In our response, we first address the relationship of drug instrumentalization with drug addiction and how this may translate on the relationship of behavioral instrumentalization and behavioral dominance to behavioral addiction. Subsequent responses then focus on extending and clarifying the idea of drug instrumentalization.

R1. From drug instrumentalization to behavioral addictions and back

Griffiths suggests the proposed model of non-addicted drug use may also hold utility for non-addictive use of other potentially addictive behaviors. We agree with this notion and would like to probe our model in relation to behavioral addictions. We argued that drug instrumentalization is a precursor of drug addiction. Although drug instrumentalization may provide a behavioral benefit, drug addiction does not, and it represents a pathological condition. In analogy to drug instrumentalization, we hypothesize that in certain situations behavioral dominance (a precursor of behavioral addiction) may be advantageous for the integrity of an individual and its chances for reproduction.

A behavioral addiction has as its core a naturally occurring behavior that becomes dominant (Shaffer et al. Reference Shaffer, LaPlante, LaBrie, Kidman, Donato and Stanton2004). For example, gambling addiction can be seen as the dominance of playing behavior; sex addiction, as the dominance of behaviors directed towards sexual intercourse. As such, the occurrence of the behavior itself may have an unquestionable evolutionary origin. We suggest that both a balanced as well as a focused behavioral strategy, can serve well to maximize the reward of an individual depending on actual environment and personality. Thereby, the establishment of one dominant behavior at the cost of all others may be an adaptive behavioral strategy in environments with restricted access to primary or secondary rewards. For example, it may be acceptable if a gold miner who seeks for gold as a secondary reward and who has no access to social communities, displays a strong bias toward the mining work rather than attempting to balance the reward. Physical and/or mental characteristics of a person may also restrict the access to rewards at any time during life. They are usually accepted as the source of a behavioral dominance when acknowledging the exploitation of “talent.” In this case, other than the obtained type of reward may become very hard to access, so that a balance in the reward is virtually not possible. Altogether, we suggest, human capability to establish a behavioral dominance is based on an ancestral adaptation to uncertain environments and/or to individual characteristics.

We argued in the target article that non-addicts consume psychoactive drugs because such consumption can change their mental state. A mental state change can then allow a more effective performance of certain behaviors and a better reaching of particular goals. Although we defined these goals on empirical and psychopharmacological grounds, they may translate easily to concepts of primary and secondary natural rewards. The suggested goals of drug instrumentalization essentially involve goal-directed behaviors as they also occur in behavioral dominance and behavioral addiction. The goals of drug instrumentalization are, thus, only a part of the goals that can potentially be served by goal-directed behavior in general and by a behavioral dominance in particular. This was noted by Griffiths in his comment and in previous work (Griffiths Reference Griffiths1996).

Social interaction may serve as an incentive by itself for social animals and humans (Matthews et al. Reference Matthews, Abdelbaky and Pfaff2005; Panksepp & Lahvis Reference Panksepp and Lahvis2007). In the absence of alternative rewards, there may develop a behavioral dominance for this behavior, for example, when people spend excessive time in interaction with friends, but neglect the partner, children, or work. This behavioral dominance may develop into a behavioral addiction, as it is seen, for example, in excessive Internet chat room use as a kind of Internet addiction (Shaw & Black Reference Shaw and Black2008; Weinstein & Lejoyeux Reference Weinstein and Lejoyeux2010), which may help to overcome a shortage in peer group access or problems with social behavior and self-esteem in the “real world” (Niemz et al. Reference Niemz, Griffiths and Banyard2005).

Sexual behavior and the resulting experience of an orgasm may serve as a strong primary reward. As secondary rewards, sexual cues may drive approach, as well as consummatory (copulatory or masturbating) behavior. Naturally given physical appearance and acquired social skills may have a strong influence on access to this particular reward. In contrast to sexual behavior involving a partner, one can assume equal access to autosexual behavior (masturbation). This behavior, however, appears driven by sexual cues. Accessibility to these cues may therefore have a major influence on the expression of the behavior and the “natural reward” obtained by it. There may possibly be two ways in which sexual behavior can be subject of a behavioral dominance. One may be related to sexual cue driven autosexual behavior. This can be observed as obsessive masturbation (Black et al. Reference Black, Kehrberg, Flumerfelt and Schlosser1997). The Internet provides an easy access to sexual cues now. In connection with autosexual behavior, this may result in excessive cue-seeking and hence time spent “consuming” images and virtual sexual contact on the Internet (Griffiths Reference Griffiths2001; Kalman Reference Kalman2008). Another type of behavioral dominance related to sexual behavior may manifest in excessive seeking of sexual partners and sexual behavior (Garcia & Thibaut Reference Garcia and Thibaut2010; Goodman Reference Goodman1992). However, this behavior requires considerably more effort than the first pattern and may be restricted to those people with a privileged access due to, for example, physical parameters, specific social skills, or high exposure to sex partners.

A high cognitive performance is required in many work-related activities in industrialized societies. Thereby, “work” may summarize all behaviors that yield a secondary reward reserve, usually in the shape of money. Money can in principle be exchanged for virtually all natural rewards; but in contrast to them, money can be stored more efficiently and accumulated as well. Although cognitive performance may by itself not constitute a behavioral dominance, “work” in general may well do. This can result in a behavioral addiction of wealth acquisition (Slater Reference Slater1980) and workaholism (Burke Reference Burke2004; Spence & Robbins Reference Spence and Robbins1992). The driving force may be a secondary reward accumulation, which uses cognitive performance only as one behavioral subcomponent.

