"The map is not the territory." A saying we quote constantly in strategy meetings. We rarely stop to ask what kind of territory we are actually dealing with—and whether the assumption of a fixed map gets in our way.

Because the phrase still assumes a territory—a stable, pre-existing reality that exists independently of your attempt to represent it, which a sufficiently skilled cartographer could eventually chart more accurately. That assumption holds for situations that are clear and complicated. It breaks for the complex.

In the complicated domain, cause and effect are real and fixed, just not obvious—they require expertise to uncover. A better map is genuinely possible there; you just need more analysis, better instruments, the right experts. In that specific world, warning people not to mistake the representation for the thing itself is exactly the right caution: people over-trust an incomplete map when a more complete one is achievable in principle.

In genuinely complex situations, there is no territory sitting there waiting to be surveyed. Cause and effect only become coherent in retrospect—Dave Snowden uses the term "retrospective coherence"—which means the pattern doesn’t exist prior to your engagement with the system. It is constructed after the fact, from probes you made into it.

And the probing itself changes the system, so the "territory" you would be mapping is partly generated by your own act of trying to map it. There is no cartographer standing safely outside, surveying a fixed landscape. Two different sequences of probes, made a week apart, can generate two different retrospectively coherent stories, both completely valid. That is not a map being imperfect. That is the map-territory split—observer here, fixed reality there—not applying to begin with.

This is also why Snowden insists his Cynefin framework is for sense-making, not categorization. In a categorization framework, you define the boxes first and then assign phenomena to them. In a sense-making framework, you observe the properties of the situation and let the assignment emerge.

A map is a categorization artifact—draw it first, navigate by it after. That is precisely the sequence complex work has to invert: probe, sense, and then respond, rather than mapping, analyzing, and responding. If you try to map first, you are treating a complex problem as if it were merely complicated in disguise, which is the most common failure mode in strategy.

The practical version: the moment someone says "before we act, let's get more data and build a clearer picture," ask whether the situation is one where a clearer picture is actually obtainable in advance, or one where the picture can only be assembled by looking backward at what a small action just showed you.

If it is the second, more analysis isn't caution, it's just hesitating.