This dissertation examines the conditions under which China is likely to accept institutional constraints on sovereignty and ratify international treaties with mandatory dispute settlement mechanisms (DSMs). Specifically, why did China ratify the United Nations Convention on the Law of the Sea (UNCLOS) and accept legalized Bilateral Investment Treaties (BITs) beginning in the mid-1990s, but vote against the Rome Statute of the International Criminal Court (ICC) in 1998? A comparative case study shows that China delegates disputes only in non-core sovereignty issue areas, where sovereignty costs are relatively low, but not in core sovereignty issue areas, where such costs are deemed to be too high. The primary driving force for China's changes regarding non-core control rights is the evolution of its new strategic interests, which increases the material benefits and partially offsets the sovereignty costs of signing and ratifying hard laws. Meanwhile, China's acceptance of the social legitimacy of an international treaty or a boundary-trespassing norm incrementally lowers the normative premiums-the ideational dimension of sovereignty costs-of delegation, and complements the material factors in driving changes. However, as China is still in a weak socialization process and has not fully internalized or accepted the normative legitimacy of alternative boundary-trespassing norms, the normative premiums of deviating from the Westphalian sovereignty cannot be significantly reduced, and China cannot accept highly legalized treaties without making reservations to lower the sovereignty costs. Accordingly, if a treaty's legalized DSM is balanced with flexible exclusion and reservation clauses, allowing states to exempt disputes regarding core sovereignty and minimize sovereignty costs, China is more likely to ratify the treaty (as with the UNCLOS and legalized BITs). If a treaty does not allow any reservations and may have negative impacts on core sovereignty rights, China will not hesitate to reject it (as with the Rome Statute).