Michael Licona, in the course of an interview discussing his 2010 book The Resurrection of Jesus, makes a number of helpful points about historical investigation.  He asserts, for example, that alleged miraculous events do not lie outside the provenance of historical enquiry: any past event that has left some kind of trace can be studied.

Drawing on the work of McCullagh, Licona outlines the following criteria for a successful historical hypothesis:-

• explanatory scope (i.e. successfully accounts for more of the data than competing hypotheses do)
• explanatory power
• plausibility (i.e. consistent with well-established facts)
• less ad hoc (i.e. relying on fewer non-evidenced assumptions than competing hypotheses)
• illumination

The process of establishing a historical hypothesis, says Licona, is very similar to that for establishing a medical diagnosis.  Consider the following illustration:-

A fifteen-year-old young man is not feeling well, and so he goes to see his GP/family doctor.  He describes his symptoms as: fever, vomiting, and pain in the lower abdomen.

The doctor has three medical students with him, and he invites them to consider what the young man’s diagnosis might be.

The first student suggests the ‘flu, since a fever would be the most common symptom of the ‘flu.  But the doctor explains that while this is correct, vomiting and abdominal pain are not usual symptoms of ‘flu.  This diagnosis lacks ‘explanatory scope’, since it cannot explain all the symptoms.

The second student comments that even though vomiting and abdominal pain are rare in ‘flu, it is still possible that this is the correct diagnosis.  The physician agrees, but says that if another diagnosis is available that can more readily account for these additional symptoms, the the ‘flu diagnosis would lack ‘explanatory power’, because you would be forcing the symptoms to fit the diagnosis.  The physician adds that in all his years as a medical practitioner, he has never seen a case of ‘flu that had this triad of symptoms: thus the diagnosis lack ‘plausibility’.

The third student applies her imagination to the case, by proposing that the boy does indeed have ‘flu, and supports this with the observation that it was the middle of the ‘flue season – thus increasing the plausibility factor.  Perhaps, suggests this student, the other symptoms are unrelated to ‘flu: the abdominal pain perhaps being cause by a recent sporting injury and the vomiting to food poisoning.  The physician remarks that the weakness here is that two non-evidence-based assumptions have been introduced.  The argument is therefore ‘ad hoc’.

The experienced doctor then explains that the triad of symptoms describe a classic case of appendicitis.  On the basis of this, he recommends the appropriate treatment.

It is important to note that the diagnosis of ‘flu is not impossible.  But the physician is going to diagnose and treat on the basis of the most likely explanation of the symptoms.

The point is this: the same kind of reasoning applies to historical investigation.  Historians are rarely able to demonstrate that a particular hypothesis is either impossible or absolutely certain.  But they can (and should) review the evidence in order to find the best possible explanation.  They are then in a position to state with what degree of certainty a given event did, or did not, occur.