When the Risk of Harm Harms

This essay answers two questions that continue to drive debate in moral and legal philosophy; namely, ‘Is a risk of harm a wrong?’ and ‘Is a risk of harm a harm?’. The essay’s central claim is that to risk harm can be both to wrong and to harm. This stands in contrast to the respective positions of Heidi Hurd and Stephen Perry, whose views represent prominent extremes in this debate about risks. The essay shows that there is at least one category of risks – intentional impositions of risk on unconsenting agents – which can be both wrongful and harmful. The wrongfulness of these risks can be established when, on the balance of reasons, one ought not to impose them. The harmfulness of these risks can be established when the risks are shown to set back legitimate interests. In those cases where risks constitute a denial of the moral status of agents, risks set back agents’ interest in dignity. In these ways, the essay shows that there are instances when a risk can constitute both a wrong and a harm.


I. INTRODUCTION
Until fairly recently, the bulk of philosophical debate had revolved around a world of certainty, where outcomes were either previously determined or determinable. Moral philosophers had largely overlooked the moral dimensions and problems specific to risks. 1 The situation has changed, and an interest in risk and its philosophical facets has intensified. Unsurprisingly, this mounting interest has been accompanied by growing disputes. Among these, the moral status of risks is currently an area marked by both contentiousness and significance.
This essay enters the fractious fray by helping to clarify the moral status of risks. It does so by weighing in on a current debate related to risk impositions of harm. Within both legal and philosophical theorization on this topic, one question has been particularly divisive: Can a risk of harm be a wrong and harm? Two prominent extremes of the debate are represented by the positions of Heidi Hurd and Stephen Perry. Hurd argues that to claim that a risk of wrong is a wrong leads to vicious circularity, 2 just as arguing that a risk of harm is a harm leads to the same uncomfortable position. By contrast, Perry holds that while a risk of harm might be a wrong, it nevertheless cannot be a harm. 3 My essay stakes out its own theoretical ground and offers a particular conception of risk of harm able to show that both Perry's and Hurd's views are mistaken. The central claim to be explicated and defended is that to risk a harm is to both wrong and harm in certain cases although not in all of them. Section II offers clarification of main concepts in operation throughout the paper. Section III argues that in certain cases to risk harm is to act both wrongfully and culpably. On the basis of the preceding, Section IV argues that if a risk of harm wrongs then it can also harm. Section V addresses some objections to my position, among them those from Stephen Perry and Heidi Hurd. A short conclusion ends this work.
II. DEFINING TERMS Some conceptual clarifications are needed from the outset. My working definition of risk of harm is the following: a risk of harm is a likelihood of harm occurring. 4 To risk a harm, then, is to increase the possibility that, that harm will come about. 5 For purposes here, I assume a hybrid position between objective and subjective risk of the following sort. To risk a harm is to increase, in a non-trivial sense, the likelihood of harm occurring. I call this position a hybrid because I believe that what counts as non-trivial risking is established by way of a relationship between objective calculations of the probability of a harm occurring (and any increase in this probability), the expected severity of the harm should it occur, as well as a judgment about the importance or significance of the results. The view adopted seeks to give recognition to the dual nature of risks, which combines both objective and subjective features. Thus, judgments regarding the risking of harm maintain an objective frequentist basis, while nevertheless being rooted in normative and value-rich conceptions of both risks and harms as they appear in different communities. 6 To see better what I mean, consider for instance that a 'risk-loving' community might tolerate certain behaviors and activities that would be deemed too risky in a 'risk-adverse' one. For that matter, a community might not recognize certain actions as being risky at all, whereas another would classify them as such. These differences can be attributed to varying thresholds for risks or different viewpoints regarding what counts as harm. Any pure objectivity regarding these matters is difficult to come by because both individual and shared beliefs, as well as varying perceptions about risks and harms, can feature heavily during evaluation and judgment. In any case, I avoid such complications here in order to deal more directly with the main question that motivates this work. I take the definition of risk of harm provided to be limited, but overall sufficient for carrying forward with this theoretical pursuit.
At first glance it seems that, by definition, one could not state that a risk of harm can be a harm. This is because if the risk of a harm is defined as the likelihood of harm occurring then risk is but the potential for harm. Risk, thus, is at its core a chance, likelihood, or probability of harm, as opposed to an actual incidence of harm. 7 If risk is a probability of harm then it is a fraction or percentage, anything between zero and one or 0% and 100%. 8 It seems, then, 6 For a position which advances the hybrid view (between subjective-objective or fact-value conceptions of risk), see that risks are merely calculations or gauges of possible negative outcomes.
Although risks are often expressed in terms of probabilities, their existence, manifestation, and impact are far less abstract. Risks are not just measurement units. For instance, risks can constrain and limit otherwise available sets of options for agents. If walking in a particularly dangerous area can be represented as a percentage -let's say a 59% risk of incurring bodily harm -then some agents will not walk in that area because they will perceive the risk associated with it as being too great or the harm too serious or both. So while risks might be measured in terms of percentages or fractions, they are also perceived in normative ways as reasons for action. Risks, then, feature normatively in agents' practical reasoning.
