My three projects here explore some semantic and metaethical problems that are unique to normative language and our normative reasoning. Ch.1 argues that the notion of a contrary-to-duty (CTD) obligation and its role in normative discourse and reasoning are not adequately captured in the standard semantics for ought-statements, developed by Angelika Kratzer and David Lewis. I show this by presenting a new puzzle, the CTD Trilemma, using a famous example from Chisholm’s Paradox. I claim that two different roles played by ought-statements in normative discourse and reasoning have to be captured in any satisfactory semantic account for normative language. I draw the distinction between axiological and deontological ought-statements in my version of deontic update semantics and how this distinction nicely solves the puzzles about CTD obligations. Ch.2 discusses the challenge that the possibility of moral dilemmas poses to the ordering semantic account of ought-statements. I critically discuss and compare the advantages and disadvantages of two ordering semantic analyses of ought-statements when it comes to moral dilemmas: the conflict and disjunctive accounts. I conclude that in order to adequately describe one’s normative predicament and reasoning in a dilemma situation, two kinds of ought-statements defined by the two accounts have to be captured in a single semantic framework. And, I show how it can be done in the ordering semantic framework. Ch.3 critically evaluates the real force of the Disagreement-Based Argument (DBA) in metaethics, exemplified by R.M. Hare’s cannibals argument and Horgan and Timmons’ Moral Twin Earth argument. I analyze the structure of the DBA and defuse its direct threat to descriptivism by disproving its most problematic assumption : that genuine disagreement between two parties requires that the parties agree in what they mean by the key terms in the dispute. I argue that the disagreement data the DBA elicits from thought experiments does not put real pressure on descriptivism, because the proponents of descriptivism can explain the data away by showing how two different linguistic communities can genuinely disagree about what to do and how to live.