Fluid Rigor

What made us human? It’s one of the most enduringly interesting questions, from mythology to science, because the apparent ordinariness of being human conceals an abyss of ignorance about how it works and how it came to be.

A profound answer to this question is the one provided by Darwin: natural selection made us human. Theories after Darwin must attempt to explain more specific aspects of human evolution. What was our selective environment like, and what were the crucial adaptations that allowed for the development of our special kind of cognition and social organization? In what order did they occur? Does evolution act on elements of human culture, or on human groups as superorganisms? Some examples of post-Darwinian origin stories: cooking made us human, running made us human (I’m partial to both), compassion made us human, schizophrenia made us human.

Perhaps you have heard of René Girard. He was an interdisciplinary scholar who proposed a theory of what made us human, a process he calls hominization. I have been reading Girard’s Things Hidden Since the Foundation of the World and have found it to be a fascinatingly troublesome theory of everything.

Narrative is not located in the text itself; rather, it is the cognitive process of model-building and model-revision that the reader engages in while processing the linear text (for more on this, see Nick Lowe’s excellent book). As I read Girard, I was alternatively pleased and irritated: impressed with the elegance and narrative power of his model, and annoyed by the inconsistency of fit his story had with the existing models living in my head. Just as I would perceive an “a ha!” of pattern-matching to one of my pet theories, I would get a sense of uneasiness as to whether Girard’s ideas were at all reconcilable with what I think is true. Is the pattern-matching feeling an identification of merely superficial similarities (like a Barnum statement), or can Girard’s simple theory explain them all, getting rid of a bunch of unnecessary (and cognitively expensive) complexity?

Girard read Darwin. Girard has even been called “the Darwin of the social sciences” (the other person to be called that was Karl Marx, incidentally). Girard (in The Scapegoat and elsewhere) maintained that even if his theory was false, it was a scientific theory; defenders of Girard often echo this sentiment precisely. However, what Girard means by “scientific” is something other than what is usually meant by that adjective.

I will present Girard’s theory and explain the ways in which it pattern-matches the ideas that I have found to be profound and enduring explanations of important puzzles – a sympathetic synthesis of Girard’s ideas with my own mental model. Then I will unsympathetically explain what I see to be the (epistemic) problems with Girard’s theory. Finally, I will attempt to synthesize these two perspectives. We will see if the “postrational” epistemic eclecticism, a fluid kind of rigor that attempts to translate meaning and knowledge between domains, has anything to say in a real epistemic mess.

Girard’s Beautiful Theory

In the beginning, there was copying. Girard proposes that all human desires (excluding instinctual desires) are copied from others. In order to learn, a human needs a model; but as he copies his model, he also copies his model’s desires, and in doing so, becomes his model’s rival. Desire is triangular: it is a relationship not merely between a desirer and an object, but between an observer, a model, and the model’s imagined object of desire.

Girard calls this acquisitive mimesis: desiring objects, social position, or a particular mate because someone else desires to have them. Girard thinks that this copying function is present, to some degree, in nonhuman animals; he describes an ape reaching for a some desirable fruit, then another ape reaching out his hand to copy him – but stopping himself, lest conflict occur.

Mimesis, in Girard’s view, is at the heart of conflict. Conflict is itself a mimetic process: when opponents fight, they adopt highly similar words and gestures and actions, as if they have become doubles. As violence reproduces itself through copying, the result is a state of war of all against all.

However, there is yet hope of peace. The copying behavior that once initiated conflict now emerges in a cooperative role. Gradually, those in battle converge, through copying, on a single person who is blamed for the trouble and violence. This is called conflictive mimesis – cooperating against rampant, unrestrained violence by sacrificing one member of the group. It’s the original trolley problem – but, in Girard’s view, the participants to not consciously know they are sacrificing an innocent to save the many. In order to coordinate together, they must deceive themselves that they serve justice, that the sacrificial victim is actually guilty.

