In 1924, the Western Electric Company launches a series of studies at its Hawthorne plant near Chicago. The aim is to measure the effect of lighting on productivity. Lighting is increased. Productivity rises. Lighting is decreased. Productivity rises too. The independent variable explains nothing. What acts is the presence of the observers. The workers know they are being studied. The measurement altered what it was measuring.
Goodhart (1975) makes it a law: when a measure becomes a target, it ceases to be a good measure. The thermometer alters the temperature of the bath. The poll alters the vote. The audit alters the performance.
Heisenberg (1927) formalises the limit case. Measuring the position of a particle disturbs its momentum. The disturbance is not a technical artefact. It is inscribed in the structure of the theory. The product of uncertainties has a floor: $\Delta x \cdot \Delta p \geq \hbar/2$. In quantum mechanics, instrument and system are not separable. Nor are they in the social sciences.
Doctrine
Observation is not neutral. It never was.
The Hawthorne effect and the uncertainty principle are the same problem seen from two different floors of the same building. The difference is that physics formalised it. The social sciences observed it without formalising it. It is not a flaw in the protocol. It is the structure of all observation.
Vecteur ouvert
Every instrument of control modifies the process it controls. Every observation report modifies what will be observed next. Every specification modifies the process that realises it.
The question is not whether measurement disturbs. It always does. The question is at what threshold measurement produces more reality than it records. And if that threshold exists, on which side we operate.
