Judging actions

I follow a religious guy on twitter. Outside of religious issues I find him a very reasonable and unusually patient twitter debater. (This later turned out to be untrue see here.) Once whilst tweeting about an issue he pointed out that we should not judge a religion by the actions of it's devotees. You'll hear similar points raised during many debates. For example many critics of Islam point to the treatment of women and homosexuals in several Islamic countries and question if this is really a reasonable 21st century stance. The oft heard response by religious commentators in more liberal countries (such as my twitterer) is that the faith is nothing to do with these cultural practices. This is certainly a reasonable point and is not limited to Islam or even to religions. For example scientists like myself get very annoyed when the scientific method is claimed to justify non-science (or nonsense depending on your point of view) such as homeopathy.

Fair enough you say merely judge the consistency of a persons actions to their beliefs based on whether any actions are in agreement with the base principles of that particular viewpoint. That's quite simple for something like science with its love of constantly updating and regularly publishing the scientific consensus. The mode of operation for science is to do your analysis and present your evidence aiming to convince people of your falsifiable hypothesis. The scientific consensus is then whatever the majority decide the totality of evidence suggests. Hence individual scientists or groups have every right to reject the consensus but if they can't over turn the scientific consensus they have no right to state that their opinion is the scientific one.

Religions have a different structure and method of operation. They often have a source of what they believe to be supreme truth which is then interpreted and applied to individual situations. One problem that can arise is equally religious groups even within the same faith can hold mutually exclusive views which they both believe to be justified in their main source. As the views they choose do not have to be falsifiable there is often no obvious way to decide who is actually correct. Both sides can exist side by side happily agreeing to differ without necessarily generating bad blood between the sides. Thou naturally it can and often is the basis and justification for violent behaviour. I could expand with specifics upon this point but I suspect that some might take the specifics as attacks on their beliefs.

So to continue lets instead use an imaginary belief. To simplify further lets assume that this has no internal divisions within it. This belief has a founding text which doesn't contradict itself and promotes nice inclusive liberal beliefs. We will call it Guardianism. There is however the issue that Guardianists are statistically proven to be more likely to commit assault than non guardianists. The scholars of Guardianism point out the beauty of their beliefs and they point out with justification that the assaulting is contrary to the bare principles of the main text. Nevertheless the abuse continues. So is Guardianism bad?

Before you jump to your answer pointing out the main text and it’s nice ideals note that right-wing groups often carefully draft texts of principles to appear reasonable. They do this partly to avoid getting banned and at least partly as extreme ideas are often rooted in what at first glance (and without further study) seem reasonable ideas. A right-wing spokesperson commenting on a racist attack by their members may very well say "Don't judge us by the actions of a small minority of our members, judge us by our main [very reasonable] agenda". I personally don't think you can realistically be expected to judge  groups of people by whatever cover text they choose to present you with. You surely must judge them by their actions. You can even be a bit more scientific than this. If you propose that a belief X leads to action Y then you should see increased correlation of X and Y in measured samples above controls.

Say for example you propose that Guardianism leads to sexual equality. Then you could plot the average womans earnings divided by the average mans over all periods of history when Guardianism experienced growths in popularity. These should all tend to one faster than otherwise similar non Guardianism (control) regions.

However correlation does not imply cause. There could be hidden factors that both promote sexual equality and Guardianism or sexual equality could itself promote Guardianism. This is not contentious for nice things like sexual equality. The issue comes when correlating nasty things like assault. If this is correlated with Guardianism more than controls (it is in this example) you have three outcomes:

1/ Guardianism causes increased likelihood of assaults

2/ Some unknown factor causes an increase in likelihood of both assaults and Guardianism

3/ Doing assaults makes people prone to Guardianism

generalizing this now. For a case of a proven correlation between a bad thing and a belief you get these three possible outcomes.

1/ The belief makes you more likely to behave bad. Not having that belief makes you more likely to behave well

2/ Some factor causes you to be more likely to behave bad and to believe. Relatedly a lack of this factor makes you more likely not to behave badly or believe

3/ Behaving bad makes you more likely to become a believer. Conversely behaving good makes you less likely to become a believer

The long and the short of this is that as far as I can see if believers behave worse than others you simply can't isolate their common factor and say its not related. No matter how much you would prefer it was not related. The relation might not be obvious but it exists in one of the three non flattering forms above.

I am of course assuming the quoted correlation actually exists. I would hazard a guess that the way to prove the vast majority of quoted social network "facts" wrong would be to just take a moment and look into them. I sort of sidestepped this zeroth check before for two reasons. Firstly if you are naive enough to believe everything you read on social media without checking then you're a lost case. Secondly if you are arguing with someone how representative a subset of a group's actions are to the whole group you're both implicitly acknowledging the existence of such actions (or more often acknowledging the increased rate) in the stated subset. I mean if someone shouts "your whole family is just as bad as your murdering brother" its a very strange rebuttal to say "don't judge me by my brother's actions" if your brother hasn't actually killed someone. It only actually makes sense if your brother is a murderer.  

Perhaps I have some logical flaw in my reasoning. If you can see it let me know and I will happily update this page. I welcome all feedback provided you understand that I won't respond to you if you aren't polite or interesting. See the contact me tab for more details.