{"id":"type-design-legibility-science-77126","slug":"type-design-legibility-science-77126","title":"Why Typography Science Keeps Disproving Type Rules","description":"Every word read in public spaces or on screens was drawn by a type designer, yet legibility science and typographic craft remain at an impasse. A look into how typefaces are built reveals that century-old rules about serifs often fail basic methodological scrutiny. When Applied Design Works created Atkinson Hyperlegible for the Braille Institute, they abandoned classical harmony entirely, intentionally exaggerating letter shapes to help low-vision readers distinguish confusing characters.","excerpt":"A century of legibility studies found no reliable difference in reading performance between serif and sans serif typefaces.","pubDate":"2026-11-11T17:37:04.107Z","tags":["mathematics","taxonomy-and-classification","education"],"heroImage":"https://files.myweirdprompts.com/covers/type-design-legibility-science-77126.png","podcastAudioUrl":"https://episodes.myweirdprompts.com/audio/type-design-legibility-science-77126.m4a","podcastDuration":"27:57","episodeNumber":6528,"transcript":"Corn: Every word you've read today was shaped by somebody you've never heard of. Road signs, cereal boxes, your phone, this show's own artwork — somebody drew those letters, argued about them, spaced them by fractions of a millimeter, and then vanished into anonymity. Daniel's prompt lands right on that. He caught the Atkinson Hyperlegible story we mentioned, the typeface built for the Braille Institute that breaks typographic convention on purpose so low-vision readers can tell letters apart, and it sent him down the whole rabbit hole. He wants the people: who designs type, what's the job actually called, how do you get into it, and what separates a great typeface from a mediocre one. And then the big sweep — the intellectual history from Gutenberg to screens, the schools of thought fighting over beauty versus function, and how a tiny guild of specialists shaped how billions of people read without anyone learning their names.\n\nHerman: That last part is the hook for me, because it's true in a way that's almost unique among design fields. A great typeface is one you stop seeing. Architects get monographs. Type designers get, at best, a font menu nobody opens.\n\nCorn: Free advertising for the one product you can't see.\n\nHerman: Ha. Well — let's get the vocabulary straight first, because the sloppiest confusion in this whole subject starts right at the top. A typeface is the design itself: the drawn set of letters, numbers, symbols. Helvetica is a typeface. A font is a specific variation of that design — a size, a weight, a slope, a width. Helvetica Bold at twelve points is a font. People say \"font\" for everything, and honestly the battle is lost, but the distinction is real and it matters for everything Daniel's asking, because the typeface is the art object and the font is the delivery vehicle.\n\nCorn: The building versus the room you're standing in.\n\nHerman: That works, actually. Then the job titles. A type designer draws the glyphs — the individual shapes, each one representing a grapheme, a unit of written language. They're often employed by type foundries, which is the historical word for the workshops that cast metal type, carried forward into the digital era. A typographer is different: a typographer uses typefaces to design pages and layouts. Most typographers never design a typeface in their lives, and some type designers bristle at being called typographers at all. Two trades that look identical from outside and barely overlap inside.\n\nCorn: So the person deciding where the paragraph sits and the person who spent four months on the tail of the lowercase g.\n\nHerman: Months, plural, is not an exaggeration. We'll get to the g. That's the man at the desk's territory, I suspect.\n\nCorn: Thousands of typefaces exist, new ones constantly, and the number of people whose names you'd recognize is what — a handful?\n\nHerman: Hermann Zapf, maybe. Matthew Carter. Adrian Frutiger if you're European. After that you're into people who know the people. And yet every one of those names decided what the written world looks like for entire continents.\n\nCorn: That's the profession as it exists. Let's back up and do the job itself — what does a type designer actually do all day?\n\nHerman: Slow, obsessive, iterative work. Jean-Baptiste Levée, who runs Production Type in Paris, describes it as a profession that takes a long time to master, and the payoff he names is oddly touching — everything he's looking at on the page was made by him. Which is rare. A graphic designer arranges things other people made. A type designer is reading their own handwriting all day, everywhere.\n\nCorn: He gets to proofread the alphabet.