The Historical Sciences are Testable
by Keith Miller
Prepared for the Kansas “Science Hearings”, May 2005
Frequent claims are made that the historical sciences (cosmology, astronomy, geology, evolutionary biology, anthropology, archaeology) deal with unrepeatable events and are therefore not experimental. Furthermore, because past events and processes are not directly observable, theories of origins are deemed inferior or less certain than studies of present processes. This view commonly finds expression in statements like: “No one was there so we can never know what really happened.” This view is false. The historical sciences are no less scientific, or testable, than the “hard sciences.”
Predictions made by hypotheses in these fields are continually being tested by new observations. Predictions are tested against each new observation or analysis. Obtaining data from a newly analyzed sample or newly described locality is no different methodologically than obtaining data from a new experimental trial. In both cases, the new observations can be tested against expectations based on previous experience and theoretical predictions. If the predictions deduced from a hypothesis are not supported by new observations then that hypothesis is modified or rejected. Scientific research proceeds by an almost continual process of hypothesis creation and testing. Many past theories in the historical sciences have been discarded with the accumulation of new observations and the development of new theories of greater explanatory power.
Like all scientific disciplines, geology and paleontology proceed by testing the predictions of existing models and theories. Predictions are tested against each new observation or analysis. Obtaining data from a newly analyzed sample or newly described locality is no different methodologically than obtaining data from a new experimental trial. In both cases, the new observations can be tested against expectations based on previous experience and theoretical predictions.
In geology, for example, the measurement and description of each new rock outcrop or subsurface core is a test of working hypotheses based on present understanding. For example, if a specific rock unit is interpreted to be part of a meandering river system, then specific predictions can be made concerning the geometry of this rock body and the characteristics and distribution of associated sedimentary rocks. In modern meandering river systems a whole complex of sedimentary environments are present: channel and point bar deposits, levees, crevasse splays, overbank flood deposits, abandoned channels, freshwater lakes, etc. Each of these environments has its characteristic spatial relationships, sediment types, depositional features, and associated biota. If the original hypothesis of a meandering river system was correct, then further exploration and sampling of the area should reveal the predicted geologic features and their predicted spatial and temporal relationships. If the new observations are contrary to these predictions, then the hypothesis must be modified, or if necessary, abandoned.