Inductive Analogy

A type of Argument by Analogy used to make empirical predictions: inferring that something will behave in a new case the way it behaved in a similar past case.

1. Past cases A1, A2, ..., An share properties P1, ..., Pn with case B.
2. In all past cases, property Q held.
∴ 3. Q (probably) holds in B too.

Example

  • This drug worked on mice (similar physiology), so it will probably work on humans
  • The last three economic downturns followed inverted yield curves; the curve just inverted again, so we should expect another downturn
  • Past medications in this class caused liver effects; the new one likely will too

What makes it strong:

  • Base cases are numerous and varied: diversity rules out coincidence
  • Shared properties are causally relevant to Q
  • The target case isn’t disanalogous in some critical respect (mice ≠ humans defeats some drug analogies)

What makes it weak:

  • Sample of one (just one prior case)
  • Shared properties are surface-level, not mechanistically connected to Q
  • Cherry-picked base cases

Inductive analogy predicts what is; Consistency Analogy argues for what ought to be. Same logical structure, different conclusions and standards.