The Convergence Analysis

The lines were assembled by this site, not by a published scholar. That is a methodological fact that deserves direct acknowledgment: a critic can reasonably ask not just whether the lines were chosen to produce a particular result, but why these lines, and not some other set that might have produced a different winner.

The generative principle. The lines were derived by asking: what does the text itself, the publication history, and the historical record jointly require of any plausible author? That is inductive logic, working from the works outward to the author, rather than selecting a candidate and identifying what they happen to fit. Each line corresponds to a documented feature of the canon or its historical context that constrains who could plausibly have written it: the Italian plays require Italian knowledge of a specific kind; the Montaigne verbal precision requires access to a specific unpublished manuscript; the 1593 publication cluster requires a specific biographical fact. No line asks a question that only one candidate can in principle answer. Each line was also identified by scholars working on questions other than authorship, literary historians, musicologists, legal historians, geographers, rather than by authorship advocates. A line whose primary proponents have an authorship stake does not qualify: the risk of motivated selection cannot be separated from the argument.

Methodological diversity as a constraint on cherry-picking. The lines span five distinct evidential types: geographic and linguistic knowledge (lines 2a, 2b), documented biographical facts (lines 1, 4, 5, 16, 17), textual and stylistic characteristics (lines 6, 9, 10, 12, 14, 15), social and institutional position (lines 3, 7, 11), and literary range and patronage (lines 8, 13, 18a). The diversity is itself part of the design: it is much harder to rig a framework across five methodologically independent evidence types than to cherry-pick within a single type. A candidate who scores well across all five types is more robustly evidenced than one who scores well only within one.

What was considered and excluded. Stylometrics was considered and excluded. The Zeta test methodology underpinning the main attribution claims has been shown to produce unreliable results under Early Modern drama conditions, where genre, collaboration, and manuscript transmission confound the signal. Including an unreliable line would add noise rather than information. Biographical correspondences of the kind assembled by advocacy groups were also excluded: a correspondence between a play's content and a candidate's biography proves nothing unless it depends on specialised knowledge that could not have been acquired by literary means.

The honest residual problem. There are certainly other defensible lines that could have been included. A framework with different but equally justifiable lines might weight candidates differently. The response to that concern is not to claim the lines are the only possible set; they are not, but to invite scrutiny of each one: is any line framed in a way that advantages a particular candidate unfairly? Is there a well-evidenced line of inquiry that was excluded and whose inclusion would materially change the results? The scoring rationale in each tooltip is there to support exactly that challenge.

The flagged lines (⚑). Four lines carry a structural warning. Lines 1 and 4 are nested: both test Florio's documented connection to Southampton's household, with line 4 the more specific version. They are not fully independent, treating them as two separate pieces of evidence double-counts a single primary source document. Lines 2a and 2b were split from what could have been a single Italian knowledge line. The split is defensible, topographical knowledge of Italy (acquired through travel) and linguistic or literary knowledge of Italian (acquired through language) are genuinely distinct and a candidate can possess one without the other, but the split doubles the Italian advantage relative to a single combined line. A sceptical reader is entitled to treat lines 1+4 as one line and lines 2a+2b as one line, reducing the framework to sixteen independent lines. On that reading, Florio's lead is smaller but survives. Line 18a is not flagged. It measures one thing, dedications received as a marker of patronage standing, and is weighed lightly on that understanding.

Two further concerns, not flagged but worth naming. Line 5 (the 1593 cluster) is the only line that fires uniquely for one candidate: it asks whether a specific set of dated biographical facts converge, and those facts are Marlovian by definition. Every other line asks a general authorship question; line 5 asks whether you fit a specific historical theory. Line 11 (feudal and aristocratic outlook) embeds an interpretive assumption, that the plays require an insider aristocratic perspective rather than knowledge of that world acquired by other means. That assumption is contested. Readers who dispute it should weight line 11 accordingly.

