As more complex and uncertain forms of health innovation keep emerging, scholars are increasingly voicing arguments in favour of public involvement in health innovation policy. The current conceptualization of this involvement is, however, somewhat problematic as it tends to assume that scientific facts form a "hard," indisputable core around which "soft," relative values can be attached. This paper, by giving precedence to epistemological issues, explores what there is to know from public involvement. We argue that knowledge and normative assumptions are co-constitutive of each other and pivotal to the ways in which both experts and non-experts reason about health innovations. Because knowledge and normative assumptions are different but interrelated ways of reasoning, public involvement initiatives need to emphasise deliberative processes that maximise mutual learning within and across various groups of both experts and non-experts (who, we argue, all belong to the "publics"). Hence, we believe that what researchers might wish to know from publics is how their reasoning is anchored in normative assumptions (what makes a given innovation desirable?) and in knowledge about the plausibility of their effects (are they likely to be realised?). Accordingly, one sensible goal of greater public involvement in health innovation policy would be to refine normative assumptions and make their articulation with scientific observations explicit and openly contestable. The paper concludes that we must differentiate between normative assumptions and knowledge, rather than set up a dichotomy between them or confound them.