Addiction and the self

Correspondence Hanna Pickard, Johns Hopkins University. Email: h.pickard@jhu.edu Abstract Addiction is standardly characterized as a neurobiological disease of compulsion. Against this characterization, I argue that many cases of addiction cannot be explained without recognizing the value of drugs to those who are addicted; and I explore in detail an insufficiently recognized source of value, namely, a sense of self and social identity as an addict. For people who lack a genuine alternative sense of self and social identity, recovery represents an existential threat. Given that an addict identification carries expectations of continued consumption despite negative consequences, there is therefore a parsimonious explanation of why people who identify as addicts continue to use drugs despite these consequences: they self-identify as addicts and that is what addicts are supposed to do. I conclude by considering how it is nonetheless possible to overcome addiction despite this identity, in part by imagining and enacting a new one. Importantly, this possibility requires the availability of social support and material resources that are all too frequently absent in the lives of those who struggle with addiction.


UNDERSTANDING ADDICTION
The view of addiction as a neurobiological disease of compulsion emerged over the course of the twentieth century partly in response to a prior view of addiction as a sin or a vice. This opposing moral model of addiction has two parts. The first part claims that drug use is a choice, even for those who are addicted. The second part is the moral condemnation of this choice. According to the moral model, addicts succumb to the temptations of a Dionysian life of pleasure, which they could instead resist, thereby revealing themselves to be people of bad character.
Compulsion is the antithesis of choice. The view of addiction as a neurobiological disease of compulsion rejects the first part of the moral model. It thereby also rejects the possibility of condemning addicts for their choices to use drugs once addicted, for, according to this view, they have none. 2 But an element of the second part of the moral model of addiction is nonetheless typically retained by the view of addiction as a neurobiological disease: the moral condemnation not of addicts, but of drugs, and, relatedly, of pleasure got from drugs.
Advocates of the disease model often emphasize that, in addiction, drug use persists even though there is no more pleasure to be had: addicts no longer "like" their drug of choice (Holton & Berridge, 2013; M.J.F. Robinson, T. E. Robinson, & Berridge, 2018;Volkow, Koob, & McLellan, 2016). 3 This emphasis is striking, given not only that very few human studies examining loss of pleasure have been conducted, but that those that have do not appear to support it. Laboratory experiments designed to mimic cocaine binges in subjects who have a history of use show only a modest decrease in reported euphoria over the period of consumption (Foltin & Fischman, 1991;Reed et al., 2009;Ward, Haney, Fischman, & Foltin, 1997); note that there is no notable decrease in subjects with no history of use (Wachtel & deWit, 2009). Additionally, a clinical retrospective study of 40 inpatients for treatment for cocaine dependence found that 100% of subjects reported persisting euphoria, with only 27.5% reporting any decrease whatsoever (Bartlett, Hallin, Chapman, & Angrist, 1997); while a community retrospective study of 100 treatment-seeking long-term cocaine users found on average a decrease in euphoria ratings since cocaine initiation, but that euphoria nonetheless persisted (Small et al., 2009). Pleasure reduced is not pleasure eradicated: on balance, taking this evidence together, there would seem to be pleasure from drugs in addiction for most addicts. To be clear, I do not claim that there are no addicts for whom pleasure from drugs is entirely eradicated: there may be. The point is rather that the existing scientific evidence does not support the claim that, in general, addiction eradicates drug euphoria, which raises the question of why advocates of the view of addiction as a neurobiological disease of compulsion are so eager to emphasize that it does. 4 One part of the answer may be that, if addicts get pleasure from use, then it is natural to infer that they use for pleasure. If so, then there is reason to question the claim that use is compelled, for it would appear goal-driven (see below). But, another part of the answer may be that an element of the moral model remains: there is an unarticulated assumption that, even if addicts do not use for pleasure, the mere fact of gaining pleasure through use -even use that is compelled -would warrant moral condemnation; hence, in order to avoid moral condemnation, it is important to emphasize that addicts get no pleasure from drugs. 5 Fighting the stigma surrounding addiction and the moral condemnation of addicts is of paramount importance; but it is not likely to be lastingly achieved through claims that are at best unjustified, and quite possibly false. The alternative to refusing to acknowledge that addicts may get pleasure from drugs is to fight moralism about drugs: to argue that there is nothing intrinsically morally wrong in using drugs (Husak, 2002; see too Pickard, 2017b) -for pleasure or for other reasons (see below). For what, precisely, is supposed to be morally wrong about using drugs for pleasure or other ends? If this challenge does not immediately give you pause, I ask you to consider that alcohol is a drug whose consumption most of us sanction and many of us enjoy and use as a means to a variety of ends, yet there is no good neurobiological or public health distinction between it and classes of drugs that are currently both illegal and morally condemned (Nutt, 2012); and that, contrary to popular belief, the majority of people who have tried drugs of all classes do not become addicted (Heyman, 2009;SAMHSA, 2005;Zinberg, 1984). This includes heroin, which can be successfully used recreationally, a phenomenon known as "chipping" (Zinberg, 1984;Zinberg & Jacobson, 1976). Of course, drug consumption carries risks, and, in some circumstances, these risks may include not only harm to self but also to others. In which case, consumption may be morally wrong because of the risk of harm it poses to others. But if so, this is contingent on probable consequences given the circumstances, not essential to consumption in itself. Drugs and drug consumption are not intrinsically bad -quite the contrary.

The multiple functions of drugs
Although it can sometimes take time to learn to enjoy and appreciate drugs (Becker, 1963;Zinberg, 1984), they are a reliable source of pleasure for many users, non-addicted and addicted alike. But drugs serve other well-known and well-documented functions too (Müller & Schumann, 2011). These include relief from pain, fatigue, stress, boredom, negative emotions, and psychological suffering; improved physical energy, cognitive ability, social connectedness, and sexual experiences; and mindaltering and self-altering experiences, including spiritual experiences. These functions are by no means exhaustive; indeed, I suggest a further function in Part II, namely, to secure a sense of self and social identity. It is also the case that different drugs typically serve different functions: opiates relieve pain and suffering; amphetamines increase energy and cognitive ability; MDMA heightens social connectedness; psychedelics produce experiences that people describe as some of the most meaningful of their lives (Belser et al., 2017;Noorani, Garcia-Romeu, Swift, Griffiths, & Johnson, 2018; see below). Indeed, even rodents select different drugs in different environments, preferring heroin in home-cage and cocaine in non-home-cage contexts (Ahmed, Badiani, Miczek, & Müller, 2018) -presumably because of the interaction between the drug-specific effects on mental state and environment-specific demands on the animal. But, nuances aside, the key point is that drugs have tremendous value to people because of what they do for us: they are a means to many valuable ends.
Once stated, this point may seem obvious. Yet its importance to understanding addiction is often overlooked. Apart from the most extraordinary of circumstances, all addiction originates in nonaddicted drug consumption, which is goal-directed. Initial and subsequent non-addicted human drug choices are typically guided by explicit anticipation of their effects on mental state, whether these are known through testimony or prior first-person experience. In other words, we take drugs by choice: their expected value is why we use them. As noted above, I do not believe there is good scientific evidence for the general claim that, once the transition to addiction has occurred, drugs no longer offer any pleasure whatsoever. But, even if this were true, most if not all of the other functions of drugs listed above are not mediated by pleasure and continue to be secured through consumption deep into addiction. Most importantly, given the strong association between chronic addiction, co-morbid mental disorders, and socio-economic disadvantage and isolation (Alexander, 2018;Compton, Thomas, Stinson, & Grant, 2007;Heyman, 2009;Maté, 2009;Regier et al., 1990), drugs continue to provide relief from pain, fatigue, stress, boredom, negative emotions and psychological suffering (even if addiction creates its own suffering) (Pickard, 2012). For some addicts, they may also offer forms of positive self-identity and social experience, which they would otherwise lack (see Part II). In other words, drugs not only have value apart from addiction. Drugs can continue to have value in addiction.

