Criticism Need Not Be Constructive

People like getting constructive criticism. And indeed getting constructive criticism is good, assuming that the "constructive" part is good. Criticism means that someone tells you about your work's weaknesses. Constructive criticism means that the person also tells you what to do about it. For example, Tim Harford writes about getting feedback on a recent talk he gave:

Bruno Giussani of TED, seeing someone praise me for speaking without slides,  immediately got to the point. “You talked about the Spitfire,” he said, “But this is an international audience. Many people won’t know what you’re talking about. You should have shown just one slide: a photograph of a Spitfire. Then everyone would have understood.”

Next time I give a similar speech, I’ll be showing one slide: a photograph of a Spitfire.

That's good advice. (As a commenter pointed out, given the international nature of the internet, it would have been helpful if Harford had included a photo of a Spitfire with his article.) But there's a mantra that criticism must be constructive - that is, don't criticize someone if you cannot also tell them what to do about the problem.

I think that is a harmful norm, for two reasons. First, once a feature of your product is identified as weak, people other than the identifying party may come up with a solution. For example, if Harford had simply been told that the audience didn't know what a Spitfire is, he could probably have come up with a solution himself. Second, even if no one can think of a solution, having a weakness pointed out leads to a more realistic appreciation of your product, which by itself can be very helpful. When several people in your movie production company report vague feelings that a specific screenplay would not make a good basis for a movie if the aim is to make money, you may not want to produce it. If tests show that the brakes on a new car are not working properly, you may not want to start selling it. These are cases where nobody points out how to solve the problem, yet knowing that the problem exists is very important.

This latter argument also illuminates why it is wrong to say, normatively, that "it takes a theory to beat a theory." The reason is that knowing you are ignorant is better than falsely believing that you have a correct theory. For example, if a treatment is actually harmful, the person who says she doesn't know whether the treatment works is closer to the truth, and will hence make better decisions, than the person who is confident that the treatment is beneficial.

"Criticism must be constructive" is the postulate of schoolteachers who do a bad job at preparing their students for the outside world, which is a harsh and nasty place. "It takes a theory to beat a theory" is the mantra of the proponents of the old theory who can't find anything wrong with a critical assessment.