What is Wrong with Addiction
Clinical criteria have trouble distinguishing addictions, on the one hand, from, on the other hand, appetites—like our appetites for food and water—and non-addictive passions that guide our lives, from serious hobbies to parenting. The simplest explanation of how addictions are distinct from non-addictive appetites and passions is that addictive behavior reveals some misvaluation by the addict, that the addict is wrong to act as she does. Psychological evidence supports this philosophical proposal by explaining how such a misvaluation is reinforced, namely by the addict’s acting in unthinking, impulsive ways. This reinforcement explains addiction’s chronic resistance to contrary evidence. This proposal neatly accounts for the questions left unanswered by standard diagnostic criteria of addiction.
addiction, compulsion, misvalue, diagnostic criteria, dependence, reinforcement, impulsivity, willing addict
The question ‘What is addiction?’ seems to ask for a clinical or biological answer. The research on addiction has progressed much faster in biology and neuroscience than our philosophical understanding of that research.1 Therefore, it can be tempting to look to the relevant psychology or neuroscience to answer our philosophical questions, which ends up treating addiction entirely as a psychological or neurological matter (Foddy and Savulescu 2010; Holton 2009, 97–111). However, many of our questions about addiction are not fundamentally biological or neurological questions. Here, I suggest that an answer to one question about addiction, the question of what addiction is, turns on the decidedly more philosophical question, “What is wrong with addiction?”
My argument focuses on a neglected distinction. Addictions look strikingly like normal appetites and strong passions that give our lives meaning. Yet our appetites and passions are not diseases, nor simple failures of rationality, nor should our passions be understood entirely as psychological or neurobiological matters.2 Because the commonly used diagnostic criteria of addiction cannot distinguish addictions from appetites and passions, I suggest that the clearest understanding of what addiction is, one that is consistent with ordinary usage and clinical criteria, comes from addiction’s being similar to other behavior but distinct from it because we understand there to be something wrong with addiction that is not wrong with similar appetites and passions. Therefore, what is wrong with addiction partially explains what addiction is.
This project is one of rational reconstruction. Terms or concepts like ‘addiction,’ which serve many practical functions and have different meanings in different idiolects, are unlikely to have informative necessary or sufficient application conditions (Sinnott-Armstrong and Pickard 2013). Therefore, what I offer is a characterization of addiction such that it fits our existing, ordinary usage and our clinical usage, capturing the clear cases of addiction and explaining why the unclear [End Page 25] cases are unclear. The advantage of such a characterization is that it will help us to understand how to extend the concept into cases that we currently find unclear, to direct us to ask the questions that will make them clearer. The project does not alone prescribe which behaviors are addictions; indeed, it counts against my proposal if it alone settles whether unclear cases, like regular, heavy caffeine use, are addictions, because it instead explains why such cases are difficult and what we need to know to settle whether it is an addiction. The characterization immediately contributes to our understanding of what addiction is, then, and opens the way to a better understanding of how we should settle difficult cases.
My argument starts with the criteria that standardize the clinical diagnoses of substance abuse and dependence. Although those criteria are inadequate to characterize addiction on their own, they do point to a philosophical way to understand what is wrong with addiction. I conclude that the best understanding of addiction will explain that it differs from appetites and passions because addiction is based on a misvalue. I then suggest that psychological findings help to explain why the motivation based on this misvalue is especially resistant to contrary evidence, in the way that addictions are.
Addiction researchers and psychologists depend for clinical diagnoses on the Diagnostic and Statistical Manual of Mental Disorders (DSM).1 The DSM makes no claim to diagnose all addictions or indeed to explain addiction per se at all— “addiction” is not itself a DSM category. The DSM-IV considers, for example, drug and substance dependence separately from gambling disorders, although both are ordinarily considered addictions (American Psychiatric Association [APA] 2000, 191–296, 663–78). The DSM-5 combines these two, but coffee, video games, and junk food—which many people consider genuine and harmful addictions—are still not included. Nevertheless, DSM-IV criteria of substance dependence (or “substance use disorder,” in the DSM-5) are widely recognized as diagnostic of addiction. To show that “What is addiction?” is not merely a clinical question, I first consider the proposed clinical answer to that question.
The DSM-IV category of substance dependence is distinguished from substance abuse. (The two are combined in the DSM-5 under “Substance-Related and Addictive Disorders.”) The DSM-IV category of abuse is, in summary, problematic substance use that is not dependence (APA 2000, 199). This diagnosis captures, say, the unaddicted drinker who has too much too often in a way that interferes with a normally functioning life: she misses work, drives drunk, and so on. Although the DSM-5 combined the two in part to avoid the implication that substance abuse is just the less harmful form of dependence—non-dependent abuse can obviously be quite harmful—there is nevertheless some distinction to draw between someone who drinks too much and the alcoholic.
The DSM-IV category of substance dependence is more directly my concern. (What I say here will apply mutatis mutandis to the DSM-5’s substance-related and addictive disorders criteria as well.) Dependence is “a maladaptive pattern of substance use leading to clinically significant impairment or distress,” that is manifested by “three (or more) of the following, occurring at any time in the same 12-month period.” These manifested criteria are:
1. Tolerance, as defined by either of the following:
a. A need for markedly increased amounts of the substance to achieve intoxication or desired effect
b. Markedly diminished effect with continued use of the same amount of the substance
2. Withdrawal, as manifested by either of the following:
a. The characteristic withdrawal syndrome for the substance…
b. The same (or a closely related) substance is taken to relieve or avoid withdrawal symptoms
3. The substance is often taken in larger amounts or over a longer period than was intended
4. There is a persistent desire or unsuccessful efforts to cut down or control substance use
5. A great deal of time is spent in activities necessary to obtain the substance (e.g., visiting multiple doctors or driving long distances), use the substance (e.g., chain smoking), or recover from its effects
6. Important social, occupational, or recreational activities are given up or reduced because of substance use [End Page 26]
7. The substance use is continued despite knowledge of having a persistent or recurrent physical or psychological problem that is likely to have been caused or exacerbated by the substance (e.g., current cocaine use despite recognition of cocaine-induced depression, or continued drinking despite recognition that an ulcer was made worse by alcohol consumption). (APA 2000, 197)
The category of dependence is, unsurprisingly, what we may ordinarily think of as addiction. This is the user who cannot quit despite good evidence that she should.
