The Aptness of What We Do Together

I. THE PROBLEM

The problem of appropriate piling-on is a trilemma of claims, all of which seem attractive at first, but cannot be jointly true. Here are the three claims:

• i. It is appropriate for each person (identically situated) among many to blame a person who is responsible for wrongdoing.

• ii. If it is appropriate for each person (identically situated) among many to blame a wrongdoer, then it is appropriate for everyone to blame that wrongdoer, regardless of how many blamers there are.

• iii. It is not appropriate for everyone (identically situated) to blame a wrongdoer if the number of blamers is sufficiently high.

Let me explain a bit further the initial attraction in each idea.

I take it that i is the least resistible claim. What counts as a blaming response is, as I just highlighted, a vexed issue, and we will survey a few possibilities shortly. But suffice to say for now: very few deny that at least some form of blaming response to wrongdoing is appropriate.⁹ Some theorists go so far as to argue that blame’s being appropriate is part of what constitutes an act of wrongdoing.¹⁰ Whether or not we ought to go that far, the fact that people take that view seriously indicates a deep relationship between wrongdoing and the aptness of blaming.¹¹

But why believe that claim ii is true? How might the idea that blame is apt for each person support the idea that it is apt for the collective, regardless of how many people in that collective? It is not in general odd to think that other people doing something can affect the morality of my doing the same. The fact that others will rescue a drowning swimmer might make my rescuing the swimmer less urgent. The fact that others contribute to climate change might make my own contribution more urgent. The fact that others have tidied the kitchen might make it more incumbent on me to tidy the kitchen. And so on.

But the aptness of reacting to wrongdoing seems to work differently. The aptness of one person’s response does not normally change just because others respond the same way. Suppose, for example, that Zara is responsible for wrongdoing, and that Ali blames Zara. By itself, this makes no difference to whether it is apt that Bea blames Zara. Assuming that Ali and Bea are identically situated, the fact that another blamer exists does not seem to alter what is apt for Bea to do. Call this feature of aptness addition-resistance.

Addition-resistance suggests that the same reasoning applies to each additional blamer. Adding a third identical blamer makes no difference to our evaluation of the first two, or to all three blaming. Adding a fourth makes no difference to our evaluation of the first three, or to all four blaming. The natural conclusion: adding an nth identical blamer makes no difference to our evaluation of n-1 blamers, or to all of n’s blaming. After all, these people are all responding to the same facts in just the same way. It is unclear why what is true of additions to small values of n would not also be true of additions to big values of n.

However – and this is the problem – claim iii calls that reasoning into question. When the numbers are immense, e.g. in the millions or billions, I doubt that many would consider the practice appropriate. Intuitively we do distinguish additions to small groups of blamers from additions to massive groups (and perhaps equivalently, between small additions and massive additions).

So what gives? Must we abandon that intuition about very big numbers (iii)? Must we abandon addition-resistance (ii)? Or, must we abandon the idea, at least when it comes to mass pile-ons, that blaming is an apt response for each person to begin with (i)?

Before outlining this essay’s response, I should call to attention the distinctiveness of this problem. Much philosophical discussion is had about situations in which individual contributions to a series of acts seem to differ in their moral valence from the series of acts as a whole. But our task is not as simple as exporting conclusions from those discussions to the case of piling-on. The latter has special features.

One puzzle in the ethics of killing may appear analogous. Millions of armed assassins queue up to wrongly kill an innocent person.¹² The target seems justified in killing, in defence of their life, each wrongful aggressor appearing at the front of the queue – each individual defensive act seems objectionless. Taken together, however, these defensive acts result in millions of deaths for the sake of a single life – the series as a whole looks objectionable. But unlike piling-on, the queue of aggressors is a one-to-many situation. In these situations, the series of acts (defensive killings) that seem individually justified are performed by a single agent (the innocent target), affecting multiple patients (the many aggressors). Appropriate piling-on is, conversely, a many-to-one situation. Multiple agents (the many blamers) do the act that seems appropriate (blaming), affecting a single patient (the person blamed, hereafter the blamee).

