| ch | title | source | state |
| 1 | The Grammar of Notation — Chomsky's footnote on paired parentheses, proved in Lean; 6÷2(1+2) has two trees; 11,808 formulas here nest at most 3 deep | Chomsky 1957; Pullum 2011; J&M 2026 | live |
| 2 | How Mathematical Prose Parses — imperatives (38x fewer than arXiv proofs; the recipes are in the Lean), 'orthogonal' as a collective predicate (Lean), ambiguity in bits | Tanswell & Inglis; Arambillete & de Groote; Balkır; Corneli et al.; Orth; Fatima | live |
| 3 | How Many Senses a Word Has — the IJL entry template (C→K→F→U) checked: co-predication must be transitive (Lean), the head sense is not the most frequent (WordNet/SemCor) | IJL v3/v4; WordNet 3.0 | live |
| 4 | Translation Keys — the operator letters across 822 pages as word clusters: one head sense per key, domain fillers, a colliding U; holology passes the test | the corpus; GCM Institutional Edition (2025) | live |
| 5 | The Machines That Read Us — the corpus's four reading tools as textbook NLP; can a classifier file the pages? 0.68 vs 0.45 baseline, macro-F1 0.39 | Jurafsky & Martin, 3rd ed. draft (2026) | live |
| 6 | Words That Went Wrong — nine words and numbers the corpus once misused (sorry, degenerate, proved, exactly, there is no, 666, 110, proof, U), each with its repair and a count of usage now | the corpus; WP-94; WP-61 | live |
| ref | Key Register — generated dictionary of the operator keys C, K, F, U: head sense, key set, fillers by domain | tools/key_register.py | live |
| 7 | Maximum Meaning — the NASA Ames paper on Sanskrit (Briggs 1985): a compressed language, a redundant analysis, 5,040 word orders with one meaning | Briggs 1985 (held) | live |
| 8 | Six Lines — Hardy & Wright's proof that √2 is irrational, which Wiedijk said no prover could read, in Lean one sentence per step (49 words, 7 lines) | Kahle, Towards the Structure of Mathematical Proof | live |
| 9 | What Types Decide — Leibniz 1684 to I❤LA: one symbol, many laws; Lean: × commutes on ℕ, not on matrices; this series uses the past tense 21× more than Mathematics in Lean | Leibniz; Avigad; Aberdein/Ganesalingam; Ganesalingam & Gowers; Macmillan; I❤LA; Stathopoulos & Teufel; Matsuzaki et al. | live |
| 10 | Where Counting Meets Chance — a primer: a chain of n steps holds with chance pn (14 steps at 95% is worse than a coin flip), a swarm of checkers misses with chance (1−q)n; applied to the Swarm Simulator, where shared intent is exactly 0.416t and the printed contraction bound has no global form | Nogueira Grossi (Swarm Simulator); Wolfram | live |