Agents, Actions, and Mere Means: A Reply to Critics

The prohibition against using others ‘merely as means’ is one of Kant’s most famous ideas, but it has proven difﬁcult to spell out with precision what it requires of us in practice. In ‘How to Use Someone “Merely as a Means”’ (2020), I proposed a new interpretation of the necessary and sufﬁcient conditions for using someone ‘merely as a means’. I argued that my agent-focused actual consent interpretation has strong textual support and signiﬁcant advantages over other readings of the prohibition. In the present essay, I respond to comments by Claudia Blöser, Irina Schumski, and Oliver Sensen. I ﬁrst address the role of maxims in relation to the Formula of Humanity, and I spell out in more detail the relation between the agent’s belief and the actual facts pertaining to the consent of the person who is used as means to the agent’s end. I then discuss the importance of the requirement that agents make their use of others conditional on the others’ consent as a matter of moral principle. I next show that the proposed reading of the prohibition, despite its focus on the agent’s maxims and practical reasoning, yields important conclusions about the moral status of external actions , and I address several cases that are purported to spell trouble for my reading of Kant’s prohibition. I end by outlining two challenges for possible (rational) consent readings of the prohibition against using others merely as means.


Introduction
Kant defines freedom as 'independence from being compelled by the arbitrary choice of another' 1 (MS 6:237; cp.V-NR/Feyerabend 27:1319-1322, GMS 4:446). 2 In line with the republican tradition in political theory, he contrasts freedom to slavery, domination, dependence, and other asymmetrical power relations in which one individual is subject to the unilateral power of another.Freedom, Kant continues in the Metaphysics of Morals, is really no different from equality (MS 6:237).
Although it has hardly been recognized in the literature, Kant's republican commitments shed light on his famous prohibition, in the Formula of Humanity, against using others 'merely as means '. 3 They reveal that the phrase reflects neither an idiosyncratic obsession with instrumentalization-as some critics have charged-nor an absolute prioritization of discretionary choice.Rather, the phrase prohibits one individual from arrogating coercive power over another and using the other at their own discretion.Thus, the prohibition is of a piece with Kant's most basic-republican-understanding of freedom and equality.
What this prohibition requires of us in practice, however, has proven hard to spell out with precision.In my article 'How to Use Someone "Merely as a Means"' (Kleingeld 2020), I propose a reading of the prohibition that articulates the necessary and sufficient conditions for using others merely as means, which I call an agentfocused actual consent account.The core of the article is exegetical and aims to develop an interpretation that best fits the textual evidence.I subsequently examine whether, on this reading of the prohibition, Kant's account runs into the problems commonly associated with other interpretations.I argue that it does not, and I respond to a possible objection.
Showing that my proposed reading avoids a particular set of problems leaves open the possibility, of course, that it suffers from other problems than the ones I discuss.This 'author meets critics' discussion in the Journal for Ethics and Moral Philosophy offers a great forum for exploring this possibility.I am very grateful to Claudia Blöser, Irina Schumski, and Oliver Sensen for serving as critics and for bringing up important issues and prompting me to develop my interpretation in more detail.I will start with a brief presentation of my agent-focused actual consent reading of Kant's prohibition (section 2).I then address the critics' comments concerning the role of maxims in relation to the Formula of Humanity (section 3), the requirement of actual consent (sections 4 and 5), the notion of perfect duties, and the moral status of external actions (sections 6 and 7).I end with a consideration of a possible (rational) consent reading of Kant (section 8).

The agent-focused actual consent account
Sensen calls my reading of Kant 'clear', 'sophisticated' and 'very plausible' and says that it has 'strong textual support' (Sensen 2024); Blöser and Schumski focus on the philosophical aspects of my reading and do not dispute my exegetical argument.Sensen disagrees with one important exegetical claim I make, however, namely the claim that the Formula of Humanity has implications for the agent's maxims.
To address his concerns and to provide the necessary background for the other discussions that follow, I will begin with a brief introduction of my proposed reading of Kant's prohibition against using others merely as means.
The Formula of Humanity articulates a categorical imperative addressed to 'you' as an agent.It does not tell you to perform specific types of actions.Rather, it tells you to act in a certain manner: 'So act that you use the humanity, whether in your own person or in the person of any other, always at the same time as an end, never merely as a means' (GMS 4:429).
You must never use the humanity in your own person or in that of another 'merely as a means' but always 'simultaneously as an end'.'Humanity', for Kant, refers to that which distinguishes humans from other animals, which is their capacity to act on the basis of rational principles and to set ends independently of inclination.Because this capacity is an inseparable part of the human being as a whole, Kant also employs the simpler expression 'using a human being', and I do the same (see Kleingeld 2020, 391).