The end of a negative stimulus or situation works as a “negative reinforcer” for the ante-ceding behavior. There are several behavioral activities that may serve this function in particular, and these activities are related to psychological stress. All have in common that they require little effort and are easy to learn; and some may involve highly automatic behavioral components that require little cognitive activity (Marks Reference Marks1990). Overall, they provide relieve from psychological stress and mood alterations by involving the individual in low-effort motor activities and/or distracting attention from mental preoccupation with real problems (Griffiths Reference Griffiths2005). Given a lack of behavioral alternatives that may lead to a controlled (and predictable) stress relief, a behavioral dominance for these behaviors may develop and can progress to a behavioral addiction. This can be seen in excessive gambling with various types of card, machine, computer, and Internet role-playing games (Griffiths Reference Griffiths1996; Rehbein et al. Reference Rehbein, Kleimann and Mossle2010). In particular, excessive “fruit machine” gambling involves easily accessible and highly automatized behaviors (Griffiths Reference Griffiths1993). Gambling addicts reported that a major driving force into the gambling was not only the chance of an effortless win, but also relief from psychological stress in situations with few behavioral alternatives. Other activities that were reported to provide relief from psychological stress and that are prone for behavioral dominance and addiction are physical exercise (Allegre et al. Reference Allegre, Souville, Therme and Griffiths2006) and eating behaviors (Grilo et al. Reference Grilo, Pagano, Stout, Markowitz, Ansell, Pinto, Zanarini, Yen and Skodol2011).

In drug instrumentalization, the pharmacological effects of a drug on the brain may account for the drug's beneficial effects toward ameliorating mental problems. However, there are also certain behaviors that were shown to provide at least temporary relief from persisting mental problems. These behaviors can become dominant and even develop into a behavioral addiction. There is evidence that physical activity and sport can have a significant antidepressant action (Martinsen Reference Martinsen1990; Paluska & Schwenk Reference Paluska and Schwenk2000). However, excessive sport may result in an exercise addiction (Allegre et al. Reference Allegre, Souville, Therme and Griffiths2006). In particular, drug-addiction may be effectively medicated or self-medicated by the establishment of a non-drug-related behavior (e.g., Roessler Reference Roessler2010; Ussher et al. Reference Ussher, Taylor and Faulkner2008). This may eventually result in the replacement of a drug addiction by a behavioral addiction (Hatcher Reference Hatcher1989).

Physical appearance and attractiveness for the opposite sex is believed to facilitate sexual behavior. There are behavioral patterns that may be seen primarily directed toward enhancing physical appearance, like exercise and body building in males (Hale et al. Reference Hale, Roth, DeLong and Briggs2010) and the suppression of eating behavior in females. Both can result in a behavioral dominance and addiction, such as, for example, exercise addiction (Allegre et al. Reference Allegre, Souville, Therme and Griffiths2006; Hausenblas & Downs Reference Hausenblas and Downs2002) or anorexia nervosa (Gucciardi et al. Reference Gucciardi, Celasun, Ahmad and Stewart2004).

R2. Extending the list of instrumentalization goals

We suggested an initial list with drug instrumentalization goals and drugs that may serve these goals. As we have learned from the comments, this list may be seen as a start rather than an end. Several commentators have suggested additional instrumentalization goals that we'd like to discuss here. Móró & Noreika, as well as Spanagel, suggested considering the use of psychoactive drugs in spiritual activity and religious rituals. Moro & Noreika argued that in particular sacramental and spiritual drug intake, which is famously associated with psychedelic drugs, may go far beyond a simple expansion of the perception horizon. We fully agree with this suggestion and add this as ninth instrumentalization goal to the list as: Facilitating spiritual and religious activities.

Móró & Noreika and Spanagel delivered a number of arguments and a mechanistic description that we want to adopt and expand here. Religion seems to be a unique human phenomenon. It provides a metaphysical base to explain the world beyond purely experience-driven cognitive constructs. Religion may provide the final reference for an individual's structure of sensory perception and thinking and in that way can give normative advice on behavior. Most important from a biological point of view may be that it provides a shared cognitive framework to explain the environment of individuals and groups. This may effectively “synchronize” perception style, thinking patterns, and behavior of social groups and foster altruism within the group. Altogether, religion provides health and well-being benefits, which make it seem an adaptive advantage (for a critical discussion see Atran & Norenzayan Reference Atran and Norenzayan2004).

How could psychoactive drugs benefit maintenance of a religion? In mono- as well as polytheistic religions, there is a god or gods with a higher understanding of the world and more-or-less universal power to influence environment and the fate of living creatures. All these religions have in common restricted contact with the supreme deity; that is, under normal circumstances and in a normal mental state, a normal human being may not have access to the “divine.” Learning from it/them or influencing the “divine will” and its ways of action is very limited. However, both are considered desirable. Many, if not all, religions have established a mediating mechanism that assures some kind of contact, that is felt to allow either enhanced insight into the divine or a perceived way to influence its action and solve problems of the individual or the group. Having a direct or indirect contact to the divine can be assumed to reinforce religious practice in the individual and strengthen the religion within the group. Móró & Noreika argue that “in culturally predefined contexts, the purpose of these [psychedelic drug induced] hallucinatory experiences is not merely to increase self-understanding and self-discovery, but also to address ultimate questions that shape or shake the fundamental worldview.” Changing the mental state can be achieved by meditation and particular thinking patterns (Van Gulick). A more powerful way, however, is by using psychoactive drugs. This practice was abandoned with the arrival of the monotheistic religions of, for example, Judaism, Christianity, and Islam in most parts of the world's population and exchanged for meditation. However, it is still in place in some natural religions, for example, in the Amazonian basin (Jay Reference Jay2010).

Why is this drug use different from the goal sensory curiosity – expanding the perception horizon? We suggest that while similar types of drugs are used, an exploratory use may be seen as undirected. Using it as part of spiritual and religious activities, in contrast, facilitates these activities in a directed way. People have clear intentions what they use the drug for and what mental state they wish to achieve.

