In the first of my "How Not to Think" blog posts, I discussed the conversational maneuver I called the "but still," which is the moment when you voice an opinion, someone offers a good and reasonable objection to your opinion, and you answer with "OK, but still." We say things like this all the time, and even a cursory consideration tells us that such an answer is counterproductive garbage, and probably an example of motivated reasoning.

This time, I'd like to elaborate on motivated reasoning, to both understand it as a useful concept and to caution against a particular way of applying it.

To explain motivated reasoning, I'll start with an anecdote: Imagine a young woman is accused of stealing a car from a stranger. She is tried in a court of law. The jury hears testimony from several witnesses explaining that they saw her break in to the car with a coat-hanger, hotwire the engine and drive away. The car is later found abandoned, and police testify that the defendant's fingerprints have been lifted from the steering wheel, and strands of hair matching her DNA have been found on the floor of the car. The defendant maintains her innocence but has no alibi.

Based on this evidence, the jury finds her guilty. Most juries probably would, and it's reasonable to assume that most impartial people in the world would come to the same conclusion. According to the evidence, she probably stole this car.

Consider, however, the defendant's parents: They know they raised their daughter right, and they're positive that she would never commit grand theft auto. Judging from all of the same testimony, revealed in exactly the same way, they find the case against their daughter highly dubious. The witnesses against her seem vague and untrustworthy to them. The parents cast doubts upon the inherent accuracy of forensic science, and they are unconvinced that the fingerprints found in the car really belong to their child. In this case, the defendant's parents are almost certainly showing signs of motivated reasoning.

In dealing with questions that are truly outcome-neutral to us -- that is to say, questions to which the answer does not affect us in any way -- we have little trouble examining the evidence fairly, applying logic and finding the conclusion that seems most likely to be true. Think of the way you solve an abstract math problem. You honestly and transparently seek the true answer. You really have no significant preference for X=3 over X=5.

Motivated reasoning, however, works backwards: We start with a conclusion that we desire to believe (such as the conclusion that our loved ones are innocent of car theft -- or, more commonly, that our cherished political values are vindicated by facts and data) and we tailor our experience of evidence and logic to fit our desired outcomes.

People who offer evidence contradicting our pre-selected conclusions seem shifty and untrustworthy. People who testify in favor of our desired conclusions, on the other hand, seem intelligent and respectable. Logical arguments that bear out our convictions seem tight and rigorous. Logical arguments opposed to our beliefs appear fallacious and full of holes.

We can all agree that motivated reasoning is not a good way to determine the facts about the world, and nobody would dispute that we should do all we can to avoid letting our psychologically-rooted biases distort our vision of evidence and reason. Unfortunately, here's where we hit a classic impasse: A person performing motivated reasoning does not know that he or she is doing so. As far as we can tell, motivated reasoning feels exactly like sound reasoning. It happens unconsciously and is mostly undetectable to the person doing it. If we were to use the example of the young woman accused of stealing a car, her parents don't believe they're being unreasonable by rejecting all of the evidence against their daughter. If we were somehow to get inside their heads, the evidence that seems so reasonable to us now would suddenly seem very thin and weak. They really believe this.

So here's the quandary: Almost everybody would agree that motivated reasoning exists and is a problem, yet almost no person recognizes it in him- or herself. The danger is that we let this knowledge of one of the gears of human psychology become an idea we're quick to apply exclusively to other people. It would be solipsistic folly for a person to think, any time he or she has a disagreement with someone, "My opinion is based on evidence and reason. Your opinion is based on the desire for affirmation of your feelings." This point of view would be treating psychology as "the science of everybody else." Surely we must all realize that we're each as susceptible to unconscious biases and desires as the next person.

If we rule out the impossibly arrogant starting point of "me = reasonable, you = irrational" in our discourse, what is left? There seem to be two options: We either treat everyone's arguments, including our own, as arbitrary side effects of psychology, or we treat everyone's arguments, despite our knowledge of the phenomenon of motivated reasoning, as logical, evidence-based and subject to review and revision. I think we really have no choice but to do the latter.

My argument is that despite how interesting it may be to contemplate the idea of motivated reasoning, you simply can't have dialectic if you assume goal-oriented thinking to be the standard. If everybody believes everyone else is arguing in bad faith, what's the point of arguing at all? If we believe juries to be heavy on goal-oriented thinking, what's the point of presenting evidence at a trial? If we believe our friends and relatives to be guilty of motivated reasoning at the outset, why bother debating this or that fact about politics, ethics or religion? And if we consider ourselves hopeless victims of this common mode of delusion, what's the point in seeking truth ourselves? Yet somehow, despite our highly biased ways of sampling evidence and proceeding through the steps of logic, we know that discussion and debate really does sometimes clarify truths and change minds. And if you're accused of a crime, evidence really does matter.

Motivated reasoning is thus left in a curious place. We know it happens. It's unfair and probably wrong to think of it as "the science of everybody else." If we assume it is everyone's basic operating principle, we can never have useful arguments. So essentially, we must acknowledge its existence, do our best to guard against it in ourselves, and then forget about it, and try to live our lives.