A review will be presented of recent theoretical and empirical results on the dynamics of temporal patterning in speech production. Empirical results are from a series of phase-resetting (e.g., Winfree, 1980) studies in which downward-directed mechanical perturbations are applied to the lower lip during the repetitive, rhythmic utterance /...paepaepae.../. These data support the existence of a central "clock" for these utterances that is bidirectionally coupled to the articulatory periphery. The theoretical work is focused on a bidirectionally coupled, two component model of speech dynamics, in which a recurrent/sequential neural network is used to control the timing of an utterance's gestural activations and a task-dynamic model is used to recruit and coordinate the articulators to produce these gestures. Emphasis is placed on the nature of the interaction between central timing processes and the evolving articulatory state.