The Individual Unfreedom of the Proletarian

Cal Fawell
University of Chicago
Marko Winedt
Hanci Lei
Neil Sehgal
Matthew Dowling
Spring 2019


In this essay, I engage with G.A. Cohen’s argument by analogy that proletarians are individually free. I grant that Cohen’s analogy successfully represents the world. I disagree, however, with his conclusion, and use Philip Pettit’s conception of freedom as non-domination to demonstrate that proletarians are individually unfree. Specifically, I argue that even though fewer proletarians leave the proletariat than possibly could, they are nonetheless “dominated”—and thus, each is individually unfree.

This essay grants the accuracy of Cohen’s analogy, and from this assumption draws the conclusion that proletarians are individually unfree. In drawing this conclusion, this essay follows the style of modus ponens. It first argues for the conditional: if Pettit’s notion of freedom holds, then the conclusion Cohen draws must fail. It then argues that Pettit’s notion holds, showing that it accurately captures our intuitions on the subject. In arguing for the conditional, this essay contends that Pettit’s criteria for unfreedom are satisfied for each individual proletarian. In arguing for the antecedent, it demonstrates a number of intuitive considerations which support Pettit’s conception of freedom. From this, it concludes that each character in the room of Cohen’s analogy is unfree. Combined with the original hypothetical stance that Cohen’s analogy accurately relates to the world, it follows that proletarians are individually unfree.

In The Structure of Proletarian Unfreedom, G.A. Cohen addresses the status of wage-earners whose only available resource is their potential to work for a salary; they are unable to produce the necessities of life themselves. These people are known as proletarians; the class of all such proletarians is known as the proletariat. Cohen begins with the view that since only a few proletarians are able to advance from the proletariat, proletarians—taken as a whole—are forced to sell their labor power. He denies, however, that this conflicts with his claim that “most proletarians are not forced to sell their labor power,” justifying this denial with an analogy:


Ten people are placed in a room the only exit from which is a huge and heavy locked door. At various distances from each lies a single heavy key. Whoever picks up this key—and each is physically able, with varying degrees of effort, to do so—and takes it to the door will find, after considerable self-application, a way to open the door and leave the room. But if he does so he alone will be able to leave it. Photoelectric devices installed by a jailer ensure that it will open only just enough to permit one exit. Then it will close, and no one inside the room will be able to open it again.[1]


The structure of this analogy is relatively simple. A group of people are imprisoned, all are initially able to escape, but ultimately only one is allowed to do so. He draws out the significance of this analogy by saying that:


Whomever we select, it is true of the other nine that not one of them is going to try to get the key. Therefore it is true of the selected person that he is free to obtain the key, and to use it. He is therefore not forced to remain in the room. But all this is true of whomever we select. Therefore it is true of each person that he is not forced to remain in the room, even though necessarily at least nine will remain in the room, and in fact all will.[2]


Here, Cohen alludes to the crucial assumption that whoever attempts to exit the room will not be interfered with from this task by the other occupants (elsewhere: “…each is free to use [the means of egress], since, ex hypothesi, no one would block his way”[3]). Cohen uses this analogy to argue that every individual in the proletariat is free to leave it. He neatly introduces this line of thought by saying that “there are more exits from the British proletariat than there are workers trying to leave it. Therefore, British workers are individually free to leave the proletariat.”[4] He calls this “argument 7.” Analogously: just as any chosen member of the room is able and free to leave it, Cohen believes that this same freedom applies to any—and therefore every—member of the proletariat.


Yet, Cohen understands that “there is a great deal of unfreedom in their situation.”[5] He invents a term for this unfreedom, naming it “collective unfreedom,” which describes the case in which “not more that one can exercise the liberty they all have.”[6] Cohen uses this device to explain our intuitions about the unfreedom of this situation, ultimately claiming that “there are very few exits from the British proletariat and there are very many workers in it. Therefore, British workers are collectively unfree to leave the proletariat.”[7] He calls this “argument 8.”


