Opponents of biotechnology ofteninvoke the Precautionary Principle to advancetheir cause, whereas biotech enthusiasts preferto appeal to ``sound science.'' Publicauthorities are still groping for a usefuldefinition. A crucial issue in this debate isthe distribution of the burden of proof amongthe parties favoring and opposing certaintechnological developments. Indeed, the debateon the significance and scope of thePrecautionary Principle can be fruitfullyre-framed as a debate on the proper division ofburdens of proof. In this article, we attemptto arrive at a more refined way of thinkingabout this problem in order to escape from theexisting polarization of views between ``guiltyuntil proven innocent'' and ``innocent untilproven guilty.'' This way of thinking alsoenables a critical review of currentdemarcations between risk assessment and riskmanagement, or science and politics, and of themorally laden controversy on the relativeimportance of type-I and type-II errors instatistical testing.