The fact that risks can be reasons for action for agents does not refute the claim that by definition risks can never be harms. Nevertheless, the observation does point to the misguidedness of construing risks in a narrow fashion -as solely the estimates or measures of future harm like the definition implies. Even a brief consideration such as the one above identifies normative aspects of risks and paves the way for their broader understanding. When conceived of or analyzed in context, risks prove to be more than their initial definition indicates. They have normative features that resist narrow categorization.
Before ending the definitional section, it is important to point out another aspect of risks that needs to be addressed. The diversity of risks and actions that can be accurately characterized as risky is far too great to allow for blanket generalizations. After all, by just being agents in the world we encounter risks regularly and many times unthinkingly. Driving, drinking, eating, walking, and sometimes even talking in certain environments can be risky business. Thus, for purposes of conceptual clarity it is important to specify precisely the risks that are relevant here.
The kinds of risks I want to consider are limited to a certain category or grouping. They are those risks imposed by one agent or multiple agents onto others. Further, I only consider those risks of harm imposed intentionally onto others and to which no consent has been given. 9 The reason for opting out of considering those risks to which one consents is simply because consent can invalidate the wrongfulness of otherwise wrongful actions (when certain conditions obtain, such as the consent is freely given, the consent counts as informed). As the famous precept goes -Volenti non fit injuria -'to those who are willing no injury is done'. This principle goes some way in capturing the reason why risks to which one consents are not obvious candidates for wrongfulness and/or harmfulness.
The above explanation extends to the choice of investigating risk impositions that are intentionally performed. The most straightforward strategy to achieve the aim of this essay, which is to show that there are some risks that qualify as being wrongful and harmful, is to select first the most promising candidates for the job. In a second step, it is to assess whether or not the kinds of risks selected can in fact count as wrongful and/or harmful. Intuitively, the kinds of risks of harm that are the most promising contenders are those with the closest connection, in terms of both mental and act elements, to harms. These are, at least at first blush, intentional risk impositions of harm. 10 In line with the general definition of risk of harm from before, this essay considers those risks of harm that are intentional increases in the likelihood that harm will befall an un-consenting agent or agents.

III. RISKING HARM AS WRONGING
Any attempt to link the above definition of risking harm to moral evaluation depends not only on the operative concepts of risk and harm, but also on the place of an agent's mental attitudes toward the risk that her action poses. Here we might note that part of the problem of risking harm, at least as viewed by some theorists, is determining why an agent can be at fault. The response is perhaps deceptively straightforward: it depends on what the agent had in mind.
In beginning the evaluation of the mental component in intentional impositions of risk one distinction is salient. To intentionally increase the likelihood of a harm occurring is not necessarily to intend to bring about a harm. 11 Martin, for instance, may intend to make the drag-race he is organizing more dangerous by placing obstacles on the path of participant drivers. He intentionally makes the drag-race riskier by intentionally increasing the possibility of harm because he thinks that the race will be more challenging and fun this way. However, this does not mean that he intends to bring a harm into existence although one could still say that he intends to increase the likelihood of harm occurring. In fact, Martin does not want to harm and would be utterly devastated should any harm actually arise. Of course, whether or not participants in the race have given their consent will feature in the moral and/or legal assessment of Martin's actions.
The example calls for a distinction between intentional action in pursuit of risk of harm and intentional action in pursuit of harm. If an agent intentionally increases the likelihood of harm with the further intention of bringing (or helping to bring) that harm into existence then she intentionally acts in pursuit of harm. If an agent intentionally increases the likelihood of harm without the further intention of producing harm then she intentionally acts in pursuit of a risk of harm. Both types of actions intentionally risk harm and both can be wrongful, but for purposes here I only consider the first sort. The reason for this is the one indicated before: Out of the myriad kinds of riskings, the most plausible category that can meet criteria of wrongfulness and harmfulness is, at least intuitively, that with the strongest link between risk and harm.
To risk a harm in the relevant sense then, is to perform an action that increases the likelihood of harm occurring with the intention of bringing about a harm. Since the harm pursued has yet to manifest the action is incomplete. Such action can be wrongful even in the context of inchoateness even if, seemingly, no harm has yet been achieved. It is sufficient that an action be committed with the intention of producing a future harm. 11 I use acting intentionally or acting with intention interchangeably for stylistic reasons. However, concern for the distinction between the two makes me note that what I have in mind is acting with intention. Further, following Joseph Raz, I take acting with intention to be acting for a reason (as things appear to one). See Joseph Raz, 'Reasons: explanatory and normative', University of Oxford Faculty of Law Legal Studies Research Paper Series 13 (2007).