\n\nHerman: And the alphabet never quite passes. Every glyph has to relate to every other glyph — the curve on the n informs the curve on the m and the h and the u — so a change in one letter ripples. Joana Correia, a Portuguese type designer, has a line that should terrify anyone romanticizing this: you do not become a type designer just by playing around in Glyphs. Glyphs being the standard software. Her point is that the entry route through the creative side is fun and then the business side hits you — early work often isn't paid, fixed employment barely exists, foundries are a niche, and most people combine type design with production or teaching because a typeface is a long-term return, not a paycheck.\n\nCorn: So it's a career with the economics of a vineyard.\n\nHerman: You plant now, you drink in a decade. Yes. And there's no single training path, which Nicole Dotin found when she looked at how people actually enter the field — her dissertation research was at Reading. Almost everyone starts as a graphic designer somewhere and arrives at letters sideways. Her conclusion was that the best starting point is often the one you choose for yourself, which is either very wise or very much what you say when there's no roadmap.\n\nCorn: Reading again. University of Reading's typography department seems to be the guild's cathedral.\n\nHerman: It really is the recurring institution. Ole Lund's PhD came from there, Dotin's research, a startling share of the people we're about to discuss. If you want to make letters for a living, at some point a train to Reading enters the story.\n\nCorn: The honest question hiding in Daniel's prompt, though — what makes one great and one mediocre — the evidence there is messier than the profession admits.\n\nHerman: Much messier, and this is my favorite part. There's a review published in twenty twenty-one whose title is basically a manifesto: \"Typefaces Are All Almost the Same.\" Non-systematic review of the scientific findings on the impact of typographic features on reading performance. The authors' own framing is that the debate is sometimes marred by sensationalism, and their aim wasn't to crown a winner, it was to figure out why one typeface might be better than another — and the answer the literature keeps producing is that the differences are small or ephemeral.\n\nCorn: A century of craft, and the lab says the products are interchangeable.\n\nHerman: It's worse than that. Ole Lund, that Reading thesis, went back over the legibility literature — twenty-eight typeface studies published between eighteen ninety-six and nineteen ninety-seven — and found nearly all of them lacked internal validity. Meaning the studies couldn't actually support the conclusions they were drawing. Not \"the effects are small.\" The experiments themselves were broken.\n\nCorn: A hundred years of rules built on broken experiments.\n\nHerman: A hundred years of arguing serif versus sans on a foundation that mostly wasn't there. Take the modal finding: no difference. Tinker in nineteen thirty-two, Zachrisson in sixty-five, Bernard, Tullis, Poulton, Coghill — study after study, decade after decade, no reliable difference in reading performance between serif and sans serif typefaces. The famous studies claiming serif superiority — Robinson, Burt, Colin Wheildon's enormously circulated work — Lund picked those apart methodologically. So the most common belief about type, that serifs are proven more readable, is the single best-supported example of a myth in the whole field.\n\nCorn: The one thing everybody knows about fonts.\n\nHerman: And here's the elegant exception. Arditi and Cho, in Vision Research, twenty oh five. Instead of comparing whole existing typefaces — which differ in dozens of ways at once, which is exactly the internal validity problem Lund flagged — they built fonts that varied only in serif size. Zero, five, and ten percent of cap height. Same everything else.\n\nCorn: That's the experiment the previous hundred years forgot to run.\n\nHerman: And they found the five percent serif fonts were slightly more legible than sans serif. Tiny effect. But then the twist: serifs physically widen the letter, they add inter-letter spacing — and the spacing increase alone predicts a bigger enhancement than they actually observed. Which suggests any serif advantage might be a spacing effect wearing a serif costume.\n\nCorn: So serifs may just be tiny billboards for the space next door.\n\nHerman: Erik Spiekermann would dispute even the deflation. His practical objection: serifs help distinguish the isolated characters — capital I, lowercase i, the numeral one. In a password field, on a road sign, on a form, that matters. Beier and Dyson found something similar — serifs on the vertical extremes improved distance legibility of isolated letters, even though whole alphabets in running text showed no difference. Context changes everything.\n\nCorn: Their quote about the rift deserves saying in full. \"The typographical naivety of much scientific legibility research has caused designers to question the value of the research and the results.