Imagine the operator standing in front of someone he wants to move. He has two hands. The right hand is the push — fear of punishment, fear of injury, fear of damage, fear of death. The left hand is the pull — promise of gain, promise of recognition, promise of belonging, promise of bliss. Most operators reach with one hand or the other and stop there. The professional knows the two hands are the endpoints of a single continuum, and the territory between them is where most of the work actually happens.
Siu lays the continuum out as a numbered progression and the verbatim list deserves to be read as one sentence: "The progressive series in the push-pulling of power is: (1) killing the target individuals, (2) eliminating them, (3) damaging them, (4) threatening them, (5) cajoling them, (6) bribing them, (7) persuading them, (8) seducing them, (9) attracting them, and (10) educating them to your view."1
Read the list across once. Notice the slope. Steps 1–4 run on fear. Steps 7–10 run on appetite. Steps 5 and 6 — cajoling and bribing — are the bridge where push and pull blur into each other. Cajoling is half-pressure and half-coaxing. Bribing is half-coercion (you accept this or you lose what you have) and half-incentive (you accept this and you gain what you want). The middle of the continuum is where the categories start to fail and where the operator's actual judgment has to do the work.
Push is the side of the continuum trained on fear. Siu's small-scale examples: the sarcastic slur, discharge from work, multiple warhead independently targeted intercontinental ballistic missiles, and eternal damnation.2 The list is comically wide on purpose. Push is operating whether you are in a meeting watching a senior person make someone else look small, or watching the strategic-deterrence community plan a counterforce strike, or watching a sermon describe the burning lake. The mechanism is identical at every scale. The target's behavior is being shaped by the prospective cost of non-compliance.
What changes across scale is the form, not the mechanism. Siu makes a quiet observation here that rewards re-reading. The amount of bodily punishment that is socially acceptable on an individual basis has declined over the centuries. The Roman fustuarium — guards beaten to death by their fellow soldiers for falling asleep on duty — has been replaced by the modern stockade and the loss of a stripe.3 Individual push has gentled.
But — and this is where Siu becomes uncomfortable — the same cannot be said for people en masse. It is within recent memory that a nation systematically decimated millions of nameless Jews in nameless production-line gas chambers over a period of several years and a single bomber killed a hundred thousand nameless Japanese over a few seconds. The instruments of push are being refined and amplified nowadays more for power on a grand scale than for power on individual coercions.4 The push has not gone away. The push has been industrialized. What was once one tribune killing one guard is now one bomber killing one hundred thousand civilians, with the operator at neither end of the bombing seeing a face. The category Siu names with the dry word push covers, at its scale extreme, the mass-extermination apparatus of the twentieth century.
Pull is the side of the continuum trained on appetite. Siu's small-scale examples: the gentle compliment, performance bonus, foreign aid, and eternal bliss.5 Same wide range. Same identical mechanism at every scale.
The pull side is where Siu's operational analysis sharpens. Modern democratic-and-socialist conditions push pull instruments toward subtlety. Inducements for the individual are so cunningly alloyed with the goals of the institution that the two are confounded by the unsuspected as one and the same. Individuals are led without their knowing.6 The institutional pull has been engineered so the target experiences his compliance as the pursuit of his own goals. The institution's growth and the individual's perceived self-interest become operationally identical, even when they are objectively divergent. Selfless workers and selfish workers both produce the institutional outcome the operator wants. Subjects are being pulled when they think they are pursuing.6
Read that last sentence again. Subjects are being pulled when they think they are pursuing. It is the cleanest single description of how modern institutional power works on willing participants, and Siu writes it like a side observation. Most operators in any given institution are pulling without realizing it; most subjects are being pulled without realizing it. The architecture works on autopilot once installed. The operator who notices the architecture is in the small minority that can deliberately engineer it. The subject who notices the architecture is in the smaller minority that can resist it.
Siu names a default and a fall-back. In general, you should lean toward the nonviolent pull end of the spectrum as options of first preference. If effective, they are more acceptable to the public, more economical all around, and less injurious to yourself.7 Pull is the cheap option. It does not generate counter-resistance the way push does. The target who has been pulled often defends the operator's interests as if they were his own. The target who has been pushed will reverse the moment the pressure lifts.
But. When the big chips are down, however, you will frequently have to fall back toward the violent push end as the old reliable. Pull is preferred under normal conditions. Push remains the irreducible reserve when the stakes exceed what pull can deliver. The operator who has only the pull end of the continuum is undefended when the big chip falls. The operator who has only the push end produces revolts and counter-revolts as a side effect of every move. The competent operator can run the full continuum and chooses which step the situation actually requires.
