The Constitution is not written in a font. It is handwriting — engrossed on parchment by Jacob Shallus, assistant clerk of the Pennsylvania General Assembly, who wrote roughly 4,500 words over a weekend in September 1787 and was paid $30. The body is eighteenth-century round hand. The We the People heading is something else entirely.
That second sentence is where the search results usually stop and where the interesting part starts. Most pages answering this question treat the document as if it had one look. It has two, they come from two different traditions, and they may not even be by the same person.
What “engrossed” actually means
The word is everywhere in eighteenth-century records and almost nowhere outside archives now. To engross a document is to produce the final, formal, fair copy — the one that gets signed. Not a draft, not a printing. One physical object, written out by hand in a script chosen for occasion rather than speed.
The Convention voted on the Constitution on 15 September 1787 and adjourned so it could be engrossed and printed. Delegates signed on the seventeenth. The gap is one working weekend, and Shallus filled it.
Consider it as a production job. Four sheets of parchment — animal skin, limed and stretched, following English legal practice. A quill cut from a large feather. Ink of oak galls, iron and gum arabic, which bites into the surface rather than sitting on it. No undo. Thirty-nine men are going to sign the bottom of it in forty-eight hours.
The body: round hand
The script Shallus used for the text is round hand — the standard clerical and business hand of the eighteenth-century English-speaking world, learned from engraved copybooks. American clerks typically worked from George Fisher or George Bickham; Bickham's The Universal Penman, published in 1743, was the most celebrated of them.
Round hand is the ancestor of the looping cursive taught in American schools. It is not the same thing, and it is emphatically not Spencerian, the flashier nineteenth-century business hand that produced the Coca-Cola wordmark. Spencerian did not exist in 1787. Any page that describes the Constitution as written in Spencerian script is off by about half a century.
What separates round hand from a typeface is not elegance. It is variation. The stroke swells and thins because a cut quill does that under pressure, and the pressure changes with the pull. Letters connect differently depending on what precedes them. A typeface is a set of repeatable parts; a hand is a performance, and the performance leaves evidence.
The heading: yes, that part really is blackletter
Here is the correction worth making, because it runs both ways.
People look at the Constitution, register something spiky and archaic, and say “Old English.” Type-minded people then correct them: no, it's cursive, it's a hand, there's no blackletter in it. The confident correction is wrong.
The large display words — We the People, and the Article headings — are drawn in an engrossed gothic style based on the German Text and Square Text exemplars in Bickham's copybooks. That is blackletter. Keith Bates of the Manchester foundry K-Type, who extrapolated a digital typeface called We The People directly from those letters, classifies the result as blackletter, and the classification is not a stretch.
So the honest answer to “is the Constitution in Old English?” is: the headings are, the text isn't, and the page deliberately mixes two lettering traditions. Formal script for the body, gothic display for the parts that had to shout. An art-direction decision, made in 1787, by a clerk working overnight.
One parchment, two alphabets, forty-eight hours. Whoever laid it out was thinking about hierarchy, not just handwriting.
The son theory
Arthur Plotnik's The Man Behind the Quill: Jacob Shallus, Calligrapher of the United States Constitution — published by the National Archives in 1987, the bicentennial year, and still the only book-length study of the man — raises a possibility that has stuck.
Plotnik suggests the calligraphic headings may have been inserted by someone else: Shallus's teenage son Francis, later a working engraver. The evidence is compositional, not documentary. The Article headings are squeezed awkwardly into the spaces Shallus left for them, in a way that suggests a second hand filling gaps — and not an especially experienced one.
Label it for what it is: a scholar's inference from the artefact, repeated online far more confidently than Plotnik put it. Nobody knows who drew We the People. That the question is open at all is the useful part.
The mistakes are still on it
The most human thing about the document is that it is visibly corrected.
