Belief Under Evidence
Learning is the reduction of entropy in belief
Shannon Entropy
H = 4.000 bits
Bits Learned
0.000
Last Surprise
--
Observations
0
Prior (before click)
Posterior (current belief)
Evidence likelihood
In Bayesian Learning, every click is an observation and belief sharpens toward where the evidence points, entropy falling a little at a time. In Dogma, a single click collapses belief onto the same fixed hypothesis (H12) no matter where you clicked, dropping entropy to 0.120 bits in one step.
Ignorance is maximum entropy: all states equally probable, nothing predicted. Each observation reduces entropy, sharpening the distribution toward truth. Knowledge compresses possibility.
belief / attractorevidence / informationdogma / collapse