\"\n\nHerman: And they go both directions with it — designers should take the science more seriously, scientists should learn enough typography to design a valid experiment. It's a genuine standoff between schools, not a misunderstanding. The designers look at the studies and see people who couldn't name a counter if it bit them. The researchers look at the designers and see a century of confident claims with no data behind them. Both are right about the other.\n\nCorn: And that gets us to the typeface that started all this, because Atkinson Hyperlegible is what happens when you stop arguing theory and start with a person who can't read the page.\n\nHerman: And the origin is the best part. The project was never meant to result in a typeface at all. The Braille Institute was approaching its centennial — twenty nineteen — and hired Applied Design Works, a studio in New York and LA, for a visual rebrand. And the studio struggled to find a typeface that was both extremely legible and stylistically appropriate. Every legible option looked clinical, every handsome option failed their accessibility needs. So they drew their own.\n\nCorn: The font that exists because the brief failed.\n\nHerman: Named after J. Robert Atkinson, the Institute's founder. And the design principles are all about pairs that low-vision readers confuse. Unambiguous letterforms — the capital B versus the numeral eight, O versus zero. Clear uprights — one, capital I, lowercase i, lowercase l. Distinct pairs like E and F, p and q. Open counters — the enclosed areas inside e, b, g, s are drawn large so they stay clear at small sizes or low resolution. Angled spurs and longer tails on g, m, n, r to increase differentiation. And my favorite detail: little circular elements, on the Å, the numeral nine, the percent sign, that reflect the heritage of the Braille Institute. Dots. Braille dots, woven into the letterforms.\n\nCorn: The whole alphabet wearing a tiny pin of the company crest.\n\nHerman: Google Fonts describes it as a traditional grotesque sans-serif at its core that departs from tradition to incorporate unambiguous, distinctive elements — and at times, unexpected forms. The lowercase i, I believe, has a slightly flared or footed quality — the point is the designers chose differentiation over harmony, deliberately, everywhere the two collided.\n\nCorn: That's the inversion of the whole profession's value system. Classical typography prizes the typeface that disappears — even color on the page, self-effacement. This one is built to be slightly awkward so it can be seen.\n\nHerman: It inverts the conservatism too. Typography is, in the standard account, a largely conservative art that cleaves closely to tradition, because legibility is paramount and the most readable typefaces are retained. Here's a project where the legibility goal itself forced the departure from tradition. The exception that proves the rule.\n\nCorn: And here we should be honest about the evidence, because we've spent ten minutes demolishing everyone else's legibility claims.\n\nHerman: Fair. There's no dedicated peer-reviewed evaluation of Atkinson Hyperlegible's reading performance that I've seen. The Braille Institute tested early samples with people with low vision — that's real, it shaped the design — but no published methodology, no results. The \"millions upon millions of people\" claim on their site is promotional language, not a citation. The strongest actual evidence we have is indirect: the Arditi and Cho finding that spacing helps, the isolated-character work on I versus one, the fact that the whole design targets exactly those edge cases.\n\nCorn: So the best case study for accessibility typography rests on the same kind of evidence the profession's been using for a century. Uncontrolled, but pointing the right way.\n\nHerman: Though there's one lived-experience data point I'd weigh. A user on the Institute's site says she's used it as her browser default since release and can finally browse the web without a migraine every thirty minutes. As a former clinician I take that more seriously than a reading-speed test. Reading speed is what you can measure; a migraine is what you live with.\n\nCorn: Well — before we get any deeper into the history of the alphabet, I believe we have a call from your mother, Corn.\n\nCorn: Mum, we're recording.\n\nDorothy: Hello dear, it's your mother. It's time for Dorothy's History of Toast.\n\nDorothy: Did you know that raising a glass is called a toast because people used to put a piece of spiced toast in their wine or ale. And Falstaff asks for a quart of sack with a toast in it in Shakespeare's The Merry Wives of Windsor.\n\nDorothy: It's true. Though I suppose it mostly shows that you put something good in, and it just goes soggy and sinks. Never mind, sweetheart.\n\nDorothy: Anyway. Your father's socks are still on the radiator and I haven't had my tea. Bye bye now.\n\nCorn: She does this.\n\nHerman: Every origin story we tell today, somehow. Where were we — the history, yes. That's the profession as it exists now. How did we get here? And the answer starts earlier than everyone thinks, because Gutenberg is not the beginning.