On Philosophy

February 18, 2007

Why Popular Doesn’t Mean Best

Filed under: Political Philosophy — Peter @ 12:19 am

I assume you have heard of reddit and digg, but let me describe them to you anyways, in case you have been living in a cave. (How are you reading this in your cave? Does someone print it out and bring it to you?) Reddit, digg, and their imitators are sites where readers submit links, which are then voted upon by other readers. Links that receive the most votes appear on the front page, while those that receive less remain out of sight. Through this mechanism these sites promised to bring to their readers the best content on the web, leaving what constituted the “best” content to the readers. Initially it seemed to work well, the early adopters found that nearly everything on the front page was to their liking. But as they have become more popular the content seems to have gone downhill, with perhaps only a handful of the links on the front page worth looking at on average. How can we explain this, given that more readers should have meant that the democratic setup would be more effective?

Well the problem lies in the nature of democracy itself. Consider the small group of early adopters. They were much alike in their interests and thus tended to vote similarly to each other. And this meant that when any one user went to the front page they found the links that people much like themselves had voted up, and so there was a good change that they would like them as well. But as you add more people the group of voters becomes less homogenous. Because of the nature of voting this means that only links with broad appeal will rise to the top. Which isn’t always a bad thing, but it means that the links that appeal strongly to only a smaller group of the readers will never make it to the front page, and this is bad. And this means that the average reader will find many links that are mildly appealing, but nothing that is really interesting to them. (Assuming that for a given person the most interesting topics for them are of a specialized nature. This is probably true of most people, because in general what people feel passionate about is some specific topic or activity, and anything that would engage them in this area will be too “technical” for those who aren’t already interested in it. For example, I have never seen anything philosophically interesting make it to the front page of reddit, although there is plenty of philosophically interesting content on the web.)

Of course the reddit and digg audiences haven’t grown large enough to encompass all kinds of people equally, for example reddit has a liberal/libertarian and atheist bias, and digg has a preference for gadgets, but they are getting there. So what is the solution? Well, ideally, you want to segregate the readers by their interests to get back to a situation more like when the site first started. One way to do this is to provide different sections for different topics, such as politics, physics, and so on. Instead of submitting links to the site as a whole to be voted on by everyone readers with narrower interests could submit links to the relevant topic section and have them voted on by readers of that topic sections alone. Both reddit and digg have done this to some extent, but on both sites there are only a few topic sections, and which topic sections exist is dictated by what the site owners feel is important, and not by what the people want (ideally the users should create topic sections, just like they submit links). And on top of that people should be able to “subscribe” to one or more topic sections, so that when they visit the site they are presented only with the links from the topic sections they are subscribed to.

Of course this is all somewhat tangential to the philosophical ramifications of these issues. The first thing that the evolution of these sites demonstrate is that democracy is not a universal solution that you can simply throw at a problem. Democracy works when you are trying to find things that are compromises, that are acceptable to most people. It doesn’t work when you are trying to find something outstanding, something that the majority of people may not appreciate. For example, you wouldn’t want to decide what was great art by democracy (because many people don’t like cubism, but that doesn’t make it bad art), nor would you want to decide what was true by democracy. Of course if you are working with democracy in such a situation you can improve your position by restricting your pool of voters, for example when deciding what was great art you could try to restrict your pool of voters to people who are known to have good taste, and when deciding what is true you can restrict your pool to people who are impartial and knowledgeable in the field. Of course it is even better to do away with democracy altogether in these situations and devise a method for deciding that is independent of what people think (for example, in the case of truth you can test your claims). I guess the real question then is: are laws, foreign policy, ect the kind of things that a compromise solution is right for, or should we trying to find a way to make them objectively the best?

Blog at WordPress.com.