Note two important points about these manifested criteria. First, the criteria are diagnostic criteria. What I mean by that is that the criteria focus on features that typify contemporary substance abuse or dependence: these are not necessary features. Second, the diagnostic criteria do diagnose clearly addictive cases without also diagnosing clearly non-addictive cases. A stereotypical heroin addict is diagnosably dependent: he is tolerant, is unsuccessful in cutting down his use, and spends much of his time in pursuit of the drug, so he easily meets at least three criteria of dependence (criteria 1, 4, 5). On the other hand, a social drinker does not pass up major obligations to drink instead, or undertake risky behavior while intoxicated, or have legal problems or persistent social problems stemming from her alcohol use. She therefore is not diagnosed by the DSM-IV criteria of dependence, which is also as we would expect.
Consider, however, the following substance whose use easily fulfills all seven criteria of dependence: food. I am quite dependent on food. I suffer hunger pangs when I do not eat for long enough, I spend a great deal of time eating or in activities preparing to eat, I give up many other things to eat regularly, I eat despite occasional indigestion, weight gain, and even acid reflux, I regularly eat more than I intend, and, if we are literal enough about the characterization of ‘tolerance,’ who could deny that I need markedly more food now than when I first started eating? Food consumption easily fulfills all seven criteria of dependence.
Why does the DSM-IV not diagnose food consumption as a dependence? If food consumption fulfills all seven manifested criteria, then it must fail to fit the DSM-IV’s initial necessary condition of dependence. Food consumption must not (normally) be a “maladaptive pattern of substance use leading to clinically significant impairment or distress,” even if it then manifests all of the subsequent criteria. This is plausible. This initial phrase (“maladaptive pattern…”) may be rather vague, but surely regular food consumption—and water consumption and sexual activity—cannot in general count as maladaptive or as leading to clinically significant problems. If such basic life activities were maladaptive, then ‘maladaptive’ would seem—impossibly—to apply to anything. The DSM-IV can deal with this first potentially troubling case, then, by reminding us of the initial, necessary condition that requires dependences to be maladaptive.
Other cases, however, require a more precise use of ‘maladaptive.’ Imagine a heavy caffeine user, ‘Jeff.’ Jeff is tolerant and suffers from typical caffeine withdrawal, a headache, and sluggishness. Jeff finds himself drinking coffee later in the day than he intended and can never seem to cut down on his coffee consumption. He knows that caffeine makes his sleep less restful, can make his thoughts flighty, and can even lead to mild tremors. Jeff thus seems to fulfill every one of the seven criteria for dependence.
Jeff fulfills the criteria, but is he dependent? Despite the DSM-IV’s classification of caffeine as a potentially intoxicating substance (APA 2000, 232), the DSM-IV explicitly denies that there can be caffeine dependence (APA 2000, 192). (The DSM-5 includes caffeine use disorder in section III, to encourage more research.) It must be that caffeine dependence does not fit the initial, necessary condition of dependence. Why not, though? Some of its effects are genuine harms and may cause Jeff “impairment or distress.”
One response is that, although Jeff has impairment or distress, such impairment or distress is not ‘clinically significant.’ ‘Clinically significant’ is a judgment required by many DSM-IV diagnoses. It defers to the clinician’s expertise to determine whether the observed symptoms (i.e., the impairment or distress) are serious enough to warrant treatment and are likely to indicate an underlying disorder, not just a ‘false-positive’ presentation of symptoms without an underlying disorder (APA 2000, 7; [End Page 27] cf. Spitzer and Wakefield 1999). But to illuminate what counts as addiction, we cannot assume a clinician’s determination of how much distress or impairment is significant: this may assume precisely the features of addiction that need to be illuminated. We need instead to clarify what should count as clinically significant; and, more centrally, we need to clarify what “maladaptive” means.
The danger of trying to clarify this initial phrase, however, is that, given that the subsequent criteria are fulfilled, whether heavy caffeine use is “a maladaptive pattern of substance use leading to clinically significant impairment or distress” seems to turn on whether patterns of caffeine use like Jeff’s seem like our antecedent—albeit rough—idea of what addictions are. That is, if Jeff’s pattern of caffeine use seems as unproblematic as regular food consumption, then we rightly conclude there is nothing particularly maladaptive about it, and any impairment or distress it may lead to is not clinically significant. Therefore, “maladaptive” and “clinically significant” exclude any diagnosis of caffeine dependence. If, on the other hand, Jeff’s caffeine use seems addictive, then its use will also seem maladaptive, and any impairment or distress will seem clinically significant. Instead of using that initial broad phrase as an independent standard to determine what can count as dependence—and therefore ultimately to clarify what addiction is—we risk using what we count as addiction to help us refine our reading of that initial phrase.
To avoid this circular explanation, we could consider whether what is harmful or maladaptive is determined by social mores or by moral or biological norms. That would clarify the initial phrase. If Jeff lives in coffee-loving Seattle, large parts of his life are dictated by his caffeine needs, but a caffeine-filled life in Seattle is a relatively easy one. If Jeff lived, instead, where caffeine use is restricted, he may come to be in significant distress, leading to genuine harms. His caffeine use would look maladaptive, and he may seem like an addict.
Indeed, social mores may play a more indirect role. If a society values extended periods of working at a high pace, regular alcohol use will not be well adapted, but caffeine use may be. If, on the contrary, late-night affability is more highly prized, then regular hangovers may be considered normal. The harms would be real, and they would be determined largely by enforcement of social mores.
If maladaptiveness is determined by community standards, that would explain why the addict also fits some other criteria of dependence. For example, if the community criminalizes possession of the substance, then legal problems are likely. (For this reason, legal problems are eliminated from the DSM-5 criteria.) If a substance is hard to obtain, then one may spend a lot of time and experience inconvenience in obtaining it. And so on. Therefore, the subsequent criteria of dependence would ‘manifest’ the initial, necessary condition.
