Abstract
Some strikes seem insufficiently discriminating. Rather than being aimed exclusively at potentially ‘legitimate’ targets (e.g., employers who, by refusing to pay a fair wage or provide acceptable working conditions, might have made themselves liable to bear certain costs), these strikes are (also) aimed at individuals who do not seem to be liable. Most problematically, such strikes invite the charge that they harm the innocent opportunistically or exploitatively. (Call this the third-party (exploitation) objection.) In other words, those who strike face the charge that they are harmfully using the innocent as a means to further their own ends, thereby violating a central and enduring principle of deontological ethics, namely, the means principle. This paper aims to set out the best case that can be made for the third-party (exploitation) objection.
I
In the winter of 2022/3, Britain faced ‘the biggest wave of industrial action … for a generation’ (Guardian, 2023a). Against the backdrop of the Russo-Ukrainian War, which precipitated an energy shock that pushed inflation to its highest levels in almost half a century, a single day in February saw over half a million workers out on strike. When members of the Royal College of Nursing struck in December, they did so for the first time in the union's 106-year history (Guardian, 2022a). In the wake of action undertaken by teachers, academics, barristers, rail workers, and others, the country braced for a ‘year of discontent’ (Guardian, 2022b). Meanwhile, in Europe, strikes among air traffic controllers, baggage handlers, and pilots grounded flights in France, Italy, Spain, and elsewhere (Travelpulse, 2023). In the United States, strikes by auto workers at Ford, General Motors, and Stellantis marked the first time that the Big Three vehicle manufacturers had been hit by simultaneous walkouts (Guardian, 2023b).
Some political theorists have defended strikes and have done so on various grounds. For example, they have defended strikes as instruments for resisting domination in the workplace, overcoming collective action problems, and exerting upward pressure on the terms of employment (Gourevitch, 2016, 2018; Locke, 1984). In addition to advancing arguments about the permissibility of striking, some theorists have devoted attention to aretaic considerations. They have argued that striking is not merely permissible but ‘a good thing, something to be celebrated and even expected of workers’ (Gourevitch, 2020, original emphasis). This, it has been claimed, is because striking involves valuable forms of ‘self-emancipation’ (Gourevitch, 2020: 112) or ‘self-management’ (Gourevitch, 2020: 119; cf. Medearis, 2020).
Theorists have also considered various objections to striking. They have considered the objection that strikers extort employers (Dobos, 2022), intimidate and coerce fellow workers (Gourevitch, 2016: 310), and unacceptably neglect the interests of third parties who rely on the services that strike disrupt, and who are not party to the dispute from which the strikes arise (Daniels, 1978; Dworkin ,1977; Locke, 1984: 174).
The present paper explores this final objection, which we can call the third-party objection to striking. The third-party objection comes in different versions. One version, which I will set aside, says that some strikes impose disproportionate burdens on third parties. (Call this the third-party (proportionality) objection.) Workers strike in order to pressure their employers into conceding to their demands, and, in the process, usually set back the interests of various third parties such as patients, students, and commuters. Suppose that a particular strike is undertaken in pursuit of an incontrovertibly just end. Suppose, for example, that the aim of the strike is to pressure an employer into providing workers with minimally decent working conditions. Suppose, further, that the burdens experienced by third parties as a result of the strike are regarded by strikers as regrettable side-effects of their actions, and not as an integral part of their plan. In this case, the third-party (proportionality) objection might be pressed. The objection would contend that the burdens endured by third parties are not proportional to the just ends that the strike is intended to achieve; rather, they are excessive.
According to a second version of the third-party objection, which I will also set aside, foreseeably inflicting burdens (even proportional burdens) upon one's clients is often contrary to the obligations one assumes by voluntarily occupying a professional position. (Call this the third-party (professional obligation) objection.) An advocate of this objection can acknowledge that one's professional obligations might include both short-term obligations to one's current clients and long-term obligations to maintain and advance professional standards. Striking might conflict with the former but might sometimes be justified on the grounds that it conduces to the latter. According to this line of argument, a strike that inflicts burdens on one's clients solely for self-interested reasons would be wrongful, but a similarly burdensome strike undertaken to protect the delivery of high-quality services could be permissible, or even required (Muyskens, 1982).
A final version of the third-party objection – the one that is the subject of this paper – arises if the burdens endured by third parties are part of what is planned. In this kind of case, the strikes that engender these burdens invite the charge that they are insufficiently discriminating. Rather than being aimed exclusively at any potentially ‘legitimate’ targets (e.g., employers who, by refusing to pay a fair wage or provide acceptable working conditions, might have made themselves liable to bear certain costs), the strikes are (also) aimed at individuals who do not seem to be liable. Most problematically of all, such strikes invite the charge that they harm the innocent opportunistically or exploitatively. (Call this the third-party (exploitation) objection.) In other words, those who strike face the charge that they are harmfully using the innocent as a means to further their own ends, thereby violating a central and enduring principle of deontological ethics, namely, the means principle.
Notice that the exploitation version of the third-party objection can apply even when the proportionality and professional obligation versions do not. 1 A lecturer might strike in order to guarantee the provision of high-quality education in the future (thereby complying with her professional obligations), and she might do this without subjecting current students to burdens that are excessive in relation to the gains associated with the achievement of that goal (thereby ensuring that the strike is proportionate), but she might nevertheless be guilty of harming her current students as means.
The potentially indiscriminate and opportunistic nature of striking is noted in the existing literature. Don Locke, for example, writes: ‘strikes are bad not only in their effect, which is the infliction of harm on both parties, but also in their intention, the intention being precisely that: the infliction of harm. Moreover, those who suffer this harm need not be restricted to the other party to the dispute; and … this may be part of the strikers’ explicit intention’ (Locke, 1984: 183). Similarly, Gerald Dworkin notes that ‘usually … the striker must cause, or threaten to cause, harm to third parties in order to force the employer to yield’ (Dworkin, 1977: 78–9). Dworkin describes this strategy as ‘calculated’ and ‘callous’ (Dworkin, 1977: 79).
In a similar vein, Muyskens writes: The refusal to work imposes inconvenience and possibly hardship on those in need of one's services. In the case of strikes by employees such as nurses, the detrimental effect of the strike on the public (those in need of nursing care) is often more immediate and more grave than on the employer. The public's inconvenience is the means by which pressure is put on the employer to come to a settlement agreeable to the striking employees. Were the public in no way inconvenienced the strike would likely be ineffectual (Muyskens, 1982: 105).
