Abstract: Is addiction a behavioral pattern, or the underlying cause of a behavioral pattern? Both views are found in prominent accounts of addiction, but theorists generally do not notice that they are taking a controversial position, let alone justify it. A third possibility is that addiction consists in both addictive behavior and its causes, though this view is less obviously present in the literature. I argue that two important considerations favor the "cause view" over the "behavior" and "hybrid" views. The first is that we want to be able to explain addictive behavior with reference to addiction. The second is that addiction seems to persist long after addictive behavior has ceased. Only the cause view can easily accommodate both data points.