The puzzle of addiction
The importance of the fact that drugs have value and that drug value persists in addiction is critical to understanding the nature of addiction and why it is so puzzling as a form of behaviour. From a clinical perspective, Substance Use Disorder (SUD) in DSM-5 (APA, 2013) is diagnosed by eleven polythetic criteria and ranges from mild (2-3 criteria), moderate (4-5 criteria), to severe (6-11 criteria), allowing for significant variation in symptomatology between people who meet the criteria for diagnosis. In addition to the pharmacological criteria of tolerance and withdrawal, the core SUD construct involves: (i) cravings and failures to limit use as intended as (ii) drugs come to occupy increasing time and attention at the expense of other pursuits and despite (iii) incurring severe risks and negative consequences. These consequences typically include drug-related mental and physical health problems, as well as loss of important relationships, social standing, employment, housing, and other significant goods. In countries that criminalise drug possession and stigmatize drug users, there is in addition the risk of criminal sanction and social ostracization. Although (i) craving and difficulties in self-control, alongside (ii) escalating use, are of course important aspects of the construct (see below), continued consumption in the face of drug-related risks and costs such as these is central to defining addiction (NIDA, 2009;WHO, 2004) and understanding both why it is a disorder as well as what is puzzling about it as a form of behaviour (Pickard, 2018). For there is no puzzle understanding non-addicted drug consumption: people choose to use drugs because drugs have value, as described above. But, once this is appreciated, then drug consumption should only be seen as going wrong or in some sense disordered and wanting explanation when the balance between drug costs and drug benefits seems to have tipped, such that costs appear to outweigh benefits and yet consumption persists. This apparent tip is critical to the nature of addiction and defines what is so puzzling about it. Put crudely, why do addicts persist in using when drugs no longer appear worth it (Bornstein & Pickard, 2020;Pickard, 2016Pickard, , 2018)?

The myth of compulsion
The view of addiction as a neurobiological disease of compulsion offers a parsimonious and powerful solution to this puzzle. To use a common metaphor, the explanation is that addiction "hijacks" the brain so that the desire for drugs is irresistible, and there is no longer any possibility of voluntary and valuebased choice. 6 On this view, there is no question that, if addicts could stop using, they would; but, they can't, which is why they don't. Neurobiological compulsion therefore explains why consumption continues when the balance has tipped and costs appear to outweigh benefits: it overrides any possibility of voluntary and value-based choice, stripping addicts of any ability to do otherwise (Pickard, 2015a(Pickard, , 2017a(Pickard, , 2018. 7 Notwithstanding its theoretical elegance and popular appeal, this solution is not empirically credible. The reason is simple: there are ever-increasing and converging lines of evidence from animal models and human addiction studies establishing that the majority of animals and humans alike respond to context-specific contingencies and choose non-drug alternatives across multiple choice settings. Addicts retain the ability to do otherwise in many contexts (Pickard, 2012(Pickard, , 2015a(Pickard, , 2016(Pickard, , 2017a(Pickard, , 2018. Research in animals has long demonstrated that, although rodents will escalate drug selfadministration in deprived settings where no alternative rewards are available, the vast majority will choose food or social rewards over drugs in forced-choice laboratory studies (Ahmed, 2010;Ahmed, Lenoir, & Guillem, 2013). Recent work suggests that social reward in particular is an especially potent alternative for rodents, with 100% opting for social over drug reward when both are immediately available, and irrespective of sex, drug class, training conditions, size of dose, length of abstinence since last dose, or "addiction score" based on a DSM-style model adapted to rodents; only delay or punishment of social reward choices affects choice (Venniro et al., 2018). 8 Put simply, if you give rodents who for all intents and purposes are "addicted" an alternative and better reward to drugs, they take it.
Human studies demonstrate similar flexibility in drug behavior. Despite the fact that we speak of addiction as a "habit" and some of its features appear habit-like (Everitt & Robbins, 2016;Everitt, Dickinson, & Robbins, 2001), there is nonetheless limited evidence in support of a generalized habit theory of addiction (Hogarth, 2018a): consumption is not automatic but remains a controlled process (Hogarth, 2018b). In other words, in many circumstances, the majority of addicts demonstrate the ability to respond to "contingencies" or "incentives" (as psychologists would put the point) or "act for reasons" (as philosophers would say) thus demonstrating that drug choice in addiction is voluntary and value-based.
As illustration, consider the following findings. In forced-choice laboratory studies offering the immediate opportunity to use crack cocaine or receive monetary reward, addicts frequently choose money over drugs (Hart, Haney, Foltin, & Fischman, 2000). In addition, contingency management treatment is highly effective (for a review see Zajac, Alessi, & Petry, 2018) offering positive rewards contingent on drug-free urine samples in the form of money, prizes, and most recently and successfully, employment (Silverman, Holtyn, & Morrison, 2016); rates of use are cost-sensitive (Ainslie, 2000); and correctional services can succeed in establishing abstinence by imposing costs for failure (Hawken & Kleiman, 2009;Levy, 2013). Lastly, large-scale epidemiological data suggest that the majority of people meeting criteria for SUD (including those with physical dependence) recover without clinical intervention by their late 20s or early 30s (Heyman, 2009) -and of course, anecdotal (Heyman, 2009) and clinical (Pickard & Pearce, 2013) stories abound of people going "cold turkey".
These multiple lines of evidence converge to establish that many addicts choose non-drug alternatives in the presence of drug options across multiple choice settings. Together, they underscore that drug consumption, even in addiction, typically remains purposive (Pickard, 2012) and involves choice (Pickard, 2015a(Pickard, , 2016(Pickard, , 2017a(Pickard, , 2018, as opposed to being compelled by craving alone. Hence, although craving, no doubt alongside limitations of self-control, is part of the solution to the puzzle of addiction, it cannot be the whole of it. Given the evidence demonstrating flexibility in behavior and responsiveness to incentives in addiction, the puzzle remains: addicts may really want to use drugs and may have to battle for control, but if they have the capacity to abstain, then, given how much is at stake, why don't they?

