If you write, you punctuate. At the very lowest level of punctuation lies the blink-and-you’ll-miss-it space that separates words; above it are the commas, colons, semicolons and dashes with which we separate clauses; and at the very top, lording it over the rest, sit the period, question mark, and exclamation mark that are used to finish off completed sentences. But punctuation doesn’t stop there. In fact, it doesn’t even start there: Before writers ever thought to separate their words with spaces or punctuate their sentences with periods, documents were routinely arranged into visible, helpful paragraphs.


Paragraphoi in a contract from the second half of third century CE.
P.Mich.inv. 1930; Recto,” University of Michigan Library Digital Collections. CC BY 3.0

The story of the birth of punctuation goes like this: At the Library of Alexandria in ancient Egypt, sometime during the third century BCE, Aristophanes, the chief librarian, was at his wits’ end. The scrolls that crossed his desk were unspaced and unpunctuated, in the fashion of the time, and were written in a script that had only a single case: there was no difference between lower- and uppercase letters so that WORDSCLAUSESANDSENTENCES ran together without breaks. Reading a text like this was an exercise in persistence and frustration.

Aristophanes’s solution was to invent what we now recognize as punctuation: he started to mark the documents he read with little inky dots—punctus, in Latin, leading to “punctuation”—that indicated short (·), medium (.) and long (˙) pauses respectively, each one wedged carefully between adjacent letters, so that reading became marginally easier. It was a tiny change with a disproportionate effect, kick-starting a continuous evolution of writing practices that have led to the spaces, punctuation marks, and other typographic conventions that define today’s written texts.

The odd thing is that punctuation had already been invented, a hundred years or more before Aristophanes’s time. At some point during the fourth century BCE, Greek readers and writers had taken to placing little horizontal lines in the left-hand margins of their documents, each one nestling under a line of text, to indicate that the line in question held some interesting change in topic. These marks were called paragraphoi, for para- ("beside") and graphein ("to write"), and they were used to call out anything from a change in subject to a new stanza of a poem or a change of speaker in a play—paragraphs, in other words, that would today be marked out with an indented or blank line. The simple stroke of the paragraphos is the earliest paragraph mark, certainly, but it is also the earliest true mark of punctuation.


Small red pilcrows (and, at right, a very large one) in a 13th-century manuscript written in England.
(Royal 13 A VII, f. 15. Image courtesy of the British Library.) CC0 1.0

As writers and readers experimented with the novel art of punctuation, paragraph marks branched out on their own. One of first new successors to the paragraphos was another very simple mark, an upside-down L-shaped symbol resembling the stylized gallows from a game of Hangman (Γ _ _ _ _ _ _). As simple as this new mark was, though, no one has definitively worked out how or why it came about. Its resemblance to the uppercase Greek letter gamma (Γ) has led some experts to suggest that it stands for the "g" in paragraphos, but this remains speculative at best, and the Γ has always languished in the shadow of a more famous sibling.

That sibling is the second of the paragraphos’s descendants. During the first centuries BCE and CE, writers working in Latin started to preface paragraphs with the letter K as an abbreviation for the word kaput, or "head," thus indicating that what followed was the "head" of a new argument or topic. For the Romans, however, the letter K, adopted from a pre-Latin alphabet, was increasingly considered old fashioned, and many writers preferred to use C in its place. K for kaput became C for capitulum, or "little head." As these little C-shaped marks grew ever more elaborate, they gained first one additional vertical stroke and later a second so that C became ¢ and later ¶, as seen here. This was the mark called the "pilcrow," from the Old French paragraphe and later the Old English pylcrafte, and it is the most iconic paragraph mark ever to have lived. It is still alive today, in fact, even if it no longer separates our paragraphs: click on the button marked with a in Microsoft Word and it will reveal the invisible characters within your document—the spaces, tabs, and line breaks that hold your words together.


A zoo-anthropomorphic initial S in a 12th-century manuscript, comprised of a human-animal hybrid. Arundel MS 98, f. 85v.
Image courtesy of the
British Library.CC0 1.0

After the collapse of the Roman Empire in the 4th century, writing and bookmaking retreated to the scriptoria, or writing rooms, of Christian monasteries. Here monks labored by candlelight, often without heating or even window panes to keep out the cold, to create some of the most beautiful books the world has ever seen, intricately illustrated and decorated with gold leaf. Even now, its gold untarnished by the passage of time, an illuminated manuscript is a thing to behold.

Monks were innovators when it came to marking out paragraphs. The pilcrow remained in common use, but the monks developed an array of additional techniques with which to mark paragraphs, sections, and chapters. Most simply, the first letter of a section or paragraph could be enlarged or moved out into the margin; later, more ambitious scribes created decorative initial letters that were so large and elaborate that they overshadowed the text itself. These initials came in a bewildering variety of forms: "historiated" letters depicted scenes from the associated story, while "inhabited" initials contained animal or human figures that are unrelated to the text. So-called "zoomorphic" initials were constructed from animals contorted into alphabetic shapes, and their tongue-twisting "zoo-anthropomorphic" cousins, as seen here, were composed of human and animal parts conjoined into a hybrid whole.


Paragraph indents in the first edition of Moby-Dick, published in 1851.
Image courtesy of archive.org. Public domain.

Printing spelled the end of the traditional paragraph mark. As demand for books increased toward the end of the medieval period, both monasteries and lay bookmakers were forced to make their production more efficient. The workers who made books were arranged according to their specialties: scribes wrote the texts; miniators sketched and painted the accompanying illustrations, or miniatures; and rubricators added decorative letters, headings, and symbols in contrasting colors—usually in red ink, or rubrica. Pilcrows, which were often placed at the start of a line to indicate a new paragraph, were typical of the marks left out by scribes and added later by rubricators.

When printing arrived in Europe in the middle of the 15th century, the earliest printers confined themselves to replicating the work traditionally done by scribes. They printed their texts in black ink only, leaving spaces for illustrations and rubricated headings and symbols to be added later so that their patrons could have their books customized to their liking. The problem was that there were so many new books to be rubricated that the rubricators could not keep up; increasingly, the spaces once filled with large initial letters and decorative pilcrows were left empty, and readers began to associate the absence of a pilcrow with the start of a paragraph. The pilcrow died and the paragraph indent was born in its stead.


Just as printing saw the pilcrow’s end, the arrival of the Internet has ushered in changes not just to the way we mark paragraphs but also to paragraphs themselves. First, more and more websites (like this one!) are separating their paragraphs by blank lines rather than paragraph indents. This makes sense: blank lines are easier to pick out than indents when scrolling rapidly through a web page, and, without physical pages to turn, paragraph boundaries are vital in helping the reader find their place in a text. In this way, the paragraph is more important than ever.

The second change is more confusing: paragraphs are getting shorter, and for no apparent reason. News websites routinely break their stories into paragraphs of only one or two sentences each, so that the centuries-old hierarchy of section to paragraph to sentence is collapsing in on itself. Soon, the paragraph and the sentence may become one and the same thing—and what then? This may finally be the end for the 2400-year history of the paragraph mark.