[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-study-links-englishs-80-percent-redundancy-to-text-structure":10,"sections":35},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},9279,"study-links-englishs-80-percent-redundancy-to-text-structure","Study Links English's 80 Percent Redundancy to Text Structure","Researchers built semantic trees from text and found their structural entropy nearly matches language models' measured redundancy.","English prose burns through far less information than its alphabet could carry - and a new paper says it knows why.\n\nResearchers used large language models to repeatedly split texts into semantically coherent chunks, building what they call a \"semantic tree\" that runs from a story's main idea all the way down to individual words. From the statistics of these trees across different kinds of writing, they calculated a new \"structural entropy rate\" - essentially a measure of how surprising the next piece of text is, based on how its ideas are organized rather than just its words. For several datasets, that number landed close to the redundancy already measured by having LLMs predict the next word in a passage. Children's stories produced flatter, simpler trees with less branching and lower entropy, while literary fiction and poetry branched out more and scored higher.\n\nThis matters because the 80 percent redundancy figure for English - the idea that each letter carries about 1 bit of information instead of the 4.8 bits an alphabet could theoretically hold - has been around since Shannon estimated it decades ago, and LLMs have confirmed it without explaining it. This paper offers an explanation: it is not just which words follow which, it is how ideas nest inside each other. That has real implications for text-chunking strategies in retrieval systems and for arguments about why some writing reads as simple and some as dense.\n\nIt is a tidy, testable idea, but it is also, so far, a clean story told on a handful of curated datasets - genre categories like \"poetry\" and \"children's stories\" do a lot of definitional heavy lifting here.","[\"ai research\",\"nlp\",\"information theory\",\"linguistics\"]","2026-10-01T04:00:00.000Z","2026-10-02T06:42:45.225Z","2026-10-02T06:42:48.693Z","published",null,[],"ai",[26,27,28,29],"ai research","nlp","information theory","linguistics",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.13194",0,{"sections":36},[37,40,44,48,53,58,62,67,72,76,81,86,91,96],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",5671,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",820,{"name":45,"slug":46,"count":47,"latest_published_at":18},"Policy","policy",430,{"name":49,"slug":50,"count":51,"latest_published_at":52},"Deals","deals",298,"2026-09-30T21:00:26.000Z",{"name":54,"slug":55,"count":56,"latest_published_at":57},"Hardware","hardware",196,"2026-09-30T13:00:00.000Z",{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",165,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",149,"2026-09-30T22:57:11.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Dev Tools","dev-tools",93,"2026-10-01T02:30:48.000Z",{"name":73,"slug":74,"count":70,"latest_published_at":75},"Software","software","2026-09-30T21:41:11.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",84,"2026-09-30T20:39:09.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",51,"2026-09-30T16:24:30.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]