The epistemological foam

Here's a simpler way (I hope) of explaining why most scientific findings are wrong, but science is still right.

The terrain of science is not knowledge, but ignorance. Scientists spend little time with what they already know, except perhaps in the classroom. They are obssessed with what they don't know.

I may call the advancing edge of science -- in every direction -- the epistemological foam. Here is where scientists dredge through mountains of observations for sparks of concordance; where they conjecture, dispute, and squabble; build long and fragile chains of inference; criticize each other's equipment, methods, analysis, and conclusions; clobber each other with contradictory theories and apparently inconsistent observations. Here, instead of knowledge, is the foam of uncertainty, a whirling boil of beliefs that are constantly merging, dividing, dissolving, devouring one another.

Start to move back from the edge, and the foam becomes less active. The bubbles of belief grow fewer, larger, more stable. At last they dissolve into a single fluid, at first turbulent, then rippled, then placid. Here is the Lagoon of Knowledge, whose warm and perfumed waters make us feel langorous and happy. (Some of us anyway. There are those afraid to swim who huddle on the icebergs of faith. Sorry, my metaphors are getting out of hand.)

Swim back out into the foam, and the skin begins to tingle. It's exhilerating, energizing, but also uncomfortable, stinging us with doubt, dissatisfaction, confusion. Out there on the agonizing edge, the foam dissolves truth out of the rock face of the universe, which distills out of the foam and flows back into the ever-growing lagoon.

Wow. Sorry about that. But I hope it makes the point. There is a great deal that we know, far more than we knew even a decade ago, incomprehensibly more than the biblical scribes could ever have imagined. Nevertheless, where science is most active, we know the least, and we are usually wrong.

BTW: One way of putting Ioannidis's argument is that the domain of ignorance is unbounded. Most scientific inferences nowadays are based on statistical reasoning, and chance playing over a boundless terrain inevitably creates spurious appearances of concordance. Of course there are other kinds of discovery, that are more certain but usually less grand. For example, send a robot to Mars, take a picture. Yep, that's really what it looks like. But that barely qualifies as a scientific discovery.