The reason appropriate piling-on is distinctive is not merely that it has a many-to-one structure. A famous set of examples mirroring that structure, Derek Parfit’s torturers, involves many people raising the current in an electrical torture device. Suppose the device is hooked up to a single innocent person. Each person presses a button in order to raise the current by a tiny, imperceptible amount. All of those people pressing together results in the victim’s agonising torture.¹³ But this many-to-one case is also disanalogous. The difficulty with the harmless torturers lies in explaining the idea that each torturer acts wrongly, despite having an imperceptible effect on the victim. In the case of appropriate piling-on, it may be true that each blamer’s contribution is imperceptible, given that so many others are contributing in the same way. But I take it one is less confident that each blamer acts wrongly. As I said, we might rather think that in holding someone accountable with accuracy, each blamer does something positively apt. Our task is to explain the different reaction to many people doing the same seemingly apt thing.

Once more, the obvious route out of the problem is to abandon at least one of its constituent claims (i-iii). In what follows, I reject claim ii. I will offer an explanation of why it is that the aptness of each person’s blaming needn’t go hand-in-hand with the aptness of everyone’s blaming.

II. THE PROPORTIONALITY VIEW

My argument against ii is based on the Proportionality View introduced at the outset: the view that blaming a wrongdoer is inappropriate when and because the harms it involves are unjustified relative to the value of blaming that person. Let me begin by clarifying why this view is grounds for rejecting ii.

Suppose, as will be argued later, that for a given act of culpable wrongdoing, there is a maximum amount of harm one can appropriately inflict on the wrongdoer by blaming them. High numbers of people blaming can together harm that wrongdoer in amounts that are unjustified relative to the value of blaming that person at all. If those amounts are reached, the practice as a whole is inapt, according to the Proportionality View. However, in those same situations, it is also possible that each person’s individual blaming act does not involve a level of harm that would be unjustified relative to the value of blaming the wrongdoer. So, each person’s blaming may not be inapt in that sense. It is therefore possible both that each ‘piler-on’ blames aptly and that all ‘pilers-on’ together blame inaptly. And that is the possibility that claim ii rules out.

Notice that judgments about appropriateness in these contexts seem to correspond to the amount of harm one would reasonably expect to befall the blamee. Let us start, again, with a single blamer: suppose Ali accurately blames Zara in response to Zara’s culpable wrongdoing. Bracketing the blame sceptics set aside earlier, most are likely to think that Ali behaves aptly. However, this is not because we imagine Ali’s blame to be harmless. To the contrary, many advocates of blame point out how it carries a ‘sting’.¹⁴ The point is that the sting of Ali’s blame can, in some way or another (more on possible ways later), be justified in the light of whatever accounts for the value of blaming Zara.

Now add the second blamer, Bea, to the set. The addition of Bea is not of major concern even if we would expect Bea’s blame to come at some additional harm to Zara. Again, we might be able to explain this by appeal to the value of blame – the value of blame is such that it can often justify harm done by adding a second blamer.

Now add a thousand blamers. At this point, one pauses before judging the thousandth blaming act appropriate. The reason I at least pause, though, is not necessarily because thousands of people blaming Zara is entirely devoid of value. Rather, the value achieved by these thousands of acts fails to justify the harm I would expect to be at stake.

To make this less abstract, think about some concrete harms and values. Often accompanying the initial shock or ‘sting’ of hostility and judgment are further distresses such as guilt, embarrassment, anger at, and disappointment in, oneself. Other harms, salient on online platforms, include social disesteem or exclusion by others. Disesteem and exclusion can result in downstream burdens, but their concomitant harms can also be more direct: consider anxiety about the possibilities of social exclusion or loss of esteem, and about whether one is the kind of person who merits these negative social reactions.

