This is the second part in a brief series of posts I’m doing on Ted Poston’s article “Social Evil”. The article is a contribution to the ongoing debate about evil and the existence of God. Traditionally, that debate has revolved around two categories of evil, natural and moral. Poston argues that there is a third category — social evil — which is distinct from these other two.
His argument for the distinctiveness of social evil was covered in part one. As we saw there, social evil arises out of collective action problems, such as those described by game theory. The example used in part one was a variation on the tragedy of the commons, involving a water shortage in LA. In the “game”, non-reduction in water use is shown to be an individually rational (maybe even moral) strategy, despite the fact that this will lead to a great deal of collective suffering. This suffering is not natural, because it is the result of human action; and it is not moral, because no individual is morally responsible for the negative outcome. Hence, it is a distinct type of evil.
Accepting that it is distinct, we now need to do three things. First, we need to see how social evil fares in light of three traditional responses to the problem of evil (natural laws, soul-making and free will). Second, we need to see how robust the phenomenon of social evil really is. And third, we need to consider social evil in light of an Edwardsian theodicy, which is the one Poston thinks most promising from the theistic perspective.
1. Social Evil and Three Traditional Responses to the POE
The identification of a new category of evil is not, in and of itself, worrying for the theist. At least, it is no more worrying than the traditional categories of evil. After all, theists have ready-made answers to the problem of evil, which they can surely trot out in response to the problem of social evil, right? Not necessarily. Poston considers three classic responses to the problem of evil and argues that they fail to account for social evil. They are: the natural laws theodicy; the soul-making theodicy; and the free will defence.
The natural laws theodicy — defended in slightly different forms by Swinburne and Reichenbach — holds that the existence of predictable and reliable natural laws, which may on occasion lead to suffering, is justified because such predictability and reliability is necessary for a range of great goods. For Swinburne, a stable set of background laws is necessary if we are to acquire the kind of knowledge necessary for moral responsibility (i.e. if we are to meet the “epistemic condition” for responsibility). For Reichenbach, a stable set of laws is necessary if we are to become free and sentient creatures (this is a more complex argument, one that I cannot summarise well in a short space).
Poston only deals with Swinburne’s version of the theodicy. His response is brief and pithy: social evil arises from collective acts of human agency, not from natural laws. So it is not clear that the natural law theodicy even begins to cover social evil. I think this is right, insofar as it goes. One thing I would point out, however, is that social evil is, effectively, a mathematical property of collective strategic interaction. The negative outcome in the tragedy of the commons is an equilibrium point — a mathematical “solution” to the particular game. Given this, a theist might be able to argue that social evil is a necessary by-product of a particular kind of human agency. And since God does not have the power to prevent such necessary by-products, we do not need a theodicy to account for it. The only problem with this response is that God probably did have the power to set the initial agency conditions which give rise to this by-product. Maybe there’s something worth exploring there.
The soul-making theodicy will be familiar to most. As proposed by John Hicks, the soul-making theodicy holds that a certain amount of evil is justified if it allows us to cultivate the virtues necessary for achieving communion with God. For example, to achieve compassion you must first suffer, to achieve tranquility you must overcome adversity, to attain courage you must first experience injustice. The list of examples could multiply.
Whatever about the good of soul-making, and the “necessity” of suffering in achieving it, there is an initial hurdle when it comes to social evil. Soul-making is an individual good: it is the individual who must overcome adversity, experience suffering and learn from their moral failures. But social evil, as described and defended by Poston, is a property of collective behaviour. What’s more, one of the key features of it is that individual action makes no moral difference to the collective outcome. Thus, it is difficult to see how individual moral growth is made possible by such dilemmas. Still, Poston isn’t completely dismissive. The Edwardsian theodicy, which will be addressed below, can be viewed as a type of soul-making theodicy, and Poston deems it worthy of serious attention.
The free will defence is, of course, the most popular and famous response to the problem of evil. It holds that free will is a great moral good, but that it has, as a necessary by-product, the (at least) occasional occurrence of evil due to its misuse. While there are general problems with this as a defence to the problem of evil (particularly in its evidential form), it is hard to see how it even gets off the ground when it comes to accounting for social evil. It is true that social evil is brought about by the aggregation of many freely-willed actions, but it is false that it is itself maliciously freely-willed. As discussed in part one, each individual can act rationally and morally from their own perspective, and yet this can still lead to collective suffering. So it’s not a by-product of the misuse of free will.
(Note: this is Poston’s response. I wonder whether my earlier comment about social evil being the necessary by-product of a particular kind of human agency might have some relevance to this debate too? I leave it as a suggestion.)