Siu adds a footnote-grade observation that ends Op#35 and points outward to a sister concept. There are a few individuals who are immune from push or pull of all kinds. Men who have subdued anger, pride, and craving cannot be enticed or pressured to do your bidding.8 The operational instruction follows: Do not waste your energy working on the likes of I-liao and Diogenes. Leave them be.9
The full development of this exception lives on the dedicated page — see The Unmovable Minority. For the operator running the Push-Pull continuum, the practical takeaway is narrow: scan for genuine immunity before committing resources to a long campaign on a target who turns out to be I-liao. The cost of the misidentification is wasted effort against the immune; the cost of the correct identification is to walk away and find a target whose drives are still attached. Siu's recommendation is resource-management, not respect.
The hidden cost of pull. Pull is the cheaper option, but the cumulative cost of running pull-without-the-operator-noticing on a target population eventually produces the kind of trust collapse that makes any future move expensive. A workforce that has been pulled while believing they were pursuing eventually figures it out, and the discovery is more destructive than ten years of honest push would have been. Siu names pull's economic advantage and underemphasizes pull's trust-decay rate.
The non-monotonic spectrum. Siu presents the ten steps as a smooth progression, but in practice some moves are not weaker versions of adjacent moves. Threatening (4) is not always less coercive than damaging (3); a credible existential threat can produce more compliance than an actually-delivered minor damage. Persuading (7) and seducing (8) are categorically different in mechanism — one runs on argument, the other on attraction — and arguing about which is "stronger" misses that they activate different channels in the target. The ten-step ordering is heuristic, not strict.
The mass-scale push problem. Siu's observation that push has industrialized at scale while gentling at the individual level is descriptive. It does not address what to do about it. The operator who can call in a thermonuclear strike but cannot socially fire one employee is operating in a world where the continuum has been bent into something Siu's smooth ten-step does not capture. The contemporary version of the spectrum may need a separate axis for visibility-of-cost-to-operator, which the Roman tribune had and the modern bomber pilot does not.
Psychology — The Operator's Cost at the Push End: Reluctance to Kill as Baseline Human Capacity — S.L.A. Marshall and Dave Grossman document a finding that complicates Siu's clean spectrum. Trained soldiers refuse to kill at rates of 75–80% even when ordered and threatened. Reluctance to kill is the baseline human capacity, not a deviation from it. The push end of Siu's continuum runs against a structural psychological resistance that the operator at the kill-step has to overcome in his own people before the push can be delivered. The operator does not just have to push the target. He has to override the inhibition in the soldier or executioner who has to do the pushing. Grossman's research names what the continuum implicitly assumes away: step 1 of Siu's continuum requires institutional machinery to construct the human capacity to deliver it, and the construction is expensive, fragile, and reversible.
The handshake between the two pages produces a finding neither states alone. Siu describes the push as a tool the operator can deploy. Grossman describes the push at its sharpest as something the operator can only deploy by building a separate machinery — training, propaganda, dehumanization, distance — that constructs the temporary capacity to override the baseline reluctance. The push end of the continuum is not free even before it reaches the target. Every move from level 4 (threatening) to level 1 (killing) requires increasing investment in the construction of the operator's own capacity to deliver. The mass-push that Siu observes industrializing at scale works because the construction has been institutionalized at scale — bomber pilots see only heat signatures, drone operators see only screens, gas-chamber bureaucrats see only paperwork. The industrialization is the construction of the machinery that overrides the baseline reluctance at population scale. The push has not become cheaper. The cost has been moved off the visible operator. This is not a minor reframing. It is the developmental account of how mass-scale push became operationally available, and it should sit beside Siu's continuum any time the operator considers stepping toward the lower-numbered moves.
History — One Operator Walking the Full Spectrum: Mao and the Chinese Revolution — Mao Zedong's career reads as a near-complete demonstration of Siu's ten-step continuum executed across a single political project. Educate (10) — the entire ideological apparatus of Marxism-Leninism translated into Chinese revolutionary terms. Attract (9) — the romance of revolution, of the New Man, of the historic mission. Seduce (8) — the implicit promise that joining the cadre meant becoming a different and better kind of person. Persuade (7) — Mao's writings, the indoctrination programs, the political-education sessions. Bribe (6) — land redistribution to the peasants who joined the cause. Cajole (5) — the United Front periods of selective alliance with the bourgeoisie. Threaten (4) — the implicit cost of being labeled counter-revolutionary. Damage (3) — the struggle sessions, the public denunciations, the social destruction of class enemies. Eliminate (2) — the systematic removal of the landlord class. Kill (1) — the executions during the land-reform campaigns and the Cultural Revolution.
