What’s wrong with paid comment?

Last week, the Guardian ran an article about “Russia’s Troll Army”. “Troll” is something of a misnomer here: the people in question are not out to provoke a reaction. Rather they are paid to promote the government’s line on political and social issue. They maintain blogs and social media profiles under pseudonyms, where they post made up incidents from fabricated lives, interspersed with political posts they are instructed to write, lauding Russia’s intervention in the Ukraine (for example). And of course they post comments on news stories, on Russian and foreign websites (for reasons unknown, they don’t seem to target Practical Ethics).

The ‘troll army’ is a new version of a propaganda battle which is (at least) decades old. In an article on CIA funding of student organisations and European cultural initiatives, Louis Menand notes that the revelation of where the funds were coming from alienated the very people the CIA was trying to court. The discovery that funding is conditional on saying the right things is repugnant. The discovery that the comments we read are funded by governments is likely to be even more upsetting. The discovery that an online ‘friend’ is a fiction crafted for spreading disinformation is likely to be especially disturbing.

But are there grounds for caring whether comment is paid for? After all, the ‘trolls’ are not saying anything that those who read them won’t hear from other sources. So why care whether they are paid to say what they do?

Louis Menand suggests that the CIA funding affected who was in a position to say what, but no one was induced to say anything they didn’t believe. In that way, the CIA program was very different to the Russian program, which pays people to say certain things. But Menand’s claim seems naïve, and the difference between the two programs (though real and significant) may be smaller than appears at first sight.

Belief formation is likely to be highly influenced by our social environment. There is some empirical evidence for this: for instance, from Dan Kahan’s work on how one’s social network influences what one believes about climate change. The influence of social network is probably adaptive. Acquiring beliefs by testimony from other agents and simply by observing their behavior enables us to aggregate the evidence gathering work of hundreds or thousands of people, thereby ensuring that we are not epistemically dependent on what we can test for ourselves. It allows us to go into (what are for us) novel situations already prepared. Since most of the time most of our peers are saying what they believe and either have evidence for their beliefs or have acquired them from others who have good evidence, we are able to be better epistemic agent by simply acquiring beliefs by observation and testimony.

When people are paid to express opinions, and also when people are placed into positions of power on the basis of their views, these mechanisms may no longer work to improve our epistemic position. On the plausible assumption that our credences (the strength of our beliefs) are sensitive to the numbers of people who express an opinion, and to their place in a hierarchy (because presumably people with better justified beliefs tend to have a fitness advantage over those with worse), paying people to express opinions, and ensuring that those with certain opinions are in leadership positions, muddies the waters. These efforts might be seen as epistemic contaminants, lowering the quality of our belief formation and making us less reliable. Of course, advertising raises the same sort of worries: it does not merely inform us; it may instead trick belief-formation mechanisms in ways that we would not want.