The sacrifice brings peace. Over many iterations of this process, humans develop a social technology of ritual and prohibition. In Girard’s view, all ritual is descended from the scapegoat sacrifice; it is used to intensify the violent upheaval, in order to bring on the selection of a single victim, and hence peace. (I think parallels can be made between this process and the “cycle of violence” that is said to typify domestic violence, in which a victim may precipitate a violent outburst in order to skip to the peaceful part of the cycle.) All prohibition is descended from collective memory of the horrors of copying (such as the prohibition of twins or of mirrors), which are forever associated with mimetic crisis – the war of all against all. The rituals and prohibitions which emerged from this experience with violent crisis are the basis for all human culture.

In addition to a psychological (desire as mimetic) and anthropological (scapegoating as origin of ritual and prohibition) elements, the theory also has a theological and ethical angle. Girard was a believing Catholic, and argued that Christ revealed, once and for all, the innocence of the victim of scapegoating, so that the mechanism would no longer work. Deprived of the ability to pretend that a scapegoat was really responsible for a violent crisis, society risked Apocalypse – the violent war of all against all that, unlike the prehistoric versions, can no longer be solved by sacrifice. (This is why Jesus declared that he came not to bring peace, but a sword.) Society after the Gospels (themselves elaborating advances from Jewish liturgical theology) must act in awareness that the guy in the trolley problem strapped to the rails is just as innocent as the many people on the other track. Girard is not arguing against utilitarianism exactly; he is arguing that Christian revelation deprives people of the ability to blame the victims of utilitarian calculations, and hence to obtain peace from sacrifice.

Mimesis and Scapegoating, Meet Peopling and Cooperative Ignorance

The Girardian concept of acquisitive mimeses and the triangular, mimetic nature of desire calls to mind what I call peopling – the special form of human self-consciousness in which we form mental models of other people, and imagine ourselves from the perspectives of these mental models. If we are constrained toward (self-)consciousness and having “others in mind,” of course the desires from these mental models leak into and inform our own desires.

The anthropological and theological aspects of Girard’s theory seem to work on the basis of what I have called cooperative ignorance – cooperating to collectively avoid knowledge in order to achieve some goal. Certainly, in Girard’s model, people cooperate to remain ignorant of the innocence of the sacrificial victim, for the goal achieving peace through the sacrifice. His reading of the Gospels is that they destroy the possibility for this form of cooperative ignorance. The Gospels change the state of common knowledge about moral justification; a preference falsification equilibrium based on that justification may no longer be maintained in the presence of this knowledge.

I was not able to synthesize Girard’s theory with my theory of ritual. Girard explicitly rejects the “functionalist” interpretation of ritual promulgated by Durkheim, in which rituals and prohibitions are analyzed according to the functions they perform for groups. The self-interested, rational, costly signaling/hedonic theory of ritual that I have elaborated goes even further in the Durkheimian direction that Girard rejects. Girard’s theory, for example, would not (as far as I can see) predict that religious communes with more costly prohibitions and obligations would outlast less demanding communes, as in Girard’s model, prohibitions are remnants of things metonymically associated with the Mimetic Crisis, the war of all against all. Nor, I think, can his model predict that communes with fewer prohibitions would dissolve through apathy and free riding, rather than in a violent frenzy. The next section will identify some other problems.

Problems For Girard’s Theory

One reason I was attracted to Girard’s theory is that it’s a bit crackpotty. I have in the past defended an aesthetic analysis of (potentially) crackpot theories – judging them as folklore or art forms. Theories that are good when judged as folklore share some traits with theories that are good when judged for epistemic value. As I noted in Puzzle Theory, theories are felt to be epistemically satisfying explanations (regardless of their truth value) if they share four main characteristics:

Girard’s theory succeeds very well at being small and elegant, a parsimonious-seeming pattern that is clean and tidy in itself. It is also easy to match to other domains; you can find it everywhere if you look, like the binary patterns of the I Ching. Girard’s theory fits well with itself, and fits well with the limits of human cognition (allowable complexity) and with what humans might want to do with it (play with it by matching it to literature and the world). However, I think it fits poorly with other domains of knowledge.