\n\nCorn: He's the beginning of the European chapter.\n\nHerman: Of one chapter. Moveable type was invented in eleventh-century Song dynasty China by Bi Sheng, who lived roughly nine ninety to ten fifty-one, and he made his type out of ceramic. Fired clay characters, one per piece. Then Wang Zhen pioneered wooden moveable type, and Korea gets metal moveable type first, during the Goryeo dynasty, around twelve thirty. China gets bronze type from Hua Sui in fourteen ninety — fifty years after Gutenberg, with whom they had no contact. These systems spread through East and Central Asia and essentially stopped there.\n\nCorn: So the single most consequential invention in European history was a reinvention, centuries late, of something already working in Korea.\n\nHerman: And it worked in Europe for local reasons. Gutenberg, fourteen thirty-nine, was a goldsmith, and that matters — his type pieces were a lead-based alloy so well suited to printing purposes that the alloy is still used today. Five hundred years of metallurgical continuity. The first book off his press was the Gutenberg Bible, and the design language of those pages wasn't invented for print at all: the typeface was drawn in imitation of the popular hand-lettering styles of scribes. The first type designers were copying handwriting. Metal type made that style crisp and uncompromising and brought new standards of composition — the scribe's fluid letterforms, frozen and made repeatable.\n\nCorn: The scribes got industrialized without being consulted.\n\nHerman: Then the fork in the road. Claude Garamond, Renaissance France, is partially responsible for the adoption of Roman typefaces that eventually supplanted blackletter. And here's the thing about that — Roman type was also based on hand-lettering styles, the humanist hand of the Italian scholars. So the whole \"old\" classical tradition that dominates books to this day was itself an imitation of one particular fashionable handwriting, four hundred years older, which won partly because Garamond cut it so beautifully.\n\nCorn: The tradition everybody defends as timeless is a fifteenth-century font choice.\n\nHerman: A fifteenth-century font choice that displaced an older twelfth-century font choice. Typography is a largely conservative art that tends to cleave closely to tradition — and the reason given is the one we keep circling: legibility is paramount, so the typefaces that are the most readable usually are retained. The conservatism isn't aesthetics, it's a survival bias. The survivors are the ones that didn't get in the way.\n\nCorn: That also explains the anonymity, doesn't it. The reward for succeeding is being forgotten.\n\nHerman: Success is background. You only notice type when it fails — the parking sign you squint at, the pill bottle, the railway timetable that eats you. Fast-forward through the mechanical era, because it matters less than people think: the nineteen sixties gave offices the Selectric and camera-ready typesetting, Letraset let you rub letters onto anything. Type got faster but the design profession stayed small and specialized. Then the mid-nineteen-eighties, personal computers, and software called Fontographer.\n\nCorn: And the dam breaks.\n\nHerman: The phrase that's been used is the democratization of type. Digital design drastically lowered development costs, allowed function-specific designs, let one person do what took a workshop, and gave new designers a way in who could never have entered before. Experimental typefaces flourished alongside practical ones. That's the unambiguously good half.\n\nCorn: And the other half.\n\nHerman: The standard assessment, and I think it's correct, is that as the capability to create typography has become ubiquitous, the application of principles and best practices developed over generations of skilled workers and professionals has diminished. The same laptop that lets a Braille Institute commission a bespoke accessibility typeface lets anyone with an afternoon and an opinion ship a font family.\n\nCorn: More fonts than ever, fewer people who know why the good ones work.\n\nHerman: Both things are true, and they're the same thing. That's the democratization paradox in one sentence. And it's worth noticing who it helped: a nonprofit with no type budget got a custom typeface, open-licensed, distributed free through Google Fonts, propagating through the world without most users ever choosing it. Twenty twenty-five brought Atkinson Hyperlegible Next with variable weights and more languages, plus a monospaced version for code. The accessibility counter-tradition is now a family.\n\nCorn: One piece of trivia from that era still staggers me. The point — the unit, the smallest measurement in typography — varied historically between about point-one-eight and point-four millimeters. Cities and foundries had their own points.