Bad Apples, Picked Cherries and Trees in Forests….
One of the great challenges earth scientists face is knowing when observations refute a hypothesis, when they support a hypothesis and when they are nearly irrelevant. For instance, Alfred Wegener’s argument for continental drift made prominent mention of his geodetic measurements of the opening of the Atlantic. These were quickly found to be in error, throwing considerable cold water on continental drift even though many other arguments that Wegener made were quite sound. Obviously in this case it was a modification of the theory that was required, not outright rejection. This is a case of one bad apple that doesn’t spoil the whole barrel.
On the other hand, sometimes a theory is presented that explains one piece of evidence (usually the evidence that author collected) but ignores other information (several mass extinction hypotheses seem to fall into this category). While this happens in the scientific community, the most extreme cherry-picking comes from outsiders who find some curiosity that, to them, overturns all of earth science (creation science museums are full of such trivialities).
While lab science is often akin to a U.S. criminal trial (where guilt must be proven beyond a reasonable doubt), field science is more like a civil trial (where preponderance of evidence determines a victor). This is because there are a host of other variables in play. In fact it is quite rare for a single contrary observation to bring down a theory that has any real breadth.
Consider the “flat slab” hypothesis for the Laramide orogeny that created the Southern Rocky Mountains between about 75 and 45 million years ago. As originally laid out by 1978 (in terms of creating the stresses to create the mountains), the shear stress from the subducting Farallon plate as it scraped the underside of North America was balanced by high normal stresses far inland, thus making mountains far from the plate boundary. (This is kind of like the pileup at a grocery store conveyer belt at the bagger’s end when there is no bagger–laying a long piece of thin foam rubber on the conveyor would fold up near the bagger’s end). A simple and elegant explanation. But when you put in values that look to make the right predictions for the stresses in the Rockies, you find that the mantle lithosphere (and indeed much of the lower crust) should have been stripped out from much of the southwestern U.S. When Peter Bird presented these calculations in 1988, this amounted to a prediction of what should be found. This is precisely what you want scientifically: you develop a hypothesis and allow it to predict something you don’t yet know.
Geochemists were pretty quick to respond to the challenge: the ancient mantle lithosphere is still there and was there during the Laramide orogeny (the main paper was Livaccari and Perry’s in 1993, but others have made similar and occasionally stronger assertions). Basically the problem is that to get the stresses high enough in the Rockies, you need high stresses at the flat plate interface. But those same high stresses should cause the mantle lithosphere to get sheared out and carried east. Unlike the bad geodetic rate across the Atlantic, which really only affected the timescale of continental drift and not its existence, this is dead center of the flat slab hypothesis and something you have to solve. So the theory is dead, right?
Er, no. And it was never fixed, either (is it fixable? Maybe; perhaps you could argue for a stronger rheology in the mantle lithosphere, or lower stresses in the Rockies. But the argument hasn’t been made). At present there is no physical model of flat slab subduction producing the stresses necessary to drive deformation in Colorado and eastern Wyoming that doesn’t remove the mantle lithosphere. Arguably most of the advocates for a flat-slab origin for the Rockies have walked away from that elegant basal shear explanation in favor of a rather mysterious lower crustal stress guide or the presence of a near-rigid Colorado Plateau, the details of which are never spelled out.
How could this be? Why does the geological community embrace with little disagreement a theory that seems to have failed at a most fundamental level?
The simple answer goes back to that analogy of a civil trial: the opponent’s case is weaker still. The only two substantial alternatives (before a paper GG wrote a few years back; we’ll leave the jury out on that for now) seemed more deeply flawed: a model where British Columbia collided with California and one where extensional collapse of thick crust in Nevada, Utah, and Arizona drove shortening to the east. So although a flat slab may have a problem in reconciling the physics, at least it seems to qualitatively explain the temporal relationships in the western U.S. Basically, geologists abhor a (theory) vacuum more than Mother Nature does, and the flat slab continues to dominate thought on the Laramide despite this crucial failing.
This is part of why geologists like multiple working hypotheses. Since there will always be some noise from stuff we don’t care about, most of the time the best we can do is say which idea works better with the observations that matter.
Unfortunately, sometimes we don’t agree on what the important observations are…which we can discuss another time…