DSM-IV (American Psychiatric Association, 1994) groups disorders into diagnostic classes based on the subjective criterion of "shared phenomenological features". We now have sufficient data to eliminate this rational system and replace it with an empiricallybased structure that reflects the actual similarities among disorders. The existing structural evidence establishes that the mood and anxiety disorders should be collapsed together in an overarching class of emotional disorders, which can be decomposed into three subclasses: the bipolar disorders (bipolar I, bipolar II, cyclothymia), the distress disorders (major depression, dysthymic disorder, generalized anxiety disorder, posttraumatic stress disorder) and the fear disorders (panic disorder, agoraphobia, social phobia, specific phobia). The optimal placement of other syndromes (e.g., obsessivecompulsive disorder) needs to be clarified in future research.Rethinking the mood and anxiety disorders 4 Rethinking the Mood and Anxiety Disorders:
A Quantitative Hierarchical Model for DSM-VA large body of research has sought to explicate the underlying structure of the mood and anxiety disorders. This research has been stimulated by two key taxonomic problems: comorbidity and heterogeneity. Comorbidity can be broadly defined as the cooccurrence of different disorders within the same individual (see Lilienfeld, Waldman, & Israel, 1994;Mineka, Watson, & Clark, 1998); it reflects the empirical overlap between constructs that are hypothesized to be distinct. Therefore, when comorbidity occurs at greater than chance levels in the population, it raises the more general issue of discriminant validity. Evidence of significant-often substantial-comorbidity among DSM disorders has led to the development of increasingly sophisticated models that attempt to account for these patterns of co-occurrence (see Mineka et al., 1998).In contrast, heterogeneity results when phenomena that ordinarily are collapsed together are found to be sufficiently distinctive to warrant their separation; in the nosological context, this frequently results in the creation of diagnostic subtypes, which is an acknowledgment that an existing taxonomic category is too heterogeneous to be maximally informative (Watson, 2003b). This type of evidence has stimulated research into the symptom dimensions underlying many of the anxiety and mood disorders, including posttraumatic stress disorder (PTSD), obsessive compulsive disorder (OCD), and specific phobia (see Watson, Gamez, and Simms, in press).In this paper, I will review this structural evidence and explore its broader implications for the classification of the mood and anxiety disorders in DSM-V. I begin by explicating the logic of the current DSM-IV (American Psychiatric Association [APA], 1994) classification scheme for these disorders; to highlight significant problems with the current taxonomy, I will recast it as a quantitative structural model. I then review earlier structural models of this domain and discuss how they attempted to address...