📚 Differing significance of states in global affairs and how – and why – state power is classified | 状态在国际事务中的不同重要性以及国家影响力的分类方式与原因
In the theatre of global affairs, nations do not all wield equal influence. Some states act as dominant superpowers, shaping international norms, while others remain peripheral, absorbing decisions made elsewhere. This stratification invites a fundamental question: how – and why – is state power classified? A surprisingly powerful lens through which to analyse such differing significance is the mathematical theory of Markov chains, where 'states' represent discrete conditions of a system and their 'power' is measured by long‑term occupation probabilities. In this article, we explore how the classification of states in a Markov process mirrors the hierarchy of nations, and we formalise the criteria used to distinguish absorbing, recurrent, transient, and periodic states.
在国际事务的舞台上，各国的影响力并不等同。有些国家扮演着主导性超级大国的角色，塑造着国际规范，而另一些国家则处于边缘，被动地接受别处作出的决定。这种层级划分引出了一个根本问题：国家的影响力是如何——以及为何——被分类的？一个出人意料的强有力分析视角来自马尔可夫链的数学理论，其中“状态”代表系统的离散条件，而它们的“权力”则用长期占据概率来衡量。在这篇文章中，我们探索马尔可夫过程中状态的分类如何映照国家的等级，并形式化地区分吸收态、常返态、非常返态和周期态的标准。
1. The world as a state space | 世界作为状态空间
Consider the global political system as a collection of mutually exclusive conditions – superpower, regional power, developing nation, fragile state. At any moment, influence can shift from one category to another through events such as elections, economic crises, or military conflicts. This is precisely the intuition behind a Markov chain: a sequence of random variables X₀, X₁, X₂, … taking values in a finite or countable state space S, where the probability of moving to a future state depends only on the present state, not on the path taken to reach it. The memoryless property may seem reductive, but it captures a key dynamic: tomorrow's global influence depends primarily on today's balance of power.
将全球政治体系视作一组互斥的条件——超级大国、区域大国、发展中国家、脆弱国家。在任一时刻，影响力都可能通过选举、经济危机或军事冲突等事件从一种类别转移到另一种类别。这正是马尔可夫链背后的直觉：随机变量序列 X₀, X₁, X₂, … 在有限或可数状态空间 S 中取值，其中转移到未来状态的概率仅取决于当前状态，而与到达该状态的路径无关。无记忆性这一性质看似简化，却捕捉了一个关键动态：明天的全球影响力主要取决于今天的权力平衡。
P(Xₙ₊₁ = j | Xₙ = i, Xₙ₋₁ = iₙ₋₁, …, X₀ = i₀) = P(Xₙ₊₁ = j | Xₙ = i) = pᵢⱼ
The one‑step transition probabilities pᵢⱼ form a matrix P, and the n‑step transition probabilities are the entries of Pⁿ. By examining these, we can classify states and understand their long‑term significance.
一步转移概率 pᵢⱼ 构成矩阵 P，而 n 步转移概率则是 Pⁿ 的元素。通过考察这些矩阵，我们可以将状态分类，并理解它们的长期重要性。
2. Recurrent and transient states: permanent players vs. passing visitors | 常返态与非常返态：永久参与者与过客
A state i is called recurrent if, starting from i, the probability of ever returning to i is 1. In global‑affairs language, a recurrent state represents a nation‑type that, once entered, will certainly reappear in the future. A transient state, on the other hand, has a positive probability of being left forever after some finite time. The distinction is the difference between a permanent player on the world stage and a fleeting presence.
若从状态 i 出发，最终返回 i 的概率为 1，则称 i 为常返态。用国际事务的语言来说，常返态代表一种一旦进入就必定会在未来再次出现的国家类型。而非常返态则有可能在某个有限时间后被永远离开。这一区别就是世界舞台上永久参与者和短暂存在之间的差异。
transient states, this sum is finite. The global significance of a recurrent state endures; that of a transient state decays.
更形式化地，令 fᵢ 为从 i 出发链终将返回 i 的概率。若 fᵢ = 1，i 是常返的；若 fᵢ < 1，则是非常返的。对于常返态，访问次数的期望无穷大：Σₙ Pⁿᵢᵢ = ∞。对于非常返态，此和有限。常返态的全局影响力持续存在；非常返态的影响力则逐渐衰减。
3. Positive vs. null recurrence: influence that matters | 正常返与零常返：实质影响力
te spaces, all recurrent states are positive recurrent, but infinite chains can harbour null recurrent states – on the world stage is theoretically certain but always just over the horizon is striking.
在常返态中，可进一步细分。若平均返回时间 mᵢ 有限，则该常返态 i 是正常返的；若 mᵢ = ∞，则为零常返。在有限状态空间中，所有常返态都是正常返的，但无限链中可能存在零常返态——这些条件虽然理论上会重现，却总是在极为漫长的等待之后才发生，几乎毫无实际意义。这与那些理论上必定重返世界舞台、但总在遥远未来才能再现的小国非常相似。
Positive recurrent states carry the 'power' of reliable, fairly frequent recurrence. In Markov chain theory, they are the only states that can have a non‑zero limiting probability πᵢ, making them the true power centres of the system.
