Addiction, Desire, and The Polluted Environment – Richard Holton’s 2nd Uehiro Lecture

By Jonathan Pugh


In the second of his three Uehiro lectures on the theme of ‘illness and the social self’, Richard Holton turned to the moral questions raised by addiction. In the first half of the lecture, he outlined an account of addictive behaviour according to which addictive substances disrupt the link between wanting and liking. In the second half of the lecture, he discusses the implications of this account for the moral significance of preferences, and for how we might structure environments to avoid triggering addictive desires.


You can find a recording of the lecture here



Two Competing Views On Addiction, and The Significance of Wanting/Liking


Holton begins his analysis by outlining two competing views on addiction. On the first view, there is nothing particularly special about addictive behaviour; such behaviour is freely chosen, and addicts should be understood as fully responsible with respect to it. This is the sort of view that Bennett Foddy and Julian Savulescu have previously defended; following their terminology, I shall refer to this as the liberal view of addiction (N.B. Holton does not use this terminology in the lecture).

In contrast, the second view (a version of which is defended by Steve Hyman) claims that addictive behaviour is the result of a pathology; on this approach, addictive substances ‘hijack the brain’, and the freedom of choice in addicted individuals is thus severely compromised, perhaps to the extent that they should not be held responsible (although Hymen himself suggests that the loss of control in addiction is neither complete nor simple). Call this view the disease model (N.B. Holton does not use this terminology in the lecture).

As Holton suggests, both what I have termed the liberal account and the disease model have some empirical support. Advocates of the liberal approach can point to evidence suggesting that (i) people typically tend to give up on their addictions as they get older, and that (ii) various dis/incentives can have significant effects on whether addicts will consume (or abstain) from a particular substance. This is hard to reconcile with the view that addicts have no say in whether or not they engage in an addictive behaviour, as a strong version of the disease model might hold.

Nonetheless, there is also empirical support for the disease model; for instance, the fact that addicts frequently relapse very soon after successfully going through the pain of withdrawal is difficult to reconcile with rational choice theory. So too is the high degree of sensitivity of the addict’s craving (and relapse) to environmental cues

A good theory of addiction should be able to accommodate all of these empirical features. Holton suggests that increased understanding about the role of dopamine in addictive behaviours has been an important breakthrough in developing such a theory. Contrary to earlier views, it is not the case that addictive substances interfere with the dopamine system in way that simply gives addicts pleasure (which they then form an instrumental desire to pursue). Rather, the picture that Holton outlines is that addictive substances can radically disrupt the ways in which dopamine regulates ‘wanting’.

To see the significance of this, we have to distinguish ‘wanting’ from ‘liking’. Roughly, to like something is to take pleasure in it; I might like the taste of ice-cream for example (NB – this is my example, not Holton’s). Typically, when we like something, we are also motivated to pursue it; for me, ice-cream has motivational salience, and we may say that I want ice-cream. The wanting system here does two things: First it gives rise to an immediate appetitive urge (for ice-cream) now. Second it gives rise to a long term diachronic disposition to pursue ice-cream when it becomes available in my environment.

In this case, my liking ice-cream has led to my wanting ice-cream. Liking and wanting usually run together phenomenologically. Moreover, as Holton suggests in his discussion, there may also be good evolutionary reasons for a close interplay between liking and wanting, given our history as opportunistic feeders. Wanting is after all a kind of learning; we learn new wants as a result of being exposed to some substance. However, Holton stresses that these new wants are not instrumental wants for the pleasure the substance gives us. Rather, this system allows us to develop new intrinsic wants for the substance itself; I want the ice-cream, not the pleasure it leads to.

Crucially though, liking and wanting are neurobiologically distinct, and they can come apart. Whilst Holton suggests that natural opiates and cannabanoids seem to be responsible for ‘liking’, dopamine has a significant role to play in regulating wants; for evidence of this, Holton refers to Kent Berridge’s work showing that dopamine depleted hungry rats are not motivated to eat available food, even though they seem to experience pleasure when the food is placed in their mouth. Moreover, when the dopaminergic system is stimulated with amphetamines in these rats, they engage in eating behaviours.


Holton’s View of Addiction


Holton’s thesis about the nature of addiction is that in addiction, wanting may become dissociated from liking because of hyper-stimulation of the dopaminergic system. That is to say that the addictive substance retains motivation salience for the addict, even though they may no longer take pleasure in it. On this understanding, addicts may not like what they feel motivated to do.

