Historians searching the Nazi archive for "the Hitler order to begin the Holocaust" have never found it. There is no signed document. No formal directive. No moment where Hitler convened his ministers and said exterminate the Jews of Europe. The Wannsee Conference of January 1942 — the famous coordination meeting — was attended by mid-tier bureaucrats and dealt with logistics, not authorization. The killing had already started months earlier. The escalation from harassment (1933) to ghettoization (1939) to mass shooting (1941) to industrial gassing (1942) happened without explicit central direction at almost every step.1
This is the paradox that the historian Ian Kershaw resolved with a phrase that has become canonical: "working toward the Führer." Hitler issued remarkably few explicit orders. His subordinates — the SA, the SS, the Gestapo, the regular military, the regional Gauleiters, the state governments — competed with each other to anticipate what he wanted and act on it. They radicalized without being told to. They escalated because they were competing for Hitler's favor and the metric of favor was who is doing the most to advance the cause as the Führer understands it. The result was a regime that produced its most extreme outcomes through a leadership architecture that never issued the extreme orders.1
The concept names this leadership pattern: a charismatic, vague-signaling leader who articulates direction (the cause, the enemy, the destiny) but refuses to specify policy, generates a competitive bureaucracy in which subordinates one-up each other in serving the perceived will, producing radicalization without explicit authorization.
Working-toward-the-Führer activates when three system conditions converge:
Vague high-intensity signal from the top. The leader sets direction at the level of speeches and visions — "Germany must be cleansed," "the Jewish question must be solved," "the East must be Germanized" — without specifying the operational policy. The signal is intense enough to mobilize action but ambiguous enough to permit interpretation. If Hitler had specified "deport all Jews, do not kill them" the architecture wouldn't have produced extermination. He didn't specify; subordinates interpreted.1
Multiple parallel power centers competing for the leader's favor. Not one bureaucracy but many, with overlapping jurisdictions and unclear lines of authority. The SA, SS, Gestapo, Wehrmacht, regional party leaders, state ministers all had partial authority over Jewish persecution. None had final authority. Each was competing with the others to be seen as the most effective servant of the cause. The competition was for favor, with the metric being concrete advancement of the vague-stated goal.1
A reward structure that rewards extremism. When subordinates compete for the favor of a vague-signaling leader, the equilibrium is whoever interprets the signal most aggressively wins. If Hitler praised harsh actions and stayed silent about moderate ones, the entire system learned to interpret his signal as endorsing harshness. Even if a moderate interpretation was equally consistent with what Hitler had actually said, the rewarded interpretation was always the more extreme one. This produces a one-way ratchet: the system can only escalate, never moderate, because moderation produces no favor-payoff and risks being seen as disloyal.1
When these three conditions hold, the system produces outcomes that none of its individual members would have explicitly chosen — outcomes that exist only as the equilibrium of competitive anticipation.
The mechanism runs through six steps:
Step 1: Vague signal. Hitler speaks about "the Jewish question" without specifying solution. Multiple subordinates hear this.
Step 2: Plural interpretation. Each subordinate has their own model of what Hitler "really wants." The SA leadership reads it as street-violence-license. The SS reads it as systematic-removal mandate. The Gestapo reads it as surveillance and harassment. The regional Gauleiters read it as local-policy permission. None of these readings is correct or incorrect — Hitler's signal didn't determine which.
Step 3: Visible action. Each subordinate acts on their reading. Smashed shops in one region. New restrictive laws in another. Surveillance operations in a third. The actions are heterogeneous because the signal was heterogeneous.
Step 4: Reward asymmetry. Hitler praises the actions he likes (or stays silent on the ones he doesn't). The praise creates the signal that this interpretation is what the Führer wanted. Other subordinates update.
Step 5: Convergent escalation. Subordinates whose actions were not praised face two options: stop (which signals disloyalty) or escalate-further (which might earn praise). Most choose escalate. Meanwhile, subordinates whose actions were praised have established that their interpretation was correct, and they too escalate to maintain their edge over competitors who are catching up.
