Why the Shape of a Particle Cluster Tells You More Than a Global Complexity Measure ✛ ✎

Original: "Structural morphology and the gravitational arrow of time"
arXiv:2607.27526v1 · 2026-07-29 · arXiv non-exclusive · independent analysis · 1 min
Two particle clusters with nearly identical global [tag:variety]variety[/tag] can have completely different internal structures, and that matters for understanding gravity's arrow of time.

Two cities with the same average population density can look completely different: one is a uniform grid of blocks, the other a labyrinth of narrow alleys. The same goes for gravity: two central configurations of 5000 particles with nearly identical global clustering measures — variety (differing by only 1.686%)—turned out to be radically different inside. Variety grows in both directions from the Janus point (the moment of minimal system size) and serves as an indicator of the gravitational arrow of time in Barbour's approach. But why is that not enough?

Variety is the ratio of the overall system size to the average distance between nearest neighbors. The larger the value, the more tightly the particles are clustered.

To probe the local geometry, the researchers introduced the six-nearest-neighbor anisotropy A6. For each particle, they looked at how evenly its neighbors are distributed: 0 means they surround it equally from all sides, 1 means they are almost in a line. The average A6 for the first configuration is 0.1331, an almost isotropic background. For the second, it is 0.3419, a difference of 156.8%. In the first, no particle had A6 above 0.5; in the second, 18.1% did — entire regions of elongated structures. So the global variety measure misses the details of internal organization: two 'cities' with the same average density can be either a quiet suburb or a tangle of congested highways. To understand their evolution, you need to look at the geometry of the streets.

The difference in global variety between the two configurations is comparable to the height difference of two people of 1 cm at an average of 170 cm, while the difference in local anisotropy corresponds to one being a weightlifter and the other a marathon runner.

V = \frac{\ell_{\text{rms}}}{\ell_{\text{mhl}}}
Ratio of the root-mean-square size of the system to the harmonic mean distance; characterizes the contrast between the overall size and tight clumps.
A_k^{(a)} = \frac{q_1^{(a,k)} - q_2^{(a,k)}}{q_1^{(a,k)} + q_2^{(a,k)}}
Measure of how stretched the local environment of a particle is: 0 for isotropic, 1 for linear.
Scientists
Albert EinsteinRobert H. DickeJohn Archibald WheelerJacob BekensteinRainer SachsJames Jeans
Concepts
topological phase gravitational constant flat space dimensionless parameter category topology homotopy spacetime curvature topological defect anisotropy topological invariant