Though I cannot answer all of these questions here, I nevertheless hope to move the debate on striking forward by presenting a fuller statement of the exploitation objection than has yet been offered.
The paper unfolds as follows. In section III, I attempt to provide a plausible rendering of the claim that, in a variety of contexts, strikes constitute pro tanto violations of the means principle. The account I present draws on Warren Quinn's influential distinction between direct and indirect harmful agency (Quinn, 1989a). In section IV, I briefly consider a variety of possible replies to the exploitation objection. In section V, I address the claim that only doing harm can be opportunistic in the relevant sense, and that strikes, as activities that merely allow harm, therefore cannot be opportunistic in the relevant sense. I defend the claim that allowing harm can be opportunistic. Section VI considers whether striking can nevertheless be permissible in virtue of the fact that third parties are sometimes liable to treatment as a means. Section VII asks whether the presumption against doing and allowing harm as a means can sometimes be overturned and whether some instances of striking that might otherwise be impugned by the exploitation objection could then be acceptable. Section VIII concludes by sketching a rough schema for determining whether any given strike can escape the exploitation objection. Before turning to the main line of argument, a preliminary section clarifies certain aspects of the paper and further situates it in relation to the existing literature.
II
I begin with five preliminary points. Firstly, striking is a highly politicized topic. It is important to stress, then, that ethical misgivings about striking need not reflect a commitment to any particular political outlook. One can object to a specific instrument of worker power (the strike), while also believing that workers should be empowered through alternative means. Indeed, opposition to striking is compatible with support for various policies and arrangements that would enhance the bargaining power of labour relative to capital. Such opposition is compatible with support for the introduction of a universal basic income, for example. Conversely, just as one can oppose striking while supporting greater equality, one can support striking while pursuing thoroughly inegalitarian ends. As Philippe Van Parijs observed at the close of the previous century, ‘recent history is replete with examples of particularistic strikes to defend or increase the privileges of powerful categories of well-paid workers’ (Van Parijs, 1995: 45). We should not, then, allow our broader political commitments to sway us, unduly, one way or the other.
Secondly, the means principle specifies a restriction on promoting the good and a constraint on attempts to minimize violations of restrictions. 2 Critics could agree that strikes are, say, effective at preventing domination while also maintaining that they are an impermissible means with which to achieve that end. Critics could also agree that certain values – such as self-determination – are realized, to some degree, through striking, and that striking thus instantiates certain goods. However, if we endorse the means principle, we might also believe that the restrictions and constraints that it specifies reflect the inviolability – and corresponding high status – of human beings. This high status is also a good; not a good that can be promoted by any act, but one that is already there, and that is to be honoured by respecting the restrictions and constraints in which it is reflected. This would mean that defences of striking that ignore the relevant restrictions and constraints, and proceed on the assumption that human beings are freely available to be used in the production of certain valuable states of affairs, are insufficiently attentive to the value that already exists in the world. 3
Thirdly, the exploitation objection constitutes an ad hominem challenge to some defenders of striking. This is because some defenders appeal to something like the means principle when characterizing and critiquing the injustices that striking is supposed to enable victims to overcome. For example, Gourevitch opines that, under capitalism, wage labourers are ‘made use of’ and ‘made into a mere instrument’ (Gourevitch, 2020: 114). The exploitation objection suggests that, by engaging in certain types of strike, workers subject innocent third parties to the same kind of disrespectful treatment that Gourevitch says workers should strike in order to oppose.
To be clear, those who press the exploitation objection can acknowledge that some employers unjustly treat their workers as means. But, as noted above, the means principle specifies a constraint on attempts to minimize violations of restrictions. Deontological ethics tells us to respect the moral limits that apply to our own behaviour; it does not generally allow us to transgress those limits in order to prevent violations that would otherwise be committed by others.
Fourthly, I presume that initiating and participating in a strike are permissible only when doing so satisfies a number of principles familiar from the literatures on just war and self-defence. In addition to the proportionality and discrimination principles that I have already alluded to, these include just cause, reasonable chance of success, and necessity. In the cases that I address, I shall simply assume that these principles are satisfied. This will allow me to focus on the exploitation objection.
To repeat: I am going to assume that, in the cases under consideration, strikers have a just cause. Their cause remains just even if pursuing it through striking is objectionably opportunistic. The exploitation objection impugns the strategy employed not the end sought. If striking is the only effective way of achieving the just cause, then, in some cases, respecting the means principle might require individuals to abandon that cause and endure injustice. That would clearly be a difficult pill to swallow, but no more difficult than recognizing (as we do) that necessary self-defence against unjust aggression is sometimes impermissible (e.g., because it would be disproportionate, or because success could be guaranteed only through the use of indiscriminate strategies). (Note that if workers can effectively resist unjust treatment only by inflicting opportunistic harm, that fact might be thought to expose an additional unjust feature of their situation, namely, that they have been denied access to morally permissible means with which to protect their rights.)
Finally, following Kant's initial formulation, it is sometimes suggested that the means principle distinguishes between treating someone as a means and treating someone merely as a means, and that it condemns only the latter (Cohen, 1995: 239–240). A proponent of the mere means interpretation might insist that while strikes might treat some people as means, they do not treat anyone as a mere means, and therefore do not violate the means principle. Since the mere means interpretation is vulnerable to powerful objections that have been well expressed in the existing literature, I shall set it aside (Parfit, 2011: 212–232; Cf. Kramer, 2011: 39–42).
III
We can begin our formulation of the exploitation objection by introducing Warren Quinn's well-known distinction between direct and indirect harmful agency. This distinguishes between agency in which harm comes to some victims, at least in part, from the agent's deliberately involving them in something in order to further his purpose precisely by way of their being so involved … and harmful agency in which either nothing is in that way intended for the victims or what is so intended does not contribute to their harm (Quinn, 1989a: 343). What seems specifically amiss in relations of direct harmful agency is the particular way in which victims enter into an agent's strategic thinking… The agent of direct harm … proposes to involve them in some circumstance that will be useful to him precisely because it involves them. He sees them as material to be strategically shaped or framed by his agency. Someone who harms by direct agency must therefore take up a distinctive attitude towards his victims. He must treat them as if they were there and then for his purposes (Quinn, 1989a: 348).