How does this model accommodate the empirical features I mentioned above? Like the liberal account, Holton’s model can deny that addictive behaviour is compelled in the sense that addicts cannot but pursue an addicted behaviour. The reason for this is that quite generally wanting does not entail action: I may maintain a (non-addictive) want for ice-cream, but (as my GP and my dentist often remind me), I have a number of competing reasons not to continually act on the basis of that want.

Of course, addictive wants seem stronger than others; indeed, the fact that an addict may know that fulfilling their want will not bring them pleasure (remember, wanting can be divorced from liking in addiction on this model) typically does not prevent addicts from acting on the basis of these wants. However, this is quite compatible with the account; the way in which addiction disrupts the dopaminergic system means that resisting these particular wants will be very hard work (even post rehabilitation). But the point is that they can be resisted on Holton’s model, if the addict exerts enough self-control. Moreover, we can increase the likelihood of success by introducing other competing motivations; for example, we might financially incentivize abstinence.

Yet, if overcoming these hedonically dissociated wants is such hard-work, how do we explain the empirical evidence (speaking in favour of the liberal model) suggesting a long-term decrease in addictive behaviours as addicts age? Holton propose three (non-mutually exclusive) hypotheses: (1) As we age, our dopamine levels generally decrease naturally, so perhaps our wants diminish; (2) as we age, the number of competing motivations to not engage in addictive behaviours increases; and (3) as our frontal brain structures mature, our self-control may get stronger.


Moral Consequences


One general point that Holton’s model raises is that we should not be thinking of the responsibility of addicts in a strictly binary sense; the relevant question here is not so much whether or not addicts are free in toto, but rather how difficult it is for them to resist their urges. Outside of this general concern about responsibility, Holton raises two further moral implications of his model of addiction.


The Significance of Preferences


First, Holton claims that the story about how we come to form intrinsic wants seems to raise problems for the authority that we afford to desires in desire-fulfilment accounts of well-being (as well as indirectly in various forms of state policy): why should we be prudentially concerned with fulfilling desires if their fulfilment is dissociated from liking and therefore pleasure? Moreover, such desires can become dissociated from cognitively driven judgments about what is valuable; the desires are simply there because of previous exposure to that which we now desire.

Of course, desire-fulfilment theorists might stipulate that there are certain sorts of conditions for the kinds of desires whose fulfilment contributes to well-being. Perhaps we should only take fully informed desires into account? However, as Holton points out, addictive desires can be fully-informed and still persist. Are such desires perhaps instead problematic because they are not desires that one would hold if one was fully rational? But, Holton responds, consumption here need not be a result of irrationality; rather consumption simply indicates that one is responding to a desire that is isolated from rationality. Fully rational selves, suggest Holton would be equally vulnerable to addiction.

Finally, Holton argues that Frankfurtian second-order desires are unable to justify the authority of preferences; on the Frankfurtian approach the relevant question becomes which of one’s relevant second order desires are strongest; is it one which takes as its object the first order desire to consume, or is it one which takes as its object the first order desire not to consume? There is little reason to assume that the strongest second order desire of addicts is the one that takes as its object the first-order desire not to consume. In fact, Holton suggests that the best candidate for a neurobiologically supported second-order desire is simply the long-term dispositions that our wanting systems can give rise to – yet, in addicts, this is often just the disposition to occurrently want the substance in the future; and these can be strengthened by further consumption.

One alternative to the desire-fulfilment approach is to adopt a hedonistic approach to well-being; in this context, this means that we should focus on liking rather than wanting when we think about well-being. However, Holton claims that attributing such normative significance to pleasure lacks neurophysiological support; for Holton the best explanation of the existence of liking is to link motivational salience with our more cognitively driven planning responses. Pleasure is the root by which our wanting system can gets a ‘surrogate goal’ into our global, diachronic planning systems. The upshot of this is that pleasure only has limited significance; it is a proxy that enables us to get what we diachronically plan to achieve, plans that are grounded in a more cognitively driven system based in beliefs about what is valuable.


The Polluted Environment


In the last section of the talk, Holton defends the claim that agents who create environments that trigger cravings have a prima facie responsibility for those wants, and an obligation to refrain from providing the kind of stimuli that will reliably trigger such cravings.

To motivate this consideration, Holton asks us to consider whether sugar is addictive. On the one hand, sugar does not contain many of the ingredients that directly stimulate the dopaminergic system in the same way as other addictive drugs. However, because sugar is widely liked, it ends us being used by the dopamine system in the way that the system evolved to accommodate; but the result is that people develop strong cravings for sugar that look a lot like addictive cravings/desires.