There are a few main ways to torture a wild theory:

1. Tie it down and force it to make predictions, then check their accuracy
2. Tie it down and force it to show how it conflicts with other competing theories, and check the fit of the data
3. Make fun of it for refusing to do either of the above

Girard’s theory is a bit hard to tie down. I suspect that a sufficiently motivated Girard scholar could sculpt an account of Girard that is in harmony with my objections. Contemporary Girard scholarship is adept at harmonizing, such as the suggestion that mirror neurons are evidence of the cognitively primordial nature of mimesis. Still, here is my attempt at the first and second theory-torturing methods.

Preferential Copying

Copying is rampant in human culture, but copying is not automatic or mindless. Children learn by copying, but even children preferentially copy some models over others. Children do not copy at random, but select models for mimesis based on whether the model intends to teach, whether the model is proficient or part of a group known for proficiency, whether the model is like the subject in relevant ways, and whether the model has high status. Human children do not merely copy; they use copying as a tool for learning. Cognitive processes much more complicated than mere imitation drive the imitative faculty.

The “herd behavior” demonstrated by adults can be modeled as a rational process: we copy others precisely in situations where their behavior offers cheap information about proper behavior, in situations in which we are uncertain, or acquiring information through other means would be too costly. Here again we have rational copying based on a complex underlying cognitive framework, rather than mere mimesis. People would have to be already pretty human in order for something like Girard’s mimetic crisis to occur, before his hominization process begins.

In both these cases, and with information theory in general, Girard’s theory seems to explain away (rather than contextualize) the importance of the non-mimetic processes found alongside mimesis. Specifically, it seems that agency is denied by treating copying as an automatic or black-box process, rather than looking at how and why people might choose to imitate, rationally or irrationally. If you look carefully an apparently contagious phenomenon, you often find rationality and social learning underneath.

I find Girard’s identification of twin prohibition (twin infanticide) as evidence of the mimetic origin of religious prohibition to be without support. In Girard’s view, people fear doubles because they associated in collective memory with the mimetic crisis. However, twin infanticide must be located within the literature on infanticide in general. Human infanticide since prehistory is not random, but preferentially enacted on those infants least likely to survive, such as children born to young mothers without partners, and children with deformities or illness. Twins are certainly less likely to survive than non-twins, and much more so in history than in the present day. Selective infanticide based on religious prohibition appears to be a cultural justification that allows individuals to practice tradeoffs that increase their fitness. The appeal to primal fear of doubles as explanation seems strange.

One desire that seems to be contagious is the desire for money. Money gets its value from the desire (or demand) of others for it. It is an amazing coordination mechanism, allowing complex trade between different people and groups. However, certain things are chosen as money (or proto-money), and certain things or not. Gold or shells may be money, but chicken entrails and fresh flowers never are. If desire is merely mimetic, any form of money should work; however, money only works when it is scarce, limited in quantity, non-perishable, etc. I do not think that Girard’s theory elucidates the history of money, which seems important to becoming human. Money is a coordination mechanism based on preferential copying – but it seems to produce wide-scale cooperation, not mass warfare of all against all.

What would mimetic theory have to say about advertising? “Provide a model of desire, and your audience will copy it.” This seems to hold for a naive view of advertising – Kevin Simler calls it “Pavlovian.” Compare Simler’s model, in which advertising elaborates legible targets for the sending of social signals, treating consumers as rational users of signals rather than automatic copiers of desires presented to them. It is not clear whether Girard’s theory perceives itself as an alternative to signaling and information theory, but there does not seem to be room for rational signaling and things like Schelling points within Girard’s elegant, streamlined account. If a target is chosen by many, in Girard’s account, it is because of the desire of others; but there are many reasons why people may choose the same object of desire, independent of being memed into it by envy or jealousy.

Is Blood Feuding So Bad?

Blood feud, the massive war of all against all, is treated as the main problem humans had to solve to emerge as human. However, in some societies, blood feuding appears more like a form of social organization (and a potentially fitness-enhancing chance to kill rivals for mates or land) than a fearsome Hell that everyone wishes to avoid. In modern survivals and ethnographic accounts, blood feuds do not seem to result in conflictive mimesis and scapegoating; rather, they are solved by negotiation and payment between the parties. When Girard claims that the interests of the community are served by and end to violence, it is not clear what agency is supposed to be responsible. Many individuals’ interests are served by blood feuding, and a theory about the end of a massive blood feud must explain why they stop fighting even though their interests have not changed. As I have indicated, “acquisitive mimesis” does not seem to explain why people fight – unless you reduce it to “status competition,” which is already explainable in non-mimetic terms.