Drugs that were and are used for this purpose include THC (Abel Reference Abel1980), cocaine (Streatfeild Reference Streatfeild2001), and nicotine (Jay Reference Jay2010). Hallucinogenic drugs used include mescaline and psilocybin that can be found in the peyote cactus and in mushrooms, respectively, in south and central Americas. In South America, dimethyltryptamine (DMT) is used as a hallucinogenic compound (Geyer & Vollenweider Reference Geyer and Vollenweider2008; Nichols Reference Nichols2004) in preparations from tropical plants (Jay Reference Jay2010). The latter were termed “entheogens,” acknowledging their application for religious purposes (Ruck et al. Reference Ruck, Bigwood, Staples, Ott and Wasson1979). Móró & Noreika suggested placing sacramental and spiritual drug use in the center of drug instrumentalization because of its deep embedding in human culture. However, in Western societies, most non-addictive drug use is predominantly nonreligious and has no sacramental or spiritual embedding any more. In another comment, Ahmed asked what the selective advantage of psychoactive drug use might have been in an ancestral environment. We speculate that an instrumentalization to facilitate spiritual and religious practice might actually have been among the first documented examples of drug instrumentalization in humans (Abel Reference Abel1980; Heath Reference Heath2000).

Another instrumentalization goal proposed for consideration by Spanagel and Kippin is drug use to adapt to peer group pressure and, as one may add, as a peer group defining behavior. We agree there is evidence that the use of drugs of particular types may form an entrance criterion for peer groups in adolescence and early adulthood and, thus, exert a social pressure to the individual. The consummatory behavior is then rewarded by becoming a member of this group. Here the opposition to adult norms, which suggest abstaining from illicit drugs and limiting intake of the legal ones, may play a larger role than the drug effect on mental state itself. Furthermore, there is no particular mental state that facilitates peer group bonding beyond that of perceived improved social interaction. Consuming psychoactive drugs may therefore be one among many other “risk behaviors” (Donovan & Richard Reference Donovan and Richard1985; Hill & Chow Reference Hill and Chow2002). As such, we do not see a clear relationship between the pharmacological effects of a particular drug and peer group formation/maintenance yet. Accordingly, we suggest to classify the choice and consumption of psychoactive drugs related to peer group formation not as an independent primary drug instrumentalization goal.

Spanagel also suggested considering the taste of a drug preparation and the seeking of sensory pleasure that it may provide as an instrumentalization goal. However, the psychoactive drug is usually not tasty itself, but rather the additional components of the preparation are (Peynauld Reference Peynauld1995). This would suggest two implications: (1) the preparation is consumed for the reinforcing effects of the pleasantness of the taste, and (2) the psychoactive effects of the drugs would play no essential role in the reinforcing effects and maintenance of the consummatory behavior. These points may be refused at the point of observation. For example, there is nonalcoholic beer and wine, which in its taste comes at least close to the alcohol-containing product. However, its consumption is not nearly that of the alcohol-containing preparation. Though it may trigger a cue-induced relapse in alcohol-dependent patients in withdrawal, it could so far not replace the alcoholic drink in non-addicts or addicts just by its taste. Similarly, decaffeinated coffee achieves similar taste as the caffeine-containing original preparation. Also, this approach did not replace the psychoactive drug–containing preparation. Accordingly, we rather argue that the taste (and other sensory parameters) of a psychoactive drug preparation may well play a role at the sociocultural level to foster and control drug consumption. However, its reinforcing effects may be too subtle and in many drugs not being present at all in order to serve as a unique goal for drug instrumentalization. Even when appreciating the taste and cultivating the social paraphernalia of legal drug consumption, we may still ask ourselves which of the named instrumentalization goals we are after when “liking the taste.”

R3. Further support for drug instrumentalization

Several authors have highlighted previously published evidence and concepts that further support drug instrumentalization theory. A very important set of evidence was highlighted by Ahmed, who discussed the evidence from animal research. Experimental animal research has shown that various species learn to self-administer psychoactive drugs when given access to them. These experiments usually use the paradigm of intravenous self-administration as a consequence of an operant behavior in a very minimal setting. In these settings (e.g., a Skinner box), animals can normally not perform goal-directed behaviors other than the one leading to drug delivery (or an activity control behavior with no consequences). As such, they cannot instrumentalize their drug seeking and consumption. Ahmed pointed out, however, that for animals in an experimental setting, drug instrumentalization also may be established when other behaviors are allowed that can be served by previous psychoactive drug consumption.

Further support for this view comes from alcohol-drinking studies in rats, which were meant to model a “bar like” situation in humans. Tomie and colleges showed in a series of experiments that single-housed male rats learn to drink alcohol before a “social opportunity” occurs in the shape of access to a conspecific animal (Tomie et al. Reference Tomie, Burger, Di and Pohorecky2004a; Reference Tomie, Uveges, Burger, Patterson-Buckendahl and Pohorecky2004b; Reference Tomie, Gittleman, Dranoff and Pohorecky2005). One can interpret these data as using the alcohol when expecting a social interaction opportunity to enhance behavior or perception of it. This effect was even more pronounced when males were allowed access to females rather than to males (Tomie et al. Reference Tomie, Hosszu, Rosenberg, Gittleman, Patterson-Buckendahl and Pohorecky2006). Although same-gender interaction may support the instrumentalization goal of improved social interaction, alcohol drinking before opposite gender access may support the goal of facilitated sexual behavior in animals. However, drinking studies in rats also showed that not all human instrumentalization goals may be served in rodents alike. The exposure to social stress rather reduced alcohol drinking in rats before and after stress exposure (van Erp & Miczek Reference van Erp and Miczek2001; van Erp et al. Reference van Erp, Tachi and Miczek2001), which suggests that the instrumentalization goal of facilitated recovery and coping with psychological stress may not be served by alcohol in rodents. These findings may also provide support for the suggestion that drug instrumentalization developed as an ancient adaptive trait. Hence, the behavior may have been established at least in some mammals, but being limited in its expression by the availability of and the access to psychoactive drugs. It also suggests that the ultimate cause for drug instrumentalization behavior may be different between instrumentalization goals.

