The Em Dash

What It Covers

The em dash punctuation mark faces modern stigma as an AI writing indicator, despite its 500-year literary history from Shakespeare to Emily Dickinson. The episode traces how this versatile mark evolved from Elizabethan theater through nineteenth-century novels, why large language models overuse it, and introduces the am dash as a human-proof alternative.

Key Questions Answered

• •Historical punctuation evolution: Italian scholar Boncompagno da Signa created the virgula plana horizontal dash in the eleventh century as a sentence-ending mark. Shakespeare's First Folio extensively used dashes to indicate speech interruptions and character thought patterns in plays like King Lear, establishing theatrical conventions that persist in modern stage writing and demonstrating early flexibility in punctuation rules.
• •Novel realism techniques: Eighteenth-century novelists like Laurence Sterne in Tristram Shandy (1759) deployed dashes to create stream-of-consciousness narratives and censor sensitive content with first-letter-plus-dashes formatting. Jane Austen used this technique in Pride and Prejudice to redact military regiment names, adding perceived authenticity to fictional stories while titillating readers with hidden meanings requiring careful interpretation.
• •Peak usage statistics: Dash usage exploded in nineteenth-century English literature, with Oliver Twist containing one dash every 224 words, Moby Dick one per 129 words, and Jane Eyre one per 90 words. Emily Dickinson wrote 1,800 poems with thousands of dashes, though her first posthumous collection editors removed 1,099 of 1,151 dashes to make work more readable for 1890s audiences.
• •AI training data theory: Large language models likely adopted excessive em dash usage between November 2022 and July 2024 when companies like Anthropic expanded training beyond web data to include millions of destructively scanned print books from the late 1800s and early 1900s. Court documents revealed this process involved physically cutting book bindings to digitize classic literature containing period-appropriate high dash frequency.
• •Human-proof alternative solution: Sydney agency Cocogun created the am dash, a curved variant resembling a mustache, available through downloadable fonts Times New Human and A Real. Users type am-hyphen to insert the mark, which remains statistically improbable for language models to generate due to extreme scarcity in training data, offering symbolic resistance to AI-generated text while maintaining punctuation functionality.