In my article, I argued for the following reading of 'using another merely as a means': You use another person 'merely as a means' if and only if (1) you use another person as a means in the service of realizing your ends (2) without, as a matter of moral principle, making this use conditional on the other's consent, where (3) by 'consent' is meant genuine actual consent to being used, in a particular manner, as a means to your end.
In support of this reading, (1) I pointed out that some translations of the Formula of Humanity give 'treat' instead of 'use' but that 'use' is the better translation of '(ge)brauchen' and better fits Kant's terminology in the relevant Groundwork examples. 4I further argued that the relevant point of reference for establishing whether you 'use' another person 'as a means' is the practical reasoning that underlies your action, most importantly the general premises from which you reason practically, that is, your maxims.(2) In support of the inclusion of the consent condition, I referred to the Groundwork example of a false promise to repay a loan, which mentions consent, and to other passages.I explained that Kant motivates the consent condition by reference to the value of rational beings as ends in themselves.Humans are rational beings, and rational beings have freedom.They are persons, not things (GMS 4:428; V-NR/Feyerabend 27:1335).Hence your freedom is limited by that of others: you may not use them at your discretion; you need their consent.
(3) For the thesis that the relevant consent is actual consent, I referred not only to the first sentence of Kant's loan example-which introduces the example as a case in which actual consent is lacking-but also to many other texts, on topics ranging from contract law to the use of servants, in which Kant writes specifically that we would be using others 'merely as a means' if we used them without their actual consent (references mentioned on pp.392, 402-403). 5ncidentally, the prohibition against using others merely as means clearly condemns the use of other human beings as slaves, but it is important to note that Kant did not always draw this conclusion himself, at least not with this general scope.There is clear evidence that during the period when he wrote the Groundwork he defended race-based slavery, in any event in the case of indigenous peoples of the Americas and Sub-Saharan Africans, and that he justified it by reference to alleged deficits in their capacity for thought and action.Elsewhere I have argued that there is good reason to believe that Kant dropped this racial hierarchy during the mid 1790s (Kleingeld 2007(Kleingeld , 2014)).It is important to recognize, however, not only as a matter of the historical accuracy but also because it is important to think through the implications, that Kant combined the prohibition against using others merely as a means with a defence of doing just that in the case of some 'races' of human beings.

Maxims
Sensen disagrees with my claim that the Formula of Humanity concerns the agent's maxims.He writes that the Formula of Humanity concerns only 'legality' and that Kant is generally (not just in the Formula of Humanity) mostly concerned with observable actions, not 'attitudes' (Sensen 2024). 6Maxims do not play a role in Kant's formulations of the Categorical Imperative, Sensen contends, except in the Formula of Autonomy, which he sees as 'unique' in its 'requirement to have a certain attitude' (Sensen 2024).Maxims are important, if at all, 'only with friends, family, partners, and close associates', where we 'also care about the attitude of the other' (Sensen 2024).
In response, I want to point out that all other formulations of the Categorical Imperative in the Groundwork, from the Formula of Universal Law to the Formula of the Realm of Ends, explicitly articulate a normative criterion governing the agent's maxim adoption.Equally importantly, Kant reformulates the Formula of Humanity as a principle that is explicitly meant to govern the agent's maxims: [T]he formula [of humanity] says: that a rational being, as an end by its nature and hence as an end in itself, must serve for every maxim as the limiting condition of all merely relative and elective ends.(GMS 4:436, quoted in Kleingeld  2020:401) 7   This would seem to be clear evidence that, on Kant's view, the Formula of Humanity concerns the agent's maxims.In my original essay, I quote other similar passages as well (e.g., GMS 4:437-8).
Furthermore, the different formulas of the Categorical Imperative do not limit the importance of maxims to one's circle of friends and family.They hold as much for interactions with strangers as they do for interactions with friends, as is confirmed by Kant's fourth example (in which an agent does not care about a person in need and considers acting on a maxim of never helping).
This point of disagreement between us also affects our approaches to other issues.For example, Sensen suggests that my focus on using another merely as a means might be too narrow in that it would fail to condemn actions that he believes it should condemn.He illustrates this with a case where 'the chemical runoff from your mine poisons the well of your neighbour' (Sensen 2024).He sees this as a case of non-instrumental harm that would-incorrectly, in his view-not be ruled out by the prohibition against using others merely as a means.On my reading of Kant, however, the fact that the runoff causes harm does not give us enough information to know how to evaluate your action morally.We need to know more about you.For example: Did you know that the runoff was poisoning the well?If not, were you culpably ignorant, e.g., due to negligence, or was it an accident?What maxim were you acting on, and what was your practical reasoning?If you were deliberately using your neighbour's property as a convenient place to dump your chemicals, without asking her, on a maxim of self-interest, then you were using her merely as a means.By contrast, if you were negligent-acting, say, on a maxim of indifference-then you did not use her (merely) as a means, but you nevertheless violated the demands of the Formula of Humanity by failing to promote (the humanity in) her as an end.In both cases you failed to meet the demands of the formula, but only in the first case did you use her merely as a means.The mere fact of the harm does not permit us to discriminate between these cases. 8 do not believe that this result makes my account of 'using merely as a means' 'too narrow'.In fact, it is Kant's stated position that not all violations of the Formula of Humanity violate the prohibition against 'using merely as a means'.Sensen seems to regard it as a defect that my account is 'only of local importance'-by which I take him to mean that it applies only to a subset of all actions-but it is 'local' by design.Kant distinguishes cases in which agents 'contravene' the humanity in others from cases in which they 'fail to promote' it.On my reading, non-instrumental wrongdoing falls under the second category.It is not clear to me why Sensen believes that my interpretation suggests that I am 'giving up a literal reading' of Kant's Formula of Humanity (Sensen 2024); I don't believe I am. 9efore turning to Sensen's other comments (in section 6), I will first discuss comments by Schumski and Blöser that also concern the details of my proposed reading as such, namely its focus on the agent and its criterion of actual consent.

