Conspiracy Theory: Why it works?

Nowadays, when conspiracy theories such as 5G Networks spreading Covid, Pizzagate, and a secret cabal taking over the world are ubiquitous, one perplexed at how any rational person could fall for it.

It is fascinating to observe how most of us have an implicit prior tendency to easily believe in conspiracy theories rather than follow the rational path. For example, why when in the lottery, we see an identical set of winning numbers two weeks in a row, we inclined to embrace conspiracy and believe that all lottery officials are corrupted. Rather than contemplate that such coincidences are not improbable and mathematically, each combination of numbers is exactly as likely to come up as any other.

In this article, I will try to lay down my thoughts on how the Bayesian paradigm helps to explain the way conspiracy theories work and why those theories are sometimes successful. If you are not familiar with the Bayesian concept, let me explain it to you by example from Jaynes(2003). Suppose some dark night you walk down a street and suddenly see a jewelry store with a broken window, and a gentleman wearing a mask comes crawling out through the broken window. Of course, you don’t hesitate at all in deciding that this gentleman is dishonest. But now suppose you have been told that the owner was coming home from a masquerade party, and as he walked by his store, a passing truck threw a stone through the window. This new evidence did not make the gentleman’s dishonesty certain, but it did make it extremely plausible. That is, given a new piece of evidence we update our prior beliefs.

In the case of the lottery example, suppose we have so-called flat prior beliefs and think that all combinations are equally probable. Thus, seeing the same combinations twice in a raw is something that does not affect much our degree of belief. However, in real life, our prior beliefs are spiky, and we assign an ample amount of mental weight to a few theories. And if you do happen to give a high weight to one of the conspiracy theories, unless it is designed to survive the winnowing process, your degree of belief in the conspiracy will drive down each time you encounter inconsistent evidence. That’s how conspiracy theories grow and die-out gradually.

Now, if this is a case, why some conspiracy theories survive a long time? Let me refer to Ellenberg(2014) to answer this question. Suppose you learn from the trusted influencer that a secret cabal kidnaps children, slaughter, and eat them to gain power from their blood. Let call this theory Q. At first, because you are following your influencer a while, you assign that theory a reasonably high probability, let say 0.2. But then you do your research and see that there is no such investigation by police or other higher law enforcement body, as well, the majority of society thinks that theory is a hoax. Thus, each of these pieces of information is pretty unlikely, given Q, and each one knocks down your degree of belief in Q.

That’s why your influencer isn’t going to give you theory Q alone. He is going to add to it theory N, which is that the government, Justice Department, and the news media (let do not forget Hollywood stars) are in on the conspiracy together, with the newspapers and cable networks feeding false information to disguise the influencer’s theory. Of course, the prior probability of the combined Q + N theory is small, because it is harder to swallow both Q and N at the same time. But as the evidence flow, even though it kills Q alone, the combined theory Q + N remains unshakeable. How Jordan Ellenberg calls it, the theory N acts as a kind of Bayesian coating to Q, keeping new evidence away from it. This property is common in all successful conspiracy theories. They are wrapped in just enough protective blankets that they remain consistent with many possible observations.

Unfortunately, social media and algorithms designed behind it do not alleviate the situation. Competition for our attention drives the algorithms to feed us with the information that we agree with apriori, and after a while, we are soaked in a deceptive veil without even trying to listen or accept other people’s opinions. Social media provides each of us with our reality, which we are comfortable with, and the principle of contradictory opinions does not work anymore. The algorithm feeds as only with the posts, videos, and groups of people that share our opinion, not the opposite. Thus, the intricate line between subjective and objective reality vanishes, and we fall into the dangerous fallacy that our subjective reality is the objective reality.

Let me end the article with an old saying in mathematics — “When you are working hard on a theorem you should try to prove it by day and disprove it by night’’. This advice is not just for mathematics; It is always beneficial to question (without cheating) all our beliefs, starting from the social, political to philosophical. And at the end of the day, if our beliefs are unshakable we understand better why we believe in what we believe.

Ellenberg, Jordan (2014), How Not to Be Wrong: The Power of Mathematical Thinking. The Penguin Press.

Jaynes, Edwin (2003), Probability Theory: The Logic Of Science. Cambridge University Press.