In 338 BC Rome beat its greatest enemy — the Latin League, a confederation of Italian cities formed specifically to block Roman expansion.
And then it had a problem that most victors handle badly.
Crush the league's members, and you leave a power vacuum. Some new enemy will grow in it, possibly a worse one. Swallow them whole, and you dilute Rome's power and prestige, and give yourself far too much territory to protect and police.1
The solution they came up with became the phrase — divide et impera — and the method by which they forged an empire.
They broke up the league, and then, crucially, did not treat the pieces the same.
No single city was left powerful enough to challenge Rome, which retained the central position.
(As the saying goes, all roads led to Rome.)1
The differentiation was not administrative untidiness. It was the mechanism.
The key to the system was that if an independent city proved itself loyal enough to Rome or fought well enough for Rome, it won the chance of being incorporated into the empire.1
Which converted the political landscape into something new:
The individual cities now saw it as more in their interest to gain Rome's favor than to ally themselves elsewhere. Rome held out the prospect of great power, wealth, and protection, while isolation from Rome was dangerous.1
And the outcome, in one sentence:
the once proud members of the Latin League now competed against one another for Rome's attention.1
Note what Rome did not have to do. There was no ongoing suppression, no garrison in every city, no policy of keeping anyone weak. Rome supplied a ladder and let the cities climb it against each other. The energy that had been directed at Rome was now directed sideways, and it was volunteered.
Greene's justification is a claim about what groups do on their own.
within any organization people naturally form smaller groups based on mutual self-interest—the primitive desire to find strength in numbers.1
Left unchecked, those subgroups become power bases that threaten the whole. And the danger is not disloyalty; it is arithmetic:
The formation of parties and factions can be a leader's greatest threat, for in time these factions will naturally work to secure their own interests before those of the greater group.1
The remedy has a precondition most people skip:
To do so you must first establish yourself as the center of power; individuals must know they need to compete for your approval. There has to be more to be gained by pleasing the leader than by trying to form a power base within the group.1
That sentence is the whole design. It is not about preventing coalitions. It is about making them a worse investment than the alternative.
When Elizabeth became queen, England was a nation divided — feudal remnants meant many competing power centres, and the court itself was full of factions.1
Her solution had two halves that had to run together.
Weaken the nobility by deliberately pitting one family against another.
And, at the same time, occupy the centre — making herself a symbol of England itself, the hub around which everything revolved.1
Inside the court she applied the same rule to individuals. When she saw that first Robert Dudley and then the Earl of Essex believed themselves her favourites, she quickly cut them loose.1
Greene draws the general instruction from this, and it is the least comfortable line in the chapter:
The temptation to maintain a favorite is understandable but dangerous. Better to rotate your stars, occasionally making each one fall. Bring in people with different viewpoints and encourage them to fight it out. You can justify this as a healthy form of democracy, but the effect is that while those below you fight to be heard, you rule.1
The candour there is worth registering. Greene is naming the cover story and the function in the same sentence.
The version that translates most directly to ordinary work is a film set.
Alfred Hitchcock faced enemies on all sides — writers, set designers, actors, producers, marketers — any of whom could put their ego before the quality of the film. Writers wanted to show off literary skill, actors wanted to look like stars, producers and marketers wanted commercial results. The whole crew had competing interests.1
His solution, Greene says explicitly, was Elizabeth's.
He took the central position. His carefully crafted public celebrity was part of it: publicity campaigns always involved him as spokesman, and his bit-part appearances in his own films made him an instantly recognisable, endearingly humorous figure.1
Then he placed himself in the middle of every stage of production — writing the script before the shoot, editing the film after it — and kept all the filmmaking departments, even the producer, a little out of the loop.1
information about every detail of the film was kept in his head, his drawings, and his notes. No one could bypass him; every decision went through him.1
The costume example shows how fine-grained it was. Before shooting, Hitchcock set out in detail the look of the leading lady's costumes. If the costume designer wanted to change anything, she had to go through him or be caught out in rank insubordination.1
In essence, he was like Rome: all roads led to Hitchcock.1
Greene's generalisation is the practical core, and it is more specific than "be in charge."
