The Invisible Craft

You are not, at this moment, looking at these letters. You are looking through them.

That is the peculiar condition of the craft: it succeeds exactly to the extent that it goes unobserved. A page of well-set text produces no impression at all — no admiration, no irritation, nothing that registers as an experience — and this absence is not a failure of ambition. It is the entire objective, stated most memorably by a typographer in 1930 who compared the printed page to a wine glass. A vessel of beaten gold draws attention to itself. A crystal goblet lets you see the wine. Typography, she argued, should be the crystal.

It is worth understanding what has to happen for that invisibility to be achieved, because it is a great deal more than choosing a typeface.

The invention that actually mattered

The conventional account credits the mid-fifteenth-century breakthrough to movable type. This is not wrong but it identifies the wrong component. Movable type — the idea of assembling a page from reusable individual characters — had existed in East Asia for centuries.

What was new was a manufacturing process for producing characters that were identical, in quantity, to tolerances that let them lock together under pressure without shifting. A hardened steel punch was cut by hand for each character and struck into softer copper to form a matrix. The matrix was then placed into an adjustable hand mould, into which molten alloy was poured. The mould's crucial property was that it could adjust in width while holding height and depth exactly constant, so that an i and an m differed in the one dimension that could vary and in no other.

That mould is the invention. Everything else — the press, the oil-based ink, the composing stick — was adapted from existing trades. Printing began as a precision metalworking problem, and the people who solved it were goldsmiths.

Two centuries of settling

Early printed letterforms imitated the manuscript hands they replaced, which is the normal behaviour of a new medium. The forms that survive were not the most faithful imitations but the ones that acknowledged the new process.

In Venice in the 1470s, a printer cut a roman letter of extraordinary evenness, balancing the weight of stroke and counter so that a mass of text produced a uniform grey tone on the page rather than a pattern of dark and light patches. This quality — the even texture of a block of text seen from a distance at which individual letters are not legible — became the principal measure of quality, and it remains so.

Three decades later the same city produced two further innovations that were commercial rather than aesthetic. The first was a sloped, condensed letter fitting more words into a line. The second was the small-format book: portable, affordable, intended to be carried and read privately rather than consulted on a lectern. The economics of the first enabled the second, and the second altered who could own books and where reading happened. A change in letter width is not usually a social development, and this one was.

Anatomy, briefly

The vocabulary is worth having, because it makes visible what is otherwise merely felt.

The x-height is the height of the lowercase body relative to the capitals. A large x-height reads better at small sizes and looks coarse at large ones. Counters are the enclosed spaces inside letters; when they close up, text turns to mud. Aperture describes how open the gap is in letters like c and s, and it is what determines whether a face remains legible when it is small, low-contrast, or seen in poor conditions. Stroke contrast — the difference between thick and thin — signals period more reliably than any other feature, running from almost none in the earliest romans to the extreme, hairline-thin contrast of the early nineteenth century, which depended on smoother paper and better presses and which looks elegant in a heading and punishing in a paragraph.

Then there is the part that is not about letters at all. Measure is line length, and it is the single most consequential decision on a page: somewhere between forty-five and seventy-five characters per line, because beyond that the eye loses its place returning to the left margin, and below it the rhythm of reading breaks into fragments. Leading is the vertical space between lines, which must increase as the measure lengthens. These two together do more for comprehension than any choice of typeface, and they are the two that non-specialists never adjust.

Legibility is not readability

The distinction matters and is routinely collapsed.

Legibility is a property of the typeface: how easily individual characters are distinguished. It is what matters for a road sign, a dial, a label glimpsed once.

Readability is a property of the setting: how comfortably continuous text can be read for an extended period. It depends on measure, leading, contrast, margin and consistency far more than on the letterforms themselves.

A highly legible face, set at an excessive measure with insufficient leading, produces an unreadable page. This is the ordinary condition of a great deal of text on screens, where the width is determined by a browser window rather than by anyone's judgement, and where the reader's fatigue is invariably attributed to the screen rather than to the setting.

A note on claims

Since the field attracts confident assertion, one caution.

Specialised typefaces marketed as improving reading for people with dyslexia have been tested repeatedly, and controlled studies have generally failed to find an advantage over ordinary well-designed faces at appropriate size and spacing. What does help, consistently, is larger text, shorter lines and increased spacing — which are setting decisions, available for free, in any typeface. The lesson generalises: in this discipline, the dramatic intervention is usually less effective than the boring one.

The discipline of not being noticed

Everything described here — the mould that held two dimensions constant, the even grey of a well-cut roman, the seventy-character line — serves a single end, which is to remove every occasion on which a reader might become aware of the surface rather than the sense.

There is a legitimate objection, and it should be stated. A crystal goblet is the right ideal for a novel and the wrong one for a poster, a title, a warning, a sign in an unfamiliar station. Sometimes the vessel is the point, and a tradition that treats self-effacement as the highest virtue has trouble accounting for the work that must arrest attention rather than release it.

But for continuous text — for the paragraphs that carry nearly all transmitted thought — the old ideal holds. The measure of success is that you have read this far without once considering the shapes doing the work, and that if I had set it badly you would have stopped several pages ago, blamed the argument, and never suspected the spacing.