An action is wrongful when on the balance of reasons one ought not to perform it. 12 To see why an action can be wrongful simply in virtue of the fact that it is carried out in pursuit of harm consider that acting with the intention of bringing about harm is acting, at least in part, for the reason of bringing about harm. 13 Such intentional action is pro tanto wrongful because to act for the reason of ensuing harm is to act for a wrong reason. The action of risking harm is only pro tanto wrongful because there might be reasons that could justify or excuse it, like for instance, self-defense.
Nonetheless, to seek harm is to be guided by the aim of harming. In the absence of justifying and explanatory circumstances, to attempt harm is to act wrongfully. Following R.A. Duff, actions which aim to harm are formed around the intention of bringing about harm, are oriented towards it, and guided by the wrong that their intention involves. 14 Intentions become a crucial element in the evaluation of inchoate actions because they imbue actions with moral significance.
Recall that the object of moral evaluation is intentional action, which is performed in order to increase the risk of harm for the further intention of bringing about harm. What this means is that an agent who is acting for these reasons is in the process of attempting to bring about harm. The agent has not yet achieved her final aim. It is because the assessment of action comes before completion that intentions play such an important role. This is not to say that intentions would not count if the action was completed. It is plain that they would. However, when actions are inchoate, intentions become principal means of evaluation. 15 Only through the consideration and assessment of intentions can the moral significance of actions that are incomplete be discerned. Since in the realm of incompleteness actions have not reached their final destination one is not able to assess their consequences in their entirety. Nonetheless, potential consequences should be foreseeable 12 A. P. Simester, 'Enforcing morality', in Andrei Marmor (ed.), The Routledge Companion to Philosophy of Law (New York: Routledge, 2012), p. 481. 13 As Thomas Nagel writes: 'to aim at evil, even as a means, is to have one's action guided by evil […] But, the essence of evil is that it should repel us. If something is evil, our actions should be guided, if they are guided by it at all, toward its elimination rather than toward its maintenance. That is what evil means. So when we aim at evil we are swimming head-on against the normative current'. especially in cases of non-trivial and objective increases in the risk of harm (those under consideration here). Together with potentially harmful consequences, intentions help to clarify the moral importance of inchoate actions by capturing the reasons for which they are performed.
Although I cannot defend the claim here, there are good reasons to think that intentions should play a larger role than expected consequences in the moral evaluation of inchoate actions because the former are within the control of agents (agents with capacity) while the latter are not. 16 An agent who possesses requisite capacity is one who is capable of responding to reasons. 17 Because the agent possesses capacity she should be guided by the balance of reasons and is responsible for intentionally acting on reasons that deviate from it. 18 Setting aside open debates about free will and determinism, when a responsible agent intentionally risks a harm (intending to bring about a harm) she, presumably, has control over the formation of her intentions through deliberation, as well as control over her subsequent actions. However, the responsible agent may not be able to exercise control over the materialization of her intentional action into either risk of harm or harm. There are always factors outside the control of the agent which can interfere with her aims.
To put it in other words, the way in which the world reacts to the intentional action of risking harm is more up to the world than up to the agent. Such things as luck, character, and opportunity may increase or decrease both the probability of actions being performed and the probability that the harm sought by an agent will come 16  about. 19 Thusly, although risk of harm (i.e., probability of harm occurring) is calculable it is always a matter of uncertainty.
The action of risking harm with the requisite ulterior intention of producing harm is wrongful when the balance of reasons does not support such an action. If the action of risking harm is Y then Y-ing is wrongful if on the balance of reasons one ought not Y. If the agent performs Y then she can be culpable for this. The culpability of the agent is a function of the culpable mental state with which the action of risking is committed. In law, the combination of an agent's guilty mindmens rea -and bad actionactus reus -gives reason to assign blame to the agent. This is illustrated in the requirement of culpable purpose with which inchoate crimes such as attempt, solicitation, and conspiracy must be committed. 20 The above points to the fact that judgments of wrongfulness for incomplete actions are not dependent on whether or not the actions succeeded with respect to their aims. Both wrongfulness and culpability can be established prior to act-completion and in the absence of outcome materialization. However, when the action is that of risking harm with the further intention of producing harm successrelated complications arise. These can be put in the interrogative. If the wrongfulness of the action of risking harm is not dependent on the successful production of harm then are we not divorcing harm from risk? Further, if the wrongfulness of the action of risking harm is not dependent on the successful production of harm, is it nevertheless dependent on the successful production of risk?
To the first question, at this point no claim has been made about the appropriateness of considering a risk of harm as a harm. Nevertheless, the apparent divorce between risk and harm concerns the action of risking harm and the success in final outcome-harm production. So, the claim here is that the wrongfulness of risking does not depend on final harmful outcomes. To the second question, the success of producing risk is taken to be necessary. The assumption is that what counts as relevant risk is a non-trivial increase in the likelihood that harm will occur or, perhaps more accurately (as will become clear), that more harm will occur. Cases considered are 19 Stephen J. Morse, 'Reason, Results, and Criminal Responsibility', University of Illinois Law Review 363 (2004) those where the success in risk production is present because what is wrongful is to be unjustifiably exposed to certain kinds of risks.