Step 6: Ratchet. The system has no mechanism to walk back. Moderation would require Hitler to explicitly forbid extreme actions, which he refuses to do because his signal is intentionally kept vague. So the equilibrium moves only one direction: harder, more thorough, more comprehensive. The endpoint of the ratchet is whatever extreme outcome the most aggressive interpretation can sustain operationally.1
The leader never had to order the Holocaust. The system anticipated it. The bureaucracy delivered it. The signature is never on the order because the order was never written. The order existed only as the equilibrium of a competitive bureaucracy interpreting a vague signal.
The Wilson-derived account above attributes the concept of "working towards the Führer" to Kershaw, who named the dynamic in his Hitler biography. Kershaw himself drew the phrase from a primary source: a routine speech by Werner Willikens, State Secretary in the Prussian Agriculture Ministry, delivered to a meeting of representatives from Länder agriculture ministries on 21 February 1934. The speech is the regime's own description of its operating principle, articulated thirteen months into the regime's existence by a mid-level official in a routine bureaucratic setting.2 Willikens' formulation is worth reading in his own words:
"Everyone with opportunity to observe it knows that the Führer can only with great difficulty order from above everything that he intends to carry out sooner or later. On the contrary, until now everyone has best worked in his place in the new Germany if, so to speak, he works towards the Führer. Very often, and in many places, it has been the case that individuals, already in previous years, have waited for commands and orders. Unfortunately, that will probably also be so in future. Rather, however, it is the duty of every single person to attempt, in the spirit of the Führer, to work towards him. Anyone making mistakes will come to notice it soon enough. But the one who works correctly towards the Führer along his lines and towards his aim will in future as previously have the finest reward of one day suddenly attaining the legal confirmation of his work."3
Two things are worth noting about the primary source. First, the dynamic was observable and articulable inside the regime by February 1934 — only thirteen months in. The mechanism was not a later historical reconstruction; it was the regime's own working theory of how it operated. Second, the scholarly lineage that recovered the concept traces through the historian Hans Mommsen, who built the structuralist analysis of the regime that Kershaw later integrated into his Hitler biography.4 The Willikens-Mommsen-Kershaw chain is one of the cleanest cases in 20th-century historiography of a primary source surviving in the archive long enough to be recovered by the right structuralist reading half a century later.
The Kershaw layer adds three primary-source clarifications to the Wilson account. (1) The dynamic was self-described by regime officials, not imposed as an interpretive frame by later historians; the framework Wilson summarises was already explicit in the regime's own bureaucratic discourse by early 1934. (2) The pattern was visible in the agricultural ministry — a substantively unimportant policy area — which indicates the dynamic was not specific to ideologically-loaded domains (anti-Jewish policy, foreign policy) but operated as the general operating principle of the entire state apparatus. (3) The Willikens speech makes explicit that the system rewarded anticipatory action retroactively through "legal confirmation" — subordinates would act, and if they acted correctly the regime would later authorise their actions. The retroactive-authorisation pattern is one of the most consequential features of the regime's operating mechanism and is not visible in the Wilson summary; the primary source makes it explicit.
What this concept gives the rest of the vault:
The German persecution of Jews from Hitler's chancellorship to industrial extermination follows the working-toward-the-Führer pattern step by step.