The answer to whether we should view sugar as addictive then becomes one about the significance of the aetiology of such strong cravings; does it matter whether you have these cravings because the dopaminergic system was hi-jacked by a substance that is working through that system in a manner that the system had not evolved to accommodate (as in typical addictive drugs, but not in sugar), or is the mere existence of motivationally salient intrinsic desire sufficient for addiction?

Holton suggests that different sides to the debate on the addictiveness of sugar are arguably equivocating on the notion of addiction they employ. However, both sides agree that motivational salience is necessary. Consider now someone who is subject to such addictive cravings. Holton observes that it is clearly morally wrong to intentionally trigger such cravings in such people; indeed, he argues that third parties have an obligation to refrain from providing the kind of stimuli that will reliably trigger such cravings.

In support, Holton suggests that this is much like our thinking about pollution; we believe that those who pollute the environment should also be responsible for clearing up their mess, and that they have an obligation not to pollute in the first place. Indeed, as Holton observes, we employ a strict liability approach with regards to the offence of polluting the water supply. His proposal about the moral implications of his model of addiction is that we should extend this way of thinking to sugar – on his view, industries that create environments that are forcing people to encounter stimuli that will trigger a craving for sugar are broadly analogous to industries polluting an environment with smoke that causes others to choke.

Of course there are some disanalogies here – First, we like and want sugary foods, but we don’t want or like pollution. Second, we cannot avoid breathing but it seems that we can avoid eating chocolate. Holton concedes that there is some truth to these disanalogies but argues that they do not undermine his argument. The reason for this is that the wants in question in sugar are often dissociated from cognitive control, and they may be very hard to resist for the reasons outlined above. Accordingly, the mere fact that we want sugary foods but not pollution lacks normative force. Furthermore, there is a sense in which we can avoid pollution; we can move to a less polluted area! Rather, what this putative disanalogy suggests is that we need to think in terms of what can reasonably be expected of others. In the case of addictive wants, Holton has already outlined why such desires may be very difficult to resist; perhaps then we can only reasonably expect them to refrain from the addictive substance if it becomes rather easier for them to avoid the stimuli that may reliably trigger their cravings.




Holton’s second lecture raised fundamental questions about the nature of well-being, responsibility and third party obligations. Due to constraints of time, Holton understandably did not flesh out his sketch of a cognitively grounded approach to thinking about well-being. However, whilst his criticisms of (non-cognitivist) desire-fulfilment theories were powerful, advocates of such theories may be left with the suspicion that Holton’s own approach to well-being may be sailing a little too close to the objectivist wind for comfort. A key question then is how we should think about the source of the values that might undergird Holton’s cognitively driven understanding of well-being. More specifically, can subjectivism of any stripe play a meaningful role in well-being, given Holton’s concerns about both the formation of our wants, and the significance of pleasure?

In the Q & A, questions also addressed the implications of Holton’s account of addiction for valid consent. We may agree that the voluntariness of an addict’s behaviour is somewhat compromised by the difficulty of resisting their addictive wants; but the important question then becomes whether their voluntariness has been sufficiently impaired to render their decisions (to take drugs or to refuse rehabilitation) as invalid.

Finally, Holton’s discussion about whether sugar is addictive seemed to suggest that the debate between the liberal and disease accounts may be raising its head under another guise. In his model, Holton acknowledges that there are nuggets of neurobiological truth in both theories; indeed, the aim of his theory is to capture what is correct in both. However, it is not clear that it can provide us with an answer about whether the aetiology of the dopaminergic disruption in cravings matters. This is not necessarily a failure of the theory; however, it is perhaps suggests that it may not settle the debate between the disease model and the liberal model, at least in their spirit if not in letter. The reason for this is that it is hard to see how aetiology could matter on the liberal account; on the other hand, some advocates of the disease model might claim that ‘unnatural’ methods of influencing the dopaminergic system are more problematic than intentionally influencing the system to the same end, but doing so in a manner that the system has evolved to accommodate. Whether or not there is a plausible story to be told here remains to be seen; however, it is a salient question when we consider the prospect of sugar addiction, and the various other forms of behavioural addiction that are more frequently under consideration today. This is particular important given Holton’s views about the scope of our obligations to refrain from exposing addicts to triggering stimuli.


The Lectures series concludes tonight (Thursday 24th May) at 17:00 at the Magdalen College, Grove Auditorium, with “Illness and Attitude”

  • Facebook
  • Twitter
  • Reddit

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use the <em>, <strong> and <blockquote> tags. Links have been disabled to combat spam.

Notify me of followup comments via e-mail. You can also subscribe without commenting.


Subscribe Via Email