Abstract I argue that libertarian paternalism is in fact paternalism, or hard paternalism, rather than a form of soft paternalism. I do so on the basis of an analysis of the paternalist act according to which the paternalist act needn’t violate the will of the agent who is the target of that act and the paternalist actor need only suspect that her action may improve the welfare of that target. The paper considers and rejects interpretations of libertarian paternalism as soft paternalism. I then provide an outline as to how libertarian paternalism may be reformed in light of the finding that current libertarian paternalism is hard paternalism. The proposal is to require that the enactment of libertarian paternalist policies be conditional on opinion polling indicating support for those policies from the agents they target. 1. Paternalism In a recent work, I have argued for the analysis that S acts paternalistically towards R by doing (omitting) Z iff: (C1) S does so irrespective of what S believes the wishes of R may be; and (C2) S does so just because S has a positive epistemic standing that Z may or will improve the welfare of R (where this includes preventing his welfare from diminishing), or in some way promote the interests, values or good of R. (Ryan 2016: 124)1 I argue for C1 on the basis of consideration of a number of cases. One such case is the case of the happy paternalist (Ryan 2016: 125). The happy, say Victorian, paternalist acts for the good of his wife. Plausibly, his actions remain paternalist, even if they are acts that coincide with the wishes of his wife. For example, the husband may put a stop to visits to his wife by a cousin who he regards as a bad influence on his wife. Even though he does this irrespective of the wishes of his wife, it may be the case that his wife also wishes that these visits end. That she also wishes this to happen doesn't undercut the intuition that the husband's action is paternalistic.2 A plausible explanation as to why her wish doesn't undercut the intuition that her husband's action is paternalistic is the preemptory nature of the husband's action; his action is not conditional on her wishes. If his action were conditional on her wishes, then the action would no longer seem paternalist. C1, that the action is irrespective of the believed wishes of the target of that action, captures the pre-emptive nature of paternalistic acts generally. That C1 predicts which actions are not paternalistic acts and, in conjunction with C2, which actions are paternalistic acts provides justification for its inclusion as a necessary condition for the paternalist act.3 The reader should note that C1 contrasts with more traditional accounts of the paternalist act, according to which the paternalist act interferes with liberty or autonomy (Dworkin 2016; Beauchamp 2010; Arneson 1980).4Dworkin (2016), for example, specifically requires of the paternalist act that it ‘interferes with the liberty or autonomy’ of R. Later we shall see that the libertarian paternalist plausibly denies that the libertarian paternalist act interferes with the liberty or autonomy of the agents who are the targets of the act. This denial then serves as a basis for rejecting the categorisation of the libertarian paternalist act as a hard paternalist act. However, if the account I'm defending is correct, then the libertarian paternalist act may still be hard paternalist even if it doesn't interfere with liberty or autonomy. C2 is plausible given the conceivability of a paternalist act in which the agent acts from mere suspicion that his action may improve the welfare of the relevant agent, rather than from knowledge or belief that his action will do so. For example, suppose a father suspects there's a small chance that his son's welfare will be improved by taking vitamin tablets (Ryan 2016: 125). His son, however, refuses to take the tablets. If the father then hides a ground-up vitamin tablet in his son's food, motivated by the mere suspicion that doing so may improve his son's welfare, then plausibly the father's act is paternalistic. The moral then is that a mere positive epistemic standing that an action may or will improve the welfare of the target of that action is enough, knowledge or even belief that the action will do so is unnecessary. Note that, given the description of the example, it would be odd to say that the father believes the vitamin tablet will improve his son's welfare. After all, it is consistent with the description that he is almost certain that it won't. I use the term ‘positive epistemic standing’ rather than belief on the basis that there may be positive epistemic standings, such as suspicion, that do not involve belief. If there are, in fact, no such positive epistemic standings, then no harm is done by using ‘a positive epistemic standing’ rather than ‘a belief’. 