People spontaneously gesture when studying spatial descriptions. Doing so improves comprehension and learning. Theirgestures create spatial models of the described environments. Here, we address two questions in two experiments: willpeople gesture to study descriptions that are not inherently spatial, and will people gesture when information is presentedvisually rather than text. The answers to both questions are yes. Together, the results suggest that gestures facilitatecomprehension and learning by creating spatial-motor representations that directly reflect meaning.