What Dreams are For

I found this very intriguing sampling of the Overfitted Brain Hypothesis for why we dream.

In effect, the hypothesis is that dreams are supposed to prevent us from being too overly adapted to our specific waking workaday world.

The “overfitting” is by analogy to machine learning.  When an AI is being trained on a training data set, one danger is “overfitting”–the algorithm might get so good at parsing the training data set that it can’t handle the actual new data it was meant for.

This is a secular hypothesis, but it caught my eye because it fits with my own speculation about why mortality includes sleeping.  I think that God may have given us sleep as a way of interrupting our mortal experience so we don’t get stuck in a rut–so we get lots and lots of bites at the apple, as it were.  In other words, the scriptural wisdom of “sufficient unto the day is the evil thereof” or the consolatory folk wisdom of “tomorrow is another day.”

Ditto our hypothesis about why mortality is so richly complicated.  A straightforward test of one’s agency could probably be much simpler than this.  But all this immense variety keeps us from limiting ourselves.  God wants us to be good and great, not good and reduced.

Yet ofttimes a secret something whispered, “you are a stranger here.”  It whispered, “this is just a training data set.  Don’t overadapt.”

I believe our taste for novelty and adventure is at bottom our subconscious awareness that we are really much bigger than this.

Of course the JG theory about dreaming is that dreaming is what the spiritual world is like.  The material world with its hard edges and concrete realities imposes order on the spirit, but when we are asleep the spirit reverts to its natural state of ambiguity, signifiers without referent (almost pure meaning), symbolism, and so on.

The Meaning of Sleep