This dissertation investigates direct object scrambling in Ukrainian. Examination of various syntactic structures has demonstrated that Ukrainian speakers observe semantic/pragmatic constraints on scrambling insofar as only partitive/definite objects are shifted. However, it appears that the syntactic movement is applied only optionally given that objects can remain in situ even when appropriate semantic/pragmatic conditions for movement are met. This research investigates the hypothesis that what underlies scrambling is an obligatory grammatical process, but one that may be expressed in at least one of two ways: by syntactic movement or by prosodic (re)contouring. Apparent `optionality' of scrambling thus results from its simply being one of the means available to speakers for achieving the same end. This hypothesis is supported with novel data on scrambling and prosody collected in three experiments: elicited production of scrambled structures by children and adults, elicited production of prosodic contours by adults, and prosodic analysis of various syntactic structures by children.The dissertation contributes to modern theoretical linguistics by developing and testing an innovative approach to scrambling, an area of intense investigation in the field for over 40 years. It also constitutes the first comprehensive study in a generative framework focusing on Ukrainian syntax and semantics. In the area of language acquisition, it provides new experimental evidence probing main factors contributing to the acquisition of word order in Ukrainian.