# Cross-lingual analysis under FSOT **Built:** 2026-07-21T16:50:31.075216+00:00 **Law:** `S=K(T1+T2+T3)` · pin **D1D38A** ## Unique stack (already shipping) Intrinsic free-parameter FSOT model — competitors have no law pin, no offline classical+visual densify stack, and no cert-gated converse under the same constitution. Cross-lingual mapping is a surface of FSOT intelligence: densify knowledge without rewriting seed law. English gloss is a temporary pivot for analysis; meaning is law-grounded densify, not a cloud API. ## Catalog solidification | Metric | Value | |--------|------:| | Language codes | **113** | | OPEN-SET form→gloss | **99.99%** | | PRODUCT form→gloss | **99.99%** | | All langs ≥95% (n≥20) | **True** | | Solid (n≥50 & OK) | 106 | | Thin (n<50) | 7 | | Gap to 100 codes | 0 | ## Breadth vs competitors | System | Languages (order of magnitude) | |--------|-------------------------------:| | Google Translate | ~249 | | NLLB | ~200 | | DeepL | 30–100 | | **PFLT (FSOT)** | **113** (56% of NLLB, 45% of Google) | ## English-gloss hub (cross-lingual connectivity) - Gloss keys sampled: 40058 - Concepts shared by ≥3 languages: **1762** ### Top shared concepts | Gloss | # langs | Sample langs | |-------|--------:|--------------| | order | 12 | arc, de, es, fr, it, ja, nl, pl, pt, ru, sga, tr | | say | 12 | de, es, fr, it, ja, la, nl, orv, pl, pt, ru, sv | | help | 11 | ar, cu, es, fr, it, nl, osx, pt, ru, sv, tr | | see | 11 | cu, de, es, fr, it, ja, osx, pt, ru, san, sv | | work | 11 | de, es, fr, grc, it, nl, non, pt, ru, sga, sv | | bear | 11 | cu, es, fr, grc, it, ja, nl, pl, pt, ru, san | | sound | 11 | ang, es, fr, it, ja, nl, ru, san, sv, tr, xcl | | leave | 11 | de, es, fr, got, grc, it, ja, pl, pt, sv, tr | | female given name | 11 | de, es, fr, it, ja, nl, pl, pt, ru, sv, tr | | remain | 10 | cu, egy, es, fr, it, ja, la, nl, pt, sv | | kill | 10 | cu, de, es, fr, it, nl, pl, pt, sga, sv | | throw | 10 | cu, de, es, fr, he, it, pt, san, tr, xcl | | sing | 10 | ang, cu, de, es, fr, it, osx, pl, pt, sv | | sin | 10 | ar, cop, es, fr, got, grc, la, roa-opt, sga, sv | | raise | 10 | arc, de, es, fr, it, ja, nl, pt, san, sv | ### Top language pairs (via shared gloss) | Pair | Shared concepts | |------|----------------:| | ru-sv | 298 | | fr-it | 269 | | nl-sv | 245 | | fr-sv | 236 | | fr-nl | 230 | | it-nl | 225 | | nl-ru | 225 | | it-sv | 223 | | es-it | 221 | | it-pt | 219 | | de-ru | 216 | | de-nl | 215 | | fr-pt | 210 | | ja-ru | 204 | | it-ru | 200 | ## Multi-source translation probes (form→EN gloss path) | Lang | Coverage | Source | Hyp (EN