Five caveats
Before answering this question, there are five caveats that need stating for the sake of clarity and comprehensiveness.
First, the evidence cited bears on what must be true of a general theory of addiction. In general, addicts can respond to incentives, which means that a sweeping appeal to compulsion cannot explain the puzzle of persistent use despite negative consequences. But the evidence does not license any sure conclusion about what is true of a particular addict on a particular occasion. We all sometimes face specific circumstances in which, for one reason or another, we are unable to exercise a capacity we nonetheless have. This is true of addicts as much as any of us.
Second, the disease model of addiction cannot soften the meaning of compulsion in light of the evidence from the impossibility of not using drugs to the difficulty of not using drugs -on pain of loss of its explanatory power. There is no question that it is extremely difficult for addicts not to use drugs. But, once it is conceded that it is possible, then, given the negative consequences of consumption, simply appealing to the difficulty of abstaining does not suffice to explain why addicts persist. For, given the severity of the consequences, we would expect addicts to bear this difficulty, without further explanation as to why they do not. The point is not that this cannot be explained. Indeed, the following sub-section as well as Part II aims to explain it. The point is that it cannot be explained without resources additional to the disease model if the meaning of compulsion is softened.
Third, and relatedly, the fact that addicts respond to incentives and so use is not compelled does not entail that drug-associated cue-induced craving is no part of addiction. Unquestionably, it is. In animal models, drug-associated cues, similarly to drug priming, reinstate drug-seeking behaviour after extinction and forced abstinence (Bossert, Marchant, Calu, & Shaham, 2013;Girardeau et al., 2019;Shalev, Grimm, & Shaham, 2002) 9 . Although an association between craving and consumption in human addiction studies has long been contested and remains far from well established (Tiffany & Carter, 1998;Wray, Gass, & Tiffany, 2013), there is nonetheless evidence that in human laboratory settings, stress and drug-associated cues predict first-person reports of craving, which is associated with subsequent relapse in cocaine ex-users (Sinha, 2006); and that, outside of the laboratory in daily life, craving is associated with consumption in smokers (Serre, Fatseas, Swendsen, & Auriacombe, 2015) and cocaine users (Preston et al., 2009). Cue-induced craving not only characterises periods of active use, but endures for months, possibly years, post-cessation; and the cognitive neuroscience of addiction has made significant progress determining relevant mechanisms (Holton & Berridge, 2013;Robinson & Berridge, 1993;Robinson et al., 2018). There is no question that craving is central to the lived experience of many addicts and an important part of any theory of addiction. But craving does not compel use.
Fourth, and again relatedly, the fact that addicts respond to incentives and so use is not compelled does not entail that limitations of self-control play no part in addiction. Crucially, craving requires effort to resist, which entails both costs (that can be rationally traded off against benefits; Shenhav, Botvinick, & Cohen, 2013) and the likelihood of failure due to simple mechanistic fallibility (Sripada, 2018), predicting occasions when self-control fails and consumption results. In addition, addicts display a range of decision-making anomalies, including some that bear on the exercise of self-control, such as reflective impulsivity, and risk and ambiguity tolerance (Verdejo-Garcia, 2018; see too Holton & Berridge, 2013). Self-control can be a battle: none of our wills is made of steel. Again, this is an important part of any theory of addiction. But, in general, the evidence establishes that addicts respond to incentives, thereby exhibiting a general capacity for controlled consumption -no matter how hard-won.
Fifth, the fact that addiction is not a disease of compulsion does not establish that it is not a disease at all. It remains possible that some (although not likely all, see below) of the neurocognitive processes underlying craving or drug choices in addiction are indeed pathological. For the record, my own view is that we ought at present to be agnostic about this question, because we simply do not yet know enough to answer it (Pickard, 2018). This is both because we need to know more than we currently do about how the brain is supposed to function in order to have a standard by which to measure dysfunction; and because we need to know more than we currently do about how the brain in fact functions in addiction in order to know whether it meets the standard thereby set for dysfunction. A commitment to truth therefore recommends agnosticism. But so too does a concern for practical consequences. Although labelling addiction a disease is often claimed to fight stigma and remove barriers to funding for research and treatment (Volkow et al., 2016), empirical studies suggest the effects of disease models of mental disorder are mixed. On the one hand, the label may reduce attributions of personal responsibility (Haslam & Kvaale, 2015) and improve support for services (Perscosolido et al., 2010). On the other, there is evidence that it increases general stigma (ibid.), including perceptions of dangerousness, unpredictability, and difference -cementing a desire for social distance in others, and worsening pessimism and hopelessness in sufferers themselves (Haslam & Kvaale, 2015). It may also undermine the sense of agency and empowerment that is typically required for recovery (Pickard, 2015d(Pickard, , 2017b. On balance, therefore, we should be open to the possibility that addiction may prove to be a disease of choice at least in some instances -but we should be in no rush to embrace this conclusion out of concern for addicts.
With these caveats in hand, let us return to the question: if addicts have the capacity to abstain, then, given how much is at stake, why don't they?

Two broad approaches to solving the puzzle of drug choice in addiction
The burden of the discussion thus far has been to establish that drug choices in addiction are voluntary and value-based. In consequence, the answer to the question just posed is simple in the abstract. We need to understand how addicts weigh drug costs and benefits in their decision-making, such that on balance consumption is expected to have more value than abstinence. Broadly speaking, there are two kinds of solution, capable of working in tandem.
The first kind of solution appeals to anomalies in addicted decision-making that serve either to boost the expected benefits of consumption, or to hide the expected costs, from the perspective of addicts themselves. For example, the disposition to discount the future relative to the present is a common feature of human psychology, but addicts have steep discount rates compared to the norm (Bickel & Marsch, 2001;Bickel, Koffamus, Moody, & Wilson, 2014); meanwhile, the benefits of drug consumption are typically immediate, while both the costs of consumption and any benefits that might accrue from abstinence are typically delayed. Temporal discounting therefore serves to boost the expectation of drug benefits and minimize costs (Ainslie, 2018;Heyman, 2009), shedding light on how an addict might choose to use drugs now despite the consequences that will follow later. Similarly, addicts are notoriously prone to denial, which, in addiction, functions to block the costs of consumption from view altogether, so there is no possibility that they factor in decision-making (Pickard, 2016). Additionally, "memory sampling" of early and highly rewarding drug experiences may bias present choices towards "chasing the first high" while side-lining more recent but less rewarding drug experiences that would encourage abstinence (Bornstein & Pickard, 2020). Such decision-making biases are common features of human psychology; possibly, they are sufficiently dysfunctional in at least some cases of addiction to count as pathological and render these cases a disease of choice. But, whether these biases count as pathological or not, they can nonetheless explain the puzzle of persistent use in the face of negative consequences, by revealing how decision-making in addiction can be skewed in favour of consumption and away from abstinence.
The second kind of solution reveals benefits to drug consumption that are visible to addicts and weigh in their decision-making, but may be hidden from the perspective of outside observers. At the start of this paper, I invited you to imagine being an addict. For many of us, this is no easy task. Most people who do not have a diagnosis of SUD or who have not lived intimately with a person who does have limited real understanding of what it is like to be an addict. Similarly, most people reading this paper are unlikely to really understand what it is like to live in the life circumstances associated with addiction. Recall the striking finding I noted above, that the majority of people meeting criteria for SUD spontaneously recover by their late 20s or early 30s. This finding raises an important question: who doesn't? The answer is that the minority of people who do not spontaneously recover from addiction typically come from underprivileged backgrounds of severe adversity and limited socioeconomic opportunity (Compton et al., 2007;Heyman, 2009;Maté, 2009) and suffer from a range of mental disorders in addition to SUD (Regier et al., 1990). Addiction flourishes in conditions of poverty, isolation, humiliation, pain, and hopelessness, of the sort many of us who are more privileged will never know. If you live in such circumstances, drugs may be the only thing that brings any relief from suffering and despair. Until those of us on the outside of addiction face up to the life circumstances associated with it and the lack of real alternatives for many addicts, we will fail to see the value of drugs. People make choices relative to the circumstances in which they find themselves. And, in these kinds of circumstances, no matter the negative consequences, on balance, drugs can make life better, not worse (Pickard, 2012(Pickard, , 2015a(Pickard, , 2017a(Pickard, , 2018.
In addition, these circumstances may be compounded by a negative self-concept which is part and parcel of a self-destructive mindset. SUDs are associated with personality disorders and complex mental health needs (Maté, 2009;Pickard & Pearce, 2013), including deliberate self-harm. This can take the form of self-directed violence, such as cutting and burning; but also sexual and other forms of risktaking behaviour, overdosing, and, arguably, drug misuse quite generally. For addicts with this mindset, it is not only that the negative consequences of use may not weigh with them because they do not care about themselves; the negative consequences of use may in fact count as benefits. For they serve to express a self-destructive aim and outlook. Once again, for those of us who do not share or have much familiarity with people who have this mindset, it can be extremely hard to understand (Pickard, 2105b(Pickard, , 2015c; but, once acknowledged, it reveals how not only the effects of drugs on mental state, but indeed the negative consequences of drug use, may have hidden value for a certain kind of addict. There are two important conclusions to draw from this broad discussion of how to solve the puzzle of addiction. The first is that there is no universal solution. Different explanations of why use persists despite negative consequences will be true of different people with SUD (and even on occasion of the same person with SUD at different times in their life). Some of the explanations I have suggested can work in tandem e.g. an addict prone to temporal discounting may also value drugs because of their unique capacity to relieve suffering given the life circumstances associated with addiction. But some of the explanations are incompatible, e.g. an addict in denial cannot also consciously embrace the negative consequences of addiction as a form of deliberate self-harm. This heterogeneity is to be expected, once addiction is recognised as a puzzle of choice. For, people make choices that are superficially similar for all sorts of different underlying reasons. In consequence, addiction is a unified construct at a superficial level only: it is defined (in part) by persistent use in the face of negative consequences. At an underlying aetiological level, the construct fractures: the nature and explanation of addiction is significantly heterogeneous between addicts.
The second conclusion is that we need to be real about the value of drugs not only in general but in addiction -as well as our own limitations in recognizing it. I shall now argue that, for some addicts, persistent use in the face of negative consequences is explained precisely because addiction is a source of tremendous value that is all too easy for those of us who are not addicted to miss: it provides a sense of self and social identity.

