In the article series on the invariants of the Adaptation Coefficient, this one opens the way for a simple reason: it is the most robust of the eight individual invariants identified in the corpus, and probably the least intuitive. Instinctively, we associate adaptability with acquisition: learning a new skill, adding a tool, expanding one’s repertoire of action. The fifteen interviews conducted for this research tell a different story. What sets the most adaptable profiles apart is not primarily what they added, but what they agreed to let go of.
What unlearning is not
The first reflex, in interviews, is to confuse unlearning with resignation. The two look alike on the surface: in both cases, one stops doing what one used to do. But one of the leaders interviewed articulated the distinction with a precision that ended up structuring the analysis of the entire corpus: resignation is an endured capitulation, the abandonment of a goal under constraint, whereas adaptation is a conscious choice, the revision of a path without abandoning the destination. This same leader summarized what differentiates the two postures: adapting is not losing your heading, it is accepting that the path you initially imagined may not be the right one.
This distinction changes how an interview should be read. Someone who describes an abandonment in terms of loss, failure or outside constraint is probably documenting resignation. Someone who describes the same abandonment as a decision, even a painful one, is documenting active unlearning. The vocabulary used, and the presence or absence of an acting subject in the sentence, become clues just as important as the factual content of the account.
A skill documented fifteen times, in very different forms
What stands out, when cross-referencing profiles, is not the existence of unlearning among high-AC profiles, which was largely expected, but the diversity of forms it takes.
Some instances of unlearning concern a specific professional belief: the conviction that a given method remains superior to all others, held well beyond the point where it stopped being true, and abandoned with an explicit acknowledgment of having taken too long to let go of it. Others concern a more diffuse belief about transitions themselves: the idea that a change of role or sector forces you to start from scratch, when in fact a thread of skills and values can run across very different professions. Others still concern an entire organizational model, known exclusively before being challenged by contact with a radically different one.
One of the clearest examples concerns giving up the leader’s own omniscience: accepting no longer being the person chosen by the client on every file, in favor of a team of consultants to whom trust is delegated. This seemingly minor shift actually touches the leader’s professional identity, which is why it ranks among the most significant instances of unlearning in the corpus.
Other configurations broaden the spectrum further. One leader shows the corpus’s most striking example in terms of personal cost accepted to unlearn an inherited management practice. Another documents an unlearning driven more by scarcity of available resources than by deliberate anticipatory analysis, a reminder that one can unlearn without having fully chosen the moment. An entrepreneur who went through several successive crises illustrates an iterative unlearning, built through an accumulation of trials rather than a single decision. A leader at the head of a very large organization documents an unprecedented coupling between her own individual unlearning and an organizational unlearning carried out at scale, each movement feeding the other.
Another configuration deserves to be singled out because it introduces an additional form of vigilance: that of a leader who actively monitors her own tendency, as experience accumulates, toward reversed unlearning, the natural drift of hardening one’s certainties as experience builds up, even though that experience should in principle favor nuance. This metacognitive vigilance, rare in the corpus, constitutes a form of preventive rather than reactive unlearning.
Finally, one figure leading an institution documents unlearning by proxy: directed at transforming the organization rather than at her own personal practices, which raises an interesting analytical question about where unlearning is located, individual or collective, and whether one can substitute for the other without ever being confronted directly. One last configuration, documented in a family-business leader, shows the unlearning of a conviction inherited through family transmission and overcome after a transgenerational trauma, adding a long-term biographical dimension rarely present in the other profiles.
Two opposite ways of staying stuck
The corpus also documents, just as instructively, what happens when unlearning does not take place. Two structurally opposite configurations produce the same capped score around 40 out of 100.
The first is acknowledged resistance: a leader explicitly recognizes his difficulty in letting go, naming it as a functional character trait rather than denying it, while otherwise showing great agility in action. This is genuine lucidity about a recognized rigidity.
The second is almost its exact opposite: denial through the claim of having no fixed convictions. Asked about an abandoned conviction, one leader replies that he has always had the humility never to lock himself into deep convictions. This answer, seemingly open, actually closes the question rather than opening it: it rests on no concrete example, even though this leader’s career documents a stable managerial doctrine theorized for fifteen years, never presented as having been questioned or revised. Where the first profile names his rigidity, the second denies its very existence, which makes the flaw harder to work on, precisely because it presents itself in the clothing of flexibility.
These two mechanisms, one candid, the other more devious, lead to the same floor score on this indicator. The corpus suggests that, despite their opposite appearances, they function as two variants of the same phenomenon: the absence of a concrete, verifiable example of a conviction genuinely abandoned.
What genuinely differentiates profiles
This is the most useful conclusion of this invariant, and probably the most counter-intuitive one: it is not the posture displayed toward change, declared resistance or proclaimed openness, that differentiates the most adaptable profiles from the others. It is the presence or absence of a concrete, dated, verifiable example of a conviction genuinely abandoned. A discourse on flexibility, however sincere it may seem, is worth nothing without this empirical proof. Conversely, an acknowledged and named resistance can coexist with a relatively high AC, provided it stays localized and that other convictions have genuinely been revised.
This conclusion has a direct methodological consequence for reading any discourse on change, well beyond the scope of this research: distrust unsubstantiated declarations of openness, and systematically look for the concrete example behind the statement of principle.
Why this invariant is a necessary condition, not a condition of excellence
Within the three-level hierarchy that structures this entire research, active unlearning belongs to the necessary conditions: those found, at least in a partial or forced form, even among moderate-AC profiles. None of the fifteen profiles analyzed is entirely devoid of it. What varies, however, is the nature of the unlearning documented: chosen or forced, personal or by proxy, one-off or iterative, acknowledged or denied. This variation explains why unlearning, although necessary, is not enough on its own to move a profile into the highest zones of the Adaptation Coefficient. It opens the door to the intermediate zone; it does not guarantee going beyond it.
A hypothesis about the mechanism
Unlearning seems to precede adaptation rather than accompany it. Without unlearning, one can add skills, tools, methods, without changing the fundamental behaviors, which remain governed by the unrevised conviction. The corpus suggests that the absence of unlearning, whether named with lucidity or denied through an unsubstantiated declaration of flexibility, produces the same capped score. What matters, then, is not the declared relationship to change, but documented, verifiable proof of a conviction genuinely abandoned.
This requirement of a genuinely revised personal conviction echoes another individual invariant in the corpus, adaptation as a “third way” between resistance and submission: unlearning is neither yielding nor digging in, it is consciously choosing to revise one’s path without abandoning one’s destination.
Next article in the series: conscious segmentation between a stable core and a flexible periphery, what never moves so that everything else can.
Related articles: