Micro-domination

Abstract

This article analyses the phenomenon of ‘micro-domination’, in which a series of dominated choices are individually inconsequential for a person’s freedom but collectively consequential. Where the choices concerned are objectively inconsequential, micro-domination poses a problem for ‘objective threshold’ accounts of domination which either prioritise particularly bad forms of domination or exclude powers that do not risk causing serious harm to their victims. Where the choices concerned are subjectively inconsequential to the victim, micro-domination poses a problem for the common republican strategy of creating arenas of contestation for victims of domination, which rely on victims objecting strongly enough to a dominating relationship to sound the alarm. This kind of invigilation may systematically fail victims of micro-domination. Throughout the article, I suggest some ways of better accounting for and responding to cases of micro-domination.

Sometimes domination is obviously very serious: a state’s power to break down your door in the middle of the night and lock you in a prison cell, for instance, or a man’s power to beat or starve his family with impunity. These are serious power imbalances involving the ability to cause serious harms. But sometimes the potential harms involved are not obviously serious. My neighbour could refuse my request to borrow his screwdriver – if he does, I will have to go and buy one. Perhaps my landlord can decide whether the garden is being kept in a good condition, and can tell me to cut the grass once a week rather than once a month. Or perhaps my manager can decide whether I am dressed professionally, and can tell me to wear trousers rather than jeans.

On their own, none of these powers risk serious harm, and they may be little more than irritating. Nonetheless, I will argue that some of them, and others like them, are extremely important forms of domination. In this article I sketch out and explore the phenomenon of ‘micro-domination’, in which a series of dominated choices are individually inconsequential but collectively consequential. Contemporary republicanism risks systematically ignoring micro-domination in at least two senses.

Where the choices concerned are objectively inconsequential, micro-domination poses a problem for what I call ‘threshold’ accounts of domination, such as those offered by Philip Pettit (2012), Cécile Laborde (2010) and Christopher McCammon (2015). These threshold accounts aim to do one of two things: either prioritise (more or less strongly) particularly bad forms of domination, or keep domination a morally serious category by excluding those would-be dominating powers that do not risk causing serious harm to their victims, by whichever objective threshold harm is defined – infringing on my basic liberties, causing me to fail to meet my basic needs, threatening my basic capabilities, and so on. In the case of micro-domination, they risk failing in both tasks.

Where the choices concerned are subjectively inconsequential to the victim, micro-domination poses a problem for the common republican strategy of creating arenas of contestation for victims of domination – the ‘fire-alarm’ model of invigilation – which relies on victims objecting strongly enough to a dominating relationship to sound the alarm. This kind of invigilation may systematically fail victims of micro-domination.

Micro-domination, then, complicates two important parts of the contemporary republican project: attempts to carve out those kinds of domination that we should care especially about (or, at the lower end of the scale, attempts to keep domination a morally serious affair), and attempts to reform public and private institutions to best reduce that domination. Although they are overlooked in the literature, these problems are not insuperable, and throughout the article I suggest some ways of moving forward that better account for cases of micro-domination.

The first section offers a general account of domination, and the second section introduces micro-domination. The third and fourth sections focus on the problems that micro-domination poses for objective threshold accounts of domination, and the fifth explores the problems it poses for the use of subjective thresholds as part of some common strategies to reduce domination, sketching an alternative and superior strategy in the form of democratisation. Throughout the article, I suggest some ways of better accounting for and responding to cases of micro-domination.

Domination

Micro-domination is a kind of domination. In broad strokes, I am dominated if I am under someone else’s thumb – in their power. In narrower ones, D (dominator) dominates V (victim) if:

D and V are in a relationship that it would be costly for V to leave;

D has the power to interfere in choices that V would otherwise be in a position to make; and

V does not herself control D’s exercise of that power, directly or indirectly.

This is pretty standard republican fare, or at least is intended to be. It is roughly what Philip Pettit (2012) has in mind in his more recent work. ¹ Domination is not an all-or-nothing affair, and any given relationship can be more or less dominating along several axes. One way of appreciating this is to ask what might change to make such a relationship more dominating. Here are some options:

You could dominate me more if it became costlier for me to leave the relationship – imagine that V is D’s stay-at-home wife, and V’s skills become less and less marketable as she spends longer out of the workplace. Divorce becomes costlier for V, so she has more reason to stay married.

D could dominate V more if he acquired a greater power of interference. Again assuming that the pair are married, imagine that domestic violence laws were suddenly weakened. If D will no longer risk criminal penalties for certain kinds of domestic assault, he has a greater power to freely enforce his will.

D could dominate V more if he acquired the power to interfere in a greater range of her choices. For instance, imagine that D is V’s landlord: he can (legally) exercise power over only a narrow range of choices that V could otherwise make herself – he cannot evict her on the basis that she cooks food or reads books that he dislikes, but can evict her for withholding rent, or illegal drug use. If this protection suddenly disappeared, then D now has the power to interfere in many more of V’s choices.

