Abstraction in models raises further questions. A climate change model, for example, may abstract from other factors such as change in albedo or changes in solar radiation. Whether these abstractions are sound will depend on whether significant variations in factors such as these may be expected during the period covered by the model. Satellite data might provide evidence on the degree of variation of solar radiation over decades. Scientists might have reason to suppose that the albedo of the Earth will be more or less constant, with increases in cloud cover balanced by decreases in the extent of ice sheets.

Clearly, climate models will not produce certain results. A cumulation of evidence from models that vary one factor or another may improve certainty overall. But this brings us back to the first issue: what is appropriate evidence in prudential reasoning may very well fall short of evidence establishing a model as beyond question in the domain of theoretical enquiry. It may also fall short of the lesser standard of beyond reasonable doubt for practical purposes that would support a conviction in a criminal trial. However, it may be sufficient for prudential reasoning and decisive for precautionary reasoning. In the absence of any evidence that we should take precautions against CO2 reduction rather than doing nothing, the standards of prudential reasoning will be sufficient.

What drives contests on the practical implications of scientific discoveries? Some Hyper-sceptics of global warming suggest that scientific clubbiness and desire for reseach funding can only explain the better than even chance that new climate studies will confirm

The cases of ‘Hyper-scepticism' considered here involve misplaced standards of doubt, coupled with greater or lesser misuse of scientific evidence and reasoning. Clarity on professional responsibility in theoretical and practical domains and clarity on standards of rational evidence would provide less room for opposition based on Hyper-scepticism.