A SENSE OF SELF AND SOCIAL IDENTIT Y
In his classic book Theory of Addiction (2006), Robert West tells of an informal study he conducted. West asked a group of chronic smokers who had just quit that week whether they identified as smokers or as ex-smokers. Half said they identified as ex-smokers. This is extremely optimistic (to put it mildly).
Smoking cessation rates are notoriously poor: one-year relapse rates are approximately 75% (ibid.). In addition, the subjects themselves, as individuals, had smoked cigarettes for years if not decades: induction therefore strongly predicts that, in their own case, they will continue to do so. However, at a six-month follow-up, West found that 50% of the people who had identified as ex-smokers in that first week were still abstinent, compared to 0% of the people who had identified as smokers. Unless the individuals who had identified as ex-smokers somehow "knew" they would remain abstinent (against the odds and induction alike), unlike the individuals who identified as smokers, the intriguing hypothesis to emerge from this study is that the act of self-identification as an ex-addict increases the chance of successfully becoming an ex-addict. Subsequent larger-scale studies conducted by West and others complement this idea. Non-addict identities are associated with ongoing abstinence (Tombor, Shahab, Brown, Notley, & West, 2015) and, especially when positively evaluated by the individual in question, decreased substitute unhealthy behaviours and increased self-efficacy (Buckingham, Frings, & Albery, 2013;Frings & Albery, 2015). In conjunction with these studies, new theories have emerged that posit an ex-addict or recovery identity as a crucial mechanism for the transition from active use to stable abstinence (Buckingham & Best, eds., 2017), such as e.g. the Social Identity Model of Cessation Maintenance (SIMCM; Frings & Albery, 2015) and the Social Identity Model of Recovery (SIMOR; Best et al., 2016). Although these interwoven research projects are preliminary, the instinct behind West's informal study appears to be sound: an ex-addict or recovery identity promotes cessation and abstinence alike. Why?

Self-Categorization Theory and Social Identity Theory
Self-Categorization Theory (SCT; Turner, 1987) and Social Identity Theory (SIT; Tajfel, 1982) combine to provide a helpful theoretical framework (Tajfel & Turner, 1979) for understanding this research. Social kinds can be understood as ways of categorizing individuals according to a range of properties such as gender, race, religion, nationality, family role, profession, political affiliation, disease status, recreational interest, and many more. Although the reality of (some) social kinds is debated, and the exact criteria for membership for (some) social kinds is contested, what matters for SCT and SIT is that social kinds in general are associated with specific sets of beliefs, values, and behaviours, to which members are expected to conform in virtue of their membership. Call these social-kind-based expectations group-specific norms.
Group-specific norms may or may not be part of the criteria for membership of a social kind; and they are rarely if ever writ in stone: they can change over time. Nonetheless, they are highly informative. When we identify an individual as a member of a social kind, we expect them to conform to its groupspecific norms: to share the beliefs and values of the group and to behave like members of the group do. These expectations are not merely descriptive: they are not based simply on what the majority of group members are in fact like. They are based in part on normative standards: group-specific norms prescribe what group members are supposed to be like (Knobe, Prasada, & Newman, 2013;Leslie, 2015Leslie, , 2017. For this reason, if a person self-identifies as a member of a social kind they are likely to (explicitly and implicitly 10 ) self-regulate to ensure their beliefs, values and behaviour conform to the group-specific norms in question (Hacking, 1996;Turner, 1987). 11 Conformity functions to protect individuals from external judgements that they have failed to be a good member of the social kind and hence, in addition, from the possibility that they are no longer identified by third-parties as members of the group, or are rejected by members of the group. Crucially, fear of failed identification and rejection will be especially strong if group membership is an important source of value for the individual.
SCT and SIT therefore offer an elegant framework for understanding why an ex-addict or recovery identity promotes cessation and abstinence. The group-specific norms associated with these identities involve beliefs, values, and behaviour that coalesce around a commitment to not use drugs. In so far as these identities are an important source of value, individuals will be fearful of identity loss and group rejection, and thereby motivated to abide by norms of non-consumption.
But the success of SCT and SIT in explaining the power of an ex-addict or recovery identity invites the following question: might there be a comparable effect in self-identifying as an addict? In other words, perhaps addicts persist in using despite negative consequences because they self-identify as addicts and these identities are an important source of value, thereby motivating them to abide by norms of addicted consumption. For, after all, using drugs despite negative consequences is what addicts are supposed to do.

Sources of value in an addict identity
The success of this explanation depends crucially on the assumption that there is value in identifying as an addict. Why would that be? The answer is no different in theory from the value of any kind of social identification. It can provide social reward, community, and express a set of values; create structure and purpose in one's life; be a ground for self-esteem; and, perhaps most elusively but intriguingly, anchor a sense of self when otherwise there would be emptiness within. Relatedly, sustaining an addict identity thereby protects against loss of these various goods.