These findings may have wide-ranging implications for basic research on neuronal mechanisms of drug addiction. Until recently (Ahmed Reference Ahmed2010), it was considered that self-administration is the most effective model to study drug addiction (Richardson & Roberts Reference Richardson and Roberts1996; Spealman & Goldberg Reference Spealman and Goldberg1978). However, it has been known for a considerable time that the characteristic escalation of consumption and loss of control by adverse consequences is observed only in a minority of drug-consuming animals (Deroche Gamonet et al. Reference Deroche Gamonet, Belin and Piazza2004; Vanderschuren & Everitt Reference Vanderschuren and Everitt2004). Based on this and other evidence (Cantin et al. Reference Cantin, Lenoir, Augier, Vanhille, Dubreucq, Serre, Vouillac and Ahmed2010; Lenoir et al. Reference Lenoir, Serre, Cantin and Ahmed2007), Ahmed suggested a biological resilience to addiction, which may protect most of the rodents and humans from becoming addicts and which may allow them to control their consumption of psychoactive drugs when using them in a systematic way. Overall, we agree with Ahmed's conclusion that “this research suggests that psychoactive drug instrumentalization is not unique to humans and probably corresponds to an ancient behavioral trait, at least in the mammal class.” At that point it can be only speculated that there may have been a parallel selection for psychoactive drug instrumentalization and the resilience to addiction. Given the abundance of animal studies in very restricted environments, we call for more research in understanding the mechanisms of operant and non-operant drug self-administration in relation to drug-facilitated behaviors in more realistic environments.

Considering the adaptive advantage of drug instrumentalization, Reid suggested better consideration of Niche Construction Theory (Kendal et al. Reference Kendal, Tehrani and Odling-Smee2011; Laland et al. Reference Laland, Odling-Smee and Feldman2000; Reference Laland, Odling-Smee and Myles2010), which was only mentioned as a potential factor when the drug instrumentalization theory was initially laid out. Reid argued that the “influence of the organism, humankind in this case, on the environment has led to the construction of a niche where modern drugs and drug-culture are part of the environment.” We fully agree with this view. We also agree with Reid's notion that “drugs can then be seen as part of the fitness landscape over which selection pressures operate.” Reid further argued that “natural selection will favour those who use drugs, then the trait of drug liking and its genetic underpinning will be promoted.” We have argued that a controlled use of now widely available psychoactive drugs may be beneficial for the individual in this niche. Selection pressure may, hence, operate toward drug instrumentalization – that is, toward the most effective use of available psychoactive tools. At that point we disagree with Reid in that we believe that not the trait of “drug liking” may be promoted but the trait of “drug instrumentalization,” which may include efficacy of the drug memory and its interplay with non-drug memories. The manifestation of this niche and the pressure for drug instrumentalization may be recognized in the search of individuals and pharmaceutical companies for cognitive enhancers (Husain & Mehta Reference Husain and Mehta2011). The principle of instrumentalization of these drugs may be the same as we proposed it for potentially addictive drugs. Accepting this, one might further expand this thinking and suggest to search for pharmaceutically engineered and non-addictive “little helper” (Hesse Reference Hesse2010; Sahakian & Morein-Zamir Reference Sahakian and Morein-Zamir2007) for all of the proposed instrumentalization goals that are currently served by addictive drugs.

At the conceptual level, Warburton suggested that the proposed drug instrumentalization model may have its special application in the nicotine utilization hypothesis (NUH), which was proposed earlier (for review, see Warburton). We agree with this view, which is supported by self-reports of smokers who claim that smoking helps them concentrate and think (Russell et al. Reference Russell, Peto and Patel1974), as well as from experimental evidence on cognitive performance (Mumenthaler et al. Reference Mumenthaler, Taylor, O'Hara and Yesavage1998).

R4. From interoceptive drug discrimination to drug instrumentalization

The ability to discriminate the effects of psychoactive drugs is an essential prerequisite for drug instrumentalization. Goudie, Gullo, Rose, Christiansen, Cole, Field, & Sumnall (Goudie et al.) noted that “drug discrimination research grew out of work on ingestion, before shifting to a broader conceptualisation of interoceptive stimuli detection.” They suggest that for a better understanding of drug instrumentalization, a conceptual shift away from ingestive behavior may be required. We agree. Although ingestion of plant preparations containing psychoactive drugs was historically the behavior that led to the bioavailability of the drug and to subsequent changes of the mental state, other ways of preparation and application became available for a number of psychoactive drugs. These include, for example, smoking/inhaling, snorting, and injecting. Although some of them circumvent ingestion and lead to direct availability of the drug to the bloodstream or nervous system, it may still represent a consummatory behavior.

We have focused in our assessment of the discriminative drug effects on the mental state, which reflects the mode of action of the brain function. Expanding this view, one should certainly acknowledge that the discriminative stimulus properties of virtually all psychoactive drugs have a peripheral component, which may interact with mental state changes. Most psychoactive drugs affect not only molecular targets in the CNS, but also often – with even higher sensitivity – ones in the body periphery. Not only the expected effects of the drug on mental state, but also its side effects on peripheral state, are part of the semantic memory about a psychoactive drugs. As such, the peripheral drug effects and their interoceptive detection may be an important part of judging the drug effects and self-titrating the dose for potential instrumentalization.

