My question takes as an assumption that using sources as authorities is a practical necessity even while acknowledging that appealing to an authority can be considered a logical fallacy in deductive argumentation.

In some contexts it is an accepted practice to cite the claims or works of other people, and in others it is not. I find it a perplexing problem to decide on a method to assess whether to accept a source as an authority because I have not been able to develop a method that is both universally-applying to this problem, and can be demonstrated as generally reliable when tested. I have considered that it may be necessary, or at least more useful, to classify proposed sources and use different assessment methods for different classes to optimize reliability.

Here is an unordered list of heuristic questions I have considered:

  • Do the people pertinent to the source have verifiable credentials?

  • Do the people pertinent to the source have a record of fraudulence?

  • Do other verified authorities, assuming that any can be established, cite the source as an authority?

  • Do the people pertinent to the source have the same view as the consensus view of other verified authorities? [Assuming there is a consensus and that there are other verified authorities]

I feel that this list is not well-formed, and incomplete, and I would appreciate input from other skeptics on how to approach the issue of assessing whether to treat sources as authorities.

1 Answer 1


Sources are not necessarily used as "authorities": for example in many cases sources report verifiable facts which we can check independently, e.g. most systematic reviews contain the "recipe" necessary to reproduce their results (or a large part of it); even peer-reviewed articles or blog posts might contain enough to be able to independently verify their conclusions. Therefore, the best heuristic in my opinion is:

  • Is the source explaining how they got those results?
  • Can I reproduce those results?
  • Have other sources reproduced those results?

When they are, there are a few more heuristics you can use:

  • Are the assertions the reference makes "extraordinary"? The "amount" of proof necessary depends on the claim: while someone might have really invented a "cure for cancer", extremely strong evidence of the fact is needed to be convincing. At the other end of the spectrum, nothing more than wikipedia is needed to support trivial facts.

  • Is the source appropriately quoted and referenced with the correct level of importance? There are many papers, for example, that find traces of an effect and make very weak claims about seeing it. Sometimes people misinterpret the conclusions, so we might get an answer that claims stronger certainty than the paper.

More in general systematic reviews contain many such heuristics by which they "weigh" the papers the are consolidating, maybe you can find more there.

You must log in to answer this question.