Evaluative conditioning (EC) is defined as the change in the evaluation of a conditioned stimulus (CS) due to its pairing with a valenced unconditioned stimulus (US). According to propositional accounts, EC effects should be qualified by the relation between the CS and the US. Dual-process accounts suggest that relational information should qualify EC effects on explicit evaluations, whereas implicit evaluations should reflect the frequency of CS–US co-occurrences. Experiments 1 and 2 showed that, when relational information was provided before the encoding of CS–US pairings, it moderated EC effects on explicit, but not implicit, evaluations. In Experiment 3, relational information moderated EC effects on both explicit and implicit evaluations when it was provided simultaneously with CS–US pairings. Frequency of CS–US pairings had no effect on implicit evaluations. Although the results can be reconciled with both propositional and dual-process accounts, they are more parsimoniously explained by propositional accounts.