There are more fanciful arguments that can be made on behalf of the free will defence. For instance, one could follow Plantinga and argue that, although social evil is not explained by the malicious freely-willed actions of human agents, it is explained by the malicious freely-willed actions of supernatural agents. But this is an ad-hoc and explanatorily redundant suggestion. The nature of rational incentives, and the logic of game theory, give us everything we need to explain social evil. We have no need to postulate supernatural agents. Hence, this does not seem like a plausible argument.
If all this is right, then social evil seems to present a unique challenge to the theist. Is there any other way to account for it? Possibly, but before we see how we’ll briefly consider the “robustness” of the phenomenon.
2. The Robustness of Social Evil
In the example given in part one, social evil arose from the individually rational and morally blameless behaviour of many agents. This is the paradigmatic case of the phenomenon. But as Poston argues, the phenomenon is reasonably robust and can arise in non-paradigmatic cases. In particular, it can arise even if the individual behaviour underlying the social evil is blameworthy.
The argument is as follows. Moral evil arises from the misuse of free will. But to be classed as moral evil there needs to be some appropriate responsibility-relation between the agent performing the act that gives rise to evil, and the evil itself. To put it more simply, the blameworthiness of the agent’s action must account for the evil that actually results from that action. If the evil that arises is disproportionate to the blameworthiness of the action, then it is not truly an instance of moral evil. But, of course, this is what happens in the case of collective action problems: individual behaviour can be minimally blameworthy, but the collective consequences of this behaviour can be morally devastating.
Using Russell Hardin’s book One for All as a guide, Poston gives the example of conflicts that are based on in-group out-group labelling (or “norms of identification”). In these cases, tremendous suffering can arise, but this suffering is disproportionate to the blameworthiness of the individual behaviour. The reasoning is (roughly) as follows:
Being identified with a group can have personal benefits (e.g. being recognized as a member of the aristocracy in 17th century England), but identification brings with it norms of exclusion.
Norms of exclusion can fuel serious and escalating social conflicts due to competition between groups.
The individual decision to identify with the group (and not lose that identification) is blameworthy, but not so bad as to account for the tremendous social conflicts that arise.
So, there is still social evil in these cases: individually blameworthy acts give rise to a disproportionate amount of group suffering.
Poston gives a slightly more detailed example of this — again based on an account found in Hardin’s book — from on the Croation War of Independence. Interested readers might like to check this out.
I guess I can appreciate where Poston is going with this, and, of course, I sympathise with the overall project. That said, I do wonder whether this stretches the notion of social evil ever so slightly. It seems to me that cases like this just mirror those in which minimally blameworthy conduct gives rise to disastrous unintended consequences. Such cases can arise even when there are no social/strategic interactions. For example, I might rush a drug to market, believing it to be wholly safe based on existing trials, because I am eager to make some profits. This is not morally commendable behaviour, to be sure. But if it turns out that the drug had unknown and devastating side effects, that could not have been discovered through further testing, am I really to be blamed for all the evil that occurs? I tend to think the epistemic condition for complete responsibility is not met in such a case. But then how do we classify that? I’m happy to say it can’t be explained away by appeal to free will, but it still raises interesting classificatory questions.
Anyway, this is a minor point. Let’s move on.
3. The Edwardsian Theodicy
As I said earlier, despite his initial dismissal of the soul-making theodicy, Poston thinks that the notion of virtue-building might provide the basis for a solution to the problem of social evil. In this respect he appeals to the work of Jonathan Edwards in The Nature of True Virtue and develops something he calls the Edwardsian Theodicy.
The central premise of this theodicy is that the truly virtuous person has a love for being in general. In other words, they don’t try to benefit particular persons through their actions, but rather they try to benefit all persons, equally and without prejudice. The claim is that such a general feeling of love is not compatible with individual non-cooperation in cases like the Prisoners’ Dilemma and the Tragedy of the Commons; the truly virtuous person is the one who cooperates, despite of the dynamics of the game. Thus, the claim is that being presented with such dilemmas gives people the opportunity to develop the Edwardsian virtue.
To put this more formally:
- (7) Evil can be theistically justified if it gives people the opportunity to develop the virtue of general love.
- (8) The social dilemmas from which social evil arises give people the opportunity to develop the virtue of general love (because individual cooperation expresses the virtue).
- (10) Therefore, social evil is theistically justified.
Grant for the sake of argument the motivating premise (7). The question that concerns us is whether (8) is true. We have sketched a basic argument in its favour, but does this argument hold up to close scrutiny? Poston argues that it does not. To be precise, he argues that the two-player Prisoners’ Dilemma does support the Edwardsian virtue, but multiplayer dilemmas do not.
The argument in relation to two-player dilemmas is easily made. Take the following version of the Prisoners’ Dilemma. The numbers in the box represent the cardinal payoffs for the players. In order to account for the Edwardsian virtue, we assume that the truly virtuous person would choose whichever strategy yielded the highest overall social utility. In other words, they would sum up the utility for each player (including themselves) in each of the four possible outcomes, and pick the strategy which is guaranteed to yield the highest number.
