This paper focuses on linguistic prosodic units related to boundaries between syntactic units. Specifically, rules for predicting the location of such boundaries, and factors affecting their location, are discussed. Examples are given on how prosodies can be used for syntactic analysis. Addressing the question of prosodic units and their distribution, two theories, both based on a hierarchy of units, are contrasted. A third theory, suggested as a possible basis for further refinement and testing is stated as follows: an NP that is not a single unstressed pronoun ends with a phrase- or clause-level boundary; NP's with embedded clauses have boundaries before the clause. A study conducted to test this rule is described, and several flaws in the rule are pointed out. Areas in the study of prosody that need further research are pinpointed. (AM)