A post from Shady Characters

The Pilcrow, part 1 of 3

This is the first in a series of three posts on The Pilcrow (¶). Continue to PART 2 or view ALL POSTS in the series.

This is a pilcrow: ¶. They crop up surprisingly frequently, bookending paragraphs on websites with a typographic bent, for instance, and teaming up with the section symbol in legal documents to form picturesque reference marks such as §3, ¶7. The pilcrow even appears in Microsoft Word, where it adorns a button which reveals hidden characters such as spaces and carriage returns. (Click on that button, in fact, and a multitude of pilcrows will appear, one at the end of each line of text.)

For all this quiet ubiquity, the pilcrow gets short shrift in typographic reference books. Take the trouble to look it up and it becomes apparent that in most cases the humble pilcrow warrants only a few lines, dismissed briskly as a “paragraph mark”.1 The more generous definitions might run to mentioning that it has fallen out of common use and that it is sometimes used to indicate a footnote.2 No mention of where its reverse-P shape comes from, or its name; for the pilcrow, this is as good as it gets.

This is a crying shame. The pilcrow is not just some typographic curiosity, useful only for livening up a coffee-table book on graphic design or pointing the way to a paragraph in a mortgage deed, but a living, breathing character with its roots in the earliest days of punctuation. Born in ancient Rome, refined in medieval scriptoria, appropriated by England’s most famous modern typographer and finally rehabilitated by the personal computer, the story of the pilcrow is intertwined with the evolution of modern writing. It is the quintessential shady character.

The orthographic world of ancient Greece was a sparse old place. A literate Greek of Homer’s time reading a contemporary manuscript would be faced with an UNBROKEN­STREAM­OF­LETTERS, all of the same upper case (because at that time there was no other case)3; not only that, but texts were composed in a style called boustrophedon, or ‘ox-turning’, where the lines ran alternately left-to-right and then right-to-left across the page in the manner of a farmer driving his oxen across his field.4 Perhaps most cruelly, the visual signposts of punctuation which today we take for granted were completely absent, and it was the reader’s unenviable lot to tease out words, clauses and even sentences from this densely-packed zig-zag of characters.

Boustrophedon writing at Gortyn, Crete
Boustrophedon writing at Gortyn, Crete, circa 6th-5th centuries BC. (Photography by Phillip Hughes.)

Despite some recent scholarly murmurs to the contrary,5 it is generally held that the painstaking task of interpreting a document like this would have been accomplished by reading it aloud. Physically pronouncing the syllables helped a practised reader to decode and retain their meaning, and to discover the rhythms and cadences lurking in the unbroken text.6 In these ancient times the written word was very much an adjunct to spoken language, and silent reading was the exception rather than the rule.

Aristophanes of Byzantium, librarian of the great institution at Alexandria in the 3rd century BC, was the first to give readers some room to breathe, so to speak, when he created a system of marks for augmenting texts written according to the rules of classical rhetoric.7 A skilled orator would habitually pause after each unit of rhetorical sense, and so Aristophanes defined a series of dots to call attention to where these pauses should take place — a boon in particular for the increasing numbers of non-native readers (such as the Romans) attempting to decipher Greek literature.8 The 2nd century BC grammarian Dionysius Thrax described the system in his essay The Art of Grammar:

There are three dots: final, middle, underdot. And the final dot is a sign for a complete thought, while the middle is a sign taken up for a breath, and the underdot is a sign for a thought which is not yet complete, but is still wanting.

How is a [final] dot different to an underdot? In duration; for in the [final] dot there is a long pause, while in the underdot it is quite short.9

The so-called ‘middle’ (‘·’), ‘under’ (‘.’) and ‘final’ (‘˙’) dots were each placed after corresponding rhetorical units called the komma, kolon and periodos. Though it took centuries for these marks of punctuation to crystallise into more familiar visual forms, their modern names are not so far removed, and the ‘comma’, ‘colon’ and ‘period’ are still very much relevant today.

