Old thoughts on replication

I’m redoing and expanding my old katydid animation, in response to a new piece of research that expands the parallels between the katydid cochlea and ours. More on that in a couple of days.

While looking at the old version, I came across an abandoned set of thoughts in 2016. The abandoned thoughts are worth reprinting.

= = = = = START REPRINT:

“Social” “scientists” are weeweed up because they’ve found that replication rarely works.

I hadn’t thought about this before; simply accepted the idea. It makes superficial sense. If you follow the same methods, you should get the same results.

In fact it makes no sense at all, because replication is perfectly impossible. It SHOULD happen in a computer or an abstract equation, but it NEVER happens in living reality. If you achieve replication, it means you’ve removed all life from your situation and reduced it to a computer program.

NOTHING IN NATURE expects replication.

All sensory mechanisms involve a considerable amount of repeated observation, or observation from different angles. Two eyes, two ears, two vestibular systems, two nostrils, lots of pain sensors, and so on. Many critters have more than two of the ‘basics’, like the katydid’s pair of 90-degree polarized stereo ear-pairs or the clam’s hundred eyes.

But these repeated observations are NOT designed to verify that two views or two moments are IDENTICAL. Our brain DOES NOT EXPECT the two signals to be identical, because they can’t be. That’s the WHOLE FUCKING POINT of multiple sensors. No matter how you look or listen or perform, every view of an event is different, and every event in a natural sequence, no matter how periodic the sequence, is different. Our brain calculates the differences based on spatial and temporal baselines and templates, and tries to find the pattern or meaning.

Good example right now. I got up too early after a scary dream. As I’m writing this, the coffeemaker is running its usual set of bubbles and snorts. At one point in each perk, it emits a sort of moaning sound. Because I’m twitchy from the dream [ie my template is distorted], I mislocated the moaning as a voice outside the window. Had to get up, walk slowly into the kitchen, and detect the intensity and phase of the moaning sound as I approached the stupid coffeemaker. I didn’t expect the moaning to be the SAME at each position of my body and head; I expected it to be DIFFERENT, and I knew what type of DIFFERENCE to expect. Sure enough, the moaning was just the stupid machine. Verified by LACK of replication. If the moaning had been the same at each step, I’d be fucking scared with good reason.

And even closer to the point, carefully designed systems for VERIFYING EVIDENCE, as in police and courts and juries, don’t expect replication. They expect, and calculate, a deep triangulation. Spatio-socio-temporal parallax. Cops and courts have always understood that different witnesses have different views based on innate abilities, prejudices, and position in life. They also understand that a professional criminal follows patterns but doesn’t replicate his behavior. A forger who always signed every check “James F. Halbert” wouldn’t get very far, because only a few people are named “James F. Halbert”. [Especially in quotes.] An armed robber who always aimed his gun 63 degrees ENE by the compass wouldn’t scare most tellers. A successful robber aims his gun at whoever seems most vulnerable, which is a complex judgment that he makes in an instant.

Nothing in the design of Nature, and nothing in the design of other human systems for analyzing evidence, expects replication. Only “science” expects replication and gets upset when it doesn’t occur. Well, what else did you expect?

= = = = = END REPRINT.