IV. RISKING HARM AS HARMING
The above aimed to show that acting in order to increase the likelihood of harm in order to produce harm can be both wrongful and culpable. When agents act for the wrong reason of harm production they act wrongfully and culpably. The question now is, can we say that agents who act in these ways and for these reasons also bring about harm in doing so? More precisely, can we say that risking harm in the ways just described harms? If a risk of harm (under conditions specified) is a wrongful action, then the risk of harm (under conditions specified) can also be a harmful action, for surely to commit a wrong against someone can sometimes harm that person. That is, there need not exist a harm that is distinguishable from the wrong. To say: 'You wronged me' can mean that you have harmed some interest of mine. This interest, if it is a legitimate one, can make an action both wrongful and harmful. Inasmuch as an agent has a legitimate interest not to be wronged in a certain way that wrongdoing harms her. So, to act with the intention to commit a harm can be itself harmful in addition to being wrongful.
To give an example, if John grabs Jane with the intention of raping her then the act of grabbing her is wrongful because it is done with the intention to commit rape and is harmful because it violates Jane's interest not to be wronged in this way (i.e., be made the target of sexual assault). Jane is worse off on either a historical worsening or a counter-factual test. 21 Had John not acted in the way that he did, Jane would not have been the victim of his assault and her sexual autonomy and dignity would not have been violated; similarly, before John's action Jane was not a victim of assault. The act of grabbing fits within the category of risk of harm as defined here because it constitutes a non-trivial increase in the likelihood that (more) harm will befall Jane. John has embarked on the path of producing harm and at the point of grabbing Jane, assuming that he 21 Perry, 'Harm, History, and Counterfactuals'. has inflicted no physical injury, his action qualifies as an imposition of risk with the aim of harming.
Consider a different scenario. If John forcefully grabs Jane with the intention of pushing her away from incoming traffic he does not wrong her. This is because his action is now guided by an, all things considered, right reason. As for the issue of harming, Jane may indeed be harmed, assuming that the grabbing is forceful and bruising, but the harm is now justified because it is in Jane's best interests overall.
Being made the target of harm is what is wrong with the unconsented intentional imposition of risk of harm onto another person. Such imposition constitutes a setback to interests inasmuch as agents have an interest to not be made targets of harm and thusly wronged in this way. This interest can be viewed in terms of a dignitary interest, 22 but it is possibly related to a host of fundamental values such as autonomy, justice, self-determination, and the respect one ought to show for fellow human beings. This core human interest demands that we treat others in such a way as not to violate their inherent moral worth. 23 I doubt that much controversy could arise from a claim that there is a duty not to act wrongfully towards others, which correlates with a right not to be wronged. 24 An agent can be harmed through intentional action aimed at harming not because she has an interest not to be harmed in the way that the risk threatens to harm her (although the agent clearly has that right), but because an intentional action aimed at harming treats that agent without due respect. Such unfair treatment through active pursuit of harm can itself be harmful. The reason for this can be grasped if one understands agents as connected to other agents and the world they inhabit. Agents are agents in the world and an agent's relationship to the world is as much part of her life as her body or her wealth is. Agents are harmed not only when they suffer physical injury or material damage to their property, but also when their rightful place in the world is denied or jeopardized. When someone acts in such a way as to deny that agent's equal moral standing in the world, through action aimed at harm for instance, then that agent is harmed because her relationship with the world around them is now changed for the worse. By protecting agents' interest in dignity one protects their rightful moral status in the world.
One need not wait until the risk of harm materializes into harm to determine that an interest of the agent has been damaged. For the cases considered, before such actualization occurs dignitary interests are set back by exposure to the risk. It bears mentioning that other kinds of risk impositions can also damage agents' dignitary interests (e.g., reckless or negligent risk impositions). Moral agents are bearers of interests in dignity, and when risk exposures impinge negatively upon these then they too can be wrongful and harmful. Having said that, determining which kinds of risks damage dignity and which do not is beyond the ambition of this work.
Although the exact scope of appeals to dignity is notoriously difficult to specify, a few things can be noted. First, no one can violate another's dignity in the absence of a wrongful action. Thoughts alone cannot count as dignity violations. An agent must perform an actus reus violating moral law, which in our case is imposing a risk onto another as part of the pursuit of harm. Second, although dignitary interests can be set back in many different ways, what is at stake when it comes to dignity is the type of regard that can be rightly demanded from others. A violation of dignity will strike the core of an individual's moral worth -the recognition of her equality.
No doubt that more should be said about the interest in dignity and how it can be set back. However, dignity is a particularly difficult concept to exact although, or precisely because, it is so fundamental. 25 It is not possible in the scope of one article to offer even the sketch of a conception of human dignity. However, one need not 25 See Meir at note 23. provide such an account in order to show that risks can be harmful. Inasmuch as some risks (the best candidates being those which fit the conditions set) negatively affect the interest in dignity then these risks are harmful. This is part of the reason why the central claim I posit is modest. It only states that at least some risks can be harmful. In order to entertain this possibility it is sufficient to point to how we can conceive of risks as harms on the basis of an interest in dignity.