Cohen means to hold argument 7 (that proletarians are individually free) and argument 8 (that proletarians are collectively unfree) together in a state of non-contradiction in order to satisfy our intuitions about freedom on both collective and individual levels. It is the aim of this essay to show that, on the contrary, argument 7 fails. To do this, Philip Pettit’s conception of liberty as non-domination is enlisted. In short, Pettit holds that X dominates Y if X has the capacity to interfere on an arbitrary basis in certain choices that Y is in a position to make. There are three components to Pettit’s view of domination: one is dominating if they have (a) a capacity to interfere, (b) on an arbitrary basis, (c) in certain decisions another is able to make. Critically, domination is an inherently potential characterization of unfreedom. It is the capacity for a certain kind of interference, not the interference itself.


With respect to (a), the capacity to interfere, Pettit claims that this interference must be intentional and must worsen the other’s situation. He claims as much when he says that “when I interfere I make things worse for you, not better. And the worsening that interference involves always has to be more or less intentional in character: it cannot occur by accident.”[8] Thus, in situations where only accidental or positive interference is possible, it violates Pettit’s view to say that there is domination.


In describing how a choice situation may be worsened, Pettit provides three variables: options, expected payoffs, and outcomes. Understanding the idea of worsening options is pretty straightforward—Pettit explains this as “changing [for the worse] the range of options available.”[9] For our purposes, the elimination of options fits this criterion. By “worsening the expected payoffs,” Pettit means the attachment of punishment to a certain course of action in order to discourage it. By “worsening the outcomes,” Pettit means attaching punishment to a course of action that has already occurred in order to negatively affect the actual payoffs.


Pettit qualifies (b), the condition of arbitrariness, by studying the relative locations of the agent deciding and the person affected (“the other”). He understands an act to be arbitrary if “it is subject just to the arbitrium, the decision or judgment, of the agent,” and moreover, that the action is done “without reference to the interests, or the opinions, of those affected.”[10] In other words, an arbitrary action is one which denies the status of the affected party as a meaningful human being by disregarding their wants and needs. On the other hand, Pettit sees non-arbitrary decisions as those which track, or take into account, the preferences and welfare of the people liable to interference.


Pettit explains (c), or “certain decisions the other is able to make,” as a way of compartmentalizing different domains of freedom and domination. He says that the most salient aspect of this clause “is that it mentions certain choices, not all choices. This highlights the fact that someone may dominate another in a certain domain of choice, in a certain sphere or aspect or period of their life, without doing so in all.”[11] In other words, the other can be dominated at work while free in the home, or vice versa; the other can be dominated politically or socially, but not in their decisions of which music to listen to, etc. Pettit holds that domination—and therefore, freedom—can vary in extent, intensity, and across different domains. Domination can be in the form of absolute power over another in many critical domains, limited ability to interfere in largely inconsequential domains, and everything in-between. The intensity of domination varies along both of these dimensions, and though Pettit fails to give an explicit framework for determining an order of severity, he acknowledges, at the very least, that loose hierarchies exist.


It is critical to understand that Pettit’s conception of domination is such that interference does not need to be actual in order for there to be unfreedom. At the heart of this view is the fact that the mere ability to interfere engenders unfreedom through domination. He writes that:


The possession by someone of dominating power over another—in whatever degree—does not require that the person who enjoys such power actually does not require even that the person who enjoys that power is inclined in the slightest measure towards such interference. What constitutes domination is the fact that in some respect the power-bearer has the capacity to interfere arbitrarily, even if they are never going to do so.[12]


Here, Pettit explains that unfreedom can occur even without actual interference. Not only does he believe this, but also that unfreedom can occur even where actual interference seems very unlikely. He strongly emphasizes that the mere ability of arbitrary interference, however “small” that interference may be, causes a proportional amount of unfreedom.


