Sarah Conly provides a number of convincing arguments in favour of what she terms ‘coercive paternalism’ (henceforth ‘CP’). CP is coercive in the sense that it stipulates that the state sometimes ought to intervene in order to force people to refrain from certain actions, even if those people would, in the absence of such intervention, perform those actions voluntarily. However, it is also a benign form of paternalism in the sense that Conly only argues in favour of state intervention in cases where, due to failures of instrumental rationality, people will fail to choose an appropriate means to their own autonomously chosen ends; CP is thus a paternalism of means, rather than ends.

CP has some striking applications. For instance, Conly claims that the state ought to ban cigarettes on the grounds that smoking represents a significant obstacle to the achievement of good health, and that people typically smoke as a result of cognitive error.1 She also recommends banning trans-fats, and limiting portion sizes in fast food restaurants.

Although I learnt much from Conly's discussion of cognitive biases and their relation to paternalism, I am not wholly convinced by her arguments in favour of CP. More specifically, I do not believe that Conly is able to completely absolve CP from the charge of perfectionism in the way that she claims; in my view, CP can be understood as veering towards a back-door perfectionism, and this is problematic.

Broadly, perfectionist theories claim that there are certain objective values, and that a person's life goes well to the extent that he or she is able to realise those values, regardless of his or her own subjective desires. This might seem troubling, because on perfectionist theories, the lives of persons might go better if they are coerced to pursue objective values that may be alien to them.

Conly explicitly distinguishes CP from perfectionism, claiming that the goal of CP is: … the advancement of individual welfare … construed as the maximization of the fulfilment of subjective ends (p. 103).1

Thus, Conly claims that while CP may mandate prohibiting those actions which serve to undermine the agent's pursuit of her own autonomously chosen ends, it does not mandate the pursuit of any particular end which might be alien to the individual. CP can only be used to promote the pursuit of ends that the agent already values.

However, it seems that CP can still be understood as being perfectionist in spirit (if not in letter). To explain why, it must first be acknowledged that some of our ends are mutually exclusive; we each often find ourselves valuing different ends that we cannot effectively pursue simultaneously. One end in particular that can come into conflict with the enjoyment of our other ends is that of health, which Conly identifies, in Wikler's words, as something that ‘… is valued for its own sake, and (that) is a means to almost all ends’.2 For instance, the activities that the gastronomes engage in to achieve their end of consuming the best tasting foods will often be inimical to their achievement of an optimum level of health.

Now, when CP mandates coercive measures in order to facilitate our health, it does not mandate the imposition of an alien value upon us; after all, health is a near universal value. However, while CP does not impose alien values, it may assume that we ought to prioritise certain of our values over others in a way that we do not ourselves endorse, and this seems to be similar to what we find worrying about perfectionism. Yes we value health, and we also make mistakes about how to achieve good health; but it seems that we can also rationally prefer the enjoyment of things that are bad for us, such as the ‘… whiskey, cigarettes, and fried food’3 that Feinberg urged us to protect from the advance of paternalism.

The paternalist might respond that health is a more fundamental value than our enjoyment of these other things, and that CP is thus justified in forcing us to prioritise this end over others. This though seems to smack of exactly what we seem to find troubling about perfectionism.

Alternatively, the paternalists might stress that they only intend to ban things that threaten health and which we only consume on the basis of cognitive error. Smoking falls into this category for Conly; she claims that people would not want to smoke if they were not subject to various cognitive biases (p. 169).1 I confess that I do not share this view, although I shall not pursue the point here. Regardless of this point of disagreement though, I do agree with Conly that paternalism might be warranted in preventing actions that threaten health and that are only carried out on the basis of cognitive error.

However, I also believe that we must take great care about what we presume that people are consuming only on the basis of cognitive error. More specifically, it is crucial that we avoid defining our terms in such a manner that it becomes impossible for agents to choose some action that poses a risk to their health without them being accused of making a cognitive error in weighing their values in that way. Health is a central value, but it is also one of many, and we should safeguard the theoretical and practical possibility of agents autonomously prioritising the pursuit of other values over the pursuit of their health. So, although I agree with Conly that there is a good case for banning trans-fats, insofar as it is nigh on impossible to see any value in the consumption of trans-fats per se, I believe that we can (and indeed many do) value the ends of enjoying whiskey, cigarettes and/or fast food without falling into cognitive error.