Firstly, Popper displayed a great disliking for the traditional fact-based approach to science as exemplified in the work of Francis Bacon. In Bacon’s approach, scientific hypotheses are derived from collected data through inductive reasoning. Popper’s point was that induction itself is not a logical, but a psychological process, that is not vital but circumstantial to the process of generating hypotheses. While we basically agree with this position, our impression is that Popper has downplayed the practical role of induction as an often indispensable accessory to the human imagination in the formulation of new hypotheses. We have arrived at our detailed reconstruction of the internal chronology of the mythical axis mundi not through ‘wild and unfettered speculation’, but through the traditional painstaking process of data-gathering on a large scale, comparative analysis of these data and the identification of recurrent patterns by means of induction.

Secondly, our opinion is that Popper’s theory about the monopoly of falsification in scientific theory is only valid for one particular class of hypothetical statements – namely, for general statements imputing a certain quality to an entire group of phenomena. To this class belong the ‘nature laws’ of lightning, aurorae, earthquakes and so on as modelled by physicists. But there are other classes of hypothetical statements that evade Popper’s comments. Hypothetical statements concerning past or future events can neither be conclusively verified nor conclusively falsified, but must be ranked according to relative probability only. To this class of statements belongs the hypothesis of an intense auroral storm at the end of the Neolithic age. Factors that might reduce the probability of this hypothesis include evidence that a high-energy density auroral storm could not produce forms detected in myths and ancient artefacts, that human beings could not possibly have seen such forms, that a storm of this magnitude could not possibly have happened or did not happen, and so on.

Thirdly, Popper proposed to exclude from the definition of science all statements that are not falsifiable. For the reason stated above, we disagree that falsifiability of claims is the only criterion to science. 'Science' is primarily a lexical term, the definition of which must be described (by lexicologists), not prescribed (by logicians). As such, we define science in the traditional, Baconian sense as the entire intellectual process from the acquisition of ‘raw’ data through the recognition of patterns and the systematic organisation of these data (via a balanced combination of ‘splitting’ and ‘lumping’) to the formulation of general theories with the capacity to test or predict the data. Science is simply the ‘art’ of reducing a large amount of data to a smaller set of ‘laws’, ‘tendencies’ or principles that add comprehension to knowledge.

In conclusion, geophysical and plasma-physical models of earthquakes, lightning and aurorae are falsifiable exponents of hard science; our comparative analysis of traditional cosmologies and myths, partly resulting in our reconstruction of a near-universal template of creation mythology, is a falsifiable exponent of ‘soft science’; and the interdisciplinary hypothesis that many creation myths can be explained in terms of past geophysical and atmospheric events, as a theory of past events, is neither conclusively falsifiable nor conclusively verifiable, but is a historical model, of a scientific character, that must be graded relative to other theories of myth according to probability. The plasma-physical model of creation mythology and art-historical data, such as geometric petroglyphs, is simply a working hypothesis for a large set of ancient data that we consider the most convincing, the most complete and the most economic theory of the nature of traditional mythology in general.

This position is not fundamentally different from the work of archaeoastronomers who ‘reconstruct’ past observations of comets and meteor showers on the basis of textual records. In each case, the impetus to produce a ‘scientific’ model is given by historical data, including human artefacts or writings.