It is now possible to apply Pettit’s view to Cohen’s analogy. In Pettit’s view, any one person in the room is dominated by any other person in the room. This is true because each person has the relevant capacity to interfere in the relevant way with every other person. Suppose we initially select one person from the hypothetical room, as Cohen does. It is true that they have the capacity to interfere in the relevant sense—intentionally and harmfully. This person could choose to leave the room, thus rendering every other inhabitant finally trapped. In Pettit’s terminology, this certainly worsens their choices, since it removes their choice to leave the room. It is especially clear that this is a worsening of their choices when one recalls that the room stands for the proletariat, and exiting it stands for ascending to a higher class. It is further possible that this person could make their exit in order to intentionally worsen the lives of the other occupants. There is nothing stopping the person from growing resentful of their peers, and seeking to harm them by leaving the room for the sake of finally imprisoning them. Even if this sounds unlikely, it is important to remember that “what constitutes domination is the fact that in some respect the power-bearer has the capacity to interfere arbitrarily, even if they are never going to do so.”[13] Domination is not a question of probability, but rather capacity. Thus, by having the ability to leave the room, the selected person has the ability to intentionally interfere in the lives of the others for the worse.


Second, this person may perpetuate this act on a totally arbitrary basis. Nothing forces the person to track the interests of the other occupants of the room when deciding whether or not to leave. They are free to make their decision to leave without any regard for the wants or needs of the others, and can reason purely from their own preferences and needs. Thus, the selected person has the capacity to interfere arbitrarily. Finally, their interference is, indeed, in “certain choices that the others are willing to make.” It is a real decision whether or not the others choose to leave the room. The capacity to arbitrarily interfere centers on this very locus of choice, and thus meets the third criteria for domination.


Therefore, the selected person has the capacity to interfere—on an arbitrary basis—in certain choices that the others are in a position to make. But this is just the definition of domination—thus, the selected person dominates the others, which means that each of the others is in a position of domination. In Pettit’s view, this amounts to saying that each of the people in the room is unfree. But as Cohen writes, all this is true of “whomever we select.”[14] Thus, each person in the room dominates all the others, which means every person in the room is in a position of domination. Therefore, every person in the room is individually unfree. This conclusion is the negation of Cohen’s in his “argument 7.”


It is worth briefly noting that Pettit generally formulates the dominating relationship between persons. He says that “while a dominated agent, ultimately, will always have to be an individual person or persons, domination may often be targeted on a group or on a corporate agent: it will constitute domination of individual people but in a collective identity or capacity or aspiration.”[15] Thus, the fact that the chosen agent is able to arbitrarily interfere in the same particular domain of choice for multiple other people does not mean it is not domination.


A more serious objection to the claim that each person in the room is dominated and therefore unfree, however, is the thought that reciprocal domination over an identical domain of choice is impossible. It seems, prima facie, that because each person is dominating all the others just as much as the others dominate them, all have the same choice options; since the idea of domination seems to intuitively rely on asymmetrical relationships, it would seem that the idea of reciprocal domination is internally inconsistent. This worry may be mitigated in a number of ways. First, it is important to keep in mind the precise definition of domination: X dominates Y if X has the capacity to interfere, on an arbitrary basis, in certain choices Y is in a position to make. There is nothing in this definition which logically excludes it from being a reflexive two-place relation. The intuitive connection with asymmetry can be explained by the fact that the capacity for arbitrary interference is often made possible by asymmetries such as wealth, power, status, and others.


Though this is the case, Pettit explicitly mentions that reciprocal power to interfere may be used as a strategy for achieving non-domination. He mentions two possibilities: defense and deterrent. He writes that “the strategy of reciprocal power is to make the resources of dominator and dominated more equal so that, ideally, a previously dominated person can come to defend themselves against any interference on the part of the dominator.”[16] In other words, the dominated party would be able to use their resources to counter the arbitrary interference of the other, and thus, this very possibility would eliminate the capacity of arbitrary interference and dissolve the dominating relation. However, it is not clear how this defensive interference would be possible in Cohen’s analogy, since he assumes that if one tried to leave, “no one would block his way,” and once the original person left the room, no one else could leave in order to finally imprison them.[17] By the very nature of this particular kind of interference, defensive interference—within Cohen’s analogy—is impossible.