Middle and high points in Codex Sinaitucus
Middle and high points in John 5 from Codex Sinaitucus, circa 4th century. This image is a little disingenuous; although in Aristophanes’ native Greek, Codex Sinaiticus was written many centuries after his death, and this image was chosen for its legibility rather than its contemporaneity. (Image courtesy of the British Library.)

Unlike their modern counterparts, which impart semantic meaning to a text, Aristophanes’ dots were meant only as aids to reading aloud. A ‘·’, for instance, did not demarcate a rhetorical komma as such but only the pause for breath that a reader would take after reading one aloud. The importance attached to the individual reader’s interpretation of a text was so great, in fact, that not a single surviving manuscript prior to the Middle Ages has been found to be punctuated in the writer’s own hand;10 only when a reader prepared a text to be delivered aloud would it be annotated as such. Even now, many marks of punctuation still function wholly or largely as vocal stage direction: parentheses are the typographical signposts of a spoken aside, the exclamation mark implies a surprised, rising tone of voice, and the question mark is as much about inflection as it is about interrogation.

Aristophanes’ system found use only fitfully,7 and later, as Greece was usurped by Rome with characteristically brutal efficiency, it had to contend with the Roman disdain for punctuation in general. Cicero, for instance, the 1st century BC orator, philosopher and politician who crops up with indecent frequency in any discussion of punctuation or grammar, looked down his aquiline nose at it. He considered that the end of a sentence “ought to be determined not by the speaker’s pausing for breath, or by a stroke interposed by a copyist, but by the constraint of the rhythm”.11 While the practice of writing in the boustrophedon style had long since passed, Roman experiments in SEPARATING·​WORDS·​WITH·​DOTS, especially in monumental inscriptions, had been abandoned by the end of the 2nd century and Roman texts of the time were most often written in monotonous, unspaced scriptio continua.12

Paragraphoi in Menander's Sicyonians, circa 3rd century BC
Paragraphoi in Menander’s Sicyonians, circa 3rd century BC. Here, each paragraphos indicates a change in speaker somewhere on the corresponding line. (Image courtesy of Institut de Papyrologie, Paris Sorbonne.)

With all this emphasis on reading aloud, it might come as a surprise that the paragraph — a purely semantic construct, with no counterpart in spoken language — had been marked up in texts even before the advent of Aristophanes’ multifarious dots. The paragraphos, from the Greek para-, ‘beside’, and graphein, ‘write’, was a mark of punctuation that first appeared around the 4th century BC,7 and which took the form of a horizontal line or angle in the margin to the left of the main text. The exact meaning of a paragraphos varied both with the context in which it was used and the proclivities of the author, but most often it marked a change of topic or structure: in drama it might denote a change of speaker, in poetry a new stanza, and in an everyday document it could demarcate anything from a new section to the end of a periodos.13 In some uses the symbol itself marked the start of the new section; in others, it served only to draw attention to a break elsewhere on the line.14

The paragraph proved to be more robust in the face of changing tastes in punctuation than word, clause and sentence breaks had been, and by the 2nd century AD paragraphs were marked in a number of ways. The paragraphos soldiered on, albeit in a growing variety of forms such as ‘Γ’ and ‘γ’,15 while some texts dispensed with a mark altogether and chose instead to outdent and enlarge the first few letters of each paragraph to yield litterae notabiliores — literally, ‘notable letters’.12 Still others inserted the letter ‘K’ for kaput, or ‘head’, to mean the ‘head’ of a new argument or thesis,11 and it is this particular paragraph mark which would eventually give rise to the pilcrow.

'K' for 'kapitulum' in Cicero's In Verrem, circa 1st century BC/AD
‘·K·’ for kaput, or the ‘head’ of an argument, in Cicero’s In Verrem. Circa 1st century BC/AD. (Image courtesy of Universität Giessen.)

