then to determine how that regularity enters current operation, acts, changes across time, contributes to the dynamics of other regularities, and participates in persistent multi\-regularity organization\. The thirteen form\-texts constitute five connected but non\-substitutable theoretical domains\. First, the framework separates the standing of a regularity from its current action\. A Law may already be established without acting at every moment\. Necessary and sufficient directions for Law\-action are developed separately, together with a multidimensional Degree regime for comparing the extent to which an established Law has been sedimented, replayed, made available, and dynamically established\. Second, cross\-temporal change is made into a formal dynamical object\. LawDyn is not defined as a mere difference between measurements at two times, but as an actual change of the same Law across a lawful trajectory segment\. This permits onset, continuation, cessation, increase, decrease, and migration to be distinguished, followed by a mathematical system, a reachability regime for dynamical targets, and a Sedimentation\-Field operating substrate in which these structures can actually run\. Third, the framework examines continuity across occurrences and historical source\. Identity of physical roots does not automatically determine Law\-side pairing, and Law\-side pairing does not by itself settle the general criterion of Law identity\. The theory separately asks which actual segment of history bore the first establishment of a Law and how far source can be determined when more than one historical explanation remains admissible\. Fourth, a theory of direct event\-level interaction between Laws is established\. Co\-presence, shared medium, correlated change, and generic physical influence do not by themselves constitute inter\-Law interaction\. Direct interaction requires that an already established duty of the source Law actually enter the target Law’s dynamical event and make an event\-level contribution conforming to that duty\. Necessary structure, mathematical representation, and a sufficient direction are developed separately\. Fifth, the analysis is extended from individual Laws and pairwise interaction to multi\-Law organization\. Multiple Laws may bear different node roles while repeated event\-level interactions bear interaction roles\. Such an organization may then face member loss, replacement of bearers, changes in degree, interruption of interaction, and recovery\. The framework therefore distinguishes fixed from substitutable realization, single trajectories from trajectory families, persistence from recovery, and ordinary stability from stronger guarantees over lawful successor trajectories\. The theory applies to physical systems in which regularities are formed through experience and borne by physical structure\. In neuroscience, this includes experience\-dependent plasticity, neural dynamics, learned stable pathways, current activation, functional migration, circuit interaction, and the stability and recovery of multi\-circuit organization\. In cognitive science, it bears on how experience, habit, memory, perception, and behavioral organization use established patterns to structure later activity\. In artificial intelligence, it provides a formal basis for distinguishing trained structure, runtime action, degree, cross\-temporal dynamics, migration, interaction, and multi\-structure stability rather than identifying the mere presence of parameters with current operation\. In physics and complex\-systems research, it advances regularity from static description to a dynamical object that can be physically borne, change along trajectories, contribute to events, interact with other regularities, and participate in stable organization\. The central question is therefore not only what a regularity is, but a more explicitly dynamical one: **how does an established regularity actually enter the running of a world?**