The handshake reveals the structural finding that Siu's static taxonomy hides. Mao did not stay at any one step. He moved up and down the continuum across phases of the campaign, calibrating which step the political moment required. The educate step (10) was the constant background; the kill step (1) was deployed at specific moments against specific targets. The Pull instruments built the long-term legitimacy that made the Push instruments locally tolerable. The Push instruments made the cost of refusing the Pull legible enough that more people accepted the Pull. Push and Pull were not alternatives in Mao's hands. They were the two endpoints of a single continuum that he ran across the entire spectrum simultaneously, with different targets at different steps at any given moment. The operator who reads Siu's continuum as a choice between hands has not yet understood it. Mao's career is the historical demonstration that the continuum is run as a combined operation, with the operator selecting which step applies to which target population at which moment, and rotating the steps as the political conditions evolve. The cost of running only one end of the continuum — only Push, only Pull — would have been Mao's defeat at multiple junctures. The combined operation is what made the campaign survivable.
1. Identify the actual stakes before choosing the step. It is Tuesday morning. You face a target who is not behaving the way you need them to. Before deciding which step to deploy, name the actual stakes. Catastrophic if non-compliance continues — you are looking at the lower-numbered steps. Costly but recoverable — you are at steps 4–7. Annoying but tolerable — you are at steps 8–10. Most operator failures involve choosing a step from the wrong stakes-band. The senior who thunders at a junior over a missed deadline is using a step-4 move on a step-8 problem and burning credibility for nothing.
2. Default to the highest-numbered step that has a reasonable chance of working. Pull is cheaper and more durable. The first move should be at step 9 or 10 unless the situation forecloses those options. Educate them to your view — make the case, show the data, walk the target through your reasoning. Attract them — make compliance materially or socially desirable. Operators who skip the high-numbered steps because they are slower forfeit the durability advantage that pull-end compliance produces.
3. Reserve push for the irreducible cases. The fall-back to push is structural, not aspirational. Some situations require step 4 (threaten) or even step 3 (damage) because no pull-end move will land. The competent operator does not pretend push is unavailable. He uses it sparingly, visibly, and only when the lower steps have failed or the time horizon does not permit them. The operator who has never used push is undefended. The operator who uses push as the default has destroyed the cooperative substrate that makes pull work.
4. Watch the bridge zone (steps 5–6) carefully. Cajoling and bribing are where most modern professional power-craft actually operates. They are also where the categories blur and the operator can run a push instrument while believing he is running pull. I'll make sure you're recognized for this — and of course it would be a shame if your team's headcount got reduced. That sentence runs both ends of the continuum simultaneously. Notice when you are doing this. It is sometimes the right move; it is also sometimes the move you choose because you don't want to admit to yourself which end of the continuum you actually need.
5. Audit your own range across a year. Most operators run only two or three of the ten steps and never use the others. The professional has at least seven steps available and chooses among them. If you have run only steps 8–10 for a year, you are operating with half the spectrum. If you have run only steps 1–4, you are wearing out the cooperative substrate and your pull-end credibility is gone. The annual audit is whether you have used the full continuum, not just whether you have it conceptually available.
6. Scan for the immune before committing. Some targets cannot be moved by any step on the continuum. The Diogenes exception is rare but real. The cost of running a long campaign against a genuinely-immune target is total — every move you make is wasted. Scan early. If the target shows the unmistakable signature — equanimity in the face of threat, indifference to inducement, no visible attachment to outcomes you can manipulate — walk away and redirect resources to a target whose drives are still attached.
The ten-step continuum is most usefully read as a complete inventory of the moves available to any operator at any scale. Every operator is running some subset of these ten on the people around them at all times. Most operators are unaware they are doing it, run only two or three steps habitually, and are running a step different from the one they think they are running. The honest application of the continuum is therefore a kind of mirror — which of these ten am I actually running on the people I work with, and is it the step the situation calls for? The mirror is uncomfortable because most operators discover that their default step is somewhere in the bridge zone (5–6) and that they are using bribery while telling themselves they are persuading. Siu does not say this kindly. He says it operationally. The continuum exists; you are running on it; the only question is whether you know which step you are on. The professional knows. The amateur does not. The cost of not knowing is paid by the people you are running it on, and eventually by you, when the trust they have extended on the assumption that you were running step 7 is withdrawn after they discover you were running step 5.
The Push end of the continuum has industrialized at scale. The Pull end has industrialized at scale too — modern marketing, propaganda, and platform-engineered behavioral capture run pull instruments at population scale that no twentieth-century institution could have approached. What does the contemporary version of Siu's continuum look like when both ends have been industrialized — and is the resulting environment one where individual operators can still meaningfully choose a step, or has the choice moved to the level of institutional design?
The Diogenes exception is presented as rare. Modern conditions may be producing a new and broader version of it — operators who have voluntarily made themselves uninfluenceable through digital opt-out, asset minimization, or professional irrelevance. Are these contemporary Interstitialists growing as a population, and how does that change the continuum's reach?
Mao ran the full continuum simultaneously across different targets. Most modern executives run two or three steps. Is this difference a sign of underdevelopment, of changed institutional conditions, or of correctly-narrowed scope for non-revolutionary contexts? Probably some mix, but the diagnostic question for any operator is which of the three explanations applies to their own range.