It is useful to contrast my view that risks can be harmful when they negatively impact dignity with John Oberdiek's position, which shares some similar aims. His conception advances the view that what is wrong with the action of imposing risks onto others is its impingement on autonomy. 26 The conception of autonomy he puts forward finds its roots in Joseph Raz's writings. In Raz's words: 'autonomy is exercised through choice, and choice requires a variety of options to choose from' 27 ; further, autonomy 'requires not merely the presence of options but of acceptable ones'. 28 Although I agree that risk impositions can sometimes limit otherwise available sets of options for agents or make some options less acceptable, I have doubts about the fit between Oberdiek's views and what is the wrongfulness or harmfulness of, at least, the class of risks with which I am concerned, which is that of intentional risking for harmful purposes.
To illustrate why Oberdiek's position raises doubts consider his own example, which belongs in the above mentioned category.
Oberdiek asks us to imagine a case when someone fires a gun at you. He writes: At the point in time that the gun is fired, but before the bullet emerges from the gun barrel, it is plausible to suppose that you are at risk of being shot. The question now is whether any interest of yours has been impinged, and specifically, whether your autonomy has in any way been diminished. To avoid irrelevant considerations, add to the example that you are completely unaware that you have been shot at. Has your autonomy been diminished? Under at least one possible scenario, the answer seems to be clearly 'yes'. If the bullet whizzes just past your head, then while you're lucky not to have been hit, it remains the case that your freedom of (safe) movement was significantly restricted -one move to the left would've spelled certain doom -and surely that constitutes a diminution of your autonomy. 29 In his article, Oberdiek aims to show that a right against risking can be established. The example he provides instantiates a risk (of being killed or injured) and seeks to show what is wrong with being put at such a risk. It seems clear from the example, however, that on Oberdiek's view what is wrong or morally significant about this case -someone almost killing another person is that this takes away a valuable option of the attacked, which in this instance is the latter's freedom of (safe) movement.
At least when it comes to such egregious acts, Oberdiek's position seems a bit odd. It might be the case that moving to the left would spell the end for the person who is about to be shot, but surely what is wrong about someone putting another at risk of severe harm (possibly an attempted murder) is not the contingent fact that the latter's freedom of movement is diminished or that some otherwise acceptable option has been removed or made unacceptable. The point is not about the logical conceivability of Oberdiek's position; rather, it is about its ability to capture what is at stake with certain kinds of risk impositions.
To clarify matters, I believe that Oberdiek's position is both insightful and right in pointing out that when risks impinge upon autonomy they are wrongful. In fact, at the beginning of the essay I adopted a not too dissimilar position to signal the error in understanding risks as mere gauges of future harm and disregarding their clear normative impact on practical reasoning. Nevertheless, I hold that Oberdiek's reliance on autonomy diminution to explain the wrongfulness of the kinds of risk impositions that have been the protagonists of this work does not fit the moral gravity of these cases.
For example, imagine a victim of an attempted murder explaining her situation in Oberdiek's terms. The victim might say: 'One move to the left and I would have been dead', but the meaning of the assault for them would not come from the fact that the option of moving to the left was taken away or made unacceptable. Rather, it would come, on a more plausible view, from the fact that someone had the audacity to try to kill them and could have succeeded in 29 Oberdiek, pp. 373-74. doing just that. This is not to rest the case on what a victim might or might not feel; rather, it is to point out that a reasonable person in like circumstances is likely to pinpoint the wrongness of the risk she was exposed to in the attack on her life itself. Before knowing whether the harm will be produced, whether the victim is aware or not of the peril and even if it is somehow known that the bullet will miss the target, the victim's interest in dignity will have been set back because she has been targeted and exposed to a risk of death or injury. In this case, the reduced number of acceptable options for the victim is, at least to my mind, of incidental concern.
One can easily imagine a case consistent with Oberdiek's specifications when someone attempts to fire a gun at a person in a coma, a person whose autonomy is already severely impaired. In this example, at the point that the gun is fired, but before the bullet emerges from the gun barrel, the comatose is put at a risk of harm. However, neither the person's autonomy generally nor her freedom of movement particularly is constrained by the action of risking. Nevertheless, we would still want to say that the risk imposed onto her is morally significant, wrongful and/or harmful. This example points to the fact that although risking can impinge on another's autonomy (understood as having acceptable options to choose from) and in this way, harm, it need not do this in order to be harmful. In the comatose person's case, the moral significance of the risk imposed onto her can be found in the meaning of the risk suffered and what that expresses. An account that relies on the reduction of acceptable options cannot capture the wrongness of risk impositions in this and similar cases.
