On Characters’ Characters

by F.

The ancient Greeks called it character, Freud thought of it as personality, and nowadays we call it the fundamental attribution error. What am I talking about? The tendency we have to make inferences about people’s dispositions based on small samples of behavior. This is why the “entrance” of a character in a story is so important. Even though they are often wrong, first impressions count.

Take the beginning of The Road Warrior. Remember that first scene in the car? We see Max, dirty and disheveled, driving in his car. Then there’s a reverse angle on his dog, who in sitting in the passenger seat. And the dog is wearing a red bandana. The dog looks at Max, lovingly. And right then, I think we make some assumptions about Max: even though he looks dirty and beat up and dangerous, his dog loves him. And he loves his dog—he dressed him in a bandana. He can’t be all bad, right?

Another example. Robert Redford made a movie called An Unfinished Life with Lasse Hallstrom directing. After a couple of beautiful establishing shots of a farm in Wyoming (and Bart the Bear, in a supporting role), we see Redford’s character, Einar Gilkyson, milking a cow. In the background, Einar has set up a few bowls for his cats to drink from. And slurping away with the cats is a raccoon. Cassi and I were watching this together, and I turned to her: “I like him already.” She felt the same way.

This is a human tendency called the fundamental attribution error. That is, we underestimate the extent to which human behavior is shaped by situations and overestimate the extent to which it is shaped by dispositions. We see Einar or Max and we think we know the way they are. And we do this without thinking: it is automatic.

Now, it is called the fundamental attribution error, rather than something like the fundamental attribution mechanism because these attributions are often wrong. Take an example described in a recent textbook, Social Cognition by Ziva Kunda. In one study,

participants were told that a psychiatrist had interviewed applicants to the Peace Corps, and were asked to estimate how well these interviews could predict the applicants’ performance as community organizers working for the Peace Corps in third-world countries…. Their estimates were compared to actual correlations obtained from an earlier study on Peace Corps volunteers…. [P]articipants thought a brief interview with a psychiatrist could predict performance in a new role in a foreign country with considerable accuracy; they estimated the correlation between the interview and performance at about .60 [which is pretty high]. In fact, the correlation was an almost negligible .07.

And yet, this is what we do, even when we try to correct for it. What’s the point for storytelling? Two things. First, the audience will infer the character’s disposition from that initial encounter. If you want the audience like the character, he should be giving milk to cats and the neighborhood raccoon. Second, the audience will tend to discount situations and focus on dispositions generally, so each few second we see the character will lead to inferences about what he or she is like.