AIM Framework glossary
Evidence Probability
Canonical definition
Evidence Probability, denoted P(E), is the marginal probability of observing a piece of Evidence across all Hypotheses in the Hypothesis Space, , functioning as the normalising denominator in Bayes' theorem. In the motivational special case, Evidence Probability is the overall probability of observing a particular piece of evidence regardless of which Source Hypothesis is true, representing how likely the Individual would be to experience a particular want under any circumstances, whether it arose from Appetites, Intrinsic Motivation, or Mimetic Desire.
Explanation
Evidence Probability captures how common or surprising a particular experience of wanting is. If you find yourself wanting coffee every morning, that evidence is highly probable regardless of source, it could be Appetitive (caffeine dependence), Intrinsic (enjoyment of the ritual), or Mimetic (workplace norm). High Evidence Probability means the want is unsurprising and therefore less informative about its source. Low Evidence Probability, an unusual or unexpected want, provides stronger signal about which source is likely operating. In Bayes' theorem, Evidence Probability functions as a normalising factor that ensures updated Beliefs remain valid probabilities.
Epistemic status
Theoretical Construct (Latent Variable).