In my previous blog post, I explicated my current model for the origin and function of prior clusters, which I associate with Jung’s archetypes and the general idea of cognitive subagents. In this post and the following, I want to examine a specific archetypal dyad in the context of an enchanting Spike Jonze film.

The Classical Anima and Animus

The anima is a number of things, but the most expressive compression seems to be to identify it with a prototype of the opposite gender’s psyche. Jung writes:

Every man carries within him the eternal image of woman, not the image of this or that particular woman, but a definite feminine image. This image is fundamentally unconscious, an hereditary factor of primordial origin engraved in the living organic system of the man, an imprint or "archetype" of all the ancestral experiences of the female, a deposit, as it were, of all the impressions ever made by woman-in short, an inherited system of psychic adaptation. Even if no women existed, it would still be possible, at any given time, to deduce from this unconscious image exactly how a woman would have to be constituted psychically. The same is true of the woman: she too has her inborn image of man.

Now, you might be asking yourself: ok, so is the anima just a device for predicting the behavior of the opposite sex? If so, it would certainly be useful, especially in societies with significant preference divergence across genders. However, this probably doesn’t compress losslessly enough. Many Jungians seem to think that the anima has much greater impact on a man’s preference; an impact one can purportedly infer from inadequate equilibria in his rationalized policy.

What would this look like? Well, in my previous post I hinted at my hypothesis that a top-level agentive preference can generally be decomposed as a preference over rotations between subagents. If we keep around a model of the opposite gender in order to improve fitness in games with real instantiations of that gender, why not incorporate it in the rotation? It would allow for other models to become much sparser, which should be optimal as long as subagent-switching costs are sufficiently low.

An anima, then, would be a feminine subagent whom we could rotate into a scenario for one of two reasons: to provide insight or to model action. The latter case seems to have some empirical evidence; I can’t seem to remember the book I read this anecdote in, but apparently, when a group of burly miners became trapped in a mineshaft, after a few hours some of them rapidly and abruptly began inhabiting “feminine” roles by comforting others, maintaining morale, and conserving resources. Assuming that the subagent model is correct, evidence like this should induce us to somewhat strongly update towards concluding that there is in fact something like a feminine subagent.

The former is probably a bit more difficult to be confident in. Why would a feminine subagent be particularly useful for a man outside of social contexts? What valence does gender even take on outside the context of the world of man? These are imposing questions which need to be answered before the insight capacity can be taken at face value. But a lot of Jungians seem to do just that. Let’s entertain why.

An Evolutionary Hypothesis for the Animus/Anima

In order to give the Jungians a fair hearing, I’m going to try to develop a toy model of gender which cashes out in gendered subagents which provide fundamental insight.

In the state of nature, women get pregnant quite often (take the Pumé — women who survive childbearing have 7 +/- 1 pregnancies over their careers). Together with the rearing of out-of-womb embryos (infants until age 1), this means that women likely spend a majority of their time in some state of child-rearing-induced effective disability. What does this mean? Well, two activities are largely out of the question: hunting and killing other males. Why?

Projectile weapons are a relatively recent invention (invented first ~50k years ago); for most of our developmental history, killing is something which was done with the hands or with simple hand tools, like a stick or rock. A pregnant or nursing woman simply would not be able to compete with men when it came to killing with the bare hands. Of course, there’s the matter of upper body strength as well, but that’s more of a chicken-or-egg quagmire. Did body morphology follow role specialization or vice versa? My hunch is the latter but in any case, it should be clear that training as a killer would not be very worthwhile for female planning to have 7 pregnancies in her prime years.
 

So what is a woman to do? Well, one might guess that the matter of hunting/killing would induce a general specialization pressure across genders. If men are spending a good amount of time hunting and preparing to kill other humans, they will be incentivized to specialize in tasks that share skills and procedures with hunting and warfare. As a result, women will be incentivized to specialize in tasks that are maximally far from hunting and killing, and thusly tasks are partitioned among the genders. One can think of a “space” of tasks, where the axes represent the degree to which basic skills are drawn on and points are tasks; here, the hunting/killing split induces a planar cut through space, whereupon each gender specializes in tasks on “their” side of the cut.

