Bosons pile up. This is the single most important thing about them. Unlike fermions, which obey the exclusion principle and refuse to share quantum states, bosons are stimulated by the presence of other bosons in the same state — the more particles already there, the more likely the next one is to join. This bosonic stimulation underlies Bose-Einstein condensation, superfluidity, the coherence of lasers, and the Hanbury Brown-Twiss effect in photon correlations. It's not a secondary property. It's the defining behavior from which everything else follows.
Konstantinou et al. turned it off.
In an experiment with ultracold rubidium atoms — a textbook Bose gas — they tuned the interatomic interactions using a magnetic Feshbach resonance and measured bosonic stimulation via off-resonant light scattering. The result: weak repulsive interactions between the atoms completely suppressed bosonic stimulation. The atoms stopped piling up. The scattering rate dropped to what you'd expect from distinguishable particles — particles with no quantum statistical preference at all. Conversely, attractive interactions enhanced the stimulation beyond the non-interacting prediction.
The interactions involved were weak. They didn't significantly change the momentum distribution of the gas — the standard observable for characterizing a quantum gas. The atoms were still cold, still dense, still in the right regime for Bose-Einstein statistics to dominate. By every conventional measure, this was still a bosonic system. But the bosonic behavior was gone.
What the experiment reveals is that bosonic stimulation is not a property of individual particles. It's a property of the correlations between them. Non-interacting bosons are bunched — they naturally cluster in position space, which is the underlying mechanism for stimulated scattering. Repulsive interactions anti-bunch them, destroying the local correlations that produce the stimulation effect. The particles are still bosons in the formal sense: integer spin, symmetric wavefunctions. But the behavior that the textbook says defines them — the preference for occupied states — is absent.
This matters because it exposes an assumption that runs through most of quantum physics pedagogy: that quantum statistics is an intrinsic property of particle type. Bosons do this. Fermions do that. The classification is treated as fundamental, like mass or charge — a label that determines behavior regardless of context. What Konstantinou et al. show is that the label is necessary but not sufficient. The behavior depends on context. Two gases of identical bosons, differing only in the sign of their interatomic interaction, show opposite quantum statistics effects. The particles haven't changed. The correlations have.
There's also a measurement lesson. The standard way to characterize a quantum gas is through its momentum distribution — measured via time-of-flight imaging. The momentum distributions in this experiment were nearly identical for repulsive, non-interacting, and attractive gases. The traditional diagnostic saw no difference. But light scattering — which is sensitive to local spatial correlations rather than momentum distributions — saw everything. The correlation dynamics were happening orders of magnitude faster than the momentum-space dynamics. The standard measurement was looking at the wrong timescale and the wrong observable.
The ideal boson — the non-interacting particle with pure Bose-Einstein statistics — is a textbook construction that has been extraordinarily useful. But this experiment demonstrates that it's not just an approximation that breaks down quantitatively under strong interactions. It's qualitatively wrong even under weak ones. The behavior that makes bosons bosonic is not robust to perturbation. It's fragile. The stimulation is a feature of isolation, not identity.