{"title":"Em dashes and AI artifacts in pre-computer literature","measuredAt":"2026-08-16","license":"CC0-1.0 — use it, no attribution required, a link is appreciated","canonical":"https://claudewatermark.xyz/guides/moby-dick-87-em-dashes","method":{"endpoint":"https://claudewatermark.xyz/api/process","request":{"text":"<sample>","rewrite":false},"sampleChars":20000,"sampleOffsetChars":3000,"preprocessing":"Strip HTML where present, discard the Project Gutenberg licence header, collapse whitespace.","note":"The endpoint needs no key and no account. Re-running this should reproduce these counts exactly, provided the same offset is used."},"findings":{"totalWords":36173,"works":10,"totalDeterministicArtifacts":0,"emDashesPer1000WordsRange":{"min":0,"max":26.4},"caveats":["One sample per work. Sufficient to show the deterministic classes do not fire on clean human prose; NOT sufficient to publish a false-positive rate.","Project Gutenberg text has been normalised over decades, which is a favourable case for the artifact classes. Wikipedia and an IETF RFC were tested separately as harder cases and also returned zero.","Non-breaking spaces are byte-exact but occur in ordinary web copy (measured on bbc.com, gov.uk, smashingmagazine.com), so that class is not evidence a chat interface was involved.","None of this concerns Anthropic's statistical watermark, which nobody outside Anthropic can currently detect."]},"rows":[{"work":"Pride and Prejudice","year":1813,"source":"gutenberg:1342","words":3423,"deterministicArtifacts":0,"emDashes":0,"smartPunctuation":61,"emDashesPer1000Words":0},{"work":"Frankenstein","year":1818,"source":"gutenberg:84","words":3587,"deterministicArtifacts":0,"emDashes":8,"smartPunctuation":32,"emDashesPer1000Words":2.2},{"work":"Moby Dick","year":1851,"source":"gutenberg:2701","words":3366,"deterministicArtifacts":0,"emDashes":89,"smartPunctuation":252,"emDashesPer1000Words":26.4},{"work":"Great Expectations","year":1861,"source":"gutenberg:1400","words":3783,"deterministicArtifacts":0,"emDashes":23,"smartPunctuation":247,"emDashesPer1000Words":6.1},{"work":"Alice's Adventures in Wonderland","year":1865,"source":"gutenberg:11","words":3854,"deterministicArtifacts":0,"emDashes":28,"smartPunctuation":269,"emDashesPer1000Words":7.3},{"work":"War and Peace (translation)","year":1869,"source":"gutenberg:2600","words":3535,"deterministicArtifacts":0,"emDashes":7,"smartPunctuation":163,"emDashesPer1000Words":2},{"work":"The Picture of Dorian Gray","year":1890,"source":"gutenberg:174","words":3684,"deterministicArtifacts":0,"emDashes":25,"smartPunctuation":180,"emDashesPer1000Words":6.8},{"work":"The Adventures of Sherlock Holmes","year":1892,"source":"gutenberg:1661","words":3635,"deterministicArtifacts":0,"emDashes":14,"smartPunctuation":326,"emDashesPer1000Words":3.9},{"work":"Dracula","year":1897,"source":"gutenberg:345","words":3707,"deterministicArtifacts":0,"emDashes":0,"smartPunctuation":64,"emDashesPer1000Words":0},{"work":"Ulysses","year":1922,"source":"gutenberg:4300","words":3599,"deterministicArtifacts":0,"emDashes":90,"smartPunctuation":96,"emDashesPer1000Words":25}]}