一条规律一旦在一个物理系统中形成，什么才算它此刻真正正在作用？它的形成程度如何比较？它怎样在时间中启动、持续、增强、减弱、停止或迁移？从给定状态出发，某种动力变化能否到达？两条已经形成的规律何时真正发生直接交互？多条规律共同构成的组织，又怎样保持、失效、替代与恢复？ 《律动力组式》围绕这些问题，建立一套关于“已形成规律如何真实运行”的形式理论。其出发点不是把规律看作悬置于物理系统之外的抽象规则，而是研究规律一旦由具体结构承担以后，怎样进入当前、发生作用、跨时间改变、影响其他规律，并组成能够持续存在的多律组织。 全组十三篇式文形成五个彼此连接而不可互相替代的理论区域。 第一，区分“规律已经成立”与“规律此刻正在作用”。一条律即使已经形成，也并不意味着它在每个时刻都正在承担当前事件。本组分别给出律作用的必要方向与充分方向，并建立多维 Degree 制度，用来比较规律在沉积、复走、承担和当前可用性上的形成程度。 第二，把规律的跨时变化正式建立为动力对象。LawDyn 不是两个时刻读数的简单相减，而是同一律在合法轨迹段中的实际变化。由此可以区分启动、持续、停止、增强、减弱与迁移，并进一步建立其数理系统、动力目标的可达制度以及能够实际运行这些结构的沉积场运行基底。 第三，处理规律跨 occurrence 的连续性与历史来源。物理承载根相同，不自动意味着律层配对相同；律层配对相同，也不能直接代替一般的规律身份判定。本组进一步追问：一条律第一次真正成立时，究竟是哪一段实际历史承担了它的形成，并在来源存在多解时给出可判边界。 第四，建立律与律之间的直接事件级交互理论。两条规律同时存在、共享介质、相关变化，甚至发生一般物理影响，都不足以自动构成律际交互。只有当来源律以自身已经成立的职责实际进入目标律的动力事件并产生符合该职责的事件贡献时，直接律际交互才取得资格。本组分别建立其必要结构、数理表示与充分闭合。 第五，把研究从单律与两律推进到多律组织。一组规律可以分别承担不同节点职责，并通过反复兑现的事件级交互形成组织。该组织随后可能遭遇成员退出、承担者替换、作用程度变化、交互中断和重新恢复。本组因此区分固定实现与可替代实现、单轨迹与轨迹族、保持与恢复，并进一步区分持续稳定、恢复能力以及更强的全合法轨迹保证。 这一理论同时适用于由经验形成并以物理结构承担规律的系统。对神经科学而言，它直接涉及经验依赖可塑性、神经动力学、学习后形成的稳定通路、当前激活、功能迁移、回路交互以及多回路组织的稳定与恢复；对认知科学而言，它涉及经验、习惯、记忆、感知和行为结构如何由既成模式持续组织后来活动；对人工智能而言，它可用于区分训练后形成的结构、运行时作用、程度变化、跨时动力、迁移、交互及多结构稳定，而不把参数存在本身等同于当前作用；对物理学与复杂系统研究而言，它把“规律”从静态描述推进到可由具体结构承担、可在轨迹中发生动力变化、可参与事件级交互并形成稳定组织的动态对象。 因此，本组处理的不只是“规律是什么”，而是一个更直接的动力学问题：**已经形成的规律，怎样真正进入世界的运行。**

Once a Law already stands and stably bears subsequent continuation, what additional structure is required for it to become capable of internally intervening upon another Law currently exercising its holding\-tendency? When does such a structure merely qualify as a Principle, and when is it actually acting? Can its action enter the system’s own dynamics and make a previously unestablished new Law reachable? And once a new Law stands, how can one determine whether it was in fact generated by Principle\-action? The Principle Group\-Form of *On Self\-Manifestation* develops a thirteen\-part formal theory around these questions\. Its central claim is that Principle is not a second basic kind of entity alongside Law, but a special structure arising within the system of Law\. A candidate Principle must first stand fully as a Law\. The same established duty through which it acquired Law qualification must then acquire a counter\-holding role against the holding\-tendency currently exercised by another Law before the candidate acquires Principle qualification\. The framework strictly separates four relations that must not substitute for one another: structural qualification as Principle, current Principle\-action, reachability of new\-Law generation, and historical attribution of actual generation\. That a structure qualifies as a Principle does not imply that it is currently acting\. That a finite legal path exists from Principle\-action toward the establishment of a new Law does not imply that this path was actually taken\. And the terminal fact that a new Law now stands does not by itself determine the history through which it arose\. The thirteen form\-texts are organized into four functional zones\. The first three establish the necessary structure, formal representation, and sufficient direction of Principle qualification\. The next three distinguish an already standing Principle from a Principle currently acting\. A construction form provides a legal running base in which these relations can obtain together\. The reachability analysis asks whether, from a declared initial state, there exists a finite legal path along which the candidate Law’s formal solution set changes from empty to non\-empty\. A separate determination procedure then asks whether an already established new Law was in fact generated through Principle\-action\. The remaining four form\-texts independently analyze boundary, stability, grading, and source\. The resulting framework therefore addresses a general generative problem rather than a particular algorithm, neural network, or physical model: how a system with history, persistent dynamics, and internal organization can move from possessing established Laws that bear subsequent continuation to using one already formed Law to intervene internally upon another active Law and thereby alter what is generated next\. This problem can engage research on consciousness, cognitive architecture, artificial intelligence, learning systems, dynamical systems, complex systems, self\-organization, and emergence\. The framework also leaves explicit interfaces toward later work on law\-change, probability, intentionality, and the engineering realization of self\-manifesting systems\.