One alternative model to the scapegoating mechanism is the observation that people in low-complexity hunter-gatherer or pastoral societies appear to engage in collective murder in order to prevent dominating individuals from taking over. They engage in less dramatic tactics as well – mockery, disobedience, or just getting up and leaving – but capital punishment for attempts at domination appears to be common in the ethnographic record. Is this scapegoat punishment? It seems unjustified to say so. While the pattern in similar, the killings typically do not follow chaos or war of all against all.

Peter Turchin’s theory of the formation of mega-empires suggests that only grave threats from nearby enemies with different modes of subsistence (settled farmers vs. raiding nomads) could allow hierarchical, non-egalitarian systems to evolve – forms of social organization in which it harder to kill dominant individuals, or that make people less interested in killing them, or both. In this theory, large-scale violent rivalry between enemy groups is what helps people transcend the kinds of killings that Girard might identify as scapegoat killings. The specific rivalry along steppe boundaries allows the development of new levels of coordination. There are echoes of Girard’s ideas here, but also echoes of the opposite.

Literary Theory

The moving parts of Girard’s theory form a harmonious and tidy whole, and match to many patterns outside of the theory. However, the theory does not harmonize with other maps that purport to explain the same domain. It is like a Le Corbusier house that is prone to mess because its excessive simplicity constrains the kinds of objects that can be perceived as harmonious within it.

Finally, Girard’s interdisciplinary theory brings to mind that of Julian Jaynes. Both are theories that seek to unite evidence from mythology, anthropology, ritual, and psychology in order to support a grand theory of human origin (or at least cognitive change). I think there are important differences, however, in the way that each uses literary evidence. Jaynes, for instance, identifies the frequency of the use of different cognitive verbs in myths from different time periods, and catalogs the mental processes attributed to the participants in the stories. He does not take the myths to be reporting true events; he introduces them for reasons other than to “assert the truth of the matter,” as in hearsay evidence. Girard, however, takes myth and literature much more literally, admitting reports as evidence of (some reflected degree of) their own truth. For Girard, descriptions of rivalrous copied desire in Shakespeare and Dostoevsky are evidence that these are fundamental to human psychological reality. Myths of murdered kings are records of actual scapegoating events, with the innocence of the victim hidden. As Girard was primarily a literary scholar, it is not surprising that his evidentiary case takes a literary form. Perhaps we can examine his theory as science imagined through the cognitive structures of literature. I do not say this as a slur: literature has its own wisdom. And more people read literature than read science.

What Do We Do With René Girard?

Here is an interesting story: it has been reported that Peter Thiel’s investment in Facebook was motivated by his engagement with René Girard’s ideas. Quite a successful prediction!

Here we might be tempted to make a case for Girard’s theory on the basis of its apparent utility, even if we have strong reservations about the metaphysical underpinnings of a theory. At any rate, I am humbled that smarter people than I have been able to find satisfying meaning in Girard’s theory. It may very well be my failing, rather than Girard’s theory’s failing.

What if Girard’s theory is successful precisely because it carves out extremely important parts of reality in a memorable way – like literature? If we suspect it’s not true in a scientific sense (and, of course, half of Science is not true in a scientific sense), should we promote it anyway for the beauty of its compression?

I think Girard’s theory itself would refuse a utilitarian justification like this. Girard’s theology asserts that an encounter with truth is dangerous but necessary. The revelation of the Gospels is justified by truth, in Girard’s account, even though it will bring on Apocalypse – the possibility for internecine war without sacrificial end. On its own terms, then, Girard’s theory does not allow false beliefs to survive even if this revelation causes widespread killing. It is not a respectful treatment of Girard’s theory, on its own terms, to treat it as an aesthetically pleasing cognitive hack. It is truth, or it is nothing.