nucleation

TERM-033: fidelity, as observational equivalence over response trajectories

1 min read

Properties4
StatusActive — by TERM-015 — fidelity, as graded, directed transport distortion
Tagsphilosophy-of-language, probabilistic-modeling
Linepragmatic-transport
Used byCLAIM-046 — Different hidden-state models can produce identical observations, so pragmatic identity and fidelity should be defined on observable responses before any equivalence of internal states is assumed, CLAIM-051 — Fidelity is always fidelity with respect to some selected structure: a specified family of communicative observations, interventions and permitted continuations, CLAIM-083 — Pragmatic translation fidelity is the preservation, under corresponding communicative conditions, of the ways an utterance transforms interpretation, interpersonal alignment and the possibilities for subsequent interaction, CLAIM-065 — Pragmatic fidelity is approximate preservation of the ways an utterance conditions, transforms and constrains later interpretation and interaction, across transient framing and learned interpretive dispositions, CLAIM-067 — If a transport exactly intertwines a generating set of communicative operations, it preserves every trajectory composed from them

Graph View

Definition

A35 §8 weakened intertwining to “observational simulation”, O_t(U_t(Φ(s))) ≈ O_o(U_o(s)), because “Exact dynamical equivalence may be too strong for translation. The original and target audiences need not possess corresponding internal states at every intermediate point.” A39 §5.3 stated it over sequences: P_o(Y_1:k | a_1:k, c_o) ≈ P_t(Y’_1:k | a’_1:k, c_t), “preservation of the distribution of possible response trajectories under corresponding action sequences”, which “does not require a shared latent state space”. Its functional L_obs(τ) sums a distance over a finite family of intervention sequences.

What it is not

Not TERM-007, which compares internal states through Φ; A39 demoted that to an optional refinement. Not TERM-021 either: A24 compared single judgements across contexts, this compares trajectories under interventions (TERM-018). The manuscript went back to the context form for L_obs (§6) and kept intertwining in §9; trajectories appear only in §8 (“how selected invariants … evolve together”).

Why it is still Active

TERM-015 replaced equivalence as the organizing concept, at the owner’s U20. Matching over response trajectories survives as the zero-distortion limit of TERM-015 and of TERM-043 (A129 §3: “TERM-033 supplies the limiting case”).

Backlinks