正常返态拥有可靠且相对频繁重现的“权力”。在马尔可夫链理论中，它们是可以拥有非零极限概率 πᵢ 的唯一状态，因而是系统的真正权力中心。
4. Absorbing states: the ultimate superpower | 吸收态：终极超级大国
recurrent state: once entered, the probability of leaving it is zero (pᵢᵢ = 1). In global‑ s. Alternatively, it might model a collapsed state that cannot recover. The rest of the system, over time, will eventually be absorbed into such a state with probability 1 if the chain is absorbing.
吸收态是常返态的一个特例：一旦进入，离开的概率为零（pᵢᵢ = 1）。用国际事务来类比，吸收态可以代表一个已变得如此主导、以至于没有别的国家能挑战其地位的霸权超大国。或者，它也可以模拟一个无法复苏的崩溃国家。在一个吸收链中，系统的其余部分终将以概率 1 被吸入这样的状态。
For an absorbing Markov chain, the fundamental matrix (I – Q)⁻¹ gives the expected number of times the process spends in each transient state before absorption. The 'power' of an absorbing state is absolute; it dictates the ultimate fate of all other states.
对于吸收马尔可夫链，基本矩阵 (I – Q)⁻¹ 给出过程在吸收前处于每个非常返态的期望次数。吸收态的“权力”是绝对的；它决定了所有其他状态的最终命运。
5. Periodicity: cyclical influence | 周期性：循环的影响力
it is periodic. In world politics, some states might exert influence in predictable cycles – ate’s power is rhythmically constrained, never establishing a steady, continuous presence.
如果返回状态 i 只能在 d(i) 的倍数上发生，则状态 i 具有周期 d(i)，其中 d(i) 是所有使得 Pⁿᵢᵢ > 0 的 n 的最大公约数。若 d(i) = 1，状态是非周期的；否则是周期的。在世界政治中，一些国家的影响力可能以可预测的周期出现——比如联合国安理会非常任理事国的轮换或周期性的经济繁荣。周期性常返态的权力受到节律约束，永远无法建立起稳定而持续的存在。
Periodicity often complicates the calculation of long‑run behaviour, because limits of Pⁿ may not exist when states cycle through a fixed pattern of lengths. Aperiodicity is required for convergence to a stationary distribution.
周期性常常使长期行为的计算复杂化，因为当状态按照固定长度的模式循环时，Pⁿ 的极限可能不存在。要收敛到平稳分布，必须是非周期的。
6. Ergodicity: the stable, predictable hegemon | 遍历性：稳定可预见的霸权
n where all states are ergodic, there exists a unique stationary distribution π = (π₁, π₂, …, πₖ) such that πP = π and Σ πᵢ = 1. This distribution reveals the long‑ nfluence reaches a stable equilibrium, allowing analysts to predict, for example, that a certain type of regional power will dominate 40% of future time steps.
若一个状态正常返且非周期，则称为遍历状态。在一个不可约链（所有状态都遍历）中，存在唯一的平稳分布 π = (π₁, π₂, …, πₖ)，满足 πP = π 且 Σ πᵢ = 1。该分布揭示了长期花在每个状态上的时间比例，且与起点无关。在我们的比喻中，遍历状态代表系统中那些影响力达到稳定均衡的组件，使分析家可以预测，例如，某种类型的区域大国将在未来的时间步骤中占据 40% 的份额。
limn→∞ Pⁿᵢⱼ = πⱼ > 0
This is the mathematical equivalent of a stable, multi‑polar world where power is shared in a balanced, predictable manner.
这相当于一个稳定、多极世界的数学版本，其中权力以均衡且可预测的方式共享。
7. Communicating classes: alliances and blocs | 互通类：联盟与集团
chable from the other. Communication is an equivalence relation that partitions the state space into communicating classes. A closed class, once entered, can never be left. In global affairs, this models a tight alliance or integration bloc – , it may find itself permanently classified within that bloc, unable to unilaterally exit without altering the very structure of its interactions.
如果 i → j 且 j → i（即彼此可互达），则称状态 i 与 j 互通（i ↔ j）。互通是一种等价关系，将状态空间划分为互通类。一个闭类一旦进入，便永远无法离开。在国际事务中，这可以模拟一个紧密的联盟或一体化集团——一旦某个国家加入货币联盟或集体安全安排，它可能发现自己被永久地归类在该集团内，无法在不改变互动结构的情况下单方面退出。
An irreducible chain consists of a single communicating class. If there are multiple closed classes, the chain exhibits path‑dependent fate – the initial condition can completely determine which echelon of ‘state power’ the system ultimately settles into.
不可约链仅包含一个唯一的互通类。如果存在多个闭类，链就会表现出路径依赖的命运——初始条件可能完全决定系统最终落入哪一层次的“国家权力”集团。
8. How state power is ‘classified’ in practice | 实践中“国家影响力”如何被分类
Just as international relations theorists classify nations using indices of military spending, GDP, diplomatic reach, and soft power, Markov chain analysis classifies states using the simple but rigorous criteria of return probability, expected return time, and period. The calculations rest on transition probabilities, which themselves are derived from observed data or structural assumptions. The table below maps global‑affairs concepts to Markov chain properties of a state.
