Bad comments are your fault

Popular Science doesn't like the quality of its comments. It says there are too many "trolls and spambots." So, the magazine has taken the single most effective anti-troll, anti-spambot counter-measure: It's closed down its comments. In so doing, it's made a powerful statement to the entire Internet: "We don't know how to run a commenting feature, so shut up."

Let me put this more bluntly: If the comments on your site's content are broken, it's your fault.

In its announcement of the policy change, Popular Science cites research that shows that if a discussion is highly polarized, readers of that discussion tend to develop strong, polarized opinions that affect their trust in the article itself. Negative comments can lead to a reduction in the credibility of the article, even if those comments are entirely unfounded. Thus, the trolls don't just ruin the conversation, they can hurt the cause of science. That's an additional reason to worry about troll-filled discussions.

The options

Bora Zivkovic, the blog editor at Scientific American, cited the same research in a post in February, but he came to a different conclusion. He recognizes the problem and writes, "If commenters think your commenting thread is a public space where they can do whatever they want because nobody's watching, they will do whatever they want. And that is not pretty."

"One option," he says, "is to give up on comments entirely, and perhaps completely shut down the commenting functionality, trying, at the same time, to find and track discussions wherever they may be happening." Zivkovic chose the other option: He's rigorously moderating the discussions on his personal blog hosted by Scientific American.

Key to this, Zivkovic is enforcing some sensible rules intended to bring about a particular type of discourse. It's Scientific American, so he's not looking for witty banter or amusing GIFs. The discussions on his blog, he says, "are about science and about the way science and society interact. This also means that the content of our site tries to present factual information about the world, as best discerned from scientific data." Troll and your comment will be removed. Cry that you've been censored, and he "will laugh out loud."

Nor will Zivkovic allow comments that are off-topic. If an article is about, say, a new hypothesis about the relation of geography and the variation of species, a comment debating the theory of evolution itself will count as off-topic. "Evolution is a fact," Zivkovic writes. "Questioning the fact of evolution is not a part of discussion of a particular new finding or mechanism ... The same goes for climate change."

That sounds pretty harsh and perhaps even unscientific. It's neither. It sounds harsh because we've been told that all authentic conversations have to be open to every point of view. But that's just baloney. Conversations require a great deal of common agreement in order to enable some useful disagreement. In this case, a useful disagreement is about what a post is about, not about whether science itself is valid.

The venue

It sounds unscientific because we've been told that scientists are open to new facts and hypotheses. And that's true ... but those facts and hypotheses have to follow the rules of science itself. So, your "feeling" that something is true or your Biblical citations are not welcome in a scientific discussion. They are perfectly welcome elsewhere. Write your own post about them. Similarly, comments from scientists claiming that miracles are impossible because they violate natural law should count as trolling at a faith-based site.

Even among people who share a set of beliefs—for example, among scientists—there will certainly be arguments about what constitutes an on-topic and legitimate criticism of a post. Topics are not so cut and dried that it's always obvious when one has strayed away from them. All rules are subject to argument. The important thing is to have rules, to make them clear and public, and to enforce them as best you can.

That's not just how commenting systems work. It's how knowledge itself works.