2. Libertarian paternalism Libertarian paternalism is the view developed by Cass Sunstein and Richard Thaler (2003: 1162) according to which paternalist intervention is permitted to help an agent make the choice that promotes her welfare, where she would want to make this choice herself, but fails to do so because of weaknesses of her will or mind. The weaknesses of will and mind that the libertarian paternalist intervention is motivated to protect the agent from include the bad effects of procrastination, framing effects, the status quo bias, amongst others (Sunstein and Thaler 2003). Where there is libertarian paternalist intervention, choices are framed on the basis of findings in the psychology of decision making in such ways as to influence the choice of the relevant agent so as to make that agent better off. A libertarian paternalist act is accepted as paternalistic, in some form, as it entails an intervention motivated by the welfare of another agent. The intervention is libertarian, because, as the authors insist, it does not curtail the freedom of the other agent to choose, as traditional accounts of the paternalist act maintain. Even if the choice facing that agent has been framed so as to encourage her to make a good choice, that agent is still free to choose otherwise. A justification for this type of intervention is that in many cases there is no ‘neutral setting’ that will allow agents to make decisions in such a way that non-relevant features are prevented from influencing decisions. To use a favoured example, the libertarian paternalist does not advocate the banning of an opt out option when it comes to making pension contributions, rather given the libertarian paternalist’s stated commitment to freedom of choice and what we know about human psychology and the benefits of making pension contributions, they argue that there should be an opt out option but that it should not be the default option (Sunstein and Thaler 2003: 1184). Pension options generally should be framed in such a way as to attempt to influence or nudge the object of the paternalist intervention towards that which will make him or her better off.5 3. Why libertarian paternalism is hard paternalism, not soft paternalism This article shows, on the basis of the definition of the paternalist act provided in the first section, that libertarian paternalism is in fact hard paternalism. As discussed previously, a paternalist act requires that the paternalist actor is motivated to act on the basis of a positive epistemic standing that the action will or may improve the welfare of the agent who is the target of her action. The libertarian paternalist satisfies this requirement. What's more, the requirement neatly accounts for the libertarian paternalist act as a paternalist act in a way that nearby alternative conditions don't. Here's why. Libertarian paternalist acts of the sort typically offered as examples are such that they involve interventions that effect a large number of agents. The libertarian paternalist is motivated to improve the welfare of those who are the target of her intervention. Requiring that the libertarian paternalist be motivated by knowledge that her action will improve the welfare of the objects of her action predicts counterintuitive results and so should be rejected in the absence of a good reason not to do so. The libertarian paternalist simply doesn't know that her intervention will improve the welfare of each of the individuals who are the target of her action. Even saying that she is motivated by a belief that her action will improve the welfare of the targets of action is too strong. Given that her action has a large number of targets, it would be odd for her to believe, and indeed it seems unnecessary to require her to believe, that each one of the agents targeted by the libertarian paternalist act will be benefited by the act.6 Requiring, rather, that the libertarian paternalist is motivated by a positive epistemic standing that the targets of her action will or may be benefited by that action avoids such a counterintuitive result. This straightforward account can be contrasted with the usual (soft paternalist) account of how libertarian paternalism is paternalistic. As I show, the latter account fails to show how libertarian paternalism is paternalistic. In fact ‘libertarian paternalism’ has been described as ‘soft paternalism’ in literature on libertarian paternalism (Rebonato 2012: 24; Whitman and Rizzo 2007: 412). Others have used ‘libertarian paternalism’ and ‘soft paternalism’ interchangeably (Schnellenbach 2012). Dworkin (2016) has questioned whether libertarian paternalism should be thought of as paternalist at all, while Sunstein's (2014: 20) own position is that libertarian paternalist acts are generally soft paternalist acts.7 According to soft paternalism, intervention is permissible, if it is to ensure that an agent is acting knowledgeably and voluntarily (Dworkin 2016). The soft paternalist is motivated by the same sort of consideration as the hard paternalist, the good of the person who is the target of the paternalist action, but she limits herself to violating the other’s autonomy in order to ensure that the target knows what he is doing. If the soft paternalist knows that a man walking off the end of a bridge knows what he is doing and wants to walk off the end of that bridge, and so commit suicide, then the soft paternalist will not stop him from doing so.8 On the view of paternalism, or hard paternalism, as discussed previously, the paternalist act needn't violate the autonomy or liberty of the object of the