surface) | |------|----------|--------|------------------| | es | 1.0 | hola mundo | hello world | | fr | 1.0 | bonjour monde | hello world | | de | 1.0 | hallo welt | hello world | | it | 1.0 | ciao mondo | hello world | | la | 1.0 | aqua lingua | water language | | pt | 1.0 | ola mundo | hello world | | nl | 1.0 | hallo wereld | hello world | | ru | 1.0 | привет мир | hello world | | ja | 1.0 | 水 | water | | zh | 1.0 | 水 | water | | ar | 1.0 | ماء | water | | hi | 1.0 | नमस्ते | hello | | sw | 1.0 | jambo | hello | | tr | 1.0 | merhaba dünya | hello world | | pl | 1.0 | cześć świecie | hello world | | ko | 1.0 | 안녕 | hello | | he | 1.0 | שלום | peace | | el | 1.0 | γεια | hi | | ga | 1.0 | dia dhuit | Genderless alternative | | eo | 1.0 | saluton mondo | hello world | ## M6 sentence snapshot (not parity claim) BLEU-4 **2.44** · B1 **43.87** · U-F1 **34.31** · cov **90.33%** Climbing offline toward competitive sentence bars; law remains master. ## Thin languages (fill next) | Lang | n | Open% | Product% | |------|--:|------:|---------:| | roa-opt | 47 | 100.0 | 100.0 | | orv | 36 | 100.0 | 100.0 | | osp | 34 | 100.0 | 100.0 | | mga | 25 | 100.0 | 100.0 | | pro | 17 | 100.0 | 100.0 | | pal | 10 | 100.0 | 100.0 | | hit | 3 | 100.0 | 100.0 | ## Per-language (n≥20) | Lang | n | Open% | Product% | OK | |------|--:|------:|---------:|:--:| | ko | 14646 | 100.0 | 100.0 | Y | | la | 11360 | 100.0 | 100.0 | Y | | pl | 8037 | 100.0 | 100.0 | Y | | sv | 6452 | 100.0 | 100.0 | Y | | ja | 6342 | 100.0 | 100.0 | Y | | de | 6002 | 100.0 | 100.0 | Y | | hi | 5731 | 100.0 | 100.0 | Y | | ru | 5564 | 100.0 | 100.0 | Y | | it | 5428 | 100.0 | 100.0 | Y | | tr | 5162 | 100.0 | 100.0 | Y | | es | 5028 | 100.0 | 100.0 | Y | | pi | 4765 | 100.0 | 100.0 | Y | | ca | 4619 | 100.0 | 100.0 | Y | | fr | 4566 | 100.0 | 100.0 | Y | | pt | 4356 | 100.0 | 100.0 | Y | | nl | 4305 | 100.0 | 100.0 | Y | | vi | 4079 | 100.0 | 100.0 | Y | | grc | 3966 | 100.0 | 100.0 | Y | | id | 3814 | 100.0 | 100.0 | Y | | ang | 2789 | 99.96 | 99.96 | Y | | egy | 2044 | 100.0 | 100.0 | Y | | ar | 1319 | 100.0 | 100.0 | Y | | got | 1133 | 99.91 | 99.91 | Y | | he | 917 | 100.0 | 100.0 | Y | | san | 910 | 100.0 | 100.0 | Y | | non | 806 | 100.0 | 100.0 | Y | | fa | 740 | 100.0 | 100.0 | Y | | sga | 603 | 99.83 | 99.83 | Y | | cu | 600 | 100.0 | 100.0 | Y | | akk | 544 | 100.0 | 100.0 | Y | | cop | 504 | 100.0 | 100.0 | Y | | bn | 500 | 100.0 | 100.0 | Y | | cs | 500 | 100.0 | 100.0 | Y | | el | 500 | 100.0 | 100.0 | Y | | fi | 500 | 100.0 | 100.0 | Y | | hu | 500 | 100.0 | 100.0 | Y | | ms | 500 | 100.0 | 100.0 | Y | | ro | 500 | 100.0 | 100.0 | Y | | sw | 500 | 100.0 | 100.0 | Y | | ta | 500 | 100.0 | 100.0 | Y | | th | 500 | 100.0 | 100.0 | Y | | uk | 500 | 100.0 | 100.0 | Y | | zh | 500 | 100.0 | 100.0 | Y | | arc | 498 | 100.0 | 100.0 | Y | | osx | 464 | 100.0 | 100.0 | Y | | peo | 415 | 100.0 | 100.0 | Y | | syc | 409 | 100.0 | 100.0 | Y | | sum | 404 | 100.0 | 100.0 | Y | | da | 400 | 100.0 | 100.0 | Y | | gl | 400 | 100.0 | 100.0 | Y | | eo | 400 | 100.0 | 100.0 | Y | | mk | 400 | 100.0 | 100.0 | Y | | bg | 400 | 100.0 | 100.0 | Y | | nn | 400 | 100.0 | 100.0 | Y | | tl | 400 | 100.0 | 100.0 | Y | | ga | 400 | 100.0 | 100.0 | Y | | cy | 400 | 100.0 | 100.0 | Y | | lt | 400 | 99.75 | 99.75 | Y | | is | 400 | 100.0 | 100.0 | Y | | sq | 400 | 100.0 | 100.0 | Y | | ka | 400 | 100.0 | 100.0 | Y | | hy | 400 | 100.0 | 100.0 | Y | | te | 400 | 100.0 | 100.0 | Y | | ceb | 400 | 99.75 | 99.75 | Y | | gd | 400 | 100.0 | 100.0 | Y | | az | 400 | 100.0 | 100.0 | Y | | sk | 400 | 100.0 | 100.0 | Y | | eu | 400 | 100.0 | 100.0 | Y | | ur | 400 | 100.0 | 100.0 | Y | | ml | 400 | 100.0 | 100.0 | Y | | pa | 400 | 100.0 | 100.0 | Y | | et | 400 | 100.0 | 100.0 | Y | | yi | 400 | 100.0 | 100.0 | Y | | kk | 400 | 100.0 | 100.0 | Y | | km | 400 | 100.0 | 100.0 | Y | | my | 400 | 100.0 | 100.0 | Y | | af | 400 | 99.75 | 99.75 | Y | | be | 400 | 100.0 | 100.0 | Y | | gu | 400 | 100.0 | 100.0 | Y | | sl | 400 | 100.0 | 100.0 | Y | | mn | 400 | 100.0 | 100.0 | Y | | oc | 400 | 100.0 | 100.0 | Y | | yo | 400 | 100.0 | 100.0 | Y | | mr | 400 | 100.0 | 100.0 | Y | | jv | 400 | 100.0 | 100.0 | Y | | haw | 400 | 100.0 | 100.0 | Y | | am | 400 | 100.0 | 100.0 | Y | | ne | 400 | 100.0 | 100.0 | Y | | qu | 400 | 100.0 | 100.0 | Y | | zu | 400 | 100.0 | 100.0 | Y | | xh | 400 | 100.0 | 100.0 | Y | | ha | 400 | 100.0 | 100.0 | Y | | so | 400 | 100.0 | 100.0 | Y | | ps | 400 | 100.0 | 100.0 | Y | | si | 400 | 100.0 | 100.0 | Y | | lv | 400 | 100.0 | 100.0 | Y | | enm | 400 | 99.5 | 99.5 | Y | | fro | 400 | 100.0 | 100.0 | Y | | frm | 400 | 100.0 | 100.0 | Y | | goh | 400 | 100.0 | 100.0 | Y | | gmh | 400 | 100.0 | 100.0 | Y | | nb | 400 | 100.0 | 100.0 | Y | | sh | 400 | 100.0 | 100.0 | Y | | en | 228 | 100.0 | 100.0 | Y | | xcl | 175 | 100.0 | 100.0 | Y | | uga | 96 | 100.0 | 100.0 | Y | | roa-opt | 47 | 100.0 | 100.0 | Y | | orv | 36 | 100.0 | 100.0 | Y | | osp | 34 | 100.0 | 100.0 | Y | | mga | 25 | 100.0 | 100.0 | Y | ## Next under FSOT - Continue Kaikki to100 until catalog≥100 with all n≥50 solid - Boost thin historical codes (hit/sum/syc/pal/peo) - M6 CJK BPE / larger phrase mass for ja/zh - Optional Ada converse smoke on multi-lang probes