Moral psychologists typically explain moral cognition by invoking abstract moral values like fairness and loyalty. We argue that this approach is inadequate, particularly in accounting for moral disagreements between people who share similar values. Our new paradigm addresses this problem by instead modeling moral cognition as a function of relevant nonmoral beliefs. One hypothesis stemming from this “Is→Ought” model is that relevant beliefs will predict moral judgments on controversial issues better than abstract moral values like the “moral foundations”. We found strong support for this hypothesis: Across two studies of U.S. adults (N = 361), nonmoral beliefs outperformed values in predicting moral judgments on each of eight controversial topics. When controlling for political affiliation, relevant beliefs were the only robust predictors. These findings support the view that people’s moral divisions are determined not, as presently assumed, by their values, but instead by their different views on relevant details.