Goudie et al. further noted that the ability to detect different drug states may not be sufficient to facilitate other behaviors. We agree with this point and like to point out further mechanisms. A systematic drug instrumentalization requires a memory for mental states. This ability was demonstrated experimentally in humans and rats by testing the discriminative stimulus properties of a drug. This can at least in part be dissociated from its rewarding effects and may occur without any euphoria at all. Animals and humans were shown to be capable of recognizing the unique discriminative stimulus properties of psychoactive drugs and to use this information to guide goal-directed behavior (e.g., Hodge et al. Reference Hodge, Grant, Becker, Besheer, Crissman, Platt, Shannon and Shelton2006; Preston & Bigelow Reference Preston and Bigelow1991). Drug instrumentalization also needs a comparator function that determines whether the one or other mental state is preferable in a particular set and setting (Zinberg Reference Zinberg1984). This may involve not only comparing mental states, but also assigning them an emotional significance and value. Thereby, significance and value are dynamic – context-specific. For example, behavioral disinhibition and reduced anxiety is a valuable mental state in a social context, but not in a professional work environment. Based on this, we propose that mental states may have a value and that humans are able to assign it to the representation of this state and retrieve it later on.

Finally, it requires a mechanism of transition, something that brings the organism from one acknowledged mental state into another in a somewhat predictable way. A behavioral sequence that can result in such a transition in humans and animals is the self-administration of psychoactive drugs (Skog Reference Skog2000; Wiens et al. Reference Wiens, Zitzmann, Lachance, Yegles, Pragst, Wurst, von Holst, Guan and Spanagel2008). As Van Gulick pointed out in his comment, psychoactive drugs are not the only way to change the mental state. In fact, the mental state of human beings changes constantly in a “natural way” by intrinsic mental activity or as a consequence of the individual's interaction with the environment (e.g., by sensory perception). There are memories of the different mental states as part of the autobiographic memory. However, there are only limited ways to change these mental states on purpose along the three axes of (a) quality of change (i.e., into a particular direction), (b) quantity of change (i.e., the magnitude of change) and (c) time (i.e., with predictable times of onset and duration). In particular, under adverse conditions, such as a depressed mood, it may appear difficult to find any way to do it “naturally” (Snaith Reference Snaith1993). In that, the systematic seeking and consumption of psychoactive drugs might have appeared as a unique instrument to change mental states on purpose.

R5. Drug instrumentalization establishment during lifetime

An important aspect of how psychoactive drugs are instrumentalized is the age at which people establish drug consumption and possibly instrumentalization. Banaschewski, Blomeyer, Buchmann, Poustka, Rothenberger, & Laucht (Banaschewski et al.) have pointed out that an early initiation of psychoactive drug consumption, which starts now in early adolescence, appears to be associated with a reduced ability to use drugs purposely in a temporally stable, non-addictive manner later in life. It seems that the earlier in adolescence consumption starts, the higher is the lifetime consumption and risk of developing addiction (Behrendt et al. Reference Behrendt, Wittchen, Hofler, Lieb and Beesdo2009; Grant & Dawson Reference Grant and Dawson1997). A necessary conclusion from this would be to tailor prevention strategies that aim to delay age of consumption onset. We fully agree with this suggestion and like to expand this view from drug use to drug instrumentalization. We suggest asking people, in surveys, not only when they started to experiment with a drug, but also when and how they first establisheded a systematic and effective drug instrumentalization. Experimenting with psychoactive drugs is dangerous. Based on some semantic knowledge, novices try to explore different drugs at different doses with unpredictable individual responses. Because of this, self-titration almost necessarily involves occasional overdosing. Establishing an early drug instrumentalization, however, derives its hazards still from other sources. First, the brain is still in an important developmental period. Repeated exposure to the drug during this period may have significantly more irreversibly damaging effects on brain function leading to behavioral impairments than when drug exposure hits a fully matured adult brain. Second, in adolescence, salience of natural rewards is more highly reflected by enhanced activation of the brain's reward system (van Leijenhorst et al. Reference van Leijenhorst, Zanolie, Van Meel, Westenberg, Rombouts and Crone2010). At the same time, puberty imposes a lot of new demands and goals on the individual, which may result in considerable stress (Banaschewski et al.). Getting access and establishing a position in the peer group, approaching the opposite sex, increasing cognitive efforts in school, expanding the perception horizon and building/maintaining an “attractive” physique are just some of the tasks at that age. All of these can potentially be served by drug instrumentalization. In addition, there appears a perception/memory bias toward the positive effects of drugs.

Although predominantly positive expectancies of drug effects at younger ages predict future drug use, in adults, negative expectancies shape future consumption as well (Boys & Marsden Reference Boys and Marsden2003; Boys et al. Reference Boys, Marsden, Griffiths, Fountain, Stillwell and Strang1999; Leigh & Stacy Reference Leigh and Stacy2004). We have argued that drug instrumentalization is essentially a reinforcement-driven behavioral sequence. Given the enhanced reward sensitivity, this would make adolescents particularly vulnerable to the establishment of drug instrumentalization. We suggest that an early onset of drug consumption, may, thus, especially in socially and/or genetically “at risk” adolescents (Blomeyer et al. Reference Blomeyer, Treutlein, Esser, Schmidt, Schumann and Laucht2008; Laucht et al. Reference Laucht, Treutlein, Schmid, Blomeyer, Becker, Buchmann, Schmidt, Esser, Jennen-Steinmetz, Rietschel, Zimmermann and Banaschewski2009) lead to early onset of drug instrumentalization to help solving adolescent developmental tasks. This may be further facilitated by additional adverse life events during childhood (Andersen & Teicher Reference Andersen and Teicher2009). Once drug instrumentalization is established to enhance performance of one particular behavior (e.g., disinhibition and peer group bonding by alcohol), a behavioral strategy may be derived. This strategy may be drug instrumentalization – that is, the knowledge that psychoactive drugs may be a ready means to address any problem in life. Therefore, we suggest introducing a systematic monitoring of drug instrumentalization as early as monitoring experimental consumption. Both need to be, if not avoided, then delayed in their onset. There may be a chance to address both separately in prevention and support programs.

