AbstractThis paper is concerned with the relationship between alcohol dependence and depression, and more specifically with the issue of which criteria should be used in distinguishing primary and secondary disorders. The data presented indicate that borderline or subclinical depression frequently precedes the onset of alcohol dependence. This finding is discussed in terms of previous studies, which show that depression is generally secondary to alcohol dependence. Such studies have focused almost exclusively on severe forms of depressive disorder, and, it is argued, used inappropriate criteria to designate the onset of alcohol dependence (e.g. withdrawal phenomena and alcohol‐related problems). In contrast to previous studies, the present paper indicates that subclinical episodes of depression should be assessed by clinicians as possible predisposing factors in the development of alcohol dependence. In addition, it is argued that research designed to establish whether alcohol dependence is a primary or secondary disorder when it occurs in the presence of an accompanying psychiatric disorder should measure onset in terms of changes in behaviour such as giving‐up existing interests in order to spend time drinking and consuming more alcohol than former compani