Our compressed perception

by pchvykov1 min read6th Apr 2021No comments



One of the most important and amazing feats that our brain learns to do in this complex world is to find compressed representations of the vast amounts data constantly streaming in through our various senses. As such, we may view many of the idiosyncrasies of our world-views and behaviors as side-effects of our individual habits for such compression. These habits depend on the actual objective patterns found in nature, as well as on our cultural and social context, which chooses to prioritize some patterns over others. A musician, for example, will tend to hear and be able to describe a sonata in much more detail than a lay-person, who may just categorize it as “classical music” and call it a day.

I intuitively think of it like this: if we represent the full space of actual sensory inputs as a continuous 2D space, then our brain finds some network or mesh of learned approximations that somehow covers this space (see pic). Then (unless we are in “curious learning mode”) any incoming sensation “snaps to” the nearest known point on this mesh, and the original input gets immediately forgotten. So when I see a brown wooden table with light blue stripes and four thick bent legs, all I will typically see is just “table,” immediately discarding all other sensory information.

This perception mesh will cover certain regions of the space of possible sensory inputs more densely than others, and some regions that we have never experienced may lack any coverage (read: lack any interpretation — like for Zen Koans). Over our life-time, this mesh may coarsen as we find more and more efficient ways to compress our experiences into just a few facts that actually matter for our survival. This may then subjectively feel like “the world has become gray and boring where nothing new ever happens” — because we learn to efficiently ignore the little things that don’t kill us. By the same token, when we are climbing a mountain, where little things can actually kill us, then we quickly start caring about much more details of our surroundings, and notice that world became rich and beautiful.

Another cool idea this raises is that since these meshes of compressed representation never agree exactly for any two people (though some of the dominant cultural concepts may generally align), we tend to have different understanding of the same situation — we can literally “see different things” when looking at the same scene. This leads to misunderstandings and disagreements, which may sometimes be very difficult to pin-down, since these compressions are performed before we ever become conscious of them. This is what allows the language of mathematics to be so powerful: its rigid structure enforces the particular compressed representation given by some theory to be the same, regardless of who uses it.

Next time, I’ll try to argue that all mindfulness practice aims at learning to “unpack” these compressed representations we are used to, and thus gives us the ability to “see the world as it really is” if and when we so choose. This in turn can help us to enjoy life and resolve disagreements, as well as to see things we did not think existed (things we used to ignore).

[Cross-posted from my blog https://medium.com/bs3]


New Comment