Slippery slope arguments—the idea that otherwise beneficial reforms should be rejected lest they beget further undesirable ones—are ubiquitous in political discourse. We provide a learning-based policy-feedback mechanism to explain why slippery slope dynamics arise. Additionally, we provide conditions under which, in equilibrium, sophisticated agents will successfully manipulate policy to either induce or prevent a slippery slope dynamic.