Unfortunately, the conclusion that dependence—and whether substance use is maladaptive—depends solely on a person’s community’s mores runs into problems. Imagine the stereotypical heroin addict of the first example and put him in a community of stereotypical heroin addicts in which heroin is abundant. Now, the addict—call him “Renton”— does have tolerance, and arguably even withdrawal,2 but Renton does not count as dependent because he does not fulfill the manifested criteria of dependence. His obligations are dictated by the daily routines considered normal by his community, so he passes up no obligations that he should be fulfilling, he spends as much time in pursuit of heroin as everyone else does, and any nascent desire to cut down his heroin use is quite publicly frowned upon. His use is entirely in line with the community mores, which means he would not—on this reading of “maladaptive”—be diagnosable as heroin dependent. This counterintuitive result counts against understanding “maladaptive” as referring solely to social mores.3
A more plausible alternative is that a community’s standards generally purport to reflect some moral or perhaps biological standards. These moral or biological standards are, for example, that some effect of the substance (e.g., hangovers, lowered productivity) is harmful or wrong, and therefore to be discouraged. These community-independent standards, when enforced by a community, would lead to a substance user’s fulfilling many of the subsequent criteria of dependence. The standards themselves, however, are independent [End Page 28] of the community’s enforcement, so some substance use is maladaptive even if no community views it as such.
This explains why community mores seem relevant to but not definitive of what is dependence-causing. For a substance to be dependence-causing, the community must be right that the substance is harmful or wrong, perhaps harmful or wrong in particular ways. Yet, even on this view of “maladaptive,” Renton is still not diagnosable as dependent. He does not fulfill enough, or perhaps any, of the subsequent criteria of dependence, so he is not diagnosable as dependent. And, as I illustrate in the next section, we do need to say more than that addictions cause some harm, because many non-addictive passions and appetites also cause harm.
This attempt to make sense of “maladaptive” as it applies to difficult cases, like Jeff’s and Renton’s, points to a fundamental problem with using DSM criteria to define or even characterize addiction. The DSM is a diagnostic manual intended to be used by clinicians and researchers in actual communities, where heroin is not in abundance, its use is outlawed, it is viewed as harmful, and so on. The DSM is not a compendium of psychiatric definitions (McHugh 2005). The seven subsequent criteria of dependence may be a good guide to how addiction manifests itself in many contemporary Western communities, but they are not a guide to what addiction itself is.
This examination of the DSM criteria shows the limit of what a diagnostic manual reveals about what addiction is. The initial DSM-IV condition was certainly promising, if inadequately developed, but the seven criteria of dependence have turned out not to be necessary, even if diagnostically useful. Most significant, the criteria fail to distinguish addictive dependences from non-addictive dependences and—as I will develop—from passions. If we want to know whether Jeff is a caffeine addict, or why Renton is an addict, the DSM-IV criteria will not answer us. And if we jettison those subsequent criteria as unnecessary for addiction, then the initial DSM-IV condition will need to be more specific than just “a maladaptive pattern of substance use leading to clinically significant impairment or distress.”
The Wrong of Addiction
“Maladaptive” is used throughout the DSM-IV, enough that it may start to seem unremarkable in any particular disorder. But what could “maladaptive” mean in the case of dependence?
To get at the answer to that question, I first make some assumptions about what sort of thing addiction is likely to be. First, there must be some pattern of behavior—obtaining the drug, using the drug, clearing impediments to using the drug; otherwise, there is at most only a potential addiction. Second, that pattern of behavior must issue from some relevant motivation: a non-smoker made to chain smoke at gunpoint is not an addict. Addiction, then, is the motivation (or set of motivations; for simplicity, I speak of just one motivation) on which the relevant pattern of behavior is based.
Further, addictions seem to be very difficult to stop. That may be enough to characterize them as ‘compulsive,’ although some may have a stricter standard of what is necessary for a disorder to be compulsive (e.g., Pickard 2012)—or may doubt that addicts find it very difficult to stop. I do not engage the issue, except to say that addictions seem to be compulsive in just the way that many non-addictive appetites and passions are compulsive, that is, they are all very difficult to stop, despite potentially strong incentives to stop.4
This initial characterization of addiction has not distinguished addictions from non-addictive appetites and passions. Appetites and passions also motivate patterns of behavior and, intuitively, are also compulsive. This point may be obvious for appetites, but it is also true for passions. My passion for running occupies my time and thoughts in the compulsive way an addiction could, but I am unlikely addicted to running (which is not to deny the possibility of a running addiction; see Chapman and De Castro 1990; Kanarek et al. 2009). Or, consider parents’ passion for their children— they pass up work and social opportunities for them, spend much time with them or preparing for their needs—yet few would (non-metaphorically) call one’s devotion to one’s children an ‘addiction’ (Roiphe 2009).
Now, perhaps there is no deep distinction to make between addictions and appetites and passions. Certainly, as a point about the language, [End Page 29] we need not make any such distinction, and we may not always have done so (Alexander and Schweighofer 1988); someday we again may not. But we currently do, and, as I develop below, there is something interesting in the distinction we do seem to make, particularly the distinction between the appetites we all have and those that only some of us have, the ones that we develop that then become compulsive. Therefore, however ‘real’ this distinction between addictions and similar non-addictive appetites and passions may be, it is a distinction worth understanding. It certainly will shed light on what addictions are.
When the DSM uses the term “maladaptive,” this may then be getting at the general thought that addictions, unlike other appetites and passions, are harmful or wrong or bad in some way. Of course, the DSM’s own particular use of “maladaptive” and “clinically significant impairment or distress” reflects the DSM’s diagnostic role. After all, the DSM is intended to diagnose those who are (or may soon be) under some psychiatric or clinical supervision, and its manifested criteria come largely from that same, supervised population. What is true of addicts in treatment, however, may not be true of addicts in general. So, for example, it may be true of addicts in treatment that they are in distress over their addictions, but addicts not in treatment may not share this trait—hence, their being less likely to seek treatment (cf. Berkson 1946).
