ÖйúAƬ

David Balding is wrong to suppose that Bayesian probability theory presents the only satisfactory method for reasoning about factual uncertainties in courts of law (THES, October 24). This is like assuming that the only satisfactory method for reasoning about space is Euclidean geometry.

Just as there are non-classical geometries so too are there non-Bayesian probabilities. The utility of such non-Bayesian systems is well illustrated in J. B. Kadane and D. Schum's A Probabilistic Analysis of the Sacco and Vanzetti Evidence. And paradoxes derivable in the Bayesian analysis of courtroom reasoning about factual issues were demonstrated in my book, The Probable and the Provable.

Jonathan Cohen Emeritus fellow Queen's College, Oxford