shuffling-contradiction-resolution

IN derived (depth 1)

Created 2026-08-25T03:50:14+00:00 · Reviewed 2026-08-25T04:28:09+00:00

The apparent contradiction between the 2025 empirical control (shuffling unembeddings *destroys* the full orthogonality structure) and the set-inclusion finding (shuffled unembeddings *reproduce* the child-parent⊥parent sub-orthogonality) is resolved by scope: the *full* four-condition Theorem 8 structure is genuine and shuffle-destroyed, while a *specific* sub-condition (consecutive-level parent⊥child-parent) is a set-inclusion artifact—meaning the geometric framework captures the genuine full structure, not the spurious sub-structure.

Summary

The scrambled-embedding tests confirm that the full geometric structure of hierarchical concepts is genuinely encoded in the embeddings, while the one weaker sub-relationship that survives shuffling is just a mathematical byproduct of how the vector sets overlap, not evidence of real meaning. This matters because it shows the framework correctly separates the genuine conceptual geometry from spurious artifacts, so the orthogonality results stand as meaningful evidence about how subordinate concepts are actually organized rather than an illusion created by the math.

Justifications

SL — Both base beliefs are IN and describe shuffle experiments, but they operate at different scopes: the control destroys the *full* orthogonality (all four conditions a–d of Theorem 8), while the set-inclusion result reproduces only the *partial* (d) condition. The emergent insight is that the geometric framework's strength lies in capturing conditions (a)–(c) that are shuffle-destroyed and therefore genuinely semantic, while acknowledging that condition (d) is partially geometric.

Antecedents (all must be IN):

  • IN park-2025-empirical-controls-shuffle-random — Three empirical controls rule out artifacts: (1) shuffling unembeddings destroys the structure, (2) random parent selection yields non-zero cosine (ruling out high-dim geometry), (3) independent 70% train splits break set-inclusion orthogonality only in shuffled embeddings.
  • IN park-2025-set-inclusion-spurious-orthogonality — Shuffled unembeddings reproduce orthogonality between child-parent and parent vectors purely from set inclusion (Y(z) ⊆ Y(w)), not semantic meaning, demonstrating that set inclusion alone is insufficient to explain the geometric structure.
  • IN park-2025-hierarchical-orthogonality-theorem-8 — Theorem 8 states that for subordinate concepts z ≺ w, the parent vector is orthogonal to the child-minus-parent residual: ℓ̄_w ⊥ (ℓ̄_z − ℓ̄_w), and sibling contrast vectors are mutually orthogonal.