Pettit himself admits that the “ideal” of defensive interference will rarely materialize. Instead, he discusses retaliation as a second way of using reciprocal power to reduce domination. He says that the dominated agent may be able “at least to threaten any interference with punishment and to impose punishment on actual interferers.”[18] Cohen’s analogy, however, does not speak to such retaliatory interference. Given that the others would be locked in the room, it is hard to see how they would be able to affect the escaped person. Here, though, the analogy breaks down, since there is nothing necessarily physically separating someone who “recently exited the room” (e.g., joined the corporate workforce, or gained control of a company) from the wrath (e.g., physical attack) of their previous fellows. Still, it is not the original situation of reciprocal domination which engenders the possibility of retaliation. The relevant kind of domination—that is, leaving the room and trapping the rest—does not enable retaliation. Ultimately, then, this response has no bearing on the claim that the reciprocal domination must be self-annihilating, since it is not in virtue of this that retaliation is possible.


For example, suppose that A and B work in a factory. A hates B, and so A decides to harm B by getting promoted to become the owner of the factory—eliminating B’s ability to do so. B cannot retaliate by also coming to own the factory. The job has been filled; the “room” has been “exited.” Yet, B can destroy A’s property, harass A’s person, or even threaten A’s life. This retaliation does not stem from the original position of reciprocal domination. Thus, Pettit’s only two ways that reciprocal domination might eliminate itself do not find application in Cohen’s scenario. We may safely conclude, then, that all the people in the room are dominated by at least one (and in fact all) of the others, which just means—in Pettit’s view—that each person in the room is individually unfree.


It will now be the aim of this essay to motivate Pettit’s view independently from its application to Cohen’s argument. I shall do this by arguing that, even on the negative view of liberty, it is intuitively desirable to expand the horizon of freedom from mere non-interference to non-domination in Pettit’s sense. To begin with, certain kinds of obvious unfreedoms cannot be recognized as such under a negative conception of liberty which only recognizes interference as unfreedom. For instance, consider a slave whose master has not exercised their capacity to arbitrarily interfere in the slave’s life for the worse, and gives no indication that they ever will. Obviously, the slave is unfree, for the slave is a slave. Pettit remarks that:


The observation that there can be domination without interference connects with the theme highlighted in the last chapter, that slavery and unfreedom is consistent with non-interference: that it can be realized in the presence of a master or authority who is beneficent, and even benevolent.[19]


In other words, the example of the slave under a benevolent (non-interfering) master shows first that a strictly negative view of liberty fails to account for this obvious instance of unfreedom. The strength of Pettit’s view lies in the fact that the reasons we conclude this slave to be unfree are simply that the criteria of domination are met—and this shows the intuitive strength of his view. In elucidating these intuitions, Pettit quotes Richard Price as saying that “individuals in private life, while held under the power of masters, cannot be denominated free, however equitably and kindly they may be treated.”[20] Like Pettit, Price focuses on the persons “under the power of masters,” though not actively interfered with. Being under the power of another, in Price’s sense, certainly seems to imply the master’s capacity to arbitrarily interfere. Analyzing the unfreedom of the slave leads Price to conclude that it is this capacity to be arbitrarily interfered with which engenders unfreedom.


Pettit himself characterizes domination as leading to this kind of slavish relationship. In domination, he says, “the powerless are at the mercy of the powerful and not on equal terms. The master-slave scenario will materialize, and the asymmetry between the two sides will become a communicative as well as an objective reality.”[21] Pettit also finds examples of this intuition in Machiavelli and Montesquieu. Machiavelli describes the power of a free community as “the power of enjoying freely his possessions without any anxiety, of feeling no fear for the honor of his women and his children, of not being afraid for himself,” and Montesquieu defines liberty as the “tranquility of spirit which comes from the opinion each one has of his security, and in order for him to have this liberty the government must be such that one citizen cannot fear another citizen.”[22] In both characterizations of freedom, the common denominator is that when the conditions of domination are met—that is, when one has to “bow and scrape,” appease their dominator, and maintain “eternal discretion”—there is personal unfreedom (and vice versa). In other words, it is strongly intuitive to claim “freedom if and only if non-domination.” What Pettit’s definition does is make this intuition explicit.


