Picture the Russian Communist organizational map across roughly thirty years, from the 1905 stirrings to the consolidation of Soviet power. At each phase, the Bolsheviks face a primary enemy and a secondary one. They ally with the secondary against the primary. They win. They turn on the secondary. They identify the next primary. They ally with whatever new secondary is available. Repeat.
Siu names the sequence directly. "At the beginning, they united with the liberal bourgeoisie against the Tsarists, then with the Mensheviks and Social-Revolutionaries against the liberal bourgeoisie, then with the smaller farmers against the large landowners, and finally with the peasants against the kulaks."1
Read the sequence again. Each ally became a target. The cycle did not pause for sentiment, for promised reciprocity, or for the shared blood of the prior victory. The promise of the alliance was always conditional on the alliance's use-value at that phase. When the use-value expired, the ally became the next obstacle. The Russian Sequential Elimination is one of the cleanest examples in modern political history of what coalitions actually are when read against their own internal arithmetic rather than against the ceremonial language of mutual commitment.
Siu opens the section with the historical generalization. "History has shown that most one-time national allies become enemies sooner or later. It all depends on the exigencies of the moment."2
The empirical evidence he cites is hard. "In an analysis by Melvin Small and J. David Singer of fifty interstate wars and forty-three colonial and imperial conflicts between 1816 and 1965 in which more than a thousand battle fatalities occurred, a fifth of the 209 pairs of opposing enemies had previously been allies. Of the 136 nations with more than one experience in war, four-fifths of the ninety-five pairs with some experience as enemies had been allies at least once. Citing hundreds of treaties between 1535 and 1968, Laurence Beilensen showed that political treaties will be broken if it is in the national interest to do so. He concluded that no reliance should be placed on the long-term honoring of treaties by any nation."3
Four-fifths of the studied enemy-pairs had been allies at least once. The data does not say allies might become enemies. It says, in this dataset, they will, with high enough frequency that any operator who treats coalitions as durable is operating against the historical record.
Siu's response is not to refuse coalitions. Coalitions are sometimes essential. The response is precautionary architecture — five specific practices that turn coalition from a romantic gesture into a managed exposure.
Precaution 1 — Make a prior determination as to whether outside help is essential.
"Unless circumstances leave no alternative, it is preferable not to seek allies. The offer for help might come from the wrong quarters with shady ulterior motives, be of the wrong kind, or prove unreliable at a critical moment."4
The first precaution is the most ignored. Operators reach for coalition because help is offered, not because help is required. Once the coalition is formed, withdrawing is harder than not entering. Siu names the worst version of the trap: "You should be wary of an alliance with a party who can do without your help, who may well end up disproportionately more powerful than before, leaving you exhausted in the process. You might then be reduced to the status of being a satellite or even being completely assimilated by him or her."5
The second worst version: "It is also prudent not to join a strong movement, which seeks only your supplementary presence to add mass to its attack on a common opposition for demands of its own. Even if it is willing to add your demand on the list, yours will still remain an afterthought and might be the first to go as a bargaining chip during the subsequent negotiation."6
Both warnings are about asymmetric coalition. The party who needs you less keeps the agenda. Your demand becomes their bargaining chip. By the time you notice, you have already spent the resources that would have let you operate independently.
Precaution 2 — Establish a prior agreement on contributions and rewards at every stage.
"Be suspicious of those who refuse to be committed to anything, as well those who are willing to promise the moon. Both types are unreliable when the going gets difficult."7
The named-and-quantified contribution-and-reward schedule is the coalition's load-bearing structure. Without it, every disagreement becomes a renegotiation. The party with the most patience for renegotiation usually wins those, which is rarely the party Siu's reader is supposed to be.
The two unreliable types — refusal-to-commit and promise-the-moon — are diagnostically different but operationally the same. The refusal-to-commit partner will exit at the first hard moment because nothing was binding them in. The promise-the-moon partner will exit at the first hard moment because the discrepancy between their promised input and their actual capacity becomes visible, and they would rather leave than be exposed. Avoid both.
