Inductive Reasoning
Inductive reasoning generalises from observed cases to unobserved cases, or from evidence to a probable explanation. This domain tests inductive strength, cogency, enumeration, analogy, causal inference and the limits of induction.
Core rules
- Induction is ampliative: the conclusion goes beyond the premises, so it is never deductively guaranteed.
- An inductive argument is strong if true premises make the conclusion probable; it is cogent if it is strong and all premises are true.
- Enumerative induction generalises from a sample. Its strength depends on sample size, representativeness and freedom from bias.
- Causal induction uses Mill's methods; statistical induction uses frequency data. Inductive conclusions are always revisable.
Worked example
Every observed crow in a survey is black; therefore all crows are black. This is enumerative induction. The premise is true, and the conclusion is probable if the sample is large and representative, so the argument is strong; if the survey is biased, it is weak. Even a strong version is not deductively valid: one non-black crow would falsify the universal conclusion.
Traps
Do not call a strong inductive argument valid. Do not generalise from a small or biased sample: that is hasty generalisation. Do not confuse absence of evidence with evidence of absence. Correlation does not establish causation without further controls.