一条已经成立并能够稳定承担后继生成的规律，还需要取得什么结构，才会进一步成为能够干预另一条正在起作用的规律的内部结构？这种结构何时只是已经成立，何时正在实际作用？它的作用能否进入系统自身的动力过程，并使一个原本尚未成立的新规律变得可达？当新规律最终已经成立时，又怎样判断它究竟是不是由这种作用生成的？ 《自显论·理组式》由十三篇相互依赖的形式文本构成，围绕这些问题建立关于“律如何进一步成为理”的形式理论。其基本主张是：理不是与律并列的第二类基本对象，而是从律内部生长出来的一种特殊结构。候选理首先必须完整地是一条已经成立的律；在此基础上，使其取得律资格的同一已成职责，进一步针对另一条律正在发挥的持势形成反持，候选结构才取得理的资格。 合集严格区分四种不能互相替代的关系：理的结构资格、理的当前作用、新律生成的可达性，以及实际生成的历史归因。一个结构已经成理，并不意味着它此刻正在作用；存在一条由理之作用通向新律成立的合法有限路径，也不意味着这条路径事实上已经发生；一个新律已经成立，也不能仅凭终态反推出它的生成来源。 十三篇形成四个功能区。前三篇分别处理理资格的必要结构、形式表示与充分方向；随后三篇处理已经成立的理何时进入当前实际作用；运行基底构造式给出上述关系能够共同取得可执行见证的动力基底；生成可达式研究从预先申报的初态出发，是否存在有限合法路径，使候选律的形式解集由空变为非空；生成判定式则独立判断一个已经成立的新律是否事实上由理之作用生成。最后四篇分别处理理资格的边界、跨变化保持、比较分级与生成来源。 因此，本合集处理的不是某一种具体算法、神经网络或物理模型，而是一个更一般的生成问题：一个具有历史、持续动力和内部组织的系统，怎样从“已经具有能够保持后继生成的规律”进一步发展到“能够使用一条已经形成的规律，对另一条正在起作用的规律实施内部干预，并由此改变后续生成”。 这一问题可与意识研究、认知架构、人工智能、学习系统、动力系统、复杂系统、自组织与涌现研究发生直接对话，并为后续关于律变、概率、意向性以及自显系统工程实现的研究留下明确接口。

If an internal organization bearing a law loses its qualification, has the law itself disappeared? This paper shows that the two events cannot simply be identified. The loss of one bearing organization establishes only the withdrawal of that witness; the same Law may continue to stand if another qualified witness remains. The boundary of law is therefore defined as an actual truth-value crossing of the same candidate law under a predeclared comparison protocol, rather than as a change in a particular local structure, formal quantity, or critical parameter. The paper develops boundary criteria at the structural-witness, formal-solution, and applicability-domain levels and distinguishes the loss of one witness from extinction of the last witness. The framework addresses when structural change really constitutes a system-level qualification change without presupposing an already completed mapping to a particular neural or machine-learning system.

