On Causation and Probability

Since this is my final (or maybe my second to last, I have to count them up soon) blog post, I decided it was worth it to think about some of the larger philosophical topics that have kind of stayed behind the scenes in the course. The first is the issue of causation. We talked a lot about the issue of inferring causation versus simple correlation due to homophily effects, but we never actually defined what causation means. In traditional philosophy, A causes B if and only if A occurring always results in B. By this definition, almost none of what we've discussed in the course is actual causation since much of it is based upon 50 (or worse) chances. So is causation really the correct word to use when discussing topics like the ones in the course?

It is probability, really, that throws a thorn in the side of this type of causative reasoning. Consider the flipping of a coin. Half the time you get a heads-up coin, half the time you get a tails-up coin. Does flipping the coin cause a heads or cause a tails? The answer, of course, is that it causes neither. It causes the coin to fly through the air and land one side up or another.

Can this line of reasoning be applied to social sciences? Perhaps it is worse to think of obesity as causing obesity in networks as it is to think of obesity as causing a potential for increased obesity. After all, if obesity in person A always resulted in obesity in their contacts B, C, and D, the obesity epidemic would be far, far worse than it is today. It may be best to think of friends with obesity (or friends that smoke, or are anorexic) as risk factors within networks, rather than agents for contagion spread. While the distinction may be small, at least philosophically it's an important one to make.