Zero is Most Likely: Skepticism, Rational Priors, and the Fine-Tuning Argument

Rational skepticism requires that, absent evidence, we assign most probability to nonexistence. I shouldn't believe in telekinesis without evidence; Neanderthals weren't obligated to believe in germs or gravity. In Bayesian terms, this means prior probability distributions should peak at zero and decay monotonically with distance from zero, what I call "zero-peaked priors." This framework has profound implications for the fine-tuning argument. It shows that parameters whose life-permitting range includes zero (like the cosmological constant) provide little evidence for design. But it also defeats the two strongest objections to fine-tuning: the normalization objection and the multiverse objection. The normalization objection assumes improper priors (uniform or log-uniform distributions). Zero-peaked priors are proper by construction, dissolving this objection. The multiverse objection requires both many universes and high variance between them—but zero-peaked priors make both improbable. Moreover, I show that if life-bearing universes were randomly sampled from a varied multiverse, most would contain many weak forces alongside their few strong ones.