Within your group, factions may emerge quite subtly by virtue of the fact that people who are experts in their area may not tell you everything they're doing.1
The remedy:
If you are to lead, you must occupy the center. Everything must flow through you. If information is to be withheld, you are the one to do it. That is divide and rule: if the different parts of the operation lack access to all the information, they will have to come to you to get it.1
And a disclaimer he is careful to add and which the practice does not obviously support:
It is not that you micromanage but that you keep overall control of everything vital and isolate any potential rival power base.1
There is a second answer in the chapter, offered almost in passing, and it points the other way.
Major General Edward Lansdale, America's principal counterinsurgency expert through the 1950s and '60s, held that the key to defeating an insurgency was to stamp out government corruption and bring the people close to the government through popular programs. That would deny the insurgents their cause, and they would die of isolation. Force just played into their hands by supplying a grievance.1
Greene transposes it directly:
Think of the people in your group who are working primarily for their own interests as insurgents. . . . You can always work to divide such factions once you know about them, but the better solution is to keep your soldiers satisfied and contented, giving the insurgents nothing to feed on. Bitter and isolated, they will die off on their own.1
That is a different strategy wearing the same chapter's clothes: not prevent coalitions from forming but remove the grievance coalitions form around. Greene calls it the better solution and then returns to the manipulation for the rest of the section.
The smallest-scale application, and the one most people will actually use:
Start by seeming to take your opponents' side on some issue, occupying their flank. Once there, however, create doubt about some part of their argument, tweaking and diverting it a bit. This will lower their resistance and maybe create a little inner conflict about a cherished idea or belief. That conflict will weaken them, making them vulnerable to further suggestion and guidance.1
Dividing a person's mind against itself, using agreement as the entry point. It is the same architecture — occupy the centre of their position, then let the parts compete.
You run something. Two or three subgroups are consolidating, information is moving sideways without reaching you, and decisions are being pre-cooked before they arrive.
Before reaching for any of this, run Lansdale's question first, because it is cheaper and Greene himself says it's better: what grievance are these factions organised around? Factions form around a real deficit surprisingly often — a promotion process nobody trusts, a resource allocation that looks arbitrary, a decision made once and never explained. Remove the grievance and the coalition loses its reason to exist. That path costs you nothing in trust.
If there's no grievance to remove, then the structural version, which has a specific and non-obvious core: you are not managing people, you are managing where information converges.
Ask where the whole picture currently exists. If the answer is "in three separate heads," you don't have a leadership problem, you have a topology problem. Hitchcock's control came from being the only person holding the complete film. Not from authority — from being the only place the parts could be assembled.
Then make the ladder real, because this is what separates Rome from mere suppression. Differential treatment only produces competition-for-your-favour if the promotion is genuinely available and visibly earned. Rome actually incorporated cities that proved loyal. If the ladder is decorative, you get resentment rather than competition, and resentment builds exactly the coalitions you were trying to prevent.
Then rotate. Greene's advice to occasionally make a favourite fall is the part that reads as cynicism and has a defensible version: a single indispensable lieutenant is a single point of failure and a rival power base at the same time, and the fix is distributing genuine responsibility rather than staged humiliation.
And know the cost you're incurring, because it is real and delayed. An organisation where information flows only through the centre is one where the centre is the bottleneck, where nobody below can act without checking, and where the whole thing degrades the moment the centre is unavailable. Hitchcock's method produced films that were entirely his. It also meant nobody around him ever became capable of making one.
The Roman settlement of 338 BC and its differentiated treatment of Latin cities is well documented and genuinely foundational to how Rome expanded. Elizabeth's factional management and Hitchcock's total production control are both well attested. [POPULAR SOURCE]
The chapter never distinguishes this from the strategy it shares a title with. Dividing an enemy coalition so you can defeat the pieces, and dividing your own subordinates so they cannot organise, appear under one heading three paragraphs apart. The targets are different — enemies versus dependents — the consent conditions are different, and the ethical weight is entirely different. A reader who takes "divide and conquer" as one technique will apply the war version to their own team.
