Preface

What this project is

It’s a study of how much of Shakespeare’s English survives in the writings of six American Founders: Adams, Franklin, Hamilton, Jefferson, Madison, and Washington. The methods are corpus linguistic. The substance is literary, political, and historical. The result is a set of findings about who absorbed what, in what way, and what they did with the material once they had it.

The headline finding will not surprise readers who know the period: John Adams leads six of the eleven rankings the project ran, including all three direct-evidence measures; Benjamin Franklin sits a step behind. What surprises is how they got there. Adams quotes Shakespeare directly, names him, returns to favourite lines across decades of correspondence. Franklin almost never names Shakespeare and almost never quotes him. His prose carries an older English register, the late-Stuart and early-Hanoverian idiom Franklin learned in his half-brother’s print shop in 1722. That register overlaps with Shakespeare’s more than the later Founders’ prose does, but the overlap is older-English-in-general, not specifically Shakespearean. Adams inherits a writer. Franklin inherits a way of writing.

Two other findings come out alongside. The first: Adams alone of the six Founders extends his Shakespeare engagement from quotation into the realm of political character-type. Six passages across the corpus invoke a character who can only have come from Shakespeare (Falstaff, Shylock, Hotspur, Lady Macbeth). All six are Adams. The other Founders, when they reach for character-as-type political comparisons, reach for Roman figures (Brutus, Caesar, Cassius) whose Shakespearean credit cannot be cleanly distinguished from Plutarch or from an eighteenth-century classical education.

The second: the difference between the Founders who quote Shakespeare and the ones who don’t tracks closely with when each came of age. Adams turned twenty in 1755, with a Harvard education and Shakespeare on the desk. Hamilton turned twenty in 1777, fighting a war that demanded the political reading list of Hume, Locke, Vattel, and Polybius. The Founders who finished their formative reading before the 1765 cut produced the great bulk of the project’s Shakespeare engagement. The ones who finished after produced almost none.

Hamilton and Madison are the project’s two least Shakespearean writers. Their record is the most thorough Shakespeare absence in the corpus. The contrast they create with Adams and Franklin at the top is what makes the rest of the project’s findings mean what they mean.

Who this site is for

The site is written for the general reader. Most chapters spend more time on biographical scenes (Adams in his Braintree diary at twenty-three, Washington in his General Orders at Valley Forge, Franklin in his half-brother’s Boston print shop at sixteen) than on statistical method. Where the methodology matters, I’ve tried to explain it in plain English; where the technical detail would slow the read, I’ve put it behind a foldable “for the methodologically curious” panel that opens on a click.

For readers who want the full statistical apparatus (G log-likelihood, Bonferroni correction, sample-size-matched lexical richness, Configural Frequency Analysis, the Stefanowitsch framework that the methodology rests on), the scholarly paper is downloadable from the site. It is the document the methodology lives in; the site condenses and translates it. Both are free under permissive terms.

How the site is organised

Three layers, in the spirit of America’s Public Bible, the digital-humanities project this one’s structure is modelled on:

• The essays are the commentary. Nine chapters of long-form prose. The introduction lays out the question. The methods chapter explains how the project asked it, in plain English. Five substantive-findings chapters describe what the data turned up (two modes of inheritance, the Shakespeare-only characters, reading by generation, the convergence of measures, and the Hamilton silence). A bibliography chapter rounds it out.
• The case studies are the per-finding deep dives. One specific passage or pattern per page, with the biographical context behind it. Adams quoting Brutus’s tide speech five times across thirty-eight years. Washington paraphrasing Henry V at Valley Forge. Franklin using one specific contraction at two and a half times the rate of any other Founder, beginning at sixteen. Adams at eighty-six reaching for Antony’s line over Caesar’s body to indict Franklin’s political theology.
• The explorer is the interactive layer: the catalogue of every traceable Shakespeare reference (filterable, searchable), and the Honour Test comparison view for any of fourteen common abstract nouns. These let the reader poke at the corpus directly rather than read about what others found in it.

There is no required reading order. Most readers will start with the introduction and then jump to whichever case study catches their eye. Readers in a hurry can read just the five substantive-findings essays (Two Modes, The Shakespeare-Only Characters, Reading by Generation, Eleven Ways of Looking, and The Hamilton Silence) and get the project’s argument in roughly forty-five minutes.

A note before you read

This site is two projects at once. The visible one is the Shakespeare-and-Founders study described above. The invisible one is an experiment in directing an AI coding agent (Anthropic’s Claude Code) to build a serious digital-humanities project end to end. The Python pipeline, the prose of the essays, the case studies, and the interactive components were substantially drafted by Claude under my editorial direction. The findings themselves are statistical facts about the corpus that exist independent of who wrote them up; the framing, judgment, verification, and editorial work are mine. The full disclosure is in About this project, which I’d recommend reading before the substantive essays if the question of how the work was made matters to you.

Acknowledgments

The Founders’ corpus is drawn from the Founders Online archive maintained by the National Archives, an astonishing public resource that makes a project like this one possible at all. The Shakespeare corpus comes from Project Gutenberg. The methodological framework is taken from Anatol Stefanowitsch’s open-access textbook Corpus Linguistics: A Guide to the Methodology (Language Science Press, 2020). The site’s structure and design language are modelled on Lincoln Mullen’s America’s Public Bible (Stanford University Press, 2023), with permission of inspiration only.

All historical portraits and document scans used here are in the public domain or used under permissive Creative Commons licenses. Full attribution is on the credits page.

The site, the research repository, the corpus, the analysis scripts, and the three text deliverables are available under open terms; see the papers page for downloads.

Continue to the introduction, which puts the question more carefully, or skip ahead to the Two Modes essay for the central substantive finding.