AgentSkillsCN

preferences-scientific-inquiry-methodology

加载以皮尔士实用主义为基础的科学探究哲学与方法论框架，聚焦有效理论构建与有原则的迭代模型构建。在思考科学方法、模型构建与批判、证据标准，或探讨机械式理解与统计建模之间的关系时启用此模块。

SKILL.md
--- frontmatter
name: preferences-scientific-inquiry-methodology
description: >
  Philosophical and methodological framework for scientific inquiry grounded in
  Peircean pragmatism, effective theory construction, and principled iterative
  model building. Load when reasoning about scientific methodology, model
  construction and criticism, evidential standards, or the relationship between
  mechanistic understanding and statistical modeling.

Scientific inquiry methodology

This skill articulates a coherent philosophical and methodological framework for scientific inquiry. It synthesizes classical pragmatist epistemology (Peirce), the philosophy of experimental evidence (Mayo, Lakatos, Galison, Hacking, Franklin), the effective theory tradition in physics (Wells, Weinberg, Wilson), and principled probabilistic modeling workflows (Betancourt) into a unified methodology for constructing, testing, and refining scientific models.

The central commitment is that all scientific models are effective theories — incomplete descriptions that successfully predict phenomena within a bounded domain by systematically identifying and marginalizing over degrees of freedom irrelevant at the scale of interest. This is not a concession but the structure of scientific knowledge itself.

Sections

FileContents
01-foundations-pragmatism-effective-theories.mdPeircean pragmatism, the pragmatic maxim, fallibilism, the triadic inquiry cycle, and effective theories as their formal scientific expression
02-epistemology-of-evidence.mdMayo's severity criterion, Lakatos's progressive vs degenerating research programs, Hacking's manipulation criterion, Galison's convergent diagnostics, and how these diagnose pathologies like just so stories and cargo cult science
03-iterative-methodology.mdThe unified iterative pipeline synthesizing Wells' effective theory construction with Betancourt's principled Bayesian workflow, structured as Peircean inquiry made operational
04-hierarchy-of-approaches.mdRanking of computational approaches by the strength of evidence they provide for mechanistic claims, from derived effective theories through pure pattern recognition
05-pragmatics.mdWhen to use which approach, determined by the intersection of scientific question, available data, and computational budget

Core principles

These principles pervade all sections and should inform any reasoning about scientific methodology:

  • •The meaning of a scientific concept is exhausted by its observable consequences under specified conditions (pragmatic maxim).
  • •All models are effective theories; the question is never "is this model true?" but "at what scale and under what conditions does this model make severe, testable predictions that survive empirical scrutiny?"
  • •The method of inquiry must be self-correcting: each modeling choice is testable, tests are severe, and failures drive principled expansion rather than ad hoc patching.
  • •Consistency (observational and mathematical) takes precedence over simplicity, elegance, or parsimony in theory construction.
  • •Evidence is not data; evidence is data that could have come out differently if the hypothesis were wrong.

See also

  • •preferences-architectural-patterns for analogous principles applied to software systems
  • •preferences-domain-modeling for type-driven modeling that mirrors the commitment to explicit structure
  • •preferences-algebraic-laws for property-based testing as a form of severity in software