Actual consent and agential focus
Schumski asks just how 'radical' my agent-focused reading of Kant is, and in particular, if it matters at all, for the purpose of moral evaluation, whether the agent has the 'worldly facts' right.Is it enough that the agent believes they have the consent of the person being used?If so, then my reading would be 'radical' ('radically' agent focused, I suppose).Or must the agent's belief also be true?Then she would call my reading 'moderate'.I defend what Schumski calls the 'radical' view, but I would like to say more to avoid a possible misunderstanding.
Consider the following case.Immanuela, a medical researcher and good Kantian, is morally committed to using only test subjects who have given genuine actual consent in the experiment she is running.She takes due care to obtain free and informed consent, and when she runs the experiment, she fully believes all her test subjects have given genuine actual consent.Now suppose it later turns out that in one case the consent was actually given not by the test subject herself but by her trickster identical twin.Immanuela observed the morally required limiting condition, and to the best of her knowledge, she received all of the test subjects' genuine, actual consent.In terms of my reading of Kant's criterion, she did not use her test subject merely as a means.And indeed, in a case like this, she acted morally correctly.She was deceived.Now consider Maleficent, a different medical researcher who often cuts corners and uses test subjects in her experiments even when she fully believes they have not given their consent.Clearly, she uses them merely as means.This would remain true even if, unbeknownst to her, her test subjects were to actually give their consent to being used in the experiment-for instance if they were to give it to the researcher's trickster identical twin, who then kept all the records to herself (perhaps to test her suspicions about her sister's behaviour).When we morally evaluate Maleficent's manner of acting, I believe we should say that she acted morally incorrectly.She went ahead and used test subjects without having obtained their consent and without using their consent as a moral constraint, and thus she used them merely as means.
These results do not mean, however, that the agent's belief is all that matters.In both cases, the agent's belief is false due to deception by a third party.But if the agent herself bears moral responsibility for her false belief, for example if this belief results from a culpable failure, then the agent may well be using another merely as a means even though she believes she has obtained actual consent.
Schumski presents a pertinent scenario, namely a case ('sexual consent') in which a man has sex with a woman, believing that she consents even though she expresses 'visible discomfort' of a sort that ought to make him realize that she does not, or at least not genuinely, consent.Schumski wonders whether, on my agent-focused account, this man uses the woman merely as a means, given that my account is focused on the agent and in this case the agent believes he has obtained the woman's consent (Schumski 2024).
In response, I would like to emphasize that on the reading I have proposed, it is the agent's responsibility to obtain genuine actual consent.Schumski presents the 'sexual consent' case as one in which the man can and ought to realize, given the woman's visible discomfort, that she does not genuinely consent.On this construal of the case, the false belief that the woman consents results from the man's culpable failure to obtain genuine consent, for example because he pressured her into saying 'yes' or failed to take seriously her 'no'.In such cases, the agent uses the other merely as a means.Thus, the agent-focused actual consent account can indeed 'catch' this case, to use Schumski's phrase.
In general, Schumski views the 'radical' account as 'unusual' but 'perfectly sensible'.She believes that it has a radical implication, however, namely that 'the CI is not designed to assess actions at all' (Schumski 2024).I discuss this issue in section 6.

Actual consent, genocidal dictators, and permissibility
Blöser rejects my reading of the prohibition against using others merely as means on account of the role it attributes to actual consent.Her diagnosis of what she sees as the problem is as follows: 'one can actually consent to being used merely as a means, so actual consent (even if included in the agent's maxim) cannot be sufficient to make the use permissible' (Blöser 2024).She objects that the prohibition, on my reading of it, therefore 'rule[s] out too little' (Blöser 2024).Her comments in subsequent sections, including her argument that a 'possible rational consent' reading of the prohibition is preferable, are premised on this diagnosis (see below, section 8).
