Unawares

Building on Thinking Backward, I had the following interaction with a friend:

me: One observation I've made regarding "in contact with". Touch an object... like your desk. It's quite easy for you to feel your fingers touching something else. Now touch your face. It becomes much more difficult to discern "finger touching face" and to discern it from "face touching finger." Rather, you feel there is one thing - contact - that is associated with neither face nor hand... or both. With a lot of effort, you can discern the two, but the default is not to. This fascinates me. How is it that the perception of two things (face and hand) are instantly merged into one contact?

Ron: That's actually not that hard to understand. All of your raw sensory input is interpreted by your brain to build a model of the world. Your awareness is of that model, not the raw input. [emphasis added] This is why you don't perceive the blind spot where the optic nerve attaches to your retina [1]. The model is informed both by the inputs (i.e. what you experience, especially when you're young) and the design of your brain, the latter of which has evolved over many millions of years to maximize your reproductive fitness, and not necessarily to faithfully reflect reality. The ability to easily distinguish between touching yourself and touching/being-touched-by something else has a lot of survival value, and so your brain evolved to put those into two different experiential categories (sometimes called "qualia"). Another manifestation of the same effect is that most people can't tickle themselves.

[1] It's not just the optic nerve blind spot. If you suffer from macular degeneration you get an additional blind spot right in the center of your field of vision, but you don't perceive this as a dark spot. Instead what you see is that the world gets "squnched-inaround the center of your field of vision as if being sucked into a vortex. Small objects can vanish entirely when you look at them directly. It's very weird and disconcerting.

That made me think of my AI architecture and specifically about my relatively unresolved thinking about Context (pdf). Combining all of the above, we can say the following (which makes no sense unless you’ve read all of the above):

System 1 integrates all sensors into a single model of the current state (including the individual’s position within the external state). Figuring out (“recognizing”) individual elements in the sensors’ views is necessary but not sufficient for this task… so "Recognition” sells it short in my old model. It seems to be more like the operating system level discussed in Thinking Backward. A summarization of the model… with survival-related areas highlighted for action… is the Context. That’s likely a very broad input to System 2 (“Reasoning”) with only a very narrow (just 10 bps!) back channel. This makes both architectural and evolutionary sense.

_______

Ron also advised:

The citation you want is Daniel Dennett's "Consciousness Explained".  It's a bit of a slog to read, but it's as respectable an academic citation as you can hope to find in this line of inquiry.

The TL;DR is that consciousness -- the subjective experience of existing as a unified mental whole that exists at the present moment in time and has agency and free will -- is an illusion.  What actually happens is that your brain concocts a back-story to explain its own actions that make it *feel* to "you" as if "you" are making decisions even though the actual decision-making process exists beyond your control and even your awareness.  Bizarre as this sounds, there is some pretty compelling psychophysical evidence to support it.

I’ve acquired a copy and will write about it once I’ve slogged through.