Blog Post 1)
Does ANYONE have a valid definition of “addiction”? Posted on 7 October 2013 by Carl V Phillips | 28 Comments
by Carl V Phillips
Sorry for the radio silence. Was traveling, and then exhausted and sick. During my travels, I had a few interesting debates about the concept of addiction, and promised to write the following, which I finally have completed. The funny thing about all debates about addiction that I have ever participated in is that whenever someone defending the use of that term (pretty much always in the context of nicotine, since those are the conversations I am in) is challenged to explain what it means, they cannot. Quite often – in a bit of rather patent irony – they also get mad at me for daring to suggest that it is not well defined, even as they fail to offer any useful suggestion about how to define it.
Addiction exists in the same sense that happiness or beauty exists: It is a “know it when I see it” concept we are familiar with, and most people use the common language (i.e., informal) term. But it does not have a scientific or even medical definition. If words like “beauty” are used in a scientific context, the author needs to define how he is using the word, because the common language notion is not precise enough. Those who use “addiction” in a scientific context, without explaining what they mean by it, are creating confusion (intentionally, in most cases, I would guess).
For example, in the debate about THR, it is often claimed that those who quit smoking by switching to a low-risk product are still addicted. But are they? That depends on what the word means. But even when concrete claims like that are being made – so the word is not an aside, but the crux of the discussion – no one actually explains what they mean by the word. (Aside: Note that usually the most effective quick response to the silly “but they are still addicted!!!!” rhetoric is not to attempt the actual thoughtful discourse presented in this post, but to just say something like, “So what? They are addicted to something that is almost harmless” or the snarkier, “So you would rather someone die from lung cancer than stay addicted?”)
It is definitely possible to create a valid definition of addiction. Most everyone agrees that there is such a thing, though not necessarily that it covers vaping or even smoking. Consider the following (labeling it Scenario 1 to refer back to it): A person uses meth. He knows he needs to go to work today (sober), or he will lose his job, which is his only means of support, paying not only for his food and housing, but also his meth. If he smokes meth now he will have no money in a week. Yet he fails to resist the urge to smoke now, in spite of the dire immediate consequences.
That seems to be addiction according to any typical understanding of the term. But smoking cigarettes does not produce any experience similar to that story. Indeed, using a term that evokes the image of Scenario 1 to describe tobacco use is misleading in itself, and that evocation is often intentionally used to make smoking seem worse than it is. Still “addiction” might be defined to include smoking also. Perhaps not. And if so, does it include e-cigarette use too? To address these points, we need an actual definition.
I have thought a lot about this over the years. I have a pretty good idea of what the definition must include and what it must exclude. So here is the challenge to those of you who are sure there is a defensible definition: Can you provide one that fits the following parameters, or argue that some parameters are not reasonable expectations for the definition? Note that the list is really a lot shorter than it looks, as summarized at the end.
Requirement 1: The definition has to cover a situation like Scenario 1. If anything fits this definition, it has got to be that. This probably goes without saying, but it is included for completeness.
Requirement 2: The definition cannot be so broad as to cover such behaviors as breathing and eating, or spending time with your family and trying to get your work done. If the term is to be useful, it cannot be so broad. Yet I would estimate that half of the definitions someone offers for the term include behaviors that are biologically necessary or are the most positive behaviors people engage in. Note that taking refuge in the caveat “but it not biologically necessary” does not address the second of these (and anyone who offers a definition and then realizes they need that caveat has clearly not really given their definition any serious thought).
Put another way, the definition cannot be equivalent to saying “the benefits of the activity outweigh the costs by such a huge amount that the actor will not stop doing it.” This is the definition of “highly beneficial” not “addictive”, and yet many proposed definitions of addiction do not make any distinction between those.
Requirement 3: The definition cannot be so broad as to include eating dessert, driving, travel, or mountain climbing. I am thinking of proposed definitions like “they keep engaging in a behavior even though it is bad for their health”. (Note that this is not to say that once you had a definition, it cannot be found to apply to climbing or eating for some people; it just means that if the proposed definition includes all those activities for everyone, it is clearly a fail.) Roughly half of the activities we habitually engage in are bad for our health on net, and we clearly would not call them addictive just because of that. Setting a minimum risk level (to try to include cigarettes and meth but exclude burgers) is no solution, since there are not-necessarily-addictive activities that are riskier than smoking. Also, imagine medical breakthroughs that lowered the risk of smoking to below the proposed threshold: Any proposed definition of addiction that would be changed by finding a magic cure for lung diseases, even though the behavior did not change, is a fail.
Note, however, that the harmfulness of the behavior seems to be part of the common language notion of addiction. The way the word is used, it obviously refers to something substantially bad happening to the addicted person (see Requirement 5). If so, this suggests that the use of snus cannot be addictive, whatever the behaviors and urges involved, because it is close enough to harmless that the risk is not measurable.