In response, I will highlight the importance of the clause, in my account of the prohibition, that you must make your use of others dependent on their actual consent 'as a matter of moral principle'.Thanks to this clause, the problem that Blöser diagnoses does not actually arise.As for the issue of permissibility, in my 2020 essay I indicated that it is necessary but not sufficient for moral permissibility that one avoids using others merely as means (Kleingeld 2020, 407).
Blöser uses a variation on my 'genocidal dictator' scenario to explain what she sees as the problem with my agent-focused actual consent reading.Imagine a genocidal dictator who subjects people to dangerous medical experiments and assume that she includes in her maxim the limiting condition that she will use only those who consent.Blöser parenthetically suggests that this limiting condition is motivated by the dictator's confidence that she will find enough consenting act utilitarians (Blöser 2024).So, the dictator's preference for consenting act utilitarians would seem to be a matter of convenience: this way, she will not have to deal with resistance.Blöser asserts that this dictator's action seems morally wrong and that she would be using her test subjects 'merely as means'.Because Blöser believes that my account of the prohibition would not yield this conclusion, she concludes that it rules out too little (Blöser 2024).
My account of the prohibition does entail that the dictator uses the consenting individuals merely as means, however.The reason is that the dictator in Blöser's scenario fails to adopt the limiting condition as a matter of moral principle.This is an explicit element of my reading of 'using merely as a means' (Kleingeld 2020, 398), but Blöser omits it from her own reformulations of my interpretation.10The dictator in her example seems to adopt the limiting condition as the road of least resistance, and in any event not as a matter of moral principle.Thus, she uses her test subjects merely as a means.Blöser's scenario in fact resembles the very example I used in my 2020 essay to illustrate this point.I discussed an example in which the genocidal dictator uses consent as a preference-based selection criterion rather than as a moral constraint (Kleingeld 2020, 405), and I concluded that this dictator uses the test subjects merely as means, in line with Blöser's analysis of the case.
When I introduced and explained the importance of the clause 'as a matter of moral principle', I referred to several passages in which Kant states that the limiting condition is required as a matter of valuing rational beings as ends in themselves (GMS 4:436-8; Kleingeld 2020, 401).Thus, the direct connection with the 'humanity' in persons, which Blöser says is missing from my account (Blöser 2024), is built into it from the start.You ought to adopt the limiting condition as a matter of valuing humanity as an end in itself; that is, you ought to adopt it as a matter of moral principle, not personal preference.The genocidal dictator in Blöser's scenario, by contrast, seems to restrict herself to using consenting subjects as a matter of convenience, that is, as a personal preference, not a moral principle.If she did properly value human beings as ends in themselves, she would not be a genocidal dictator in the first place.
In my 2020 essay, I mentioned that there are normative conditions that govern the giving of actual consent and that agents who ask others for consent should take these into account.This imperative does not 'undermine' the agent-focused actual consent account's ability to provide concrete moral guidance, as Blöser suggests (Blöser 2024); if anything, it makes it more concrete.After all, as a moral subject, I am conscious of my moral duties to myself and others, but I also know that you likewise have these moral duties to yourself and others.It would be wrong to ask you to consent to proposals to which I already know you ought not to consent.And if I do nevertheless ask and you do nevertheless consent, I cannot rid myself of my moral responsibility for this immoral project by referring to your consent.
In the final section of her comments, Blöser objects that the agent-focused actual consent account cannot properly criticize exploitation and structural injustice.To illustrate her point, she introduces cases of people who agree to buy overpriced water from a price gouger and people who agree to work in sweatshops.She seems to assume that because they agree to buy the water and work in the sweatshop, these cases present difficulties for my account.
My own analysis of these cases would focus on the maxims of the price gouger and the sweatshop boss and on the question whether consent in such cases counts as genuine consent.By definition, price gougers charge exploitive prices, and sweatshop owners employ their workers under oppressive conditions.They deliberately take advantage of the dire circumstances that force those in need of water or income to enter exploitive transactions under duress.When setting their prices high and their wages low, they act on maxims of self-interest, factoring in that their buyers and workers have no viable alternatives for obtaining their basic necessities and using them as sources of undue income.Although I recognize that there is room for debate, as Blöser mentions, I would argue that consent that is coerced under such circumstances does not count as genuine consent.This means that price gougers and sweatshop owners, as these terms are commonly understood, fail to limit their use of others, as a matter of moral principle, to what others genuinely consent to.Thus, the agent-focused actual consent account offers a straightforward condemnation of such practices.

Maxims, actions, and perfect duties
Both Schumski and Sensen raise interesting questions about the implications of my reading of 'using merely as a means' for the moral evaluation of external actions.Kant's prohibition, on my reading of it, focuses on the agent's maxims and practical reasoning.What, if anything, does the prohibition entail concerning the (im)permissibility of outward actions?It might seem that on a 'radically' agent-focused reading, the Categorical Imperative is-in Schumski's words-'not designed to assess actions at all' (Schumski 2024).Sensen asks more generally whether my account 'covers all perfect duties' and how attitudes and actions relate (Sensen 2024).In this section, I clarify what I see as the implications of the Formula of Humanity for an account of perfect duties and the moral status of external actions.