Kippin noted that achieving efficiency and accuracy of mental state changes during the establishment of drug instrumentalization may be a crucial problem during experimental consumption, which might strongly affect later drug taking. We agree with Kippin's suggestion that during this learning process, people experiment in a trial-and-error way with toxic drugs. To find an optimal dose range for a desired mental state change can be achieved in a “bottom up” and “top down” self-dosing approach. This does indeed result in a considerable number of overdosing episodes. It may become obvious, for example, in weekend binge drinking in young adults in Western societies. Overdosing usually results in a highly aversive episode, which is usually remembered well; and the type of drug or the high dose is avoided in the future. Experimentation with new psychoactive drugs, as well as dose self-titration for the desired effects, matures into knowing the preferred drugs and optimal doses in most people when reaching adulthood and facing a new life episode (Leonard & Das Eiden Reference Leonard and Das Eiden1999; Miller-Tutzauer et al. Reference Miller-Tutzauer, Leonard and Windle1991; Nilsen et al. Reference Nilsen, Holmqvist, Hultgren, Bendtsen and Cedergren2008).

R6. Drug instrumentalization in addictive and non-addictive drug use

Pickard asked whether drug instrumentalization is limited to a non-addictive drug use or whether there may be also one in diagnosed drug addicts. Pickard suggested that “it may not be adaptive…but it may yet be an instrumental means to desired ends.” This would imply that addicts are not only passive victims of a disease, but may also within an overall negative health benefit still maintain “islands” of subjectively perceived benefits. There have been several attempts to recognize potential benefits of addictive drug consumption, mostly derived from subjective reports. Alexander (Reference Alexander1987; Reference Alexander1990) has argued that an addictive drug use may be beneficial to cope with “integration failure.” Davies (Reference Davies1997) argued that also in classified addicts, people chose drug taking not because there is a compulsive drive to do so, but because there are personal benefits. Considering single behaviors, we agree that in drug addiction, there may also be some subjective benefits derived from the drug use. Who can deny that, for example, addicted alcoholics may forget their severe problems, which they did not find other ways to resolve, during binge drinking episodes? However, we strongly argue the importance of considering the overall situation and perspectives of the individual.

Almost all psychoactive drugs can cause severe damage to the brain and body periphery. Chronic alcohol consumption can, for example, cause cognitive deficits and neuronal loss (Harper Reference Harper2007; Parsons Reference Parsons1998; Ward et al. Reference Ward, Lallemand and de2009); chronic ecstasy consumption can destroy serotonergic terminals in the brain and induce cognitive deficits (Gouzoulis-Mayfrank & Daumann Reference Gouzoulis-Mayfrank and Daumann2009; Seiden & Sabol Reference Seiden and Sabol1996); nicotine consumption can cause various forms of cancer (Ray et al. Reference Ray, Schnoll and Lerman2009), and high doses of androgenic–anabolic steroids may cause myocardial infarction (Wood Reference Wood2004). Chronic consumption of psychostimulants can lead to cardiovascular problems, cerebral atrophy, and a decline in cognitive performance (Pascual-Leone et al. Reference Pascual-Leone, Dhuna and Anderson1991; Volkow et al. Reference Volkow, Hitzemann, Wang, Fowler, Wolf, Dewey and Handlesman1992). Many psychoactive drugs enhance pre-existing psychopathologies in vulnerable individuals (Robbins & Everitt Reference Robbins and Everitt1999; Wood Reference Wood2004). Chronic consumption of cannabis/marijuana may not only cause cognitive deficits (Solowij Reference Solowij1998), but also induce schizophrenia (e.g., Andreasson et al. Reference Andreasson, Allebeck, Engstrom and Rydberg1987; Reference Andreasson, Allebeck and Rydberg1989; Negrete Reference Negrete, Nahas and Latour1993). Accordingly, the effects of the high drug exposure on personal health, social environment, private economy, and the public are usually so severe that an overall beneficial effect or an adaptation must be denied.

We have argued that drug addiction is preceded by drug instrumentalization, or at least the attempt of it. A crucial mechanism for the transition from controlled drug instrumentalization to drug addiction is the “over-instrumentalization” of psychoactive drugs. We suggest that over-instrumentalization should describe the attempt of a person who “successfully” instrumentalized one or more psychoactive drugs to meet increasing demands for goal achievement by increasing the frequency and/or dose of the consumed drug. Although drug instrumentalization can be seen as a dynamic process with respect to changing goals and drugs used for it, we assume that there is generally a sigmoid dose/frequency-efficacy function in drug instrumentalization. There may actually be quite a small dose window for the optimal mental state for each drug serving a particular instrumentalization goal. If this dose window is left by increasing the dose of the drug, no instrumentalization may be possible any more. For example, overdosing oneself with alcohol at a social occasion may have sedating effects. The mental state change with a too high dose may then be less supportive for the goal of social interaction than no drug-induced mental state change. In that, the attempted over-instrumentalization is actually a mis-instrumentalization, as Kippin suggested. Kippin discussed a number of consequences of mis-instrumentalization with which we fully agree. One of the most adverse effects is the increased risk of losing control over drug consumption beyond any utility for instrumentalization and to develop a drug addiction. As such, drug instrumentalization requires a fine-tuned learning process to get established and works only when the window of psychoactive drug doses is carefully self-titrated.

R7. Sociocultural determinants of drug instrumentalization

Goudie et al. noted that, in its initial shape, drug instrumentalization theory did not consider cultural factors. They argued that drug use is not a result of modern societal demands alone, but to such demands interacting with specific cultural contexts that determine attitudes and expectancies. We fully agree with this view and would like to expand the theory by a cultural dimension. Drug instrumentalization is a dynamic process over time, across and within cultures, groups, and individuals. We would argue that cultural factors essentially determine which drugs are available for legal or illicit consumption, approved dosage ranges and corresponding behavioral effects, and finally for which goals the available drugs are allowed to be used.