The special presumption against direct harmful agency roughly approximates the Kantian injunction to refrain from using others as means (Quinn, 1989a: 350). As Tadros notes, to use someone as a means ‘is to incorporate them into one's plans and projects, or to manipulate them to serve one's ends’ (Tadros, 2011: 140). It is for this reason that the presumption against direct harmful agency can be referred to as the means principle. 4
In the remainder of this section, I take a first stab at rendering plausible the claim that some types of strike involve direct harmful agency, and, more specifically, that they involve harm that is opportunistic (or exploitative) in nature. (When I say that p ‘harms’ q, I shall mean either that p does harm to q or that p allows harm to befall q. I return to this distinction in section V.)
Consider the following series of cases:
Bakery 1
A baker's assistant, Andy, strikes for better pay from his boss, Bertie. Without Andy's assistance, Bertie is unable to bake sufficient produce to satisfy customer demand. Due to the shortfall in supply, a customer, Charlie, cannot buy from Bertie's bakery his weekly loaf of pumpernickel. Instead, he buys his bread from Beatrice's Bakery, a rival business located next door to Bertie's. The quality of Beatrice's pumpernickel is equal to that of Bertie's, it costs the same amount, and Charlie enjoys it just as much. Charlie is the only affected customer.
Bakery 2
The same as Bakery 1, except that there is no alternative bakery from which Charlie can purchase his pumpernickel.
In this case, Andy's strike has adverse consequences for Charlie as well as for Bertie. Bertie suffers reduced profit, and Charlie suffers from an absence of pumpernickel. Although Andy's strike has these adverse consequences for Charlie, Andy might plausibly claim that they are a product of indirect rather than direct agency. Andy could maintain that his aim is to adversely affect Bertie by reducing her bakery's productive capacity and thereby reducing her profit. Given the absence of an alternative, Andy can predict that a consequence of the bakery's reduced productive capacity will be that Charlie goes without pumpernickel. But it need not be to Andy's purpose that Charlie goes without pumpernickel. Suppose that, having learnt of his plight, a friend takes it upon himself to bake Charlie some pumpernickel, and that the loaf he provides, free of charge, is as good or better than that which Charlie would otherwise have purchased from Bertie. This need not frustrate Andy's plans. For Andy, all that might matter is that Bertie's profit shrinks, and that this puts pressure on Bertie to increase his wages, not that Charlie goes without pumpernickel.
Bakery 3
Andy goes on strike, thereby reducing Bertie's productive capacity. Bertie is unable to furnish Charlie with the pumpernickel that he has ordered and paid for in advance. No one compensates Charlie for his loss. Frustrated, Charlie writes a negative review of Bertie's bakery, the appearance of which puts pressure on Bertie to increase Andy's pay. Andy goes on strike precisely in order to provoke such a review.
In this case, Andy seems to harm Charlie and to do so opportunistically. It is to Andy's purpose that Charlie is frustrated. If Charlie were adequately compensated and refrained from writing a negative review, Andy's plan would be undermined.
Bakery 4
The same as Bakery 3, except Charlie does not write a review. Nevertheless, there is a danger that people will learn of Charlie's negative experience, that there will be further such experiences in the future if the strike is continued or repeated, and that the bakery will earn a bad reputation. Anxiety about how such a reputation would affect her business puts pressure on Bertie to increase Andy's pay. Andy goes on strike precisely in order to create this anxiety.
Again, Andy seems to harm Charlie, and the harm seems to be opportunistic in nature. In order for Andy's plan to succeed, Charlie must have a bad experience that others can potentially learn about and be influenced by.
Bakery 5
The same as Bakery 3, but rather than writing a review, a bereft Charlie – along with other occasional customers who, while not impacted by the present strike, are able to put themselves in Charlie's shoes – realises how much he values Andy's services and so decides to encourage Bertie to increase Andy's pay. 5 Andy goes on strike precisely in order to force upon Charlie, and others, recognition of his value.
Here, too, the harm that Andy seems to inflict upon Charlie appears to be opportunistic in character. Harming Charlie is to Andy's purpose. If Charlie were not harmed, Andy's plan to communicate his value to Charlie and others could not succeed.
These artificially simple cases foreground the particular features that we are interested in – namely, manifestations of opportunistic harm – and reveal how those features can animate the activity of striking. This makes it easier to see how the relevant features can also be among those possessed by the more complex real-world strikes with which we are familiar. Bakeries 3 and 4, for example, seem to reflect what happens in many academic strikes. Lecturers know that their strikes will set back the interests of students, and it is often to lecturers’ purposes that their strikes have this effect. If the strikes did not frustrate student interests, students would not complain, the reputation of the university would not be threatened, and the mechanism through which lecturers often seek to exert pressure on their employers would not operate. The strikes can be expected to work as planned (i.e., according to the Bakery 3/4 model) only if they set back student interests. 6
Bakeries 3 and 4 also seem to reflect what happens in some medical strikes. Medical professionals in public health services know that their strikes will cause disruption that will upset and frustrate patients, that such frustration will often be directed at the government, and that ministers will be pressured into ending the disruption by conceding to the demands of the strikers.
Some strikes, or the motivations of some strikers, may have more in common with Bakery 5. Believing that their efforts are inadequately appreciated, some workers may withdraw their services in an attempt to make vivid their contributions. Once again, the adverse consequences for those affected seem not to be merely incidental but to be central to the strikers’ plan. If, say, striking nurses were replaced with equally competent substitutes, and patients continued to be cared for as usual, their plan (as conceived along the lines of the Bakery 5 model) could not succeed. It is only when patients see appointments cancelled, procedures delayed, waiting times extended, and so forth, that they, and observers in the wider population, might be forced to revise their valuation of the nursing profession. Once again, the hardship experienced is instrumental to the achievement of the strikers’ goals. The striking nurses exploit the suffering endured by their patients, using it as a means to advertise their worth.
This has been a provisional attempt to make sense of the claim that some strikes involve opportunistic harm. Let me now bring out some complexities and identify some challenges.
IV
In this section, I identify and address some of the main challenges that the exploitation objection is likely to encounter. My aim is not to provide decisive rebuttals – doing so would involve, inter alia, unmanageable excursions into the metaphysics of harm – but to suggest some reasons for doubting that these challenges are fatal.
