Cognitive neuroscience likes to assign functions to parts of the brain. This bit does planning. This part does short term memory. This piece perceives faces.
Does that bit really do planning? If you cut it out and held it, would it plan for you?
No. And that other bit doesn’t perceive faces without messages from the eyes and other parts, and further interpretation by yet more parts and the whole person.
Brain sections don’t plan, remember, or perceive. A person plans, remembers, perceives. Those sections of the brains participate. They’re essential, but not sufficient.
Maybe I develop a service that does credit scoring for a loan company. “We do the credit scoring,” I might say.
“Who are your users?”
“The mainframe app calls our service, without any personal identifying information, and we reply with a score.”
Um. My service doesn’t “do credit scoring,” now does it? It participates in credit scoring. But without that mainframe app and the rest of the system, my service does nothing, like a frontal lobe on the dissection table.
It’s tempting to attribute functions of a system to parts of that system, parts that are uniquely active in that function. Those parts may be necessary, but they are not sufficient.
In fancy words, this is called the “mereological semantic fallacy.” Mereology is the examination of part-whole relations. Semantic has to do with meaning. This is a fallacy that attributes meaning to parts, when that meaning doesn’t exist without the whole.
It takes the entire loan company software system and the people entering the data and the credit bureaus we interface with in order to do credit scoring. It takes a whole person to plan, remember, perceive.
We don’t design systems, we design subsystems.
Jabe Bloom
Care about your environment. Practice responsible change (backwards compatibility, advocacy). Get to know the people or software who use your service. Because it doesn’t do anything alone.
This post comes from Turvey, Lectures on Perception (the part you can get free in the Kindle Sample).