Müller & Schumann's (M&S's) theory can be formulated in a weak form – that drug-induced mental states facilitate behaviour; and a strong form – that such states facilitate reproductive fitness. We believe that the weak form of the theory is easier to defend.

State memories are clearly critical for the theory. M&S suggest that such memories are based on ingestion-related phylogenetically old learning mechanisms. However, drug state/discrimination research (NIDA Research Monograph Reference Glennon, Jarbe and Frankenheim1991) shows that animals can detect at least 40 different drug states, dose-specific states, centrally and peripherally mediated states, and states in which animals discriminate between closely related drugs (e.g., amphetamine and cocaine); and also that when animals learn about drug mixtures, the components are counter intuitively processed in parallel rather than as a novel “gestalt.”

Such findings suggest that the brain detects very specifically many different types of interoceptive states. This makes sense in evolutionary terms. It allows organisms to detect different states that need regulation to maintain the constancy of the milieu interne and to detect different states arising from drugs that are all essentially toxins. Drug discrimination research grew out of work on ingestion, before shifting to a broader conceptualisation of interoceptive stimuli detection. We suggest that M&S's theory requires a similar conceptual shift away from ingestive behaviour.

The fact that the brain can detect many drug-induced states supports M&S's theory that many drugs may be used instrumentally. An important point, such an account of drug use parsimoniously explains reports of the use of drugs such as antipsychotics and antidepressants, which are not typically conceptualised as used drugs (e.g., Tarasoff & Osti Reference Tarasoff and Osti2007). As M&S note, because the aetiology of psychiatric disorders is little understood, it is possible that such drugs are used to induce a “less aversive mental state”; this is also suggested by radical preclinical (Colpaert & Koek Reference Colpaert and Koek1996) and clinical (Moncrieff & Cohen Reference Moncrieff and Cohen2009) theorists and supported by recent research (Mizrahi et al. Reference Mizrahi, Bagby, Zipursky and Kapur2005; Moncrieff et al. Reference Moncrieff, Cohen and Mason2009).

The ability to detect many drug states is a necessary but not sufficient condition to assert that drugs are taken to facilitate fitness and/or social functioning. We are sceptical that drug use reliably enhances social functioning or reproductive success. It is a widely held belief that alcohol and drugs convey these benefits, but the evidence is unclear. Cooper (Reference Cooper2006), reviewing research on the alcohol-sex link, suggested that several third-variable explanations could account for the association, notably expectancies and sensation seeking. Sensation seeking prospectively predicts greater alcohol/drug use, better social functioning and reproductive success, and stronger positive drug expectancies (Gullo & Dawe Reference Gullo and Dawe2008). Sensation seeking is also highly heritable and, given its mesolimbic dopamine system basis, might provide a more parsimonious and evolutionarily plausible account of some non-addictive drug use that is consistent with Nesse and Berridge's (Reference Nesse and Berridge1997) alternative model.

In many of the instrumentalization examples discussed (e.g., facilitated social interaction and sexual behaviour), drug use is an ineffective “quick-fix” compared with long-term strategies. For example, reading the Kama Sutra might facilitate sexual behaviour related to reproductive success better than getting intoxicated. Similarly, exercise might be of more value, as well as improving cognitive performance, reducing fatigue, and improving physical appearance, with no risk of addiction. However, reading the Kama Sutra while jogging requires more time and effort than drinking. This probably reflects delay discounting, a universal highly heritable human/animal trait (Anokhin et al. Reference Anokhin, Golosheykin, Grant and Heath2011) believed to underlie sensation seeking/impulsivity.

Expectations concerning alcohol and drug effects have been shown to have some predictive value in determining drinking behaviour (Morawska & Oei Reference Morawska and Oei2005). However, the relationship between expectancy type (positive or negative) and drinking is far from clear-cut (Jones et al. Reference Jones, Corbin and Fromme2001). Expectations of the benefits of drug use exist before a drug is first taken. However, even adolescents are aware of the negative impact of excessive use, as used in educational programs for the young about negative consequences of alcohol/drug use and ways of tackling risky situations – such as protection against peer pressure. M&S acknowledge that positive and negative expectancies prior to initial alcohol and drug use would be held as semantic memories. However, for a theory of non-addictive drug use, M&S need to consider further how competing expectancies determine which specific drug-related behaviours occur.

M&S suggest that once an individual learns that drug use can be adaptive, drugs may then be used in greater quantities which may lead to addiction. Alcohol/drug use may start as an adaptive process to facilitate interactions, but it can lead to levels of consumption that impair interactions (Rehm & Gmel Reference Rehm and Gmel1999). This is a critical theoretical issue, but the instrumentalization theory does not consider all the likely negative consequences of heavy drug use, which should act as a warning signal to individuals before heavy usage becomes habit-like. A major issue for the theory is that it concentrates on adaptive drug use. There is little discussion of what happens in the critical transition from social to hazardous drug use, whether it be non-dependent or dependent hazardous use.

The theory does not consider cultural factors. Use is not a result of modern societal demands alone, but to such demands interacting with specific cultural contexts that determine attitudes and expectancies. The United Kingdom, Spain, France, and Germany are presumably similar societies in terms of demands on workers, although cocaine use is less prevalent in France and Germany than it is in the United Kingdom and Spain (United Nations Office on Drugs and Crime 2010), despite the fact that in evolutionary and socioeconomic terms and in cocaine supply, there is little difference between these European countries. Similarly, methamphetamine use in the United States is predominantly a “white problem,” even after controlling for job status and education (Anglin et al. Reference Anglin, Burke, Perrochet, Stamper and Dawud-Noursi2002; Cartier et al. Reference Cartier, Farabee and Prendergast2006), again implicating cultural factors in use.

Our comments clearly do not imply that instrumental drug use does not occur. The challenge for the theory is to define its limits, particularly the strong fitness-related form of the theory, and to clarify the relevance of enhanced fitness or facilitated behaviour versus other factors.