7 November 2022
This is my dog.
When I arrive home he comes running to greet me. It fills me with joy.
He’s a good friend.
He is actually the very definition of friendship. Loyal, unconditional, I can always count on him.
Errrr…
no, I don’t have a dog,
I just want to explain The Abstraction Stack™.
There’s no dog
I guess we can all agree there’s no such thing as a dog.
There’s this animal, sitting by my side. It was born one day and will die another day. It’s unique. There’s no other like him. In the Tangible Reality™, things and moments are all singular, multiple and impermanent.
But because of the similarities of this animal here with other animals, someone thought a long time ago that it was convenient to assign a common symbol to all of them.
Someone chose the letters d, o and g for this particular type of animal. Well, someone chose it but only for a handful of friends, because it seems that someone else had the same idea and came up with another one with the letters p, e, r, r and o.
As much as I tried, I couldn’t find who chose them, nor when or why.
I would love to send a big Thank You! note 🙏 to those people that came up with the idea of naming things. Just imagine:
“Hey, look at that animal with four legs and hair that is running there!”
“You mean the one with black and white stripes on the back?”
“No, I mean the brown one, longer hair.”
“The one with the horns?”
“No! The smaller one!”
And here comes a third guy, who was observing the dialogue in complete despair 😠:
“The dog! For God's sake. He’s referring to THE DOG! Don’t you see!?! Look at it, that one. See? Call this a DOG! Man… I cannot stand this life anymore. 🤬”
Someone please give this guy a hug 🫂. And a tip of the hat 🎩 for his 200 IQ, because he was clever enough to use a short symbol for a common thing, optimizing throughput.
But…
hmmm… 🤔
if the Spanish chose 5 letters instead of 3…
does this mean English are more clever than Spanish?
That’s not possible… 🤨
Maybe dogs are more frequent in England than in Spain…
I don’t know…
But…
“Dog” has the same number of letters than “Yes”…
Is then the probability of finding a dog in England equal to the
probability of an English guy agreeing with what you say?
In Spanish “si and “no” are equally short, giving them
equal frequencies…
Does it mean that it’s easier to agree with a Spanish guy than with an
English guy?
This one is definitely true…
Language, even at base of the Stack, the most utilitarian, holds endless wonders.
Language is like living daily in a house built thousands of years ago.
Wondering at the choices of the architects,
time travelling to remote days.
every wall echoing the voices of primordial times.
Accountants?
The base of the Abstraction Stack I will call the Accounting Layer. In the end, language was invented by accountants.
No…
Noooo….
Now that I think of it, language was probably invented by the software
engineers that designed the databases for the Neolithic accountants.
Instead of binary states in a silicon transistor, we made up a different
encoding on clay slates.🤯
Yes, language was clearly invented by software engineers:
“You take this slate and write this symbol here if they are cows and the number of cows in the field beside it.”
“But I have sheep, not cows.”
“Well, in case of sheeps you use this other symbol… no… not there, here… no… let… let me…”
When the software engineers designed the code for cows, sheep, dogs and wheat they were solving a very narrow problem: record keeping—are the names of all common livestock and commodities always monosyllabic? 🤔
The symbol “dog” will never be able to capture the complexity of the tangible object that it represents, nor will it consider that no two dogs are the same. It’s a hyper-hyper-HYPER simplification of the Tangible Reality™.
The base of the Stack is very narrow. Like VERY NARROW.
Don’t blame the engineers, though. They optimized for a very narrow problem.
“You told me record keeping! Nobody said anything about poetry! Tell me, tell me where in the requirements doc is this?”
This smells like problem, and you know where this is going…
Very old tech,
designed for a different use case,
being worked around to enable more and more features,
and used concurrently by billions of users…
and you complain about mainframes?
Analyzing language is like analyzing the instruction set of the processor we’re equipped with. We are off to a troublesome start and, as we build up our stack, our pains will only grow.
To be continued—as long as we got each other, of course…
Afterword