引论：易经与现代决策科学的范式相遇 / Introduction: The Paradigmatic Convergence of I Ching and Modern Decision Science

第 1 章 引论：易经与现代决策科学的范式相遇

Chapter 1 Introduction: The Paradigmatic Convergence of I Ching and Modern Decision Science

---

1.1 算法时代的三千年文本：在确定性与敬畏之间

When DeepMind's AlphaFold predicts protein structures with near-perfect accuracy, when reinforcement learning agents defeat human champions at Go and poker, when recommendation algorithms shape the daily epistemic diet of billions, a pointed question emerges: 在这样一个由数据、算力与优化算法主导的时代，一部成形于西周初期、以蓍草和龟甲为工具的占筮文本，究竟还能在决策科学的严肃讨论中占据何种位置？This is not a rhetorical flourish, but the precise question this book sets out to answer. If the I Ching (《易经》), also known as the *Yijing* or *Book of Changes*, has nothing to contribute beyond cultural nostalgia or exotic metaphor, then the most honest verdict is silence. If, however, certain features of this text anticipate—or at least resonate structurally with—problems that contemporary decision science is only beginning to formulate, then the silence would be intellectually costly. The thesis defended throughout this volume is that such structural resonance is demonstrable, and that the dialogue between the *Yijing* and modern decision science is neither anachronistic mysticism nor a forced analogy, but a legitimate cross-paradigm inquiry with concrete epistemic payoff.

三千年是一个难以用日常经验衡量的时间尺度。如果把《易经》的经文部分（卦辞与爻辞）视作一个连续诠释过程的起点，那么到今天为止，这部文本已经经历了至少四个诠释范式的更迭：先秦的卜筮实践、汉代的象数体系、魏晋至唐的玄学化与象数深化、宋明的义理之学，以及清代的汉学回归与近代以来的多学科解读。在每一轮更迭中，《易经》都被以那个时代最具解释力的概念框架重新阅读：天人感应、阴阳五行、太极元气、理气心性、朴学考据、系统论、信息论、复杂性科学、认知心理学、风险管理。这种持续的"再语境化"（recontextualization）本身就说明，《易经》不是一部可以被归档的古典文献，而是一个仍在生成意义的诠释事件（hermeneutic event）。

The problem of "paradigm incommensurability" raised by Kuhn (1962) and refined by Feyerabend (1975) is relevant here. Two conceptual systems may be incommensurable at the level of vocabulary, and yet structurally isomorphic at the level of formal relations. A 64-element combinatorial space defined by six binary choices is formally identical whether one calls its elements "hexagrams" or "6-bit codewords." The question is not whether Leibniz, who corresponded with the Jesuit Joachim Bouvet and recognized the binary structure of the *Yijing* in 1703, was right about a hidden affinity (he was), but whether the *Yijing*'s internal logic of interpretation—its reliance on positional reading, contextual correspondence, and the dynamic weighting of changing lines—can be reconstructed as a formal theory of decision under conditions of incomplete information, bounded rationality, and dynamic environments. This is the working hypothesis of the present book.

我们并不主张《易经》是一部被埋没的数学论著，也不认为其作者是先于莱布尼茨的二进制发明者。真正值得严肃对待的是以下事实：《易经》在三千年的诠释史中，发展出一套对"时机—位置—关系"（timing-position-relation）的精致语义学，这套语义学在结构上与现代决策科学中关于情境决策、动态博弈、鲁棒性优化、机会识别等问题的最新讨论存在未被系统阐发的对应关系。本书的任务，就是把这种对应关系从隐喻层次提升到形式层次，并在必要时发展出可被检验的命题。

Throughout, we adopt a principle of methodological austerity: 一切无经验内容、无法被证伪、纯粹诉诸神秘体验的解读方式，都不在本书的讨论范围之内。这并不是为了回避《易经》与神秘主义传统的深厚历史联系，而是因为本书的论域是"决策科学"，而一门严肃的决策科学必须可被检验、可被反驳、可被推进。

---

1.2 《易经》诠释史的多元维度：从卜筮到哲学到决策隐喻

The interpretative history of the *Yijing* is unusually rich precisely because the text is unusually underdetermined. The hexagram judgments (卦辞 *guaci*) and line statements (爻辞 *yaoci*) are short, often cryptic, and frequently admit multiple readings. This underdetermination is not a defect to be corrected but a feature to be explained: the text functions as a *generative grammar* for situational interpretation rather than a closed propositional system. The brief survey that follows identifies the major interpretive layers that have accumulated over three millennia, with particular attention to those features that will reappear in our later formal reconstruction.