正如国际关系理论家使用军费开支、GDP、外交影响力和软实力等指标对国家进行分类，马尔可夫链分析使用返回概率、期望返回时间和周期性这些简单而严格的标准来对状态进行分类。这些计算基于转移概率，而转移概率又源自观察到的数据或结构性假设。下表将国际事务概念映射到状态在马尔可夫链中的性质。
| Global affairs concept | Markov chain property |
| Permanent superpower | Positive recurrent, aperiodic (ergodic) |
| Unchallengeable hegemon | Absorbing state |
| Rising but fragile power | Transient state |
| Cyclical influencer | Periodic recurrent state |
| Bloc member | State in a closed communicating class |
9. A worked example: a three‑state model of global influence | 实例演练：一个全球影响力的三状态模型
Consider a simplified world with three categories: Superpower (S), Regional Power (R), and Fragile State (F). The transition matrix P is given below.
考虑一个简化的世界，有三种类别：超级大国 (S)、区域大国 (R) 和脆弱国家 (F)。转移矩阵 P 如下所示。
S R F
S [ 0.9 0.1 0.0 ]
R [ 0.2 0.7 0.1 ]
F [ 0.0 0.4 0.6 ]
fication of each state. State S leads to R (p=0.1) and to itself, but can it return from R? o R or staying in F are possible, so F → R, but R is not in {F} alone. Actually, from F, pFR = 0.4, pFF = 0.6, so F can reach R, but R cannot reach F? Wait, R → F with p=0.1, so R can reach F. Then F can reach R, so they should communicate! Let’s re‑ (p=0.1). Thus F ↔ R. Since S ↔ R, all three states communicate. The chain is irreducible.
我们检查每个状态的分类。状态 S 可到 R (p=0.1) 和自身，但它能从 R 返回吗？从 R 我们可以以概率 0.2 回到 S，所以 S ↔ R。状态 F 从 R 可达，但从 F 我们不能到达 S（到 S 的概率为 0）。因此类为 {S, R} 和 {F}。{S, R} 是闭互通类，因为没有路径离开它去 F。{F} 不是闭的：从 F 出发，只能转移到 R 或留在 F，所以 F → R，但 R 不在单独的 {F} 中。其实，从 F，pFR = 0.4，pFF = 0.6，所以 F 可到 R，但 R 也能到 F：R → F (p=0.1)，因此 F 可达 R，R 可达 F，所以 F ↔ R。由于 S ↔ R，所有三个状态互通。该链是不可约的。
All states are recurrent (finite, irreducible chain) and aperiodic (self‑ ry distribution by solving π P = π. The limiting probabilities π_S, π_R, π_F give the long‑ the system is a superpower, regional power, or fragile state. These represent the relative ‘significance’ of each category in steady state.
所有状态都是常返的（有限、不可约链）且非周期（S、R、F 都有自环，保证非周期性）。因此该链是遍历的。我们通过求解 π P = π 找到平稳分布。极限概率 π_S、π_R、π_F 给出了系统长期处于超大国、区域大国或脆弱国的时间比例。这些代表了稳定的状态下每个类别的相对“重要性”。
10. Why classifying state power matters – in both domains | 为何对影响力分类很重要——在两个领域皆然
In international relations, classifying state power helps anticipate alliance shifts, potential conflicts, and the stability of the international order. In Markov processes, classification determines whether long‑run predictions are possible, whether the chain will eventually settle, or whether it will endlessly cycle. Both disciplines benefit from a rigorous taxonomy: it turns qualitative intuition into quantitative analysis. Knowing that a state is transient, for example, tells us that its current prominence is a passing phase, just as a temporary economic miracle may fade without fundamental structural change.
在国际关系中，对国家影响力进行分类有助于预见联盟变化、潜在冲突和国际秩序的稳定性。在马尔可夫过程中，分类则决定了长期预测是否可能、链最终是否会稳定下来，或者是否会无休止地循环。两个学科都受益于严格的分类学：它将定性的直觉转化为定量分析。例如，知道一个状态是非常返的，就告诉我们它当前的显赫只是一个过渡阶段，就像一次临时经济奇迹可能在没有根本性结构变革的情况下消退。
Ultimately, the differing significance of states in any complex system – be it global politics or a stochastic process – rests on their capacity to influence future trajectories. The mathematical framework of state classification gives us a precise language to describe and measure this capacity, making it an indispensable tool for A‑Level Further Mathematics students exploring the dynamics of chance and change.
归根结底，在任何复杂系统中——无论是全球政治还是随机过程——状态的不同重要性都取决于它们影响未来轨迹的能力。状态分类的数学框架为我们提供了一套精确的语言来描述和测量这种能力，使之成为 A‑Level 进阶数学学生探索机会与变化动态时不可或缺的工具。
Published by TutorHao | Further Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图