On documentary discovery and forgery. A single primary source document, however compelling, faces an immediate challenge in this field: the history of Shakespeare scholarship includes serious forgeries, the Ireland papers of the 1790s, the Collier forgeries of the 1840s, that were initially accepted by experts. That history means any smoking-gun discovery will attract the forgery question as a matter of course, regardless of its actual provenance. The convergence framework is the appropriate response to both sides of that problem. A forger would need to have fabricated not one document but agreement across multiple independent methodological strands drawn from different archives, different document types, and different scholarly traditions, a qualitatively different and far more implausible undertaking. Equally, a genuine documentary discovery does not stand alone: it either fits the existing pattern of evidence across the lines or it creates new tensions that need explaining. Convergence gives you a way to evaluate a new document rather than simply accept or reject it on impact. This analysis does not pre-empt documentary discovery. A properly authenticated document that connected a specific candidate to the authorship directly would require this framework to be reassessed, and it would be.

The strongest evidence against motivated selection is, finally, the results themselves. A framework designed to produce de Vere as the winner would not score Florio ✓ on Southampton connection (line 1), Italian linguistic knowledge (line 2b), pre-1603 Montaigne (line 6), and botanical knowledge (line 14), while scoring de Vere ◑ or ~ on the same lines. It would not score Bacon ✓ on legal knowledge (line 7) and theological range (line 13). It would not score Mary Sidney Herbert ✓ on literary dedications received (line 18a). A rigged framework does not hand the strongest results on multiple lines to candidates other than the intended winner.

Convergence of circumstantial evidence across methodologically diverse lines is not the same as primary source proof. A candidate who scores ✓ across many lines has a strong circumstantial case, multiple independent questions about the authorship return answers consistent with their candidacy. That is meaningful. It is not, however, the same as a document produced during the period of composition connecting that candidate to a specific play.

The framework is designed to show where the evidence points, not to establish who the author was. No candidate is proved by this table. The candidate who scores highest has the strongest circumstantial case; they do not have a demonstrated attribution.

Each line counts as one point. That is a choice, not a default, and it deserves an explanation.

The alternative, differential weighting, where some lines count for more than others, sounds more sophisticated. It is actually more dangerous. To assign a weight to a line, you must first decide how much that type of evidence matters relative to the others. That decision is not derivable from the evidence itself: it is a prior judgment. And prior judgments in the authorship debate almost always carry an advocacy position inside them. A reader who believes aristocratic insider knowledge is the key to the plays will naturally weight line 11 heavily; a reader committed to Italian travel as the clinching criterion will weight lines 2a and 2b accordingly. Neither is acting dishonestly, but both have already imported the conclusion into the weighting before the evidence is examined. Differential weighting doesn’t refine the analysis; it relocates the advocacy from the candidacy claim to the methodology.

Equal weighting removes that bias. Every line poses a question about what the works require; each answer counts once. The framework doesn’t decide in advance that biography matters more than textual knowledge, or that chronology outweighs legal expertise. It asks its questions and adds up the answers. If the result strikes a reader as wrong, the challenge is available at the line level, the tooltip for every cell shows the evidence and the reasoning. A reader who thinks line 11 is overweighted can discard it; a reader who thinks lines 2a and 2b double-count Italian knowledge can treat them as one. The underlying evidence is all there. What equal weighting prevents is the invisible move of pre-deciding which questions matter most before looking at who answers them.

The two frameworks on this site, Price's literary paper trail and this convergence analysis, measure different things, and they can and do pull in different directions for some candidates. That disagreement is a diagnostic signal, not a flaw in the method.

When both frameworks point the same direction, the case is as robust as this site's methodology can make it: multiple independent analytical lenses returning the same answer. When two agree and one dissents, the dissent is worth examining; it may reflect a genuine evidential gap on that dimension, or a contested interpretive assumption built into that particular approach. When they diverge substantially, the honest reading is that the evidence presents differently under different analytical approaches, and no single score should be treated as a verdict.

Two examples from this table illustrate the diagnostic value. Bacon scores strongly on the paper trail framework; he left an extensive documented literary paper trail, which is precisely what Price's framework rewards, but more modestly on convergence, because the authorship-specific evidence (Italian knowledge, Southampton connection, literary register match) is largely absent. That pattern, strong on documentary presence but less strong on authorship-specific evidence, is itself informative: it shows what the Baconian case does and does not establish. Mary Sidney Herbert's paper trail score is among the strongest of all candidates on this site; her convergence score is lower, in part because several convergence lines assume public theatrical engagement that her gender and social position make structurally harder to document. That pattern illuminates a limit of the convergence framework for candidates whose access to literary life was structurally constrained, not a limit of the candidate.

Reading the frameworks together, rather than relying on any single score, gives the most accurate picture of the evidentiary state for each candidate.