Regardless, the general thought that there is something harmful or wrong or bad about addictions seems promising, even as applied to addicts who are never in treatment and are never in distress over their addictions. ‘Harmful’ or ‘bad’ and ‘wrong’ need not entail (on some accounts of morality) that addictions are immoral. They may just mean that there are some standards—perhaps moral, perhaps biological—that are violated in cases of addiction and that are not (in general) violated in non-addicted patterns of behavior. Addictions are bad or wrong in some (yet unclear) way. (I use ‘wrong’ in this relaxed sense of ‘failing to reach some standard’ that will cover the general idea of harmful, bad, or wrong.) What seems promising about this thought is that the wrongness of addictions would explain why we criticize addictions but do not criticize (in general) a person’s passions, which are not (in general) wrong to have or pursue. If the suggestion clarifies what addiction is, that clarification will count strongly in favor of thinking it is correct.
What is the wrong that would make a pattern of behavior an addiction? It cannot just be the bad or harmful effects. Even non-addictive patterns of behaviors (and their underlying motives) may cause harm: chronic running harms the knees, one’s hobby may expose one to harmful chemicals, and children require sacrifice of projects, pleasure, and peace. If such harms are to distinguish addiction from similar non-addictions, addiction must at least involve some more specific wrong.
A wrong of addiction is often noticed in its effects, but these are not specific to addiction. An addict may lose her job, family, get into legal trouble, and go bankrupt, which indicates a problem, but that problem may be that she drinks too much or gambles too much, not that she is addicted to either. Moreover, addictions need not have such dramatically adverse effects.
The better known, and more immediate, effects of many addictive activities or substances, particularly of psychoactive drugs, are the effects of removing a person to some extent from complete contact with her environment and decision making, what we loosely call ‘intoxication.’ Because many drugs of abuse are intoxicating, it may be tempting to think that the wrong of addiction is just the wrong of regular intoxication, even if it does not lead to job loss or family breakup.
Regular intoxication itself may indeed be a wrong. But, although they prominently co-occur, it is important to keep intoxication distinct from addiction. One key way in which they differ is that intoxication could be a reasonable goal, a desirable state, whereas the same cannot generally be said about addiction. The point may be obvious. If not, consider an example. I have untreatable, excruciating pain that keeps me from living anything close to a normal life. The right kind of intoxication allows me to ignore the pain. Perhaps the intoxication itself keeps me from living a normal life, so I am in many ways no better off, but at least I am not in excruciating pain. Although intoxication is not the only option available to me, it may be a reasonable one. [End Page 30]
Even when intoxication is intended and reasonable, addiction is not therefore intended and reasonable, even if regular intoxication in fact results in addiction. For the addiction itself to be intended would require quite a special story, one like the following (cf. Frankfurt 1988, 14–15). My eccentric grandmother believed that our entire family is alcoholic, and she wanted us all to seek treatment. She leaves me a large inheritance to be collected only after my successful completion of the treatment program. However, the treatment center will not admit someone who has no addiction to alcohol, and they are very good at screening out people who will use up their resources but who are not really addicted. Because I can only complete the program if I start it, and I can only start it if I have an addiction, then it may be reasonable for me to intend to have an addiction. Such cases are rare, however, whereas cases in which it is perhaps reasonable to intend intoxication are, unfortunately, widespread.
Given this distinction between addiction and intoxication, it is clear that not all addictions involve intoxication. Gambling, for instance, may involve no intoxication. It may involve some states in which one is less than fully aware of oneself and one’s surroundings (Burke 2008; Castellani and Rugle 1995; Evans and Coventry 2006), but that is also true when one is engrossed in a movie or an interesting project (Csikszentmihalyi 1990). It would therefore be a stretch, or a metaphor, to say that gambling addictions are intoxicating.
Even the addictions that do involve intoxication need not do so regularly. One could become intoxicated by nicotine, but nicotine intoxication, that is, having so much nicotine that one is effectively removed from one’s environment or decision making, is a rare state for the regular smoker. Perhaps the occasional cigarette is intoxicating, but, even if one does become intoxicated by nicotine, it would be strange to think these uncommon and even unpleasant states are what are wrong with the addiction.
It is an empirical question whether every addiction is regularly intoxicating, and there is a clear counterexample in nicotine. Indeed, there may be counterexamples for any addictive substance which one takes just to ‘feel normal.’ Even highly intoxicating substances may be used by an addict at regular low doses so as not to cause intoxication, though using regularly ‘just to feel normal’ may still characterize addiction (Khantzian 1997; Koob and Le Moal 2001).
Much of what I said about intoxication could also be said about a drug’s other effects. Increased pleasure, lessened pain, increased concentration, and decreased anxiety, all these are effects of various addictive substances or activities, but no one of them is the effect of all addictions. Gambling need not lessen pain, and smoking just to feel normal does not increase one’s pleasure.
The one class of effects that may seem to be common to all addictions includes effects that are at least in part dependent on prior use. For example, concentration increases by smoking in part because, given regular nicotine use, concentration wanes as nicotine levels drop. As with intoxication, however, these effects may not be necessary to all addictions, particularly not to well-regulated addictions. Moreover, those effects that are the elimination of discomfort or withdrawal cannot be the wrong of addiction. That would seem to have things backwards. Withdrawal’s elimination is surely a good of drug use. Between withdrawal and withdrawal’s elimination, withdrawal is the one to avoid.
What if it is not the withdrawal that is the wrong, then, but the antecedent craving (Morse 2000, 2006; Wallace 2003)? Craving can certainly be present long before withdrawal, but it, too, may not be present in the well-regulated addiction. Note, however, that we have ventured from effects like intoxication and withdrawal, and, by looking at craving, we are now looking more closely at what addiction itself is rather than at any effects of addictions. This is for the best. Given that there look to be no effects common to all objects of addiction, any wrong of addiction must be in addiction itself, not in such effects.
To distinguish addictions from passions, consider that addicts, at least when it comes to their addictions, seem to make bad choices, choices to pursue the object of the addiction instead of more worthwhile goals (Volkow, Baler, and Goldstein 2011). [End Page 31] Some of those bad choices may account for the ‘manifestations’ of dependence that the DSM specified: continued use despite decisions to cut down, legal problems, health problems, and so on.
