Source
Toner and Tu (1995), LIT-1670, pp. 2–9 of the arXiv preprint adap-org/9506001; read in NOTE-1228.
What was actually shown
The puzzle. Vicsek et al.’s rule reduces, at zero speed, to Monte Carlo dynamics of the 2D XY model, which has no long-range order at T > 0; yet their simulations find an ordered phase. “In light of the Mermin-Wagner theorem for equilibrium systems, its existence must depend on fundamentally dynamical, non-equilibrium aspects of the model” (p. 2).
What could have come out otherwise, and did not. The linearized theory of eqs. (1, 2) gives χ = 1 − d/2, so “v⊥ fluctuations grow without bound (like L^χ) as L → ∞ for d ≤ 2” (p. 4): without the nonlinearity, the flock is the equilibrium case. Power counting then shows the convective vertex λ(v⊥ · ∇)v⊥ and the σ₂ pressure term relevant below d = 4 (p. 5).
The exponents. A “Galilean invariance” protects λ, and in d = 2 the noise strength Δ and D∥ are also unrenormalized, which gives z = 6/5, ζ = 3/5, χ = −1/5 (eqs. 9, 10; p. 7). Then ⟨|v⊥|²⟩ = constant − L⊥^{−2/5} g(L∥/L⊥^{3/5}) (eq. 16, p. 8), which “goes to a finite constant as L → ∞”: “long ranged order is stable in this model in d = 2”.
The two unproved steps. The exactness needs the cubic vertex |v⊥|²∂∥v⊥, with linear eigenvalue 3 − d, to stay irrelevant to d = 2; the authors argue it from the φ⁶ term in φ⁴ theory and “strongly suspect” it (p. 8). If it fails, rotational invariance still gives χ = (z − d)/3, and χ < 0 in d = 2 if z < 2, which they “expect on physical grounds”, since “the motion of the “birds” enhances the mixing” (p. 9).
The mechanism, as a heuristic. In Lagrangian coordinates the convective term drops out, “However, at different times, the “neighbors” of one particular “bird” will be different depending on the velocity field itself. Therefore, two originally distant “birds” can interact with each other at some later time. It is exactly this time dependent variable ranged interaction which stabilizes the ordered phase” (p. 3).
What this does not say
- That the exponents are exact. That is conditional here, and is standardly reported to have been revised by Toner’s later reanalysis (not held, not read).
- That any finite flock is ordered. Eq. (16) is about L → ∞; how large a flock must be for the statement to bite is not given.
- That the order is the members’ doing in any sense beyond the field equations. The birds appear only through a conserved density and the coarse-grained velocity; the change-of-neighbours story is a gloss on the convective term, not a calculation about members (CLAIM-280).
- That every flocking rule is in this class. The universality claim is asserted, not derived from a microscopic rule.
- Anything about the transition into order, which is left open (pp. 9–10).