Categorization has a large impact on how people perceive the world, especially when used to make inferences about uncertain features of new objects. While making these inferences, people tend to draw information from only one possible categorization of a new object; in addition, people are sensitive to pre-existing correlations between features. Here, we explain these trends of feature inference using a priming-based cognitive process model, and show that our model is distinguished in that it can explain not only these two main trends, but also cases where people seem to reverse the first trend and base inferences on information from multiple categories.