Now, whether the addict genuinely makes choices is a fraught issue. Some of her behavior looks unfree and perhaps, in virtue of that, is not really a choice. But addictions, like passions, are greater patterns of behavior that are broadly intentional, guided by what the person wants or chooses or even values (Levy 2006). The addict may intentionally buy an extra pack of cigarettes for tomorrow or decline a teetotaler’s dinner invitation, even if she would be helpless to resist—thus possibly not choose—an immediately available temptation. Addictions do not prevent the addict from making all choices, nor are addicts simply responding to irresistible stimuli—as if the heroin user uses whenever he unpredictably happens to run across heroin (cf. Sripada 2014). Likewise, for the non-addictive passions: a parent may not freely choose to grab his toddler who steps off the curb, but he intentionally positions himself between the toddler and the road.
The addict’s larger patterns of behavior are guided in ways that are open to assessment, even if some actions within that pattern are not genuine choices.5 For example, the pattern may be harmful, or worthwhile for anyone to have, or a waste of time, or fine if you are into that sort of thing, or misguided, or amusing, and so on. Corresponding with many of these assessments are assumptions about the motives: when I say watching spaghetti Westerns is fine as an amusement but not something worth doing every day, I am at least assuming that amusement is a possible motive in this case. Focus on these patterns and their underlying motives as a way of getting at what is wrong with addiction.
My focus on an underlying motive does not imply (nor do I think) that the addict simply suffers from some flaw in her thinking, whether that flaw is characterized psychologically or neurologically (Graham 2010; Schroeder 2010). There are many problems with such a view, but consider an empirical one: it is difficult to find any such flaw in the addict’s thinking that would clearly distinguish addictions from passions (cf. Volkow et al. 2011). Addictions are clearly neurologically reinforcing, but the suggestion cannot be that only addictions are neurologically reinforcing, because appetitive and passionate behaviors are also neurologically reinforced. And the neurological evidence does not (yet) show unequivocally that addictions are reinforcing in a greater or different way than passions are (Carter and Hall 2010; Iguchi and Evans 2010; Kalant 2010; cf. Everitt and Robbins 2005; Foddy and Savulescu 2010; Hyman 2005; Montague, Hyman, and Cohen 2004; Schroeder 2010). Indeed, part of the explanation why certain things are addictive is that they co-opt the same neurological mechanisms of reinforcement used by our natural appetites and passions (Bray et al. 2008; Childress et al. 1999; Kelley and Berridge 2002; Nestler 2005; Hollerman and Schultz 1998; Tobler, Fiorillo, and Schultz 2005; Schultz, Dayan, and Montague 1997).
Even if it turns out that there is an underlying neurological difference between addictions and other patterns of behavior, this does not entail that the difference between addictions and passions is entirely a biological or psychological discovery to make. The incredible biological and psychological research on addictions is research done on what we have an antecedent—if rough—understanding of. The research may eventually change our ordinary understanding, but we undertook the research with a rough understanding of which things count as addictions and which as passions or appetites. As George Ainslie puts the point, addiction research “doesn’t do much to change our concepts” related to addiction; rather, such research “only lets us see in detail the…process that used to happen inside a sealed box” (Ainslie 2000, 77). Such discoveries about the differences between addictions and passions or appetites are made in the light of the distinction we already understood.
That distinction we already understood, I assert, is that addicts engage in patterns of behavior that we assess as based on some incorrect or misguided motive. The smoker and the passionate philosopher both value something quite highly. The philosopher’s pattern of action is based on an unusually high (and statistically rare) but permissible value, one we all likely agree is fine if you [End Page 32] are into that sort of thing. The addict’s pattern of action, on the other hand, is misguided in some way, based on a misvalue of the importance of smoking.
I use ‘values’ and ‘valuing’ with some trepidation because these terms are often thought of as moral, and I do not intend that. I intend ‘value’ as a morally neutral term that specifies the motives underlying a pattern of behavior as the abstract, relatively stable, non-instrumental ends or goals that organize that pattern of actions. My doing philosophy is motivated by my ‘valuing’ philosophy (or valuing resolving puzzles, or finding truth, or engaging with smart people—for simplicity, I am speaking throughout as if there is only one operative motive) and one’s smoking is motivated by one’s ‘value’ of smoking (or of dealing with anxiety, or of looking cool, or whatever the actual motive is).
Using ‘value’ and ‘valuing’ in an ordinary way, we can further distinguish, as we must, between the values we claim to have and the values we in fact do have, those we demonstrate in our actions. These need not align, and my claim is only about those values that do motivate some pattern of behavior, not about what we say motivates our patterns of behavior.
We misvalue when our pattern of behavior is motivated by some abstract, relatively stable, non-instrumental ends or goals that we should not have, or that we should not weight as heavily as we weight them. This, I claim, is a trait that addictions do not generally share with appetites and passions.
What this means is that, in describing a pattern of action as an ‘addiction,’ both in ordinary usage and in clinical usage, we are making a claim that the motivation underlying the relevant pattern of action is wrong to have (or to act on). For example, in claiming that one is addicted, I ipso facto claim that she is wrong to value the object of her addiction more than its alternatives: the heroin addict is motivated to spend more time and money on heroin than he should, and the gambler is motivated to spend more time and money on gambling than she should. The wrong of addiction is that the motive on which the addict’s pattern of behavior is based is a motive that she should not have, or that should be less important than other motives.
These are very general claims about one’s motives—even ignoring my presentational simplification that a pattern of action is motivated by a single motive. We may further ask what such motives are and why they are wrong to have. The concept of addiction itself is neutral on these questions, and I suspect that people may use ‘addiction’ as I claim they do while disagreeing about what exactly is wrong in the underlying motive. For example, one person may believe that the wrong motive of the heroin user’s pattern of actions is that it esteems pleasure well beyond what is reasonable, whereas another believes it soothes his anxiety instead of confronting it, and a third believes it is just a waste of time. To disagree that this is an addiction, to claim that this is not an addiction, would be to claim instead that there is no such mistaken value underlying the pattern of action: maybe it is too bad that heroin is vilified and has side effects, and maybe there are better things to do with one’s time, but the motive that explains this pattern is not mistaken, not wrong to have, so a pattern of heroin use would not be an addiction.