This motley collection of paragraph marks was typical of the state of punctuation at the dawn of the first millennium: written by one person and marked up by another (who most likely shared Cicero’s distaste for the form), texts were punctuated inconsistently or not at all. Punctuation was, however, about to be well and truly shaken up by the biggest upheaval since Rome’s fall from Republic to Empire. The arrival of Christianity would change the face of written language on a grand scale, and almost as an afterthought, it would kick-start the pilcrow’s journey from ‘K’ for kaput to a fully-formed mark in its own right.

Rosendorf, Theodore. “Pilcrow”. In The Typographic Desk Reference, 66+. New Castle, DE: Oak Knoll Press, 2009.


Unknown entry 
Lupton, Ellen, and Abbott Miller. “Period Styles: A Punctuated History”. In Design Writing Research. London: Phaidon, 1999.


Encyclopaedia Britannica. “Boustrophedon (writing Style)”.


Fenton, James. “Read My Lips”. The Guardian.


Saenger, Paul. “Silent Reading: Its Impact on Late Medieval Script and Society”. Viator: Medieval and Renaissance Studies, no. 13 (1982): 367-414.


Brown, T. Julian. “Punctuation”. Encyclopaedia Britannica.


Humez, Alexander, and Nicholas Humez. “Bang! The Dot Meets the Family”. In On the Dot: The Speck That Changed the World. Oxford University Press, USA, 2008.


Kemp, J Alan. “The Tekhne Grammatike of Dionysius Thrax: Translated into English”. Historiographia Linguistica 13, no. 2/3 (1986): 343-363.


Parkes, M B. “Introduction”. In Pause and Effect: Punctuation in the West, 9+. University of California Press, 1993.


Unknown entry 
Parkes, M B. “Introduction”. In Pause and Effect: Punctuation in the West, 10+. University of California Press, 1993.


Johnson, William A. “The Function of the Paragraphus in Greek Literary Prose Texts”. Zeitschrift für Papyrologie Und Epigraphik 100 (1994): 65-68. https://doi.org/10.2307/20189008.


Pearse, Roger. “More on the Paragraphos Mark”. Roger Pearse, November 10, 2010.


Parkes, M. B. “Glossary of Technical Terms”. In Pause and Effect: Punctuation in the West. University of California Press, 1993.


52 comments on “The Pilcrow, part 1 of 3

  1. Comment posted by Troy Curless on

    Thank you so much for this article. When I was a young man I independently created the boustrophedon style thinking I had invented something totally unique and inherently superior to traditional left right left right English. I even flipped the alphabet to avoid any ambiguity. The one thing I do though is to start new paragraphs at the side of the page that is closest to the side where the last paragraph ended.

    It was not until this day and your article that I discovered this existed in ancient history. I have always wanted a word processor capable of using the style I thought I invented, eventually I bought a graphics tablet.

    1. Comment posted by Keith Houston on

      Hi Troy,

      I’m glad you enjoyed the article! I’m amazed to hear that you invented boustrophedon independently.

      Stay tuned for more on the pilcrow — parts 2 and 3 are still to come, and the interrobang is up next.

    2. Comment posted by Sue Siler on

      Oh, thank goodness, the interrobang is still alive! After I first learned of it, I used it and taught it to students and generally promoted it to everyone I knew, but I made little headway in popularizing it. Then the “keyboard” replaced the typewriter, and I pretty much gave up as I could no longer “type” one. I can’t wait to find out how you construct one these days (other than when writing by hand).

    3. Comment posted by Keith Houston on

      Hi Sue,

      Thanks for the comment!

      I’m not sure I’ll be able to show you how to type one, but I’ll definitely be talking about interrobangs on the typewriter keyboard.

    4. Comment posted by MadArchitect on

      The alt code for an interrobang is ALT+8253, although its implementation differs from font to font, so it may not work depending on what font you’re typing in at the time.

