Accuracy isn’t enough

Somebody cited this quote from physicist Feynmann:

“We would always like to present things accurately, or at least precisely enough that they will not have to be changed when we learn more.”

Precise and accurate are distinct words in metrology, but this still isn’t enough. When we learn more we change our assumptions about what ISN’T there. When we present what we THINK is there, we’re naturally inclined to feel that we’ve finished the job. There’s nothing else beyond our description. Confidence slows down our later realization that something IS there.

Example: Two parts of the immune system were very recently seen in the brain and skull. (1) The skull contains marrow, just like leg bones, and the marrow contributes immune cells to the brain through small openings in the skull. These marrow cells are coordinated with the legbone marrow when ordered up by the immune system. (2) The lymphatic system has vessels in the brain, which connect to the rest of the lymph vessels.

Both of these are subtle but visible to the naked eye if you KNOW WHAT YOU’RE LOOKING FOR. They don’t require electron microscopes or MRIs to see them. Despite their visibility, centuries of anatomists never recorded them, or perhaps saw them and dismissed them as meaningless variations. We were ASSUMING that the skull was simple and solid, and the brain had no connection to the immune system. We “knew” what ISN’T there, so we didn’t bother looking for it.

If we’re operating the other way around, blackboxing the system’s BEHAVIOR, we’re less likely to be blinded. Especially true in anatomy and medicine where the behavior is caused by an infinitely complex set of feedback loops that are absolutely unobservable.

In engineering and in life, the PURPOSE of a feedback loop is to make the system work regardless of the internal details, within limits. You can change some of the details, remove some of the pieces, and the system will quickly adjust to the new configuration.

In other words, we should stick with verbs, not nouns. Transfer functions, not statistics.

THEORY KILLS. EXPERIENCE SURVIVES.

= = = = =

Sync: Immediately after posting this, I checked email and saw an ad for an article in NYTimes:

Research has piled up in recent years to debunk the supposed benefits of moderate drinking, concluding that no amount of alcohol is good for you.

For 40 years all of “science” has operated on the UNQUESTIONED AND UNQUESTIONABLE assumption that one molecule of tobacco somewhere in the universe is instantly deadly for everyone, while a few drinks are helpful. These assumptions were driven by paychecks and fashion, not forgivable theorigenic blindness. Simple observation of the TRANSFER FUNCTION leads to the opposite conclusion. Heavy drinking kills people FAST, often in the 30s or 40s. Heavy smoking kills SLOW, usually taking two or three years off lifespan. Alcohol also eats away the brain and leads to early senility, while tobacco improves brain function.

= = = = =

Carver tells us to focus on the transfer function.

Look about you.
Take hold of the things that are here.
Talk to them.
Let them talk to you.