You need someone. Not romantically, not as a friend — you need what they have. Their access, their budget line, their signature, the thing you cannot do yourself.
And you will not say it that way, even in your own head. You will call it a partnership. A collaboration. Working together. Greene puts his finger on exactly this reflex:
As a description of human relationships, however, the word "use" has ugly connotations, and in any case we always like to make our actions seem nobler than they are. We prefer to think of these interactions as relationships of assistance, partnering, friendship.1
Then the claim the whole chapter rests on:
This is not a matter of mere semantics; it is the source of a dangerous confusion that will harm you in the end. . . . The art of forming alliances depends on your ability to separate friendship from need.1
The argument is not that you should be cynical. It's that you already have a need, and dressing it as affection means you will choose badly — you will pick the person you like over the person who has the thing.
Greene's diagnostic is the most immediately usable part of the chapter, because each error is one most people can recognise having made.
Too many. The instinct that more allies is better. Having numerous allies increases the chances we will become entangled in other people's wars.1 Every alliance is also a claim on you, and claims compound.
One big one. The opposite error, and the more seductive. Allies like that tend to get what they can from us and then drop us when our usefulness is exhausted1 — and Greene's example is pointed, because it is his own hero: just as Louis dropped the Swiss. Machiavelli's line is quoted directly: the forces of a powerful ally are useful to those who have recourse to them but are perilous to those who become dependent on them.1
The friendliest. Choosing the person who seems warmest and most likely to be loyal. Our emotions lead us astray.1
And the positive formulation:
the perfect allies are those who give you something you cannot get on your own. They have the resources you lack. They will do your dirty work for you or fight your battles. Like the Swiss, they are not always the most obvious or the most powerful.1
That last clause is the real content. The right ally is defined by the shape of the gap, not by their standing — which is why Louis XI, with a weak army, needed the Swiss and not England.
The construction rule follows:
Be creative and look for allies to whom you in turn have something to offer as well, creating a link of self-interest.1
And then the payoff Greene actually cares about, which is about the ending rather than the beginning:
To lose such allies of convenience will not destroy you or make you feel betrayed. You must think of them as temporary tools. When you no longer need such tools, there is no love lost in dumping them.1
Read that twice, because it contains the chapter's whole moral position and its whole vulnerability. The stated benefit of an interest-based alliance is emotional insulation: it cannot hurt you when it ends. That is a real and underrated advantage, and it is also an argument constructed entirely from the perspective of the person doing the dumping.
The chapter's sharpest distinction comes out of the Bowen case rather than the Louis XI one, and it is more useful than the cynical material around it:
A false alliance is created out of an immediate emotional need. It requires that you give up something essential about yourself and makes it impossible for you to make your own decisions.1
A true alliance is formed out of mutual self-interest, each side supplying what the other cannot get alone. It does not require you to fuse your own identity with that of a group or pay attention to everyone else's emotional needs. It allows you autonomy.1
That is a much better test than "is this person useful," and it is not really cynical at all. The question is whether the arrangement leaves you able to decide things. Groups that demand fusion produce alliances you cannot exit and cannot think clearly inside — see Bowen and the Refusal to Take Sides.
Greene's instruction for those groups is behavioural rather than attitudinal: you deftly avoid false alliances by taking provocative actions that make it impossible for people to entrap you. Shake up the dynamic, target the troublemakers and controllers, and then — once you can stay rational inside the group — you can seem to join an alliance without worrying about your emotions running away with you.1
And the reward he promises is worth marking, because it is the opposite of what most people expect from refusing to take sides:
as the person who is simultaneously autonomous and part of the group, you will become a center of gravity and attention.1
The chapter's historical frame is supplied by Machiavelli's Discourses, quoted at length, and it is bleaker than anything Greene writes himself.
Rome, he says, did not overlook fraud, and could not at the start have been more deceitful than she was in the means she took to acquire allies, since under this title she made them all her servants. She used the Latins' arms to subjugate the neighbours and build her reputation, and having subdued them, grew until she could beat anyone.1
The line that lands:
Nor would the Latins ever have realised that in reality they were mere slaves, if they had not seen the Samnites twice defeated and forced to accept Rome's terms.1
The Latins found out what the alliance had been by watching what happened to somebody else. That is the actual mechanism of the strategy at scale, and Greene reproduces it approvingly without asking which side of it his reader is likely to be on. The vault holds the Roman case itself at Divide and Rule and the Central Position.
