Argues that the probabilities invoked in Anglo=American law do not conform to the mathematics of the classical calculus of chance but have a determinate non-POascalian structure whih is closely connected with certain other patterns of human reasoning, and notably with the strucutre of inductive reasoning in experimental science.