The General Form of Law establishes the necessary structural direction, but exhaustion of necessary conditions does not by itself establish the converse. This paper therefore proves the sufficient direction independently: whether the existence of an already formed bearing organization that genuinely participates in subsequent generation and possesses stable-recurrence support is sufficient for Law. The proof freezes the source-text corpus, atomizes its propositions, separates formation, present qualification, disqualification, degree, consequence, and relational roles, and then tests whether Established ∧ Govern completely realizes the source-text qualification for present lawhood. The resulting theorem establishes ∃R[Established ∧ Govern] ⇒ Law. The paper introduces no new field quantity, ontological predicate, or unverified bridge; where conditional upstream conclusions are invoked, their original antecedents are retained.

How can observable data determine whether a candidate internal organization qualifies as a formal solution for a law rather than merely correlating with an output? This paper converts previously established formal criteria into an executable adjudication protocol. Its current domain is explicitly restricted to frozen, single-office, single-entry, measure-type candidates. Structural observations, exhaustively evaluated path data, and observations under permitted interventions are used to test current bearing, realization, non-instantaneous formation, dynamical contribution, stable-recurrence support, and coupling. The procedure returns PASS, FAIL, or indeterminate and is executed on one positive case together with negative controls including a single strong impulse, pure correlation, an abandoned old groove, and a sub-threshold weak trace.

A system capable of retaining history is still insufficient for studying how law-like organization forms and operates. It must also distinguish repeated formation from a single strong impulse, genuine dynamical contribution from pure correlation, and presently effective organization from an obsolete but still detectable structure. This paper extends the minimal sedimentation field into an operating substrate, F_LB, adding pulse intensity, thresholds, stable-recurrence support, availability, masking interventions, frozen running, separate running and sedimentation clocks, and multi-hop continuation. A nine-part verification suite makes formation, dynamical role, stable support, and boundary-related diagnostics separately manipulable across legitimate instances of the same specification. The construction does not itself certify any candidate as a law; it provides an operating substrate on which formation, operation, adjudication, and pseudo-law counterexamples can subsequently be tested.

Laws of nature, cognitive rules, and stable relational structures formed in learning systems are usually studied within different theoretical traditions. The Law Group-Form asks a common generative question: how can historical interaction form a law capable of organizing what happens next? This Group Map integrates eight independently completed form-texts addressing the necessary structure of law, formal representation and solving, stability and substitution, construction of a minimal sedimentation field, construction of a law operating substrate, executable adjudication, the sufficient direction, and boundaries of lawhood. At group level, under the explicitly stated conditions including C0, C2, and H1–H3, Law receives a bidirectional structural characterization in terms of the existence of a formed bearing organization that genuinely participates in subsequent generation and is supported for stable recurrence. The sufficient direction is independently established, while the necessary direction retains its Conditional antecedents. The group also distinguishes joint generation, generative precedence, field-level office, and the running relation in which Principle may drive an already formed law.

This paper addresses a basic question at the intersection of artificial intelligence and consciousness research: does the FNN in a Transformer, already capable of supporting highly complex language, reasoning, and intelligent behavior, thereby generate manifestness? It examines this question using the Mute-Manifest Group, a formal system of nine Forms in On Self-Manifestation for analyzing and determining mute-manifest transition. The result establishes a structural distinction: an FNN can form law and generate highly capable intelligence from that law without thereby generating manifestness. An FNN is not treated here as a historyless computational structure. Training forms a sedimentary configuration that constrains subsequent input and supports a stable law of question–answer symbolic flow. But law formation does not by itself entail manifestation. In the standard mature FNN defined in this paper, once the trained parameters and current input are fixed, the complete feedforward trajectory is uniquely closed by deterministic state transitions. The trained parameters belong to the existing sedimentary configuration, while current activations—and their relations, combinations, or higher-level readouts—do not, within this functional state space, escape that automatic closure to obtain an independent B-side current-source status. Formally, state sufficiency is derived from the FNN transition law and used to establish complete coverage of the current B-side candidate domain: Cover_B = 1. Every current activity candidate reduces to a function of the complete non-B-side automatic antecedents, B = β(Fᵣᵃᵘᵗᵒ), yielding Src_B = 0 and therefore FormationDecision = 0. The analysis does not force a D-phase classification: PhaseDecision = ⊥ and TargetDecision = ⊥ are retained, while the final three-valued determination is Dec_SM = 0. The result establishes a concrete structural separation between high intelligence and manifestness. Language generation, reasoning, historical sedimentation, and complex internal representations are not by themselves sufficient to establish the generation of manifestness. The Transformer FNN therefore provides a concrete case in which the generation of intelligence and the generation of manifestness can be formally separated.