Precaution 3 — Maintain a continuous followup.
"Should some member of the coalition appear to be intentionally getting too far out of line, take immediate and effective compensatory moves. Should unforeseen problems necessitate a revision of the allocation of contributions, roles, and rewards, reopen the original agreement formally with all of the collaborators present."8
Continuous followup is what Siu has substituted for trust. The original agreement does not survive contact with reality. The coalition member who is "getting out of line" is not betraying you; they are responding to changing exigencies. The question is whether the response will be metabolized through formal renegotiation (Siu's prescription) or through covert defection (the path of least resistance for most parties). The leader who does not maintain followup is choosing covert defection by default.
Precaution 4 — Protect against the two-on-one strategy.
"If you contemplate using it yourself against others, then go on the assumption that your erstwhile collaborators are already suspicious of the situation."9
This is where the Russian Sequential Elimination case gets named. The two-on-one is not exotic; it is the normal grammar of multi-party coalition. Three parties stay aligned only as long as no two of them find joint elimination of the third more profitable than continuing the three-way. The moment that calculus tips, the coalition reorganizes into the two-on-one configuration. Siu's instruction is symmetric. If you intend to do this to others, assume they suspect you. If you do not, assume someone else is contemplating it against you.
Precaution 5 — Conduct your operation so that at the time of joint victory your own reserves are not unbalanced.
"The closer to victory against the common foe you stand, the more closely should you scrutinize your own fortunes, adjust your balance, and toughen your resiliency."10
The fifth precaution names the moment most operators get killed. The coalition wins. The shared enemy collapses. In the moments just before and just after victory, the coalition's internal arithmetic shifts: the raison d'être (defeating the common foe) is gone, and the only remaining structure is the relative power of the coalition members. The party who arrives at victory most depleted is the easiest to dispatch in the post-victory reorganization. The Russian Communists understood this. Their phase-by-phase elimination was not opportunism after the fact; it was planned during each prior phase, with reserves preserved for the next phase even while the current alliance was deployed.
Siu closes: "Only in this way can you really reap the harvest of coalitions."11
Scene 1 — The Help-Required Test. Before accepting any offered alliance. Sit with the question: if this offer were withdrawn tomorrow, would I be unable to achieve my objective? If the answer is no — if I have alternatives, even worse ones — the offer is a temptation, not a necessity. Most operators accept temptations because the alternative looks like work. The work is usually cheaper than the alliance bill that arrives later.
Scene 2 — The Schedule Negotiation. First week of any coalition. Before any joint operation begins. Sit with the partners and write down: at the end of phase one, who delivers what and who receives what? At the end of phase two, what changes? At the end of victory, how is the spoil divided and how is the post-coalition relationship structured? If anyone refuses to negotiate this — anyone — re-read precaution two and reconsider whether to enter the coalition at all.
Scene 3 — The Quarterly Followup. End of every quarter, on every active coalition. List the inputs each partner has contributed during the quarter, against the inputs they had agreed to contribute. List the outputs each partner has received, against the outputs they had been promised. Where there are gaps, name them in writing and circulate the document to all coalition members. Coalition members who object to having the gaps named are signaling their intention to operate the coalition's discrepancies asymmetrically. The signal is information.
Scene 4 — The Two-on-One Map. Once a year, on every active multi-party coalition. List every partner pair other than you-plus-someone. For each pair, ask: is there a joint move these two could make that would benefit them more than continuing the current configuration? If the answer is yes, you are the third member of an unstable triangle and your reserves had better be fresh. If the answer is no, the configuration is stable for now; revisit next year.
Scene 5 — The Pre-Victory Audit. At the moment victory becomes visible — when the common foe is two months from collapsing rather than two years — stop. Audit your own resources. Are you depleted? Are your partners less depleted? If both are true, the post-victory reorganization will not favor you. Either replenish in the time available or begin negotiating the post-victory settlement before the victory arrives, while your partners still need you.