Turning to values, one set widely associated with blaming are motivations in the blamee to remedy their fault based on a moral understanding of that fault.¹⁵ Another set of values, salient in public fora, include the expression and display of respect and solidarity towards victims of wrongdoing.¹⁶ A third involves victims themselves exercising a voice, having a platform publicly to share their account of moral injury, and to demand the responses that wrongdoers owe them.¹⁷

Those are weighty goods! They certainly seem worth achieving at significant cost. Suppose, then, that respect and solidarity towards victims, or motivation to make amends and reform, inevitably come at non-trivial levels of guilt, anxiety, and loss of face. That may be a price worth paying.¹⁸ This, I think, is a key reason why the harms expected of paradigm one-to-one blaming interactions do not threaten the aptness of those interactions. There may be other reasons too: some believe that the harms of blaming are themselves valuable, e.g. because Zara deserves blame as someone responsible for wrongdoing.¹⁹ Others believe that harms are a necessary part of something valuable, e.g. that guilt is necessary for moral understanding, or social exclusion is necessary to show respect for victims.²⁰ But these are routes to the same conclusion: the values of blaming are standardly able to justify the harm it does.

Furthermore, those same values are also able to justify multiple blaming responses to the same instance of wrongdoing. To be sure – when more people blame, the harm requiring justification tends to increase. Not only can additional blaming interactions each contribute their own ‘sting’, as discrete psychologically distressing experiences; they can also compound the harms of a single interaction. When many rather than few blame, threats to social standing, for example, will seem and become more real. At the same time, however, the addition of blamers may enhance some of the values outlined above. The fact that multiple people confront a wrongdoer about their wrongdoing can underscore the seriousness of that message to the wrongdoer. Multiple exhortations and reminders of a person’s need to reflect on their behaviour and respond accordingly can provide more powerful motivation than a single interaction. We might also think that solidarity to victims is better communicated when several members of a community speak out.

Albeit enhanced in these ways through multiple interactions, the values just described are hardly sacrosanct. There are, surely, restrictions on our ability to appeal to them to justify a practice we engage in.

Consider four restrictions familiar to wider ethics of harm, but also to the ethics of public shaming.²¹ First, the blamee might show that they have gained pertinent moral insights and motivation, for example through recompense and willing dialogue with those calling them to account. As the values sought have already been secured, they do not justify additional harmful blaming. Additional harms are in this case gratuitous. A good example is the old saying – ‘Once a meme, always a meme.’ A ‘memed’ person continues to face distressing critical content no matter how much apology, reconciliation, reform, and so on, has taken place.²²

In other situations, the harm done might be counterproductive. The blamee may become too intimidated or traumatised to absorb the message, or to engage in the moral self-reflection required of them. Self-harm in response to online shaming is a more extreme though equally recognisable instance.

Third, the burdens may be of such a magnitude that reform and respect for victims cannot support their imposition. Here, the practice is straightforwardly disproportionate. Long-term harm to mental well-being, including self-harm, are often disproportionate in this way, as are some levels of socioeconomic damage. (Confronting someone about their minor breach of etiquette isn’t worth blacklisting them from future job prospects.)

A blaming practice is not off the hook even if it manages to be productive in achieving worthwhile goals, and even if it does not involve pointless further intervention once achieved; for those same goals might have been achieved in less destructive ways. Suppose that in order to understand the seriousness of their offence, it is sufficient for a culprit to face the consequence that others will refuse to support them until they apologise and explain themselves. Additionally exposing them to unsettling threats of violence is unjustified as unnecessary harm.²³

The Proportionality View holds that if any such objection applies – if its harms are not justified relative to its value – blaming is, for that reason, inapt. I have just been illustrating why this view is grounds for rejecting claim ii in the light of how the addition of blamers affects both harm and value. In the light of those effects, proportionality restrictions may apply to harm caused by multiple rather than single contributions. Adios addition-resistance.²⁴

The foregoing is only a sketch of the thesis, to be developed further below. Natural question-marks will arise about how to explicate the proposed relationship between aptness, on the one hand, and the harms or goods of blaming on the other. For example, some would argue that it can be apt to blame harmfully even if this does not secure reform, as when we blame psychopathic or other morally hardened people incapable of reform. Others might argue that it can be inapt to blame harmfully when this does secure reform: it seems inapt, for example, to pile more harm on a wrongdoer to incentivise lots of other people’s reform.²⁵ How, more generally, are we to separate the question whether blaming is apt, or inapt, from whether its outcomes reveal it to be a good or bad idea?