Pettit’s view is further strengthened by accounting for the intuition that a state of freedom should foment equality among the agents by whom it is enjoyed. Pettit often alludes to this intuition by showing the converse: in a dominating relationship, the individuals are not able to look each other in the eye. The dominated person may only assert their equality on the pain of being interfered with, which is to say, they cannot. Similarly, the dominating agent’s power suggests a condescending mindset. More to the point, it seems that domination dehumanizes the dominated, while non-domination forces the would-be dominator to see the other as a person. Pettit says that in a state of non-domination,


You do not have to live either in fear of that other, then, or in deference to them. The non-interference you enjoy at the hands of others is not enjoyed by their grace and you do not live at their mercy. You are a somebody in relation to them, not a nobody. You are a person in your own legal and social right.[23]


Here, Pettit argues that when one is free from domination, one secures one’s personhood. By this argument, Pettit seeks to directly align non-domination with freedom, as opposed to appealing to the converse, as above. Non-domination seems to imply fundamental equality and restores the relationship to that which holds between equal persons, instead of that which holds between master and slave. Because non-domination shares this intuitive property of freedom, and because of the (above) strong intuitions that the conditions for freedom are those of non-domination, Pettit’s view of freedom of non-domination ought to be accepted.


Finally, it should be noticed that the intuitive support for Pettit’s conception of freedom does not trivially apply to the people in Cohen’s analogical room. Since the people in this room are on equal standing, it may seem like there is no room for the master-slave relationship to emerge, that people will be able to meet each other in the eye, and that there can be no degradation of personhood as there is when the domination is truly one-sided.


A thought experiment will make clear how the intuitive strength of Pettit’s view holds, even in contexts like Cohen’s analogy in which there is reciprocal inter-domination. Imagine the following situation: persons A, B, and C are in the room. Person A wishes to have an affair with person C and afterwards leave the room. However, person B is jealous of this, and plans to leave the room as soon as the affair commences—if it does—in order to forever imprison A, denying A’s ultimate wish for escape. B makes this known to A. Now, A must appease B by not having an affair with C in order to fulfill their ultimate wish to leave the room. Thus, B dominates A, as A is unable to exercise their individual freedom: in order to exit the room, A must “bow and scrape” to B’s preferences. The mere fact that A also dominates B does not mean that A does not acutely feel the unfreedom of their domination by B. A cannot look B in the eye, because B stands between A and A’s desires. B’s whim determines the fate of A’s life. That A may leave the room, forever trapping B and C, does not negate this fact.


All this thought experiment seeks to demonstrate is that the intuitions which locate freedom with non-domination are still present in Cohen’s analogy. Succinctly, these intuitions are first that someone is free if they are able to exercise their freedoms without the approval of another, and unfree if not (which is equivalent to defining freedom as non-domination). This is exactly the case in the analogy of the room: person A is unable to exercise their freedoms because they lack the approval of person B. Backtracking for a moment, the second intuition is that freedom as non-domination accords with the intuition that freedom produces equality. The inverse, that unfreedom induces inequality, also holds in Cohen’s room: precisely because of the possibility of situations such as the above thought experiment, the people in the room will be unable to look each other in the eye. Since this is the inverse, it is hardly a rigorous proof; however, even the inverse demonstrates a correlation between equality and non-domination which certainly does not malign the suggestion that they are equivalent.


Suppose instead that the people in the room were individually free, as Cohen claims. Then they would enjoy equality amongst themselves, according to the rule: if there is freedom, then there is equality. When Cohen speaks of the possibility of solidarity among the members of the room—for instance, those who want to rise, not out of the proletariat, but with the proletariat—he asserts the possibility of the consequent, implying the situation is one of genuine freedom. However, two considerations must be noticed. First, the above thought experiment is meant to show that this possibility is nontrivial, and must be argued for. As the experiment shows, there is also a great possibility that there is profound inequality—an inability to “look each other in the eye”—amongst the members of the room. Insofar as this denies the consequent, it speaks to the possibility of the original situation being one of unfreedom. Second, it is important to remember that even if Cohen were able to succeed in proving the possibility of equality amongst the members of the room, it would be fallacious to conclude from this that they were free. Therefore, both of the original intuitions which support Pettit’s particular notion of freedom (respectively, the potentiality of interference and the relationship between freedom and equality) remain relevant to Cohen’s analogy of the room. Since these intuitions remain intact, Pettit’s concepts should still be held in the analogy of the room. Thus the individuals in Cohen’s analogy are individually unfree.