It might be objected that, in the cases described in the previous section, benefits are withheld, but no harm is suffered. One who presses this objection might contend that harm is a matter of making someone (or allowing someone to become) worse off than they were, whereas withholding benefits is a matter of refraining from making someone better off than they are. The objector might add that the means principle cannot be violated merely by refraining from making someone better off than they are.
Can there not be instances of withholding benefits that leave some people worse off than they were? If there can, we could distinguish between instances of withholding benefits that involve harm and instances of what we might call merely withholding benefits, which do not involve harm. Some instances of striking seem to fall into the former category. (Whether these instances of striking involve doing harm or allowing harm, and whether allowing harm can violate the means principle, are questions to which I return in the next section.) Indeed, Bakery 3 – and the subsequent cases that build on Bakery 3 – seem to be instances of withholding benefits that involve harm. Charlie is denied the benefit of pumpernickel that he has already paid for, and he is not compensated. Charlie is therefore worse off than he was. Moreover, this feature is mirrored in real-world academic strikes where students fail to receive education for which they have already paid. These students no longer have their money, but nor do they have that for which they gave their money away in exchange. In that sense, they are worse off than they were. In medical cases, where patients endure some injury or illness that, due to a strike, deteriorates, it similarly seems that we are dealing with instances of withholding benefits that involve harm. Patients are denied the benefit of medical care, their injury or illness deteriorates, and they become worse off than they were. Harm can also be involved when appointments are made and then, because of a strike, cancelled. Harm is involved in these cases when the process of scheduling, cancelling, and rescheduling subjects the patient to an anxiety-inducing cycle of anticipation and frustration.
Some other strikes seem to be instances of withholding benefits that involve harm by virtue of the effect that they have on people's plans. For example, when rail workers strike, they withhold a beneficial transportation option that they could have offered, but, in doing so, they might also scupper the plans that people have made. Put differently, they might move individuals from a reality in which they had viable plans to a reality in which they have unviable plans, thereby making them worse off than they previously were.
Now, striking workers might try to justify their use of these strategies by insisting that they are used to pursue just ends. They might claim that they are overworked and underpaid, and that any expectation of continued and uninterrupted service under such conditions is unreasonable, especially when interruption can potentially mitigate the unjust treatment that they suffer. They might add that although students might have a pro tanto right to receive the education that they have paid for, and patients a pro tanto right to be cared for, these rights must compete with the opposing rights of lecturers and nurses to try to secure fair and equitable terms of employment.
It is not obvious that this argument can succeed in showing that the special presumption against specifically opportunistic harming is overturned. As Quinn notes, when a particular act takes an opportunistic form, the fact that it takes that form can strengthen one's rights not to be subjected to it, ‘perhaps giving [those rights] a power to stand against moral forces to which they would otherwise give way’ (Quinn, 1989a: 346). If the harms suffered by students, say, arrived via indirect agency, the right of lecturers to challenge unjust treatment might win out in competition with the opposing right of students to receive the education that they have paid for. But when the harms in question have a direct, opportunistic, character, the latter right is bolstered and will be less easily defeated.
Strikers might also claim that they are striking for the good of their clients. Striking doctors, for example, might say that meeting their demands is necessary to ensure that patients receive adequate treatment. Perhaps some invocations of this argument are made in bad faith, amounting to little more than a veiled threat to deliver suboptimal services unless one gets the pay or other perks that one desires. 7 But even good faith invocations of the argument can fail to defuse the concern about opportunistic harm. One reason for this is that the individuals harmed by a strike will not necessarily be the same individuals that benefit from the strike (if any do) (Muyskens, 1982: 107). For example, future students might see gains in their education that are bought at the expense of present cohorts who see their timetables cleared by wave after wave of lecturer strikes. In that case, present cohorts suffer opportunistic harm for the benefit of those who come later.
A third line of defence appeals to the fact that strikes have a three-party-structure: they involve employers, workers, and those reliant on the services that the strike disrupts. This line of defence argues that strikers do not make use of individuals who would otherwise be uninvolved; they do not, to return to Quinn's formulation, ‘involve them in some circumstance that will be useful’. Rather, the strikers’ employer has already involved these individuals in potentially harmful circumstances. The employer agrees to provide a service to a customer or client while mistreating the workers whom she relies upon to deliver that service. It is therefore the employer who renders the client vulnerable to suffering harm. The employer both promises that the service will be provided and is responsible for the unjust treatment that prompts workers to withdraw their labour, thereby ensuring that the service is not delivered. Thus, any moral complaints from or on behalf of third parties should be levelled exclusively at the employer.
Suppose it is true that the employer, not the worker, has promised a particular service, and that the employer has treated her workers unjustly. This does not seem sufficient to absolve the strikers of wrongdoing in cases akin to Bakery 3–5. Through her unjust behaviour, the employer has plausibly made the client vulnerable to not receiving the service that he has been promised, but, in the types of case under scrutiny, it is the strikers who choose to treat the client in a specifically opportunistic manner. To use Quinn's terms, the strikers involve the client in the sense that they strategically frame his participation in a particular way: they treat his vulnerability to harm as an opportunity to be exploited for their ends.
Some further comments might help us to think about the exploitation objection more clearly. Note that strikes are likely to produce several different consequences, each of which conduces to their success. To illustrate, Andy's strike might further his ends by reducing Bertie's productive capacity and by prompting Charlie to write a negative review. Suppose Andy foresees that his strike will produce both consequences and that each will further his ends. The morally best-case scenario here is that, upon examining his motives, Andy finds himself able to truthfully declare that he is striking only in order to bring about the first consequence and not the second. In order to reach this conclusion, Andy would have to determine the circumstances in which the strike can reasonably be expected to succeed. If Andy's investigations reveal that the strike can reasonably be expected to succeed regardless of whether Charlie writes a negative review, it might be plausible for Andy to say that he is striking only in order to bring about the first consequence. If so, then Andy might be able to escape the charge that he is harming Charlie opportunistically. But suppose Andy instead discovers that Charlie writing a negative review is crucial to the strike's success. In this case, Andy could continue to maintain that he is striking solely in order to bring about the first consequence only if he conceded that his intention in striking is something other than successfully accomplishing its stated aim.