    5. Comment posted by Ak on

      Anyway, it’s part of Unicode so it can be represented here: ‽

    6. Comment posted by Troy Curless on

      I started doing a bit of research on Boustrophedon over the last day. I found one other gentleman who is very similar to myself he also thought he had invented the style, like myself he apparently taught himself to be ambidextrous and like me he is a developer of sorts. Apparently he has created a Boustrophedon reader that formats text files, but as of yet the closest word processor is Abiword, though it is only roughly capable of doing so. I use a wacom tablet now, but have kept paper journals for many years. I attempted to contact the gentleman that created the reader via email, but he has not responded yet.

    7. Comment posted by Troy Curless on

      Oh btw on his site he suggest the Hitite culture may have used the style 1000 years prior to the Greek culture. I plan to do more research, because I think there are a few other people like us.

    8. Comment posted by Keith Houston on

      Hi Troy — thanks for the extra details, and good luck with your research! Maybe there’s a boustrophedonic word processor out there waiting to be popularised.

  2. Comment posted by Tricia White on

    Thank you for this article, I found if fascinating – can’t wait for parts 2 & 3

  3. Comment posted by Peter Myre on

    Very interesting and well written. Scholarly, but entertaining. I look forward to seeing your future articles.

    1. Comment posted by Doug Steele on

      Don’t you mean ‘Scholarly AND entertaining?’

    2. Comment posted by Johnny Storm on

      Hah, I think the implication is that the quality of being scholarly is diametrically opposed to that of being entertaining.

  4. Comment posted by John Boardley on

    Great article, Keith. You’ll also find some information about the pilcrow in Bischoff’s ‘Latin Palaeography’, where he suggests that the pilcrow or chapter mark is a medieval transformation of a sign that looked something like this: ¬

    1. Comment posted by Keith Houston on

      Hi John,

      Thanks for the suggestion! I’ll add it to my reading list. The alternative derivation of the pilcrow sounds intriguing.

  5. Comment posted by Roger on

    Hi Sue,
    regarding interrobangs, you can produce one using the charmap program that comes with windows, or you can copy and paste this into notepad for future use:

  6. Comment posted by Rudy Rooz on

    I don’t have access to Ref. 11 so, I’m puzzled by the claim for the letter ‘K’ representing ‘kaput’ or ‘head’. I’m neither an epigrapher nor an etymologist, but I would’ve thought the ‘K’ was a Greek ‘kappa,’ based on surrounding the text; unless it was inserted at a much later time. In which case, it would be the Latin ‘caput’ for head. If, however, it’s a later Teutonic inscription, the word for ‘head’ would be ‘kopf’.

    1. Comment posted by Keith Houston on

      Hi Rudy,

      I’m afraid I have to defer to Pause and Effect (ref. 11) on this one. The manuscript shown in the ‘‘·K·’ for kaput’ image comes from a plate in that book, and M.B. Parkes refers to it when describing the capitulum (or kapitulum), which is elsewhere cited as kaput.

      The script in the image does have a very Greek feel to it, but again Parkes’ transcription of it shows that it’s Latin. Just like you, I can’t claim any professional knowledge of the subject, but I suspect that early Latin scripts must have owed a lot to the contemporary Greek alphabet.

      Thanks for the comment!

    2. Comment posted by Tom Recht on

      I would guess that the K is originally a Greek kappa, and stands for kephalaion, ‘head’, ‘topic of an argument’. I don’t think the Latin word caput was ever spelled with an initial K. Of course, it would be easy for Latin speakers to associate this K with their own calque of the Greek word kephalaion, namely capitulum.

      A small quibble, by the way: it’s possible that “a literate Greek of Homer’s time reading a contemporary manuscript” would be reading in boustrophedon, but there’s no way to know, since there aren’t any manuscripts surviving from that period (whenever exactly “Homer’s time” was). There are only inscriptions, and these come in all imaginable directions – right to left, left to right, boustrophedon, sometimes even boustrophedon with alternate lines upside down!

    3. Comment posted by Keith Houston on

      Hi Tom,

      I’m afraid I’m neither a Greek nor a Latin speaker, so I can only defer to you in this instance. In all cases, though, I’ve tried to quote or paraphrase what seem to me to be credible references, so I do hope I haven’t published anything misleading or outright incorrect. In particular, I’ve used Pause and Effect a lot, and it seems to me that Parkes is considered the authority on this subject — see, for instance, his discussion of ‘K’ for kaput/kapitulum.

      Anyway, I hope you like the site in general, and thanks for your comment!

    4. Comment posted by John Cowan on

      All Old Latin words starting with /ka/ were originally spelled with k, though only a few survived into the classical period, like K. for Caeso or calendas. I expect you are probably right about the Greek origin of this particular K.

  7. Comment posted by Angela Taormina on

    Fascinating. Thank you for all your work researching and writing. I look forward to future installments.

    From another type geek, but one without the discipline to do the research you’ve shared with us…

  8. Comment posted by Ron on

    Great post! Very interesting. Waiting for the following parts.

  9. Comment posted by Karen Wise on

    Wonderful, thank you! I first heard the word “boustrophedonic” to describe those old dot-matrix printers that did indeed print the first line of type across from left to right, the second line back from right to left, and so on. As far as I can tell, the only thing anyone does boustrophedonically these days is mow the lawn.

  10. Comment posted by Bart on

    Fascinating article! I am definitely looking forward to the sequel.

  11. Comment posted by Rudy on

    Hi Keith, re my previous ‘K’ question, I just located “Pause and Effect” in Google books. Apparently, such marks were inserted much later by a scribe (teacher or student).

    Unfortunately, I can’t provide the (horrendously long URL) link due to pasting being disallowed in this Comment box. :(

    [A pain, btw]

  12. Comment posted by Keith Houston on

    Hi Rudy,

    With regard to the kapitulum, that’s my understanding too. It crops up fairly often in Parkes’ book, although there’s a lot of other material to be waded through first.

    I’m not sure which browser you’re using, but Google Chrome does seem to have some problems with pasting long strings into text areas like this one. You might want to try an alternative browser such as Firefox (I’ve just tried it, and it doesn’t seem to have the same problem), or a URL shortening service like bit.ly. I hope this helps!

    1. Comment posted by Keith Houston on

      Hi Linda,

      Thanks! Part 2 of the pilcrow entry will be available this coming weekend. And I enjoyed your article — it covers a lot of stuff I’ll be talking about in the future.

  13. Comment posted by Rudy on

    Re: cut-and-paste into the Comment form, I’m using Safari 4.1.3 on Mac OS X and, astonishingly, it seems to work today! But because I’ve never seen the problem elsewhere, I thought it was something you’d deliberately disabled. Sorry ’bout that.

    Anyway, for the interested reader, here’s the link to p.12 of Ref 11 in Google books:-

    [I’m not game to try and enter it as an HTML hyperlink]

  14. Comment posted by JB Christy on

    Oh, no! There goes my free time for the foreseeable future. I’ll be lurking on Shady Characters, breathlessly awaiting future installments.

    THANKS for your careful research, the amazing photos of early manuscripts, and your engaging writing.

  15. Comment posted by Luke Jones on

    Thank you for the incredible insight into the pilcrow and other characters I’ve never been aware of.

    As an aside, you’ve created a beautiful site here. It’s nice to see such a clean and easy-to-read site on the web.

    1. Comment posted by Keith Houston on

      Hi Luke,

      Thanks for the comment! I’m glad you like the site. Check back later today for the second instalment of the pilcrow’s story.

    2. Comment posted by Luke Jones on

      All ready to read now thanks. Looks perfect on the iPad.

    3. Comment posted by Keith Houston on

      That’s good to hear! I don’t have an iPad on which to test the site, so it’s handy to hear that it looks okay.