Coalitions degrade along observable markers, often months before any party formally defects. The early signs:
When two of the five are present, defection is in advanced preparation. When all five are present, the coalition is already over; only the formal announcement remains.
The five-precautions framework is empirically robust across coalition types — interstate alliances, corporate joint ventures, political coalitions, social-movement coalitions. The Small/Singer database is the most-cited piece of evidence; the Beilensen treaty study is the second. Beyond these, the framework's predictive power is highest at the precaution-one stage. Operators who decline coalitions when help is not strictly required avoid most of the downside the other four precautions are designed to manage. Operators who say yes to every offered alliance accumulate a coalition portfolio that becomes structurally unmanageable, regardless of how diligent their followup.
The Russian Sequential Elimination case is the canonical worked example because it is unusually well-documented and unusually clean. Other historical cases follow the same template at different scales: the Whig-Tory rotations in eighteenth-century British politics, the multiple Ottoman-European alliances of the nineteenth century, the Comintern's serial alliances and ruptures during the interwar period, the post-2003 Iraqi political coalitions. The framework predicts these cases as instances of a structural pattern, not as anomalies.
Siu's framework treats coalitions as inherently expedient. This is empirically defensible at the level of nations and major political parties. It is less defensible at the level of long-running ideological movements that sustain genuine cross-faction commitment for decades. The civil rights coalition in the 1960s, the South African anti-apartheid coalition, the European Christian-Democratic alliances of the postwar period — these include cases where coalition partners maintained commitment well beyond the exigencies of the moment. Siu's framework, applied uncritically, would have misread these cases.
A second tension lives in Precaution Four. Siu instructs the reader to expect the two-on-one move from others and to consider deploying it themselves. The instruction is symmetric, but the moral weight of deploying it is not equivalent to the moral weight of expecting it. A reader who internalizes only the deployment instruction risks producing the very pattern the precautions are designed to defend against. The coalition framework can be read as a defense manual or as an offense manual. Which reading the operator adopts becomes the page's most consequential downstream decision.
Two domains illuminate the coalition framework from beneath the operator's frame. One supplies the metaphysical claim that makes the empirical pattern intelligible. The other supplies the individual-scale psychological mechanism that compounds, across thousands of micro-decisions, into coalition-scale defection.
Eastern-Spirituality — Anicca: Impermanence as Constant Flux
Picture a Burmese monk on a meditation cushion in 1955. He is watching the breath. He notices that each breath is a complete cycle — arising, sustaining, dissolving. There are no breaths that do not dissolve. Every formation, however briefly stable, is on the trajectory of its own dissolution. This is not a meditation theme; it is a description of how things are.
"Everything that arises passes away. Without exception. No storage, no exceptions, no escapes. What arises in one moment has dissolved by the next."12
Now apply this to coalitions. A coalition is a formation. It arises from the convergence of two or more parties' interests around a shared exigency. While the convergence holds, the coalition has structural integrity. The convergence does not hold permanently because the conditions producing it (the shared enemy, the shared scarcity, the shared opportunity) are themselves impermanent. As the conditions shift, the coalition's structural integrity dissolves. Beilensen's treaty study and Small/Singer's enemy-pair analysis are documenting the operational consequence of a metaphysical claim that Buddhist analysis made twenty-five centuries before Beilensen's first century of treaty data.
Siu's five precautions are the operational accommodation of anicca. They do not assume coalitions are betrayals waiting to happen; they assume coalitions are formations subject to dissolution, and they prescribe practices that make the operator's exposure to the dissolution survivable. Precaution one (refuse coalitions when not essential) is non-attachment to the formation itself. Precaution two (named contributions and rewards) is structural agreement that anticipates the formation's eventual dissolution and pre-distributes the costs and benefits. Precaution three (continuous followup) is the recognition that the formation requires active sustaining and will not maintain itself. Precaution four (two-on-one map) is the recognition that the formation's dissolution often follows internal reorganization rather than external attack. Precaution five (pre-victory audit) is the recognition that the formation's dissolution accelerates at the moment its founding condition (the shared exigency) is satisfied. See Anicca: Impermanence as Constant Flux.
