Over the last 40 years certain psychosocial treatments have made great strides on the long road towards becoming a reliable clinical science. A mark of this maturation is the recent enabling of sufferers from anxiety disorders and depression to help themselves when guided by particular manual or computer self-help systems, indicating that several effective therapeutic mechanisms have become reasonably defined. Still unknown, however, is exactly how many separate therapeutic paths are being trodden to improve anxiety and depression, and how much various tracks converge on common mechanisms. Careful dismantling studies may in time reveal that some behavioural, cognitive and interpersonal therapy procedures are in fact proxies for administering as yet dimly-perceived therapeutic principles. Progress would accelerate further once therapists were required to specify how much they improve their patients on simple and widely-agreed measures of outcome and cost, and once they learn to speak a broad-accepted common language.