Three Aesop fables sit in this chapter, and unusually, they carry positions the prose does not.
The Dog, the Cock and the Fox. The cockerel, threatened by the fox, sends it to wake the "doorkeeper" sleeping at the foot of the tree — who is a dog. Aesop's moral: sensible men, when their enemies attack them, divert them to someone better able to defend them than they are themselves.1
The Fox and the Billy-Goat. The fox, stuck in a well, talks the goat into climbing down, then climbs out on his horns and leaves, ignoring the shouted reproach about their agreement of mutual assistance.1
The Lion and the Wild Ass. They hunt together, the lion divides the spoils into three, and takes all three — the third on the grounds that refusing him would be a great source of harm to you. Moral: always calculate your own strength, and not enter into an alliance with people stronger than yourself.1
Notice what happens across the set. The first is advice for the weak. The second is the technique from the exploiter's side. The third contradicts the chapter's own advice about picking powerful allies to fill your gaps — and it is the only piece of text in this chapter written from the position of the ally who gets used.
The chapter closes its Keys with a move worth naming for its own sake:
you will of course be attacked for playing the Alliance Game. People will accuse you of being feckless, amoral, treacherous. Remember: these charges are strategic themselves. They are part of a moral offensive (see chapter 25). To advance their own interests, your accusers are trying to make you feel guilty or look bad. Do not let them get to you.1
Greene is using his own Strategy 25 to inoculate the reader against criticism of Strategy 27. Any moral objection to the chapter is reclassified, in advance, as a manoeuvre by someone with an interest.
That is an unfalsifiable defence, and it is deployed by cross-reference to another chapter of the same book. It is also exactly the technique Strategy 30 attributes to Machiavelli — building a text so that the reader's resistance is disarmed before it forms. Worth holding onto: the book defends itself with its own instruments, and does not mention that this is what it is doing.
The practical rider that follows is more honest: the only real danger is that your reputation will eventually keep people from making alliances with you1 — which concedes that the strategy has a cumulative cost, and then dismisses it on the grounds that self-interest rules the world.
You have a need you cannot meet alone, and you're about to go to the person you get on with best.
Stop and name the gap in one sentence, without a person's name in it. I cannot get a decision out of legal. I have no distribution. I don't have the credential this room respects. The sentence has to describe a capability, because that is what determines who the right ally is, and it is almost never the person you'd choose socially.
Then find who fills that specific gap and check what you have that they need. If you cannot state their side of the trade, you do not have an alliance — you have a favour, and favours run out at exactly the wrong moment. Louis needed an army; the Swiss needed a patron against Burgundy. That is a link, and it held for years because neither side was relying on affection.
Then keep the arrangement legible to yourself. Write down, at the start, what it is for and roughly when it ends. This sounds cold and it is actually the protective part: an alliance whose purpose you have never articulated is one you will keep resourcing long after it stopped paying, out of a loyalty nobody asked you for.
Now run the false-alliance test on every group you're already inside. Not do I like these people but: can I still make my own decisions here? If holding a position would cost you your standing in the group, you are in a false alliance, and the price is being unable to think.
And then decide, deliberately, where you sit on the ending. Greene's advice is to dump tools when they stop being useful and feel nothing. You are entitled to a different answer, but make it a decision rather than a drift — because the chapter's own Reversal tells you to check a prospective ally's history for signs of greed or of using people without giving in return, and the behaviour it recommends is what that check is designed to catch.
The Louis XI campaign, Rome's Latin League, and the Warring States alliance logic are all accurate as history. Bowen's family-systems work is real and well documented. [POPULAR SOURCE]
The chapter recommends the behaviour its own Reversal tells you to screen for. The Keys instruct you to treat allies as temporary tools and to dump them with no love lost. The Reversal instructs you to look at your prospective allies' past for signs of greed or of using people without giving in return, and to be wary of people who speak well, have apparently charming personalities, and talk about friendship, loyalty, and selflessness.1 A reader who follows the Keys becomes exactly the person the Reversal warns about, and the chapter never notices — it addresses the reader only as the one running the screen, never as the one being screened.