Social reward and a community where you belong
Social reward is a potent source of value for many species; recall the animal study described above, wherein 100% of "addicted" rodents opted for social over drug reward (Venniro et al., 2018). But, importantly, this was a forced-choice study: social reward was offered as an alternative to drugs. Notwithstanding the fact that addiction can destroy relationships with family members, friends, and colleagues, social reward is often an accompaniment to drug consumption for human addicts -not an alternative. This is true in myriad small and familiar ways. Think, for example, of how adolescents may be initiated into peer groups through willingness to use drugs; or how smokers may enjoy the in-group camaraderie of huddling outside together for a cigarette; or how alcoholics can find not only alcohol but company and solace in the pub. But it is even more pronounced for addicts who are members of some of the most vulnerable and marginalised drug user communities, such as long-term homeless poly-drug heroin users.
As poignantly documented by Philippe Bourgois and Jeffrey Schonberg in their book Righteous Dopefiend (2009) based on a decade of fieldwork, many such addicts have long lost most if not all relationships with individuals and communities who are not drug-using. Their user community may be all that they have left; and the relationships between members of such communities can be deep and meaningful. Living on the margins of society, addicts may love, protect, and care for each other, while they face their collective daily need for drugs in a context of poverty, homelessness, disease, disability, and police harassment and violence. Their identity as addicts is precisely what binds them together: quitting using would involve quitting the community and these relationships. Meanwhile the lack of treatment, housing, health care, and employment opportunities, combined with the intense stigma surrounding this form of addiction, means that in reality, abstinence does not guarantee a better life, let alone the possibility of replacement relationships of comparable commitment and meaning. Hence, for addicts such as these, their addiction is a way of belonging to a group and maintaining social bonds, when they would otherwise face the prospect of severe social isolation and loneliness.
Self-identifying as an addict can therefore have value because it brings with it a community of people who care.

Deviance and romanticization
So-called "deviant" sub-cultures hold allure for people who find themselves at odds with -and their values and lifestyles thus pigeon-holed by -conventional society (Becker, 1963). It is possible that, for some addicts living on the margins of society, their defiance in so living is itself of value to them (Bourgois & Schonberg, 2009). But it is not just on the margins that drugs hold allure. As Owen Flanagan has argued (2018), drug consumption (even heavy consumption) is an important part of many conventional rituals, rites of passage, and celebrations. For illustration, consider the writer Pete Hamill's description of growing up Irish in New York: "And so the pattern had begun, the template was cut. There was a celebration and you got drunk. There was a victory and you got drunk. It didn't matter if other people saw you; they were doing the same thing. So if you were a man, there was nothing to hide. Part of being a man was to drink" (Hamill, 1994, p. 57). 12 As for many men, drinking for Hamill is both a core part of a masculine identity and ubiquitous within his community.
Drugs are also reliably glamorized and sexualized in the media and across artistic genres, perhaps especially in autobiographical literature and photography (Goodman, 2018; see too Flanagan, 2013Flanagan, , 2016. 13 Consider the following quip about drinking from Augusten Burroughs: "A Ketel One martini please, very dry with olives, I want to say. "Um just a selzer with lime," I say instead. I might as well have ordered warm tap water with dirt. I feel that uncool" (Burroughs, 2003, p. 137). 14 Not only drug consumption, but also addiction itself, is not infrequently romanticized within our culture as part of a life that is worth wanting. Of course, the accuracy of this depiction ought to be in serious dispute for the majority of real-life cases; 15 the point is that there is an element of social cachet to depictions of addiction, which can contribute to the value an individual derives from an addict identity -whether they live on the margins of society or maintain a more conventional lifestyle.

Structure and purpose in daily life
In addition to the association with social reward and social cachet, addiction can create structure and a sense of purpose. For many addicts, their daily routine is defined by their addiction. Life revolves around securing and consuming drugs: securing drugs may either follow a set pattern or require creativity and cunning; consuming may be valuable not only because of the psychoactive effect of the drug but because of the pleasure in enacting rituals surrounding use (Zinberg, 1984). As with membership of other social kinds defined by e.g. family role, profession, or recreational interest, being an addict shapes a person's daily life.
This point emerges forcefully in a series of qualitative interviews that Anke Snoek conducted with addicts over a number of years: "Waking up in the morning […] I feel … my body feels … I got to get drugs. How am I'm going to get it, who am I going to rob, who am I going to rort, 16 where am I going to get money" (Snoek, 2017, p. 145).
"Your whole life revolves around amphetamines and speed. You have it when you get up and you have it … before you go to bed, and spend weeks or whatever not eating, you stay in your clothes all day every day.
[…] After a while it becomes just a part of life, just a way of life. You use it like you clean your teeth every day" (Snoek, 2017, p. 145).
"I will sit there and drink it and drink it and drink and drink and then blackout, wake up, drink and drink and drink and drink, blackout, wake up, drink, drink, drink, every day" (Snoek, 2017, p. 145) Drugs define what addicts think about, plan, and do with their day. Indeed, the loss of this structure and purpose can be part of what is so difficult about quitting: "It's hard, you know. You've just got to replace it and get in another routine sort of thing (…) it's a way of life since I was … you know, before I was a teenager so if you're doing something all your life and then suddenly you got to change everything about it, it's really hard" (Snoek, 2017, p. 147).