Finally, and most relevant for our purposes, D could dominate V more if he gained the power to interfere in choices that are more consequential. Quite what this means is up for debate – a debate which will occupy a good deal of this article – but the intuition is obvious enough. Health and safety regulation is a good example of this. Imagine that D is V’s employer in a chemical plant, and that he has the power to determine what kind of protective apparel she can wear at work. His choice will be consequential, because different options will give V more or less protection, but he may be prevented from giving her no protection from those chemicals. Lower these safety standards, and D gains the power to make a new and more consequential choice, one that leaves V at risk of serious harm. Because of this, his domination is more severe. It is common for republicans to identify a range of choices that are supposed to be especially important for people to make themselves, such that the power to interfere in any of them makes for especially dangerous forms of domination.

Micro-domination

Micro-domination involves the power to interfere in choices that are not consequential in this last sense – they do not on their own risk serious harm, say – but where the set of many such dominated choices is consequential, whether the same choice repeated over time or many similar choices. Tom O’Shea coins the term in the course of discussing the relationship between disability and domination, and the way that aspects of this domination might – if they are not careful – fly under the radar of otherwise vigilant republicans. For him, micro-domination is ‘the capacity for decisions to be arbitrarily imposed on someone, which, individually, are too minor to be contested in a court or a tribunal, but which cumulatively have a major impact on their life’ (O’Shea, 2018: 136). And, slightly later, on the importance of paying attention to these minor decisions, ‘If the republican goal is a citizenry resiliently protected from arbitrary power, then the threat posed by a few major channels of domination is not necessarily greater than the contribution of countless smaller tributaries’ (O’Shea, 2018: 137).

My use of the term will be wider than O’Shea’s, as will become clear, but since I take myself to be dislodging the same problem he identifies, his example is a good starting point. A psychiatric inpatient finds themselves interfered with in various choices that they describe as ‘mundane, routine, noncrisis kinds of matters’ (O’Shea, 2018: 136) – when they’re allowed to use the phone, who they can spend time with, and so on. Individually, none of these instances are very consequential, either objectively (any one of them would not cause serious harm, etc.) or subjectively (any single one of them would be little more than irritating, say). But the overall effect, as the inpatient describes, is oppressive and stifling, compromising the equal status of the patient and their nurses. It is not as grand or impressive as the power that a king has to arbitrarily put his subjects to death, but it doesn’t have to be: the accretion of small intrusions into the patient’s free choice is consequential and harmful, even though none of those component intrusions may themselves be consequential.

What makes a choice inconsequential? We could try to identify an objective threshold upon which we could say, for instance, that D’s ability to choose whether V wears gloves or no gloves when handling dangerous chemicals is consequential, but D’s ability to choose whether or not V can wear colourful jewellery at work is not. V being left unprotected might cause her serious physical harm, or cause her welfare to drop below some minimum level, whereas her inability to wear the colourful jewellery would not. Micro-domination, then, involves the power to interfere in a number of these objectively inconsequential choices in such a way that the effect on the victim tips over the threshold, and becomes consequential.

Alternatively, we could try to use a subjective threshold. On that picture, micro-domination involves the power to interfere in choices the outcome of which – for any given choice – the victim just doesn’t care about all that much. V might (rightfully) care a great deal whether or not she gets chemical burns, but be merely irritated that she can’t wear jewellery. But the victim does care a great deal when enough of those subjectively inconsequential choices are invaded. The two kinds of thresholds – objective and subjective – are related to different problems for some mainstream approaches to domination. I look at objective thresholds in the third and fourth sections, and subjective thresholds in the fifth section.

Objective thresholds and prioritisation

First, take micro-domination to involve the power to interfere in choices that are, on their own, objectively inconsequential. Take any account of what it is for a choice to be consequential or important for the chooser – regardless of what he or she may think about it – and that account will involve a threshold above which a choice is consequential and below which it is not. Micro-domination, then, is a term for forms of domination which have a certain shape relative to that threshold.

As an example, imagine that you think there is something particularly bad about lasting physical injury. So you decide to care especially about domination in which the dominator has the power to choose whether or not to inflict those lasting injuries on their victim. We have a threshold with two features: the selection criterion is lasting physical injury, and the purpose of selection is to pick out domination that we should care especially about. D is a hanging judge in a capricious justice system, and has the uncontrolled power to choose whether or not to sentence V to death – D’s domination clears the threshold, and is straightforwardly something that we should care especially about. But imagine that D is V’s employer in a factory, and has the power to make only much smaller and more constrained decisions. He cannot choose to physically harm her, much less sentence her to death. He cannot even choose to make her work in conditions that might result in her being injured by the end of a shift. D can choose when to schedule V’s shifts and workload, however, and can choose every week whether or not to give her the most repetitive tasks on the production line. A week on those repetitive tasks won’t cause any lasting injury. A career on those repetitive tasks, without any variation or break, certainly will: it’s sometimes called a cumulative trauma disorder. As such, for the physical-injury-oriented observer, D’s ability to choose V’s shifts every week is an example of micro-domination. No individual instance of that interference counts as consequential, and so a snapshot of D and V’s relationship, or a small enough set of dominated choices, might not prompt further investigation or concern. But considered over time, D’s ability to regularly interfere in that choice crosses the physical injury threshold, and so should trigger both investigation and concern. The selection criteria risk missing the wood for the trees, and overlooking relationships of micro-domination that those criteria were intended to pick out.

