I have a major issue here. You state that reliability should be allowed to be flexible across diagnostic categories. Nothing could be further from the truth. The fact is that reliability sets the UPPER limit on validity, something that many seem to forget. A diagnostic category can be no more valid than it is reliable. Thus, by loosing standards of reliability you automatically detract from the validity of a diagnostic category. Would you state the validity of certain diagnostic categories should be allowed to fluctuate? I would hope not. Moreover, the values for Kappa used by the DSM reflect research thresholds and these values can be flexible depending on the aims and design of research, but in a clinical situation where the life and future of a person is at stake it is important to be as precise and accurate as possible. The DSM does not concentrate on validity studies, which is a major problem.