The account helps us to understand the difficult cases of the sort I suggested, although it will not simply resolve them for us. If we wonder whether a person genuinely is an addict, as we may in the case of Jeff, our heavy caffeine user, we are first wondering whether Jeff is motivated to use more caffeine than he should. Making that judgment is difficult for all of the reasons that making any normative judgment about what people should be doing given their circumstances and alternatives is difficult: What else should Jeff be doing instead? Why is he drinking so much coffee? Is he postponing the boring inevitability of bedtime, or is he just accomplishing more while awake? What supports my proposal is that the debate about whether one is motivated to do something—drink coffee, in this example—more than one should do it will reflect the debate about whether the person is addicted: it will be incoherent to claim that a person is addicted but there is nothing that she is doing wrong, nothing she should be doing instead, nothing about her motivation that we would correct. [End Page 33]
Notice that this does allow for the judgment that, all things considered, the addict should remain an addict. That may seem puzzling because I just claimed that, in saying she is an addict, we are saying that there is something wrong with her motivation. But saying that there is something wrong with one’s motivation is not to say that, all things considered, one would be better off without it, or better off with an alternative motivation. The prolific mathematician Paul Erdös used amphetamines daily, and, after stopping for a month to win a bet with a friend that he could quit, said to the friend upon restarting his daily habit: “I’d get up in the morning and stare at a blank piece of paper. I’d have no ideas, just like an ordinary person. You’ve set mathematics back a month” (Hoffman 1999, 16). If mathematics is valuable enough, perhaps we should be thankful that Erdös was an addict, because his addiction led to so many insights. If his motivation in using was merely to concentrate, I may even doubt that he is an addict at all. But there is room to think that one is both an addict and is better off as an addict than not.
My account may explain the distinction between appetites and passions, but many other non-addictive patterns of behavior are based on misvaluations—I may run because I believe it makes me immortal—so misvaluation is not yet a complete explanation of addiction. Further, the example of Erdös raises another issue. Erdös was able to quit using amphetamines for a month, and we often take quitting to be a mark that one is not an addict. (Erdös himself said that his quitting proved to himself that he was not an addict.) In the next section, I complicate this account of addiction, which also explains why an ability to quit (or its lack) is indicative but not dispositive of whether one is addicted.
To explain what addiction is, and why it is more than a pattern of behavior based on a misvaluation, I draw in some psychological evidence. I do this by way of observing that an addict’s misvaluations would chronically and stubbornly resist contrary evidence. Non-addicts generally come to accept evidence of their misvaluing something— eventually. If we discovered conclusively that running marathons led to drastically shortened lives and attentive parenting caused dangerous psychopathy, then (eventually) those passions would be kept in check. Those passions are, in fact, kept in check by much weaker contrary evidence: marathon training often yields to the realization that one has had regular, non-debilitating injuries, and parents, aware of the dangers, try to resist becoming ‘helicopter parents’ (Cline and Fay 1995). Indeed, even our strongest non-addictive appetites, for food, drink, and sex, are resisted when the benefits of abstinence or postponement seem worth the sacrifice: witness religious fasts, diets, and the general absence of impulsive public sex.
The benefits of abstaining from addictions, however, do not often stop the addict.6 If the putative addict quits simply in response to learning evidence of its deleterious effects, we would be hard pressed to call her an addict. This is also true if she continues despite genuinely acknowledging the evidence: imagine the smoker who, although she realizes it will destroy her health, is the executive of a tobacco company whose livelihood depends on her actively endorsing smoking by continuing herself to smoke.
Moreover, even when addicts do try to respond to the evidence, they have difficulty keeping their resolutions to end their addictions, even forming such resolutions in response to the perceived benefits of abstinence. Addicts often ‘quit’ multiple times, each time followed by an eventual relapse. This is a common enough occurrence that many have taken this to be the distinguishing feature of addiction, calling it a “chronic, relapsing disorder.”
An addiction’s resistance to contrary evidence shouldn’t be overstated. The addict’s misvaluations are not forever fixed. Those values change both in response to things other than evidence (e.g., to medication) and, sometimes, in response to evidence. An addict may eventually quit, and the health costs of smoking may ultimately play some role in her quitting, but she does not quit simply in response to such evidence.
My position until this point is this: addiction is the misvaluation on which the addictive pattern [End Page 34] of behavior is based. Our ordinary and clinical view of addiction also adds to this position that this misvaluation would chronically resist contrary evidence. The question is why addicts chronically resist contrary evidence.
Many suggestions of what is wrong with addiction—for example, mistaken motivation, self-deception, neurological changes—should be seen more properly as explanations of how an addict avoids or resists contrary evidence by explaining how such a misvaluation is reinforced. This topic, which deserves and has received a great deal of attention in psychological and neurobiological research, can complete a philosophical account of addiction as a compulsive misvaluation. Moreover, the empirical research can be incorporated into a general account of addiction without assuming that the empirical findings are all there is to such an account.
I will not do justice to the research on addictive reinforcement, but I do mention some possibilities. One explanation of addiction’s chronic resistance to contrary evidence is that the addict feels conflicting desires that weaken her willpower. This would directly account for chronic relapse by the addict whose more powerful desire is sometimes to quit and sometimes to use. Such conflicting desires are no doubt often present (Ainslie 1992, 2001; Wyvell and Berridge 2001), but conflicts are present for the non-addict, and it is probably a false generalization that all addicts have internal conflicts (Burroughs 2003; Frankfurt 1988; Middendorp 2009). An addict may never desire or try to quit. It therefore cannot be an internal conflict that makes an addiction resistant to contrary evidence.
Another tempting explanation is that the addict has strong desires or cravings that are or create a certain amount of ‘cognitive static’ when one thinks about things other than the object of the addiction (Goleman 1985, 40; cf. Watson 2004). Such ‘static’ presumably makes it easier to focus on the object of the desire, or things associated with the object, and to pursue those (Goldstein 1994). This cognitive-static explanation nicely explains addiction’s seeming irresistibility without making the addict into a marionette, controlled possession-like by an addiction. Unfortunately, it also fails to explain why addiction’s misvaluations are resistant to change. Not only would such a cognitive-static model also apply to non-addictive passions, but the mechanism by which cognitive static motivates is importantly underdescribed. ‘Cognitive static’ sounds suspiciously like a recurrent, anxiety-inducing thought that intrudes into one’s ordinary life, in a way similar to the thoughts of obsessive–compulsive disorder. To simplify somewhat, obsessive–compulsive disorder’s obsessive thoughts are anxiety-inducing for the sufferer, and her compulsive behaviors are attempts to reduce the anxiety. The addict is not a sufferer of obsessive–compulsive disorder. An addict’s cravings need not be anxiety-inducing, nor need her actions be intended as anxiety-reducing. The addict’s repeated thoughts about the object of her addiction may be enjoyable or invigorating or relaxing instead. What the addict wants is to obtain the object of the addiction—which perhaps she cannot stop thinking about—not to obtain relief from the anxiety of thinking about that object.