These sorts of objective thresholds are used widely by theorists working on domination, although rarely explicitly. We can split them into two rough groups according to the purpose of the threshold: attempts to prioritise the kinds of domination that we have a special reason to act upon, and attempts to keep domination morally serious by excluding various apparently trivial relationships of uncontrolled power. Both are at least prima facie valuable goals, but both are vulnerable to overlooking micro-domination in the sense above. I take them in turn, dealing in this section with objective thresholds aiming to prioritise certain kinds of domination over others, and in the next section with objective thresholds that attempt to keep domination morally serious.

A good (and particularly strong) example of the first kind of objective threshold can be found in Pettit’s work, where he suggests a useful test for determining the obligations of a state towards its citizens. The state will have done its duty – and can be expected to do no more – in reducing domination when each citizen:

can look others in the eye without reason for the fear or deference that a power of interference might inspire; they can walk tall and assume the public status, objective and subjective, of being equal in this regard with the best. (Pettit, 2012: 84)

He calls this the ‘eyeball test’, and it sets an upper limit to the republican state’s duties: protecting people against forms of domination which don’t compromise this ideal of status equality is unnecessary. To do any more is not prohibited, except to the extent that any further action would risk increased state domination (not an idle worry, from the republican perspective), but it is an improper target of a republican state. The threshold matters a great deal, then, and part of what Pettit has in mind with this test is a way of addressing the worry that a state committed to non-domination would have an essentially unending task – it ‘makes the goal of a republican theory of justice more accessible than it might otherwise be’ (Pettit, 2012: 84). ²

Rather than setting the limits of the state’s obligations, other theorists merely want to find a way of prioritising especially urgent or harmful forms of domination. Cécile Laborde writes:

While domination is always an evil for republicans (it denies a basic interest in minimum autonomy or ‘discursive control’), it is a tragedy when it results in the denial of basic human capabilities (such as subsistence). Therefore the priority of global justice should be to reduce those forms of global domination which grant one set of agents the potential to deny basic capabilities to others. (Laborde, 2010: 54)

Laborde leaves the specific ‘basic, universal and objective human interests’ (Laborde, 2010: 56) she has in mind vague, but they include at least health and subsistence, and are intended to be relatively thin. Ian Shapiro takes a similar approach, although with reference to the resources required to satisfy basic interests, rather than basic capabilities. In identifying ‘which kinds of domination should concern us most’, he writes, we should ‘include access to the sources of security, nutrition, health, and education required to become effective adults in the prevailing economic system and to participate fully in a democratic political order’ (Shapiro, 2016: 22).

Whether in the language of obligations or priorities, the thresholds above are all meant to be action-guiding. They tell us which kinds of domination need addressing, or need addressing first, and which can be left alone, or left alone for now. Micro-domination causes difficulties for each one of these objective threshold approaches, because it causes difficulties for any objective threshold approach – certainly any threshold approach that urges us to direct our attention only or chiefly to examples of domination which cross that threshold. Using Pettit’s approach, for instance, we might think that the power of O’Shea’s psychiatric nurse to interfere in the trivial decisions of their patients isn’t the sort of thing that the state should concern itself with. I can still look you in the eye without fear or deference, even if you have the power to tell me when exactly I can make a phone call. But (as O’Shea himself argues) the nurse’s power to interfere in a wide range of those individually trivial decisions may very well cause their patients to fail the eyeball test. The risk here is not that Pettit couldn’t accept this judgement – he could – but that it is harder to spot, especially when we take seriously the idea of these threshold accounts as practical, action-guiding tools. Imagine that a committee of Pettit-inspired republicans decides to run an audit of the power of psychiatric nurses over their patients. One plausible way of doing this would be to make a list of those nurses’ powers and go through them, deciding whether or not they allow a patient to pass the eyeball test. The power to decide when a patient can have dinner may well pass the test, and the power to decide when a patient should go to bed may too. The power to decide when a patient can use the toilet? Perhaps tougher, but still a pass. This checklist approach is compatible with the use of the eyeball test and, because it fails to consider the set of powers as a whole, it will pass over micro-domination in silence. This piecemeal, power-by-power scrutiny is a common and intuitive way of assessing whether or not a relationship is dominating, especially in legislative or incremental periods of change. The relationship between individual police officers and individual members of the public, for instance, usually takes place against the backdrop of a proposed set of new police powers, or a newfound appreciation of the dangers of some power or set of powers in particular. Discussion of the arbitrary power of employers, such as it is, often takes place against the backdrop of a union legal campaign tightly focused on one set of powers, or a corporate campaign to depart from the existing status quo (California’s recent Proposition 22, for instance). If non-domination is going to be of much use in these sorts of discussions, then the possibility of micro-domination makes this power-by-power approach a risky way to go about it.

