There Are No Anthropic Paradoxes: Normal Anthropics for Normal People

Anthropic reasoning argues that an observer should, all else equal, reason as if they were randomly selected from some set of observers. The three dominant theories—the Self-Sampling Assumption (SSA), the Strong Self-Sampling Assumption (SSSA), and the Self-Indication Assumption (SIA)—yield radically different verdicts, and each is alleged to produce paradoxes, from observers gaining telekinetic powers to armchair philosophy settling cosmological questions. This paper argues that anthropic reasoning is straightforward. First, I show that the three popular theories differ along three binary dimensions, implying a taxonomy of eight theories rather than three. I argue that one of the five “missing” theories—reasoning as a randomly selected observer-moment from all actual observer-moments, without a reference class—is preferable on theoretical grounds: it avoids appeals to merely possible observers and to arbitrary reference classes. I then show how it produces the thirder answer to Sleeping Beauty without committing to the implausible cosmological claims implied by SIA. I present a novel argument against SIA on grounds that it either violates the Reflection Principle and conservation of expected evidence or the Principal Principle in the Sleeping Beauty Problem. I then review other claimed anthropic paradoxes, such as UN++ and Lazy Adam, and show that none function under conventional Bayesian updating. I conclude by arguing that the Doomsday Argument provides very little or zero information about the timing of human extinction, and generates no genuine paradox.