A sense of self and self-worth
The provision of structure and purpose in daily life is connected to another source of value in selfidentifying as an addict: a sense of self and self-worth. Consider the following first-person report by an addict in recovery named Kate, which I quote at length for its striking illustration of this point: "Just as a person can feel loss of identity when they lose a long-standing job, or their children have grown and left home, it is also very common, I believe, to feel loss of identity when recovering from a drug-addicted lifestyle.[…] I had established myself as a druggie. My friends and family knew me as such, and in a way I was proud of my varied life experiences and my street-smarts. […] I'd had an older boyfriend who had introduced me to the drug scene, and who I learnt a lot of drug-taking practices from. I took pride in the fact that I knew more about drug taking than most my own age, by virtue of hanging out with older and more experienced drug users. At age 18 I already knew how to cook and filter different drugs for IV use, and how to prepare poppies to extract the opium, I knew dosages and strengths for illicit use of prescription meds, I knew all sorts about scoring and smoking dope and lots of quirky little tricks for increasing your buzz. I was proud of that knowledge base. […] Seeing as I'd not done much else with myself over those formative years of early adulthood, I didn't have a heck of a lot else going on with my sense of identity.
[…] I began to leave my drug identity behind, but felt like I didn't have much else to equate myself with, there was a real void.
[…] I felt not so much like I missed the druggie lifestyle, but that I was starting to lose my grip on who I was, and was finding it hard to function. I was tempted to return to old habits for it was all I knew.
[…] I think a lot of the time when a person is trying to quit drugs and keeps relapsing, the personality/identity side of it is overlooked, I think that's just as difficult to gain control over as the physical drug usage.
[…] I would encourage anyone who's dissatisfied with a drug-chasing lifestyle to change it. But you've got to want it. And you've got to work at becoming someone and something else, otherwise you're still empty, and your inner self is still left craving and wanting." 17 There are two distinct but equally important themes in this testimonial. The first is that being an addict provides a sense of who she is; the second is that it is a source of self-worth.
To begin with self-worth: Kate is explicit that she took pride in her knowledge of drug pharmacology and her skill in drug delivery: she knows more and is better at this than her peers. Importantly, drug know-how is not the only possible source of pride for addicts in relation to their addiction. Mothers who are addicts can take pride in managing their consumption so that it interferes minimally with parenting (Kearney, Murphy, & Rosenbaum, 1994); presumably, analogous pride is in principle available to any addict who manages their consumption so that it interferes minimally with other important roles and activities. In contrast, some addicts may take pride precisely in pushing their body to its limits -there can be a kind of machismo in adopting a reckless devil may care attitude towards drug use (Flanagan, 2016(Flanagan, , 2018. To be clear, I do not claim that addicts like Kate who take pride in their addict identity embrace it wholeheartedly and are devoid of all feelings of regret and shame. Ambivalence is not only possible but in all likelihood the norm. The point is rather that it is possible to find a sense of worth in the fact that, if you are an addict, then at least you are a good one. This is a straightforward consequence of the fact that addict is a social kind and so subject to group-specific norms that set standards by which people can then measure themselves and others. Of course, what exactly the standards are for being a good addict (or a bad one) may be up for debate. Kate focuses on drug know-how; other addicts may focus on drug management or drug machismo. But this disagreement does not tell against the fact that there can be standards to which addicts may aspire qua addicts, and which can therefore be a source of self-worth (or self-criticism) in relation to their addict identity.
The second theme strikingly illustrated by Kate's testimonial is that being an addict provides her with a sense of who she is: "I had established myself as a druggie" and "didn't have a heck of a lot else going on with my sense of identity".
We often think of addiction as involving identity loss -as "spoiling" the self (Goffman, 1963) and destroying all that was meaningful in a person's life (Mackintosh & Knight, 2012). Correspondingly, we may think of recovery as reclaiming this past self and life. One problem with this image is that it is too black-and-white: it fails to recognize how some addicts maintain other meaningful elements of their identities and life despite their addiction (Kearney et al., 1994). But another is that it fails to recognize that, as in Kate's testimonial, addiction can be experienced as an identity gain (Dingle, Cruwys, & Frings, 2015). There are few if any meaningful social roles or group identifications for Kate apart from her addict identity: "I didn't have much else to equate myself with". Her addict identity is therefore of value because it provides a sense of who she is when otherwise she would have none.
Why would an addict have no alternative social identity? There are at least three possible explanations. The first is that their addiction started early in childhood or adolescence -too early for the person to have had the chance to develop alternative identities through taking up a normal range of social roles and activities. In this kind of case, there may simply be no (adult) past identity to be reclaimed. The second is that the social scaffolding required to support a past identity no longer exists. For example, suppose that, prior to your addiction, the crux of your identity was as a factory worker. But all the factories are now closed. Alternatively, perhaps the crux of your identity involved drug-unrelated social roles and relationships. But living as an addict has cost you your friends and your family -these relationships are beyond repair. In cases like these, there is no possibility of reclaiming your past identity in the here-and-now: you cannot enact it, even if you quit drugs. The third is that years if not decades of living as an addict have created a psychological barrier between you and a past identity. That identity used to be yours. And, let us assume, there are no external barriers to reclaiming it. But there is an internal barrier. The psychological gap between you then and you now is too great to be traversed. Although the explanation is different in all three cases, the upshot is the same: there is no alternative social identity ready and waiting to be reclaimed in recovery. A non-addict identity must instead be actively imagined and created if the person is to have any alternative sense of social identity at all. As Kate says, "you've got to work at becoming someone and something else".
I shall turn to how this is possible shortly, in conclusion. But I want first to offer a more speculative thought, pulling a final theme out of this testimonial. For Kate does not merely report that she lacks any kind of alternative social identity with which to replace her addict identity. She writes of there being a void in her sense of self -without her addict identity, her inner self is left empty.
What does this mean? Thus far in this sub-section, I have explored how self-identifying as a member of one or more social kinds can be part of what constitutes one's sense of self-worth and who one is. But, however important this social identity is to one's sense of self, for most of us, it is not all of it. There is also a sense of self that underlies one's social identity. This idea of an underlying sense of self is extremely difficult to theorise and operationalise. 18 But we can recognise it. It is, for example, what we want for our children. We want them to have a stable, core sense of who they are, that is not simply constituted by whatever social roles and group identifications they possess at any particular period of time, but is able to persist as a sort of anchor through changes in these roles and identifications. This stable, core sense of self may be partly constituted by personality traits. But, again, these traits are not all of it. Like our social roles and identifications, our personalities can change significantly over time, and yet we can have a stable, core sense of self that transcends these shifts.
What more is part of this sense of self? Recent work in social cognition has explored folk beliefs about the true self (for a review see Strohminger, Knobe, & Newman, 2017). Cross-culturally, the folk appear to believe that some characteristics of the self are more authentic than others: a part of who a person really is, deep down, in contrast to characteristics that may be part of the self but are less essential, more peripheral. One of the most striking findings to emerge from this research is that the characteristics attributed to the true self are in general positive and in particular morally good. One possibility therefore is that the sense of self I have been describing can be understood as awareness of the true self in one's own case. 19 Many of us are lucky enough to feel that deep down, we are good, capable, and reliable. We do not think we are perfect. We know we have many flaws. But we nonetheless have a sense of ourselves as having worth. And we fundamentally trust ourselves to deal with what the world throws at us and to survive it. Of course, fulfilling our social roles and being good members of the groups with which we identify, as well as possessing virtuous personality traits, can feed this sense of ourselves as good, capable, and reliable. But this sense of self is nonetheless distinct: an inner awareness that deep down you are OK -that your true self is indeed positive and morally good. This is precisely what addicts who have a negative self-concept and self-destructive mindset (as described above) appear to lack. From a development and psychotherapeutic perspective, this is no surprise. The conditions necessary for acquiring this core, stable sense of self likely include growing up in a loving, supportive, stable environment -while addiction is associated with severe adversity (including childhood mistreatment) and starkly limited socio-economic opportunity (often a cause of instability).
There is no doubt a great deal more to be said about what a stable, core sense of self consists in and how it is normally acquired. But if this sketch is correct even in outline, then an identity as an addict has value not only because it provides a sense of who one is when no alternative social identities are available. An addict identity can also function as a psychological defence mechanism -a barrier protecting against the experience of looking inwards and finding, not a stable, core sense of self as positive and morally good, but only a dark and empty void where that self should be.
To summarise the argument presented thus far: in general, addicts persist in using drugs despite negative consequences because doing so has value. And, in particular, self-identifying as an addict can be valuable in multiple ways that are all too easy to miss if one is on the outside of drug user communities. Members of these communities may have no realistic alternative source of social reward and relationships, structure or purpose in life, self-esteem, and sense of self and social identity. But once a person self-identifies as an addict, the explanation of why they would continue to consume despite the severe negative consequences caused or compounded by drugs may turn on a simple fact: they are an addict, and that is exactly what addicts are supposed to do.