How could this power-by-power approach be avoided? As an alternative, we could insist on relationships, rather than individual powers, being subject to scrutiny. At least in principle this would prevent us from missing the wood for the trees in the above sense, by trying to keep in view the whole set of powers of interference that a candidate dominator has over a candidate victim of domination. But it is difficult, and in some cases perhaps impossible, to consider relationships as a whole when trying to pick out domination. In cases where it is the repeated domination of choices over time that matters, for instance, the situation is particularly tricky. As in the cumulative trauma example above, the accumulation of small injuries over time can be just as damaging, if less spectacular, than an individual serious harm. And particularly if our benchmark involves the kind of intersubjective equality targeted by the eyeball test, this metaphor seems to hold up well. Pettit himself discusses, and shares, the traditional republican concern that rational adaptation to domination over time will tend to create the figure of the ‘cringing, toadying, fawning sycophant’ (Pettit, 1999: 57) – the very opposite of someone who can look others in the eye without reason for fear or deference. That adaptation takes the repeated exposure to a dominating will, over a childhood, or a career, or a lifetime. To the extent that we often cannot predict how long someone will spend in a dominating institution – workplace, hospital, prison, school, etc – we can’t easily or always distinguish domination that will genuinely not risk us failing the eyeball test, and domination that hasn’t yet caused us to fail that test. That is, at any one moment we can’t readily distinguish between objectively inconsequential domination (on Pettit’s terms) and micro-domination. The same can be said for basic capabilities and basic interests, with the relevant modifications. There will be relatively uncontrolled powers of interference which do not individually risk seriously harming those capabilities or interests, but which are nonetheless part of a series of such powers – or the same power over time – which do risk serious harm.

This discussion of power-by-power and relationship-by-relationship approaches to identifying domination may seem at odds with the fact that many – perhaps most – republicans take a third option: focusing on domination as the absence of a certain kind of social status, the status of the free citizen. Pettit’s eyeball test is the result of this attempt to ‘take as a guiding heuristic the image of the liber, or “free person”, in the republican tradition’ (Pettit, 2012: 82), and he is not alone in grounding his republican programme with reference to the status of a free and undominated person. Republicanism-in-practice, on this view, isn’t chiefly a matter of identifying those people, those powers or those relationships that are dominating and guarding against them, but about asking whether people have a social status that prevents them being subordinate to others. In principle, the task of finding out whether I have that status will involve looking at all my relationships with others in their entirety and over time, and so will sidestep the problems with micro-domination discussed above. ³ And, again in principle, much of the task of reducing domination will involve broad-brush measures to give everybody that kind of secure social status in all their relationships: material security, secure access to justice and to political engagement, and so on. On this picture, republicans could expect to sidestep the problems with micro-domination above.

Things are not quite so simple in practice – successfully identifying and protecting a certain kind of undominated social status is easier said than done. Again taking Pettit as an example, much of what he says can indeed stay at the more general level of providing an undominated social status without descending too far into the weeds of the relationships in which the problems caused by micro-domination become sharper. He proposes, for example, entirely general programmes of what he calls infrastructure, insurance and insulation. Pettit’s infrastructure programmes include universal education, secure financial institutions, a decent transport system, and so on; insurance programmes secure at least a minimum level of medical and material wellbeing, preventing anyone from falling to such a point that they might be vulnerable to domination by others; his generic insulation programmes establish criminal law, protecting people from (among other things) being dominated by their stronger or more violent neighbours (Pettit, 2012: 110–122). Although the term isn’t ideal, we could say that these status-based programmes are ‘victim-centric’, designed to equip citizenship with the resources and the protections to guard those who hold it from domination whatever the source, without reference to any particular would-be dominators or would-be dominating relationships.

But Pettit himself recognises that there are limits to what can be done from this altitude, and supplements his generic programmes with special ones. ‘Special insulation’, he writes, ‘is the sort of protection that is required in relationships like those of wife and husband, employee and employer, debtor and creditor, where there are often asymmetries of power’ (Pettit, 2012: 114), and will require a close examination of the dynamics of those relationships, the duties and powers of each party, the options for voice and exit on the part of the would-be victim of domination, and so on. In short, even strong proponents of a status-based approach recognise the need to adopt a relationship-by-relationship approach (or indeed power-by-power approach) to flesh out significant parts of a republican programme. This descent from a status-based approach brings back in the problems described above.

This move away from a generic, status-based analysis of domination will also be a constant temptation when republicanism meets the real world. We can of course be guided by an image of the kind of status a citizen would hold in that society, but a more ‘applied’ republicanism will almost inevitably find itself focusing onto specific relationships, laws and formal and informal powers in the real world, or find itself adjudicating between rival options when deciding how to best change that world. This is not a bad thing, and I suggest that it’s a necessary part of putting an ideal of non-domination to work. Consider three good, recent examples, each of which makes use of Pettit’s account of domination: Eleveld argues that various forms of welfare-to-work conditionality dominate welfare claimants (Eleveld et al., 2020: 263–280); Crummett (2020) argues that prosecutorial discretion in the US is seriously dominating, and suggests less dominating alternatives; Roberts (2015) mounts a republican critique of the European Union’s data retention rules. All three are working with Pettit’s status-based approach, and all three could be said to be engaged in precisely the sort of ‘normative and institutional research program’ (Lovett & Pettit, 2009: 11) for which he has called. When status-based accounts of domination are put into practice they in fact descend and, to some degree at least, must descend to the level of individual candidate dominators, dominating relationships and dominating institutional arrangements. To the degree that they do so, if they make use of threshold accounts of domination they risk overlooking micro-domination.

