Critical Reasoning Framework — Cheat Sheet
Master Execution Order
- Conclusion (5-step Derivation Process)
- Premises (Empirical, Logical, Normative, Concession, Prescriptive)
- Assumptions — Good / True / Happen
- Gap Test — applied to ALL assumptions
- Weakening — assumption-based + paragraph-by-paragraph
- Vulnerability Ranking — ALL assumptions, most → least
- Failure Modes — 6 common patterns
- Exam-Ready Answer
Step 1 — Conclusion (5-Step Derivation)
| Step | Action |
|---|---|
| 1 | List ALL candidate statements |
| 2 | Scan for linguistic cues: therefore, the solution is, should, must |
| 3 | Apply “Remove and Collapse” test — if removing a statement collapses the argument, it’s the conclusion |
| 4 | Distinguish diagnostic (“what’s the problem?”) from prescriptive (“what should be done?”) |
| 5 | Eliminate false candidates: premises, sub-conclusions, concessions, background context |
Pitfalls: confusing premises with conclusions, mistaking evidence for thesis, selecting emotionally strong statements, choosing background information.
Step 2 — Premises
Premises support the conclusion. Types:
| Type | Description | Example Signal |
|---|---|---|
| Empirical | Observable fact or data | “is,” “has been,” “shows” |
| Causal | Claim about cause-effect | “causes,” “leads to,” “creates” |
| Normative | Value-based claim | “should,” “ought” |
| Concession | Admitted counterpoint | “This does not mean,” “Of course” |
| Prescriptive | Recommendation | “must,” “have a responsibility” |
Key distinction: Premises are what the author presents as fact. Assumptions are what the author leaves unstated.
Step 3 — Assumptions: Good / True / Happen
🔵 GOOD (Value Assumptions)
What is assumed to be desirable or worth pursuing?
- Hidden values, moral priorities, standards of success
- Signal: the argument prescribes action → there’s always a value assumption
- Pitfall: treating values as objective facts; assuming shared moral standards
🟢 TRUE (Definitional / Factual Assumptions)
What is assumed to be factually correct?
- Hidden definitions, classifications, representational assumptions
- Signal: the argument uses loaded terms (“misuse,” “healthy”)
- Pitfall: accepting definitions automatically; confusing metrics with reality
🔴 HAPPEN (Causal Assumptions)
What chain of events is assumed?
- Causal mechanisms, predicted consequences, outcome assumptions
- Signal: the argument connects an action to a result
- Pitfall: confusing correlation with causation; inevitability claims; ignoring alternatives
Step 4 — Gap Test (Applied to ALL Assumptions)
The 4-Step Process
| Step | Action |
|---|---|
| 1 | Identify connection — Which premise does this assumption connect to which part of the conclusion? |
| 2 | State the bridge — “For [premise] to support [conclusion], it must be true that [assumption].” |
| 3 | Deny the assumption — If the assumption were FALSE, would the premise still support the conclusion? |
| 4 | Rate the gap — Critical (argument collapses), Significant (weakens substantially), Minor (survives) |
Step 5 — Weakening
Goal: Reduce confidence in the conclusion (NOT prove it false).
Assumption-Based Methods
| Method | Core Idea |
|---|---|
| Alternative Explanation | The observed effect may have another cause |
| Cause Is Not Necessary | The effect can happen without the proposed cause |
| Implementation Failure | The solution may fail operationally |
| Scaling Failure | Success at small scale may not generalize |
| Unintended Consequences | The intervention may cause new harms |
| Countervailing Forces | Other factors may offset the expected benefit |
Paragraph-by-Paragraph Method
- Extract the implicit claim from each paragraph
- Challenge that specific claim using the toolkit above
- Cumulative effect: reduced confidence in the overall conclusion
Step 6 — Vulnerability Ranking
Rank ALL assumptions using three criteria:
| Criterion | Question | Weight |
|---|---|---|
| Contestability | How easy to challenge with plausible alternatives? | High |
| Counterexamples | How readily available are real-world contradictions? | High |
| Centrality | If this fails, how much of the argument collapses? | Highest |
Pattern: Causal (HAPPEN) assumptions are typically most vulnerable → Definitional (TRUE) in the middle → Value (GOOD) most resilient.
Step 7 — Failure Modes
| Failure Mode | Signal | Diagnostic Question |
|---|---|---|
| Correlation ≠ Causation | “Because A happens with B, A causes B” | Could the direction reverse? Is there a third factor? |
| Hidden Definition Shift | Loaded terms like “misuse,” “healthy,” “efficiency” | How is the author defining this term? |
| Overgeneralization | “Many” → “all”; one pilot → nationwide | Is the sample representative? |
| False Dichotomy | “Either X or Y” | Are there other options? |
| Normative Leap | Facts → “should” without moral bridge | What value assumption bridges the gap? |
| Inevitability Claims | “Will inevitably,” “guarantees,” “eliminates” | What could prevent the predicted outcome? |
Final Memory Formula
Derive the conclusion (5-step), identify premises, uncover ALL assumptions (Good/True/Happen), apply the Gap Test to every assumption, weaken using both assumption-targeting and paragraph-by-paragraph methods, rank all assumptions by vulnerability (contestability × counterexamples × centrality), and scan for reasoning failures.