Nasreddin Hoca arrives at a public bath in a strange village, dressed in poor clothes. The attendants take one look at him and decide he is not worth their time. He is handed a sleazy, torn, dirty rag for a towel. Nobody helps him scrub. Nobody offers soap or perfume. He bathes himself in indifference and leaves. On his way out, to the astonishment of every attendant in the place, he gives each one a gold piece.
The following week he returns. The attendants have remembered him. This time they bring fresh towels, new soap, perfume. They scrub and massage him. They help him dress. Their palms tingle with the anticipation of another gold coin. On his way out, he gives each one a penny.
The attendants protest. Nasreddin Hoca explains: "The gold pieces I gave you last week were for the manner you treated me today. The pennies I gave you today are for the manner you treated me last week."1
This is Siu's opener for the spec on how an operator should treat the cadre below him, and the parable is doing more than it looks like it is doing. It is naming the central problem of cadre treatment: the people around you who push you into power are also the people who can push others into your place. The talents that built your team are the talents that can replace you. Whether the talent points up at your benefit or sideways at your replacement depends entirely on the architecture you build for them.2 Nasreddin Hoca's parable is a compressed lesson on the asymmetry of treatment. The first lesson is that the timing of your appreciation determines whether the appreciation lands. The deeper lesson is that people are calibrating you continuously; what they feel is reciprocity is the foundation of the loyalty you depend on.
Siu's first operational instruction is unsentimental. Since people are more likely to help those from whom prior help had been received, you should make it a point to provide unsolicited assistance to individuals within your cadre on a regular basis. Maintain a small reserve of funds, personnel vacancies, and informal favors. Parcel them out unprompted to cadre members who are wrestling with a difficult task whose failure would be personally embarrassing to them. Or — Siu's preferred move — suggest another member of your cadre to go over and offer help. The second move is the better one. The reciprocity flows in two directions instead of one. The cadre member you helped is now indebted to the operator. The cadre member you sent to help is now in a relationship with the cadre member he helped. This would also increase the spontaneous reciprocal interactions within your unit, thereby strengthening its teamwork immeasurably.3
Read that move twice. The operator does not always intervene directly. He sometimes sends another cadre member to do the helping. The reciprocity-network the operator is building does not run only through him. It runs across his cadre. He is the common node, but the loyalty he is generating is partially networked into the cadre's relationships with each other. A cadre that owes each other favors is a cadre that holds together when the operator is not in the room.
Siu drops one sentence into the spec that changes the whole frame. Of equal practical significance is the development of much greater persistence in the behavior of animals subjected to a variable schedule of partial rewards than of those subjected to a fixed schedule of full rewards of the same total amount and kind.4
The finding is from operant conditioning. Pigeons trained on a fixed reward schedule — every tenth lever-press produces food — stop pressing the moment the food stops. Pigeons trained on a variable schedule — sometimes every fifth press, sometimes every twentieth, never predictable — keep pressing for hours after the food has stopped. The variable schedule produces deeper learning and longer extinction. It is the most reliable behavioral cement that experimental psychology has documented.
Siu does not name what he is recommending in operational terms, but read the sentence again with a manager's ear. Reward your cadre on a variable schedule of partial rewards. Sometimes the bonus is large. Sometimes it is small. Sometimes the recognition comes immediately. Sometimes it does not come for months. The cadre member never knows quite when the reinforcement will arrive or how big it will be. He keeps pressing the lever. The reciprocal version of Nasreddin Hoca's parable is running in the background of the entire reward architecture. Your cadre's persistence is a function of the unpredictability of your appreciation, not its size.
The honest version of this finding is uncomfortable. The cadre member is being trained by mechanisms identical to the ones experimentalists used to train pigeons. The mechanism does not require malice on the operator's part — most operators run variable reinforcement unconsciously, simply because life is variable — but the effect is the same whether intended or not. Cadre members who are intermittently rewarded develop deeper loyalty than cadre members who are reliably rewarded.
