There's also no problem defining and measuring addiction using discrete definitions such as withdrawal symptom severity, rehabilitation success rates or neurochemical signature. The problem arises when researchers (or anyone) then believes that this is the definition of addiction, when in reality it is a definition contrived for the purpose os (valid and useful) research. In a different context, it might be useful to think of these as indicators of... Usefulness is contextual.
Discrete language is fine. It just isn't the way we communicate normally, and it's impossible to use only discrete language to describe things we don't understand fully.
