A "slippery slope" argument that something is dangerous is, imho, similar to use a corellation to argue that there is a direct causal link between two observati

A "slippery slope" argument that something is dangerous is, imho, similar to use a corellation to argue that there is a direct causal link between two observations. Neither are sufficient to prove such claims. But, both _may_, depending on other information, be evidence supporting the claims, even if they do not prove the claim outright. In other words, slippery slope areguments (like correlation arguments) are fallacies if they are claimed as definitive proofs for a claim being made, but are not fallacies if used as supporting evidence for a claim or if arguing that the claim should be considered a possibility that may need to be considered or investigated. In bayesian reasoning, both can be used as evidence for some claim, that creates a new set of (posterior) probabilities for a set of mutually exclusive claims based on a set of prior probabilities. For instance, lets say you are concerned that some president may end up as a dictator. In a democracy, the prior probability may be relatively small. Then, lets assume the country abolishes term grants more power to the president in some time of emergency. In that case, one could argue using a sliding slope argument that the risk that the posterior probability of the president ending up as a dictator had increased after being granted more powers.