As a final point, status-based threshold approaches also suffer from the same temporal problems as relationship-by-relationship approaches to domination. Just as we cannot always tell how long someone will spend in a dominating relationship (and so whether their time in that relationship is likely to cause them to fall below our threshold), one’s social status is often a function of what one’s life looks like, rather than what one’s month or year looks like. To use the workplace as an example, a familiar refrain in the US when discussing raising the minimum wage is that some jobs aren’t meant to be careers, and that fast-food workers being paid a pittance and poorly-protected are students, or teenagers, or otherwise people just getting started on their working life. The implication is that while a lifetime as a poorly-paid and mistreated service worker might compromise your social standing such that we should do something about it, having a period of poorly-paid service work is just something that everyone has to go through before they go to college and get a decent job. This life-stage approach to bad work has a good pedigree – it mirrors Paine and Lincoln’s republican acceptance of apprenticeships and wage labour as a period of economic infancy while saving up to buy land or a workshop of one’s own (Anderson, 2017: 22–32) – but the economic picture on which it might be based quickly became outdated after the Industrial Revolution. It may now seem obvious that poorly-paid service work cannot be sensibly assessed as if it were a short-lived stepping stone, but the calculation is harder with emerging labour practices: platform work, live-in caring work, zero-hour contracts, working without the hope of an adequate pension, and so on. Again, these temporal questions bring back in the risk of overlooking micro-domination in the form of cumulative effects of dominating powers, and do so even when we explicitly try to consider people’s overall social status.

Objective thresholds and ‘cheap’ domination

The other, quite different, reason to create an objective threshold is in order to keep domination morally serious, and this kind of threshold is also vulnerable to systematically misidentifying cases of micro-domination. In this case, the aim is not to prioritise certain kinds of domination, but to work a threshold into one’s definition of domination – those candidate relationships that fall under this threshold are simply not counted as domination at all. ⁴ The clearest expression of this goal comes from Christopher McCammon:

[A] useful conception of domination needs to make sense of domination’s close association with injustice. Thus the second constraint: our conception shouldn’t tell us that lots of morally unserious or clearly permissible forms of human interaction count as domination. If it does, we should get a new one. (p. 1031)

In other words, we need to avoid what McCammon (2015: 1033) calls the ‘cheap domination’ problem: in order to properly call me a dominator, there must be ‘something about my relation to you that might ground a legitimate complaint’ (McCammon, 2015: 1035). ⁵ McCammon convincingly argues that several mainstream theories of domination run into this problem, producing examples of domination that do not ground such a complaint. The power that a neighbour with the only swimming pool for a hundred miles might have over me – I may try to stay in their good books so that they continue to let me swim in it – is not liable to keep anyone awake at night, even though they might be able to leverage that power to ask for small favours every so often. And yet, argues McCammon, it satisfies Lovett’s conditions for domination. I am in an ongoing relationship with that neighbour that would be (subjectively) costly for me to leave, and they have some degree of power over me that isn’t adequately constrained by external rules. In a similar vein, Cécile Laborde (2017) has used the example of a lovestruck man who would do anything for his (unreciprocating, perhaps unaware) tennis instructor. He is totally at her command, should she ever choose to issue one, but something seems very wrong with saying that she dominates him – it is just not morally serious.

If cheap domination is a problem, and deserves a response that is not simply ad-hoc dismissal of these cheap cases, then many of those responses will involve an objective threshold. McCammon’s own solution is not straightforwardly an objective threshold; it involves deciding whether or not the would-be dominator can ‘attach costs to non-cooperation higher than the costs of cooperation for a very wide range of forms cooperation might take’ (McCammon, 2015: 1040; emphasis mine). This is a subjective threshold rather than an objective one, since it doesn’t matter for McCammon why the costs of non-cooperation are prohibitively high for me.

But many intuitive solutions to the problem of cheap domination do make use of objective thresholds. Cheap domination is low-stakes domination. The basic interests that Laborde uses to distinguish between more-and-less tragic forms of domination also work for her to rule out the cheap domination of the tennis student by his instructor – we can find a list of basic interests which, if threatened, allow us to say that we are examining a relationship serious enough to count as domination. The idea that objective thresholds are helpful for ruling out this kind of morally unserious domination is present but less explicit elsewhere: in O’Shea’s more recent work, for example, he argues that while the capacity to deny me something I need constitutes domination, the capacity to deny me something I merely strongly want would be ‘too lax’ and ‘implausible’ (O’Shea, 2019: 12). In order to rule out these too-cheap forms of domination, we need an account of what people need, rather than what they merely want. Both Laborde’s account of basic needs and O’Shea’s answer – focusing on those capabilities required to function as equal citizens – are plausible, but using either to distinguish between cheap and genuine domination raises the same problem as before. We risk mistaking morally serious relationships for morally unserious ones, and so discounting them as domination, when they exhibit the familiar kind of structure relative to our threshold. In other words, we again risk ignoring forms of micro-domination in the pursuit of keeping domination morally serious.

One way to proceed in the face of these risks would be to simply stop trying to keep domination morally serious, and to include all sorts of morally trivial relationships in the scope of our conception: besotted tennis students and their teachers, neighbours of pool-owners and the pool-owners themselves, the whole lot. This shouldn’t be dismissed out of hand. The case against doing so (as given by McCammon) is essentially that as a matter of fidelity to common usage, a conception of domination upon which lots of morally unserious relationships are dominating would be like a conception of domination upon which slavery is not dominating. We would have missed the mark, and whatever we are now talking about, we’re not talking about domination.