paternalist act.9 Rather, what is required for an act to be paternalist is that the act is carried out irrespective of what the wishes of the target of the paternalist act are believed to be. This marks a difference from soft paternalism, according to which action is justified to confirm the will and knowledgeability of the target of the action. A libertarian paternalist act does plausibly meet this paternalist condition, given the pre-emptive nature of the libertarian paternalist act. Unlike the soft paternalist, the libertarian paternalist doesn't advocate an intervention to check the will and knowledgeability of her target, she acts to promote an outcome she thinks will be good for her target. While it's true that she does show deference to the will of the object of the paternalist act, that deference applies only after the intervention has been made.10 Her initial action is a hard paternalist action.11 This is in contrast to the soft paternalist, whose intervention on the one hand depends on not knowing whether a target agent is acting knowledgeably and voluntarily and on the other hand merely amounts to checking if a target agent is acting knowledgeably and voluntarily. Therefore, standard accounts of soft paternalism do not capture libertarian paternalism. Returning to the pension case, the libertarian paternalist calls for intervention to frame pension options in such a way so as to encourage an outcome that the libertarian paternalist suspects will or may make the objects of her intervention better off. The libertarian paternalist might claim that she doesn't do this, that what she's advocating is intervention to help the agent do what she would want to do if it weren't for weakness of mind or will.12 However, considering the examples of proposed paternalist interventions – pension choice architecture, speed lines to create the impression that one is going faster than one is when driving, the careful placement of fruit in canteens and shops – the libertarian paternalist project isn't generally to help people think more clearly or ensure, as the soft paternalist might, that one's choice is informed. The project, based on examples given of libertarian paternalist interventions, is to make certain outcomes that may benefit their targets more likely. The libertarian paternalist might begin to respond to the charge that her position is a hard rather than soft paternalist position by rejecting Dworkin’s definition of soft paternalism and instead appealing to what Feinberg (1986) writes of soft paternalism. Feinberg (1986: 12) writes that ‘[s]oft paternalism holds that the state has the right to prevent self-regarding harmful conduct… when but only when conduct is substantially nonvoluntary, or when temporary intervention is necessary to establish whether it is voluntary or not’ (Feinberg’s own emphasis). The libertarian paternalist might claim that in cases in which an agent is subject to weaknesses of mind, such as cognitive biases, or weaknesses of will, the agent’s conduct is substantially non-voluntary. Such a claim, however, is implausible. In neither personal nor legal settings do we take the fact that someone has been biased or, say, lazy to imply that their action is non-voluntary. If the libertarian paternalist wants to argue that all such actions are non-voluntary, the burden is on her to make that case.13 4. The future of libertarian paternalism If we accept the preceding argument, and therefore accept that libertarian paternalism is hard paternalism, then libertarian paternalism is subject to the same criticisms that have been made of hard paternalism in general. This doesn't, however, automatically mean that we should reject libertarian paternalism. In fact, a number of responses are open to the libertarian paternalist. One option is simply to defend hard paternalist action, or at least some hard paternalist actions or a form, say the libertarian paternalist form, of hard paternalist action. Such defences are current in the literature on paternalism (Conly 2013; Ryan 2016). I will not, however, explore this option here. Another option is to soften libertarian paternalism without losing what's valuable about libertarian paternalism. A criticism of hard paternalist action is that it doesn't give appropriate consideration to the wishes of those agents who are the target of the action. Libertarian paternalism is advocated as a policy tool for governments (Sunstein and Thaler 2003). In that domain, governments are applying psychological findings and using their position of power to promote outcomes that they believe are best for the people they govern or some subgroup of the people govern. A particular concern with this is the appropriateness of governments in liberal democracies adopting such a stance towards people they govern (Schmidt 2017). One way of addressing this concern is to require approval of particular proposed libertarian paternalist policies from those who are the intended targets of those policies.14 If they were to endorse particular proposed libertarian paternalist actions, then the issue of the appropriateness of the stance described wouldn't arise. Such a move would not only go a significant way towards addressing the described concern