Abstract beliefs earn their keep by cashing out in sensory predictions

It's tempting to say only beliefs that directly anticipate a sensory experience count as real beliefs. But this cuts too much. To predict on which clock tick a dropped bowling ball hits the ground, you need propositional beliefs: "gravity is 9.8 m/s²," "this building is 120 meters tall." Neither of those is itself a sensory experience.

The test isn't whether a belief is directly a sensory anticipation. It's whether the belief chains to one. Gravity isn't visible, but "the second hand is on the 1 when I hear the crash" is. The abstract belief earns its place by eventually connecting to something you can check.

Beliefs one or more steps removed from experience are fine — as long as the chain connects somewhere.

Source claim: Propositional beliefs that aren't directly sensory experiences are legitimate when their inferential chain leads to specific, checkable sensory predictions.