nucleation

QUESTION-218: Can a collective mode that is dynamically closed, but is not a weakly coupled block of the units carrying it, be one system, and if so what should a boundary test take as the unit: a subset of members or a function of the whole state?

4 min read

Properties6
StatusOpen
Version2
Tagsindividuation, complex-systems, mathematics
Linedistributed-agency
RefinesQUESTION-035 — Relative to a stated target, when is a combined form one system, where does its boundary lie, and how long does it last?
BearingTHEORY-822 — For a Markov chain and a partition of its states, strong lumpability, every state of a block sending the same probability into each block, is the column-equitability of the master-equation operator and makes the block label informationally closed; when the chain is irreducible, an informationally closed label that is itself Markov is strongly lumpable, with no aperiodicity needed; weak lumpability gives a Markov label that need not be closed; and the same holds on the causal-state chain of any stationary process, THEORY-820 — For a stationary discrete-valued process X and any coarse-graining Z of its past, informational closure at every horizon is the same condition as zero one-step transfer entropy from the micro past, Barnett and Seth's dynamical independence, and as causal closure, Z's ε-machine coinciding with its υ-machine; for a coarse-graining of the present state it also makes Z's causal states a coarse-graining of X's, which does not imply closure, and the closed coarse-grainings are not closed under common refinement, THEORY-758 — For a self-adjoint diffusion operator T with spectrum in [0, 1], the diffusion-wavelet levels are spectral subspaces spanned by the eigenvectors with λ^s ≥ ε for an increasing sequence of powers s, so they form an exactly nested chain of T-invariant subspaces whose differences are bands of the spectrum, about one octave each, not spatial blocks; Coifman and Maggioni's compressed algorithm, which squares the operator on each orthonormal basis, thresholds T^{2^{j−1}} at level j and not T^{2^j−1} as the paper states; and since each level is T-invariant, each is an exactly closed linear coarse-graining whose localized basis is a choice the level does not fix, THEORY-755 — For a linear system x(t+1) = Ax and a partition of its variables, the vector of block sums evolves autonomously from every initial state iff the partition is column-equitable, every column of each block of A summing to the same value; the aggregate operator then keeps a subset of A's eigenvalues, and the blocks may interact as strongly as they like, so exact aggregation is not decomposition

Graph View

Why it is a question

CLAIM-157’s unit is already a projector, “the pair’s joint mode, not either member” (CLAIM-164), but it is found by grouping members into a block whose edge is weakly coupled. THEORY-257’s candidates are “linear subspaces mixing all units, not partitions”. A5, after Shpak et al. (LIT-1242, read in NOTE-979): aggregation, a map whose dynamics close, is not decomposition, a split into weakly coupled subsets, and “systems that are both aggregable and decomposable are a special case.”

CLAIM-388 makes the split exact. In CASE-120, with uniform coupling, the sum of all units is closed at every coupling strength, and no partition is nearly decomposable. So a block test finds nothing below the whole and a closure test finds the mode.

What the record holds

  • CLAIM-157, CLAIM-164 and THEORY-661 on the block side.
  • THEORY-257 and THEORY-733 on the closure side.
  • A third kind of unit, from Shpak et al. (THEORY-755, read 2026-10-20): a partition of members whose block sums jointly close, which happens iff the partition is column-equitable, while the blocks interact strongly. In CASE-120 every partition is one. It is a grouping of members, as a block is, and closed, as a mode is, and it is neither.
  • Published cases of the same kind (NOTE-977). On a single-peak fitness landscape, Hamming-distance classes aggregate exactly although their members “did not communicate with one another at all”. Schema marginals close on the linkage-equilibrium manifold although rates between classes are as large as rates within them.
  • Coarse levels of an operator (THEORY-758, diffusion wavelets). Each level is an exactly closed subspace, but its localized basis is a choice the level does not fix. On three tight clusters the localized scaling functions each sat on one cluster. On a circle, with no gap, no localized coarse basis exists. So closure belongs to the subspace, and blockhood to a basis that only a gap supplies, as CLAIM-157 and CLAIM-164 would have it.
  • Rosas et al. (2024, LIT-1306, read in NOTE-1030) take the unit to be a function of the whole state, up to computational equivalence. Informational closure, causal closure and Barnett–Seth dynamical independence coincide (THEORY-820); for a partition of a Markov chain, closure is strong lumpability, the equitable condition of THEORY-755 on Pᵀ (THEORY-822). Two warnings: closure is cheap, since an i.i.d. noise component is closed; and two closed coarse-grainings taken together need not be closed, contrary to the paper’s sublattice claim. The paper offers no independent mark of unity.
  • CLAIM-157’s guard: clearing a test is not being an agent or a self. A5 kept it: “A newly discovered distributed collective mode should not automatically be treated as an individual, much less an agent.”

What would count as an answer

A5’s table gives the four cells: strong block and strong closure, block without closure, closure without block, neither. An answer says which cells the record’s “one system” covers, with an independent mark of unity (CLAIM-157’s compensation under perturbation, or entrainment) measured in a closure-without-block case. If the mode compensates and entrains and the block test misses it, that meets CLAIM-157’s defeated_if (“forms that do compensate and entrain but never clear it”). If it does not, the block is the better unit and closure tracks something else.

Note of 2026-10-21: closure gives no mereology, and is monotone in neither direction

The exchange of 2026-10-20 (A6–A9) built its paper on the two warnings this question already records from THEORY-820. A6 called the second “the biggest discovery”. Two facts were checked from the exact three-bit chain. (A, C) and (B, C) are closed, though (A, B) is not, and A⊕B, a coarsening of (A, B), is not closed (1 bit), so the closed coarse-grainings are neither upward- nor downward-closed in the refinement order. That is filed as CLAIM-409, with the reason: refinement changes the future that closure asks to be predicted. A8 and A9 drew from this that closure “cannot be a sufficient criterion” of a unit or an agent, which is this question’s open half and CLAIM-157’s guard, and offered no independent mark of unity either. TERM-068 separates the senses of “closure” the exchange ran together.

Backlinks