It's very similar to being untestable. If a hypothesis provides a prediction, but can alter the interpretation after the event to fit the facts, then it's providing no predictive power. e.g. If someone predicts that <insert deity> rewards/punishes people when they die (which alters their near death experience), but then amends their interpretation of good/bad to fit the data, then it can't be used to predict anything.
> If a hypothesis provides a prediction, but can alter the interpretation after the event to fit the facts, then it's providing no predictive power. e.g.
Is this an accurate and comprehensive description of the situation we have before us though?
> If someone predicts that <insert deity> rewards/punishes people when they die (which alters their near death experience), but then amends their interpretation of good/bad to fit the data, then it can't be used to predict anything.
That you can identify one attribute of this nature is no guarantee that all attributes (which is not known) are of this kind.
