nucleation

CLAIM-128: No single ranking of translations holds across tasks: where distortion profiles trade off, only a task's priorities rank them, and the theory of fidelity fixes no weighting

3 min read

Properties9
StatusProposed
Version2
Rolethesis
Warrantinterpretive
Tagsmodel-comparison, philosophy-of-language
Linepragmatic-transport
Rests onCLAIM-106 — Pragmatic fidelity is a graded, potentially asymmetric property of transport, not an equivalence relation; equivalence is its zero-distortion limit, CLAIM-371 — Under the componentwise order on distortion profiles, two profiles are incomparable exactly when they trade off, and then positive weightings can rank them either way, while every positive weighting agrees with a dominance
ComplementsCLAIM-033 — In published translations, renderings that lose distinctions along one dimension commonly gain specificity along another, so loss is not one-dimensional, CLAIM-240 — A holistic fidelity judgement is fidelity under the judges' default task, a weighting of the profile's components that can be estimated on some renderings and tested on others, so it is well defined and consistent with task-relativity, and a defeat condition that uses it must name the population and fix the weighting before the test, and across populations with unequal access to the source must fit the weighting only on the components each can perceive, CLAIM-347 — Frustration among an organization's demands is an ordinal fact before it is a quantity: the jointly satisfiable sets of demands form a downward-closed complex whose minimal non-faces are the minimal conflicts, relief is strict enlargement of that complex for the same declared demands, reorganizations that relieve different conflicts are incomparable, and dropping or suppressing a demand is not relief, CLAIM-354 — Freedom of will and autonomy come in degrees along several dimensions, among them viable reachability, control cost, robustness to disturbance, governance reconfigurability, revisability and motivational integration, and agents compare by inclusion of what they can viably reach under matched conditions: a partial order with incomparable pairs, not a scalar, and a gradient only of a declared measure
UsesTERM-043 — fidelity, as a profile of directed distortions compared component by component

Graph View

The claim

Fidelity is a profile of directed distortions, compared component by component (TERM-043). One transport is at least as faithful as another when it distorts no component more. When one transport is better on one component and worse on another, neither is at least as faithful as the other. A129 §3: “When the profiles trade off, the transports are incomparable without selecting additional priorities.” Its example: “A translation that preserves social footing but discards rhyme need not be ranked unconditionally above or below one that does the opposite.”

A ranking appears only when a task fixes weights, D_w = Σ_j w_j D_j, and a different task fixes different weights. So no ranking of translations holds across tasks. A129 §3: “Fidelity is a family of directed comparisons whose meaning depends on which structures the communicative task requires.” It connects the vector form to “our earlier rejection of a universal equivalence relation”, which is CLAIM-106: fidelity is graded and directed, and equivalence is its zero-distortion limit. This claim rests on that one, because a profile of graded distortions is what makes trade-offs possible.

The claim recurs. A151 §5: “two translations can preserve different structures without being forced into a universal ranking.” A178 Ch9.7: “Partial orders on fidelity profiles”. A203 §22: “no universal weighting is assumed”. It complements CLAIM-033, which found at A75 that a translation can broaden the readings along one dimension while narrowing them along another: that is one trade-off, in the region picture.

What it does not say

It does not say how the components are measured. Only D_obs has a formula, and D_causal needs a correspondence on interventions that nothing yet fixes (QUESTION-005). It does not say the components are independent. Whether they move together under iteration is QUESTION-013. If they do for every task the manuscript studies, trade-offs never arise there and the claim is defeated for those tasks.

It does not say a scalar is illegitimate. A task-relative weighting is what the claim allows. It denies only that one weighting is fixed by the theory of fidelity itself.

Note of 2026-10-19: split under ADR-043

The order-theoretic part of this claim is now CLAIM-371, with its proof: under the componentwise order, two profiles are incomparable exactly when they trade off, positive weights can then rank them either way, and no positive weighting reverses a dominance. That part is demonstrated and Active. This claim keeps the thesis built on it, that a ranking of translations exists only relative to a task and that the theory of fidelity fixes no weighting, and is interpretive. Its defeat condition was already about the thesis and is unchanged.

Note of 2026-10-30: what holds it back

No fidelity profile has been measured for any compared pair of transports on any task. Only D_obs has a formula (QUESTION-005), and whether components move together is QUESTION-013, so whether trade-offs occur at all, the first branch of the defeat condition, has no observation. The order-theoretic part is CLAIM-371 and is Active.

Backlinks