Claims › Concept · How a verdict is made

Convergent paths

Is agreement between different derivations evidence about the answer?

Several different derivations arriving at the same answer is evidence about the answer, not noise to be averaged away.

Where it sits

1 Free tag Mayo a name finds, it does not vouch
2 Topic Annotation · Severity · Architecture closed list, one per claim
3 Theme How a verdict is made derived from the topics
4 Concept Convergent paths · 4 claims a note and links, or no build
5 Geometry 1 1 1 0.75 source, use, dissent

How its claims lean on each other

1 · Free tag2 · Topic3 · Theme4 · Concept5 · Geometryweight, 0.5 → 2The claimsMayoMayoAnnotationAnnotationSeveritySeverityArchitectureArchitectureHow a verdict is madeHow a verdict is madeWhat scale does not buyWhat scale does not buyCoder agreementCoder agreementSeveritySeverityConvergent pathsConvergent pathsAlso: Measurement artefactAlso: Convergent pathsAlso: Convergent pathsAlso: Convergent pathsC020C020 · Thirty differently worded answers to the same prompt are not thirty disagreements: they can be one conclusion reached thirty ways.C014C014 · Convergent, independent checks can warrant a conclusion more strongly than any single one of them.C042C042 · Two selections made by different methods err in different ways, so their agreement is evidence about the truth.C025C025 · When several different chains of reasoning arrive at the same answer, that agreement is itself evidence for the answer.C042 confirms C014

The claims

  1. C020

    “Surface-level variance […] masks high semantic convergence.”

    Thirty differently worded answers to the same prompt are not thirty disagreements: they can be one conclusion reached thirty ways. Measuring the wording instead of the conclusion makes a model look far less stable than it is.

    Vertical Validation ¶11 · ¶11 · assumed · weight 1 · uses-method-in C019 · open C020

  2. C014

    “An overall inference can be more reliable and precise than its premises individually.”

    Convergent, independent checks can warrant a conclusion more strongly than any single one of them. This is the epistemic ground for aggregating multiple fallible runs or judges instead of demanding a single infallible one.

    Mayo 2018 · p. 15, Excursion 1, Tour I, §1.2 · used · weight 1 · open C014

  3. C042

    “Nor must we lose sight of the constant tendency of the inductive process to correct itself. This is of its essence. This is the marvel of it. ...even though doubts may be entertained whether one selection of instances is a random one, yet a different selection, made by a different method, will be likely to vary from the normal in a different way, and if the ratios derived from such different selections are nearly equal, they may be presumed to be near the truth.”

    Two selections made by different methods err in different ways, so their agreement is evidence about the truth. It is the convergence argument the programme uses when independent runs or coders reach the same verdict.

    Peirce 1931–1958 · CP 2.729 · quoted in Mayo 2005, p. 315 · assumed · weight 1 · confirms C014 · open C042

  4. C025

    “Self-consistency leverages the intuition that complex reasoning tasks typically admit multiple reasoning paths that reach a correct answer […] .”

    When several different chains of reasoning arrive at the same answer, that agreement is itself evidence for the answer. The method samples many derivations and keeps the one they converge on.

    Wang et al. 2022 · §1 Introduction · used · weight 0.75 · open C025

Reading cards on this concept

All 15 concepts →All the claims, one by one →