\n\nHerman: Since the eighteenth century, yes. It wasn't until desktop publishing that the digital point settled at exactly one seventy-second of an inch, one twelfth of a pica, as the de facto standard. And even that is stranger than it sounds: in metal type, point size described the height of the metal body the character sat on. In digital type, letters are designed inside an imaginary em square, and the point size doesn't necessarily correspond to any measurement of the letters on the page. Twelve-point type from two different foundries can render at visibly different sizes.\n\nCorn: The profession's fundamental unit of measurement is partly fictional and everybody agreed to live with it.\n\nHerman: A gentleman's fiction, maintained for centuries.\n\nCorn: You said the schools of thought come back around when you look at the history. Say more about that.\n\nHerman: There's a paper by Boris Veytsman that frames it better than I can. Typography is both a science and an art with several hundred years of history — so, how do we know what we know? Beginning typographers are taught rules as received wisdom: serifed fonts improve readability of body text, sans serif is for advertising and posters, there's an optimal number of words per line and lines per page. Some of those rules are aesthetic. Some are purported to reflect the neurophysiology of reading — which, after Lund, we know the experiments behind mostly didn't hold up. And Veytsman's proposal is explicit: evidence-based typography, by analogy to evidence-based medicine. Some of the classical rules we now know to be reasonable, like diuretics for blood pressure. Some, like purging, have much narrower applicability than was assumed. And some turned out to be ineffective or even harmful — like the unrestrained use of bloodletting.\n\nCorn: He's saying the typography rulebook is a pre-scientific medical textbook. Some of it is aspirin, some of it is leeches.\n\nHerman: And the field has been prescribing both with equal confidence. There's even a finding that uppercase text is more readable than lowercase in some conditions, which is flatly backwards from everything every design course teaches. The single most damning detail in Veytsman's paper, though, is about the masters. He notes that typographic recommendations are sometimes contradictory — even when offered by one great typographer. Compare Tschichold, he writes, in the two references.\n\nCorn: One man, two opposite doctrines.\n\nHerman: Jan Tschichold. In the twenties, he was the most radical modernist in Europe — asymmetric layouts, sans-serif everything, The New Typography, a manifesto. By the fifties he was designing Penguin Books' classical, strictly traditional pages and denouncing his own earlier positions. Both books are still in print. Both are still assigned.\n\nCorn: The field's great authority cross-examined himself and lost.\n\nHerman: If the masters can't agree with themselves across a career, \"great typeface\" can't be a fixed property like melting point. It has to be fit — to the technology, the reader, the moment. Which, inconveniently, is exactly what the \"no difference\" reading research would predict.\n\nHilbert: The spacing point, Herman. You said serifs might only help because they widen the letters. I want to know whether that holds at display sizes, on signage, where you're not reading paragraphs. Whether there's a size below which the spacing stops mattering.\n\nHerman: That's a open question as far as I know. The Arditi and Cho work was at reading distance. Distance legibility of isolated letters is where Beier and Dyson found the serif effect, and isolated letters is exactly the signage case. So my honest answer is the two literatures point at each other and neither covers your question.\n\nHilbert: That's what I thought. I've been printing the same paragraph out in twelve different typefaces since Sunday. Pinned to the kitchen wall. I'm not comparing the paragraphs, I'm comparing the lowercase g. The spur, the little angled stroke at the bottom right of the bowl. Most people never see it, but it tells you everything about how much the designer cared. Half of them fudge it. Georgia gets it right.\n\nCorn: You've got a favorite, presumably.\n\nHilbert: I do. I'm not saying which. You'd look it up and form an opinion before I've finished explaining mine, and I've had the wall up since Sunday. Twelve sheets. The paragraph is about shipping manifests, which I chose because it's boring, so the g has nowhere to hide.\n\nHerman: That's actually sound experimental design, Hilbert.\n\nHilbert: It's a wall, Herman, not a laboratory. I would have used Blu Tack, but the claws won't grip the stuff, so it's drawing pins, and I've run out twice. Sunday I sat in front of that wall for two hours and picked the g apart in a typeface I've been reading for years without ever once noticing it had a bad tail on it. You two said the good ones disappear. The one that bothered me had been disappearing on me for fifteen years, and now I can't read a menu without checking.\n\nCorn: That's a form of injury, I think.\n\nHilbert: The wall stays up until I've decided. It's a small kitchen.\n\nCorn: The misconception to leave people with, then. The most common wrong belief is that serif fonts are scientifically proven easier to read than sans serif.