but it would also confer a degree of legitimacy upon such policies.15 Of course, it is impractical to require that the intended targets vote on every libertarian paternalist act proposed by governments. Nevertheless, that's not the only mechanism by which governments might be alerted to the public's approval or disapproval of a proposed libertarian paternalist act. Opinion polling that could reasonably be expected to provide a representative sample of the relevant target group could be one such mechanism. If a poll found approval for a particular proposed libertarian paternalist action, and those governing wouldn't undertake such an action without such approval, then those governing couldn't reasonably be criticised for acting irrespective of the will of the targets of their action. In other words, such an action would not be a hard paternalist action as it would not satisfy C1. But what about those who disapprove? A government sensitive to the will of those whom they govern would have reason to tailor proposed libertarian paternalist acts so as to avoid subjecting those who disapprove to such acts. Based on opinion polling this needn't be difficult to do. If those who disapprove in polls are found to be representative of identifiable cohorts, then those governing could tailor their acts to avoid subjecting those cohorts to those acts, if practical.16 This may of course not be practical. Nevertheless, if a government were to act in the way described, tailoring proposed libertarian paternalist acts when practical, then their libertarian paternalist actions would target those who approve of such acts, even if people who disapprove were unintentionally subjected to those acts. Whether such a reformed libertarian paternalism would still be rightly thought of as paternalism of any sort is open to question, but that is not my concern here. Reformed libertarian paternalism, one that depends on the approval of potential targets, is an elegant response to the realisation that unreformed libertarian paternalism is in fact hard paternalism.17 Footnotes 1 This analysis draws on and challenges Dworkin's (2016) own analysis of the paternalist act. 2 Seana Shiffrin (2000) also holds that a hard paternalist act needn't conflict with the will of an agent. On Shiffrin’s account, a paternalist act needn't be motivated by the welfare of the agent targeted by the paternalist act. Rather Shiffrin holds that the behaviour is directed at the target’s own interests or ‘interests or matters that legitimately lie within’ her control. I reject Shiffrin's account for a number of reasons. One reason is that Shiffrin's account predicts counterintuitive results. The following example illustrates this point: ‘Suppose Arthur is an artist who has left his paintings with Mindy to mind. Mindy judges Arthur’s paintings to be aesthetically suboptimal and highly values the aesthetic. Arthur’s artistic work is a matter that legitimately falls within Arthur’s area of control (and not within Mindy’s). Aesthetically Mindy believes she can improve on Arthur’s work and on that basis she makes various changes to the paintings. Mindy isn’t in the least interested or concerned with Arthur’s interests – we can even say that she generally wishes Arthur ill’ (Ryan 2016: 134). Intuitively Mindy's act is not paternalistic, although Shiffrin's account predicts that it is. Shiffrin might respond by saying that her account is not intended to be an exceptionless account of the paternalist act. Her account is couched in terms of how paternalism may be characterised. Even so, other things being equal, we still then have reason to prefer an elegant two condition analysis over Shiffrin’s four condition characterisation. 3 While C1 does not require that the paternalist act violate the will of the target of the act, it preserves the insight that the paternalist act generally poses a threat to the target's autonomy. This is as we would expect from a condition that requires that the act is carried out irrespective of the wishes of the target of the act (Ryan 2016). 4 For further discussion of these authors with regard to this requirement, see Ryan 2016 and Groll 2012. 5 The authors point out that there is scope for disagreement between libertarian paternalists on the respective weight to afford the libertarian and paternalist components (Sunstein and Thaler 2003: 1185–86). The particular libertarian paternalism that Sunstein and Thaler are advancing is one that ‘… attempts to ensure, as a general rule, that people can easily avoid the paternalist’s suggested option’ (Sunstein and Thaler 2003: 1186). 6 Many of us would, for example, opt in to pensions regardless of the architecture of the choice. 7 Mark White (2013: 90–95) argues that libertarian paternalism is soft paternalism, but via a different argument than this article provides. 8 Mill (1863) provides essentially the same example and the characterisation of soft paternalism above is based on Mill’s discussion of soft paternalism (Dworkin 2016). 9 Note that those in the debate use ‘hard paternalism’ and ‘paternalism’ interchangeably while ‘soft paternalism’ is treated as a different category altogether. I follow this usage. 