In the case of the two-player game, this is very clearly the strategy of cooperation. Why? Because it yields an overall utility of 5 or 6, versus an overall utility of 4 or 5 for the strategy of defection. In other words, cooperation dominates non-cooperation: it yields an equal or higher overall utility, no matter what the other player does. Thus it is true that this social dilemma gives the player an opportunity to develop the Edwardsian virtue.
This brings us to the multi-player dilemma. The situation is much trickier here. A crucial assumption made when presenting the water shortage game in part one was that the threshold of cooperation needed to achieve the social good was vague. In other words, there is a penumbral region of cooperation/defection where a change to one individual’s behaviour will not make a difference and where it is not clear whether the social good will be obtained. This is a realistic assumption, though it is often left out of game theoretic presentations of the multiplayer dilemma. It also has a significant effect on the plausibility of the Edwardsian theodicy.
The problem is that once the threshold for realising the good/evil is vague, it becomes very difficult to express the Edwardsian virtue of general love through cooperation. Indeed, given the vagueness, individual defection might make things better off for everyone. This is because, for example, the individual could make the world a more beautiful place by using the water for their flowers. In fact, it’s worse than that. Not only is it possible for individual non-cooperation to be the virtuous move, it can be shown that individual non-cooperation is the rational strategy from the Edwardsian perspective.
Look at the payoff matrix below. It represents the ordinal payoffs to a randomly selected player (player i) given three possible states of the world: (a) we are below the threshold of cooperation; (b) we are in the penumbral region; and c) we are above the threshold. As can be seen, granting these three possibilities, and granting that the aim is to secure the highest overall social utility, it is true that individual defection is the rational strategy. It dominates defection across all three possibilities.
The one aspect of this that may niggle is the claim, represented by the payoffs, that defection dominates cooperation when we are below the threshold. The reason for this is that if we are below the threshold, we don’t know for sure whether our decision to cooperate will make the critical difference. If it does, then that’s all well and good, but if it doesn’t the social evil isn’t going to be avoided and we miss the opportunity to make the world a more beautiful place. As Poston puts it:
The reason this situation arises is that each individual can bring about a Pareto improvement of resources by defecting. Since an individual doesn’t affect the social outcome but does affect some item of value in the world, each individual faces the prospect of a wasted sacrifice and a loss of some item of value by cooperating.
But, of course, since defection is the rational strategy, the social evil will arise. Thus, being truly virtuous doesn’t help us to avoid evil. The Edwardsian theodicy looks less promising in this light.
To sum up, there’s a lot to be said for Poston’s article. It highlights an interesting social phenomenon; it identifies and defines it as a new type of evil; it argues, somewhat persuasively, that traditional theodicies fail to account for that type of evil; it identifies a new type of theodicy; and it argues that, promising though it may be, this new theodicy also fails to account for the social variety of evil.
Two critical points to conclude. First, as pointed out by Chris King in the comments to part one, although Poston criticises others for making unrealistic assumptions about social dilemmas, it may be that his analysis incorporates an unrealistic assumption of its own. Specifically, it assumes that individual decisions to cooperate/defect are made simultaneously and independently from the decisions of others. This is a classic assumption in game theoretical modelling, but it is often socially unrealistic. If it is true that individual decisions influence other decisions, then it is much less clear that (a) the tragic elements of social evil identified by Poston will arise; and (b) that social evil is completely distinct from moral evil. If a player knows he might influence others through his actions, then he might be doing something morally wrong/right by defecting/cooperating.
The other point, less important than the first, is that although Poston thinks his analysis of the multiplayer dilemma defeats the Edwardsian theodicy, I’m not so sure. I have to say, I’m not too enamoured by the theodicy in the first place, but leaving that aside it seems to me that Poston’s argument relies on the belief that expressing true virtue must prevent evil if the theodicy is to be successful. As he puts it:
Ultimately, the reason the Edwardsian defense fails is that each truly virtuous person can know that her act will bring about a better world but the collective result of each person making a better world is unintended disaster. Even with the truly virtuous, the road to hell is paved with Pareto improvements.
But that seem slightly off to me. Surely, the whole point of a theodicy is to justify the occurrence of evil, not to point to ways in which it can be prevented. Even if the Edwardsian virtue leads to suffering, we still get a chance to express and develop it through our individual action. If the virtue is sufficiently worthwhile, that might be enough. What’s troubling in this case is the obvious inconsistency between the expression of the virtue (viz. defection in the multiplayer PD) and the outcome. Clearly, the outcome does nothing for human being in general, even if the individual actions that led to that outcome were intending it. Maybe that’s where the real problem lies.