Here is a related point. Andy's strike might further his ends to some degree by reducing Bertie's productive capacity and to a further degree by prompting Charlie to write a negative review. If Andy is genuinely intending to exploit only the first of these two mechanisms, then, to the extent that it is possible to do so, he should relinquish any gains he receives via the second mechanism. If he instead chooses to hold on to those gains, it seems he cannot, in good faith, claim not to have harmed Charlie as a means.
V
One might agree that there is a special presumption against opportunistic harming but think that only doing harm can be opportunistic in the relevant sense. One might further think that, since striking is a largely passive activity where one simply refrains from providing some service, striking merely allows harm, and therefore is not something that can be opportunistic in the relevant sense.
In this section, I will consider two replies to this claim. The first (which I shall not rely on) points to how seemingly passive forms of behaviour can interact with active forms of behaviour such that the resultant harm is attributable to active behaviour. The second (which I shall defend at some length) seeks to establish that merely allowing harm can be objectionably opportunistic in the relevant sense.
In order to illustrate the first reply, consider
Interview
On Wednesday, Jessica asks Jane if she will drive her to a job interview the following Tuesday. Jane agrees, but when Tuesday rolls around, she fails to show up. Jessica misses her interview.
Might some strikes fit this mould? Consider, for example, a student who misses multiple classes as a result of his lecturer's strike. Although refraining from teaching is passive, it might be said that the lecturer is failing to provide something that she has previously actively committed to provide. Although the relationship between the student and lecturer has a degree of mediation that was lacking in Interview, and although the student/lecturer scenario is extended over a longer period of time, the relationship between the parties might seem morally similar. The lecturer promises the university that she will teach, and the university promises the student that he will be taught. In light of this agreement, the student declines opportunities to be taught at other universities, and thereby renders himself dependent on this particular lecturer (and on others at the same institution). We might therefore be tempted to characterize events in terms of something being done to the student by the lecturer.
A lot seems to hinge, however, on what exactly the lecturer promised. What was the precise nature of the contract? Did the lecturer commit to teaching no matter what or only under certain conditions? Moreover, there might be a mismatch between the contents of the university/lecturer contract, on the one hand, and the university/student contract, on the other.
In light of these complications, I shall not rely on the first reply. However, it should be noted that there do seem to be cases where the reply is applicable. Earlier, I drew attention to the potentially anxiety-inducing cycle of anticipation and frustration to which patients are subjected when appointments are scheduled, cancelled, and then rescheduled. Consider doctors and nurses who schedule (or agree to participate in) a procedure and then take part in a strike that causes the procedure to be cancelled. Any anxiety suffered by patients is attributable to something that has been done to them by the doctors and nurses in question.
Admittedly, the details of cases like this are perhaps not always as I have described them. Some appointments might be scheduled by administrators rather than by the medical professionals who perform them, and without the latter's explicit consent. Nevertheless, it would be surprising to learn that no real-world cases match the description given in the previous paragraph. The first reply, then, is likely to have force in some real-world cases.
The second reply, to which we should now turn, suggests that even allowing harm can be objectionably opportunistic. Quinn attempts to illustrate this with a case involving a doctor who deliberately leaves his patients untreated in order to learn more about the disease with which they are afflicted (Quinn, 1989a: 336, 345–6). The doctor merely allows the patients to suffer, but since this is done with the intention of learning more about the disease, his behaviour counts as objectionably opportunistic.
One shortcoming of Quinn's case is that the doctor might have an obligation to treat his patients, an obligation acquired by choosing to take up his post and thereby agreeing to treat his patients. If this is true, then our analysis of the current case should mirror our analysis of Interview: the harm to patients, like the harm to Jessica, should be attributed to a doing, because the doctor has invited others to rely on his services and then acted in a way that is contrary to their legitimate expectations. But what we are asking here is whether the means principle can plausibly be interpreted as condemning pure allowings. Put differently, we are trying to determine whether pure allowings can be opportunistic (in the relevant sense). Let us consider, then, a similar case with a protagonist who lacks a professional obligation to assist. (As we shall see, this case, and the one that follows it, will also turn out to be unsatisfactory for our purposes. However, identifying the shortcomings of these cases will help to point us in the right direction.)
Disease
A patient is infected with a rapidly progressing disease. The hospital to which she has been admitted is in an emergency situation and no doctors are available to treat her. Irene, a bystander, can easily provide treatment but refuses to do so in order that doctors can learn more about the disease once they eventually arrive on the scene.
We might attempt to circumvent this difficulty by revising the case further. Consider
Disease 2
The same as Disease, except personal protective equipment is currently unavailable. Since treating the patient without such equipment would put Irene in considerable danger, Irene lacks a natural duty to provide treatment. Irene is a heroic individual who would usually be willing to help the patient despite the risk to her own health. Irene declines to help so that doctors can learn more about the disease.
However, it seems important that, in a very nearby possible world in which protective equipment is available, Irene (or someone else occupying a similar position) does have a duty to treat the patient. To put the point slightly differently, the patient lacks a claim to treatment that she possesses in a very nearby possible world, and she does so despite not having waived or forfeited that claim.
To make the point vivid, suppose that Irene initially has access to a mask, and therefore has a duty to provide treatment. She puts the mask on (having not yet decided whether or not to provide treatment), and, as she does so, the strap breaks, meaning that she can no longer secure the mask to her face. A duty to provide treatment existed a second earlier but is now suspended.
There is something tragic about this situation. We cannot guarantee satisfaction of the demands for assistance arising from the patient's moral status without trampling over the demands for (some degree of) independence arising from Irene's moral status. It follows that if Irene were to offer (supererogatory) treatment, this would not merely be morally good (in the way that generously buying my colleague a cup of coffee each morning might be morally good), but morally reconciliatory, in the sense that it would allow for the simultaneous satisfaction of the competing pair of demands. In short, we might think that the suspension of a normally present duty leaves some kind of moral residue. It ensures that, although providing treatment is supererogatory, Irene has more reason to treat than she would if there were no suspended duty in the background. We might think that Irene's disregarding of that reason could shape our intuitions and characterizations of the case. If so, then the case is not pure enough to focus attention on the question of opportunistically allowing harm.
Let us consider a further set of cases where there is no preexisting duty to act suspended in the background, and thus no question of a moral residue affecting how we characterize the situation. Consider
Streets
On her way to work, Laura passes several homeless people on the streets. Laura knows that a cold weather front is moving in, and that if these people are not assisted, their condition will worsen throughout the day. Laura, I presume, does not have a duty to assist all (or perhaps any) of these individuals.
