ABSTRACT
The obvious way for a scientist to study any concept is to try to measure it; for measurement, numeracy is central to the scientific method. A great advantage of
numbers is that they combine easily according to well-known rules of addition and multiplication. Consequently, if a concept occurs in several contexts, measurement of it enables the contexts to be combined. And this is exactly what is wanted in a trial. There are many causes of uncertainty-was he at the scene?, is he of the right blood type? —and overall judgement is possible only by combining all the different causes, by putting all the pieces of evidence together. As scientists, then, we should like to measure the uncertainty that is omnipresent in the trial. If this could be done, the various uncertainties could be combined by the rules of arithmetic to reach a final uncertainty concerning guilt embracing all the evidence. So the question is: how can uncertainty be measured?
