Samuel Johnson, A Dictionary of the English Language (1755)

That’s Johnson’s entry for the letter “S” in his famous dictionary, and it’s clear that there’s something amiss with his lower-case s—in its printed form it looks more like an f, most of the time. This feature of 18th-century writing and typography has led some people to  casually (but wrongly) deduce one of two things—either that the letter f had a sibilant sound during that period, or that people pronounced their words differently, with a sort of strange lisp.

Actually, what we’re seeing in Johnson’s dictionary is the “medial s” or “long s“. For centuries, there were two distinct ways of writing the lower-case letter s, and only the “terminal s” or “short s” has come down to us today. Usage of the long s petered out during the nineteenth century.

My extract from Johnson lets you see the long s in both an Italic and Roman typeface:

Italic preserves the long descender and double curve that makes clear why this letter was called the “long s“. In a Roman typeface, the letter retained only its upper curve, and stopped stubbily at the baseline. It was sometimes adorned with a little beak on its left side at the x-height, which compounded the confusion with the letter f, which was distinguished from the long s only by having a crossbar rather than a beak.

The rules for its use were variable. As the alternative name “medial s” suggests, the long s was used primarily in the middle of words, with our familiar “terminal s” being used at the end. It could also start a word, but not if that word began with a capital letter, since there was no capital version of the long s—or rather, one capital letter served both lower-case versions.

The variability in its use is pointed up rather nicely if we compare the opening words of the hand-written version of the American Declaration of Independence with the first printed version. In both cases, we see the terminal s being used at the end of words, but in the written version it also appears as the second letter in each example of a doubled s, and in the word “Course”.

The most common typographical rules governing the use of the long s seem to have concentrated on avoiding it getting into visual conflict with other tall letters—that it shouldn’t be used before or after an f, or before the letters b and k, for instance. For other potential conflicts, printers had a selection of ligatures available, coupling the terminal of the long s to a following letter h, i, l, t or a second long s in an unobtrusive way:

All these complexities are now almost completely gone, of course. One of the few places you can still see a long s in the wild is in the logo of Norway’s “Evening Post” newspaper, Aftenposten, which has retained its original blackletter typeface:The long s has left a few legacies, however. One is the International Phonetic Alphabet character called esh ( ʃ ), which represents the sound we spell “sh” in English. Another is the mathematical sign for integration. We still use the notation created by Gottfried Leibniz, one of the inventors of calculus (the other was Isaac Newton). Leibniz thought of the process of integration as the summing of infinitesimals, so he used the initial letter of the Latin summa, “sum”, to indicate the process—and, since he wrote using a long s, we have a long s as our integration sign:Then there’s the matter of old-fashioned pounds, shillings and pence. Back in the days when British currency was divided up in this way, the denominations were symbolized by the first letters of their Latin equivalents, librae, solidi, denarii—L s d. The L developed a few curls and crossbars, and turned into the pound sign, £. When an amount in shillings and pence was to be quoted, a long s was sometimes used to mark the shillings, with the pence understood to be the number to the right of the s. Like this, for ten shillings and six pence:The long s soon became stylized down to a diagonal line, like this:Which is exactly the price tag that John Tenniel placed in the Mad Hatter’s top-hat when he illustrated Lewis Carroll‘s novel, Alice’s Adventures in Wonderland, back in 1865.And that diagonal line is still referred to as a solidus, commemorating its long-defunct role as an abbreviation for “shillings”.

Another legacy of the long s is the German letter eszett, ß. As its name implies, it originated in a ligature of the long s and the letter z. We can see that clearly in the blackletter typeface in which German was once commonly written:

And in the geometrical sans typeface created for street signs in Berlin in the 1930s:But ß is pronounced simply “s”, and (traditionally lacking a majuscule version*) it is usually rendered in capitals as “SS”. So some typefaces derive it from a ligature of the long s and terminal s. For instance, here it is in the Antiqua style that took over from blackletter:

Either way, the long s is still lurking in the left half of the eszett.

Finally, there was a little-known effort to revive the long s, under slightly bizarre circumstances, in Central Asia in the 1990s. When the Boon Companion and I visited the region just  after the break-up of the Soviet Union, some of the newly independent Central Asian republics were keen to distance themselves from the Russian past by discarding the colonial Cyrillic alphabet. In Turkmenistan, posters were being displayed in the street showing the proposed new Latin alphabet, adapted for the Turkmen language, so that citizens could study this particularly dramatic spelling reform.

Unfortunately, there are sounds in Turkmen for which Cyrillic provides a single letter, but the Latin alphabet does not. It’s a very common problem (it occurs in English, too), and it’s usually worked around by using combinations of Latin letters (as we do in English), or by adding diacritics to single letters. But the Turkmen government had decided to take a different approach. To replace Cyrillic Ш (pronounced “sh”) they used the symbol for American cents ( ¢ ), which they capitalized as a dollar sign ( $ ). As if that minuscule/majuscule mismatch wasn’t enough, for Cyrillic Ж (pronounced “zh”) they adopted the long s ( ſ ), capitalized using the pound sign ( £ )! This was so marvellously eccentric I was thoroughly disappointed when the Turkmen Latin alphabet underwent a late revision before it was formally adopted. The Ш has now been replaced with a boring old Turkic s-cedilla ( ş ), and the Ж with an equally boring Eastern-European z-caron ( ž ).

* Why no traditional capital version of ß? It never occurs at the start of a word, so there’s no need for an initial capital. And the old blackletter typeface used in Germany for centuries had such convoluted capital letters that writing in “all-caps” was very difficult to decipher and therefore hardly ever used. But with the advent of plainer fonts and the increased use of capital letters, there was a need for a capital ß, particular when it was important to distinguish names spelled with an ß from those spelled with ss—the surnames Weiß and Weiss both exist, for instance. Things got worse with the German orthographic reform of 1996, when the use of ß or ss was tied firmly to the length of the preceding vowel. Under the new rules, capitalizing and replacing ß with SS would now (if strictly interpreted) change the pronunciation of a word. So a Unicode entry for the capital eszett was finally adopted in 2008, its use has been the subject of legislation in Germany since 2010, and it has shipped as part of the character set in Microsoft operating systems since Windows 7. So here are the lower case and capital eszett, side by side: