Possible-world semantics for modal languages of various kinds has been a major focus, and indeed motor, of research in philosophical logic, formal semantics, and philosophy of language over the past several decades. This talk advances an a priori argument to the effect that the orthodox application of possible-world semantics to modal language---as manifest in its permissive attitude toward truth-aptness---is fundamentally flawed, and advocates a strictly Tarskian approach to truth in its stead. Some consequences of adopting such a revisionary approach will be discussed, both with respect to the formal languages themselves and to their application to natural language.