CONCLUSION
Recall that the subjects in Rob West's study who identified as ex-smokers could not have based this identification on rational consideration of the evidence, for the evidence is not in their favor. Their selfidentification must rather be seen as expressive of a commitment to be an ex-smoker (Marušić, 2015). As we saw, this is a commitment which, once sincerely made, increases the likelihood of actually becoming an ex-smoker. How is it possible to sincerely make this commitment?
Being a smoker is a drug-specific identity that is somewhat unusual. In contrast to other drug-specific identities like e.g. being an alcoholic or a heroin addict, it is less likely to conflict with and thereby erode an individual's other significant social identities. The subjects in West's study are likely to have had a range of alternative identities to provide them with a familiar and continuous sense of self while they abandoned their identity as a smoker. Hence, the commitment to being an ex-smoker may not be based on rational consideration of the evidence, but it is nonetheless straightforwardly imaginable. What would life really and truly be like for them not to be a smoker? Quite a lot like it is now, except that they don't smoke. Who would they be, if they are an ex-smoker? Pretty much who they are now, minus their identity as a smoker. With this image of oneself as an ex-smoker clearly in mind, a commitment to be an ex-smoker can be understood as a decision to act in accord with this self-image. Given that ex-smokers don't smoke, this means one no longer smokes. 20 In contrast, for people whose identity as an addict provides their main if not only source of social reward and community, structure and purpose in life, self-esteem, and sense of self and social identity, the act of imagining their life and identity in absence of their addiction is far from straightforward. Which people and relationships replace the people they know and care for? What activities fill the day? And most of all, who would they be? Abandoning their identity as an addict represents an existential threat: there may be no alternative social identities or underlying stable, core sense of self to fall back on. If so, then, for addicts whose addiction is constitutive of their sense of self and identity, quitting is in effect a transformative experience of the sort elucidated by L.A. Paul (2014). If it is not possible to imagine yourself without imagining being an addict, then you lack not only a rational but also an imaginative basis for committing to a self-identification as a non-addict. If you make this commitment, it is as if you jump blind, with no certainty that you will survive the fall. This is terrifying, and renders utterly comprehensible why a person might choose to continue to be an addict rather than face this existential threat.
Yet people do sometimes quit drugs and recover from addiction -even people, like Kate, who may have little sense of self or identity apart from their addiction. How?
There is no doubt as much variation in the path to recovery as there is heterogeneity in the nature and explanation of addiction. But the only way to avoid jumping blind is to find a way to imagine a life and a self that is authentically yours, but without addiction. I want to conclude by suggesting three ways this is possible.
The first is obvious, but limited. Person-centered language can offer a small, first step. "Person with a substance use disorder" serves to name drug use as a problem without making it an essential part of the person by grammatical design. Having written this paper exploring the meaning of self-identification as an addict, it is time to kick away the ladder and embrace this alternative form of discourse so far as possible. 21 But although words matter, they are not all that matters. From a first-person perspective, thinking of oneself as a person with SUD as opposed to as an addict offers nothing more than a skeleton of an alternative identity -there is as yet no detail about who this person is.
Somewhat paradoxically, the second way to imagine a life and a self that is authentically yours but without addiction is drug-derived. Psilocybin-assisted therapy is a new treatment for addiction currently in development in clinical trials with alcohol and nicotine dependence (Bogenschutz, Forcehimes, & Pommy, 2015;Johnson, Garcia-Romeu, Cosimano, & Griffiths, 2014). 22 The treatment typically occurs over 12-15 weeks. It begins with weekly counselling sessions designed to build motivation, develop rapport between subject and therapist, and prepare the subject for the psilocybin sessions. The first of these is a moderate dose (20 mg/70 kg) delivered under the supervision of a therapist after approximately one month of counselling. It is followed by fairly intensive debriefing in the hours or days after the session, so that the subject can explore and consolidate the experience. A maximum of two further psilocybin sessions are offered during treatment, with the possibility of taking a higher dose if desired (30 mg/70 kg). Weekly follow-up counselling as well as debriefing post-session are continued for the duration of the therapy.
Although it is important to note that thus far the trials are open-label so efficacy is not known, cessation rates are extremely high and substantially exceed those for all other addiction treatments (ibid). Qualitative analysis of debriefing and follow-up sessions reveal that subjects themselves find the psilocybin sessions extraordinarily meaningful and believe they have a profound effect on their recovery for a range of reasons (Neilson, May, Forcehimes, & Bogenschutz, 2018;Noorani et al., 2018). 23 But one recurring theme is a change in sense of self and identity: "There was a part there [during the second session] where I felt gone… it wasn't like last time [during the first session] where I always remembered I was Adam.
"For a few seconds, it was just like 'I'm me, and there are no defining characteristics!' … that made me realise that I'm not a 'smoker'" (Noorani et al., 2018, p. 4).
"It felt like I'd died as a smoker and was resurrected as a non-smoker. Because it's my perception of myself, and that's how I felt. So I jumped up and I said 'I'm not a smoker anymore, it's all done'" (Noorani et al., 2018, p. 4).
"[I emerged out of a cocoon, shaking out a pair of magnificent wings, which was] me revealing myself, like actually showing myself to the world. This is who I am, this is who I really am" (emphasis in original; Noorani et al., 2018, p. 4). "[T]he self-defeating talk is gone, the extremely negative talk is gone, the depression [and] the anxiety is gone, the sleeplessness is gone … I learned to be me. And that it's ok to be me. I learned that I'm gonna make mistakes and everybody makes mistakes" (my emphasis; Noorani et al., 2018, p. 11).
For these subjects, psilocybin-assisted therapy seems to have created on the one hand an experience of shedding a previous identity as an addict, and on the other an experience of a stable, core sense of self that is positive and morally good. As a result, they do not have to imagine themselves apart from their addiction. That work is done for them: psilocybin-assisted therapy is actually transformative of their sense of self and identity. They emerge from it with a sense of who they are which is free of their past addiction. What they must now do is live in a way that is true to this self.
The third way to imagine a life and a self that is authentically yours but without addiction is to do the hard work of imagining it for yourself, despite the fact that, from your current position, you cannot make the imaginative leap. How do you do this?
As Agnes Callard has emphasized (2018), the answer must be that this is not an act but a process. You work towards it, feeling your way slowly but surely towards a new sense of self and identity. This is no easy task, for any of us. But there are two major obstacles that people with SUD who lack alternative identities face.
On the one hand, the imagination of an alternative life and self requires hope and creativity -and takes time and energy. You must actively tell a new story about yourself that breaks free from your past, in effect surprising yourself with what might be possible for you in future (Pickard, 2015d). Living with chronic addiction is not only destructive of hope and creativity. It also typically makes it significantly more difficult to complete any long-term project whatsoever -imaginative or otherwise. For, to adapt a vivid metaphor of Jeanette Kennett's, addiction is like a series of grenades exploding in one's path (Kennett, 2009).
