Follow these steps:
1. Identify the claim chain
List each distinct claim in the reasoning sequence, from premises through intermediate steps to conclusion.
2. Ask what supports each claim
For each claim, identify its warrant:
- Another claim in the chain
- An observation or measurement
- A definition
- A logical axiom
- Nothing visible
3. Follow chains to bedrock
Trace backward until reaching an endpoint:
- Direct observation: "I measured X"
- Definitional truth: "by definition of the term"
- Logical axiom: "law of non-contradiction"
- Absence: no warrant found
4. Mark floating claims
Flag any claim whose chain terminates in absence rather than bedrock — unsupported assumptions.
5. Assess chain integrity
- Are bedrock warrants appropriate for the claims they support?
- Are floating assumptions plausible?
- Does the overall structure hold?