Tensions in Practice: An untouched baseline forecast in tension with guidance that changes behavior¶

Six people choosing between two invented destinations

All six people initially plan to go to A. A conditional forecast says that if plans stay unchanged, A will receive six and B zero. Keeping that forecast private leaves the declared plans unchanged. Publishing it as guidance triggers a stipulated response: three people switch to B, producing a three–three split. The public message is useful because it changes the target, so its conditional baseline cannot be evaluated as though that response never happened.

Observe the baseline path

Keep the forecast from changing the choices it describes.

Use information to alter the path

Publish guidance that induces a more balanced allocation under the declared response.

Why these aims pull against each other

A forecast can be used as an observation or as an intervention. Once its publication changes choices, the resulting outcome answers a different question from the unchanged-plan baseline.

Compare the arrangements

Keep the baseline private

Retain the conditional forecast in a private evaluation record. In the declared setting no other event changes anyone’s plan.

What it protects: The realized six–zero split can be compared with the unchanged-plan baseline without this message affecting it.
What it costs: The information does not help participants coordinate a more balanced split.
When it fits: Fits a shadow evaluation whose purpose requires leaving this response channel inactive.

Illustration note: The comparison isolates publication only because other influences are held fixed in the toy. A real private forecast can still affect behavior through other channels.

Publish conditional guidance

Publish the six–zero conditional baseline along with its unchanged-plan condition. By the stipulated response rule exactly three people switch to B.

What it protects: The information contributes to the declared three–three allocation.
What it costs: The observed outcome no longer directly tests the untouched baseline; evaluation needs an account of the publication response.
When it fits: Fits a guidance task where changing the allocation is intended and the response assumption is appropriate.

Illustration note: The message does not falsely claim six–zero will persist after people react. The three–three result is a separate prediction of the stated response model, not a repair to a knowingly wrong forecast.

What this illustration does—and does not—establish

The source supplies the structural tension. This bounded example makes a particular relation inspectable; the aims, conditions and residual costs are part of the comparison.

The six people, initial plans and exact response are invented. The structural source does not establish a real-world response magnitude or guarantee balancing.
This uses the source’s explicitly qualified feedback extension of performativity, not the claim that every forecast constitutes a fact merely by being uttered.
No simultaneous measurement of the untreated counterfactual is obtained from the published run. The private arrangement is a distinct stipulated scenario.

Source entries

Performativity

Prime · Source of the tension

The canonical tension motivates this comparison. The setting and arrangements are declared editorial illustrations, not observed findings.

The line between feedback-driven constitution and ordinary description is a matter of degree, not kind

Economic performativity says a model can reshape its referent, but every description has some feedback potential, and most have negligible amounts.

Read the source section

The source’s feedback scope

Performativity is the right lens for institutional and status facts (money, marriage, citizenship, corporate personhood) and for cases where description feeds back into the described; it is the wrong lens for brute physical facts that obtain regardless of any human act.