Fluctuations and Response

Relates equilibrium fluctuations to susceptibilities and linear responses of the system.

Conceptual context

At equilibrium, a system is not microscopically motionless. Its observables continually vary around mean values even when macroscopic controls appear constant. Those variations contain physical information: they reveal how easily the system can be displaced by a weak external perturbation. $\sigma_A^2$ summarizes fluctuation amplitude, while $\chi$ measures linear response to $h$. Their connection is not accidental. The same degrees of freedom that fluctuate spontaneously are those that respond when a field is applied. $T$, encoded through $\beta$, sets the scale between equilibrium noise and sensitivity.

🟢 Essential level

A system at equilibrium is never microscopically still. The instantaneous value of an observable wanders around its $\langle A\rangle$, and $\sigma_A^2$ measures the strength of those variations.

If fluctuations are large, the system is usually easy to displace with a weak $h$, so its $\chi$ is also large. This connection allows response to be predicted without strongly forcing the system: one can observe spontaneous behavior instead. $\beta$ introduces the thermal scale, meaning that the same variance can correspond to different responses at different $T$. Equilibrium noise therefore becomes a physical probe rather than an unwanted nuisance. This reading is the starting point of linear response theory: the same mechanism that produces spontaneous fluctuations also governs the response induced by an external field.

A concrete historical example is Johnson-Nyquist thermal noise: in 1928, John B. Johnson experimentally measured the voltage that spontaneously fluctuates across a resistor due to thermal agitation of its electrons, and Harry Nyquist explained that same year, using statistical-mechanics arguments, that the noise intensity is fixed by the $T$ and the conductor's own electrical resistance.

🔵 Formal level

Basic formulas

The spread of an observable is defined from its $\langle A^2\rangle$ and $\langle A\rangle$:

Variance of an observable

The result is $\sigma_A^2$, which is nonnegative and has squared observable units. In canonical equilibrium, static response to the conjugate field is related to this spread:

Susceptibility-fluctuation relation

Real student questions

Is a fluctuation the same as experimental error? No. Error belongs to measurement; fluctuation may be a genuine equilibrium property.

Why can variance never be negative? Because it averages squared deviations from the mean.

Does large susceptibility imply instability? Not always, but it may signal proximity to a transition or extreme sensitivity.

Can the energy relation be used for an isolated system? Not directly. The formula involving $\sigma_E^2$ assumes canonical fluctuations.

What if the field is large? The response becomes nonlinear and may show saturation, curvature, or hysteresis.

Final synthesis

Equilibrium fluctuations reveal system sensitivity. Their variance, scaled by $\beta$, determines $\chi$ and predicts linear response before the system is strongly disturbed.

Coherence between fluctuation, response, ensemble, and thermal scale must be preserved in every physical interpretation. This means checking that the same observable, the same conjugate field, and the same temperature are used when comparing variance and susceptibility. When these conditions hold, equilibrium noise becomes a reliable predictor of induced response.