Three considerations pull in the other direction. The first is just a restatement of the problem of micro-domination: sometimes it is very hard to tell morally serious from morally unserious relationships, and any effort to do so will at least risk systematically excluding a certain kind of domination relationship from analysis of it in those terms. The fact that the resulting conception of domination will be less useful is at least a pro tanto reason to reject McCammon’s condition.

The second is to question McCammon’s appeal to intuition. It is not that unintuitive to think that some forms of domination aren’t morally serious. Consider a high-flying group of bankers who are guaranteed their very substantial salaries – local employment law makes them extremely hard to sack – but whose large Christmas bonuses come from a pot of money divided between them at the end of the year at their boss’s discretion. A mix of status anxiety, greed and rivalry means that they spend most of their time toadying up to their boss in a way that would make a courtier at Versailles cringe. They will dress like him, laugh at his jokes, vote how he votes and work long, unpaid hours to try to impress him. The boss happens to be nice enough. If he were a sadist, though, he could successfully ask them to do almost anything. They are under his thumb, even though they risk nothing worse than a bruised ego and having to make do on their high salaries. Are they dominated? Intuitions may differ, but it certainly doesn’t seem to me to be out of the question. ⁶ But it is not morally serious, at least not according to either of the objective thresholds above. I don’t think it debases the term to say that this boss dominates his employees, but that for other reasons (their guaranteed salaries, their ability to drop out of the race to impress him, that they merely want and do not need their bonuses) we should not care very much about that relationship. O’Shea and Laborde both distinguish between more and less consequential and serious forms of domination, as well as – implicitly or explicitly – trying to establish a floor beyond which a relationship no longer counts as domination at all. I’ve already argued that attempting to establish that first threshold can lead to problems of its own, but it is of obvious practical use to know which forms of domination demand urgent attention. If it does not stretch the notion of domination beyond breaking point to imagine morally unserious forms, then there’s no corresponding practical reason to want the second, lower threshold.

The third consideration is that sometimes, political philosophy is normatively surprising. It may sometimes be the case that what everyone considers to be examples of McCammon’s (2015: 1033) ‘ordinary, innocuous human interactions’ turn out to be morally serious after all. That is, those intuitions might be wrong. This touches upon a more general methodological issue in political philosophy that I have neither the space nor the ability to solve here, but a narrower discussion will be sufficient for my purposes. Relying in this sense upon pre-philosophical intuitions about which relationships are ordinary and innocuous, and so could not be examples of domination, is a good way to find out that domination is an exceptional deviation from a basically non-dominating norm. The history of republicanism is, among other things, a history of terrible intuitions about which sorts of widespread social relationships might constitute domination. Republican complaints in the past have included that it is dominating to treat citizens as if they were slaves (whose own treatment, in contrast, was non-dominating or a benign, developmental kind of domination), and to treat men like women (whose own treatment was non-dominating or a benign, developmental domination). These articulations of domination were often radical, but they were also clearly not radical enough, as republicans today recognise. We do not need any particularly developed theory of ideology to suggest that suitably entrenched systems of domination can quite easily seem like common sense, especially to those who benefit from them.

This is neither to say that republicans like Laborde, McCammon and O’Shea cannot find that lots of ordinary kinds of relationships are dominating – they quite explicitly can, and do – nor that we are especially likely to conclude at some point that it is dominating to have a nice pool and let your neighbours use it only with permission. But it should weaken the case for sticking tightly to our intuitions about which kinds of relationships are commonplace and innocuous, and therefore the case for establishing a firm lower bound to what can count as domination.

Of course, the problem of excluding micro-domination at this lower end of the scale can instead be lessened, but not eliminated, in the same way as when distinguishing between more and less consequential forms of domination – making sure to look at whole relationships over time, rather than individual powers of interference.

Subjective thresholds

Another kind of problem for republicans emerges if we instead take micro-domination to involve a subjective threshold between consequential and inconsequential interference. The example of an objective threshold I gave in the last section – where interference in a choice was consequential if it could result in lasting physical injury – made no reference to what the victim herself thought about that interference. In contrast, what I call ‘subjective thresholds’ are precisely about what the victim thinks about that interference.

I understand O’Shea as being concerned about one type of subjective threshold in particular: whether or not the victim cares enough about a single instance of interference that they think it is worth contesting. On a reasonable interpretation, in fact, this is how he characterises micro-domination as a whole: ‘… the capacity for decisions to be arbitrarily imposed on someone, which, individually, are too minor to be contested in a court or a tribunal, but which cumulatively have a major impact on their life’ (O’Shea, 2018: 136).

Creating arenas in which the victims of domination can contest the decisions made for and about them is an absolutely central part of the modern republican strategy for reducing domination. Very roughly, the strategy is this. D may have various sources of power that allow him to interfere in V’s life – he may be her husband, her employer, her nurse, and so on. In order to limit D’s exercise of that power, an obvious suggestion is to prohibit certain forms of interference he would previously have been in a position to make. In order to work, there must be some outside force – usually the state – willing to step in and enforce them. Assuming that the state must keep its eye on many people in D’s position, and assuming that V wants to experience as little domination as possible, one simple and obvious method of enforcement is to give V the power to send up a flare and have her case heard in a neutral setting. She could take her husband to court, or take her employer to a tribunal, in case either relevant dominator uses their power in a way that she wants to contest. If this system works well, it reduces the power of those dominators, and so reduces their domination.

