Skip to content

Epistemology and scientific reasoning

Curriculum status

These entries are architecture specifications, not final lesson plans. They define teaching intent, logistics, outputs and scientific boundaries; release-ready settings, worksheets, notebooks and answer keys still require empirical validation.

Return to the module catalogue

Family source cluster: [T1 Ch.1; O1 Model Types; O7; R4]

EPI-01 — Observation, measurement and inference

Concept / theme Displayed variables differ in how directly they are measured, computed or inferred.
Audience / context Advanced school to doctoral. Formal practical, workshop or modular course. Prior modules: FND-02, DAT-02.
Logistics 45 min; 4–12; boards: 0–1; software: None; optional shared annotation board; prepared dataset: Case traces or claims. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Conceptual seminar. Stages: 1–3, 10–12.
Spikeling relationship 7 — Conceptual analogy; 5 — Recorded-dataset analysis

Learning outcomes.

  • Identify and predict the principal behaviour described in displayed variables differ in how directly they are measured, computed or inferred.

  • Configure or document label a multi-panel trace as command, measured electronic signal, model state or inferred feature and record data provenance and transformation steps.

  • Measure, calculate or compare label a multi-panel trace as command, measured electronic signal, model state or inferred feature using an explicit operational rule.

  • Interpret the result and state why a clean plot does not collapse distinctions between direct and indirect variables

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Label a multi-panel trace as command, measured electronic signal, model state or inferred feature.

4. Acquire or inspect data provenance and transformation steps.

5. Produce evidence-layer annotation.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: A clean plot does not collapse distinctions between direct and indirect variables.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate label a multi-panel trace as command, measured electronic signal, model state or inferred feature. The reusable output is evidence-layer annotation.

Sources / connections / priority. Sources: [T1 Ch.1; O1 Model Types; O7; R4]. Natural follow-ons: EPI-02, IMG-05, EXT-05. Development priority: Core module. Boundary: A clean plot does not collapse distinctions between direct and indirect variables.

EPI-02 — Model validity, analogy and underdetermination

Concept / theme Different mechanisms or parameter sets can generate similar observations, and useful analogies have explicit domains.
Audience / context Advanced undergraduate to doctoral. Formal practical, workshop or modular course. Prior modules: FND-02, CMP-01.
Logistics 60 min; 4–12; boards: 0–1; software: None; optional shared annotation board; prepared dataset: Case traces or claims. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Epistemological seminar. Stages: 1–3, 10–12.
Spikeling relationship 7 — Conceptual analogy; 8 — External computational comparison

Learning outcomes.

  • Identify and predict the principal behaviour described in different mechanisms or parameter sets can generate similar observations, and useful analogies have explicit domains.

  • Configure or document compare two models/interpretations that fit the same trace and specify discriminating evidence and record observed features, candidate explanations, test predictions.

  • Measure, calculate or compare compare two models/interpretations that fit the same trace and specify discriminating evidence using an explicit operational rule.

  • Interpret the result and state why spikeling’s phenomenological success cannot establish biological mechanism

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Compare two models/interpretations that fit the same trace and specify discriminating evidence.

4. Acquire or inspect observed features, candidate explanations, test predictions.

5. Produce validity/analogy matrix and proposed discriminating test.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Spikeling’s phenomenological success cannot establish biological mechanism.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate compare two models/interpretations that fit the same trace and specify discriminating evidence. The reusable output is validity/analogy matrix and proposed discriminating test.

Sources / connections / priority. Sources: [T1 Ch.1; O1 Model Types; O7; R4]. Natural follow-ons: CMP-09, EPI-03. Development priority: Core module. Boundary: Spikeling’s phenomenological success cannot establish biological mechanism.

EPI-03 — Causality, falsification and negative results

Concept / theme Causal claims require interventions, controls and alternatives; null or negative outcomes can constrain explanations.
Audience / context Master’s/doctoral methods. Formal practical, workshop or modular course. Prior modules: MET-03, STA-02.
Logistics 60 min; 4–12; boards: 0–1; software: None; optional shared annotation board; prepared dataset: Case traces or claims. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Conceptual/methodology seminar. Stages: 1–3, 10–12.
Spikeling relationship 7 — Conceptual analogy; 5 — Recorded-dataset analysis

Learning outcomes.

  • Identify and predict the principal behaviour described in causal claims require interventions, controls and alternatives; null or negative outcomes can constrain explanations.

  • Configure or document evaluate claims from a network or sensory dataset and redesign a falsifying test and record claim, intervention, control, alternative outcome.

  • Measure, calculate or compare evaluate claims from a network or sensory dataset and redesign a falsifying test using an explicit operational rule.

  • Interpret the result and state why statistical association and temporal precedence alone do not establish causality

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Evaluate claims from a network or sensory dataset and redesign a falsifying test.

4. Acquire or inspect claim, intervention, control, alternative outcome.

5. Produce claim-evidence table and falsification protocol sketch.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Statistical association and temporal precedence alone do not establish causality.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate evaluate claims from a network or sensory dataset and redesign a falsifying test. The reusable output is claim-evidence table and falsification protocol sketch.

Sources / connections / priority. Sources: [T1 Ch.1; O1 Model Types; O7; R4]. Natural follow-ons: EPI-04. Development priority: Advanced specialised module. Boundary: Statistical association and temporal precedence alone do not establish causality.

EPI-04 — Open hardware, transparency and responsible communication

Concept / theme Transparent instruments and code enable inspection, critique and reuse but do not remove the need for validation.
Audience / context Undergraduate to educator. Formal practical, workshop or modular course. Prior modules: MET-06, EPI-01.
Logistics 45 min; 4–12; boards: 0–1; software: None; optional shared annotation board; prepared dataset: Case traces or claims. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Seminar/workshop. Stages: 1–3, 10–12.
Spikeling relationship 7 — Conceptual analogy; 5 — Recorded-dataset analysis

Learning outcomes.

  • Identify and predict the principal behaviour described in transparent instruments and code enable inspection, critique and reuse but do not remove the need for validation.

  • Configure or document audit a result’s hardware/software/data provenance and rewrite an overclaim and record version, parameters, files, claim language.

  • Measure, calculate or compare audit a result’s hardware/software/data provenance and rewrite an overclaim using an explicit operational rule.

  • Interpret the result and state why open-source status is not itself evidence of accuracy, safety or biological validity

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Audit a result’s hardware/software/data provenance and rewrite an overclaim.

4. Acquire or inspect version, parameters, files, claim language.

5. Produce open-method checklist and responsible summary.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Open-source status is not itself evidence of accuracy, safety or biological validity.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate audit a result’s hardware/software/data provenance and rewrite an overclaim. The reusable output is open-method checklist and responsible summary.

Sources / connections / priority. Sources: [T1 Ch.1; O1 Model Types; O7; R4]. Natural follow-ons: OSC-02. Development priority: Core module. Boundary: Open-source status is not itself evidence of accuracy, safety or biological validity.