To the latter point, my suggestion is that the moral significance of the risk imposed through firing a gun at someone is that it expresses a deep lack of respect for that person. In other words, it treats that person as a target for harm. I believe that such an action can be both wrongful and harmful. It can be wrongful because without excuse or justification it is wrongful to aim at harm; it can be harmful because the victim's interest in being treated with due respect is set back. What matters then is that placing another at a risk of death or injury expresses a lack of respect for that person's inherent worth as a human being. In such cases, what is necessarily impinged is the interest in dignity, while autonomy might also be constrained in the process, if incidentally.
It should be noted that the conception of harm which has been in operation here is, in its fundamentals, the Feinbergian sort. 30 It understands harm as the 'thwarting, setting back, or defeating of an interest' of a person. 31 In turn, one's interests consist of all those things in which one has a stake -those constitutive elements of a person's well-being. 32 Simply stated, harms and benefits connect to interests in the following way: other things being equal, an agent is better off if her interest is advanced, and she is worse off if her interest is set back or thwarted. Feinberg distinguishes between two types of interests: ulterior and welfare. 33 Roughly, ulterior interests are ultimate aims of agents, such as: achieving fame, becoming a sports anchor, or having a family. The peculiarity of these kinds of interests is that they represent personal aims of agents planning the course of their lives. By contrast, welfare interests are those interests that count as necessary means for the fulfillment of ulterior interests (whatever they are or turn out to be). In this category we find interests in life, bodily integrity, health, a sense of security, some degree of freedom from coercion or intervention, emotional stability, and so on. According to Feinberg, when these latter kinds of interests are damaged or violated, a person is very seriously harmed.
Although dignitary interests are not explicitly included in Feinberg's list of welfare interests, dignity is a basic requisite for the accomplishment of any life plans and, for this reason, seems to have a secure place among these. Welfare interests are those interests that are common to most human beings, and they are no more than basic requirements for human flourishing. Moreover, having one's dignitary interests undamaged is needed to ensure fair treatment as a moral agent, which is a prerequisite for all other kinds of treatments (e.g., legal and societal). Although the protection of dignity is certainly not enough for wellbeing, in its absence a person can be severely harmed. 30  One ought to be careful not to regard dignity as something one cannot lose or something that cannot be diminished because it is inherent in all human beings in virtue of their humanity. 34 Such an understanding of dignity cannot capture its vulnerability and, thus, the reason why it needs protecting in the first place. A relational conception can, however, place dignity within its wider network of moral and ethical human relationships, apply it between persons, and can, by this measure, capture the fact that dignity can be damaged by treatment that goes against it. 35 Framing dignity and determining what counts as a violation of dignitary interests is a difficult endeavor that ought not to be taken lightly. Although the boundaries of appeals to dignity remain to be drawn at a later time, it is sufficient for now to have at least made the case that certain risk impositions, such as those that target others for serious harm, can justifiably qualify as setbacks to dignity interests.
Having advanced the view of risks being harmful because of dignitary impairments, there is yet another sense of harm to consider -the normative sense -a wrong that counts as a harm. Drawing again from Feinberg, it can also be said that wronging another will almost always also set back a legitimate interest of that person 36 . As Feinberg writes: 'To say that A has harmed B in this sense is to say much the same thing as that A has wronged B, or treated him unjustly'. 37 In this way, there is little that separates wronging from harming, although there are still instances when someone can be harmed (in one sense), but because they have given their consent they will not have been wronged. In such cases, no perpetration of (non-material) harm occurs.
In spite of such complications, according to Feinberg: 'There can be wrongs that are not harms on balance, but there are few wrongs that are not to some extent harms. Even in the most persuasive counterexamples, the wrong will usually be an invasion of the interest in liberty'. 38 What this means is that even for those rare 34  cases when on the balance of interests, an agent ends up being advantaged by a wrong, a kind of harm will still have been perpetrated -that which captures the interest violated by the wronging. To see how the argument works, consider the example of the broken promise. Suppose that the unjustified violation of a promise wrongs the promisee, but that by some fortuity the promisee actually ends up being benefited as a result of the promise having been broken. Even though the promisee is now better off than she would have been had the wronging not taken place, there is still a sense in which she has been harmed because, in this case, her interest in being able to pin down or determine the future for herself was negatively impacted. 39 In light of the above-described two senses of harm, if an agent acts in order to increase the possibility of harm or in pursuit of harm intending to inflict harm then she may be already committing a harm. With the case for this particular view made, it is now time to consider whether there is something odd about claiming that risking a harm can be both a wrong and a harm. Heidi Hurd states that to risk a wrong cannot be a wrong (or harm) on pain of vicious circularity. Stephen Perry believes that risking harm cannot be harm.

V. OBJECTIONS TO RISKING AS WRONGING AND HARMING
Heidi Hurd claims that to say that a risk of wrong is a wrong leads one into vicious circularity. She writes that: 'Upon pain of vicious circularity, one cannot, as a conceptual matter, construe risks as wrongs. For to risk is to risk a wrong; and what is wrong cannot therefore be to risk'. 40 Correspondently, her claims ought to hold for risks of harms as well and indeed other theorists share her view.
