It's likely that the problems with statistics are inherently due to the nature of knowledge, so alternative formulations are not likely to help much.