Requirement 4: The definition cannot just be “uses a drug” or something similar. Often in political rhetoric (some of which pretends to be science), “addiction” is just used as an inappropriate substitute for “use”. It is a political trick: Take a word with nefarious implications and use it to refer to mere existence, and thereby tar a population with it without ever actually making any substantive claim. It reminds me of my days in Berkeley at the dawn of what would come to be disparaged as “political correctness”, when the chatterers tried to declare “racist” to mean something like “gets any benefit from the fact that there are racial disparities” or merely “is white”. Obviously a word misused like this becomes so broad that it loses all real meaning – except that it does not lose innuendo because the nasty implications of the word linger, even though the new definition has no nasty implications. Cute, huh?
But regardless of whether it is intentional rhetoric or just the sloppy language of non-thinkers, it is clear that addiction does not just mean use. We already have a word for that.
Requirement 5: It must necessarily be a bad thing that someone is addicted. That is, if all you know is that a person is addicted to X and nothing more (including what X is), you know that this is a circumstance that is substantially worse than if it were not true. To be consistent with the common language usage, it must be that “addicted to X” is bad independent of whether “does X” is bad (which relates closely to Requirement 4). The way the word is typically used, it implies something bad in itself, beyond any badness of the activity. Witness the common anti-THR refrain “but they are still addicted”, meaning “something is still bad when of ex-smokers are using a low-risk alternative, even though the activity itself has no substantial downsides.”
This is a rather more complex condition than it might seem at first blush. It subsumes Requirements 2, 3, and 4. This makes those redundant, but I went ahead and included 2, 3, and 4 as separate points to emphasize their implications which might be a bit too subtle if they were just subsumed under 5.
Separating those more undeniable specifics also allows for some aspects of Requirement 5 to be relaxed if anyone wants to argue that the proposed scientific definition need not be a negative epithet like the common language usage. If that is one’s approach, it is important to keep in mind the political uses of the term (and the analogy to “racism” noted above), and to make clear when using the term that it is not necessarily a bad thing. For example, the Chicago School economic definition of addiction by Becker et al. – which is the only candidate for a real scientific definition of addiction I recall ever having read in literally decades of interest in the subject – was proposed in the context of how it can be beneficial and a rational choice to become addicted.
Requirement 6: The definition cannot appeal to untestable claims, let alone absolute claims that are clearly false. In particular, any definition that includes a phrase like “cannot stop” is a fail. First, it is probably not true: if you could credibly tell someone that, say, you would torture his mother or child to death if he smoked another puff, he would stop. No doubt there are a handful of highly dysfunctional cases where even this consequence would not stop someone, but so few that the term would be almost vacuous. Second, even without such extreme scenarios, it is impossible to know how someone would react under every possible realistic circumstance, and thus a universal such as “cannot” can never be shown to be true, and so any proposed definition that uses it actually includes nothing.
Requirement 7: The definition must be based on behavior. The phenomena that are referred to as “addiction” have to do with the actions and volitions of the actor, not the object of those actions (a drug or whatever) itself. This is not to say that the implications of the object cannot be included in the definition (e.g., “addiction means that someone does X, and X has property Y” where Y might refer to a health hazard), but the characteristics of the actor (“does X”) are primary and any “property Y” is secondary.
In particular, this requirement excludes an appeal to “brain porn”. (That term refers to the recent quasi-science of measuring people’s brain activity under particular circumstances and drawing worldly conclusions, ignoring that there is no basis whatsoever for relating the images to reality – hey, it makes cool pictures and uses fancy equipment, so it must be science, right?). It is easy to make the case that a definition of addiction cannot refer to dopamine receptors, PET scans, and the like: Recall Scenario 1. Now consider some proposed definition of addiction that is based on brain chemistry, and imagine that you measured the brain activity of the person in the Scenario and found that none of the conditions were met. Would you say “oh, I guess he was not addicted after all”? Of course not. Similarly, if someone had all the brain activity in a proposed definition, but could easily take-or-leave the behavior, we would agree that is not addiction. So while brain porn might (might!) offer a prediction about whether addiction is present, it cannot be a defining characteristic.
Additionally, this requirement excludes defining addiction in terms of merely facing withdrawal symptoms (often called “dependence”) or having an acquired tolerance. Those experiences might be part of the reason why someone is addicted, but they are clearly separate phenomena. If someone has these but can take or leave the behavior nonetheless (which is a fairly common pattern for, say, caffeine) that cannot be addiction.
Requirement 8: The definition cannot just beg the question but using other ill-defined terms. If “addiction” is partially defined by “having a compulsion”, it is necessary to define “compulsion”. (Also, if someone “has a compulsion” but does not act on it, is that still addiction? You can go either way on this, but need to be clear about it.)
So, that is a very long list, but it also can be almost completely summarized as this short version: A definition of addiction must be based on behavior and must not be so broad as to include every strongly desired behavior nor every behavior that creates health risks nor all use of drugs; meeting the definition must either be inherently bad or a case must be made that addiction is not necessarily a bad thing. And, of course, it actually has to be a definition.
As I suggested, I have some candidate ideas. But before offering those I would like to see if those who insist that there is a clear and obvious definition can tell me what they think it is. I await your replies. But based on my experience of the universal failure to actually answer the question, I will not hold my breath.