The disclaimer does not survive the examples. Greene says "it is not that you micromanage" immediately after describing a director who specified the leading lady's costumes in advance and made any change an act of insubordination. That is micromanagement, and calling it central-position control does not change what it is. The honest version of the claim would be that centralised information requires a degree of micromanagement and that this is the price.
And the Lansdale alternative is under-weighted to the point of being buried. Greene explicitly calls satisfying your people the better solution, gives it two sentences, and spends the rest of the section on manipulation. That imbalance is itself informative about what the book is for: the grievance-removal strategy is more effective, less costly, and offers the reader no sense of being a strategist.
One further cost the chapter doesn't count. Every case here is a leader with a long tenure and total legitimacy — an empire, a monarch, an auteur. The technique's dependence on an unchallengeable centre means it degrades badly in any structure where the leader's position is itself contestable, and in that setting deliberately setting subordinates against each other reliably produces the coalition it was meant to prevent, now aimed at the leader.
Against Strategy 5: The Command-and-Control Strategy and Strategy 6: The Controlled-Chaos Strategy, this page sits in the middle of a three-way contradiction the book never resolves.
Strategy 5 demands unity of command and one voice. Strategy 6 demands decentralisation and Auftragstaktik — subordinates who act on intent without waiting for instructions, because "omission and inactivity are worse than resorting to the wrong expedient." This page demands that information flow only through the centre and that no one below be able to bypass you.
Strategy 6's Prussian officer who knows when not to obey is structurally impossible in Hitchcock's production. You cannot both distribute the whole picture so subordinates can improvise and withhold it so they have to come to you.
The reconciliation available and unstated: Auftragstaktik distributes the intent while divide-and-rule withholds the coordination. Tell everyone what the objective is; do not let them assemble the operational whole among themselves. That is coherent — and it is also a description of a workforce that can act decisively without ever being able to organise, which is a more troubling thing to have designed on purpose than either chapter admits.
Against Hitchcock and the Detached Buddha and Strategy 12: The Grand Strategy, Hitchcock now appears three times in this book from three chapters — pre-production control producing on-set calm, the indirect first move with the film-less camera, and the central position here. Greene never notices he has explained the same working method three times with three different lessons attached. The unifying fact is simple and none of the chapters states it: he front-loaded every decision, which simultaneously made him calm, made him unreadable, and made him indispensable.
Business → Information Topology as the Real Org Chart. Every organisation has a formal structure and an information structure, and only the second one determines who can actually act. Hitchcock's authority was not on any call sheet; it came from being the sole location where the complete film existed.
The insight the pairing produces: you can read the true power distribution of any organisation by asking where the whole picture is assembled, and the answer is frequently not the person at the top. This also explains a failure mode that looks mysterious from inside — the executive who has authority, issues decisions, and finds nothing changes. If the operational whole is assembled two levels down, that is where the decisions are effectively made, and adding hierarchy does nothing. The corollary is uncomfortable for anyone building a resilient organisation: the same concentration that produces control produces a bus factor of one, and the two cannot be separated because they are the same property described twice.
Psychology → Competition for a Single Source of Approval. Rome's cities competing for favour, Elizabeth's courtiers, Hitchcock's departments — all describe a structure in which recognition flows from one place and peers become rivals for it. The dynamic is well known in families with a parent who plays favourites, and it produces reliable effects: intense vertical loyalty, weak lateral bonds, and a persistent low-grade insecurity that motivates hard.
The insight neither domain gives alone: this arrangement is unusually productive and unusually poor at surviving its centre. The energy is real — competing for approval genuinely does generate effort, which is why the structure keeps getting rebuilt. But the lateral relationships that would let the group function without the centre never form, because forming them was exactly what the design prevented. So the collapse when the centre goes is not a failure of succession planning; it is the structure completing itself. Anyone choosing this arrangement is trading the group's independent survival for the leader's control, and the trade only looks free because the bill arrives long after the decision.
Rome didn't keep the Latin cities weak — it made them compete for something worth having, and the energy that had been aimed at Rome turned sideways and was volunteered. Which means the durable version of this is not suppression but a ladder that is genuinely climbable, and that every version without a real ladder produces resentment instead of competition, and resentment builds the coalitions you were trying to prevent.