USL Interactive Dictionary
Griffin Keystone LLC · loading…
Translate

Type English. See it in USL.

A word, a phrase or whole sentences. Every word gets its USL form, meaning and breakdown, grammar is marked by Rule 53 (past, future, not, with…), each step says how it was decided, and every link between ideas is listed with how strong its evidence is.

Start here

A language where every word is built, and every link has to be earned.

English words are accidents of history. Nothing in fog tells you it is made by cold. In USL every word is assembled from 26 letters by the same rules, which puts unrelated-looking ideas side by side. Then research decides whether they really share something.

Phase 1 · The alphabet, one letter at a time

26 letters: sound, mouth and score

Each letter is said as one short syllable, not its English name. A hyphen separates letters while you learn (Rule 25); spoken fluently they blend into one word, with the stress on the middle letters.

Quick jump · all 26 letters
Show every letter with its scores
Phase 2 · How a word is built

Six things to know

1 · Every word ends in O, A, U or I

At least two letters, no maximum (Rule 3). Two-letter forms are grammar words (Rule 53), so real words have three or more. New base words are four letters (Rule 39).

2 · Middle letters lead

The middle letters count double in a word's score (Rule 4), and they are chosen first, from the meaning (Rule 5).

3 · Case is meaning

UPPERCASE is the felt reading, lowercase the physical one, with the same score (Rule 17). KVI love · kvi heart.

4 · Prefixes and pointers

Z- the ending of, G- chronic, Y- potential, A- the beginning of, F- intense, Q- inverted (Rule 18). DA- this, HU- you, IU- I/my and more (Rule 19).

5 · Field endings

-VE emotional, -MA mathematical, -TE technical, -SO social, -BI biological, -PH physics (Rule 20).

6 · Size numbers

A number 1–10 in front is physical size, subatomic to cosmic, never how bad or good something is (Rule 21).

A word means what its definition says. The letters give it a position and a family, not a hidden message (Rule 44).

Phase 3 · Marks and fields

The seven marks go on A (Rule 8)

In four-letter words, the first letter is the field (Rule 45)

Observed in the four-letter words, not decided in advance. Unlisted letters are free for new fields. This does not apply to three-letter words.

C, D, E, H, I, O, U, V, W and X were free and showed no pattern, so Rule 52 designed them from WordNet, Princeton's sorting of English nouns and verbs into categories. A new word's field comes from the most common meaning of the English word. The letters are labels: only E's confirmed sense (deliberate structure) matches its field (made things).

Phase 4 · Grammar (Rules 53 and 54)

Existing words first, two letters for the rest

USL already has a set of small grammar words, mostly three letters starting with B (BOI is, BOO the, BJU not, BXO and…). They are ordinary dictionary words. Only where nothing existed were new grammar words made, and those have two letters. Every real word has three or more, so a two-letter word is always grammar (B1).

Time goes before the action (B2)

Nothing = now or always. ZA past · PAU will · YA may · SA going on.
I will not forget you → IBI PAU BJU ⟨forget⟩ HBO.

Order before the action: time, then can/must, then not. The space matters: ZA KVI (two words) is “loved”, ZKVI (one word) is “the ending of love” (Rule 18).

“To be” is BOI (B6)

When “is” links a thing to a description: the fog is cold → BOO CUO BOI UCO. Without BOI, “fog cold” could be read as a describing word before its word. As a helper (is running, was taken) it's carried by SA or ZA instead.

Can, must and their softer forms

PVI can · TDI must · Rule 54 adds YPVI could · YTDI should · YPAU would: Y- (potential) on the existing word, the way v10.7 built grammar words.

Not, no, never (B4)

BJU not · BDU no · BJO never, each its own word.

Pronouns (B5, Rule 54)

IBI I · HBO you · WYO they · OU it · WO them · MU we/us · VU he/she · BDA this · BZO that · BDI here · BDO there. Pointers join a word to say whose: IUKVI my love, MUKVI our love.

