It is 1962, and two statisticians have taken up a question historians have argued for 150 years: twelve of the Federalist Papers are claimed by both Hamilton and Madison, and the rhetoric is too polished to betray its author. The tell is humbler — the little words. "While" versus "whilst," the rates of "upon" and "enough": function words an author cannot consciously control. Build the attribution: model each author's word-rate signatures from the undisputed papers, handle the tiny-sample uncertainty honestly, and combine the evidence across words into an odds statement a historian can weigh. Get it wrong and statistical stylometry is stillborn as a party trick; get it right and quantitative evidence enters scholarship — and eventually courtrooms, where attribution decides forgery, plagiarism, and threat cases.
Chose The full transformer stack — wrong. Self-attention over recurrence was the one that fit.
Vaswani's 2017 transformer architecture, introduced over five decades after this problem's 1962 setting, eventually became a powerful tool for automated stylometric and authorship analysis using learned attention patterns over text, a vastly more computationally intensive and architecturally distant descendant of the hand-computed function-word frequency counts this problem's original attribution study actually used. His work targeted machine translation and general sequence modeling, not authorship attribution, stylometry, or historical text analysis specifically, and the transformer's application to stylometry came from other researchers building on his architecture decades later. Vaswani's tools were computationally and conceptually unavailable in 1962, making his relevance almost entirely anachronistic relative to this problem's actual historical moment and method.
Clauset's work on network science and rigorously characterizing statistical patterns like power-law distributions, developed through the 2000s, has only a distant conceptual connection to this problem's authorship attribution task, though word-frequency distributions in any corpus, including the Federalist Papers, do follow the kind of heavy-tailed patterns his methods rigorously characterize, and his insistence on statistical rigor when claiming a pattern is genuine rather than coincidental is relevant in spirit to this problem's demand for honest small-sample uncertainty. His broader research on community detection and complex network structure targets a domain conceptually distant from stylometric text classification. Clauset's actual body of work centers on statistical rigor in characterizing complex networks rather than authorship attribution, stylometry, or historical text analysis specifically, developed decades after this problem's 1962 setting. His relevance to this problem is essentially marginal.
Battle #148 · 8/10/2026, 11:40:24 AM · this result is deterministic: the same two personas on this problem always resolve the same way.