1933 — early signal. Hitler becomes chancellor. He gives speeches denouncing "Jewish influence" but signs almost no anti-Jewish laws in the first year. The first wave of violence is street-level: SA goons smashing Jewish shops, harassing Jewish businesspeople, beating up Jewish students. None of this is centrally ordered. The SA is competing with the SS for prestige; one of the metrics of prestige is visible anti-Jewish action. Hitler's response is to praise the SA's energy while officially saying these are "uncoordinated excesses." But the praise outweighs the disavowal, and the SA escalates.1
1935 — Nuremberg Laws. After two years of escalating subordinate-action, the regime formalizes citizenship discrimination. The Nuremberg Laws were drafted by lawyers anticipating what Hitler would approve, then submitted to Hitler who approved them. The laws weren't Hitler's idea in any specific sense; they were the bureaucracy's interpretation of what Hitler wanted, codified into legal form. Hitler signed.1
1938 — Kristallnacht. The November pogrom is a textbook anticipation-ratchet event. A Polish-Jewish refugee in Paris assassinates a German diplomat. Goebbels uses this as a pretext to encourage street violence; the SA goes wild. Goebbels is doing his own anticipation of what Hitler wants; Hitler himself was on a train when Kristallnacht broke out and his actual reaction was mixed (annoyance at the property damage, satisfaction at the demonstration of regime-force). But Goebbels read the signals correctly enough that the action was praised. The bureaucracy learned: spontaneous-feeling mass violence is what the Führer wants.1
1941 — Einsatzgruppen. When Germany invades the Soviet Union, special operational units accompany the regular army with orders to "pacify" occupied territory. The actual orders are vague. The Einsatzgruppen interpret pacification as systematic killing of Jews in conquered villages. They shoot ~1.5 million people over the next year. No explicit order from Hitler authorized this; the Einsatzgruppen leadership were interpreting what was wanted. They were rewarded with promotions, praise, additional units. The bureaucracy learned: systematic killing in conquered territory is what the Führer wants.1
1942 — Wannsee. Mid-level officials coordinate the logistics of European-wide systematic killing. The phrase "Final Solution" enters the bureaucratic record. Industrial gassing facilities are constructed. None of this required an explicit Hitler order. The previous decade of anticipation-ratchet had established the trajectory; the Wannsee participants were coordinating execution, not authorizing initiation.1
By 1942, the system was running on its own momentum. Hitler could have stopped it by explicitly forbidding it, but explicitly forbidding it was the one thing his leadership style couldn't do — it would have required him to issue specific policy direction, which would have broken the working-toward-the-Führer architecture across all other domains. The architecture that gave him power-amplification was the same architecture that took specific decisions away from him.
For recognizing working-toward-the-Führer dynamics in any organization:
Diagnostic 1: Does the top articulate direction without policy? When the senior leadership speaks about goals ("we need to dominate this market," "we must transform our culture," "we must eliminate inefficiency"), do they specify how, by when, with what tradeoffs? If the speeches are intense but policy-free, the conditions are present.
Diagnostic 2: Are there parallel power centers competing for senior favor? When multiple managers, divisions, or coalitions have overlapping authority over the same goal, with promotions and resources flowing to whoever shows most progress, the competition-architecture is in place.
Diagnostic 3: Are extreme interpretations rewarded relative to moderate ones? Look at the promotion pattern. The most aggressive employees in service of the senior's goal — those who pushed hardest, accepted the most collateral damage, refused to settle for partial progress — are they the ones rising? If yes, the asymmetric-reward mechanism is producing the ratchet.
Diagnostic 4: Are outcomes drifting further than anyone seems to have ordered? Layoffs that exceed what was announced. Customer-treatment that violates stated values. Strategic moves that contradict explicit memos. If the organization is doing things nobody can point to an order for, working-toward-the-Führer is probably operating.
Intervention question: if you find yourself inside such an architecture, you have three options. (1) Match the equilibrium and rise. (2) Refuse to match and accept that you will be passed over. (3) Try to change the architecture by demanding explicit policy from the top. Option 3 almost never works because the senior leadership benefits from architectural ambiguity and will resist clarification. Most people end up choosing between options 1 and 2, often unconsciously.
How this concept is misapplied:
The concept comes from Ian Kershaw's two-volume biography, which Wilson cites extensively. Kershaw developed it in academic work in the 1990s before popularizing it in the biography. The framework is well-supported in the Holocaust historiography (Christopher Browning's Ordinary Men, Saul Friedländer's Nazi Germany and the Jews, Hans Mommsen's "cumulative radicalization" thesis all converge on similar dynamics). Wilson presents the Kershaw version because it has the most compact name.
Open question: how much explicit Hitler-direction was actually present? Some historians (the "intentionalist" school) argue Hitler was more explicitly directive than the working-toward-the-Führer frame implies. They point to specific Hitler statements (in Mein Kampf, in private table-talk, in coded references in speeches) that they interpret as explicit orders for genocide. The "functionalist" school (where Kershaw mostly sits) sees these as still-vague and reads the operational reality as bureaucratic anticipation. The debate continues. Wilson is a functionalist by default.
Open question: does the framework transfer to non-charismatic-dictator contexts? Probably yes, with modifications. Corporate examples (Boeing safety culture pre-737-MAX; various Silicon Valley product-decisions producing harm; Enron's anticipatory-fraud architecture) all show similar patterns with non-charismatic leadership. The architecture is the key variable, not the leader's specific personality.