Siu quotes G. E. G. Catlin on a particular kind of follower who deserves separate handling. "The devoted follower in the train of his leader, the affectionate friend, the weaker brethren, the submissive subjects, attain their purposes and are satisfied in the success of their hero; his glory is theirs, and through his will they prevail and become more fully persons."5
This is the spiritual-extender follower. He is not in the relationship for material reward. He is in the relationship because his self extends through the leader's self, and the leader's success is therefore his own success at a deeper level than any utilitarian transaction can reach. These followers are exceptionally valuable and deserve separate treatment. The operator who gives the spiritual extender material rewards has misread the relationship; what the spiritual extender wants is participation in the leader's becoming, and the operator who provides that participation gets a level of devotion that no money can buy.
The corollary case is the zealot — the cadre member who exceeds the bounds of judgment and even the law on behalf of the leader and the cause. Siu's instruction is direct. Especially when he sacrifices himself in so doing, do not add to his misery. As leader, you should implicitly express your appreciation by assisting him in his troubles. The 1954 UAW Kohler strike: cadre members were arrested, tried, convicted, and sentenced to jail for striker violence. The union provided their legal defense and continued their pay while they served their sentences. When questioned, the union president said "I for not one second will defend what he did, because I think he was wrong. He was punished. He should have been punished. Things that we did were to help his family. His family didn't make the mistake."6 One stroke. Respectful of the law, human-hearted toward the family, appreciative of the cadre member's loyalty, and reassuring to the rest of the cadre that they too will not be abandoned. The architecture rewards the zealot for excessive loyalty without endorsing the excess itself.
Siu prescribes social distance with precision. The power distance between you and your subordinates should be kept distinctly evident, so that a kind of respectful fear of you is ever present.7 Not glowering. Not abusive. Be always courteous and even considerate. But — and the negative instructions matter — do not socialize with your subordinates on too frequent a basis. Socialize with your peers and outsiders. Think twice about employing close friends and relatives as your direct subordinates; if they need jobs, help them get one in some other organization. The best way to make use of social and intimate friends is not as subordinates but as confidants on selected matters.8
The reason behind the rule is that close friendship and direct subordination are operationally incompatible. The friend cannot give the operator unfiltered information because the operator is also his boss. The boss cannot fire the friend without losing the friendship. Both relationships degrade when forced through the same channel. The operator's friends should sit outside the chain of command where they can be honest with him; the operator's subordinates should sit inside the chain of command where they can be effective for him. Putting friends inside the chain destroys both functions.
Siu names a precise distinction that is rare in the management literature. The injection of fear for ensuring the unity of an organization goes back to antiquity. Augustine recommended fear in fifth-century campaigns against the Donatists. The instruction is universal: some fear is necessary for unity. But the intensity of fear among your immediate cadre should never be so great that it interferes with the performance of duties. It should be particularized against intentionally interfering with your own rise in power.9
The two instances Siu walks through are operationally precise:
Do not punish for ill fortune or incompetence. The chief of state who punishes field commanders for losing battles produces an army whose commanders are too distracted by fear of defeat to fight the next one. The competent move is to transfer the incompetent without punishing them. Replace; do not destroy.
Do punish curry-favor with opponents. The cadre member who currys favor with your opponent or your peer at your expense is forming a coalition against you — wittingly or otherwise. He must not only be dismissed but dismissed in such a fashion as to discourage others from entertaining the notion.10 The 1968 case of the Air Force engineer who testified before Congress about the C-5A overrun and was applauded by the press — and immediately fired by the Air Force secretary — is Siu's example. The cadre member's act was admirable on its merits; it was operationally treasonous against the operator above him. The architecture punishes the second, not the first.
The Han Tzu compression sits in the middle of all this and deserves to be lifted whole. "Never ennoble anyone in such a way that he may molest you. Never trust anybody so exclusively that you lose the capitol and the state to him."11
Siu names four operational considerations that go beyond standard management-textbook delegation principles. Each one is a place where the operator's interest in his own continued power diverges from the institution's nominal interest in efficient delegation.