If such ‘cognitive static’ is not an unpleasant state, like anxiety, that the action attempts to relieve, then the cognitive-static model sounds most like a phenomenological description of desires in general, of which addictive desires are instances. Cognitive static would ‘compel’ action only by continually reminding the actor of her desire to act in some way. It would not explain addiction’s resistance to change, however, to say that addictions rest on desires.
Similar reasons count against some less common explanations of addiction’s resistance to contrary evidence. If addiction’s desires are ‘strong’ or ‘weighty’ or ‘irresistible,’ that does not amount to an explanation. Such putative explanations may not hold true of all addictive desires. Even if the explanations do hold true of all addictive desires, they seem to be a restatement of the explanandum: addictive desires may be especially strong or weighty or even irresistible, but why? The question remains: why do addictions chronically resist contrary evidence?
To explain addiction’s chronic resistance to contrary evidence, take seriously the thought that the addict may be unable to keep a resolution to quit (Holton 2009). This is not because addictive motivations [End Page 35] overpower or seduce, but because such motivations are simply unaffected by contrary resolutions—indeed, they are unaffected by any thought at all about one’s preferences, including thoughts about the evidence against the addictive behavior. Therefore, one may prefer or even resolve not to take more of a drug, or to gamble anymore, but the addictive value—the abstract, relatively stable, non-instrumental end or goal that motivates the action, whatever that end or goal may be—is unaffected by that preference or resolution.
Some highly habitual cases may clearly be cases of motivation unaffected by any thoughts or resolutions. If a chain smoker leaves her cigarettes where they always are next to her on the desk, she may habitually pick one up and light it while thinking entirely of something else. Indeed, she may do this even if she prefers not to smoke or even explicitly resolved not to smoke before foolishly leaving her cigarettes next to her. Less habitual, but equally unthinking, cases are central to understanding addiction, as I consider next.
Call an action ‘impulsive’ if it issues from no significant thought about what to do at all. Of course, even impulsive actions are caused, and, insofar as it is genuinely action rather than a mere reflex, it is end- or goal-directed; but, an impulsive action does not originate in one’s thoughts about what to do. For example, I have been trying to quit smoking, but I have told no one. One day, I run into a friend who offers me a cigarette. I unthinkingly, habitually say ‘sure’ and reach for the cigarette. I may realize that I am accepting, but I do not debate with myself in any way what I am doing. I simply act. Of course, as I realize what I am doing, I could stop, I could return the cigarette before lighting it, I could decide to take a deep breath and recite my list of reasons that I am trying to quit smoking: impulsive actions need not preclude all further thought or decision. (Of course, each of my subsequent actions may themselves be impulsive, too.)
Even if, immediately before the impulsive action, I had been considering whether I would have a cigarette if available, my action may still be impulsive. Imagine this same situation, in which I impulsively decide to take the cigarette, but in this new situation I have been walking across campus thinking about whether I would like to have a cigarette. I am considering the pros and cons as I form my hypothetical preferences or resolutions, although I think this is an entirely intellectual exercise because I do not actually have any cigarettes. But then my friend shows up, and the story takes off where it did before: I impulsively respond to his offer. It is more plausible in this case that there was deliberation about the action, but even here the action is impulsive, because my taking the cigarette bypassed my thoughts about smoking—those thoughts about smoking that I had just had.
On this proposal, addiction’s misvaluation chronically resists contrary evidence because, regardless of prior thought or any resolution to quit, many of the addict’s addictive actions are impulsive (cf. Jentsch and Taylor 1999; Robinson and Berridge 2003). My suggestion is that addicts sometimes act impulsively, not on the basis of current or prior thoughts about what to do. Even such impulsive actions are guided, in part, by the same ends or goals that explain the rest of addiction’s pattern of behavior, so the smoker impulsively reaches for a cigarette. Such impulsive action keeps addicts from being able to carry through on their preferences or resolutions to stop, and may lead them to act even without explicitly deciding to act.
Such impulsive actions, in the case of addictions, then serve to reinforce the misvaluation that underlies the addiction. The primary way in which some impulsive actions can reinforce a misvaluation is that most if not all of the typical objects of addiction are neurologically reinforcing. For example, using amphetamines makes one more likely to notice and remember it, to overestimate its pleasurable properties, and to underestimate related costs, even after one has quit using it (Wyvell and Berridge 2001). In this way, impulsive actions are reinforcing in the way that deliberate use of those substances is reinforcing. Because the addict’s motivation guides even impulsive actions—impulsive actions, after all, are not unguided seizures or reflexes—impulsive actions are part of the addictive pattern of behavior based on the same misvalue, so the addiction is [End Page 36] self-reinforcing through such impulsive, unthinking behaviors.
On this account of reinforcement, impulsive actions would maintain the addict’s misvaluation in the face of contrary evidence. Contrary evidence can only influence a person’s thoughts or preferences about what to do. If impulsive action bypasses those thoughts and leads to reinforcement of the initial misvaluation, then contrary evidence will have a hard time changing a person’s misvaluations. The addict thus continues to be motivated in her pattern of behavior based on a misvalue, and her motivation is resistant to change.
Notice that this account is neutral on when such patterns of behavior are reinforced enough—or are even pattern-like enough—that they constitute an addiction. Which is as we may expect: addictions rarely have clear, determinate beginnings, and we may legitimately wonder whether someone’s behavior is yet sufficient to constitute an addiction. Even when one’s first gamble or first drug use makes it likely that this will become a pattern, it is not yet a pattern, and he is only a prospective addict. Many things could yet happen to interrupt the pattern before it begins, keeping the reinforcement from occurring, keeping the person from becoming an addict.
