A brief overview of the nature of modeling in science generally, and cognitive science in particular, is provided. Given some phenomenon we wish to understand, a model is another structure which is relevantly similar but better understood. Models are standardly abstract state-determined systems; these divided into two kinds, the computational and dynamical. Connectionist models are a special class of dynamical models (high-dimensional, homogeneous, neural). A useful alternative to the standard opposition between mainstream computational cognitive science and connectionism is to see cognitive scientists as either computationalist or dynamicist in general orientation. Connectionists straddle this divide even though their models are dynamical systems.