Even though what I have argued in the previous sections should point to the conclusion that risks of harm can be both harms and wrongs, I want to address Hurd's objection directly and in so doing buttress my position. As I see the matter contra Hurd, circularity need not arise if one is careful in employing a distinction between an action-harm -call it H 1 -and an outcome-harm -call it H 2 . The action of risking harm is a harm, H 1 , distinct from the harmful 39 Ibid,p. 264. outcome, H 2 , which may or may not arise as a result of H 1 . The same distinction can hold for wrongs.
Circularity is avoided because: to risk a harm H 2 is to harm H 1 . In this sense, no circularity arises because any risk that is a wrong and/ or a harm will be different in kind from the ultimate wrong or harm at which the risk aims. The harms and wrongs at stake are, then, non-identical, and they can and should be distinguished from one another. The circularity objection stems from a misunderstanding concerning the nature of risks that can count as being wrongs or harms; this misunderstanding collapses the distinction between the latter and the kinds of wrongs and harms at which the risk aims.
Think of harm H 1 as a kind of proto-harm or inchoate harm, a seed-harm if you will, from which another harm, likely more pernicious, may or may not arise. This way of considering harm faithfully captures the idea of inchoateness which applies not only to incomplete actions, but also to incomplete harms. It is important not to disregard these seed-harms in part because they too can constitute harm. In addition, without some conception such as seed-harm, in those cases where the outcome harm H 2 does not manifest, we would have little to say about a perpetrator who got lucky.
To illustrate the latter point, consider the following example. An agent, D, takes aim to shoot and kill her mother, M, from a distance, but a car blocks her view and stops her from going through with the murder. Does D's action (aiming to shoot and kill) cease to be wrongful because a car got in her way? There seems to be no compelling reason to believe that the moral significance of D's action is altered by a car. Is M not worse-off than she was before her daughter's actions? My suggestion is that M is indeed worse off. In being made a target for harm, M's dignitary interest has been set back. Her daughter's treatment was profoundly unfair and violated her equal moral worth as a human being. Notice that in this example, aiming to shoot and kill is a harm H 1 distinct from H 2 , which does not manifest.
There are two other objections related to Hurd's claim about circularity that can be leveled against my account. The first is the problem of infinite regress, and the second is that of double counting. 41 The two problems relate to Hurd's circularity objection be- cause in trying to rebut her claim by using the distinction between H 1 and H 2 , another concern has arisen: it can be argued that when evaluating the harm suffered by an agent, if that harm is the final result of a risk of harm, then, given the multiple number of harms, all harms ought to be added together.
As an example, this could mean that an agent is worse-off if she suffers a broken limb as a result of an action that risked the breaking of her limb rather than if she breaks her limb absent a pre-existing harm. This conclusion is both problematic and rather counter-intuitive; a broken limb is a broken limb after all. Further, one could construe the harm of the broken limb as the broken limb plus an infinite series of ever-increasing risks of a broken limb. 42 To put it in other words, if an agent suffers a loss that is the result of a risk or series of risks of a loss, then that agent could be said to suffer the sum total of every distinct harm. This could mean that she suffers more (possibly infinitely more) than an agent that suffers the same loss without going through a chain of risks of loss.
The above captures the problems of double counting and infinite regress, which Claire Finkelstein addresses in her essay, 'Is Risk a Harm?' Finkelstein develops there a preference-based account on the basis of which she argues that a risk of harm is a harm. Although I am clearly sympathetic to the thesis that a risk of harm is a harm, her brand of preferentialism proves to be a problematic ground for arguing in support of it. John Oberdiek has astutely captured what are the most significant problems of Finkelstein's preference-based conception. 43 As such, it will not advance the state of the discussion to treat these issues again. Even in light of such defects, there are some insights from Finkelstein's conception that do not depend on preferentialism. I will employ them here.
The infinite regress and double counting objections can be avoided by detailing the preceding distinction between action and outcome-harms. If we properly distinguish between H 1 and H 2 and construe H 1 as the original seed-harm from which H 2 may or may not arise we can see that when H 2 does in fact manifest, H 1 will be assimilated within H 2 . The same will happen with all the other risks of harms which might precede the final outcome-harm. Hurd refers to this line of reasoning as the Absorption Thesis. 44 This view pays proper regard to the idea of a final outcome-harm arising out of or emerging from a risk of harm. To continue with this maybe not entirely felicitous metaphor of seed-harm from before, when a plant grows out of a seed the whole of the plant includes the seed from which the plant came. Analogously, there is a continuity between H 1 and H 2 such that H 1 and H 2 cannot be decoupled to the extent that H 2 is a result of H 1 . Thus, the separation and adding of harms does not appropriately belong to this understanding of risk.