We will revisit these concerns in earnest later on.²⁶ Another question mark to be raised about the Proportionality View is that it is not the only possible account of our problem. Some may be instead inclined to reject claim i, to reject claim iii, or indeed to reject claim ii on grounds other than those adduced here. Let us next consider these alternative strategies.

III. AGAINST THREE ALTERNATIVES

IV. APTNESS AS INTERNAL VALUE

So far, I have outlined how the Proportionality View explains the problem of appropriate piling-on and argued that it does so more appealingly than some other explanations. But what explains the view itself? Why surmise, in the first place, that blame can be rendered inapt by the harm it does? To answer this, I now offer a general account of what determines blame’s aptness, and explain some of its virtues.⁴¹

V. APTNESS & OUTCOMES

The aptness of blaming obviously comes apart from its good and bad effects. We raised the examples of inapt blaming to excellent reformative effect (piling-on to make an example of one wrongdoer in order that many others reform), and apt blaming with little to no reformative effect (blaming psychopaths or the morally hardened). How does the argument given handle these problems?

First recall the idea that blame’s internal value is limited by the activity’s minimal description – ‘A blames Z for x-ing’. Apt blaming is thus not a matter of just any good effect: the goods to be effected within that structure must relate closely to A, Z, and those closely connected to Z’s x-ing. Although reform of other wrongdoers is a good outcome of blaming, that outcome does not relate closely to the blamee’s understanding of their own reformative duties. To put it differently, reform of people agentially distant from Z’s x-ing is another instance of an external value of blaming, located outside of the activity’s basic structure. As such, it does not contribute to aptness on the Internal Values Account.

Second, the account allows that aptness is multidimensional. Because blaming has more than one internal value, the practice might be defective with respect to one such value while fulfilling another. Thus, while blaming the morally hardened might be defective in securing their reform, it might nonetheless instantiate other values like the blamer’s understanding of their wrongdoing, and acknowledgment of the claims of victims who belong to the history of that wrongdoing. Blaming the morally hardened can be apt in these latter respects despite being inapt in the former.

That second element of the account sets it apart from fittingness accounts. Fittingness accounts hold that aptness is a ‘relation of fit’ between a particular response, on the one hand, and the facts to which it responds, on the other. For example, it is a fact about Mount Everest that it is beautiful. Responses which ‘fit’ that fact – e.g. smiling with wonderment at the way the sunlight reflects off the crevices in the rockfaces – are apt. Responses that do not ‘fit’ – e.g. recoiling with disgust – are inapt. Or consider a funny joke. Sobbing inconsolably is inapt because it does not ‘fit’; laughing is apt because it ‘fits’. Whether something is apt, on this view, is just a question of whether it constitutes an act that ‘fits’, rather than about outcomes. Applied to blame, suppose one holds that some response, e.g. disappointment or anger, ‘fits’ a responsible act of wrongdoing. In that case, whether blaming is apt is just the question of whether it is a kind of disappointment, anger, etc. The good or bad outcomes of that response are neither here nor there. This in part is why Amia Srinivasan holds that counterproductive anger can be apt.⁴³ Srinivasan argues that appreciation of wrongdoing is good in itself, regardless of what this does to redress wrongdoing.

Some might have worried that relating aptness to harmful outcomes comes at the cost of denying that insight – that certain reactions are apt just because they constitute appreciation of moral violation. However, the Internal Values Account concurs that appreciation of wrongdoing is an internal value. It only adds that aptness is multidimensional. A blaming practice can be apt in instantiating each person’s appreciation of wrongdoing. It can, at the same time, be inapt when it comes to instantiating other internal values, for example the blamee’s shared appreciation of that wrongdoing, and their redressive motivation based on it.

This point helps to substantiate my contention in Section III that beyond adopting fitting attitudes, it can also be apt to seek felicitous communication of those attitudes.⁴⁴ There is indeed reason to view both values as complementary. Appreciation of wrongdoing is after all incomplete without appreciation of the duties of reform that wrongdoing creates.⁴⁵ Something would be amiss if a blamee, upon gaining that appreciation, lacked any motivation to make amends. But, for the same reasons, it also wouldn’t be ideal if a blamer appreciated an instance of wrongdoing with nil motivation to ensure that this understanding is mutually recognised and acted upon, by urging the blamee to apply their agential capacities to its subject matter. To be sure, whether communication is ultimately felicitous in gaining that uptake is a question about the outcomes of accountability interactions. But these outcomes derive some of their value from the fact that they are brought about by a joint agential procedure aimed at mutual understanding.