1. Compatibility. The recipients of your transferred power must be susceptible to your continued influence and compatible with your operating pattern. The negative case is Giovanni Giolitti as prime minister of Italy in 1913. He was the master of trasformismo — backstage deals, promised favors for votes, the manipulation of personalities rather than parties. His secure grip on Italian politics was historic. Then he extended suffrage to the working classes. The 1913 elections brought millions of new voters. Giolitti's methods of controlling notables no longer worked for the masses. Radical deputies refused to play trasformismo. Giolitti had to step aside.12 His delegation method worked perfectly on a political class that shared his pattern. The expansion brought in actors who did not, and the delegation architecture broke. The lesson: do not delegate to people whose operating pattern is incompatible with the methods you control them by.
2. The Golden Section. Sometimes you must decentralize substantial authority to an ambitious subordinate in order to get a job done. The dilemma is between giving him enough authority to accomplish what you need and keeping him short enough that he cannot overflow his boundaries. Siu's mathematical heuristic is the Golden Section — the ratio at which the smaller portion is to the larger portion as the larger is to the whole — 0.38 to 0.62.13 When weight distributes at this ratio, the minority contributes meaningfully to joint decisions, and the majority retains leadership without being able to squelch the opposition. Adjust the ratio toward 0.43/0.57 if the junior collaborator is much weaker; toward 0.33/0.67 if much stronger. The ratio should not fall below 0.25/0.75 nor above 0.45/0.55. Outside that band the architecture either pre-determines the outcome (the junior has no real input) or invites usurpation (the junior has too much).
3. Class-vs-Individual Struggle. A clever subordinate who is losing the individual contest will sometimes try to transform the contest into a class contest. They are calling on the class of race, sex, or religion to stamp you out as an individual enemy. You are no longer fighting the person as an employee, with your corporate top management as judge. You are fighting such persons as representatives of a class, with the rest of the class as judges.14 Siu's defensive instruction is preventive: always so conduct yourself in your corporate life so that when accused, the relevant class will repudiate one of its own members and proclaim its confidence in you. If your moves keep getting reframed as class struggles regardless of how you conduct yourself, the operating environment has become hostile to your continued power, and the right move is to continue your growth in power in another setting.
4. The Deputy Danger. A clearly-designated second-in-command is a structural risk. He may move on you directly. He may be used by your opponents as a rallying point even if he does not move on you himself. This is one of the reasons why the Manchu emperors of earlier centuries never named a crown prince until on their deathbed.15 Two practical defenses. Decentralize authority more or less equally to an echelon of at least six individuals; do not leave an impression that any one of them is "more equal" than the rest. Or maintain a base of support independent of this group — Juan Peron's 1945 case, where the descamisados (the shirtless ones of the laboring masses) rose in his defense after his fellow officers had him arrested, and the junta had to release him. The lesson: never let your power flow only through one channel that another operator can occupy.
The variable-reinforcement honesty problem. Siu recommends running operant-conditioning mechanics on cadre members without addressing the ethical asymmetry of the move. The cadre member is being trained by methods used on pigeons; he does not know it; the training works because he does not know it. The operator who tries to be honest about this with the cadre member compromises the mechanism, and the cadre member who learns the mechanism is running on him often becomes the operator's most determined opponent. The honest version of cadre-treatment ethics is harder than Siu admits.
The friendship-displacement problem. Siu's prescription that friends should not be employed as direct subordinates is operationally sound and socially expensive. Operators who follow it strictly find themselves with a small number of true friends and a large number of professional acquaintances. The strict version of the rule is one of the loneliness costs of senior leadership, and most operators violate the rule because the loneliness is unbearable. The violation is one of the most predictable career failure modes.
The class-vs-individual asymmetry. Siu's analysis of class-struggle reframing assumes the operator is being framed unjustly. Sometimes the framing is accurate — the operator's behavior really is patterned by race, sex, or class biases he has not examined. Siu's defensive prescription works in both cases without distinguishing them, which means the prescription protects bad actors as effectively as good ones. The honest application requires the operator to do the prior work of distinguishing the two cases, and the prior work is not in the spec.
Han Tzu's loop. Never ennoble anyone in such a way that he may molest you and never trust anybody so exclusively that you lose the capitol — these are sound defensive principles that, run consistently, produce an operator surrounded by cadre members none of whom has been ennobled or trusted enough to be deeply effective. The defensive optimum and the operational optimum are in productive tension that the spec does not resolve.
