You Became What You Were Fighting: The Replication Problem in Leadership

"I've been doing this a long time."

The sentence landed in a Tuesday meeting, mid-afternoon, aimed at a direct report who had proposed an alternative timeline for a rollout. It was spoken with the kind of calm authority that sounds like competence and functions like a door closing. The person who said it had spent three years, earlier in their career, sitting across from a director who used that exact phrase to end conversations. They had described that director, in a coaching session, as "someone who made everyone around her smaller by making herself the only person in the room whose experience counted." The description was precise. It was also, without a single word changed, an accurate description of what just happened in the Tuesday meeting.

Bandura (1977) established that behavioral patterns are acquired primarily through observation and modeling, and the finding has a specific implication for leadership that most leadership development programs skip past because it is uncomfortable. The leadership behaviors most deeply encoded in a person's repertoire are the ones they were exposed to under conditions of high emotional arousal: stress, fear, admiration, the desperate desire to survive a room. Those are the conditions where learning is most efficient and least conscious. A person does not have to admire the leader they are replicating. They do not have to want to replicate them. The encoding happened through exposure, the way an accent gets absorbed by living in a place long enough. Deciding to speak differently does not change the accent when the accent was acquired below the level of decision.

The leaders most at risk are the ones who swore they would be different. The determination creates its own rigidity: a hypervigilance about the obvious version of the pattern that leaves the subtle version unguarded. The meeting where the direct report's timeline gets overridden does not look like the old boss's behavior, because the old boss would have been dismissive. The new version is collaborative, engaged, supportive, and arrives at the same outcome: one person's judgment replacing another person's judgment, with the replacement justified by seniority rather than examined on its merits. The language has been updated. The operating system has not.

Argyris and Schon (1974) drew a distinction that sits at the center of this problem. They separated "espoused theory," what a person says they believe about leadership, from "theory-in-use," what a person actually does under pressure. The gap between the two was, in their research, nearly universal. A leader who espouses collaborative decision-making will, under time pressure, consistently override team input while describing the override as efficiency. The override is not hypocrisy. It is a behavioral program installed during an earlier career, under a leader who decided things quickly, and the program runs when conscious deliberation does not have time to intervene. Tuesday afternoon, mid-rollout discussion, tired, behind on three other deliverables: that is when the program runs. Not on the retreat day when everyone is reflective and the espoused theory has room to operate.

Fleenor et al. (2010) added a finding that makes the self-correction problem sharper. Studying the gap between leaders' self-ratings and ratings from their direct reports, they found that agreement was lowest for the competencies leaders valued most. A leader who places high value on empowering others is the leader most likely to overestimate how empowering their actual behavior is. The importance of the value inflates the perception of living it. Caring about empowerment feels like practicing empowerment, and the feeling is convincing enough that the leader stops checking, which means the gap between intention and impact widens in exactly the area where the leader is most confident it has closed.

The ironic version of this, the leader who swore to be different and ended up the same, gets the most attention because it makes the best story. But the more common version is quieter and, in organizational terms, more expensive. A regional manager at a logistics company was promoted three times in six years. She was effective, well-liked, and had never once experienced a crisis about her leadership style because she had never questioned it. Her leadership looked like the leadership above her, which looked like the leadership above that, which had been installed by a founder in 1997 who believed that clear direction and tight accountability were the hallmarks of operational excellence. The regional manager agreed. She agreed because the model worked, and it worked because the organization had been built around it, and the organization's success confirmed the model, and the confirmation loop was so complete that questioning it would have required her to question the thing that made her successful, which is a thing almost nobody does voluntarily. She was not fighting anything. She was not performing anything. She was executing a program that had been installed through success rather than trauma, and the installation was invisible because it arrived wearing the clothes of competence rather than the clothes of damage. This is the version of the replication problem that fills most leadership pipelines, and it does not produce the disorienting mirror moment because there is no mirror. There is just a hallway of people who all learned to lead from people who all learned to lead the same way, and the hallway feels like culture.

This is where the replication problem becomes structural rather than personal. A leader replicating an old pattern is doing so inside a system that often rewards the replicated behavior. The organization needs decisions made quickly. The team has been conditioned to defer. The time pressure is real, and in that context, the override produces results that look like leadership from every angle except the one occupied by the person whose judgment was replaced. The system does not punish the replication because the system was built by people running similar programs, and the programs, operating at scale across an organization, create a culture where speed reads as competence and consultation reads as indecision. Padilla, Hogan, and Kaiser (2007) described this as the environmental leg of what they called the "toxic triangle," and the word "toxic" is misleading because the leaders in question are often excellent. They are excellent at the version of leadership the system selects for, which is the version their predecessors modeled, which is the version being replicated, which is why the loop is so difficult to see from inside it.

The part that resists clean advice, the part most worth sitting with, is that awareness does not close the gap at the rate most people expect. A leader who can say "I know I override my team when I'm stressed because my previous director modeled that behavior and it was installed during a formative professional period" has accomplished something real. The accomplishment is diagnosis. Diagnosis is necessary and it is also, on its own, a resting place that can feel like the destination. The vocabulary of self-knowledge creates a sensation of change that the nervous system registers as completion: the pattern has been named, the origin has been identified, the analysis is coherent. Whether the behavior on Tuesday afternoon is any different is a separate question that the analysis, no matter how sophisticated, cannot answer by itself.

There is a practice for this, and it moves through four stages that each address a different piece of the problem. The first stage surfaces something most leaders have never examined at the granular level: the specific behavioral moments where their leadership most closely resembles the leadership they experienced. Not the general philosophy but the precise thing: "When a direct report pushes back on a deadline, a surge of irritation arrives and the override follows, and that sequence is the same sequence that played out in 2019 with a director whose name still produces a tightness in the shoulders." The specificity matters because generalities can be analyzed away. Specifics stick. One person sat with the first prompt for forty minutes and wrote half a sentence, and that half-sentence, once it existed on paper, became the starting point for the most significant behavioral change observed in five years of this work.

The remaining stages address the emotional driver underneath the replicated behavior, the measurable gap between self-perception and team experience, and the concrete structural change that makes the pattern harder to run even on the days when awareness lapses. Because awareness will lapse. A system that depends on perfect self-monitoring is a system designed to fail on Tuesdays, and Tuesdays are when it matters.

The full framework, including the four-stage practice, is at [counterplay-ten.vercel.app](https://counterplay-ten.vercel.app). J. Fraser works with leaders navigating the replication problem.