In the light of this discussion, let’s return to the general challenge: What is the difference between blaming’s being apt and its simply being a good idea in view of its expected outcomes? Whether blaming is apt depends only on the standards for realising internal values. As noted, some of those values relate to outcomes, but others are procedural. In contrast, whether blaming is a good idea depends on its full range of expected outcomes. The full range can involve internal values (at least those that are outcome-related) but also external values, as well as any other normative feature of outcomes.

It follows that doing something apt can sometimes be a good idea – when and because instantiating internal values happens also to be recommended by its full range of expected outcomes. But that co-occurrence is never guaranteed. It could always be a good idea to do something inapt – because the full range of outcomes could always recommend doing something other than instantiating internal values.

VI. INTERNAL & EXTERNAL PROPORTIONALITY

Having worked through an account of the relationship between the aptness of blaming and disproportionate outcomes, we might be left wondering how it differs from prominent accounts of proportionality in online criticism.⁴⁶

Billingham and Parr in particular discuss proportionality in detail. They focus on justification conditions of public shaming, arguing that shaming is unjustified if it is (inter alia) counterproductive, or unnecessary, or causes harms outweighing its goods. The Proportionality View holds that the same objections to a blaming practice can make it inapt.

The two accounts appear alike if we subscribe to the idea that blaming and shaming can do similar harm. I suggested in Section III.A that blaming can do harm even if each blamer intends not to shame but sincerely to call someone to account. I suggested in Section III.C that blaming can do harm even if each blamer does not communicate with the wrongdoer. There’s no reason to exclude these harms from the Billingham-Parr proportionality conditions of justification. It might seem, therefore, that all blaming practices condemned by those conditions are inapt, and vice-versa, that all blaming practices condemned by the Proportionality View are also unjustifiable.

But the two views in fact differ extensionally. Whereas the Proportionality View is indexed to a subset of values, the Billingham-Parr view is not. The Proportionality View constrains the pursuit of internal values only. It sets, if you like, a standard of internal proportionality: which determines whether a practice’s ills are morally supported from the inside; or supported by the point or purpose of the practice. The Billingham-Parr view constrains the pursuit of the full range of possible outcomes of public criticism. That, I noted in the previous section, is a much wider-scope constraint, involving both internal and external values, as well as other normative facts affecting the overall justification of criticism. We can call it a standard of external proportionality: a standard that determines whether we ought to tolerate a practice from the outside, whether or not it is morally supported from the inside.

Billingham and Parr carve out a narrower standard that might again appear closer to internal proportionality. They distinguish between narrowly and widely proportionate criticism. Narrowly proportionate criticism harms people less than or as much as they are liable to be harmed – its harms needn’t infringe anyone’s rights. Widely proportionate criticism harms people beyond the extent to which they are liable: while that criticism infringes people’s rights, it is proportionate in the sense that the goods achieved by infringing those rights are important enough to justify doing so.⁴⁷

But this distinction between narrow and wide proportionality in blaming also comes apart from the distinction between internal and external proportionality. Were they the same distinction, all (and only all) harm that meets proportionality standards for appropriate blaming is harm that a blamee is liable to suffer: harm that does not infringe the blamee’s rights. Yet that is not true, at least according to the argument of this essay.

Granted – much harm that we do to a person through appropriately blaming them is harm that the blamee is liable to suffer. That much is true for the simple reason that wrongdoers incur duties to be accountable for their wrongdoing in virtue of having committed it. These are the duties I’ve been emphasising as part of internal value – to reflect on what one did, and to be motivated by that reflection to meet additional duties like apology, compensation, and reform. A wrongdoer’s rights are not normally infringed by the costs of being blamed in the service of those duties.

