Appeal to Authority

Is, of course, only a fallacy when the person appealed to is not really an authority on the topic in question, e.g., "Even the Pope says Python is a great programming language!"

The other day I saw someone defending this sort of appeal by saying it is a form of probabilistic reasoning: "If the experts agree that X is Y, then X probably is Y."

Someone who refused to go along replied, "History argues strongly against that principle: see geocentrism, phlogiston, and ether."

Think seriously about this for a moment: it is wrong to say that the experts are usually correct, because this fellow can come up with, over all of human history, three times when the experts were wrong!

No doubt we can come up with more, many more. But for every such example, I'd bet we can come up with a thousand where the experts have been found correct. The problem, in fact, is that the experts are so overwhelmingly correct that we really only notice the occasional situations in which they are wrong.

UPDATE: As Tom points out in the comments, it is also a fallacious appeal to authority to use an authoritative opinion as a premise in a deductive syllogism:

P: Brian Kerighan says Python is a great programming language, therefore...
C: Python is a great programming language.

I did not mention this case because I almost never encounter it: the appeals to authority I see on the Internet are usually done in a marshaling of additional evidence for the truth of some point.

Comments

  1. "[Appeal to authority i]s, of course, only a fallacy when the person appealed to is not really an authority on the topic in question"

    Er, no. Appeal to authority for "X is Y" is a fallacy any time it is treated as a proof that "X is Y."

    But that doesn't conflict at all with "[i]f the experts agree that X is Y, then X probably is Y."

    Why?

    Because there's a difference between the standards for formal logic and the standards for broad general statements of common sense.

    If I want to know some fact about Oakeshott, there's a good chance I'll consult an expert on Oakeshott -- one Gene Callahan, perhaps -- and take his answer as dispositive, or at least likely correct.

    If I want to logically prove some factual claim about Oakeshott, "Gene Callahan says" will not be sufficient.

    ReplyDelete
    Replies
    1. 'Appeal to authority for "X is Y" is a fallacy any time it is treated as a proof that "X is Y."'

      Well, I was taking it for granted that looking for proof at all in any sort of empirical argument is an error! And that is where appeals to authority are usually introduced.

      Delete
    2. Yeah, that's what irks me: virtually no one offers up the authority as some ironclad, deductive guarantee of the conclusion being correct, and yet people respond as if they did exactly that, as if they only have to find one case where that expert (or expertise in general) were wrong.

      It would actually be a rather strange world if experts *weren't* more likely to be right than wrong (and thus for their opinion to lack inferential value): that would mean that studying a topic (or doing other things that move you in the direction of expert) does nothing to your chance of being right about it! It would correspond to a structureless, max-entropy world.

      Delete
  2. Are appeals to popularity also valid as a form of probabilistic reasoning? Inserting into your argument: "If the majority agree that X is Y, then X probably is Y."

    ReplyDelete
    Replies
    1. It depends on the question doesn't it? On certain things, the opinion of the majority is very good: guessing the number of jellybeans in a jar.

      Delete
    2. Yeah, the problem (when one dismisses majority opinion has having weight) is that they're not sampling over the set of all opinions people hold. If you look only at *controversial* issues, well, the majority doesn't look so great. But if you take a random topic, sampled evenly across all topics, you'll get something like "what is the color of this item?", for which they're little disagreement.

      Delete
  3. Anonymous6:25 PM

    There are areas in which ""If the experts agree that X is Y, then X probably is Y" is true and areas in which it is not true. For example, Philip Tetlock's work shows that there are areas where expert predictions are only slightly more accurate than the proverbial monkey throwing darts, and worse than simple rules of thumb (e.g. assume that whatever has being happening will keep happening). A lot of political and economic prognostication falls into this category.

    ReplyDelete
    Replies
    1. Absolutely true, Josiah, although I would prefer to say that in such fields there are no real experts, only faux experts.

      Delete

Post a Comment

Popular posts from this blog

Libertarians, My Libertarians!

"Machine Learning"

"Pre-Galilean" Foolishness