In the Groundwork, Kant introduces the distinction between perfect and imperfect duties provisionally, merely for the purpose of illustrating the application of the Categorical Imperative (GMS 4:421n.).He does not explain what he sees as the mark of perfect duties, but it is clear enough that they are strict duties, that is, duties that allow for no latitude (GMS 4:424) Kant further indicates that on his view there are both internal and external perfect duties (GMS 4:421n.),and thus some perfect duties concern external actions.He does not elaborate.
On the face of it, it could seem strange to ask whether, on my reading of it, the prohibition against using others merely as means 'covers all perfect duties', even when the question is restricted to perfect duties to others.There is no independently existing set 'out there' against which we can measure whether a particular interpretation 'covers' the whole set, nor does Kant give us a complete set in the Groundwork that can be used as a benchmark (and the Metaphysics of Morals is not Sensen's point of reference).The set of perfect duties must be determined by applying the Categorical Imperative.Thus, one possible response to Sensen's question is that my reading does cover all perfect duties because the set of perfect duties just is the set of strict duties that is derived by applying the formula-whether on my reading or any other.
This rather formal response does not address the status of external actions, however, so more needs to be said.I would like to show that the prohibition against using others merely as means has clear implications for the moral status of external actions, even though the prohibition as such governs the agent's maxims and practical reasoning.I first introduce Schumski and Sensen's central concerns and then develop a longer argument in response.
What Schumski sees as a possible problem for my reading of Kant's prohibition is that it seems to miscategorise cases of non-instrumentalizing wrongdoing.Her example is reckless driving, such as speeding for fun.Doing so does not entail using others as means, and hence it does not violate the prohibition against using others merely as means.Nevertheless, she observes that the moral prohibition against reckless driving does not come with latitude: it is strictly prohibited, which she takes to mean that it is a perfect duty not to drive recklessly.In the Groundwork, however, Kant associates violations of perfect duty with violations of the prohibition against using humanity merely as a means.As a result, Schumski wonders whether my reading of the Formula of Humanity can show that non-instrumentalizing wrongdoing is strictly prohibited (Schumski 2024).
Sensen's core concern is the general issue of whether, and if so how, my proposed account can determine the moral status of external actions.Although he praises the exegetical argument of my paper, he sees it as his role as critic to 'raise critical questions', which he does in abundance.Sensen defends a moderately consequentialist theory based on an empirical account of human needs (see Sensen 2021, 113), and consequently he finds himself agreeing more with Derek Parfit than with (my reading of) Kant.Rather than addressing his many questions separately, I will focus on the following main questions: 1. Whether the three conditions jointly suffice to make certain external actions morally wrong (Sensen 2024); 2. if so, whether an external action is always morally wrong if the agent uses others merely as means (as an example of a case in which he considers it morally right to use others merely as means, Sensen introduces a scenario-the 'quick thinker'-in which someone uses others merely as means in the service of his personal interests but does so in a way that happens to benefit them [Sensen 2024, Sensen 2021, 104]); 3. conversely, whether an external action can be permissible if the agent uses others merely as means in the process (as I claim is true in the wannabe hero case); and 4. whether, in order to settle these issues, we should not first spell out a criterion for morally right external actions.
As is familiar, Kant distinguishes the question whether an action is 'in accord with duty' from the question whether it is performed 'from duty' (GMS 4:397-398).This distinction suggests not only that he sees his theory as having implications for the moral status of external actions as such, but also that the deontic status of the external action does not always track the deontic status of the agent's internal underlying maxims.The shopkeeper in Kant's famous Groundwork example charges children the right price: this is his external action, and this action is in accord with duty.The shopkeeper does not act from duty, however, but from the maxim of selfinterest: this is his underlying motivation.
But how can we determine the deontic status of outward actions independently of the specific maxims of their agents?Kant does not explain for any of the formulas how he moves from a criterion that pertains to maxims to an account of the moral status of outward actions.I believe that we can make some progress towards an answer, however, by thinking through the results of applying the Categorical Imperative-in the case at hand, the Formula of Humanity's prohibition against using others merely as means.The argument that follows is somewhat tentative, however, given that Kant says very little about how he understands perfect and imperfect duties in the Groundwork.
