What would be sufficient to show of some apparently higher-level property that it is ‘nothing over and above’ some complex configuration of more basic properties? This paper defends a new method for justifying reductions by demonstrating its comparative advantages over two methods recently defended in the literature. Unlike its rivals, what I’ll call “the semantic method” makes a reduction’s truth epistemically transparent without relying on conceptual analyses.