Scholars frequently distinguish between “hard” and “soft” forms of judicial review: “hard” review gives courts the final say on constitutional questions, whereas “soft” review attempts to strike a balance between judicial and legislative supremacy by giving the legislature the ability to override or set aside what the courts have done. However, the binary hard/soft distinction does not convey the full spectrum of possibilities between the extremes of legislative and judicial supremacy. On the one hand, so-called “hard” review can in practice operate like soft review because, even in systems of nominally hard review, the government usually retains the ability to effectively override the courts via constitutional amendment. On the other hand, courts have increasingly turned to options that are harder than “hard” review. A small but growing number of courts—the Taiwanese Constitutional Court among them—have claimed the power to invalidate constitutional amendments as unconstitutional. Yet even judicial review of constitutional amendments does not exhaust the options at their disposal. Analysis of Taiwan’s experience with judicial review over the last thirty years holds a number of lessons about judicial power and constitutional politics. First, “soft” review, “hard” review, and even review of constitutional amendments are merely intermediate points on a spectrum of judicial power that ranges from no judicial review to the judicial initiation of constitutional amendments. Moreover, the spectrum itself is dynamic and prone to expansion. In the face of new challenges, courts find new ways of asserting themselves, and in so doing, they continually remake the frontiers of judicial power. Second, there is no point on this spectrum where a court—or any other institution—is guaranteed the last word. Whenever courts assert themselves, other institutions have the ability to respond, and vice versa. Different responses are available at different points on the spectrum, but a response always exists. It is a defining characteristic of politics that, for every move, there exists a countermove. Finality is as elusive in constitutional politics as in any other form of politics. Third, the celebration of “dialogic” judicial review—in which constitutional questions are resolved via interaction between courts and legislatures rather than by judicial fiat—is problematic because, in some sense, all judicial review is “dialogic”. The fact that courts are never guaranteed the last word means that every plausible institutional configuration can be characterized as “dialogic” in the sense of allowing or inviting some degree of interaction, or “dialogue,” between the courts and the legislature. Finally, a court’s position on the spectrum of judicial power is not necessarily fixed by the constitution and can instead be highly fluid. As the recent history of the Taiwanese Constitutional Court demonstrates, it is possible for a court to traverse the entirety of the spectrum in a relatively short period of time without any corresponding change in the constitution itself. Major shifts in judicial power can occur without formal constitutional change because judicial power is at least as much a function of the political and institutional context as of the constitutional text.