This is equally true for the end of the addiction, although defining the end of an addiction is even more complicated by various therapeutic and ideological positions about how addictions end, or whether they ever genuinely do. (For example, is abstinence the only end of an addiction, or is more limited use?) My account asserts that an addiction is over when there is no longer a pattern of behavior typical of that addiction, although we may still discuss when the pattern ends. This proposal allows that the reinforcement that underlay that pattern of behavior may remain, that the addicts may remain more impulsive or unthinking in ways that will allow the pattern of behavior more easily to restart. This, too, is as we would expect, because relapse is a common, if not universal, feature of addiction.
I have proposed herein that addiction is the misvaluation on which a pattern of behavior is based, and that this misvaluation would chronically resist contrary evidence. Such misvaluation is resistant to contrary evidence because addicts act in impulsive, unthinking ways that reinforce the misvaluation.
This position, albeit schematic, already seems to succeed in distinguishing addiction from non-addictive dependences and passions by identifying something wrong with addiction. One wrong is the misvalue on which the addiction-related pattern of behavior is based. That pattern of behavior is motivated by a value that one shouldn’t have, by a misvalue.
Moreover, the addict’s misvaluation chronically resists correction. An inability to change one’s misvaluations in response to contrary evidence is a particularly pernicious wrong because it impedes one’s ability to self-govern in a way that rational creatures do, in a way that responds to changing conditions and evidence. Importantly, this inability to self-govern in response to what is best would account for the ways that addiction manifests itself, as illustrated in the DSM-IV criteria from which this discussion began.
This explains why we often test whether someone is addicted by asking if she could quit. If this pattern is an addiction, then there is something mistaken about acting in this way at all, so there are presumably reasons that it is worth quitting. If one nevertheless does not quit, particularly when the reasons to quit are made explicit, that suggests that one’s underlying motivation has been reinforced in such a way that, even though the motivation is mistaken and the person has seen this and perhaps even resolved to quit, the reinforcement is too strong. Being able to quit does not prove that one was not addicted, just as failing to quit will not show that one was addicted, but the reasoning behind using such a test fits the account I have provided.
I mention two upshots to my account. First, the account explains how one could be addicted to something good, like taking pain medication when in pain. The worse my pain is, and the fewer the [End Page 37] side effects of the medication, the harder it is for me to overvalue the medication and undervalue the side effects: the harder it is to be addicted, although it is certainly still possible. This topic deserves its own development, but it seems to be what a correct characterization of addiction would say, and it follows naturally from this account.
The account also has the benefit of explaining less stereotypical cases of addiction. For example, the addict who is happy with his addiction, who does not want to change, is a puzzling case for accounts of addiction that take motivational conflicts or strong cravings to be central to addiction, because someone happily addicted may have no such conflict and no particularly strong cravings, particularly if he regulates his addiction well. But, because happily valuing something does not entail that one should value it, my account explains how he may still be an addict.
Finally, my position indicates that the puzzling questions of whether Renton and Jeff are addicts ultimately turn on whether Renton and Jeff are right to value heroin and coffee, and act on those values, as much as they do. For example, when we discuss whether Renton is an addict, we consider whether his values are correct, even given his options and his community. Likewise with Jeff’s caffeine use: we address the apparently clinical question of whether Jeff is addicted by considering the more philosophical question of what, if anything, is wrong with his actions.
Feature Article: What is Wrong with Addiction
Commentary: Words, Worlds, and Addictions
Response: Addiction by Any Other Name
The author owes particular thanks to Pamela Hieronymi, A.J. Julius, David Jentsch, Calvin Normore, Walter Sinnott-Armstrong, and Gideon Yaffe, as well as to Ben Chan, Owen Flanagan, Diane Kierce, Bob Kane, Brent Kious, Doug MacLean, Jeff Sebo, Nandi Theunissen, Stephen John White, and to audiences at UCLA, the Rocky Mountain Ethics Congress, Rice University, Loyola University—Chicago, the Catholic University of America, and the undergraduates at UCLA and Duke who took my courses on addiction.
1. The current (as of 2013), 5th edition is the DSM-5, although most of the contemporary literature on addiction developed under the DSM-IV, which was current from 1994 to 2013. The differences between the two are largely irrelevant to my overall claim—with exceptions mentioned in the text—so my claims focus explicitly on the criteria of the DSM-IV.
2. Criterion 2b would diagnose withdrawal if the addict used the drug “to relieve or avoid withdrawal.” In Renton’s dystopia, however, it is unclear that the addict would use heroin in order to relieve or avoid withdrawal, although his use would in fact relieve or avoid withdrawal. This is an important distinction, one that ultimately weakens views of addiction that hold that an addict necessarily acts in response to felt or anticipated withdrawal (Koob and Le Moal 2001).
3. If Renton’s case seems too farfetched to warrant our attention, consider a so-called ‘functional alcoholic’ instead: by some standards he is dependent, and by others he is not. Our own community standards regarding alcohol are more complicated and conflicting than those regarding heroin, though, so it is harder to show the alcohol case succinctly and convincingly.
4. I choose not to discuss willpower explicitly, both because I think addictions and non-addictive appetites and passions interfere with willpower, but also because the term ‘willpower’ brings to many minds certain presuppositions that may be misleading, for example, that there is a mental faculty corresponding to the will (however understood) that is exercised in resisting temptation, or that willpower is adherence to one’s own explicit resolutions (Holton 2009), rather than that (as I find plausible) ‘willpower’ is a shorthand for a collection of self-control techniques that one practices until they are more or less routine. To eliminate confusion, then, I omit the term ‘willpower,’ although the related ideas are also discussed.
5. Even for these actions that are, in some sense, unintentional, it is certainly not a consensus view that they cannot be assessed (Arpaly 2000; Arpaly and Schroeder 1999; Sher 2009; cf. Hieronymi 2008).
6. A well-worn quote in addiction literature has it that an alcoholic told Benjamin Rush, “Were a keg of rum in one corner of a room, and were a cannon constantly discharging balls between me and it, I could not refrain from passing before that cannon, in order to get at the rum” (James 1890, 543). This strikes me as too hyperbolic to merit the attention it receives, but it can at least be conceded to contain this truth: the benefits of abstinence probably must be greater for the addict to stop his behavior than for the non-addict to stop hers (Chiu, Lohrenz, and Montague 2008).