Every day, you read an average of 100,000 words on screen. Emails, dashboards, technical docs, interfaces, notifications — a continuous stream of text your brain absorbs without ever really pausing on the shape of the letters. And yet those shapes are working constantly in the background: they signal credibility, build trust, evoke an era, betray a brand. A font chosen at random is a broken promise before a single word has been read.
There's a brutal proof of this phenomenon: Comic Sans. The typeface was created in 1994 by Vincent Connare for the speech bubbles of Microsoft Bob, an interface designed for children. It was never meant to leave that context. It became the universal symbol of digital bad taste, plastered on medical certificates, funeral home storefronts, and official announcements from major corporations. It's not that Comic Sans is a bad font — it's a comic-book font being used completely out of context. Understanding typeface families is precisely how you avoid that kind of accident.
Two systems to make sense of it all
Fonts aren't classified by taste or chance. Two reference systems were formalised last century, and both remain relevant today.
Francis Thibaudeau published a simple classification in 1921, based on a single morphological criterion: serifs — those small horizontal strokes finishing the stems of letters. He identified four families: Elzévir (triangular serifs), Didot (hairline serifs), Égyptiennes (rectangular serifs) and Antiques (no serifs at all). Robust, pedagogical, easy to remember.
Maximilien Vox went further in 1952: ten families, incorporating historical, formal and cultural criteria. This is the classification definitively adopted in 1962 by the Association Typographique Internationale (ATypI). It remains the academic standard.
In practice, the two complement each other. Thibaudeau for morphology, Vox for depth.
Thibaudeau's Four Families
Elzévir — the legacy of the printing press
Elzévir typefaces have triangular serifs and a visible contrast between thick and thin strokes. These are the fonts you picture when you think "book" — sober, legible on paper, carrying a certain literary authority.
Common examples: Garamond, Sabon, Palatino, EB Garamond.
Where they work: publishing, long-form journalism, cultural and institutional identities, content where credibility matters more than modernity.
What's changed: they remain highly effective for body text in e-books and reading apps. On very contemporary interfaces, they can feel dated unless paired with a deliberately editorial design approach.
Didot — contrast at its peak
Didot typefaces push contrast to the extreme: strokes so thin they're barely perceptible against massive bodies. Spectacular at large sizes, risky in body text.
Common examples: Didot, Bodoni Moda, Playfair Display, Libre Bodoni.
Where they work: fashion, luxury, cosmetics, cultural press, high-end branding. Vogue has used Didot since the 1950s — it's become a visual signature in its own right.
What to know: at small sizes or on low-resolution screens, the hairline strokes disappear or "vibrate". Use with care — always check how it renders on mobile.
Égyptiennes — industrial robustness
Born during the Industrial Revolution in the early 19th century, Égyptiennes have rectangular serifs as thick as the letter stems themselves. They impose. They assert.
Common examples: Rockwell, Clarendon, Roboto Slab, Zilla Slab.
Where they work: headlines, posters, interfaces that need a strong personality, tech branding. Many developer newsletters and documentation sites use a slab serif for headings — the effect is immediately readable and distinctive.
What to know: over long texts, their uniform weight tires the eye. They're built for titles and pull quotes, not paragraphs.
Antiques — modernity stripped bare
No serifs. Consistent strokes. The essentials, nothing more. Antiques (or sans-serifs) now dominate the entirety of contemporary digital design.
Common examples: Inter, Roboto, Helvetica, Arial, SF Pro (Apple), Segoe UI (Microsoft), Noto Sans.
Where they work: interfaces, mobile apps, websites, technical documentation — anything read on screen. Inter is probably the most downloaded font in the world right now. It was designed specifically for readability on small screens and dashboards.
The Helvetica story: designed in 1957 by Max Miedinger, named after the Latin word for Switzerland. It's everywhere: the New York City subway, US tax forms, the logos of Lufthansa, Toyota, American Airlines, Nestlé, and a large proportion of Fortune 500 companies. There's an entire documentary dedicated to it — Helvetica (2007) by Gary Hustwit. Half of typographers love it, the other half despise it, and everyone recognises it within a second.
The Vox-ATypI Ten Families
The Vox classification adds historical depth that Thibaudeau never set out to cover.
Humanists
The earliest typefaces of the Venetian Renaissance, late 15th century. Organic forms, low contrast, a tilted axis recalling the scribe's quill.
Examples: Jenson, Centaur, Adobe Jenson.
Today: almost exclusively used in publishing or projects seeking a deliberate historical atmosphere.
Garaldes
From the work of Claude Garamond and Aldus Manutius. Finer than Humanists, more balanced, more elegant.
Examples: Garamond, EB Garamond, Adobe Garamond, Cormorant Garamond.
Today: ideal for long texts, cultural publications, institutions projecting seriousness and refinement without coldness. A little-known fact: in 2014, an American high school student calculated that if the US government switched from Times New Roman to Garamond for all its printed documents, it would save around $234 million in ink per year. The number went viral. The methodology was imperfect, but the typographic intuition was sound: Garamond is one of the most economical typefaces in terms of space and ink.
Réales
The classical period of the 17th and 18th centuries: more rationality, less trace of handwriting. The axis becomes vertical, forms become uniform.
Examples: Baskerville, Times New Roman, Georgia, Utopia.