This distinction is essentially the same as that made between ‘snap inspection’ and ‘fire-alarm’ oversight in political science, analogies first made with reference to the US Congress’s oversight of various executive agencies (McCubbins and Schwartz, 1984). ⁷ Snap inspection oversight is ‘comparatively centralized, active, and direct … with the aim of detecting and remedying any violations … and, by its surveillance, discouraging such violations’ (McCubbins and Schwartz, 1984). Fire-alarm oversight, on the other hand, ‘involves less active and direct intervention’, and ‘establishes a system of rules, procedures, and informal practices that enable individual citizens and organized interest groups to examine administrative decisions…to charge executive agencies with violating congressional goals, and to seek remedies from agencies, courts, and Congress itself’ (McCubbins and Schwartz, 1984: 166). Snap inspection oversight is costly, will involve investigating a great number of innocent decisions or cases and can only ever invigilate a sample of decisions or cases. Fire-alarm oversight, at least in principle, will mean that time spent on oversight will be spent looking into cases where there has already been a complaint.

There are other potential benefits of fire-alarm oversight to republicans in particular. Unlike snap inspections, fire-alarms demand some vigilance on the part of victims, and the cultivation of vigilance as close as the contemporary republican project gets to the traditional republican concern with designing institutions that promote civic virtue. Pettit, for instance, writes that ‘the price of liberty, in the old adage, is eternal vigilance, in particular vigilance in looking at those in power and in challenging, where necessary, their claims and initiatives’ (Pettit, 2012: 52), and calls for institutions that can help encourage citizens to develop that vigilant character.

In any case, there are several reasons for republicans to lean towards fire-alarm-style arenas of contestation when it comes to constraining would-be dominators. We already have some of these limited powers of contestation; the republican strategy is to make them stronger, and make them much broader. The focus on contestation as a check on domination is one of Pettit’s many contributions to the literature – his own approach involves constructing systems of ‘contestatory democracy’ (Pettit, 2000) at the level of the state, an ‘editorial’ dimension to supplement the ‘authorial’ one provided by traditional, electoral democracy. But republicans argue for these contestatory measures at all levels.

Even those republicans who lean heavily on contestatory mechanisms acknowledge some problems with this approach. Pettit himself writes that ‘the act of triggering them can have costs of its own’ (Pettit, 2012: 115). There are costs to raising the alarm. Basic, everyday costs of time and energy, but more serious costs as well. As long as your dominator has some uncontrolled power over you, raising a complaint risks retaliation – ask any tenant who has tried to make their landlord abide by minimum housing standards. Whether that retaliation is legal or illegal, formal or informal, will often make little difference to the person at the receiving end. So, while people might not like domination or the interference it involves, they may tolerate low levels of it rather than trigger potentially costly processes of contestation.

Contestatory mechanisms won’t always work to root out instances of domination in decisions that the victim doesn’t care about all that much. It is tempting to regard this as a shame, but not a crying one. If contestatory mechanisms fail in this way, then at least they self-select to fail only for the least bad forms of domination.

Here micro-domination rears its head again. Some relationships of domination will exhibit the following structure: no power of interference into any individual decision will be so bad that the victim is prepared to bear the costs of contesting that decision, but the power of interference into the set of such decisions will cause significant subjective harm to that victim. D might be able to interfere in lots of decisions that V would otherwise be in a position to make – which, in the long term, V cares about very much – and V may have the ability to contest any one of those decisions in a neutral forum. But if she only has the opportunity to contest individual decisions (rather than D’s power as a whole), or if a more general process of contestation can only be activated after challenging some individual decision, and if none of those individual decisions are themselves subjectively bad enough for V to contest, then she falls through the cracks.

This weakness is not just limited to formal mechanisms of contestation. Alex Gourevitch discusses a similar issue with relying upon the power of workers to leave their jobs. In broad strokes, some republicans place a great deal of value on the threat of exit to constrain the dominating power of employers: at least with a suitable welfare state (or basic income), workers can informally contest any decision their employer makes by threatening to leave (Pettit, 2008; Taylor, 2017). But, as Gourevitch (2013: 608) writes, ‘threatening to leave a job is a kind of nuclear option that is simply not credible in many low-level disagreements’. Threats – at least explicit threats – normally have to be made at a specific time and in response to a specific decision. But if what really gets to you is the accrual of petty, insignificant intrusions of your boss’s will into your daily life, then it may never seem like the right time to trigger such a significant kind of contestation, one that might well backfire. This provides an obvious example of the harm of micro-domination: we have domination and we have a victim of domination who is subjectively harmed and would rather not be dominated. This is not a thorny case of a monk or soldier who appears to consent to dominating structures of authority – the worker merely acquiesces to a situation they wish they could change. Of course, a worker in that position might snap, deciding one day that they want to contest the pattern itself rather than any individual instance of interference, cost be damned. But a strategy that relies upon that happening will not reliably challenge dominating relationships of this kind.