Open question: what breaks the architecture? Cases where the ratchet was reversed are rare. The Stalin succession (where Khrushchev de-Stalinized the bureaucracy partly by explicit moderation-policy) is the closest case. The lesson: explicit moderate-policy from a new senior figure can break the architecture, but only if the new figure has the authority of the old and is willing to suffer the political cost of explicit direction.
Kershaw and Christopher Browning (in Ordinary Men) converge on the bureaucratic-anticipation reading but split on emphasis. Kershaw focuses on the leader's signaling architecture (vague-but-intense direction). Browning focuses on the implementers' psychological process (ordinary men becoming killers through small-step normalization). The two are compatible — Kershaw explains how the system was set up; Browning explains how the people inside the system became capable of the actions the system required. The interesting split is with the intentionalist historians (Lucy Dawidowicz, Daniel Goldhagen). Goldhagen in particular reads the same evidence as showing explicit cultural-German anti-Semitism producing direct intentional killing — no architecture required. Kershaw's response: Goldhagen has to explain why German anti-Semites also didn't produce a comparable killing program in 1925 or 1875, when the architecture wasn't there. The architecture is what scales the latent attitude into actual extermination.
Three handshakes, each generating distinct insights.
Behavioral-Mechanics — Shadow Governance Infrastructure: The shadow-governance concept describes how the NSDAP built parallel governmental structures before taking power. Working-toward-the-Führer describes how those structures operated once in power. The two are the input and the output of the same architectural design. Shadow-governance pre-staged the parallel-power-centers; working-toward-the-Führer ran them in anticipation-competition mode. What this connection produces: a unified theory of how movements take and exercise power. The architecture isn't separate from the takeover plan — it is the takeover plan, played forward into governance. Movements aspiring to power should be examined for shadow-governance signs (parallel structures, redundant authority) because those structures will become the working-toward-the-Führer machinery if the movement succeeds.
Behavioral-Mechanics — Menticide: Joost Meerloo's menticide concept describes systematic psychological coercion. Working-toward-the-Führer is the organizational analog of menticide — the leader's vague-signaling reshapes the bureaucracy the way an interrogator's pressure reshapes the prisoner. Both produce subjects who want to do what the coercer wants, who internalize the coercer's frame so deeply that overt orders become unnecessary. What this connection produces: a recognition that organizational architectures can perform menticide on their members. Working in a working-toward-the-Führer organization for years reshapes employees' moral instincts in the same way that prolonged interrogation reshapes prisoners' attitudes. The damage compounds and is hard to reverse. Anyone leaving such an organization will need active reconstruction of their pre-organization moral frame.
Cross-Domain — Dehumanization Precedes Atrocity Temporal Ordering: The cross-domain frame on atrocity-causation argues that dehumanization always precedes mass violence temporally. Working-toward-the-Führer adds the mechanism: dehumanization gets installed through the same architecture that produces the atrocity. The leader's vague-signaling about the enemy installs dehumanization in subordinates while leaving the leader plausibly deniable. By the time the violence happens, the subordinates are operating from a fully dehumanized frame that the leader can claim never to have ordered. What this connection produces: a model of how dehumanization scales beyond what the original signaler could explicitly produce. The dehumanization isn't just propagated — it's amplified by the anticipation-competition. Subordinates compete to demonstrate the strongest enemy-rejection, and that competition rapidly produces more extreme dehumanization than the original leader ever articulated.
The Sharpest Implication. The most uncomfortable thing this concept says is that charismatic leaders who never write explicit orders are not less responsible for their organizations' atrocities — they are more responsible, because the architecture they designed is the proximate cause of the outcomes. The defense "I never told anyone to do that" is, in working-toward-the-Führer cases, an admission rather than an exoneration. It means the leader knew their signaling would produce the outcome through anticipation-competition and chose not to issue the moderating order that would have stopped it. Applied to modern leaders: the question "did the CEO order the harm" is the wrong question. The right questions are "did the CEO maintain an architecture in which harm was rewarded and moderation punished, and did the CEO take action to prevent the harm once the pattern was visible." If the answer is yes-then-no, the CEO is functionally responsible regardless of whether their signature is on the directive.
Generative Questions