Today: Times New Roman was designed in 1931 specifically for The Times of London, to maximise characters per line. It's legible, neutral, universal — and so tied to academic documents and default word processor settings that using it deliberately in design reads as the absence of a choice. Georgia, by contrast, was designed in 1993 by Matthew Carter for screen reading. In the era of low-resolution displays, it was a marvel of legibility. It holds up well today.
Didones
Didot + Bodoni = Didones. Extreme contrast, rectilinear hairline serifs, absolute verticality. The typographic incarnation of the French Revolution — rigorous, spare, imperial.
Examples: Didot, Bodoni, Walbaum, Bodoni Moda (available on Google Fonts).
Today: fashion, luxury, premium cosmetics, high-end editorial. Open a fashion magazine or a luxury brand lookbook, and there's a good chance you'll encounter a Didone in the headlines.
Mécanes (Slab Serif)
Industrialisation in letterform. Rectangular serifs, low contrast, near-uniform stroke weight. Solid, impactful, slightly brutal.
Examples: Rockwell, Clarendon, Roboto Slab, Courier New, Zilla Slab.
Today: widely used in tech and startup branding (they signal reliability and solidity without the coldness of pure sans-serifs), display work, signage, newsletter headlines. Some, like Courier, are associated with terminals and hacker culture — using them deliberately in an interface calls back to that world.
Linéales (Sans-Serif)
This is where most contemporary digital design lives. Four major sub-families: grotesques (Helvetica, Akzidenz), neo-grotesques (Univers, Folio), geometrics (Futura, Gill Sans) and humanists (Frutiger, Myriad, Inter).
Examples: Inter, Roboto, SF Pro, Segoe UI, Helvetica, Futura, Open Sans, Noto Sans.
Today: humanist sans-serifs (Inter, Roboto, Frutiger) are the default choice for accessible interfaces. Their proportions draw on calligraphic lowercase forms, producing a sans-serif that's legible at every size — without the geometric coldness of Futura or the ambiguity of some grotesques. For digital accessibility, they're generally the safest option. Note: watch out for the "rn" vs "m" disambiguation problem in some geometric typefaces.
Incised
Between serif and sans-serif: very small, triangular or barely suggested serifs, inspired by letters carved into stone or metal.
Examples: Optima, Trajan, Albertus.
Today: Trajan deserves a special mention. Designed in 1989 by Carol Twombly, it became the Hollywood movie poster font — to such an extent that there's a satirical short film called Trajan Is the Movie Font (2007). Gladiator, Troy, Titanic, The Da Vinci Code, Avatar... Whenever a film wants to project epic scale, gravitas, or antiquity, Trajan appears. It's a nearly unique case of one font achieving cultural dominance over an entire industry.
Scripts
They mimic cursive handwriting with a quill, letters connected to each other.
Examples: Pacifico, Great Vibes, Allura, Pinyon Script.
Today: invitations, artisan packaging, logotypes, brand identities evoking tradition or craftsmanship. Critical accessibility note: scripts are often illegible for dyslexic and visually impaired users, and for anyone reading on a small screen at normal speed. Never use them for functional text. Never deploy them as body text. Their digital accessibility is structurally limited.
Manuaires (Casual)
Inspired by handwriting, but with unconnected letters — more legible than scripts as a result.
Examples: Caveat, Patrick Hand, Kalam, Indie Flower.
Today: educational content, children's interfaces, annotations, interactive illustrations. They communicate spontaneity and informality — effective in the right context, risky elsewhere. Only use for short text at generous sizes.
Fractures (Blackletter / Gothic)
The famous "gothic" letterforms: angular, dense, historically loaded. They were the first typefaces ever printed (Gutenberg, 1455), and remained dominant in Germany until 1941 — the year Hitler abruptly banned them, ironically calling them Judenlettern (Jewish letters), when they had nothing of the sort.
Examples: Fette Fraktur, Old English Text MT, UnifrakturMaguntia.
Today: extremely specialised use. Historical reconstructions, fantasy-medieval settings, some metal band logos, newspapers playing on their heritage (The New York Times has kept a stylised blackletter in its masthead). Outside these clearly defined contexts, they're inadvisable: low legibility, culturally heavy, inaccessible.
Typography and accessibility: what the research actually says
The received wisdom — "serif is more legible than sans-serif" — comes from 19th-century studies of printed text. For screen reading, the reality is more nuanced.
The factors that genuinely influence digital legibility are, in rough order of importance: body size (16px minimum for body text), line height (1.4 to 1.6 times the body size), contrast between text and background (minimum 4.5:1 per WCAG AA, 7:1 for AAA), column width (45 to 75 characters per line), and finally the intrinsic quality of the typeface's design.
The serif vs sans-serif question matters less than commonly assumed — provided the chosen typeface is well-drawn, well-configured, and used within a clear visual hierarchy.
What never works: a very thin font at very small size on a very light background. It's the most common failure mode in "minimalist" interfaces that mistake sparseness for illegibility.
What this changes in practice
For a front-end developer or designer, understanding these families transforms how you choose a typeface. It's no longer about finding "something nice" in Google Fonts, but about asking three questions: What register do I want to project? Will this text be read continuously or scanned? Who is my audience and what are their reading constraints?
A Garalde for long editorial text. A humanist sans-serif for a data interface. A Didone for a premium branding headline. A slab serif for a tech newsletter title. And never, under any circumstances, Comic Sans on a professional document — unless you know exactly what you're doing, and you're doing it with irony.