Current international and EU environmental policies increasingly promote the participation of non-state actors in environmental governance as a means to improve policy effectiveness and implementation. To date, six EU directives have been adopted that contain provisions on public participation in environmental decisions, including the 2000 Water Framework Directive and the 2003 Public Participation Directive. Those calls for democratic innovations are legally non-binding, i.e. are EU soft law. Plausibly, EU initiatives promoting participatory institutions are expressions of meta-governance. Yet how effective is the EU in meta-governing through legally non-binding soft law? The empirical evidence is surprising: Participatory requirements challenge existing policy networks, historically grown administrative cultures, and predominant views on the merits of participation (democratic innovation vs. technocratic stakeholder management tool); at the same time, soft law is unlikely to enforce policy change. Curiously, however, some member states have indeed changed their modes of water governance in response to the two directives. Why? So far, processes of top-down activation (as opposed to bottom-up activism) have received moderate scholarly attention only. Based on research carried out in a larger EU-funded project, this paper compares systematically the emergence and sustenance of, and resistance to, participation in three EU countries: the UK, Denmark and the Czech Republic. Implementation patterns are counter-intuitive (the CR as good, the UK as moderate, and Denmark – despite its long collaborative and corporatist traditions – as non-implementer). When democratic innovations are introduced, member states do so in response to perceived democratic and legitimacy deficits, despite the EU’s focus on policy effectiveness, yet the practical operationalization of participation through regulatory agencies is governed by policy output considerations, resulting in largely disputed, technocratic, post-political hybrids which neither enhance democracy nor policy delivery. To explain these outcomes, I test three theories, taken from organisational studies, EU policy implementation research, and discourse analysis.