Why historical science is the best kind of science
There are sciences, such as physics and chemistry, where we can perform experiments. There are other sciences, such as the science of planetary motion (and astronomy in general) where we cannot do this, but we can still carry out repeated observations in well-controlled circumstances, and devise theories with whose help we can make definite predictions. All of these are what I will call rule-seeking sciences. At the other extreme, we have sciences such as palaeontology and much of geology, which one might call historical sciences. With these, the aim is not so much to establish general rules, as to unravel and explain the specifics of what happened in the past. It is usual to regard the rule-seeking sciences as the most rigorous, to which the others should defer. This shows a deep misunderstanding of how science works, and, time and time again, when historical and rule-seeking sciences have come into conflict, it is historical science that has triumphed.
A few examples. The rocks clearly show (for detailed arguments, see e.g. here) that the Earth, and by implication the Sun, must be at least many tens of millions of years old. Lord Kelvin, the leading physicist of his time, argued that this was impossible, because using all sources of energy then known, the Sun could not have kept going for more than some twenty million years. There was nothing wrong with Kelvin’s reasoning, apart from the small fact that there was a massive source of energy (nuclear fusion) of which he had no inkling.
The fossil record shows catastrophic discontinuities, most famously the sudden disappearance of the [non-avian] dinosaurs, for which there was no physical explanation at the time when they were discovered.
The close similarities in geology and in the kinds of fossils found on both sides of the South Atlantic, and even (as Francis Bacon observed 400 years ago) the way their coastlines fit together like jigsaw pieces, show that South America and Africa were formerly joined together. Yet geologists long resisted this conclusion, because they could not see how the continents could have moved through the solid ocean floor, and had not realised that the ocean floor itself is capable of moving.
The Ice Ages were well established from the historical record, because of such clear physical traces as scratches on rocks, transported boulders, and the ways that vanished glaciers had reshaped valleys, long before they were explained in terms of subtle shifts in the Earth’s rotational axis.
What about reproducibility, prediction-making, and testing against observation, traditional hallmarks of good science?
All we need to be able to reproduce is our observations, not necessarily the event that caused them. We cannot duplicate the asteroid impact that killed the dinosaurs, but we can duplicate the observations from which we infer that it occurred. We cannot duplicate the formation of the Cretaceous limestones of Europe and North America, but we can repeatedly confirm that they contain similar microfossils, showing them to be of the same age. And when we speak of prediction-making in science, we are using the word “prediction” rather loosely, to include relevant information about the past. Thus when William Halley used Newton’s physics to work out the trajectory of the comet that bears his name, he “predicted” that the comet would have appeared previously around 1531 and 1607, in accord with recorded observation. As for observational testing, it is as true for historical science as for any other kind of science that when our ideas are contradicted by new results, we need to modify them. For example, there were sceptical palaeontologists who maintained, on the basis of radiometric dating, that some dinosaurs had survived the asteroid impact, but more accurate dating of the relevant strata showed that this was not so.
Historical science also draws together different kinds of observations, linked by their common cause. It is interesting that we have found shattered rocks, and evidence of tsunamis and wildfires, all occurring in the same layer of sediment laid down some 66 million years ago. It is even more interesting that this layer carries excess iridium (the calling card of an impacting asteroid), and matches the death date of the dinosaurs.
Rule-making science can seem impersonal and remote, because it excludes the contingent, the way things just happened to be, and it is just such contingencies that dominate our own lives. Historical sciences, on the other hand, explore the sometimes dramatic implications of contingencies, such as the fact that the asteroid discussed above happened to land in a shallow coastal sea, rather than the deep ocean where it would have done far less harm.
All of which stands on its head the common presumption that deep time geology is inferior because it is merely a historical science. Not only is it in many ways more interesting; it is also more certain.
I thank @Carol Cleland and @Craig Jones (both U. Colorado, Boulder), @Maarten Boudry (Ghent), @Michael Fugate (UC Riverside) and @Tom Scharle for useful discussions. Grand Canyon image via lovetheseptics [sic]. Pangaea map from UK coolgeography A-level materials. Hadrosaur image(Creative commons), by Debivort via Wikimedia