What the pairing reveals — that neither concept produces alone — is the operator's relationship to coalition betrayal. Without anicca, coalition betrayal feels like personal failure: the partners broke faith, the operator misjudged character, the trust was misplaced. With anicca, coalition betrayal becomes the expected dissolution of an impermanent formation, neither personal nor surprising. The operator who has metabolized anicca can deploy Siu's precautions without bitterness, because the dissolution is not a moral event; it is the formation's nature. This is the difference between a tragic coalition operator (constantly betrayed, constantly surprised) and a cold coalition operator (always exposed, never surprised). The difference is whether the operator has fused their identity to coalition durability or has accepted that the formation was always going to dissolve and has structured their reserves accordingly.
Psychology — The Turncoat in Each of Us: A Self-Betrayal Taxonomy
Picture Meerloo's small barber in occupied Holland in 1940. He had spent decades cutting the hair of cabinet ministers and opposition leaders. He was "always very courteous and agreeable, eager to please his clients." When the Nazi general arrived in his chair, he lathered the man's face himself. When the SS officers followed, he served them. When the Nazi-collaborator organization invited him to buy a membership card, he paid the fee — "he thought of welfare as a special tax on business." When old acquaintances warned him this would be called collaboration, he answered: "I am a barber, and I live as a barber. I have absolutely no interest in politics. I only want to serve my clients."13
The barber did not betray anyone in any single moment. He betrayed by accumulation. Each decision (lather the general's face; serve the SS man; pay the membership fee; explain the payment as a tax) was small enough to not feel like a moral choice. The pattern of decisions, viewed from outside, was collaboration. Viewed from inside, it was just a barber being a good barber.
Coalition defection works the same way at scale. The partner who eventually defects rarely makes a single dramatic betrayal decision. They make a sequence of small decisions, each of which is locally rational and locally self-justified, that compound into a defection visible only retrospectively. The under-attendance at meetings, the bilateral conversations with outside parties, the gradual prioritization of their own interests over the coalition's — these are individual barber-scale decisions whose compounding produces the coalition-scale outcome. Each decision is internally explicable. The pattern is what the precautions are designed to detect. See The Turncoat in Each of Us: A Self-Betrayal Taxonomy.
What the pairing reveals is why precaution three is the operationally hardest of the five. Continuous followup catches the barber-scale decisions before they compound. Without continuous followup, the operator only sees the compounded outcome — the defection — and then must reckon with what to do after the fact. Followup is the technique by which coalition partners are kept from each becoming the small barber, drifting into a pattern of collaboration with whatever new exigency is in front of them. It is not a trust-display; it is a structural intervention against the natural human tendency, documented in the turncoat taxonomy, to follow the path of immediate convenience until the cumulative pattern becomes the identity. The five precautions do not assume coalition partners are bad-faith. They assume coalition partners are humans, and humans, under sustained pressure, drift into self-betraying patterns through a sequence of small decisions none of which felt like betrayal at the moment of making.
The Sharpest Implication
If the empirical record is correct — four-fifths of studied enemy-pairs were once allies — then coalition durability is a fiction operators tell themselves to make the present alliance feel meaningful. The fiction has uses. It motivates the joint work. It binds the parties to each other in the short term. But the fiction must not be confused with the structural reality, which is that the alliance was always temporary and the parties' interests will eventually diverge.
The implication for any reader currently in or considering a coalition is behave as if the coalition will end on a timeline you cannot fully predict. Build your reserves so that you can survive its end. Maintain your independent capabilities so that you do not become assimilated. Negotiate the terms of dissolution at the time of formation, not at the time of crisis. Treat your partners with full operational courtesy and zero strategic naivety. Most coalition tragedies happen to operators who could not hold these two postures simultaneously and chose courtesy over reserves. Reserves outlast courtesy.
Generative Questions