An easy way to test to see if an argument is flawed.

Don't just look at the conclusion, and assume that if it's a conclusion you agree with then the argument must be logically valid.

Instead, ask yourself "what other conclusions follow from this?".

It's especially worth doing this with your own arguments, before setting them out.

For example, if you are advancing an argument against the death penalty, look at what you have said and see if there is anything in it that specifically would *not* apply to prison sentences. If your argument could just as well be used to oppose imprisonment as execution, there *must* be something wrong with it, and you can set about working out what.

If you are advancing an argument against discrimination on grounds of sex, race etc, look at what you have said and see if there is anything in it that specifically would *not* apply to discriminating on grounds of intelligence.

If you are advancing an argument against discriminating against a certain group, reread your argument, taking out every reference to that group and substituting "murderers". If it still applies, there's probably something wrong with it.

If you are advancing an argument for abortion rights, look at what you have said and see if there is anything in it that specifically would *not* apply if foetuses were fully intelligent and capable of arguing philosophy and pleading to be spared (unless you'd still support abortion rights if that were the case).

If you are advancing an argument against American military involvement in Iraq, look at what you have said and see if there is anything in it that specifically would *not* apply to intervention in WWII or in Kosovo (assuming you think those were good decisions, of course).


Being right for the wrong reasons is all very well, but it's likely to lead to fuzzy thinking, and that's likely to lead to being wrong about other things in the future.

Of course, this approach isn't infallible; it will still produce false positives. But it's a useful check - far too often people assume that if an argument reaches a conclusion they support, it must be a valid argument.