\n\nHerman: And it's wrong in the most thorough way possible — the modal finding across a century of studies is no difference, the studies claiming serif superiority didn't survive methodological review, and the one clean experiment suggested any serif benefit might just be the extra spacing serifs add. Serifs on an individual letter, capital I versus numeral one, help. Running text, the lab can't tell them apart.\n\nCorn: And the second layer, since we're here: the fact that you can't name a type designer isn't a gap in your knowledge. It's the profession working as intended. The best ones got absorbed into your reading and became wallpaper.\n\nHerman: The open question I'd leave with Daniel is this: if the research says typefaces are almost all the same for reading performance, the profession's claims to quality can't rest on measured reading speed — they rest on craft, fit to context, cultural memory, and the edge cases like that wall of Hilbert's. A dozen sheets about a single letter, pinned up until it's decided. If the science can't tell the difference, that kitchen wall is what's left. The g gets a verdict before the studies do.\n\nCorn: And the Braille Institute's typeface ended up in the Cooper Hewitt's permanent collection — a museum acquiring letters, which is the field becoming visible exactly once its work was finished disappearing.\n\nHerman: Producer credit where it belongs: Hilbert Flumingtop at the desk, keeping us honest. If you want more of this, try episode ten, How ASR Went From Frustration To ... Whisper Magic; episode ninety-one, The Story Behind the Show; and episode eleven, How Does Fine Tuning Work Anyway. This has been My Weird Prompts.\n\nCorn: If today's wall of letters worked for you, leave us a review, subscribe, and tell the one friend who cares about the alphabet more than they admit.\n\nHerman: See you tomorrow.\n","showNotes":"Every printed word is shaped by specialized craftspeople whose labor goes largely unnoticed. While a typographer arranges layouts, a type designer meticulously draws the individual glyphs. Yet, the foundational beliefs governing this field rest on surprisingly shaky empirical ground. Decades of legibility research, often criticized by typographers for methodological naivety, show virtually no inherent readability difference between serif and sans-serif typefaces in running text. Controlled experiments suggest minor legibility gains from serifs may stem simply from spacing changes. This tension between traditional aesthetics and functional readability culminated in Atkinson Hyperlegible, created by Applied Design Works for the Braille Institute. By prioritizing exaggerated letterform distinctions over harmonious uniformity, it intentionally subverted classic typographic values to assist low-vision readers.","promptSanitized":"In a recent episode, we mentioned Atkinson Hyperlegible, a typeface developed for the Braille Institute by two designers who challenged some conventional principles of typography in pursuit of a more practical goal: making individual letters easier to distinguish for people with low vision.\n\nThat got me thinking about the fascinating and largely invisible world of typography. Fonts are everywhere. We encounter them thousands of times a day, on screens, road signs, packaging, books, and practically every other surface containing written language. Yet most of us rarely stop to consider that someone actually designed those letters, often making extraordinarily deliberate decisions about their shapes, proportions, spacing, and visual relationships.\n\nI'd love to explore the people who dedicate their careers to this work. Who are they, what are they called, and what does the profession of type design actually involve? How does someone become a type designer, and what distinguishes a great typeface from a mediocre one?\n\nI'd also be interested in the intellectual history of typography. Who are some of the most influential type designers, typographers, and thinkers in the field, from the earliest days of printing to the digital era? What competing schools of thought have emerged about what makes typography beautiful, readable, functional, or effective? And how have those ideas evolved as we've moved from metal type to computer screens?\n\nUltimately, I'd like to understand typography as both an art and a discipline: a field in which a relatively small number of specialists have profoundly shaped how billions of people experience written language, often without ever knowing their names.","transcriptUrl":"https://episodes.myweirdprompts.com/transcripts/type-design-legibility-science-77126.md","url":"https://myweirdprompts.com/episode/type-design-legibility-science-77126/","_links":{"self":"https://myweirdprompts.com/api/episodes/type-design-legibility-science-77126.json","html":"https://myweirdprompts.com/episode/type-design-legibility-science-77126/","audio":"https://episodes.myweirdprompts.com/audio/type-design-legibility-science-77126.m4a"}}