Or, Why It’s Better To Be 50% Right Than 100% Wrong.
Consider some complicated, thorny issue. Poverty in the developing world, say, or cancer, obesity, global warming… People think about the problem, and eventually, someone comes along and says, “Aha! Look at this evidence here! X causes poverty/cancer/obesity. Let’s stop doing X.”
And then, the others reply, “That can’t be it. The issue is so complex. You’re missing the bigger picture. There’s a broader context… there are lots of other things involved… you’re oversimplifying… ”
Now, all that sounds very Wise. I suspect it does for the usual reasons – the appearance of neutrality is generally considered Wise. Instead of saying, “Yes, X is bad”, or “No, X is actually fine, for reasons ABC”, you don’t actually pass judgment on X. Instead, you make vague statements about “complexity”.
However, unlike “X is bad” or “X is good”, the statement “the issue is complex” doesn’t actually improve our mental model of the world. It doesn’t make some specific hypothesis more likely, or some other hypothesis less likely. Indeed, it is a fully general counterargument that can be used against any X.
Now, it is true that, in any real-world scenario, no one cause is responsible for every effect. But we already know that – there’s no reason to keep repeating it, as if it’s some great new discovery in every instance. It’s also true that most of the effects tend to come from a small number of causes – the classic Pareto principle, where 90% of the results come from 10% of the causes. We do often see one X that, even if it doesn’t cause everything, is still responsible for the majority (sometimes the large majority) of a problem.
If the claim about X is right, we still won’t have perfect knowledge of reality, as no model is perfect. However, a model of “X mostly causes Y, with lots of still-mysterious details” is a lot better than “we ain’t got no clue”. It isn’t perfect, but perfection is both unnecessary and impossible; progress is better than stagnation, and not all greys are the same shade. The one who says “X is important… but actually, it turns out Y is more important in this context… but Z is still present if you rule out X and Y…” is moving forward in understanding, while the one who says “there is no one solution, we must look at the problem holistically” is not, for such sentences are merely repeated endlessly.
Seems sort of like an elaboration of the fallacy of the single cause. That is, when you propose a solution it is argued that it won’t solve the entire problem and thus must not be the one single cause we’re looking for.
Related: Marvin Minsky liked to rant about people who try to pass off “emergence” as an explanation, instead of trying to understanding the actual mechanisms.
“But we already know that – there’s no reason to keep repeating it”
Well, some people do. But I don’t think everyone does: indeed, it seems like quite often somebody comes up with a Great One Principle That Explains Everything. (It’s common enough for xkcd to parody it.) For the sake of those people, it probably is a good idea to keep repeating it – though of course just saying that “things are complex” isn’t a very good explanation, examples of past situations where something seemed complex are better.
There’s a whole literature on this issue, Phillip Tetlock’s “foxes and hedgehogs” literature.
Phillip Tetlock wrote a whole book and some papers about this issue, showing that ‘hedgehog’ analysts, who use one big principle, are common and make worse predictions than ‘fox’ analysts, who are also common and use a variety of principles.
People really do differ dramatically in how complicated they think issues are.