The moral pre-emption is doing structural work. By routing all criticism of this chapter through Strategy 25's exterior-manoeuvre frame, Greene makes objection self-discrediting. The technique is real and effective; it is also the one thing in the chapter with no counter-case, no Reversal, and no cost stated.
And the Louis XI case is admired for a cost paid entirely by an ally. Greene's summary is that Louis crushed his opponents by getting others to do his work for him and destroyed Burgundy without wasting a single French soldier.1 The soldiers who were wasted were Swiss, at Grandson and in Lorraine, in a war Louis engineered and then formally stepped out of by signing a truce with the duke — a truce he knew would outrage the Swiss, and about which Greene writes: what did Louis care? Friendship meant little to him; the Swiss would fight to defend their land with or without him.1 The chapter names the abandonment and admires it in the same breath. This is the book's running SIGNAL running through the ally rather than through the enemy.
Against Strategy 7: Morale Strategies and Strategy 5: The Command-and-Control Strategy there is a new and load-bearing contradiction. Strategy 7 is an extended argument for building devoted followings through emotional bonds — Cromwell's true believers, LBJ's eighteen-hour staff, Napoleon wandering the campfires and making a drum major a baron overnight. Strategy 27 states flatly that alliances infected with emotions, or with ties of loyalty and friendship, are nothing but trouble.
So Greene teaches deliberate loyalty-cultivation downward and deliberate loyalty-avoidance sideways, twenty chapters apart, and supplies no rule for the boundary. The reconciliation he never writes is presumably that you receive loyalty and do not give it — which, stated plainly, is a considerably harder position than either chapter takes on its own, and it is the position the two chapters jointly entail.
Against Strategy 21: The Diplomatic-War Strategy and Philip of Macedon and Negotiation as Advance this is convergence and escalation. Philip treats trust as a purchasable commodity inside a negotiation; this chapter extends the same accounting to every standing relationship you have.
Against Strategy 17: The Divide-and-Conquer Strategy the alliance-breaking material here is the same operation aimed outward, and the pairing produces the fuller picture: Strategy 17 attacks the joints of a structure, Strategy 27 attacks the joints between parties, and the second is cheaper because the joint is already under tension from competing interests.
Psychology → The group that requires fusion is the one you cannot assess. Greene's false alliance — the one born of emotional need, that requires giving up something essential and makes independent decision impossible — is a description of enmeshment from the strategist's side. The vault's Family System Roles holds the mechanism: a system under stress reorganises around its dysfunction and assigns each member a role, not by decision but by developmental pressure, because that configuration produces the most safety available.
The insight the pairing produces that neither gives alone: the reason a false alliance is strategically disabling is that the role precedes the reasoning. You are not choosing badly inside the group; you are occupying a position that was allocated to you, and every judgement you make from inside it is shaped to keep the system stable. Which reframes Greene's advice usefully — the instruction to "take provocative actions that make it impossible for people to entrap you" is not aggression, it is the only available way to find out what you actually think, because you cannot reason your way out of a role while continuing to perform it. That is precisely what Bowen's letters were for, and it is why he described the result as a feeling of lightness rather than of victory.
Business → Dependency risk is the same problem in two vocabularies. Greene's warning about the single powerful ally — useful to those who have recourse to them, perilous to those who become dependent — is customer concentration, platform risk and single-supplier exposure stated in fifteenth-century terms.
What the pairing produces: the danger is not the size of the counterparty but the absence of a second option, and the tell is that you have stopped being able to say no. A large partner you could replace in a quarter is an asset; a small one you could not is a liability, and headline numbers do not distinguish them. This gives a sharper diagnostic than concentration percentages: ask what you would have to stop doing if this relationship ended next month, and how long the substitution would take. If the honest answer is that you would have to change what the business is, the relationship is already running you — and note that this is the same test as the false-alliance test one domain over, since both are asking whether you can still make your own decisions.
The test of an alliance is not whether it is useful but whether you can still make your own decisions inside it — which means the dangerous ones are not the exploitative arrangements you can see, but the warm ones you would lose your standing by disagreeing with.