Recognising that micro-domination poses a problem in this sense doesn’t close the door on contestation as a means of reducing domination, but it does make some forms more attractive than others, and should shape the use of individual contestation in republican theory. Snap inspection oversight begins to look more efficient in comparison to fire-alarm oversight, so supplementing the latter with the former is more attractive. It is also possible to lower the cost of raising the alarm. An entirely general way to do so is to make people less dependent upon dominating relationships, a strategy which has the side-effect of itself making those relationships less dominating. If someone’s basic needs are guaranteed even if she divorces her wage-earning husband, then the costs of contesting the power structure within that relationship are reduced; if I am guaranteed healthy unemployment benefits or a basic income, it will be less potentially costly to sour my relationship with my manager by triggering a process of contestation. Some more specific changes are obvious enough as well. Anonymous reporting is less costly than having to present one’s case in full view of the accused, and so less individually consequential decisions will be easier to contest. Equally, it is not free to organise forums for contestation: the cost can be borne by the claimant, the respondent, the state or a mix of any of the three parties. Making the claimant pay anything more than a token amount – that is, literally raising the cost of contesting your relationship with those who hold power over you – will raise the bar above which it is subjectively worthwhile for them to bring a case. In the context of employment tribunals, the UK Supreme Court has agreed, finding that the sharply raised cost of bringing a case to a tribunal since 2010 ‘has had a particularly deterrent effect on the bringing of claims of low monetary value’, and that claimants were ‘effectively prevented from having access to justice’ (UNISON, R (on the application of) v Lord Chancellor [2017] UKSC 51, 2017).

One of the reasons that micro-domination risks being missed by processes of contestation is that those processes are almost always available only in response to specific decisions (Hsieh, 2005). This sort of arrangement reduces domination where it allows would-be victims of domination a measure of after-the-fact invigilation of those decisions, and thereby a measure of influence over decision-making in an organisation (in the general sense: a firm, a family, a state, etc). If any decision might be contested and overturned, then decision-makers will have to take that into account in their decision-making. It requires us to move from particular decisions to the power structure that produced those decisions. It is this feature that makes it particularly ill-suited to dealing with micro-domination, where the individual decisions involved are subjectively inconsequential.

An alternative would be involvement in, rather than after-the-fact invigilation of, the decision-making process – democracy, rather than contestation. In this way, paying attention to micro-domination can intervene usefully in ongoing practical debates within republicanism. To return to the workplace for an example, workplace democracy does not rely upon moving from contesting a particular decision to contesting the structure that produced it. Instead, it directly reshapes that structure, with workers participating in rather than merely influencing decision-making through ‘partial’ and ‘pseudo-’ participation, to use Carole Pateman’s (1970) terms. Of course, democracy isn’t a panacea when it comes to reducing domination, and may even still risk some informal issues of fire-alarm-style costs to participation: do you really want to bring up your idiosyncratic concerns in this meeting, or will you acquiesce on this issue to build goodwill for that other, more important decision? But democratic organisations, especially in more participatory forms, ⁸ are much better equipped to avoid overlooking micro-domination. If workers object to some position in the firm acquiring a stifling set of individually petty powers, for instance, they will not choose to empower that position in that way. There may well still be day-to-day coordinative authority in a democratic workplace, but just as at the level of the state, democracy at its best can provide a government that interferes with its citizens ‘only under their equally shared control and so without domination’ (Pettit, 2012: 281). Even fairly weak, representative forms of workplace democracy will still give workers the regular, anonymous and extremely low-cost power to electorally challenge managerial decisions in a process that isn’t triggered by objecting to any particular decision.

Contemporary republicanism can seem relatively hostile to democratising, rather than invigilating or providing contestation within, organisations at levels lower than the state. There are indeed some recent republican proponents of workplace democracy: some are friendly critics of mainstream, Pettit-style republicanism (González-Ricoy, 2014), others are less friendly critics (Casassas and De Wispelaere, 2016; Gourevitch, 2013; O’Shea, 2020). Several more theorists in conversation with republicanism recommend workplace democracy, but on the more complex basis that non-domination ought to be supplemented by some other goal – non-alienation or resistance (Watkins, 2015), or self-direction (Breen, 2015) – and that workplace democracy is the best strategy for pursuing those combined goals. But as discussed above, Pettit and Lovett, representing a broad strain of mainstream republican thought, both focus on making it easier for workers to leave a workplace (functioning as an informal system of contestation), rather than participate in its governance (Lovett, 2012; Pettit, 2008), and the most sustained attempt to apply Pettit’s republicanism to the workplace limits itself chiefly to internal, post-hoc, contestatory measures (Hsieh, 2005, 2008).

Against this backdrop, the fact that micro-domination is poorly addressed by contestation – formal or informal – should give us another reason to doubt that exit and contestation can on their own adequately tackle domination in organisations like workplaces, and another pro tanto reason to explore the democratisation of those organisations as a means to promoting non-domination.

Conclusions

Micro-domination matters, by any account of what it might be for domination to matter. But it is slippery in the sense that it presents special problems when we try to grab hold of it; it can fairly easily escape attempts to spot or reduce the kinds of domination that matter, and the strategies that work for more spectacular forms of domination risk failing in cases of micro-domination.

The way that I have described micro-domination is not as a challenge to mainstream conceptions of domination, then, but as a challenge to certain common ways of putting any conception of domination to work – those which make use of either objective or subjective thresholds. Avoiding the use of thresholds is almost certainly impossible when thinking about domination. But micro-domination gives us a reason to be wary of them, and adopt strategies – only a few of which I have sketched out above – to minimise the chance that they will systematically blind us to a whole class of dominating relationships.