On the other hand, the process of imaging an alternative life and self cannot be conducted in the imagination alone if it is to be experienced as genuinely authentic. What makes the difference between a fantasy and an imagined but realistically possible alternative self and social identity often rests, in part, on being able to try it out -not just in imagination but in reality. If you imagine yourself participating in non-drug using relationships for the first time in decades, the credibility of that imaginative act may require that, if you make overtures, however small, to people outside of your drug community, there is uptake. If you imagine yourself in a new job after decades of unemployment, the credibility of that imaginative act may require that there really is some prospect of getting one: that, e.g. if you walk into a job center, you are treated with respect. If you imagine making a home for yourself, it is hard to feel that person is authentically you if you see no chance of ever making it off the streets. Hence the process of imagining an alternative addiction-free self and social identity may be stymied not only by the psychological hardship of living with addiction. It may also be stymied by the large-scale failure of our society to provide the social support and material resources that people with SUD need to live meaningful alternative lives. In other words, they are stuck within an addict identity, not simply due to a failure of imagination, but due to a failure of social justice. Arguably, this is yet another reason why recovery rates for SUD are particularly poor for people who come from underprivileged backgrounds of severe adversity and limited socio-economic opportunity.
The aim of this paper has been to establish that we cannot understand and address addiction without taking seriously the value of drugs -in general, and, for some people, in particular for their sense of self and social identity. But the idea that transforming one's sense of self and social identitywhether through grammatical, pharmaceutical, or imaginative means -could be crucial to the process of recovery is utterly obscured by the view of addiction as a neurobiological disease of compulsion. We do people with SUD no service by our current singular focus on the disease model of addiction. We rather do service to ourselves. The moral model of addiction got us off the hook by blaming the problem on the character of people with SUD. The disease model of addiction gets us off the hook by blaming the problem on drugs. We should not let ourselves off the hook. In many cases, addiction is caused and maintained by the fact that, just down the street from us, there are people whose best prospects in life involve taking drugs and self-identifying as addicts. This will not change until we change how we think -about drugs, about addiction, and about what we owe to all people in our society, including those of us with SUD. 24 ENDNOTES 1 Owen Flanagan (2013Flanagan ( , 2016Flanagan ( , 2018 also explores the connection between addiction and identity; my own thinking has been significantly influenced by his discussion, which I draw on in Part II. 2 But note that morally condemning addicts for their choices to use drugs prior to becoming addicted remains an open possibility on the view of addiction as a neurobiological disease of compulsion. 3 Although proponents of the incentive salience theory of addiction (Holton & Berridge, 2013;Robinson et al. 2018;Robinson & Berridge, 1993) sometimes claim (or imply) that pleasure is lost in addiction, this is not a requirement of the theory: a dissociation between "wanting" and "liking" is perfectly consistent with the claim that some "liking" nonetheless persists. 4 Note too that the vast majority of studies investigate cocaine use only. There is some evidence of a similar pattern of persisting but diminishing euphoria with respect to heroin (Peele, 1975;Zinberg, 1984), but it is nonetheless an open question how much a finding from one class of drug can generalize to others. Of course, quite generally, as tolerance for a class of drug increases over time, higher doses are required for similar effects; meanwhile the lore of "chasing the first high" suggests that the intensity of early drug experiences may be relatively unique (Bornstein & Pickard, 2020). Anecdotally, my own experience of talking and working with people who are addicted to drugs is that in general they enjoy their drug of choice. Similarly, Owen Flanagan reports that, in his experience, when addicts say -as they often do -that drugs no longer do what they used to, this does not mean that drugs no longer produce euphoria (personal communication, January 2, 2020). Lindgren, Teachman, Neighbors, & Wiers, 2013). Intriguingly, this correlation appears to be especially robust when the explicit drinker identity is disavowed (Frings, Melichar, & Albery, 2016). 11 Elliott Berkman and colleagues have recently proposed the Identity-Value Model (IVM) to explain behavioral selfregulation; arguably, the model could also apply to cognitive self-regulation. IVM offers an account of how identityrelevant goal-directed behaviors (and, possibly by extension, beliefs and values) are selected over identity-irrelevant goal-directed behaviors based on calculations of greater subjective value (Berkman, Livingston, & Kahn, 2017). In effect, identity-relevance serves to increase expected subjective utility. IVM can be seen as complementing and rendering computationally tractable SCT and SIT. 12 Quoted in Flanagan (2018). 13 To cite just a few famous examples in literature: Dry: A Memoir by Augusten Burroughs (2003); Junkie: Confessions of an Unredeemed Drug Addict by William S. Burroughs (1953); Lit: A Memoir (P.S) by Mary Carr (2009); Confessions of an English Opium Eater by Thomas de Quincey (1821); Dope: A Novel by Sara Gran (2006); Infinite Jest by David Foster Wallace (1996); Trainspotting by Irvine Welsh (1993). For visual examples see: Cocaine True, Cocaine Blue by Eugene Richards (2005); Tulsa by Larry Clark (1971); and the very many examples of "heroin chic" fashion photography which can be found in any major magazine. 14 Quoted by Owen Flanagan (2018). For further discussion of this point, see Flanagan (2013) and (2016). 15 But see Flanagan (2016) for some possible counter-examples. 16 To "rort" means to scam. 17 From: http://www.stuff.co.nz/stuff-nation/assignments/how-have-drugs-affected-your-life/9513619/Drugs-were-theonly-life-I-knew. Kate's testimonial is quoted in part in McConnell (2016). 18 The social psychological literature is replete with attempts. R. H. Turner describes it thus: "Typically my selfconception is a vague but vitally felt idea of what I am like in my best moments, of what I am striving towards and have some encouragement to believe I can achieve, or of what I can do when the situation supplies incentives for unqualified effort" (1968, p. 98). Carl Rogers (1959) suggests the self-concept has three components: self-image; self-worth; and the ideal self. Other theorists see it as consisting in both self-evaluation and self-efficacy (Gecas, 1982). It may not be possible to combine these various ways of drawing the landscape into a single, coherent map. 19 Although it may be natural to interpret this idea as positing the true self as an inner object of introspective awareness, it is by no means necessary to endorse this metaphysical view in order to make sense of it. Here is an alternative interpretation: there is a web of beliefs about a person's self, some of which are more peripheral and easy to give up, and some of which are more central and hard to give up. Think of the true self as the most central of these beliefs about the person's self. Awareness of the true self in one's own case can then be understood as awareness of these central first-person beliefs. 20 For further discussion of the nature and possible rationality of the process of becoming the ideal self that one imagines, see Velleman (2002), Marušić (2015), and Callard (2018). Strikingly, one of the examples Velleman uses to illustrate this point is of a colleague who reports quitting smoking by making-believe he was not a smoker. I am neutral about the issue of rationality; the point that matters for the purposes of this paper is rather that the process is only possible at all if the ideal self is imaginable. 21 One restriction is when speaking to people with SUD who themselves strongly prefer the term "addict". In my experience, this is not uncommon. 22 Psilocybin is a naturally occurring classic serotonin 2A receptor agonist hallucinogen (otherwise known as a "psychedelic") found in many species of "magic" mushrooms. 23 These include e.g. experiences of interconnectedness, spiritual awe, and metaphysical curiosity; increased selfefficacy; decreased withdrawal symptoms and cravings; and the importance of their relationship with their therapist. 24 Versions of this paper have been delivered at the University of St Andrews, the University of Western Ontario, Princeton University, Columbia University, the National Autonomous University of Mexico, New York University, Duke University, the Society for Philosophy and Psychology, and the National Institute of Drug Abuse. I am grateful to all these audiences for their comments. I would also like to thank Serge Ahmed, Aaron Bornstein, Carl Hart, Owen Flanagan, Richard Holton, Niki Lacey, Harvey Lederman, Shaun Nichols, Peg O'Connor, Laurie Paul, Ian Phillips, Stephen Setman, and Anna Zilverstand for many helpful discussions.