In ‘Salvaging the Concept of Nudge’ Yashar Saghai performs an important clarificatory task which certainly advances our philosophical and ethical understanding of nudges in public policy, and in healthcare ethics in particular.1 In this brief commentary I identify some issues which could usefully be taken forward in subsequent discussions.

A central difficulty with ethical discussions of nudging is that insufficient care is taken to distinguish two morally important features of nudges. The first, which Saghai very properly concentrates upon, is the mechanism of nudging. Nudges rely on psychological properties of human decision-makers as the way in which their intended effects are brought about. Much of the ethical concern with nudges focuses on just this. Whatever the motive of the nudger, or the objective of the nudge, or the ex ante or ex post ratification of the nudge by the nudge, operating through influence, impulse or non-rational features of the nudgee's decision-making is wrong. The nature of this wrongness can be spelled out in different ways, and which one of these we choose has important implications for the precise ethical argument we develop. It may be that working through non-rational mechanisms fails to respect the rational core of the agent, in virtue of which she or he has human dignity—a version of Kantian ethics. It may be that we cannot agree that a choice brought about through influence rather than conscious decision-making is free in the way that moral libertarians require. Consequentialists may be concerned that nudged decisions lack the quality of internal motivation which they might think is necessary to produce long term patterns of behaviour which are better all things considered. And so on. All these different accounts of the moral wrongness of nudging focus not on the content of the nudge, but on the way it works, and on (different) normative accounts of what decision-making ought to look like. Much of the time, this normative account is implicit, and one good consequence of the nudge debate has been to force moral psychologists to be more explicit about their normative theories of practical reason.

What the defender of nudges has on his or her side is that we cannot do good ethics without good psychology: it is no use complaining about the ethics of nudge if our normative theory of decision-making rests on an impossible or unsupported psychological model. But the argument here is delicate and uncertain: the proponent of nudges can mock the critic's naive account of free choice, for instance, but must show all the same that nudging respects freedom and autonomy in all the ways which matter. A very simple argument for a nudger to make would be to say—you worry that my nudge diminishes your freedom or fails to respect your freedom. But you needn't worry. You were never free in the first place. And while this is a respectable view, most nudgers do not want to say this. Yet it is not at all clear what middle course they do want to take between psychological determinism and psychological libertarianism.

The other issue which the nudge debate raises is not the how of the nudge but the what: rather than asking is nudging in principle acceptable, we need to ask for what kinds of behaviour change it is acceptable. Often, the way this is approached is procedural. We adopt a principle of ‘least restrictive means’ and place nudge in an informal ranking of means of behaviour change from pure coercion down to mere information. The acceptability of nudging in the context of interest then is partly determined by whether there are less restrictive means which are available and effective, and partly by whether the nudge itself is demonstrably effective and acceptable. It should be noted that the evidence base for most nudge interventions is pitiful—inference from small numbers of experiments with small numbers of subjects without properly controlled evaluations with decent follow-up periods is commonplace in the literature. But the open questions remain: who should be permitted to nudge, how can they be made accountable? Are there situations in which nudging would be wrong even if the nudge is effective, harmless and the least restrictive effective means available? And in the background of all this is the question Saghai alludes to early on but doesn't develop in his paper: when does a third party have a legitimate interest in changing my behaviour by whatever means? Setting aside those unproblematic cases where I want you to nudge me to help me to achieve something I want to achieve but defeat myself by my own self-subverting behaviour, we focus on those cases of warranted paternalism where we focus not on what I do want, or what my ‘revealed preferences’ disclose, but rather on what I would want were my desires properly informed and internally consistent. The nudge debate continually slides into this territory, even though, as Saghai argues, it is not properly an issue for ‘nudges’ as such. I suspect that the reason for this is that where moral paternalism points to defects in my moral reasoning which lead my desires to be improperly informed, nudges point to defects in my judgement which lead my decisions to be improperly framed or acted upon. Nudges and paternalisms focus on the defectiveness of my decision-making as a warrant for intervention. So what sorts of defectiveness, and under what conditions, do actually warrant intervention?