However, responsibility for wrongdoing can also make a person liable to be harmed for non-blaming reasons. Suppose, for example, that a person crushed by a dogpile has been harmed to a much greater extent than is necessary for them to achieve moral understanding and redressive motivation, and for a community to speak out for the victim. Suppose, though, that this already overblamed person has yet to compensate their victim. Coercive sanctions might then be narrowly proportionate despite being internally disproportionate. Forcibly exacting compensation from the wrongdoer might not infringe their rights, for example, yet this fact would not undo the proportionality defects within the blaming practice itself.

Last, suppose that a person is overpunished for their wrongdoing: they have, let’s say, served most of a wildly excessive prison sentence. But suppose that until now, this person has never been called to account by the victim of their wrongdoing. It seems to me that this victim could be vindicated by internal value in arranging a visit to exhort the wrongdoer to apologise and explain themselves, even if that process heaps more costs upon the already overpunished blamee. So, it seems, a person might be blamed a way that satisfies internal proportionality, but not narrow proportionality. Though the overpunished blamee’s rights not to be harmed have long been infringed, the blaming practice remains felicitous as the proportionality defects are located outside of it.

In sum, proportionality in apt blaming determines how well-performing our blaming practice is; whether it is a defective, or morally corrupted, version of itself. That is how its implications differ from conceptions of proportionality often applied to content moderation and online criticism. Those conceptions are about the wider justification of these practices, independently of whether they have been corrupted in some way.

Some might wonder – how significant is the internal category we have just carved out?⁴⁸ External proportionality calculations, which count both internal and external values, seem likely to dominate internal proportionality calculations in determining how to blame people together. Does this mean that internal proportionality is normatively trivial or inert? Not necessarily. After all, a good deal is morally at stake within our accountability practices. Wrongdoers have powerful obligations to be answerable to others, and to deliver certain responses. The wider community has powerful obligations to acknowledge the facts about wrongdoing and the interest of victims. Given the importance of those things, we have at least some presumptive moral reason to honour them: to strive for their felicitous pursuit. No doubt external considerations will often compel us to do otherwise. But at a cost.

VI. CONCLUSION

Time to take stock of the main moves of this essay, by reminding ourselves of the problem it began with:

• i. It is appropriate for each person (identically situated) among many to blame a person who is responsible for wrongdoing.

• ii. If it is appropriate for each person (identically situated) among many to blame a wrongdoer, then it is appropriate for everyone to blame that wrongdoer, regardless of how many blamers there are.

i and ii preclude:

• iii. It is not appropriate for everyone (identically situated) to blame a wrongdoer if the number of blamers is sufficiently high.

But iii is plausible too.

I have argued that ii is the least plausible, because apt blaming is subject to internal proportionality standards: standards which multiple agents together can flout. As we can make sense of harm aggregating across many people’s acts, internal values can lose vindicating force through harm that no single piler-on inflicts on their own. The argument thus leaves room for the possibility both that each piler-on’s contribution is apt, and that everyone piling-on is inapt.

Albeit a widget for closing that particular door, I fear that last possibility opens a new one. When it obtains – when it is true both that each blames well, and that everyone blames badly – one dimension of internal value is secured at the expense of another.⁴⁹ Each piler-on’s sound blaming performance undermines everyone’s sound blaming performance. This poses a problem about whether to prioritise individual or collective conduct. On the one hand, all of the pilers-on morally corrupt the practice. Yet on the other, no piler-on’s behaviour is corrupted, and accountability goods are not well-advanced if none of them blames. Is it most apt, in these situations, that everybody piles-on, or that nobody does? I submit this question for future inquiry, in the vague hope that collective action theory might help resolve it.

ACKNOWLEDGMENTS

Thank you for brilliant comments to Chris Mills, Massimo Renzo, Victor Tadros, two anonymous referees and the Editor of this journal, as well as seminar discussants at the Department of Applied Philosophy & Ethics, Institute of Philosophy, Academy of Sciences, Prague. This work was supported by the National Research Centre on Privacy, Harm and Adversarial Influence Online (REPHRAIN); and the Lumina Quaeruntur Grant of the Czech Academy of Sciences (LQ300092001).

COMPETING INTERESTS

The author declares that there are no competing interests.