History — Khan's Architecture for the Same Problem: Meritocracy-Within-Subordination — The Loyalty-Capability Equation — Genghis Khan, consolidating the steppe in the early thirteenth century, faced the cadre problem at imperial scale. He came from a minor clan and could not staff an empire from his own kinship network. He had to recruit warriors and administrators from rival tribes and offer them advancement. But advancement without loyalty was lethal: a capable officer loyal only to advancement would advance, gather power, and defect to the next bidder. Khan's solution is the architecture Siu compresses into Han Tzu's caution. Advancement equals capability plus loyalty within subordination. You advance if you are capable and loyal. The advancement is conditional. You can advance — but only as far as Khan permits. You cannot advance beyond the point where you threaten Khan's position.
The handshake produces a finding neither page states alone. Khan's meritocracy and Siu's cadre treatment are running the same architecture at different scales. Khan's Great Law punishing disloyalty by death is the institutionalized form of Siu's dismiss in such a fashion as to discourage others from entertaining the notion. Khan's Within Subordination clause is the imperial form of Han Tzu's never ennoble such that he may molest you. Khan's structural innovation — opening capability-based advancement while constraining the upper bound of advancement to the operator's continued safety — is what Siu's four delegation considerations are designed to produce on the corporate scale. Reading them together is reading the same problem solved twice, eight hundred years apart, in radically different cultural settings, with overlapping mechanisms. The cadre architecture is structural to the conditions of power, not specific to any era. Operators who think their cadre problem is unique to their corporate environment have not yet noticed that Genghis Khan was solving it in 1206 and the answer is converging.
Psychology — How the Devoted Follower Becomes Through the Leader: Archetypal Bonding Patterns — Hal and Sidra Stone describe a phenomenon that turns out to be the developmental foundation of Catlin's spiritual-extender follower. Two people lock into complementary subpersonalities without conscious awareness. The good-mother subpersonality in one person bonds with the helpless-child subpersonality in the other. The protector in one bonds with the wounded in the other. The bonding self-sustains because both parties get what their subpersonalities need from the configuration. Once locked, the bonding pattern continues without renegotiation, and both parties experience the relationship as natural and unchosen.
The leader-follower bond Catlin names — his glory is theirs, and through his will they prevail and become more fully persons — is the same mechanism. The follower has bonded a part of himself (the part that wants to extend, to belong, to participate in someone larger) to the leader's projected greatness. The leader has bonded a complementary part of himself (the part that wants to be seen, witnessed, augmented) to the follower's devotion. Both parties get what their subpersonalities require from the bond. Neither party negotiated it consciously. Once locked, the bonding produces the spiritual-extender phenomenon Siu describes — the follower whose self genuinely extends through the leader, whose loyalty cannot be reduced to any utilitarian transaction.
What this reveals when both pages are read together is that cadre architecture is partly subpersonality architecture. The operator who builds a strong cadre is doing more than running the operational mechanics Siu describes; he is licensing specific subpersonality patterns in the people who join his orbit. Reciprocity, variable reinforcement, particularized fear, the Golden Section — these are operational moves, but they only work because they activate subpersonality bonds that the cadre member already had pre-installed from earlier developmental experience. The operator with charisma is the operator whose presence triggers the spiritual-extender subpersonality at high frequency in the people around him. The operator without charisma is the one whose presence triggers the watchful-employee subpersonality instead. Both kinds of operator can build cadres, but the architecture they produce is structurally different and decays under different conditions. The handshake produces a finding the operational literature mostly avoids: cadre treatment is partly a psychological licensing system for a particular kind of follower-bond, and the operator who runs it without understanding what he is licensing is at the mercy of which subpersonalities his presence happens to evoke.
1. Audit your reciprocity flow this month. Look at your last four weeks. Which cadre members did you provide unsolicited assistance to? Which ones did you arrange to have another cadre member assist? If the answer to the second question is "none," you are running cadre-loyalty through yourself only. The network is centralized. Centralized networks fail when you are unavailable. Diversify by sending help through cadre members instead of always providing it directly.
