Friday, October 14, 2005

If asked about a conditional statement, would you understand it?

A poll has found that 50% of Americans agree with the statement: "If President Bush did not tell the truth about his reasons for going to war with Iraq, Congress should consider holding him accountable by impeaching him." (via Leiter) There are two things really odd about this article, which indicate that people don't understand conditional statements. Firstly: Fertik, a quoted commentator, suggests that this result shows that "a solid plurality of Americans want Congress to consider removing Bush from the White House." But of course it shows no such thing, for this result is entirely consistent with everyone being quite certain that Bush was honest. It is transparently fallacious to infer B solely from if A then B. Clearly one must first establish A!

Secondly, I'm amazed that only 50% of Americans agreed with the statement. Worse, only 20% of Republicans! Surely even Bush's staunchest supporters ought to agree that, if he had lied (which, of course, they would deny), then he should be held accountable for it. If the (hypothetical, remember!) scenario of a democratic leader deceiving his nation into war does not constitute grounds for impeachment, I'm not sure what does. That 80% of Republicans (and 28% of Democrats!) would refuse to consider holding Bush accountable even if he lied about the war, is truly scary. Surely they don't really mean to say that.

I can only conclude that people don't understand conditional statements. Perhaps they instead interpreted the poll question as a conjunction, i.e. as asking them whether they agree that both (1) Bush lied, and (2) he thus ought to be impeached.

Perhaps some experimental philosophers would like to test this some day. They might ask Republicans whether they agree with the following statements:

"If Bush is a pedophile, then criminal charges ought to be brought against him."

or "If Bush is plotting to destroy the universe, then he ought to be stopped."

or simply: "If Bush is a bachelor then he is an unmarried male."

Would they also deny all of these obviously true conditional statements, merely because the antecedent happens to be false? Supposing not, would Fertik then conclude that "a solid plurality of Americans think that Bush is unmarried"?


  1. I was pretty shocked when I read that, too. Which is worse: the fact that American's don't understand conditionals or the fact that nearly half of Americans rejected the conditional?

  2. It suggests both that Americans don't understand conditionals and that journalists and philosophers alike don't understand polling. 50/50 suggests "I don't know" was not a possible response. 50/50 is consistent with a strongly skewed population of people with strong opinion of the matter and zero understanding of the poll question. It's randomness.

  3. Erm, "44% disagreed, and 6% said they didn't know or declined to answer."

  4. I think either they did not seriously consider the question or they are just refusing to play the game - they may well see the answer to the question to be a political statement that they dont want to be misused.
    In fact, with this in mind, one might ask why yu would EVER answer a survey purely honestly as opposed to tactically. I many take that into account next time...

    there is also the question "if he lied what exactly was he nature of that lie". Some lies might be more of an issue than others. For example if I thought all politicians lie (a common position) I might be wiling to accept a 'white lie' by a president when the action itself was a thing I suported, or if I felt others were decieved but I was not.

    I am interested to see if anyone has asked any of those questions though -

    "If Bush is plotting to destroy the universe, then he ought to be stopped."

  5. I think there's a game-theoretical element here as well. Even though I understand that conditional perfectly well, I'd be wary of actually answering the question "correctly" because I know that pollsters and pundits are mostly a bunch of morons like Fertik who'll draw totally invalid inferences from it. Which is, by the way, why I don't answer surveys.

  6. That suggestion that its treated as a conjunction is interesting. All you would have to do is repeat the survey with both questions, the conditional, and the conjunction (Bush should be impeached on the basis of his lying to the public). It would be interesting to see if there were very few people who answered yes to the first, but no to the second :)

  7. What, *you* understand conditional statements but "people" do not? Perhaps you need a more charitable interpretation of this result?

  8. If you have a suggestion, I'm all ears...

  9. I think it most likely that 50% of people don't understand conditionals. People are idiots, and moreover have no experience being held to account for their reasoning.

    Polls are ludicrously easy to manipulate -- or to be more accurate, you can prove anything through taking something out of context and deliberaturely misinterpreting it. For all we know, the poll was highly contextual, talking about specific events for example.

    There are many things I understand but "people" don't, and that's just fine because I can justify them. There are also many things I don't understand, which is also fine, because I avoid having strong beliefs on those subjects.

    The results of the poll are roughly what I would expect.


  10. haha MP you are such a cynic!

  11. In general people will accept very poor arguments if they already affirm the conclusion. This could be one of these cases, they agree that Bush shouldn't be impeached so they bite the bullet.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)

Note: only a member of this blog may post a comment.