Today I’d like to write about something a little more abstract than usual: logical fallacies.
If you were on the debate team in high school, you’ve probably studied these extensively, but unfortunately, the rest of us were left to muddle about on our own. And that’s unfortunate, because understanding some of the common logical fallacies can help us understand why sometimes what feels right is actually wrong. (I believe all the fallacies I’m discussing here are informal fallacies, in case any of you care about the difference.)
I touched on this subject back in my very first post, “Correlation is not Causation,” when I discussed my own mistaken belief that my daughter was having an allergic reaction to an antibiotic.
The problem is that our brains have not had time to catch up with the intricacies of modern life. Evolution just doesn’t work fast enough to go from agriculture to the Internet in a hundred or so years.
There are a lot of logical fallacies, and I don’t have the time to touch on more than a few (You can find some lists here and here if you’d like to read more), but I’d like to discuss some that I think cause problems when evaluating medical claims and medical problems.
First, there’s argumentum ad populum, the appeal to popularity, appeal to majority, or bandwagon fallacy. You’ve heard this one: “I’m sure acupuncture works. After all, a billion Chinese can’t be wrong!” Well, actually…they can. This is also the argument that tries to sell you perfume: “Look, these beautiful people like our perfume, and you want to be like them, so you must like our perfume.”
Unfortunately, it doesn’t matter how many people believe that acupuncture works, that still doesn’t qualify as evidence. Bzuh! Thanks for playing. As Mark Twain said, “One of the proofs of the immortality of the soul is that myriads have believed it—they also believed the world was flat.”
A related fallacy is argumentum ad verecundiam, or the appeal to authority. “Dr. Mehmet Oz says that colonic cleansing works, so it must be true. After all, he’s a doctor!” Well, that’s nice, but Dr. Oz also thinks that John Edwards can talk to dead people, so pardon me if I’m not exactly willing to take him as a sane pers–er, an expert on everything.
Just because someone is a doctor, or has a PhD, it doesn’t mean they’re always right. One doctor may go off the deep end and prescribe ground up shark fins for cancer, but that doesn’t make it right just because he somehow scraped through medical school without getting tossed out on his ear. The point of science is consensus, and the process weeds out the crazy stuff.
Of course, my favorite fallacy (yes, I have a favorite logical fallacy. Shut up.) is post hoc, ergo propter hoc (after which, thus because of which), aka correlation not causation. This is the one I fell prey to back in my first post. “Yael took an antibiotic, then a few hours later was sick. The antibiotic must have caused the illness.”
Not necessarily. Around the world, hundreds of people (hell, maybe thousands of people) take an antibiotic and then get into car accidents. However, the antibiotic didn’t cause the car accident. But (I hear you cry) antibiotics can cause illness! True. Absolutely true. People do have allergic reactions and bad side effects to antibiotics. And if you take an antibiotic and then get very sick, you should certainly see a doctor in order to figure out what’s wrong.
But…just because you got sick, it doesn’t mean the antibiotic is at fault. Just because your kid developed autism within a month after getting an immunization, it doesn’t mean the immunization is at fault. Just because one thing happened after another thing, it doesn’t prove the first thing caused the second.
Another version of this fallacy is the opposite circumstance, in which you take something and then feel better. If you take a homeopathic medication, I can almost guarantee you’ll feel better at some point afterward, because many illnesses are self-limiting and you’ll get better no matter what you do. Between that and the placebo effect, you’re likely to believe that getting better is caused by the homeopathic medicine, even though they’re unconnected. Because homeopathic medications are pretty much just expensive water.
Then there’s joint effect or the fallacy of the single cause, in which one thing is seen as causing another, although they’re actually both symptoms of a different cause. So, for example, if your child has a high fever and vomiting, you might think the fever caused them to vomit, but both the fever and vomiting turn out to be symptoms of scarlet fever.
And let’s not forget another personal favorite: demanding negative proof. This is especially popular in arguments about atheism and religion, when the religious person demands that the atheist “prove God doesn’t exist.” Sorry, Charlie, but there are two problems with that. First, the person making an extraordinary claim is the one who has to provide proof. And second…how in the world are atheists supposed to prove a negative? (If you think it’s easy, then I demand you prove the nonexistence of the Invisible Pink Unicorn, praised be her horn. She’s invisible! How do you know she’s not standing right behind you, getting ready to stab you with her horn?)
In medical contexts, it’s demands for proof that vaccines don’t cause any disease ever in any child in any circumstance. Dozens of studies have found no connection between vaccines and autism/epilepsy/whatever, but vaccine opponents always change their claim and demand proof that vaccines can’t ever do any harm. That’s a little difficult to prove.
And, of course, no such list as this is complete without cherry picking. If the discussion is of some questionable medical treatment, then you can’t win the argument by describing the 10 people who got better while ignoring the 100 people who got worse. You can’t pick your data, you’ve got to work with what you’ve got, even if you don’t like the results.
I could go on all day, to be honest (I left out the argument to moderation! And the perfect solution fallacy!), but I think it’s time for me to stop. I hope that these examples were helpful, or at least somewhat interesting 🙂