No plurals (B3), articles exist

Nouns don't change for number; say “many” (MBA) when it matters. Digits are sizes, not counts (Rule 21). BOO the · BPA a.

Word order and questions (B6, B7)

Doer, action, receiver: IBI KVI HBO = I love you. Describing word first. Questions: BHO who, BRI which, BYU where, BEU how, BGU why, WA what, WI when, or end with “?”.

Joining words (B8)

BXO and · BYO or (dictionary words). New, by convention: KO with · LI in · PU on · TU to · MO from · HA of · BU for.

Reader

Type any USL word, real or invented. The reader takes it apart the way the rules build it.

Quick build

Make a word from an English word

Real definitions come from Wiktionary. The letters are chosen by the rules, and you see why. Nothing joins the dictionary until it is approved.

Step by step
Live rule check
Discoveries

Links map

Proven and possible links between words, and anagram neighbours.

The rules

Grammar and how we verify

Every rule in plain English, with its exact wording.

Research

Research archive

Every letter, kinship, definition, finding and mistake on record, with sources. Downloadable.

The purpose

Why USL

What we are actually looking for, and what USL is not.

This copy

About the data

Research archive

USL, explained from the ground up

Start at the first tab. Every letter, every connection with a traffic light, every document and Copilot podcast we made, every mistake and how it was caught. Nothing here is peer reviewed.

Loading the archive…
Rules

Grammar, and how a claim earns its place

The purpose

What USL is for

To find places where two ideas English keeps in separate, unrelated words turn out to share one real mechanism, and to give each of those findings a word.

The problem with English

A natural language stores ideas as arbitrary sounds handed down through history. Heart and love share no letters, yet one is the body's seat of the other in a great many cultures. Fog and cold share no letters, yet cold is what makes fog. The links live in people's heads, not in the words, so the language can't show them to you or check them.

What USL does differently

Every word is built, by one set of rules, from 26 letters whose scores are fixed. That gives every word a position (two scores computed by one formula, Rule 9), a reading (case marks felt or physical, Rule 17), a family (prefixes, endings and size numbers derive related words the same way every time) and neighbours (its anagram partners). Being near another word is a place to look, never proof (Rule 48).

What we are actually looking for

Kinship: two concepts English treats as unrelated that research shows share one mechanism.

CUO fog ↔ UCO cold: KIN, causal. Fog forms when air cools to its dew point.
DFA immune response ↔ FDA inflammation: KIN, by definition. Inflammation is part of the immune response.
CJU bad faith ↔ JCU disinformation: KIN, by definition. Disinformation is false information spread in bad faith.

Sometimes the find is a whole family. Six roots in the P field each name one way that correct arithmetic produces a false finding: PGEO the same variable on both axes, PRZO Simpson's paradox, PZNU collider bias, PNEO the Will Rogers phenomenon, PZSO immortal time bias, PJZO the winner's curse.

Why the rules are strict

A pattern-finding system finds patterns whether or not they are real. In the first 70 anagram groups tested, 45 shared nothing, which is what you'd expect, since shuffling letters doesn't shuffle meaning. The rules exist so that a "yes" is earned: four research angles (Rule 22), evidence that grows with how surprising the claim is (Rule 48), and a comparison against chance before any count is believed (Rule 38).

What USL is not

Not a hidden code in English: scoring ordinary English words this way finds nothing (correlation about −0.02). Not meaning written in letters: a word means what its definition says (Rule 44). Not proof by sound: pitch and tone are learning aids only (Rules 2 and 41). Not a map of reality by itself: a dense cluster shows how the dictionary was built until research says otherwise.

One measured result, with its size: on 340 base words, the letter-based positivity score lines up with independent human ratings (VADER) at r = +0.53, and it survived a test against shuffled letter tables. That is a real but modest signal on a modest sample.

Where it is going

Toward a vocabulary an AI can use: every word with a precise definition, and every link labelled (kin, opposite, derived from, part of) with its evidence attached.