The Map Is Not the Territory: How Models Leave Things Out
The map is not the territory: see how models leave things out, spot when missing details change an answer, and choose what to check next.
The most common mistake with a map is trusting its tidy line more than the uneven ground: a shortcut marked as open may end at a locked gate. The map is not the territory, and that gap matters because a representation can guide a decision while hiding the detail that changes it.
Models are built from choices
A model is a representation made by selecting and arranging information. A bus diagram makes routes and connections easy to follow; it may distort the physical spacing between stations. The distortion is part of how the diagram works, not a claim about the landscape.
Other models make different choices. A water-cycle diagram turns a complicated set of movements into a few named stages. A school timetable turns a day into slots and subjects. Neither records every event it represents; each gives a particular structure to information so that a reader can reason with it.
Those choices shape what is easy to notice. A timetable makes a clash between lessons visible, but it does not show whether a classroom is noisy or a student needs more time to get there. A model can organise thought without containing all the facts that thought may eventually require.
Missing information can change an inference
Omitted details matter when someone draws a conclusion that depends on them. A chart showing that two events occur together does not, by itself, tell you that one caused the other. The missing explanation might involve a third factor, a sequence of events, or coincidence.
It helps to distinguish what a model displays from what a reader infers. A diagram might show arrows linking several stages. The arrows establish the relationships the diagram chooses to present; they do not automatically settle whether every real instance follows that sequence or why it does so.
Say you are given an invented classroom diagram in which a plant receives sunlight, grows, and then produces seeds. You use it to explain why a particular plant in a windowless cupboard has not flowered. The diagram has no information about that plant’s watering or condition, so it cannot identify the cause; the sound answer is to treat light as one possible factor and seek observations about the plant itself.
Representations can organise evidence differently
Two models of the same subject may disagree in appearance because they arrange evidence for different purposes. A family tree groups people by descent; a timeline groups events by date. Neither arrangement cancels the other, but each makes some relationships easier to inspect than others.
Models can also embody different assumptions. A simple economic diagram might treat people as if they respond to incentives in a consistent way. That assumption can make a line of reasoning manageable, but it leaves open whether actual people in a particular situation respond as the diagram predicts.
When models conflict, compare the claims each makes and the evidence each can accommodate. Ask whether they use the same definitions, include the same cases and rely on the same assumptions. Sometimes the disagreement is substantive; sometimes the models are answering different questions with differently organised information.
How do models shape what you can infer?
Does a diagram prove that a process always happens this way?
No. A diagram presents a selected account of a process, often in simplified form. To claim that the sequence is universal, you need evidence that addresses variation and exceptions.
What should you do when two models disagree?
Set their claims side by side and identify differences in definitions, assumptions and included evidence. If the disagreement remains, look for observations or sources that can help distinguish between the competing accounts.
Can a model be useful if some of its assumptions are false?
It can still help organise a question, provided you do not treat its conclusions as automatic descriptions of reality. Check which results depend on the questionable assumption before carrying them into another setting.
How can you tell whether a conclusion goes beyond its evidence?
Ask what would have to be true for the conclusion to follow, then check whether the model actually supplies that information. If it does not, state the conclusion more narrowly or seek additional evidence.
Use a concrete check before extending a model
When a model seems to answer a question, use these checks to decide whether its information supports the next inference:
- Check the labels: If a term is defined clearly, use that meaning; if it is vague or absent, clarify the term before comparing claims.
- Check the stated assumptions: If the conclusion relies on an assumption you accept for this question, proceed cautiously; if not, revise the assumption or seek another account.
- Check what evidence is shown: If the model includes evidence relevant to the claim, assess that evidence; if not, look for a source that addresses the missing information.
- Check for alternatives: If another explanation could produce the same pattern, keep both possibilities open; if the model rules one out with evidence, explain why.
- Check whether the conclusion changes across cases: If it holds only for some cases, narrow the claim; if it holds across the cases examined, describe that scope rather than claiming more.
- Check what observation could challenge the claim: If a possible observation would count against it, consider whether that observation is available; if none is specified, make the claim more precise before relying on it.
Imagine, as an invented exercise, that a class uses a simple diagram to explain why a fictional pond turns cloudy. The diagram links rain to soil washing into the water. The class then learns that the pond looked cloudy during a dry spell, so it cannot use the diagram alone to explain that occasion.
They check the diagram’s labels and find that it covers rain-driven runoff, not every cause of cloudiness. They list an alternative, such as disturbance of the pond bed, and decide to inspect notes about recent weather and activity near the water. The result is a narrower answer: runoff is one possible explanation after rain, while the dry-spell instance needs other evidence. That is a more defensible decision than treating the diagram as a complete account.
What is not worth the effort
It is not useful to reject a model simply because it leaves things out. Every representation selects information; the better question is whether a particular omission affects the inference being made.
Nor is it helpful to demand that every model contain every possible detail. That would make many representations too crowded to use. Spend your effort on assumptions and missing evidence that could alter the claim, rather than on omissions with no bearing on the question.
The map is not the territory, but it can organise thought
The map is not the territory: the phrase is useful because it marks a difference between a representation and the subject it helps us reason about. A model can make relationships visible, but its structure and assumptions also shape which inferences seem natural.
When a conclusion matters, name the claim, identify the evidence that supports it and look for the assumption connecting the two. If that connection is missing, the next useful move is not to force certainty from the model; it is to narrow the claim or find evidence that can answer the unresolved question.