10 Because of this deference and because libertarian paternalism only justifies intervention in cases of weakness of will or mind, I'm not claiming that the libertarian paternalist does not care at all about the autonomy of her targets. Nevertheless, it remains the case that these interventions are motivated by positive epistemic standings on the part of the libertarian paternalist regarding what may or will make targets better off. The interventions are not conditional on the support of those targeted by the interventions. 11 That a supporter of a libertarian paternalist act might not support a more stereotypical hard paternalist act obviously doesn't have the implication that a libertarian paternalist act is not a hard paternalist act. Similarly, that libertarian paternalism attaches some value to autonomy doesn't have the implication that a libertarian paternalist act is not paternalistic. The same may be true of hard paternalism even in a stereotypical case. A parent might, for example, force her son to attend school because she thinks it will be good for him in that it will develop his autonomy. 12 Of course, a Victorian paternalist might similarly justify his action by sincerely claiming that his wife is of the weaker sex, of will or mind. In fact this is a sort of justification we'd expect from a hard paternalist. (That one sort of paternalist might have good evidence for their claim and the other not doesn't bear on whether an action is hard paternalist or not.) 13 Of course, the libertarian paternalist might appeal to a different definition of soft paternalism again. While providing an alternative definition of soft paternalism is a possibility, such a definition would have to be defended against alternative definitions of soft paternalism and it would have to be shown that libertarian paternalism fits the alternative definition. Obviously an alternative definition simply shaped by the goal of accommodating libertarian paternalism would be problematic. 14 This would also test the claim that a libertarian paternalist act involves an intervention to help the agent do what she would want to do if it weren't for weakness of mind or will. Why not try to find out whether this is really so? 15 Even if there was endorsement of libertarian paternalist actions from the targets of those actions, one could still object to the appropriateness of governments undertaking such actions. For example, one might worry that such actions lead to the infantilisation of a population. 16 Similarly, in a case in which a majority disapproves of a particular libertarian paternalist act, if an identifiable cohort among the minority approve, then a government could perform a particular libertarian paternalist act aimed at improving the welfare of that group. 17 I thank anonymous reviewers who provided me with helpful feedback on this paper and an Associate Editor for this journal. This study was supported by the Taiwanese Ministry of Science and Technology (MOST 104-2811-H-031-006). References Arneson R.J. 1980. Mill versus paternalism. Ethics 90: 470– 89. Google Scholar CrossRef Search ADS Beauchamp T.L. 2010. Standing on Principles: Collected Essays . Oxford: Oxford University Press. Conly S. 2013. Against Autonomy: Justifying Coercive Paternalism . Cambridge: Cambridge University Press. Dworkin G. 2016. Paternalism. In The Stanford Encyclopedia of Philosophy , ed. Zalta E. N.. http://plato.stanford.edu/archives/win2016/entries/paternalism/. Feinberg J. 1986. Harm to Self . Oxford/New York: Oxford University Press. Groll D. 2012. Paternalism, respect, and the will. Ethics 122: 692– 720. Google Scholar CrossRef Search ADS Mill J.S. 1863. On Liberty , 2nd edn. Boston: Ticknor and Fields. Rebonato R. 2012. Taking Liberties: A Critical Examination of Libertarian Paternalism . New York: Palgrave Macmillan. Ryan S. 2016. Paternalism: an analysis. Utilitas 28: 123– 35. Google Scholar CrossRef Search ADS Schmidt A. 2017. The power to nudge. American Political Science Review 111: 404– 17. Google Scholar CrossRef Search ADS Shiffrin S. 2000. Paternalism, unconscionability doctrine, and accommodation. Philosophy and Public Affairs 29: 205– 50. Google Scholar CrossRef Search ADS Schnellenbach J. 2012. Nudges and norms: on the political economy of soft paternalism. European Journal of Political Economy 28: 266– 77. Google Scholar CrossRef Search ADS Sunstein C.R., Thaler R.H.. 2003. Libertarian paternalism is not an oxymoron. University of Chicago Law Review 70: 1159– 202. Google Scholar CrossRef Search ADS Sunstein C.R. 2014. Why Nudge? The Politics of Libertarian Paternalism . New Haven/London: Yale University Press. White M.D. 2013. The Manipulation of Choice: Ethics and Libertarian Paternalism . New York: Palgrave Macmillan. Whitman G., Rizzo M.J.. 2007. Paternalist slopes. NYU Journal of Law and Liberty 2: 411– 43. © The Author 2017. Published by Oxford University Press on behalf of The Analysis Trust. All rights reserved. For Permissions, please email: email@example.com
Analysis – Oxford University Press
Published: Jan 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.
Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.
It’s easy to organize your research with our built-in tools.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera