Here’s a thread from Saturday that seemed to be quite popular. It explains a saying that I’ve found myself reaching for a lot recently, using some other ideas I’ve been developing in the background on the intersection between philosophy of action, philosophy of politics, and philosophy of computer science.
In reflecting on this thread, these ideas have unfolded further, straying into more fundamental territory in the philosophy of value. If you’re interested in the relations between incompetence, malice, and evil, please read on.
This afternoon, I would like to talk about Hanlon’s Razor: “Never attribute to malice that which can be adequately explained by stupidity.” I think this is but one half of an important relation between essentially normative concepts.
Of course, the razor is very useful, insofar as it encourages us to look for mundane sources of failure that need not be explained in intentional terms, and thereby engaged in adversarial terms. Often, it’s used to temper conspiracism: do not presume an (ill) will without cause.
However, as many have pointed out, this is a heuristic for which their are obvious exceptions: sometimes there are real conspiracies and actual conspirators, whose wills are most definitely ill, even if they are neither irresistibly powerful nor seamlessly integrated.
I think something similar can be said for Hanlon’s razor more generally, but that this reveals something important about the logic of systemic explanations and their associated normative critiques. Here we move from the attribution of intentions to their constitution.
So, here is the dual principle to Hanlon’s, as yet unnamed: “Any sufficiently advanced incompetence is indistinguishable from malice.”
Here’s a familiar example: a police officer puts a suspect in a hold; they haven’t been trained to do this properly; the suspect dies. Hanlon treats the training lapse as proximate cause, and diffuses responsibility. Grey treats the system of training as a site of responsibility.
To extend the example: the same thing happens again, then again, etc. Lapses in training cease to be excuses once the failures are revealed to be systemic rather than accidental. Responsibility can be diffused, but it is thereby infused into the system and those who compose it.
This is the foundation of the critique of complicity: if one is part of a putatively self-correcting system, and one is capable of contributing to this correction, but fails to do so, then one is complicit in the consequences of its errors, even when one doesn’t grasp them.
The foundation of responsibility is choice, but one is not merely responsible for those consequences of one’s actions that one completely and wholeheartedly wills. The content of the action one intends exceeds the grasp implicit in one’s intention, yet the excess is not infinite.
Responsibility for failure correlates with involvement in systems that provide epistemic and practical purchase on these types of failure. Some are familiar-avoidable, others are unfamiliar-unavoidable. Hanlon considers the familiar-unavoidable, Grey, the unfamiliar-avoidable.
In accordance with the moral logic of ought-implies-can, both the range of possible mistakes and the range of permissible mistakes expand alongside our epistemic and technical capacities. However, they do not expand in the same way. The latter tends to trail the former.
The means of self-conscious action often arrive after the means of conscious action, and this entails some strange normative consequences. At least, if one believes, as I do, that there is no consciousness without the possibility of self-consciousness.
But the theory of action and complicity I’ve been outlining is already Hegelian enough without getting into the intricacies of essentially historical modes of normative reasoning.
Let me return to the question of how the intentions we can form are constituted by the epistemic and technical systems that underpin our capacities for action: What is it to send an email, exactly? What is it for this to fail? What sorts of failure are we liable for, and why?
The sender and receiver of an email each have expectations about what must be done to send an email and what it is to receive one, and these expectations are historically shaped by, but equally abstracted from, the layers of protocols and implementations that make it possible.
That’s a fairly quotidian example of an action whose significance is articulated by a wider technical context and a series of norms regarding how we handle failure within that context. This technical context then itself involves epistemic infrastructure.
Accidentally including a string in an email that induces an unanticipated error on the receiver’s phone is one thing, the receiver failing to check their spam folder is another, and the devs failing to fix the errors that lead to these failure states is something further still.
This is why the word ‘incompetence’ in Grey’s Law is important. It includes epistemic failures as much as technical ones, and the complex mixtures of the two we find everywhere in contemporary society.
Much has been said about the connection between power and knowledge (cf. Foucault), but we are beginning to appreciate the relation between power and ignorance (cf. David Graeber, Charles W. Mills, and James C. Scott). The epistemology of ignorance is a conceptual frontier.
Interestingly enough, though Graeber’s work on the connection between violence and stupidity is quite interesting, I think it undermines his critique of Foucault. This is precisely where it is important to distinguish power from domination.
Anyway, to bring it all together: building increasingly complex and fallible technical systems through which the lives of others are mediated by necessity, and then sustaining systemic ignorance of these failures, either through laziness or bloody-minded refusal, is malicious.
This is the modern face of malice: automated ignorance, dumb compliance, universal credit. We don’t need big data and deep learning to experiment with artificial stupidity, but it certainly opens exciting evolutionary opportunities for what was once merely banal evil.
I’ve been teaching medieval theology this term, and so have had some opportunity to revisit the concept of evil. I’ve always been sympathetic to the Stoics, Spinoza, and even Nietzsche to some extent on the topic of evil. They are united by their rejection of some transcendent opposition between good and evil that applies absolutely, across contexts, in favour of an immanent opposition between good and bad whose applicability is relative to contextual factors governing the genesis and cultivation of agency. Yet, as I’ve noted before, I’ve gotten more Platonist as I’ve gotten older, and less hesitant to talk about normative absolutes. So, have I changed my mind? Can we move from talk of malice to talk of evil as such?