A first result of thinking through the application of the prohibition against using others merely as means is the following.Those actions are strictly prohibited that can be performed only by using others merely as means.11Such actions are 'against duty'.An example is the action of 'transplanting organs from an unsuspecting healthy person into multiple patients who need these organs to survive'.Different transplant surgeons performing this action may be doing so on different maxims: for example, one may act on a utilitarian principle, another on a principle of financial self-interest.Whatever the details of the maxims they act on, however, since they use their victims' organs without asking, they clearly use the unsuspecting victim merely as a means.A similar analysis can be given for actions such as cheating (e.g., overcharging children in your store) and stealing, to mention two of Kant's examples.These actions are 'against duty', that is, strictly prohibited.You have a perfect duty not to perform these actions.
Second, we can generate a positive counterpart to this set of prohibited actions, namely, external actions that are in accord with what duty requires.Overcharging children is prohibited; put positively, shopkeepers ought to charge children the regular price (I bracket the option of charging less).Charging the regular price is 'in accord with duty'.It is what a virtuous shopkeeper would do, say, on a maxim of honesty or fairness.But it is possible to perform this outward action on morally impermissible maxims, such as the maxim of self-interest, as Kant's shopkeeper example illustrates.In that case, the external action is and remains in accord with duty, even though the agent's underlying maxim is morally impermissible.
Third, the moral status of the action and that of the maxim that underlies it can similarly diverge in the case of actions of imperfect duty, such as helping others in need.When you help someone in need on a maxim of self-interest, your external action is in accord with duty even though your maxim is morally impermissible (assuming your action does not aid an immoral project, such as helping a thief in need of assistance).The wannabe hero in my 2020 essay resuscitates a collapsed jogger in the service of his own greater glory.As in the shopkeeper example, his external action is in accord with duty, but his underlying maxim is impermissible.
Fourth, regardless of whether the outward action is in accord with duty, acting on impermissible maxims is and always remains morally impermissible.This is true for both violations of perfect duties and violations of imperfect duties.In the case of imperfect duty, latitude exists only with regard to acting on maxims of duty.For example, in the case of the duty of helping others in need, this duty requires that the agent adopt a maxim of helping others but does not specify when, whom, or how to help.The latitude exists only on the side of acting on the maxim of helping, however.There is no liberty or leeway to violate an imperfect duty by failing to adopt maxims of imperfect duty or by adopting maxims that run counter to the imperfect duty (for a more detailed argument, see Kleingeld 2019).
I believe that this analysis helps to address the issue of non-instrumental wrongdoing.Whether it involves speeding for fun or letting people drown, non-instrumental wrongdoing is strictly prohibited.In the Groundwork, Kant explicitly introduced the distinction between perfect and imperfect duties only for the purpose of his examples and did not mean to clarify, develop, or defend the distinction as such.But it is very clear that there can be no latitude concerning violations of duty.
In other words, non-instrumental wrongdoing is 'caught' by the Formula of Humanity: not by the prohibition against 'using merely as a means', to be sure, but that was to be expected since it is non-instrumental wrongdoing.It is fully compatible with my reading of 'using merely as a means', however, that speeding for fun is strictly prohibited.It can be done only on an impermissible maxim, and on Kant's account acting on impermissible maxims is strictly prohibited. 12he analysis in this section also suggests answers to the four questions I formulated above in reaction to Sensen's comments.The response to the first question is yes: the three conditions jointly suffice to make some outward actions morally wrong, namely those that can be performed only by using someone merely as a means.These include actions such as deceiving, stealing, and non-consensual organ harvesting as in the example mentioned above.
The response to the second question is: no, not always.If the outward action can be performed only by using others merely as means, such as in the non-consensual organ harvesting case, then the action is strictly wrong ('against duty').Whereas Sensen introduces the action of the 'quick thinker' as a counterexample on the grounds that it has good consequences for all involved, on my reading of Kant the quick thinker's action is 'against duty'.He coercively moves other people's bodies around without their consent, to suit merely his own interests.Thus, he violates the Formula of Humanity's prohibition ('never merely as a means'), and this does not change if, coincidentally, his action happens to have good consequences in particular circumstances.The man should have asked; the scenario does not stipulate that there was no time to do so. 13If, by contrast, an outward action is 'in accord with' duty, then it is permissible, even if the agent acts on an impermissible maxim and uses others merely as means in the process.
Therefore, the answer to the third question is 'yes'.Depending on the details of the case, an external action can be in accord with duty even if, given the underlying maxim and practical reasoning, the agent uses others merely as means in the process.The fact that the outward action happens to be in accord with duty does not exonerate the agent for acting on an impermissible maxim.Furthermore, contrary to Sensen's assumption (Sensen 2024), when Kant writes that such actions deserve 'praise and encouragement', he does not mean that he cares only about outward behaviour.Kant continues that very sentence by saying that these actions deserve 'no esteem' (GMS 4:398) since they have no moral worth.Giving CPR to people who need it is to be applauded and encouraged, but Kant does not regard it as sufficient, morally speaking: we ought to act from duty, on moral maxims, to ensure that our external actions are in accord with duty reliably and not merely accidentally, when it happens to suit our self-interested purposes.(Recall that the wannabe hero in my example lets collapsed joggers die by the side of the road unless reanimating them will serve his greater glory.) Finally, in response to Sensen's question whether an account of impermissible outward actions requires us first to spell out a separate criterion for doing so, the answer is no.The answers to the previous questions could be developed by reference to the Formula of Humanity.

