maanantai 20. maaliskuuta 2017

Obscurity of Unlikely Facts

In this era of vast flow of information, it has become harder and harder for anybody to keep up with any topic of interest in a comprehensive way. This particularly applies to topics which provoke heated debates and appeals to mankind's intrinsic desire to reinforce their already existing opinions and views.

To this end, I've been playing around with a thought pattern lately which I've now fancily dubbed the law of Obscurity of Unlikely Facts. The purpose of this rather simple "check-list" of a cognitive bias can be applied rather universally, though in fields where it's hard to quantify and to measure a phenomenon in an exact manner it's harder. Simply put the idea behind the whole thing is the saying "Even a Blind Chicken Finds a Kernel of Corn Now and Then". This saying doesn't seem very popular in English, but it's used time to time in Finnish.

The basic idea is, that if we can do the (false) dichomization on whatever topic to be able to say whether a certain fact is TRUE or FALSE, let's say on a political debate, usually people have a set of e.g. 5 main commandments they follow. Suppose a person with an agenda posting the following list (which might be in a narrative format but can be reduced to a list of main hypotheses):

The great Tomato debate thread example:

Arguments by tomato-specialist Foo Bar:
1) Natural tomatoes are always yellow
2) Tomato farmers have gone too far in bringing out unnatural red tomatoes
3) Farmers are backed up by an international tomato X company wanting us to eat addictive (red) tomatoes
4) All tomato farmers are evil
5) Tomato farmers want to make a profit

A pro-red tomato person counters this post in a relevant forum, and pushes in for a counter-attack. He makes a list of the hypotheses and starts collecting counter-material. He then posts the following post (shortened from what you would encounter in real-life):

Counter-post by red-tomatoist Qwerty Asdfg:
"Mr. Bar recently posted a seriously misguided list of curious facts that I feel need to be corrected. He claims that all natural tomatoes are always yellow, but we have discovered many varieties of tomatoes in the wild that are by nature red, as recently published in the International Journal of Tomatoes by Prof. Tomatogic et al. Furthermore, as we encounter these red tomatoes in the wild, it is hilarious to even propose that farmers would be the ones propagating the advance of red tomatoes over yellow ones - majority of tomato subtypes are red, as shown by Prof. Tomatogic. There is zero evidence that e.g. Prof Tomatogic would have any affiliations to company X, and looking at their catalogy at my nearby tomato store I discovered that they do offer a large variety different colours of tomatoes, including yellow ones. Bringing in X in relation to tomato farmers is just ridiculous. This all just shows how misguided Mr. Bar is in his original post; to claim that tomato farmers are evil is just a cherry on the top, and tomato farmers do not want to make a profit - they just aim to produce the healthiest of products and take great pride in their work!"

Now, for starters, both posts by Bar and Asdfg would rub me the wrong way. Both are very extreme and show no signs of lenience towards the opposing view, which in my opinion is the key to detecting a triggered argument which should be taken with a truck-load of salt. But the point of the law of Obscurity of Unlikely Facts is to systematically evaluate Asdfg's counter-arguments, disregarding how ridiculous the original claims may have been.

While Asdfg appeals to reason and even science in his shooting-down of mr. Bar's post, I find above sort of counter-posts questionable. First of all, it addresses every single point from mr. Bar directly in a list-like manner, even if it's narrative. When I see this sort of counter-post, red lights start flashing. To me, it only indicates that Asdfg has first decided that mr. Bar's arguments are to be proven wrong, and then Asdfg has collected the evidence and/or rhetoric arguments to address every single one of them. A statistical basis for the law could probably be found in False Discovery Rate or the lack of corrections for multiple testing, but that's beyond this post.

Now the key point is that when I see somebody go through their opposer's supposed arguments in this non-constructive, list-like systematic way, it makes me question Asdfg's motives - he or she has probably not even tried to understand the original arguments and/or taken the time to consider whether they can be interpreted in a manner that is true.

To summarize the law of Obscurity of Unlikely Facts:
The longer the list of arguments by person A, the more likely that at least one or more of these arguments are true to at least a reasonable extent. In reverse, the longer the list the A's opposer B shoots down systematically, the more likely the person B is biased to just counter A without a proper, through-out and fair inspection. There are multiple factors that affect this (e.g. ad hominem yes/no) or lenience in B's comments, but a systematic list vs. list discussion is very likely to include a degrading amount of cognitive bias and probably ought to be ignored altogether.

If you engage in multiple discussions of sensitive topics, I have found the above simple rule useful in filtering out discussions where the discussion is just not worth my time. Usually they degrade into a kindergarten-level "yes! no! yes! no!" poking spiced with ad hominem, so it's better to move on to a better forum/thread if one really wishes to find out information on the various views to the topic. But above rule can also be used to detect discussions that are a great source of humorous content (Dunning-Kruger effect usually increases in this case as a function of time), in which case one should go grab some popcorn, just sit back, relax and see it unfold in its glorious marvel.