The Counting House No. 6 — Writing History Without a History Degree

Before the Storm — Post 5 of 5

I want to say something about how this paper got written, because I think it matters — not just for this project, but for what is becoming possible for people who have something to say and lack the particular credentialing to say it in the expected form.

The short version: I wrote an academic paper for peer-reviewed journal submission. I am not an academic. I used an AI writing tool to help me do it. I disclosed this in the paper itself, directly and without apology.

Here is the longer version.

What I Knew and What I Didn’t

I knew my document. I had spent months with it — reading the scholarship on colonial commercial history, tracking down the relevant archives, corresponding with librarians in Worcester and Manchester, learning the mechanics of bills of exchange, understanding the Stamp Act’s commercial provisions. I knew what I had found and why I believed it mattered.

What I did not know was how to translate that knowledge into the form a peer-reviewed journal expects. The footnoting conventions. The architecture of a historical argument. The particular prose register — formal but not stiff, precise but readable — that academic history writing requires. The way you build from evidence to claim in a way that a specialist reviewer will find credible and a general reader will find engaging.

That is a craft. It is learned, usually, over years of graduate training and academic writing. I did not have those years. I had a document, a stack of research, and a deadline I had set for myself.

The Tool I Used

I used Claude, an AI assistant developed by Anthropic.

I want to be precise about what that means and what it doesn’t mean.

The historical argument in the paper is mine. The identification of the three signatories — Freame, Sayward, Hancock — was mine. The decision to contact the AAS, the Barclays Group Archives, JSA, the American Philosophical Society — mine. The recognition that the Hutchinson riot entry on August 26 transformed the significance of the August 27 date — mine, when Ashley sent me those images, and my response to what I saw.

Every factual claim in the paper has been reviewed against my own knowledge of the material and flagged for independent verification. The cited sources are real. The diary quotations are from actual diary images I have seen. The modern currency equivalents were verified against OfficialData.org. The footnotes were checked against the actual finding aids and institutional records they cite.

What the AI did was help me translate all of that into the form a peer-reviewed paper requires. It helped me understand the architecture of the argument — which evidence belonged in which section, how to frame the significance claims without overstating them, how to integrate the footnote apparatus with the prose, how to write an abstract that accurately represents the paper’s contribution. It was a writing collaborator that happened to understand historical prose conventions.

A scholar with different training than mine — one who had spent years writing academic history — would have produced the same paper by different means. They would have had the craft already. I used the tool available to me.

The Disclosure

I disclosed this in the paper. Directly, in the acknowledgments section, where it belongs:

The manuscript was drafted with the assistance of an artificial intelligence writing tool, Claude, developed by Anthropic. I regard this as neither apology nor confession. It is disclosure.

I added: A scholar with a different formation than mine — one trained in the conventions of historical prose, the architecture of academic argument, and the apparatus of footnotes and bibliography — would have translated the same research into the same paper by different means. I used the tool available to me.

I expect this will generate some friction in the peer review process. Academic publishing has not yet settled on consistent norms for AI-assisted writing, and some reviewers will view any AI involvement as disqualifying. Others will focus, as I believe they should, on whether the historical argument is sound, whether the evidence supports the claims, whether the sources are properly cited. Those things are not produced by AI. They are produced by research.

The paper will stand or fall on its argument. The disclosure is there because hiding it would be dishonest, and dishonesty in scholarship — of any kind, by anyone, with any tools — is the actual problem worth worrying about.

What This Means for People Like Me

I am in my sixties. I spent my career managing information systems for organizations that needed someone to make databases work and projects land on time. That career gave me analytical habits — the discipline of following an evidence chain, the intolerance for sloppy inference, the insistence on checking sources — that turned out to be directly useful in historical research. What it didn’t give me was the prose.

There are, I suspect, a significant number of people in the world who have done serious original research — into local history, family history, archival discoveries, material culture — and who have been unable to translate that research into publishable form because the form itself requires a training they don’t have. Some of them give up. Some of them self-publish in ways that don’t reach the scholarly audience their work deserves. Some of them sit on what they found, unsatisfied.

The tool I used changes that equation. Not by replacing the research — the research remains the hard part, the irreducible human part — but by making the translation possible for people who would otherwise be locked out of it.

I think this is a good development. I think the disclosure requirement is essential. I think the historical profession will take time to agree, and I think that time will be useful, and I think the outcome will be a broader scholarship that includes voices and discoveries it previously excluded.

That is, at least, what I am betting on with this paper.

Where Things Stand

The document is at JSA for the Freame authentication. The paper has been submitted for peer review. The American Antiquarian Society has provided the August 1765 diary images that gave the paper its most powerful evidence. The Barclays Group Archives have been contacted about Freame handwriting exemplars. Two interlibrary loan books are pending — verifications for two footnotes — and Ashley Cataldo may yet come back with the transcriptions I need for the 1768 and 1775 diary passages.

The 250th anniversary of American independence is July 4, 2026. The document will go to auction, or to an archive, or to some combination that hasn’t fully taken shape yet.

Jonathan Sayward wrote “the mob Rose in Boston” in his diary on August 26, 1765. The next morning, John Freame drew a payment order in Halifax, Nova Scotia. John Hancock endorsed it on the back.

It sat in a frame on my wall for twenty-eight years.

I am glad I finally looked at it.

Paul G. Molé is a retired database administrator and independent scholar from Stillwater, New Jersey. He is preparing “Before the Storm: Commerce, Credit, and the Coming Revolution in a Triple-Signed Transatlantic Sight Draft of 1765” for peer-reviewed journal submission.