先秦的卜筮原型（Pre-Qin Divinatory Prototype）. The earliest stratum of the *Yijing*, associated with the Western Zhou court and reconstructed in works such as 李学勤's *周易溯源* (2006), treats the 64 hexagrams as a repertoire of situational templates invoked by diviners in the field. The *Zuozhuan* (《左传》) and *Guoyu* (《国语》) preserve multiple case studies in which a hexagram is drawn for a specific question, and the resulting judgment is read against the situation at hand. Three features of this original usage are decision-theoretically significant: (1) the hexagram is read *situationally*, not universally—the same hexagram yields different readings for different inquirers at different times; (2) the moving lines (动爻 *dongyao*) introduce an element of *local perturbation* into an otherwise stable pattern, indicating that the situation is not in equilibrium; (3) the judgment and commentary are typically *prescriptive but conditional*—they say what will tend to happen and what posture (often 贞 *zhen*, "perseverance") is appropriate, without pretending to mechanical prediction. These features will resurface, with much greater formal precision, in chapters 4 and 8 when we discuss robust decision making and adaptive heuristics.

汉代象数学（Han Dynasty Xiangshu Systematics）. Beginning in the late Warring States period and culminating in the Han, a remarkable theoretical edifice was constructed around the *Yijing*: the *xiangshu* (象数) "image-number" school, associated with figures such as 孟喜, 京房, 邵雍, and 扬雄. This system maps the 64 hexagrams onto cosmological structures—the 24 solar terms, the lunar mansions, the five elements, the calendar—producing a comprehensive correlative cosmology. For present purposes, the *xiangshu* school contributes two key ideas. First, the *sequence* of hexagrams (卦序 *guaxu*) is treated not as arbitrary but as a *state-transition graph* describing the possible paths of change; the famous *Sheng Gua* (生卦) and *Xiang Gua* (相卦) sequences in 邵雍's *Qiantu Xianti* (《观物外篇》) are early examples of what we would now call *dynamical systems over discrete state spaces*. Second, the *xiangshu* tradition develops an elaborate apparatus of "correspondence" (应 *ying*, 配 *pei*) between lines, trigrams, and hexagrams, which formalizes the intuition that *meaning emerges from relation, not from isolated elements*. This relational ontology is structurally akin to what Bertalanffy (1968) called "systemic" thinking and what contemporary complexity science calls "relational emergence" (cf. Mitchell, 2009).

宋明义理之学（Song-Ming Yili Hermeneutics）. The Song Confucian reconstruction, especially 程颐's *Yizhuan* (《易传》, 1099) and 朱熹's *Yixue Qimeng* (《易学启蒙》, 1186), shifts emphasis from cosmological correspondence to moral and metaphysical principle (理 *li*). For 程颐, each hexagram illustrates a *principle of conduct* (处事之道), and the *Yijing* becomes a manual of ethical reasoning under conditions of change. Zhu Xi develops a sophisticated distinction between the "image" (象) and the "principle" (理), arguing that the former is a pedagogical vehicle for the latter. For our purposes, the *yili* school contributes the insight that *interpretation is irreducibly perspectival*—the meaning of a hexagram is not a property of the text alone, but emerges from the interaction between text and interpreter. This is, in effect, a *constructivist* and even *enactivist* theory of textual meaning, and it anticipates by seven centuries the constructivist turn in cognitive science (von Glasersfeld, 1995; Maturana & Varela, 1980). The decision-theoretic upshot is that the same "state" (hexagram) admits different "decisions" depending on the agent's position, goals, and resources—precisely the kind of context-dependence that modern contextual bandit and reinforcement learning algorithms must explicitly engineer (cf. Sutton & Barto, 2018).

清代汉学与近代转型（Qing Hanxue and Modern Transitions）. The Qing dynasty witnessed a return to philological rigor, exemplified by 惠栋, 张惠言, and 焦循's monumental *Yixue Sanshu* (《易学三书》). 焦循 in particular developed what may be called a *calculus of relations* between hexagrams, using operations analogous to addition and substitution to derive new hexagrams from old. This is the closest the pre-modern tradition comes to formal algebra. The modern period then wit