In becoming more Platonist, I’ve gained further respect and sympathy for Alain Badiou. I’m far from agreeing with him on most topics, but his work displays an admirable unwillingness to compromise on the value of philosophy and the need to return to its perennial questions. This is perhaps nowhere more evident than his insistence on defining evil, though his definition is some sort of dialectical synthesis of absolute (unconditional commitment/truth) and relative (conditioned situation/event):
Evil is the moment when I lack the strength to be true to the good that compels me.
There’s something compelling about this definition, though it’s undoubtedly peculiar. Although it’s superficially similar to Spinoza’s conception of sorrow as that which decreases the power of a given mode (or animal), it is distinguished by its focus on the will of a given subject (or rational agent) constituted by its commitment to some (singular) truth. The environmental changes that diminish the power of a species of animal are merely bad for it qua biological species, whereas the contextual factors that diminish the will of a political tendency are evil for it qua subject of truth. The substance of evil is something like the akratic friction that prevents us from living up to our commitments.
I’m not entirely sure about this definition, but I think there is something significant in Badiou’s unwillingness to reduce evil to whatever conflicts with some hegemonic moral framework, and thereby to conflict as such. Evil has historically been attributed to those agents, actions, and associated factors with which the dominant system is somehow at war, in order to locate the motivations for this war outside of the system itself, and thereby rationalise the real motivations that drive it. This form of rationalisation is precisely why so many feel the need to abandon the concept of evil entirely. It’s equally the source of my antipathy to those who preach the ‘re-enchantment of nature‘, by positing some more or less primitive layer of response to normative facts embedded in our environment.
However, I think that there is something deeper to be said about the relation between conflict and cooperation here. In almost every area of life, from aesthetic and epistemic projects to ethical and political praxis, there are not merely those who have commitments that diverge from our own, but those whose commitments are thoroughly opposed to ours. The types of adversarial relationships that result from these oppositions are mixed and manifold, determined both by the intrinsic character of the commitments and the practical contours of the contexts within which they are realised. There are radical aesthetic conflicts whose depth we cherish, and minor political clashes whose shallowness cannot mask our unbridled ire; grand theoretical disagreements that are perpetually productive, and petty practical differences that breed nothing but resentment; nemeses we keep closer than friends, and allies we treat worse than enemies. The complexities of conflict and cooperation are at least as rich as the myriad means of human industry and the endless ends to which they are put.
Wherein lies the evil, if not in these conflicts? Well, there is some evil exhibited here, but its extent is relative not to the severity of a conflict, but to the significance of the cooperation such a conflict disrupts. It’s here that our strength to pursue the ends our actions converge upon is sapped by divergences in the means through which we pursue them. If there is any true sin, it is the elevation of petty differences above the common causes that supposedly define who we are, either individually or collectively. This much is clearly compatible with Badiou’s existentialism, but I would like to push it further, by relating it to the more Hegelian theory of action sketched above.
If the content of our commitments can exceed our grasp of them, insofar as it is constituted by the variegated technical and epistemic infrastructure of the modern world, then our ability to undertake divergent commitments, let alone opposed commitments, is to some extent dependent upon prior practical convergence and theoretical agreement. There are complex relations of mutual recognition and mutual reproduction that articulate both our causal capacities to act, and the normative standards that apportion responsibility for success and failure. This includes everything from the distributed peer networks that validate epistemic and technical expertise, and the legal institutions that configure the roles we are assigned in any and every official enterprise, to the physical and economic infrastructure that underpins the production, reproduction, and modulation of every aspect of human life. Evil is not the refusal of such cooperation, but its subversion.
I’ve previously mentioned the principle of normative parity: the idea that social roles are constituted by a combination of authority and responsibility, where the possession of the former is justified by the assumption of the latter. This is to say that one gains socially mediated capacities for action only on the condition that one subjects oneself to certain norms governing their correct disposition. Here, then, is a concrete subversion of cooperation: the acquisition of authority combined with the evasion of responsibility through sublimated conflict. The face of evil is the face of an enemy who pretends they are a friend; the face of one friend who would turn you against another; the mask of authority that pretends it has one’s best interests at heart, while shirking any and every responsibility that might legitimate it.
And yet, the truth of Grey’s Law is this: there exists faceless evil. Malice requires will, and evil merely supposes its subversion. We can unwittingly twist our ties of cooperation into evil shapes. Worse, we can merely intend some minor sin, a small comparative advantage or a tiny lapse in duty, and in doing so unravel the threads of commitment from which society is spun in unpredictable and irrevocable ways. We pay for the freedoms we enjoy with ranges of competence that no single individual is capable of appreciating let alone possessing in full, and incompetence can undermine those freedoms in ways that the incompetent are almost by definition unable to anticipate. The reason that sufficiently advanced incompetence is indistinguishable from malice is that it is genuinely evil: the value of eradicating or ameliorating it exceeds any value we might ascribe to punishing the corresponding sin, though the former may sometimes imply the latter.