Thursday, September 25, 2008

Healer's Fallacy

I want to look at a particular type of causal fallacy, which we might lump under "magical thinking," which I have dubbed the "healer's fallacy." I really can't think of a better name, though I'm open to suggestions. Let me explain it first. Since the body tends to recover on it's own from most minor illnesses, it becomes easy to confuse things that actually cure the disease with things that do nothing or only harm you slightly. For example, if you get a cold, then most likely you'll recover in a week or so. So, if you decided during that week to consume twice a day a shot of vodka mixed with mustard and Worcestershire sauce, you might be led to believe that it did some good. If you took a daily dose of homeopathic medicine, you might be led to the same conclusion. Maybe it might do some good as a placebo effect, but even if there is no placebo effect, it can sill appear like it cured your ailment. This can lead to confusions about what treatments are effective in the absence of carefully constructed experiments and propagate the myth of apparently effective home remedies. Thus, I would define the "healer's fallacy" as "a causal fallacy directly resulting from apparently affecting something that would happen anyways on its own." I use "on its own" very informally and broadly. For example, another good example, would be an orgonite cloudbuster (many videos on youtube of these things). The cloudbuster is basically a solid block of solidified chemical resin, containing crystals and metal shavings with some long copper tubes sticking out, that can apparently break up clouds when placed on the ground beneath them. Of course, the thing about clouds is that they are very fleeting and unstable and given time will always dissipate "on their own" (informally speaking). Thus, with a little patience, the cloudbuster will always seem to work, apparently confirming its effectiveness.

Economics is also prone to this healer's fallacy because economies tend to grow and improve as well. There are two reasons for this: 1) because two people will usually only engage in voluntary exchange if they both believe they are benefiting. The cumulative effect of lots of voluntary exchange is greater benefit for everybody. Even if people are sometimes mistaken about what will benefit them or sometimes willingly make sacrifices for the benefit of others, the economy will still grow because a) the cumulative effect of many exchanges is towards overall greater benefit and b) those who most often make exchanges which most benefit themselves will tend to grow richer and thereby become a larger part of the overall economy. The other reason economies grow is 2) because people tend to try to improve their situation. When they do this in the context of voluntary exchange, the general result is people symbiotically figuring out how to produce more with their finite allotment of time. They thereby earn more money with which to purchase more goods and services and also produce those goods and services in ways that require less money. Even those who don't try to improve their situation benefit as prices go down, since they can afford to buy more with the same amount of money, and thus can increase their real wealth.

This is important because it can seem like policies are improving the economy, when they have no effect or even a negative effect. The economy is like a boulder rolling down a hill. Attempts to speed it up are difficult and have little effect, whereas most policies end up simply getting in the way and slowing it down. The metaphors usually used are that the economy is like a car that needs to fueled, that needs to be "jump-started" when it slows, lest it stops, or can "run out of gas." But the economy never stops: people will exchange so long as they are able. The great impetus of the Industrial Revolution was not a positive push, but the removal of barriers. It is no coincidence that the Industrial Revolution began in a country with a very recent history of many philosophers advocating political philosophies of freedom.

This tendency of economies leads to certain epistemic problems when trying to understand economies, since it can be hard to isolate when something is helping it or preventing for growing as fast as it could. This is why economists usually look at how things change: what were conditions like before a policy went into place compared to what they were afterward. But this still limits the empirical breadth of economy and makes empirical evidence difficult.

Hume observed long ago about how we can't really observe causality, only constant correlation. What causality implies is universal correlation: that a will always lead to b. But we can never observe "always" only "all times in my experience." Thus, causal fallacies are always a threat when we attempt to understand the world and the type of magical thinking that leads to the healer's fallacy is an understandable, especially concerning certain phenomena.

2 comments:

Theo Clark said...

Hi Joe,

This sounds like a particular version of the post hoc ergo propter hoc fallacy - the regression fallacy. I.e. things such as health and behaviour (as two examples) naturally fluctuate about a mean and when they move far away from the mean, say with a sore back or particularly delinquent behaviour, eventually they will naturally regress back towards it, with or without "intervention" (either that or you die or end up in prison...). But when you go to a chiropractor or punish someone, you assume it was the intervention that did it, rather than things just going back to normal. Hence the need for double blind trials to sort out real effects from apparent ones.

Joe said...

Hello Theo, I was aiming to describe something that's a very particular type of post hoc ergo propter hoc fallacy. Or it might be appropriate just to call it a type of regression fallacy, which I didn't know about before. I just wanted to emphasize how sometimes things get better on their own and sometimes people try to take credit for it by claiming they made it better. Whether this phenomenon need its own name, I don't know.