Before concluding, it is worth noting that Cohen modifies his analogy in the same paper: in the modified analogy, exactly two people may leave the room. However, this modification leaves the above arguments unaffected. Indeed, as long as there are fewer exits than people in the proletariat, nothing changes. This is because the above arguments remain unchanged if we select two people at random, treat them as a single agent, and then in a similar manner proceed to show that the others are dominated by them, concluding that since they were chosen at random, all are dominated. Pettit has no problem with this kind of strategy, saying that “while a dominating party will always be an agent…it may be a personal or corporate or collective agent: this, as in the tyranny of the majority, where the domination is never the function of a single individual’s power.”[24] Thus, considering the dominating agent as an arbitrary group of people (the exact count of which equals the number of exits from the proletariat) rather than a single individual does not threaten any of the above conclusions.

In conclusion, this essay first demonstrated that on Pettit’s view of freedom as non-domination, Cohen’s analogical backing for “argument 7” fails to prove that any proletarian is individually unfree. This is precisely because in this analogy, every person in the room is individually dominated. Assuming, with Cohen, that this analogy accurately represents reality, it is useful to step out of the analogy and towards what it depicts: Cohen begins by saying that there are more exits from the proletariat than proletarians leaving, and thus that any proletarian is individually free to leave the proletariat. Pettit’s view, however, shows how proletarians can hardly be said to enjoy freedom, since they are constantly threatened with losing it based on the arbitrary whims of their peers. Thus, holding Pettit’s view entails rejecting Cohen’s. This essay then argued for the intuitive strength of Pettit’s view, showing first that Pettit’s formulation of freedom matches with common, intuitive formulations, and then showing how the claim that ‘freedom is non-domination’ accurately tracks our intuitions about the relationship between equality and freedom. From this, in the style of modus ponens, it follows that the people in the room are individually unfree. Once again stepping out of the analogy, it follows that proletarians are each and all individually unfree.



[1] Cohen, G. A. “The Structure of Proletarian Unfreedom.” Philosophy & Public Affairs 12, no. 1 (1983): 9.
[2] Cohen, “Proletarian Unfreedom,” 10.
[3] Cohen, “Proletarian Unfreedom,” 10.
[4] Cohen, “Proletarian Unfreedom,” 13.
[5] Cohen, “Proletarian Unfreedom,” 11.
[6] Cohen, “Proletarian Unfreedom,” 11.
[7] Cohen, “Proletarian Unfreedom,” 14.
[8] Pettit, Philip. Republicanism: A Theory of Freedom and Government. (Oxford: Oxford Univ. Press, 2010), 52.
[9] Pettit, Republicanism, 53.
[10] Pettit, Republicanism, 55.
[11] Pettit, Republicanism, 58.
[12] Pettit, Republicanism, 63.
[13] Pettit, Republicanism, 63.
[14] Cohen, “Proletarian Unfreedom,” 10
[15] Pettit, Republicanism, 52.
[16] Pettit, Republicanism, 67.
[17] Cohen, “Proletarian Unfreedom,” 10.
[18] Pettit, Republicanism, 67.
[19] Pettit, Republicanism, 64.
[20] Pettit, Republicanism, 64.
[21] Pettit, Republicanism, 61.
[22] Pettit, Republicanism, 71.
[23] Pettit, Republicanism, 71.
[24] Pettit, Republicanism, 52.


Cohen, G. A. 1983. “The Structure of Proletarian Unfreedom.” Philosophy & Public Affairs 12 (1): 3–33.

Pettit, Philip. Republicanism: A Theory of Freedom and Government. Oxford Univ. Press, 2010.

  • brown
  • Logo_White
  • poli_theory_logo
  • Facebook Social Icon
  • Twitter Social Icon

© 2019 Brown University Journal of Philosophy, Politics & Economics