8
Nevertheless, Laura is a kind-hearted person who would usually be inclined to help in any way that she can. On this occasion, however, Laura declines to help. She declines to help because she knows that, later in the day, her nemesis, Lena, will be in the area viewing properties, and suspects that, since Lena has a severe aversion to the sight of suffering, the presence of homeless people shivering in the cold will deter Lena from buying a house in the neighbourhood.
Streets 2
Same as Streets, except that Laura is not a kind-hearted person and, even without the prospect of Lena buying a house in the area, would regard herself as having sufficient reason not to help. She regards the fact that not helping might deter Lena from purchasing a property as an additional sufficient reason not to help.
And
Streets 3
Same as Streets, except that, without the prospect of Lena buying a property, Laura would be conflicted about whether or not to help. The prospect of Lena buying a property settles the issue.
In each of these cases, I think that Laura acts in a manner that is objectionable because opportunistic. This is despite the fact that Laura acts passively, not in a manner that contravenes any preexisting duty to act, and not in a manner that is contrary to any suspended background duty.
Perhaps some will say that, contrary to what I have claimed, Laura does have a duty to assist the homeless people whom she encounters, and that this fully explains the intuition that she acts wrongfully. But anyone who presses this claim will have to recognize some limit to Laura's duty of assistance. How many homeless individuals is Laura duty-bound to help? Ten? Twenty? Fifty? However this question is answered, the case can easily be modified to suit. Suppose Laura can be duty-bound to assist up to one hundred homeless individuals. We can then specify that there are, say, one hundred and ten such individuals on the street that day, and that while Laura helps the first hundred, she leaves the final ten as a deterrent for Lena. If this really is Laura's motivation for leaving the final ten on the street – and she does not instead cease providing assistance due to exhaustion or the sense that she has done enough – then presumably we will still feel that Laura has acted wrongfully. We cannot explain that intuition by claiming that she has not discharged her duty of assistance.
Perhaps a sceptic will point to the fact that the individuals in Streets are homeless; that they are among the least advantaged people in society. Maybe it will be said that this is what is driving our intuition. Perhaps what is wrong about Laura's conduct is not simply that it is opportunistic, but rather that it takes advantage of those at the very bottom of the social ladder.
But the case can easily be revised to meet this concern, too. We can specify that the people in question are only temporarily homeless. Let us suppose that they are tourists whose coach has been destroyed, and who will be stranded until the end of the day. Suppose that Lena is known to be so sensitive that the sight of even this temporary suffering can be expected to create an association in her mind that will put her off the idea of living in the area. If Laura decides to leave some of these people exposed to the elements, and decides to do this not for the unremarkable reason that it is not her responsibility to help them but rather in order to deter Lena from buying a house, we will surely still want to condemn Laura's behaviour.
Laura's passivity – like Irene's – is ‘strategic and deliberate’, 9 and, further, motivated by the recognition that the hardship to be endured by those she does not help will serve her own purposes. It is this that seems crucial. Since striking is also strategic and deliberate, and motivated by the sense that the harm involved serves the purposes of strikers, we cannot appeal to the fact that strikes are passive (when they are passive) or that they do not contravene any preexisting or suspended duty of service in order to reject the claim that strikes are objectionably opportunistic.
To be sure, we might suspect that Laura's behaviour also wrongs Lena, and this might partially account for our sense that Laura acts impermissibly. However, that sense would be unlikely to disappear if we simply stipulated that Lena had made herself liable to be treated in the way that Laura treats her. It might also be pointed out that the homeless people in Streets are allowed to suffer a physical harm, whereas third parties affected by strikes need not suffer physical harm (although those affected by medical strikes often do). But this does not seem crucial. We could revise the case such that there is no cold weather front moving in, and such that Laura declines to help because she suspects that this will lead to the homeless individuals in question becoming visibly more depressed, and that this will deter Lena from buying a property. Laura's behaviour still looks objectionably opportunistic.
VI
Throughout the discussion, I have assumed that the third parties adversely affected by strikes (patients, students, commuters, etc.) are innocent bystanders. This might be contested. It might be argued that, in at least some cases, some of these individuals bear some degree of responsibility for the injustices that strikes are intended to rectify. Moreover, it might be claimed that this responsibility renders them liable to the opportunistic harming that strikes involve.
Quinn recognized that individuals might act in ways that render them liable to opportunistic harming (Quinn, 1989a: 334–345), and this insight has since been developed in the literatures on war and punishment. 10 To illustrate, suppose that the economic sanctions imposed on Russia by the West aim, among other things, to make life difficult for Russian civilians, and thereby to compel them to oppose Vladimir Putin's war against Ukraine. To the extent that Russian civilians encouraged or supported that war, we might think that they thereby made themselves liable to this form of opportunistic harming. 11
Suppose that some rail workers earn meagre wages and that they strike for better pay. Suppose, further, that part of the rail workers’ plan is to disrupt the lives of commuters, since they believe that doing so will serve their purposes. If the line of argument currently under scrutiny can be applied to this case, it must be that at least some of the affected commuters have somehow made themselves liable for the opportunistic harms that the strike involves.
How might they have done this? Perhaps it could be said that commuters pay too little for the services that they use, and that, because of this, the rail company cannot afford to pay its workers a higher wage. The company knows that if it raises fares, some commuters will reduce the number of journeys they make by train, or stop travelling by train altogether, and that the revenue lost as a result will offset the gains reaped by higher ticket prices. By refusing to pay a higher price for the service, commuters make it impossible to pay rail workers higher wages. This makes them liable for the opportunistic harm involved in the strike.
What should we make of this claim? Presumably, consumers have no duty to purchase any particular good or service at any particular price. If the price of a service is such that, for any given consumer, its purchase represents an unacceptably large opportunity cost, that consumer has the prerogative to forego purchase. This seems obvious. I contravene no duty if, seeing the price of a ticket, I choose to ride my bike rather than the train. (In fact, the consumer has the prerogative to forego purchase regardless of whether it represents an unacceptably large opportunity cost. I am not duty-bound to use and pay for a service simply because doing so would be in my interest. This also seems obvious.) Perhaps, if many people do as I do, some rail workers will see their wages shrink, or even be laid off. But I have not wronged these people simply by declining to use a service that I do not think (or maybe do think) is worth its present cost. (Unlike the strikers who withdraw their labour, I do not withdraw my custom because of the harm that this will bring to certain individuals.)