The assimilation of risks' harmfulness into final outcome-harms occurs only in those cases where the ultimate harm manifests. At that point, given the causal chains between the risks and the final harm, the latter absorbs the risks that led to its creation. The two objections, double counting and infinite regress, are thusly rebutted because when the final harm manifests we do not count both the risks that resulted in the final harm and the final harm. The final harm encapsulates the chain of risks that brought it into existence.
Having attended to Hurd's circularity problem and a couple of additional objections, there remains the argument from Stephen Perry. In a nutshell, Perry believes that a risk of harm can only affect second-order interests. He writes that 'risking cannot be regarded as adversely affecting any interest that has a strong or plausible claim to be in the set of core or primary interests'. 45 Interests that can plausibly be included in the category of core or primary interest are: 'life, health, dignity, the physical integrity of the body, autonomy and freedom of movement, the interest in not experiencing severe pain, the interest in not experiencing severe mental or emotional distress, and certain kinds of property interest'. 46 For Perry, then, a risk of harm only affects second-order interests. These are those derived from first-order interests. An example might help elucidate his thought. If I have an interest in not being injured by you then I also have an interest in you not trying to injure me or put me at risk of injury. 47  My argument aimed to show that when conditions obtain, a risk of harm can set back a core interest in dignity. For Perry, this would be a first-order interest. Again, the core interest in dignity demands that we treat others with the respect they deserve. The interest can be set back when an agent becomes a target for harm. However, in order to advance his claim, Perry supposes from the beginning that a risk of harm does not affect a core interest. He writes: 'Let me assume that no such core interest is set back by the trying or risking'. 50 Of course, in doing so, he assumes his conclusion. This is because if one assumes that a risk of harm does not set back a core interest and then one defines harm as the setting back of a core interest then one cannot help but arrive at the conclusion that a risk of harm is not a harm.
Perry's claim that the moral significance of risk is derived from the harm at which the risk aims is, of course, appealing. In the formulation 'risk of harm', it is harm that is the evident moral notion and not risk. However, the preceding has tried to show that a risk of harm need not only give rise to the harm at which the risk aims. On the path to the aimed or final harm other kinds of harm may be brought about. Even absent final harm, other harms may still be effected through the act of risking. This, of course, does not refute Perry's claims, but it does suggest that the harm at which the risk 48 Ibid. 49 Ibid, pp. 1307. 50 Ibid, p. 1306 aims (perhaps primarily) is not the sole harm that should be considered.
As Perry acknowledges, we do not know whether or not an instance of risking will materialize into the harm it seeks. 51 It is for this reason, and especially when the harm sought does not materialize, that other kinds of harms represented by the risk are worth considering. My view offers an alternative to disregarding risks simply because the harm they sought happened to not manifest.
Perry has much more to say about risks and harms, and the above only addressed one aspect of his thought, which was taken to be most pertinent to endeavors here. Regardless of whether he is right about risks affecting second-order interests, it seems clear that risks need not only affect such interests. There are certain kinds of risks, those considered in this work, which can and do impinge upon core interests of agents.
Finally, at least one more objection can be raised: claiming that risks of harm can be harmful blurs the distinction between risk and harm generally. To this, I wish to restate that the position adopted here does not suggest that all risks of harms are harms, but only that under certain conditions some risks of harms can be justifiably considered harms. Further, the kinds of harms that can be appropriately identified as pertaining to risks are original inchoate harms (H 1 ) which may or may not give rise to outcome-harms (H 2 ). These qualifications should at least assuage the above worry.
To conclude, this section addressed at a glimpse several objections against construing risks of harms as wrongs and as harms, most notably those of Heidi Hurd and Stephen Perry. Although other points against positions such as mine might be raised and most likely will be, I take the objections addressed above as the more powerful in the literature. By providing answers to rival positions I hope to have strengthened mine.

VI. CONCLUSION
The essay aimed to show that there are instances when to risk harm is both to wrong and harm. It argued that at least one category of actions risking harm can meet wrongfulness and harmfulness crite-51 Perry, 'Torts,Rights,and Risk',p. 59. ria. In addition, this work addressed strong claims about the unsustainability of positions such as mine and distinguished it from other accounts.
Although my position is narrow inasmuch as it abstains from making sweeping assertions to the effect that all risks of harms are wrongs and harms, it does offer reasons to abstain from counter sweeping claims that risks of harm are never wrongs and harms. Although much work remains to be done, this should count as a step towards the clarification of the moral status of risks. ACKNOWLEDGMENTS Open access funding provided by University of Graz. I wish to thank the Centre for Advanced Studies (CAS) in Sofia, Bulgaria for a fellowship that enabled me to develop ideas present in this work. Early parts of this essay were presented at CAS seminars and published in the CAS Working Paper Series No. 7/2015. I also thank Stearns Broadhead, Lukas Meyer, members of the University of Graz's Philosophy Department, and participants in the 2015 'Law in the Risk Society' Conference at the Utrecht Centre for Accountability and Liability Law (UCALL) for their comments on early drafts of this work. Finally, I am grateful to two anonymous reviewers for their insightful written comments.

OPEN ACCESS
This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/ by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.