The diagnosis of posttraumatic stress disorder has been criticized on numerous grounds, but principally for three reasons (a) the alleged pathologizing of normal events, (b) the inadequacy of Criterion A, and (c) symptom overlap with other disorders. The authors review these problems along with arguments why the diagnosis is nevertheless worth retaining in an amended form. A proposal for the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V) is put forward that involves abolishing Criterion A, narrowing the B criteria to focus on the core phenomena of flashbacks and nightmares, and narrowing the C and D criteria to reduce overlap with other disorders. The potential advantages and disadvantages of this formulation are discussed.