What if the rail company lowers fares and I start (or resume) using the service? Am I now wronging the rail workers such that I render myself liable to opportunistic harm? If my contribution to setting the price for labour does not make me liable when that contribution takes the form of not paying, it is not clear why it should make me liable when it takes the form of paying.
Now, there are certainly exceptions to what has been said so far. Although it is true that consumers have the prerogative to forego purchase of any particular good or service, morality does impose limits on their market discretion. Suppose that racist commuters prefer to use services with white conductors. The rail company knows this and responds by offering lower wages to black conductors. (Suppose that the law allows the company to do this, but not to exclude black conductors altogether.) In this case, the low wages received by black conductors would be not merely low but unjustly low (because attributable to racial prejudice), and, in virtue of their responsibility for creating the injustice, racist commuters would plausibly be liable to bear some harms engendered by efforts to correct the injustice, including the opportunistic harms involved in some strikes.
In real-world cases, prejudiced preferences that contribute to setting unjustly low prices for labour may often be difficult to detect. Moreover, such preferences may sometimes be reflective not of overt animus, but rather of largely subconscious beliefs about the relative competence of different groups. Although difficult to detect, the existence of such preferences does seem capable of making their owners liable to some degree of opportunistic harm.
Perhaps consumers can be implicated in wrongdoing not only when they unjustly contribute to setting a low price for labour, but also when they merely benefit from a low price that is explained by injustice elsewhere. Suppose rail workers earn what they earn not because commuters are unwilling to pay higher prices, but because, in virtue of their comparative lack of marketable skills, they enjoy little bargaining power with which to negotiate their terms of employment. In virtue of their emergence from such background conditions, we might deem rail workers’ wages to be unjust (Parr, 2025). If commuters, and not only employers, benefit from these unjustly low wages (in the form of cheaper fares), then commuters are beneficiaries of injustice. Furthermore, by availing themselves of those gains, we might think that commuters render themselves liable to bear at least some of the opportunistic harms involved in strikes that aim to rectify the injustice. If commuters benefit innocently from the injustice, then we should perhaps not say that they are liable to bear harms, since we typically think of liability as arising from wrongdoing. Nevertheless, it might be said that these commuters could have no reasonable complaint if a strike deprived them of gains that were made possible by injustice and that they therefore had no right to retain.
The claim is that if consumers have contributed to or benefitted from injustices that some strikes aim to rectify, those strikes are not necessarily impugned by the fact that they opportunistically allow harms to befall consumers. (The moral picture is more complicated if some consumers have contributed to or benefitted from injustice whereas others have not.) I have provided a rough sketch of an argument for this claim. The limitations that this argument would exhibit even if fully developed should be fairly apparent. One obvious limitation is that injustices suffered by workers do not always translate into benefits for consumers. To illustrate, suppose that lecturers earn less than they would under different, more just, conditions. (I am not claiming that this is true but simply assuming it to be true for the sake of the present argument.) Would the higher earnings that would be secured under just conditions be funded by higher fees for students or by measures that cut into the earnings of vice-chancellors and other members of senior management? I imagine that many advocates of recent academic strikes would claim the latter (which is of course not to say that this is how higher earnings for lecturers would actually be funded, if achieved, in less than fully just circumstances). If this is right, then we cannot say that the current fee structure benefits students relative to a just counterfactual.
The same point undermines the claim that even if students do not benefit from the alleged injustice suffered by lecturers, they might nevertheless contribute to the injustice in certain ways, e.g., by protesting against higher fees or by publicly declaring that they would not pursue a degree if fees were raised. If the just way of funding higher pay for lecturers (if higher pay is, indeed, what justice requires) is not via higher fees for students, then this claim is unconvincing.
Another limitation is that although some victims of strikes may be implicated in the injustices that the strikes aim to rectify, the degree of their responsibility may be insufficient to render them liable to the particular harms involved in the strike. If, for example, patients using public health services benefit from the low wages paid to nurses – not qua patients but qua taxpayers – it seems highly unlikely that any wrongdoing involved could be serious enough to make those patients liable to bear all of the opportunistic harms that have been involved in strikes (the repeated cancelling and rescheduling of cancer operations, the cancelling of chemotherapy sessions, and so forth). 12 This conviction is only strengthened if we think that, as a matter of justice, any tax increases needed to fund higher wages for nurses should be borne by the wealthiest individuals in society, many of whom have already opted out of public health care.
If individuals have benefitted innocently from injustice – and are therefore not liable to bear any harms but may nevertheless be required to relinquish their unjust gains – the related problem is that there can be a lack of equivalence between the gains unjustly acquired and the harms involved in a strike. An individual would not seem to have a reasonable complaint if, say, a nursing strike were to deprive them of income that, under just conditions, they would have lost through taxes levied to fund higher wages for nurses. Such a strike would deprive individuals of resources that they would not have possessed under just conditions and would do so precisely in order to bring about the just conditions whose absence explains why those individuals currently do possess the resources in question. The same argument cannot be made if the losses exacted are not ‘in kind’ – for example, when they take the form of cancelled operations.
VII
Before concluding, it should be noted that the presumption against opportunistically doing or allowing harm can sometimes be overturned. Threshold deontologists maintain that although we are generally forbidden from treating others in certain ways, the relevant modes of treatment can be permissible when necessary to achieve an amount of good large enough to cross a threshold set at some point on the scale by which goodness is measured. One interpretation of this view says that deontological restrictions are overridden when the threshold is reached. A different interpretation says that restrictions simply cease to operate when the threshold is reached (i.e., because in such circumstances individuals lack the rights in which restrictions are grounded). 13 For expositional ease, I shall write as though the former version is the correct one, but nothing substantial will hinge on this being so.
In the original formulations of this position, philosophers sometimes wrote as though there were a single threshold applying to all pro tanto violations of deontological restrictions, and as though that threshold was very high. Indeed, it was once suggested that the threshold was set at the level of good that could be achieved only by averting ‘catastrophic moral horror’ (Nozick, 1974: 30n), such as nuclear war or the destruction of a large city. If we understood threshold deontology in this way, none of the opportunistically harmful strikes with which we are familiar in Western liberal democracies could realize enough good to overcome the restriction against using as a means.
