It is shown how the highest levels of prosodic phrasing, φ-phrase and ι-phrase, are mapped to syntactic structure. The interface between the two is driven by the Match model, which requires an isomorphic correspondence between syntactic and prosodic constituents and assigns prosodic boundaries at both edges of syntactic constituents at once. When the syntactic structure is recursive, the prosodic structure is recursive as well. This perfect mapping can be disturbed by well-formedness conditions, a special kind of markedness constraint that bears on the prosodic constituents themselves. Constituents must have a head, be non-recursive, have a minimal weight, etc. In some cases (e.g., when syntactic constituents are too light to be matched by a φ-phrase), they even restructure the matching between syntax and prosody. Information structure is a further factor that influences prosody: Focus may require a different location for the nuclear accent, and givenness may have a deaccenting effect in the postnuclear region of the sentence. As a result, the phonological correlates of φ-phrase and ι-phrase include relative prominence of the prosodic constituents represented on metrical grids.