Above the signatures sits an errata note, written into the engrossment itself, listing the fixes: “The Word, ‘the,’ being interlined between the seventh and eighth Lines of the first Page, The Word ‘Thirty’ being partly written on an Erazure in the fifteenth Line of the first Page, The Words ‘is tried’ being interlined between the thirty second and thirty third Lines of the first Page…” You can read the whole thing in the National Archives transcription, where it appears exactly where Shallus put it.
A scraped word. Three interlineations. All of it declared in the text, because on a signed instrument an unexplained alteration is worse than a visible one. The founding document of the United States ships with a changelog.
The printed Constitution is a separate question
If the question is really “what type was the Constitution set in,” that has an answer, and it isn't the parchment.
John Dunlap and David Claypoole printed the official edition in Philadelphia and ran the text in The Pennsylvania Packet and Daily Advertiser on 19 September 1787, two days after the signing. It was set in Caslon.
Caslon is also what Dunlap used for the broadside of the Declaration of Independence on the night of 4 July 1776. Two founding documents, one printing house, one English typeface — William Caslon's, cut in London. The revolution was announced in the type of the country it was seceding from, which is either an irony or a supply-chain fact depending on how romantic you are feeling.
Be precise about what that does not mean. Caslon was not chosen. It was the roman a good Philadelphia shop had in its cases. The association came afterwards, and it is strong enough now that Caslon reads as American-founding to people who have never heard of William Caslon.
Fonts that imitate it, and what they can’t do
Two are worth knowing, and it matters which document each one comes from.
- We The People, by Keith Bates at K-Type — drawn from the Constitution's own engrossed heading, by way of the Bickham models behind it. The backslanted original is free for personal, educational and charitable use; upright versions are licensed. This is the closest thing to a Constitution font that exists.
- American Scribe, by Brian Willson at Three Islands Press — based on Timothy Matlack's hand on the engrossed Declaration of Independence, not on Shallus's. Different scribe, different document, eleven years earlier. It gets lumped in constantly; it is a Declaration face.
Both are good work and both hit the same wall. A revival of a hand has to pick one version of each letter and then repeat it. The original never repeated anything. OpenType alternates and contextual substitution soften that — American Scribe ships a lot of both — but set a long passage in either and the eye starts finding the loop: the same e, at the same angle, forever. Real handwriting never gives you that, which is exactly why it reads as real.
Check anything below those two before trusting it. Search “Constitution font” and most of what returns is download aggregators reposting each other's files and each other's history — telling you the document is Spencerian, or Old English, or the work of Thomas Jefferson. Jefferson drafted the Declaration. He wrote out neither document for signing, and he was in Paris during the Convention.
What a type designer takes from this
Three things, and the first is the oldest lesson in the trade.
Hierarchy predates typography. Shallus — or whoever drew those headings — had no size palette, no weight axis, no second family to switch to. What he had was a different alphabet, and he used it the way a designer today uses a display cut: to mark the entrance. The instinct to make the opening words behave differently from the body is older than movable type and it has never needed a style sheet.
Second: irregularity is information. Round hand carries the pressure of the pen and the speed of the hand, and a reader picks that up without naming it. Digital type deleted the variation and has spent decades buying pieces of it back — contextual alternates, randomised ligatures, variable axes. All approximations of something the quill got for free, and none of them solved.
Third, and most practical: the question “what font is that?” is often the wrong question. It assumes every piece of lettering was chosen from a menu. Plenty of the most recognisable lettering in the world wasn't — not the Constitution, not the Coca-Cola script, not most logos. Asking what font it is skips the more useful question of who made those marks, with what tool, under what deadline. In this case the answer is: a clerk, a quill, and a weekend.
Vigil is the wrong typeface for anything on this page, and it is worth saying plainly. It is a geometric sans built for screens and paragraphs — circular bowls, even strokes, no calligraphic memory at all. Set the preamble in it and you get a competent, completely ahistorical block of text. For period work use a real revival: a Caslon for the printed documents, We The People for the heading, a proper roundhand for the body. Then set your own writing in something that belongs to your own century.