However, if we think that violations of some deontological restrictions are less serious than others, that immediately puts pressure on the idea that there is a single threshold. In fact, if striking involves treatment as a means, that puts pressure on this idea. For, it is implausible to think that the amount of good needed to overturn the restriction against using people as a means by striking is as large as the amount of good needed to overturn the restriction against using people as a means by, say, killing or torturing.
Recognition of the fact that pro tanto violations of deontological restrictions can be more or less serious – depending, inter alia, on the nature of the restriction and the manner of its violation – has prompted the suggestion that thresholds are set at different levels depending on which restriction and what kind of violation are at issue. 14 This is sometimes called ‘sliding scale’ threshold deontology (Moore, 2019: 386). This is perhaps not an ideal name as it implies that there is still only one threshold, and that this single threshold moves up or down depending on the violation. This misrepresents the underlying idea (at least as I conceive of it), which is that there are multiple fixed thresholds attaching to the wide variety of different possible violations. A better name, then, might be multiple threshold deontology.
It should be reasonably clear that the multiple threshold view could potentially accommodate some instances of striking. Since the violations associated with some strikes will be relatively minor, the corresponding threshold should be relatively low, and more easily reachable by the kinds of goods that strikes with just ends aim to produce. Of course, much will hinge on what services are being withdrawn and what goods are being pursued. Some workers (nurses, doctors) will almost inevitably find their strikes adversely affecting services the disruption of which will do or allow considerable harm. The multiple threshold view could vindicate such strikes only if the good sought was exceptionally important (e.g., the preservation of such services in the long run).
Other workers will be able to discriminate between more and less important services. For example, lecturers could pursue strikes that deprive students of one or two lectures but eschew marking boycotts that prevent finalists from graduating. But workers in this category will still have to ensure that the good sought is sufficiently important to justify the degree of treatment as a means that the strike involves. Poorer workers enduring degrading conditions are much more likely to be able to target goods of the requisite magnitude. Delivery drivers forced to urinate in bottles due to a lack of bathroom breaks clearly have the right kind of objective (Guardian, 2021). On the other hand, it will be harder for barristers, doctors, and academics to identify appropriate goals.
VIII
We began with the claim that some strikes involve the opportunistic harming of innocent third parties. We should now have a better sense of how this claim might be assessed in any particular case, and of what should follow from it. The analysis offered suggests the following procedure.
First, we must determine whether the harms involved are indeed opportunistic in nature. Is the doing or allowing of these harms a regrettable side-effect of the strike or is it an integral part of the plan? If the latter, we can move to the next step.
Second, how severe is the opportunistic harm in question? In other words, to what extent does the strike treat third parties as means? Here, one must examine the nature of the sacrifice being exacted. What good of the third parties do strikers sacrifice in order to attain their ends? At the lower end of the scale are strikes that merely inconvenience their victims. At the higher end are those that jeopardise their victims’ health, well-being, and careers. Put another way, we can ask whether a strike leaves any lasting adverse mark on its victims’ capacity to realize their conception of the good. Strikes that do leave such a mark treat their victims as means to a greater extent than those that do not.
Third, have the strike's victims done anything to make themselves liable to the harms that will be suffered? Do they bear any responsibility for the injustice that the strike is intended to address? If so, how much responsibility do they bear? If any of the strike's victims lack responsibility for this injustice (or lack sufficient responsibility to be liable to the particular harms involved), we can move to the next step.
Fourth, how significant is the good that the strike can secure? Put differently, how severe is the injustice that the strike can be expected to remedy? I have not suggested a way of measuring the significance of the good to be achieved, but some questions that might be asked in this connection are: Will the strike mitigate an inequality suffered by those who are nevertheless reasonably well-off? Will it ameliorate the disadvantage of those close to the sufficiency floor identified by sufficientarians? Will it address a threat to individuals’ basic dignity or fix a deficit in their fundamental liberties?
Finally, it is necessary to compare the severity of the opportunistic harm that will be suffered to the severity of the injustice that the strike can redress. If the opportunistic harm is minor, and the injustice considerable, the strike is likely to be permissible, all-things-considered. In such cases, the presumption against opportunistically doing or allowing harm can be overturned. However, as we vary both scales – the seriousness of the opportunistic harm and the severity of the injustice – the outcome of the calculation will change. Graver opportunistic harms ground firmer restrictions, more resistant to being overturned. When strikes encounter these restrictions, it is quite likely to be the former that must give way.
It is worth noting, if only in passing, that these criteria might also be relevant to the evaluation of other practices, in addition to labour strikes (e.g., capital strikes and consumer boycotts). I should stress that I offer this suggestion somewhat tentatively. An entire paper was necessary to identify the implications for one practice of the presumption against opportunistic harming. Further research would be needed to spell out the implications for different practices.
If fully vindicated, the exploitation objection would reveal neglected moral restrictions on the conduct of workers. It would also suggest that governments should furnish workers with alternative means with which to protect their legitimate interests. Less obviously, it might also have implications for how we evaluate the conduct of employers, a topic that the present paper has largely set aside. If employers treat their workers unjustly and place them in a situation where they must sometimes choose between enduring the injustice or resisting it by perpetrating injustices of their own – thereby both compromising their moral integrity and wronging others – it might be appropriate to judge employers more harshly than we would if workers had access to permissible means of resistance. Unless we think agents are always fully absolved when conduct that foreseeably engenders harm is mediated by the intervening agency of a second party, we should perhaps criticise employers not only for the original unjust treatment but for backing workers into a corner from which they may sometimes be able to escape only by victimizing others.
Footnotes
Acknowledgements
This paper grew out of lengthy conversations with my colleagues at the University of Essex. For illuminating discussion, the author would like to thank David Axelsen, Miranda Simon, Isa Trifan, and, especially, Paul Bou-Habib. In addition, for the valuable feedback they provided on various versions of the paper, the author would like to thank Rufaida Al Hashmi, Alice Baderin, Malte Jauch, Robert Jubb, Timo Jütten, Shmuel Nili, Shuk Ying Chan, Martin Sticker, and Charlotte Unruh. Finally, the author would like to thank the anonymous reviewers who read the paper for this journal, and, for his admirable editorial work, Ryan Pevnick.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
