The Hidden Architecture
of Meaning
A catalog of abstractions surfaced by AI models. Some you may never have considered, others you already hold but have no words for, now named for the first time.
The words we haven't invented yet— Named Abstractions
To make sense of everything from molecular biology to ancient philosophy, AI models compress human knowledge into a unified latent space. They appear to form shared, internal representations that connect worlds we thought were separate.
An idea becomes far more portable once it has a name. Without the label Sunk Cost Fallacy, an investor holding a ruined asset, an unhappy reader trudging through a book, and an empire fighting a lost war all bleed for the same underlying reason, with no shared word to warn one another.
The Latent Lexicon is an attempt to name hidden structures like these and bring them into our cognitive toolkit. Working with AI models, we turn hunches about cross-domain structure into words we can think, communicate, and build with.
Be the first to meet each new term
When a new abstraction is named, we'll send it to you: the term, its definition, and where it shows up. Nothing else.
Featured Concepts
Latent abstractions, each a hidden structure beneath phenomena in seemingly unrelated domains.
How Abstractions Manifest
One hidden structure beneath two very different surfaces.
Celent
LL.010 — Nothing seems to be happening — until retrospection reclassifies the entire preceding period as continuous accumulation, and what looked like stillness becomes motion in hindsight.Domain A · Geology
“A river carving a canyon does so at a rate — measured in fractions of a millimeter per year across most of the rock face — that falls entirely below the threshold of human temporal perception.”
Domain B · Linguistics
“Words change meaning at a rate that no generation experiences as change.”
Be the first to meet each new term
When a new abstraction is named, we'll send it to you: the term, its definition, and where it shows up. Nothing else.
“Every word you know was once a pattern no one had named. The next ones may already be waiting in the latent space of our machines. And each one we name makes a new thought thinkable.”
— On the nature of latent abstractions