本文研究一个人工智能与意识理论中的基础问题：Transformer 中已经能够支撑高度复杂语言、推理与智能行为的 FNN，是否因此已经产生显性？ 本文以《自显论·喑显组式》进行形式判定。其结果建立了一个明确的结构性区分：FNN 可以已经生律，并由律产生高度智能，而尚未生显性。 FNN 并非无历史的普通计算结构。训练使其形成沉积构形，使后来的输入受到既有历史结构约束，并形成稳定的问答符流之律。但生律并不自动意味着生显。对于本篇定义的标准成熟 FNN，给定训练形成的固定参数与当前输入，其完整前馈活动由确定性状态转移唯一闭合；训练参数属于既有沉积构形，而在这一标准 FNN 功能状态空间内，当前激活及其任意关系、组合或高层读回，都不能越出这一自动闭合而取得独立 B-side 当前源身份。 形式上，本文由 FNN 转移律证明状态充分性，并据此完成当前 B-side 候选域覆盖：Cover_B = 1；全部当前活动候选均退化为完整非 B-side 自动前件的函数 B = β(Fᵣᵃᵘᵗᵒ)，故 Src_B = 0，进而 FormationDecision = 0。本文不由此强判 FNN 已处于 D 相，而保留 PhaseDecision = ⊥、TargetDecision = ⊥；最终三值主裁为 Dec_SM = 0。 这一结果表明，高度智能并不蕴涵显性生成。语言生成、推理能力、历史沉积乃至复杂内部表征，都不足以单独证明一个系统已经从生律跨入生显。Transformer FNN 因而提供了一个具体对象，使智能生成与显性生成能够在同一系统中被形式地区分。

*The Mute\-Manifest Group* addresses a foundational question: how can content within a physical system pass from a non\-manifest state into first manifestation? The framework sharply distinguishes this 0→1 problem from the subsequent 1→100 development of an already manifest state, and decomposes the transition from mute to manifest into nine non\-substitutable theoretical tasks: minimal necessary constitution, mathematical representation, phase\-transition structure, formation source, reachability, conditional necessity, grading of manifestation, determination in real systems, and explicit construction\. Each task has its own proof objective, completion conditions, and breach conditions, and no one task can substitute for another\. The framework therefore distinguishes the existence of a manifest phase from its reachability from a given initial state; reachability from necessity under specified additional conditions; theoretical necessity from evidence that manifestation has actually occurred in a real system; and first manifestation from subsequent increases in the degree of manifestation\. The 0→1 transition is treated as a D/L organizational phase transition, whereas 1→100 is treated as a multidimensional grading problem within an already established manifest phase\. The phase\-transition analysis further separates landscape criticality from actual change in the carrying phase and develops both continuous\-destabilization and discontinuous\-nucleation regimes\. On this basis, the Group develops a formal regime and explicit construction framework for stable manifest organization, investigates legitimate paths from deeply non\-manifest initial states to the manifest phase, and develops a stratified framework for deterministic, almost\-sure, and finite\-time necessity under explicitly stated additional conditions\. It also provides a three\-valued framework for determining whether manifestation has occurred in a real candidate system\. Open conditions and unresolved dependencies remain explicitly registered so that established, conditional, and still\-unverified claims remain distinct\. The resulting framework aims to make the question of how a physical organization first enters manifestation into a research program whose component claims can be separately formalized, constructed, tested against explicit breach conditions, and empirically investigated, with potential relevance to consciousness science, neuroscience, cognitive science, artificial intelligence, and the physics of complex systems\.