What might this cut look like? Well, we can make an educated guess from the kind of skills that hunting animals and killing adversaries require. Firstly, spontaneity and constant innovation on strategies. Some animals are notoriously sensitive to human hunting practices and patterns: it’s been found that widespread practices of divination emerged as a way to randomize strategies — an optimal approach when hunting animal groups that will in their migratory patterns exploit any shred of predator predictability to their advantage. Of course, the danger of predictability and gains to innovation are only amplified when seeking to kill other humans. Using the same means, strategy, or tactics over and over likely would lead to failure. In an adversarial world, improvised innovation often beats perfected routine. Creativity, focus, short-term planning, and performance-based emotional monochromaticity would be paramount.

With the “male” side of the cut having been identified, we need only to look to the opposing end of the skill space to infer the “female” side. Where does perfected routine beat consistently improvised innovation? Well, basically in every other context. Child-rearing, gathering, and sustaining the community both socially and materially are all demanding but rewarding learning environments. Honing a few strategies works in these non-adversarial domains; this is a subset of the skill space where conscientiousness, awareness, long-term planning, and emotional responsiveness are critical.

What does this get us? Well, we might expect that a pressure for men to specialize in cluster 1 and for women in cluster 2 would lead to an emotional brain that has these clusters pre-built; cluster 1 would represent a subagent optimized for male social and economic behavior, and cluster 2 would represent a subagent optimized for female social and economic behavior. Because all people need to access both clusters to a degree, and because being able to simulate others is useful, we’d expect for men to inhabit cluster 1 very frequently but cluster 2 as well from time to time, and vice versa for women. Cluster 2 would help men become more pro-social, conscientious, and think long-term, and cluster 1 would help women be more spontaneously creative, independent, and stand up to adversaries. Indeed, if one reads Jungian literature, much of the flowery language essentially amounts to just this — and voila, we’ve evo-psyched ourselves the anima and animus.

Uneven Development

Much of arrested development can be caused by an addiction to one’s strengths. Whenever skills are substitutable, it’s tempting to over-rely on the one which is best developed, even if doing so induces a sub-optimal learning policy in the long run. Learning a new skill is slow and has uncertain returns, so we’d expect agents with high time discounting and risk aversion to go all-in on a few clear winning strategies more often than not, if even these strategies extend to some contexts more poorly than others.

An instructive analogy might be that of a soccer player who is a right-dominant. The soccer player is strong in his right naturally, and reinforces that strength by constantly practicing shots with his right. He knows that there could arise situations wherein it would be advantageous to shoot with his left, were he to have a bit of live experience doing so, but doing so is expensive — it means he potentially needs to eat the cost of missing a few shots while he develops the skill. It means he — and his team — have to accept the curse of development. This can be risky. What if he incurs the curse during a critical contiguous period and gets benched, seriously impairing his long-term development? On the other hand, if he never moves up the skill tree, he gradually becomes less and less competitive. It’s a non-trivial problem, especially when the stakes are high.

Of course, they often are for the man in his youth and adolescence. He needs to develop his primary cluster for pressing in-group social reasons (his male peers care if he is in touch with his cluster 1, and are at best indifferent to facility with cluster 2), and so cluster 1 quickly becomes a strength that is relied on addictively. Unfortunately for him, training the contra sexual other risks exposing him to be insufficiently masculine, which can cause problems for his social life with other men and even — especially — women. He knows that women do like a sensitive guy, but really the kind who seamlessly integrate caring into a masculine mold; that is to say, they prefer the post-learning equilibrium to the pre-learning equilibrium. The middle? Where he vacillates between gushing softness and tough-guy machismo? Cursed.

Just as many soccer players overcome short-term penalties in order to progressively move up the skill tree, many men will keep growing their contra sexual capacities to whatever equilibrium is motivated by his surroundings (presumably this constantly increases as men negotiate relationships with girlfriends, female coworkers, wives and eventually daughters — perhaps this is related to the fact that many men become more agreeable with age). However, some will get stuck. Just as with the aforementioned soccer player, this obsoleteness doesn’t immediately lead to catastrophe. But, with time, they become more and more ill-equipped to deal with peer competition. At some point, their atrophied anima becomes an impediment — a neurosis, as a Jungian might say.

Next time, we’ll be discussing what this might experientially look like and how it might be overcome, using Jonze’s Her and personal experience as a dialectical substrate.