Book Club: The Character Gap, 2, The way we actually are

There seems to be a gap between how good we would like to be (or even think we are) and the real us. That’s the thesis of the second part of Christian Miller’s The Character Gap, which we are discussing in this ongoing philosophy book club. Miller devotes four chapters to documenting — plenty of psychological experiments in hand — how we are doing neither horribly nor particularly well in four areas related to our character: helping, harming others, lying, and cheating. The specific anecdotes and relevant research are interesting and well worth reading, but I’m going to focus here on the general findings from part II of the book.

The evidence martialed by Miller is decidedly mixed, which is his main point. He calls his generic human being — the aggregate of all the experiments he has examined — Frank. Frank is in the company of 76% of people when he voluntarily helps someone he has never met before, out of empathy for the stranger’s predicament. He is also unwilling to cheat if he is reminded of his values, and that holds even when no one is watching.

Then again, Frank is a common victim of the bystander effect: if someone is in need of assistance he won’t help if he is surrounded by people who are not helping. Miller’s main explanation for this is not that Frank has now suddenly turned into a vicious person. He is simply unsure of what to do, and does not want to risk the embarrassment of standing out, especially if it turns out that he has somehow misread the situation.

Overall, the studies considered by Miller show that Frank — and hence most of us — is all over the map: sometimes he behaves admirably, at other times despicably. Moreover, these changes are extremely sensitive to our surroundings, often in ways that we don’t consciously appreciate. Setting aside the bystander effect, it turns out that people are more likely to help strangers if they just passed a bakery from which a warm and pleasant odor of bread or pastries is emanating! This particular effect is strong, with controls helping 22% (males) and 17% (females) of the times, while bakery-triggered subjects help 45% and 61% of the times respectively.

This also means another somewhat disturbing thing: we often don’t know our own motivations for helping or not helping. We are usually not…