|When, for every new question in a particular domain, the inductive evidence from relevant past experience is ignored.||Wake up, Dear! Something’s banging on the trash cans again. I know that for the hundreds of other times you’ve checked, it’s only been cats, but could you please go out and make sure it isn’t zombies this time?|
|This notion is closely related to the concept of priors when assessing Bayesian probabilities. We seldom need to start with equal probabilities for 2 possible explanations of a given phenomenon. For example, both mice and fairies are logically possible, but when we find food items nibbled on in our cupboards, we need not give mice and fairies equal explanatory standing before we begin to explore the evidence since we have a much longer history of confirmations that mice were meddling with cupboard items.|
Case Study One
After thousands of investigations into claims of ghost hauntings, researchers have discovered many material causes, but no ghosts. Therefore, the next time someone claims to have seen a ghost, we are warranted in defaulting to the expectation of a material explanation.
Case Study Two
For each new alleged miracle, many theists demand skeptics to demonstrate it was not a miracle, supposing that, until such negative proof is provided, they are warranted in believing it was an actual miracle. This ignores centuries of claims of miracles being shown to either fall well within the parameters of natural phenomena and statistics, or being too inaccessible to the scrutiny of science. This track record not only permits, but demands a default to a methodological naturalism in which we expect a material explanation, and in which the burden of proof rests entirely on those proposing something contrary to this successful track-record of material causes.
Keep in mind that a fallacious argument does not entail an erroneous position.