Korzybski’s observation is not complicated. The word is not the thing. The description is not what it describes. The model of a situation is not the situation. These seem obvious when stated plainly, and they are obvious when stated plainly, which is why it is worth asking why the organisations and institutions that govern most of contemporary life operate as though they were not obvious at all.
The map that an organisation produces of its own activity—the reports, the metrics, the KPIs, the budget lines, the performance scores—is a representation. It was produced by a measurement system, which was designed by people who had to decide what to measure, which required them to decide what counted, which required prior decisions about what the organisation was for and how its functioning would be assessed. Each of these decisions involved a simplification. The simplification was necessary. An organisation that attempted to represent its own activity without simplification would produce a representation as complex as the activity itself, which would be useless for the purposes representation is supposed to serve. The map is simplified by design.
The simplification is not the problem. The problem is what happens when the map is no longer recognised as a simplification—when the metric becomes the measure of the thing itself rather than an approximation of it, when the report becomes the activity rather than a representation of it, when the score becomes the performance rather than an index of it. At that point the organisation is managing the map. The territory continues to exist, unmanaged, in the gap between what the measurement system captures and what is actually happening.
The gap is not incidental. It is produced by the measurement system’s design choices, and it is consistent across organisations in ways that reflect consistent choices about what to count. The things that are easiest to count—transactions completed, forms submitted, calls answered, units produced, targets met—tend to be the things closest to the surface of the activity, the most visible and the most discretely measurable. The things hardest to count—the quality of judgment brought to a complex situation, the effect of an interaction on the person who experienced it, the accumulation of small decisions that produces an institutional culture over time—tend to be further from the surface and more resistant to the measurement apparatus.
The organisation that measures what is easy to measure and manages what it measures will, over time, become increasingly proficient at producing the measurable outputs and decreasingly attentive to the harder-to-measure things the measurable outputs were supposed to approximate. The map is optimised. The territory drifts. The drift is not visible in the measurement system because the measurement system is the map.
Korzybski was writing in 1933, before the infrastructure of contemporary measurement had been built. The organisations he was observing were using simpler abstractions—balance sheets, headcounts, production volumes. The principle he identified was visible in those simpler forms. The contemporary version has the same structure at considerably greater scale and sophistication. The measurement apparatus is more elaborate, produces more data, generates more confident-looking outputs, and is therefore more convincing as a representation of the territory rather than a map of it. The elaboration of the map makes the gap between map and territory harder to see, not because the gap has narrowed but because the map has become more convincing.
The policy is the map of the behaviour it is supposed to govern. It describes what people in the organisation should do in defined circumstances and produces, in the organisation’s self-understanding, a picture of the behaviour as conforming to the policy. The picture is accurate when people follow the policy. It is also accurate, from the measurement system’s perspective, when people perform compliance with the policy—fill in the forms, attend the training, sign the acknowledgments—without their behaviour having changed in the ways the policy was designed to produce. The map shows compliance. The territory may contain something else.
The organisation that has implemented a policy on, for example, the treatment of customers in difficult circumstances, and has measured compliance through completion of the relevant training and the absence of formal complaints, has a map that shows a compliant organisation. Whether the customers in difficult circumstances are being treated the way the policy intended is a question about the territory. The measurement system has not been to the territory. It has measured the map.
This is not a failure of intention. The people who designed the policy intended the behaviour they described. The people who implemented the compliance measurement intended to verify that the behaviour was occurring. The gap is structural—the measurement system is capable of verifying the map and cannot verify the territory, and the organisation has learned to manage the map because that is what can be managed within the apparatus available.
Korzybski’s principle would require the organisation to treat its maps as provisional—as approximations that require periodic re-examination against the territory they are supposed to describe. The re-examination is not what most measurement systems are designed to produce. They are designed to produce confirmation—to verify that the organisation is performing as the map says it should be performing, to generate the signal that allows the hierarchy to confirm that the directives it has issued are being followed, to produce the reports that demonstrate to external stakeholders that the organisation is functioning as claimed.
A measurement system designed to produce confirmation cannot produce the finding that the map is wrong. It can produce the finding that the metrics are not being met, which generates a response designed to bring the metrics back into range. It cannot produce the finding that the metrics are measuring the wrong things, or that the policy has had unintended effects that the measurement apparatus is not capturing, or that the territory has changed in ways that the map no longer represents. These findings require a different orientation—one that treats the map as a hypothesis rather than a description, and the measurement as a check on the hypothesis rather than a confirmation of it.
The distinction is between a system that asks whether it is performing as its map says it should, and a system that asks whether its map is accurate. Most systems ask the first question. The second question is more uncomfortable because it introduces the possibility that the confident claims the organisation has been making about its own performance are claims about the map rather than the territory—that the organisation has been performing well on the measures it has chosen while the territory those measures were supposed to approximate has been doing something the measures did not capture.
The confidence is the operational requirement. An organisation that publicly acknowledged that its metrics were approximations, its policies maps rather than territories, its performance reports representations rather than descriptions, would be acknowledging a level of uncertainty that the people the organisation serves, and the people within it who require clear direction, cannot easily function with. The hierarchy depends on confidence. The directive issued with the qualification that it is based on a model that may not accurately represent the situation it is designed to address loses the authority that makes it a directive rather than a suggestion. The report that acknowledges its own uncertainty loses the clarity that makes it useful for decision-making at speed.
So the confidence is maintained. The map is presented as the territory. The measurement is presented as the thing measured. The report is presented as the activity. Not because the people producing these things do not know they are representations, but because the system requires them to be presented as though they were not, and the people within the system have learned to operate within that requirement without examining it.
Korzybski’s point was not that maps are useless. Maps are necessary. The navigation of any territory more complex than immediate sensory experience requires representation. His point was that the usefulness of a map depends on the navigator knowing it is a map—knowing that it simplifies, that it is produced from a particular perspective, that it was made at a particular time and may not reflect current conditions, that the territory will contain things the map does not show. The navigator who knows this uses the map as a tool. The navigator who has forgotten it uses the map as a destination.
The organisation that has forgotten it is navigating toward its own representation of itself. The KPIs are met because meeting the KPIs is what the organisation has learned to do. The policies are complied with because compliance is what the measurement system measures. The reports are clean because the reports are managed. The territory—the actual behaviour, the actual outcomes, the actual experience of the people the organisation serves—is somewhere behind the map, receiving less attention the more polished the map becomes.
This is not a problem that better measurement solves, because better measurement produces a more detailed map, not a closer relationship to the territory. The detail increases. The confidence increases. The gap between the confident description and what is actually happening remains, reshaped by the more detailed measurement but not closed by it.
The map is very good now. It is precise and elaborate and generated at speed and displayed on dashboards that update in real time.
The territory continues to exist outside it.
The organisation has not been there recently.
It is managing the dashboard.