2. Notice your reinforcement schedule. Are you predictable in your appreciation? Do you praise the same things at the same times in the same ways? Predictability produces extinction. Variable timing, variable intensity, variable form — these produce persistence. The point is not to manipulate cynically; the point is to notice that excessive predictability is not the virtue it feels like. Reliable in promises and predictable in appreciation are different operational settings. Be reliable; do not be predictable.
3. Identify the spiritual extenders separately. In your cadre, who is in it for the work and who is in it because their self extends through your role? The first kind needs material reward. The second kind needs participation in your becoming. If you treat the spiritual extender with material reward only, you have misread the relationship and forfeited the depth of devotion that was available. The signature of the spiritual extender is that they care about your success in a way that exceeds any utilitarian calculation; their face when you win is the diagnostic.
4. Practice the friendship-distance rule once. Pick the closest social friend in your direct reporting line. Move that friend out — into a peer relationship if you have no authority over them, or into a different chain of command if you do. Notice the cost. The cost is real. Notice that the operating quality of both the friendship and the working relationship lifts after the move. If you cannot bring yourself to do this, you have decided to pay a different cost — the gradual degradation of both relationships under combined load.
5. Run the four delegation considerations on your most ambitious lieutenant. Compatibility — does his operating pattern match the methods you use to influence him? Golden Section — is your authority weighting somewhere between 0.55/0.45 and 0.75/0.25 against his? Class-vs-individual — is the contest still being fought as individual? Deputy danger — is he the only second-in-command, or is there an echelon of six? If two or more of these are off, you have a structural cadre risk that your operational competence cannot offset. Adjust the architecture, not the individual relationship.
6. The C-5A engineer test. When a cadre member acts in a way that is admirable on its merits but operationally costly to you — the public testimony that surfaces an embarrassing fact, the principled stand that breaks the firm's strategic cover — what is your response? The Air Force secretary fired the engineer. Most operators would. The decision is not whether to fire; the decision is whether you are running the particularized fear architecture (you fire because the cadre member curry-favored with an external power against your interest) or whether you are running ordinary punishment of integrity. The two look identical from outside. They feel different from inside. The test is whether you can articulate the distinction honestly to yourself.
Cadre treatment is the place where the operator's stated values and his operational values diverge most visibly. The stated values are usually meritocracy, fairness, transparency, trust. The operational values — variable reinforcement, particularized fear, calibrated power-distance, deputy-danger management — are calibrated for the operator's continued power and only secondarily for the cadre's well-being. Siu's spec is unusual in that it makes the divergence explicit. Most management literature obscures it. The operator who reads Siu honestly discovers that he has been running the operational architecture all along while telling himself he was running the stated architecture, and that the cadre members below him have always known, or eventually figured out. The discovery is uncomfortable. The alternative — pretending the divergence does not exist — is worse, because the cadre members who are running on the operational architecture without admitting it produce specific failure modes that operators who admit it can prevent. The reading is harsh. It is also accurate, and the cadre members of the harsh-reader operator are usually treated better in the end than the cadre members of the operator who refuses the diagnosis. Honesty about the architecture produces better architecture.
The variable-reinforcement finding from operant conditioning produces deeper loyalty than fixed-schedule reward. Modern HR practice tends in the opposite direction — predictable performance reviews, regular bonus cycles, transparent compensation bands. Has this trend produced cadres that are demonstrably less persistent than the variably-rewarded cadres of earlier executive eras? The empirical question has not been studied directly because it is operationally inconvenient to ask.
The spiritual-extender follower is the cadre member who is most valuable and least replaceable. He is also the one most psychologically vulnerable to the operator's failures. When the leader fails, the spiritual extender fails through him in a way the utilitarian cadre member does not. What is the operator's responsibility to a follower whose self has extended through him, when the operator's own life trajectory makes that extension a liability for the follower? Most operators do not acknowledge that the question exists.
The Manchu emperors did not name a crown prince until on their deathbed. The Manchu solution to the deputy-danger problem worked at the cost of regular succession crises. Modern corporations name crown princes routinely and pay a different cost — the named successor either becomes a structural threat or becomes politically compromised by being named. Is there a third architecture neither Siu nor the Manchu emperors found, or are these two the available options and the choice is between two failure modes?