# Wajnryb, Szymczak & Cichocki (2004) — Brownian Dynamics: Divergence of Mobility Tensor

> Pulse 55 source note · logged 2026-06-20 (medulla pulse 55) · Anthrocybernetics research

**Citation**
E. Wajnryb (Inst. of Fundamental Technological Research, Polish Academy of Sciences), P. Szymczak (piotrek@fuw.edu.pl, Inst. of Theoretical Physics, Warsaw University), B. Cichocki (same). *Physica A* **335** (2004) 339–358. Received 16 Oct 2003. DOI: [10.1016/j.physa.2003.12.012](https://www.fuw.edu.pl/~piotrek/publications/bdd.pdf). PACS: 83.10.Mj; 82.70.Dd; 66.10.Cb; 47.15.Gf.

**Local copy:** `Anthrocybernetics/research/bdd.pdf` (mirror of the author-hosted preprint).

---

## What the paper actually claims

A many-body **Brownian dynamics** / low-Reynolds-number hydrodynamics paper. It derives an **analytical multipole-expansion formula for the divergence of the mobility tensor** for *N* identical spheres suspended in an incompressible viscous fluid (Stokes flow), for both an unbounded fluid and periodic boundary conditions.

- **Mobility matrix** μ_tt maps forces → velocities: `U_i = Σ_j μ_tt,ij F_j`. It is non-diagonal because of **hydrodynamic interactions** and depends on the full configuration *X* of all particles.
- The **divergence vector** `d_i = Σ_j ∇_j · μ_tt,ji` is needed for two things:
  1. stepping the **Smoluchowski equation** in Brownian dynamics simulations (Eq. 1.3), and
  2. the **memory function** *M(k,t)* that governs the gap between the short-time and long-time **collective diffusion coefficients** (D_c^s vs D_c^l). The memory strength is `α = (D_c^s − D_c^l) / D_c`.
- Brute-force finite-difference divergence needs ~3*N* evaluations of μ → slow and inaccurate. Their analytical formula reuses the **same multipole functions** already computed for μ, so divergence comes essentially for free.

### The load-bearing result (for our purposes)

> In any **two-body approximation** (Oseen or Rotne–Prager tensors), the divergence vector **d_i vanishes** — and with it **all memory effects** (M = 0). Only **many-body hydrodynamic interactions** produce a non-zero divergence, and therefore a non-zero memory function.

In plain terms: **the memory of the suspension lives in the many-body structure of the medium, not in the particles.** Strip the medium down to pairwise couplings and the collective memory disappears. This is a precise, peer-reviewed instance of "the medium carries the memory."

### Numerical memory factor α (hard-sphere suspension)

| φ (vol. frac.) | α estimate |
|---|---|
| 0.01 | 0.01% |
| 0.1  | 1%   |
| 0.2  | 3%   |
| 0.3  | 5%   |
| 0.4  | 7%   |
| 0.45 | 7%   |

Memory contribution is small at low concentration but climbs to ~7% at φ = 0.45.

### Long-time tail

The collective-diffusion memory function decays **algebraically**, `M(t) ~ t^(−3/2)` (or `t^(−5/2)` when the leading amplitude M₁ vanishes) — a hydrodynamic long-time tail, **not** exponential decay. This is the standard signature of long-range coupling through a continuum medium.

---

## Relevance to the ghoju / Anthrocybernetics frame

**Honest scope first:** this is a 2004 statistical-mechanics paper about colloid diffusion. It makes **no** claims about resonance, EM fields, consciousness, or anything speculative. The bridges below are **structural / analogical**, not empirical links. It belongs in the "real physics that the metaphor rests on" bucket, not the "evidence for the resonance framework" bucket. Use it accordingly.

1. **Field-mediated coordination without contact.** Hydrodynamic interactions (the Oseen tensor ~ 1/r) let particles coordinate their motion through the fluid **without touching**. The mobility tensor *is* the coupling kernel. This is the clean-physics version of the ghoju **"overlapping resonance / convergent incentives"** idea — coordinated effects emerging from compatible couplings through a shared medium, no orchestrator required — and of the **aether-as-medium** thread. The medium does the coordinating.

2. **Memory lives in the medium, not the nodes.** The vanishing-divergence-in-two-body result is a peer-reviewed anchor for two ghoju lines:
   - **WaterStorage (information-as-hydrology):** Swirl's `WaterStorage/` is literally "memory/storage as a property of the medium." This paper is the actual hydrodynamics of memory in a fluid suspension — the physics of that metaphor. Cite it to ground the analogy in real colloid/stat-mech.
   - **Akashic "state lives in the field":** the punchline (pairwise couplings → no memory; many-body medium → memory) supports the framing that cross-session state is carried by the structured medium (the ledger/field), not by isolated nodes.

3. **Power-law, not exponential, persistence.** The `t^(−3/2)` tail is a general consequence of long-range coupling through a continuum. It is the mathematical bridge to the resonance frame's interest in **non-exponential, scale-free persistence** (Schumann persistence, long-memory fields). Same mechanism class: long-range coupling through a medium ⇒ algebraic decay.

4. **Method note (for the course).** The "divergence vanishes in the crude approximation, so the crude approximation silently erases the very effect you're studying" is a clean **falsifiability / model-adequacy** lesson — the kind of move the Brasovan test pattern and the QDM duplicate-image trick (pulse 54) reward. Worth a course example: *the model that's too cheap to compute the medium is too cheap to see the memory.*

---

## Standing questions

- Does the WaterStorage framework's "phase grid" want an explicit **mobility-tensor / coupling-kernel** term so the "memory in the medium" claim has a calculable object behind it?
- Is there a clean way to state the ghoju "many-body → memory, pairwise → no memory" parallel as a *testable* claim about agent-state persistence (i.e., a metric that vanishes under pairwise-only context and appears under full-field context)?
- The `t^(−3/2)` hydrodynamic tail vs. the Schumann/QB-life persistence curves — is the algebraic-decay match a real shared mechanism or coincidental shared functional form? (Open — don't overclaim.)