One of the ways in which spoken languages differ from each other is in the temporal patterning of speech. These temporal patterns can be examined on several time scales, from very local (subsegmental and segmental) to global. The global patterns lasting from a quarter second to a second or so are most noticeable when we listen to speech in any language but our own. We may also become aware of rhythmic patterns in our own language when we hear it spoken by someone whose native language is something else. Presumably there is some kind of pattern in the timing of speech gestures that accounts for our perception of rhythmic differences between languages. This article examines the problem of speech patterns in time, two conceptual frames for temporal patterns in speech (symbol strings and cycles), global timing constraints (“stress-timed” vs. “syllable-timed” languages), regular “mora timing” in Japanese, harmonic timing effect, modelling rhythm with dynamical systems, and phrase edge timing phenomena.