Depending on the respective culture, only a subset of the known psychoactive drugs is permitted for legal use (Heath Reference Heath2000). Some cultures have explicit rules that determine which drugs may be used and how (Harding & Zinberg Reference Harding, Zinberg and Du Toit1977; Maloff et al. Reference Maloff, Becker, Fonaroff, Rodin, Zinberg and Harding1981; Moreira et al. Reference Moreira, Smith and Foxcroft2009). Western societies do, for example, allow for alcohol, nicotine, and caffeine, but prohibit cocaine and opiates. Islamic societies disallow alcohol but are less restrictive with psychostimulants like khat (Rehm et al. Reference Rehm, Rehn, Room, Monteiro, Gmel, Jernigan and Frick2003). There are specific occasions and places where drug use is explicitly allowed and encouraged, often to a programmed excess (Heath Reference Heath2000). These social rules can be fixed in religious and secular law terms with sanctions and punishment threatened. The rules on how to use the drugs are established in social rules of use, which are passed by social learning mechanisms from parents, peers, and media (Harding & Zinberg Reference Harding, Zinberg and Du Toit1977; Zinberg et al. Reference Zinberg, Harding, Stelmack and Marblestone1978).

There is also evidence for measures to prevent toxic doses. This can be found by the thresholds that, for example, European countries have fixed in laws on alcohol use. Although small amounts of alcohol in the blood are tolerated for taking part in public activities and work, there is a certain threshold (associated with an amount of drug to be consumed) at which legal punishment is threatened. Besides that, there are culture-specific, often mutually agreed and unwritten rules on consumption of allowed drugs, which also prevent lethal overdosing. These rules are defined by approval and disapproval of the behavioral consequences of the drug, rather than on absolute amounts consumed (Harding & Zinberg Reference Harding, Zinberg and Du Toit1977; Maloff et al. Reference Maloff, Becker, Fonaroff, Rodin, Zinberg and Harding1981). At some occasions (e.g., at a restaurant visit), alcohol drinking is accepted and encouraged to a level of increased social interaction and moderate behavioral disinhibition (correlative to a low dose of the drug); it is disapproved when inducing aggression and overtly intrusive behavior (reflecting high levels of drug intake). At other occasions (e.g., at the Munich Oktoberfest), excessive alcohol consumption may be encouraged, but is also disapproved of when the consumer drinks up to a level that endangers consciousness and vital functions. Disobedience of these rules may result in social punishment. These rules also exist for illicit drugs but are clearly less easy to encounter in their application. What they have in common is that they attempt to control the dose-range of potentially toxic substances and to prevent drug abuse and addiction (Harding & Zinberg Reference Harding, Zinberg and Du Toit1977; Zinberg et al. Reference Zinberg, Harding, Stelmack and Marblestone1978).

Wu has noted that because social considerations play an important role in drug addiction, so they may do in non-addictive drug instrumentalization. Given the “peer pressure, volatile family atmospheres, and/or doomed images of their future” people from poor socioeconomic background lose control over drug instrumentalization more easily than people from higher social classes (Wu). On the opposite end, drugs may be instrumentalized to enhance well-being and happiness of individuals and within social groups (Miller). By conceptualizing drugs as “exogenous neurotransmitters” and proposing the use of drugs to actively shape individual perception, Miller described the apotheosis of an individualistic and “neoliberal” approach criticized by Wu, who emphasized drug instrumentalization as a reactive coping strategy to compensate for socioeconomic disadvantages. This made for interesting tension in the debate. Miller, however, also stressed the use of drugs in a social context and provided some interesting and noteworthy examples of drug instrumentalization in scientific thinking. Taken together, it is undisputed that drug instrumentalization has a strong social component with many dimensions, such as the goals and behaviors drugs are used for or the type of drugs available in microenvironments.

There may be also significant differences in alternative ways to change the mental state (see Van Gulick). One may argue that people with poorer socioeconomic backgrounds may attempt to instrumentalize drugs partially to overcome consequences of this situation. We have argued that there are clearly limits to the extent psychoactive drugs can be instrumentalized. They can certainly not equalize “natural” (e.g., genetic or early developmental) or social (e.g., born in a poor environment) disadvantages. As such, there may appear more instrumentalization goals in socially unprivileged people than in privileged ones. At the same time, there are fewer alternative “tools” (see Van Gulick) and less rational control established by families and the educational system. This may explain why poorer people use more psychoactive drugs and generate a higher proportion of addicts, who may have driven drug instrumentalization to its utmost ends and beyond.

R8. The claims of the drug instrumentalization theory

Several authors suggested that the major claims of the drug instrumentalization theory may be dissected, which may help to define the limits of the theory. Goudie et al. and Pickard distinguished two essential and independent claims. The first one is that drug instrumentalization theory suggests “drug use is instrumental” (Pickard) and “that drug-induced mental states facilitate behavior” (Goudie et al.). The second and wider-ranging claim refers to the evolutionary origin and role of drug instrumentalization as an adaptation, suggesting that the behavior has a fitness-related evolutionary origin and enhances reproductive fitness. In addition, Pickard identified a third claim that says “the desired ends [of drug instrumentalization] are easy to comprehend.” Although most commentators appeared to agree with the first claim, they were skeptical about the second one (Goudie et al., Pickard, Troisi, Ahmed, Ainslie, Sullivan & Hagan). Several authors have suggested that the second claim may not even be necessary to hold the first one.

