I don't find the TFA explanation convincing; it's too abstract.
I like the explanation in terms of predator avoidance behavior. Imagine a primitive fish with eyes that can detect motion. If it sees something moving, usually it wants to get away from that thing. If you see something moving in your left eye, the best way to swim away is to send a signal for a muscle contraction in your right side, which will cause you to curl and swim to the right. So the best wiring is a direct connection from left eye to right side, and right eye to left side. The brain is built up starting from that kind of connection.
But sensory input is also reversed. If you're touched on the right side of your body, it projects to the left side of your brain, which then projects back to the right side of your body again. If the goal were to minimize wire length, everything would stay on the right side.
> I don't find the TFA explanation convincing; it's too abstract.
...meaning you didn't understand it.
It may not be a good explanation, but the problem isn't that it's too abstract.
> If you see something moving in your left eye, the best way to swim away is to send a signal for a muscle contraction in your right side, which will cause you to curl and swim to the right. So the best wiring is a direct connection from left eye to right side, and right eye to left side. The brain is built up starting from that kind of connection.
If your fish only contracts the right side of it's body, that's not going to create effective motion, and while your fish is having the seizure you've described, it's going to get eaten.
Even if there were some sort of direct eye-to-muscle connection (which there isn't) the left eye would connect to the right side of the brain, which would then connect to the left side muscle, so you'd have left eye to left muscle (via the brain), not left eye to right muscle.
This is my take. With two sensors connected crosswise to musculature that pushes forward, you get goal seeking for targets in front and avoidant behaviour for targets behind.
It's what our ancient ancestors had. And they were successful, and elaborated on that theme, and by the time they were complicated enough to start thinking "hey wait, why is our nervous system crossed up like this" it was too much bother to uncross it.
You're just making shit up. There's zero evidence for this.
You also don't seem to understand that inputs (senses) are also crossed, meaning that in your hypothetical inputs-directly-tied-to-outputs scenario, the movement produced would be in exactly the opposite direction to what you're describing, i.e. toward the predator.
And even if you found a way to resolve these problems, movement is far more complex than you're describing, and practically requires a more complex structure between sensation and reaction to coordinate muscular contractions which produce movement. The sort of reflexive muscular contraction you're describing would look more like a seizure than a contraction which produces useful movement.
I'm all for preferring the simplest explanation which explains our observations, but your explanation doesn't explain what we observe in any way.
I like the explanation in terms of predator avoidance behavior. Imagine a primitive fish with eyes that can detect motion. If it sees something moving, usually it wants to get away from that thing. If you see something moving in your left eye, the best way to swim away is to send a signal for a muscle contraction in your right side, which will cause you to curl and swim to the right. So the best wiring is a direct connection from left eye to right side, and right eye to left side. The brain is built up starting from that kind of connection.