Consent in lifesaving cases and coercion by the state
Schumski and Sensen would like to hear more about cases in which an agent gives CPR to someone who is unconscious, such as in the example of the collapsed jogger.Some authors have argued that on an actual consent account of 'using merely as a means', this would constitute using the jogger merely as a means, given the absence of actual consent.In my 2020 essay, I replied that the rescuer does not use the unconscious jogger as a means and hence does not use her merely as a means.
Schumski objects, however, that 'the use condition seems like the wrong resource' for dealing with this issue and that I should address the absence of consent (Schumski 2024).I believe it was the right resource for dealing with the original case, since it addresses that case at the most fundamental level: if the person is not used, she cannot be used 'merely as a means'.But I recognize the need for further discussion, if only to address cases in which the agent saves the jogger's life as a means to his own ends, as in my 'wannabe hero' scenario.
I submit that when joggers in parks fall unconscious and need CPR to survive, those who come to their rescue may act on a defeasible presumption of consent.This presumption is defeasible, because it may turn out, for example, that they have a living will stating that they do not want to be resuscitated.Absent proof to the contrary, however, we generally assume that people who suddenly collapse want to continue living.Although I will not argue the case in full here, I believe that this is indeed a better assumption, morally speaking, than the available alternatives, such as rolling the dice or letting people die in the absence of clear proof that they want to stay alive.On this assumption, then, a beneficent rescuer does not use the jogger merely as a means. 14Furthermore, the action of saving the life of the unconscious jogger is 'in accord with duty'.This remains the case even if, given the maxim and practical reasoning that underlies it, the rescuer uses the person merely as a means in the process, as in my wannabe hero case. 15 Incidentally, the presumption of consent to receiving CPR does not also entail that the jogger must be presumed to consent to serving as a means to the man's greater glory.He cannot ask the jogger to consent to serving as a means in his personal project of becoming a hero, since she is unconscious, and there is no reason to presume her consent in this regard.Thus, for example, it is not in accord with duty for him to post pictures, without her consent, depicting her in her unconscious state with his hands on her chest, to highlight his heroic deed. 1614 This is also what I would say in reply to Fahmy, who believes that on my account a beneficent rescuer necessarily uses the unconscious jogger merely as a means (2023,46).Fahmy further believes my account cannot handle cases of theft, because, she asserts, the owner of the stolen goods need not enter into the thief's practical reasoning (2023,46).However, this overlooks the way in which, on Kant's theory, my property belongs to me.When the thief steals something that belongs to me, he uses me merely as a means in using my property for his own purposes, just as he would use me merely as a means if he took and used my kidney without my consent.Whether he steals the kidney that belongs to me or the bag that belongs to me, as a means to promoting his ends, he uses me merely as a means.(This is not to say that my body is my property, but that both my body and my property are intimately related to me, cp.Kleingeld 2020, 391.)Fahmy in fact mentions this idea with approval in a related context (Fahmy 2023, 60, n.16).Hence I do not believe that the two cases that she presents as unsurmountable obstacles actually pose problems for Kant's account on my construal of it. 15I would like to warn against formulating this point by saying that the rescuer 'acts rightly' or 'acts permissibly', as some critics do.This formulation is ambiguous, since it does not specify whether the permissibility concerns his external action, his maxim, or both. 16Samuel Kerstein (2023, §3.3) rejects actual consent accounts (and specifically the one I attribute to Kant) with the following counterexample: a physician judges that the life of an unconscious patient can be saved only by undertaking an experimental treatment, and therefore she uses the patient in the study.Kerstein asserts that, on an actual consent reading such as mine, this physician would implausibly use the patient merely as a means.But this conclusion seems to be drawn too hastily.For one thing, as I argued above, it is defensible if not morally obligatory to (defeasably) presume consent to the treatment if this is indeed judged to be the only way to save the patient's life.For another, consent to receiving the treatment Sensen presents a different case that he believes spells trouble for my reading, namely coercion by the state.Here, he asserts, actual consent is missing, even though the action by the state 'is still right' (Sensen 2024).Within the confines of this reply I cannot explain Kant's legal and political theory in sufficient detail to give a full response.Elsewhere, however, I argue that Kant's views on this matter underwent significant development and that actual consent does play a crucial role in Kant's republican political theory of Toward Perpetual Peace and the Metaphysics of Morals.There he conceives of coercion by the state as resting on consent: not only the (imagined) consent of the parties to the original contract but also the actual consent of ('active') citizens to the laws to which they are subject.In the ideal republic, the freedom of the citizen consists precisely in living under laws to which 'he has given his consent' (MS 6:314; for further explanation and discussion, see Kleingeld forthcoming).