The most important critique was that there is as yet no clear evidence for an increase in reproductive fitness by non-addictive drug instrumentalization. We think that distinguishing different claims is useful in order to separately address the supporting evidence. We believe that there is good evidence in support of the first claim as outlined in the main text. When scanning the available evidence on psychoactive drug consumption, we noted that non-addicted drug use is a rather persistent phenomenon across several species. In humans, it is evident over long periods of time for different psychoactive drugs in virtually all habitats. In an attempt to explain this persistence, we considered whether it may have provided an evolutionary advantage. We think this claim is supported by the available evidence – though we have to admit that an explicit testing of this hypothesis and more empirical evidence are required.

The idea that psychoactive drug consumption may represent an adaptation in certain circumstances is not new, as some commentators allude to. Lende discussed evidence from interviews of Colombian adolescent drug users from which he constructed a summary variable of evolutionary benefits out of sexual and competitive benefits. We hope that this neuroanthropological approach may also guide future field research on non-addicted psychoactive drug use in humans. Ainslie suggested that there is an evolved trait that controls psychoactive drug consumption, but that it is one controlling the sensitivity to “reward, reinforcement or utility”. We disagree that sensitivity to reward alone would be able to explain the rather sophisticated and highly organized non-addicted drug use we currently observe. However, sensitivity to the “utility” of drug use may be a relevant trait for selection, which may well favor non-addictive drug instrumentalization. Accordingly, those individuals with a high sensitivity for the use of psychoactive drug effects for certain purposes and at the same time resilience for compulsive use and addiction (Ahmed) may face an advantage. Sullivan & Hagen suggested earlier that the manipulation of one's own CNS with psychoactive plant compounds may actually be an adaptation, but one restricted to circumstances of mental illness or nutritional deficiencies (Sullivan & Hagen Reference Sullivan and Hagen2002).

True adaptations that have enhanced reproductive fitness may only become obvious over long periods of time. Most animals that do self-administer and instrumentalize psychoactive drugs under laboratory conditions, have only irregular access to psychoactive drugs in their natural habitats. Judging adaptive effects on fitness may therefore be difficult. On the other hand, recorded human history may be a too short time to allow for a longitudinal analysis. In addition, early human records may not provide sufficient information on drug use (types of drugs, content of psychoactive compounds, and frequency and occasion of use) and possibly supported behaviors. It may also be hard to judge whether a known drug instrumentalization has actually increased the reproductive fitness to a measurable extent.

Acknowledging these difficulties and also possible alternative explanations (Ahmed), we understand our discussion for an ultimate cause of the behavior rather as a suggestion and directive for future research than as a proven claim. However, we and others (Lende; Lende et al. Reference Lende, Leonard, Sterk and Elifson2007) have provided a number of arguments that do – at least from our point of view – suggest that drug instrumentalization is adaptive and has an ancient origin instead of being a most recent behavior. We took this suggestion further and expanded the list of potential circumstances, assuming indeed what Pickard called the third claim that we may at least grossly comprehend what is finally good for us. Regarding drug use as a consumer behavior with distinct reinforcement mechanisms as suggested by Foxall & Sigurdsson may further enhance insight. Based on these suggestions, we would like to call for a more thorough investigation of the potential beneficial and adaptive effects of drug instrumentalization at the levels of single individuals, groups, and the population, which should go beyond self-reports or counting babies.

R9. Implications of the drug instrumentalization theory

Here, we wish to clarify that drug instrumentalization theory was primarily suggested in order to better explain the present situation, incorporating a wide range of evidence from different disciplines. In doing so, we did not intend to judge whether it is good or bad that human beings consume psychoactive drugs. If one accepts the second claim of the theory (see sect. R8) that the ultimate cause of the behavior is related to an increase in reproductive fitness, it may be hard to reject the behavior as purely maladaptive. However, because this claim still needs to be sufficiently supported with evidence (or disproved), it remains open for debate whether non-addictive drug instrumentalization may be rejected by its implications, as Swendsen & Le Moal do, or actively encouraged, as Miller suggests.

On that point, we should note that at least Western society has already made its decision on non-addictive drug instrumentalization, but under another label. Though ethical debate on cognitive enhancers is ongoing (e.g., Hesse Reference Hesse2010; Sahakian & Morein-Zamir Reference Sahakian and Morein-Zamir2007), systematic search and use of psychoactive drugs for healthy individuals is already well underway. Also, pharmacological help to achieve other instrumentalization goals, such as improved physical appearance and attractiveness has long been pursued, for example, by the development and use of anorectic drugs (Bray Reference Bray2000; Ryan Reference Ryan2000). At this point, we welcome a debate on the implications but agree with Sullivan & Hagen's precautionary view that it is still too early for political suggestions regarding known drugs with an abuse potential based on drug instrumentalization theory.

We argued that drugs are used to systematically alter one's own mental state in order to better perform non-drug-related behaviors. Van Gulick pointed out in his commentary that there are other ways to change the mental state that may not rely on psychoactive drugs, such as by “ideas and patterns of thought.” We fully agree that certain types of mental activity may change general mental state. For example, a focused consideration of the positive elements in life may yield a more pleasant mood/mental state than a thinking pattern focusing on negative elements. This change in mental state may in turn determine whether and how other behaviors are performed. It is tempting to conclude that health policy should try to replace psychoactive drug use by enabling individuals to use their mental capabilities. The problems with this simplistic advice may be that efficacy of the mental activity-induced state changes can depend considerably, for example, on intelligence and training. Psychoactive drugs, in contrast, “work” in virtually all individuals no matter how privileged one person is. Regarding the relatively wide spectrum of mental state changes (e.g., behavioral disinhibition vs. cognitive enhancement), one might also question whether cognitive activity may work in all those directions. If psychoactive drugs cannot be removed from modern life, one may at least aim toward reduction and controlled consumption. In that we think Van Gulick's suggestion of a combination of pharmacological and mental tools for instrumentalization may provide a valuable approach in health and disease.