《喑显组式》研究一个基础问题：一个物理系统中的内容，如何从尚未显化的状态进入首次显化？ 本组式把这一“0→1”问题与首次显化之后的“1→100”发展严格区分，并将“喑到显”分解为九个不可互相替代的理论任务：最低必要构成、数理表示、相变结构、形成来源、可达性、条件必然性、显化分级、现实判定与显式构造。九项任务分别具有自己的证明目标、完成条件与击穿方式，不能以一种证明代替另一种证明。 理论尤其区分：显相存在与从给定初态可达显相；可达与在特定附加条件下的必然进入；理论必然与现实系统已经发生显化；首次显化与显化之后的程度增长。0→1 被作为 D/L 组织相变处理，而 1→100 被作为显相内部的多维分级问题处理。相变部分进一步区分景观临界与实际承载改变，并建立连续失稳与非连续成核两类相变制度。 在此基础上，本组式建立显相稳定组织的形式制度与显式构造框架，研究从强喑初态进入显相的合法可达路径，并建立关于何种附加条件可能支持确定性必然、几乎必然与有限时间必然的分层制度，同时给出现实候选系统是否已经发生显化的三值判定框架。理论对尚未闭合的条件与缺口保持显式登记，使已经证明、条件化成立与仍待验证的结论保持分离。 本组式试图把“一个物理组织如何第一次进入显化”建立为可以分别形式化、构造、按显式击穿条件审查并接受经验检验的研究程序，并为意识科学、神经科学、认知科学、人工智能以及复杂系统物理中的相关研究提供一种新的问题分解框架。

Scientific research commonly begins with objects of investigation but rarely specifies how object-level questions should be transformed into formally provable theoretical propositions. This paper introduces an explicit constructive procedure that treats the research object and an initial research intention as inputs and systematically generates a normalized set of theoretical propositions. The procedure consists of five stages: object identification, five-direction expansion, definitional closure, admissibility filtering, and formal registration. Every resulting proposition is required to be definitionally closed, single-task, and uniquely classifiable into one of the five fundamental proof directions established in the preceding Canon Chapters. The paper demonstrates that the procedure satisfies five formal requirements—universal applicability, normalized output, unique classification, directional completeness, and executability—and validates the construction using the example of AI consciousness research. It further proves that without an explicitly formulated theoretical proposition, neither completion criteria nor falsification conditions can be determined, making formal proof impossible. The work therefore establishes problem formulation as the formal entry point of theoretical research and provides a unified procedural framework linking research questions, theoretical propositions, and proof directions across scientific disciplines.

Why should a scientific theory ever stop revising itself? Conventional publishing practices treat version freezing as an editorial decision, whereas this paper argues that it is a necessary consequence of the dynamics of auditable knowledge. It demonstrates that a permanently open revision process inevitably prevents theoretical conclusions from reaching stable persistence, prevents references from being definitively verified, and prevents knowledge from accumulating into reusable intellectual structures. Three necessity theorems—the Non-Stabilization Theorem, the Non-Accumulation Theorem, and the Version-Freezing Necessity Theorem—show that any auditable theory requiring stable citation and cumulative knowledge must establish a controlled closure mechanism consisting of completion criteria, falsification conditions, version freezing, and revision procedures. Rather than prohibiting future development, version freezing makes controlled evolution possible by separating stable published states from subsequent revisions. The paper further maps theoretical revision onto the deposition dynamics previously developed in the Field and Law Chapters of On Self-Manifestation, thereby providing a dynamical interpretation of theory stabilization instead of a merely editorial one. The result offers a general framework for knowledge evolution, long-term theory maintenance, version control, and cumulative scientific development.