Eleven Ways of Looking

Eleven different questions

Each row in the matrix below corresponds to a different way of asking the basic question. Each method looks at a different aspect of each Founder’s writing and ranks the six Founders from most Shakespearean to least. Seven of the measures are statistical or stylistic: do the verb forms, the sentence lengths, the metaphor patterns line up with Shakespeare’s? Three of the measures are evidence-based: how many traceable Shakespeare references does the Founder produce per million words, one measure for each of the project’s three evidence layers?

• Overall ranking: the average rank position across the ten base methods below.
• Pronoun-distribution similarity: how closely does the Founder’s use of pronouns and address terms (I, you, he, she, lord, sir, friend) match Shakespeare’s?
• Old-fashioned word survival: how many forms that had largely died out by 1750 (hath, doth, thou, thee, ‘tis) does the Founder still use?
• Metaphor pattern similarity: do the Founder’s metaphors for political life match Shakespeare’s preferred patterns (body, ship, fire, plant)?
• Statistical-style overlap: on twenty different sentence-level features (sentence length, punctuation density, emotional charge, word length), how many does the Founder share with Shakespeare?
• Use of Shakespeare-coined phrases: of twenty-four phrases popularly attributed to Shakespeare, how many does the Founder use, and how often?
• Shakespearean vocabulary: how much of Shakespeare’s distinctive word choice survives in the Founder’s writing?
• Shakespearean context patterns: of Shakespeare’s distinctive way of using common abstract nouns (the words and phrases he places near honour, power, love, death), how much survives in the Founder’s usage?
• Verified references per million words: 137 hand-verified Shakespeare references (61 direct quotations + 76 by-name) divided by each Founder’s corpus size.
• Thematic character invocations per million words: Cases where a Founder names a Shakespearean character as a recognisable type without quoting the play. The project distinguishes Shakespeare-only character types (Falstaff, Shylock, Hotspur, Lady Macbeth) from Roman names (Brutus, Caesar, Cassius) whose source is ambiguous between Shakespeare and Plutarch; this measure counts the Shakespeare-only invocations only.
• Candidate-echo density per million words: MEDIUM-or-HIGH-confidence short verbatim matches (4–5 word strings with multiple distinctive Shakespeare content words) per million words of corpus.

Eleven different questions. Eleven different data sources. Each one is a separate analysis run on the corpus, not a rescoring of the same data. They could in principle disagree. Mostly, they don’t.

Where the methods agree

The matrix reads like a confirmation. Adams takes first place on six of the eleven rows, Franklin takes first on five. Together they occupy the top two positions on seven of the rows. Jefferson is third on five rows, second on three, and tied for second on one. Washington sits fourth on eight rows. Madison and Hamilton split the bottom two positions between them across the matrix, with Madison hitting last place more often than Hamilton.

This kind of agreement across independent methods is the rough equivalent, in a corpus study, of multiple independent witnesses telling the same story. Each method is exposed to different kinds of error: the pronoun-distribution measure might be fooled by genre differences (drama uses you at much higher rates than letters do); the old-fashioned-word measure might be fooled by individual Founders who happen to write more colloquially; the metaphor measure depends on which metaphors you decide to count; the statistical-style measure depends on which features you pick. The fact that all eleven methods, each with its own plausible sources of error, point at roughly the same six-Founder ranking is what makes the ranking believable. None of the methods individually is decisive; their convergence is.

Where they disagree

Read the “Use of Shakespeare-coined phrases” row above. The pattern is different from every other row:

• Adams ranks 1, the highest. He uses fifteen of the twenty-four popularly-attributed Shakespearean phrases, with 139 hits across his life.
• Franklin ranks 5 of 6, ahead of only Madison. He uses just four of them, with eleven hits total.

On most other rows in the matrix, Franklin and Adams trade first and second. On this row they sit at nearly opposite ends. The same Founder who is the most Shakespearean on the statistical-style measures is the least Shakespearean on conscious quotation.

This isn’t noise. It is the most important finding in the project compressed into one row. The conscious-coinage measure asks does the Founder reach for Shakespeare consciously: does he reach for the phrases people would recognise as Shakespearean? Adams does, constantly. Franklin almost never does.

The three evidence-based measures added after the project expanded its data layers (verified references, thematic invocations, candidate-echo density per million words) deepen the same pattern. Adams takes first on all three. Franklin slips behind Adams and Jefferson on the verified references column (he produced almost no quotation), retains his strong second on candidate-echo density (the absorbed-mode signal he is famous for), and disappears on the thematic invocations column (he never named Brutus or Caesar as a political type the way Adams or Jefferson did).

The other measures ask about register and rhythm: the verbs Franklin reaches for, the pronouns he uses, the way his sentences fall on the page. On those, Franklin is the most Shakespearean Founder. He sounds the part without ever signalling that he is reaching for it. He absorbed Shakespeare so thoroughly that the borrowing was beneath the level of citation.

Adams collected. Franklin absorbed. Both are real kinds of influence. Neither is more “influence” than the other, but it took eleven different ways of looking, and the careful comparison of where they agreed and disagreed, to see both modes at once.

Why the convergence matters

A reader could reasonably ask: why all these measures? Why not just pick the best one and report it? The answer is that for a question like this one (about how a literary inheritance moves into eighteenth-century political prose) no single measure is the right measure. The kind of evidence the question produces is multi-dimensional. A counter of named quotations sees Adams clearly and misses Franklin entirely. A counter of register and rhythm sees Franklin clearly and misses what makes Adams’s engagement with Shakespeare distinctive.

Running the eleven methods independently and looking at the pattern of their agreement is what lets the project tell a story that doesn’t depend on any one of them being right. Where they all agree (Hamilton and Madison sit at the bottom, with Madison’s silence the more thorough), the agreement carries weight no single measure could. Where they disagree in a specific structured way (Adams beats Franklin on conscious quotation; Franklin beats Adams on absorbed register), that disagreement identifies a real two-mode pattern rather than a measurement artefact.

That is the contribution this essay records. Shakespearean inheritance in the Founders splits cleanly into deliberate quotation and absorbed register, and you can see both modes operating independently in the data. The Adams Shakespeare and the Franklin Shakespeare are different Shakespeares, and the way the eleven measures agree and disagree is the evidence for that.

The full two-mode story is in Two Modes of Shakespearean Influence. The third major finding, that two of the six Founders opted out of Shakespearean inheritance altogether, is in The Hamilton Silence. For the methodologically curious, the full description of the eleven methods is in the scholarly paper.