The term “interaction” describes situations in which one predictor’s influence on a linguistic variable is conditioned on another predictor. Informally, we can speak of interactions as cases where predictors are “more than the sum of their parts”. Mathematically, interactions are multiplicative rather than additive effects. Interactions are extremely common in linguistic data, and they are often theoretically very interesting. However, unfortunately, interactions make the interpretation of one’s models much more difficult, and many linguists commonly misinterpret their models in the presence of interactions. To combat this, the chapter walks the reader through numerous examples of interactions. Example analyses include datasets of a conceptual metaphor experiment, as well as a study on iconicity (how much a word sounds like what it means).