Possible consent and possible rational consent
Blöser and Sensen propose a 'possible rational consent' and a 'possible consent' account, respectively, of 'using merely as a means'.I would like to take this opportunity to formulate two challenges to these alternative accounts of 'using merely as a means'.
The first challenge I see is the lack of clear textual evidence.As Blöser understands the criterion, we use the humanity in others as a mere means if they 'could not rationally consent, that is, consent on the basis of [their] intellectual selves' (Blöser 2024). 17Sensen offers a pluralistic account of 'possible' consent (Sensen 2024;2021, 110-111).The evidence Blöser and Sensen cite in defence of these readings are negative formulations in which Kant claims that certain courses of action are wrong because others 'could not possibly consent'.These negative formulations do not specifically support possible consent readings as such, however, since these formulations are also compatible with different readings of 'using merely as a means', including my own (see Kleingeld 2020, 394).After all, if others could not possibly consent, then they also do not actually consent.
Furthermore, there is much textual evidence that points in the direction of an actual consent reading, ranging from the first sentence of Kant's false promising/loan example to his discussion of the use of servants (see Kleingeld 2020, 392, 402-403).On my view, a defence of a possible consent or possible rational consent reading of Kant's prohibition should therefore not only present positive textual evidence that supports this reading specifically (rather than passages that are merely compatible does not entail consent to being used in the study.In cases like this one, researchers will ordinarily ask for consent as soon as possible, and if the patient refuses to be included in the study, their data will retroactively be excluded.Therefore, I believe Kerstein's example poses no difficulty. 17Lina Papadaki similarly defends a rational consent account, specifying rational consent in terms of an end all agents are 'rationally compelled to have', namely the 'end of respecting humanity'.On her account, you do not use another merely as a means if they ought to consent to your using them (Papadaki 2016, 97-99).
with it) but also explain why Kant nevertheless emphasizes the importance of actual consent in so many contexts.
The second challenge is what seems to me to be an implication of their positions: the disappearance of strict duties.The problem here can be clarified by looking at the conditions under which Sensen and Blöser consider it justified to overrule the dissent of others who are used merely as means.Elsewhere, Sensen states quite clearly that using or treating others merely as means, for example by taking their property against their will, is justified when doing so is necessary to save lives or otherwise serve basic human needs (Sensen 2021, 107-108).Blöser's position is different, since she holds that using others merely as means is always wrong, but she argues that we do not use others merely as means if they could rationally consent to being used, for example based on their moral duty, even if they actually object.It would seem that on her view, rich misers could rationally consent to my taking some of their money and giving it to those in need, given their duty to help others in need, even though in fact, being misers, they strongly object.Blöser argues that when agents' actual consent reflects the priority they give to discretionary ends based on sensible desires, their consent is not an expression of their humanity or rational nature, and hence it is not 'protected' by the Formula of Humanity: what the Formula of Humanity is supposed to protect and promote is their humanity, not their discretionary ends (Blöser 2024). 18Relatedly, she argues that 'consent has normative power when it is an expression of one's rational nature' (Blöser 2024).
It seems, then, that on both Sensen and Blöser's views, though for different reasons, it would be morally right for you to steal money from affluent misers in the service of basic human needs (Sensen) or morally good ends (Blöser).By contrast, your stealing from good Kantians who spend their money on morally good causes, to increase your own wealth, would likely be wrong on their view, since the Kantians could not (rationally) consent to your doing so.In other words, it seems that under the principles proposed by Blöser and Sensen, stealing would not be strictly prohibited.This seems hard to reconcile with Kant's account of the immorality of stealing-and mutatis mutandis the same problem arises in the case of other perfect duties.
On the reading of Kant that I have proposed, by contrast, using another person as a means to your end against their objections is never morally justified, not even for the sake of otherwise morally obligatory ends such as the happiness of others.It does not follow, of course, that you should then give up those moral ends; you should find morally permissible ways to promote them.Furthermore, on Kant's view the genuine freedom and equality of individuals requires their joint subjection to collectively self-given laws in a republic (MS 6:314).To return to the previous example concerning basic human needs: on Kant's view, a just state should have a public welfare system (MS 6:326), such that those who are unable to provide for their own basic needs are not dependent on the charity or moral vigilantism of other individuals in the first place.Here is not the space to discuss Kant's legal and 18 Blöser claims that actual consent may be given by the 'phenomenal self' while the 'noumenal self' may disagree (Blöser 2024), but that claim does not sit easily with Kant's description of the phenomenal self as the appearance of the noumenal self (rather than describing them as two possibly disagreeing selves).political theory, but I mention this philosophical connection to underscore that his concern with individual freedom, understood as independence from compulsion at the discretion of other individuals, permeates his practical philosophy as a whole. 19pen Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.