Technical Abstract: State-and-transition models are increasingly being used to describe the drivers of change and constraints to restoration in rangeland ecosystems. These models are beginning to serve as the conceptual foundation for interpreting and managing rangeland behavior. Presently, hundreds (eventually thousands) of models can be produced that are tailored to particular soils in different regional settings. These models emphasize a variety of interacting ecological processes. Why and how should we produce such models? We address the first part of that question by arguing that state-and-transition models are an opportunity to reformulate how we conceptualize rangeland behavior. They facilitate the integration of evolving concepts in various disciplines with rangeland management, such as stability, scale, geomorphology, multiple interacting drivers, soil quality, and spatial effects. The model structure induces model builders and model interpreters to think broadly and critically about why they think the way they do about various aspects of rangelands. We address how models can be built by presenting several examples of models that vary in complexity. In doing so, we discuss the implications of designating communities and states in models, accounting for varying scales of pattern in vegetation and soils, interpreting the presence of plant communities on different soils, and dealing with our uncertainty about how those communities were assembled and how they will change in the future.