  16. Comment posted by Nan Erwin on

    Wonderful. I second the “scholarly and entertaining” comment. Hope you’ll include the many-named # at some future date.

    1. Comment posted by Keith Houston on

      Thanks! The ‘#’ will be making an appearance in due course.

  17. Comment posted by Levi Montgomery on

    Not quite sure how you’ve managed to get 30 comments on what I would have thought would be a dry and boring subject, of interest only to a few of us, um… weirdos, but I congratulate you on that, and I, too, will be lurking behind the curtains, waiting for more.

    1. Comment posted by Keith Houston on

      Hi Levi — just lucky, I suppose! I’m glad you like the site.

  18. Comment posted by Anniina Jokinen on

    Brilliant article! Thoroughly enjoyable in its lucidity, while imparting the pleasure one gets from learning something new. Thank you!

  19. Comment posted by Daniel Davies on

    What a great blog! Great post… thanks very much. Punctuation is far too often ignored, I’m definitely gonna be reading all these posts…

  20. Comment posted by Jim Forest on

    A friend in England, a member of the Orthodox Church, sent me the following comment: “In my copies of the Authorised Version (King James) Bible, you can see the pilcrow marks throughout the Gospels and Acts. Fascinated by this, I tried to find a pattern – and there is one. The marks define the passages prescribed in the Orthodox Service Books for liturgical readings. I can only assume that the translators in 1604 to 1611 got hold of a Greek manuscript which had been marked out for the use of the Deacon and Reader, and simply reproduced the pilcrows without understanding why they were there. Pilcrows also appear in some of the OT books in the KJV but I suspect there the translators were using Latin manuscripts based on translations from the Hebrew rather than Greek texts from the Septuagint.”

  21. Comment posted by Marc on

    Thank you for the explanation of the pilcrow. It is indeed a shady character. I would like to pose a question to another similar character – the obelisk aka dagger. I used to think that it symbolized the ending of a thought before the next sentence. As you know, paragraphs can run on on a general theme. Yet, I can’t remember if or where I read it or if I’m completely dreaming this up. I did some research on it online, but it too proves to be another shady character. Would you happen to have any comment on this creature? Anything would be helpful. Thanks again. marc

    1. Comment posted by Keith Houston on

      Hi Marc — I’m afraid I can’t add anything more on the subject of the obelisk/dagger. I haven’t come across the usage you mention, and personally I’ve only ever seen it used as a footnote reference and to mark the date of someone’s death. H&FJ discuss it at moderate length, and I hope to be able to cover it in more in future.

  22. Comment posted by Anthony Appleyard on

    Sometimes a wrongly-set-up computer printer will print text as boustrophedon.

  23. Comment posted by Janey on

    It’s not 6am yet, great;’Highly enjoable reading to start
    my morning.. I think I’ll swing my legs back towards my bed. Rest, but not long as I breath the fresh Air.. and wake up.

  24. Comment posted by Lawrence Bouett on

    As a surveyor and registered Professional Engineer, I have known the term “boustrophedon” for decades, but not in the context of writing; it is the term used for the manner in which sections (a square, one mile on a side, or 640 acres) are numbered within a township (six miles on a side, or 36 square miles) in the Public Lands Survey System (PLSS) that is used in the United states.

    Sections are numbered beginning with Section 1 in the northeast corner of the township and proceed West to Section 6; Section 7 is below Section 6, and the numbers then proceed East through Section 12. Likewise, the sections are numbered in a boustrophedon through Section 36.

    Thank you for a fascinating article.

    1. Comment posted by Keith Houston on

      Hi Lawrence,

      That rings a bell — I think I must have read about that in the course of my research. Thanks for the note! It’s a bit off-topic, but can I ask what happens when a township exceeds six miles on a side? Is there a larger boustrophedon coordinate system that supersedes the first?

  25. Comment posted by Christian Walter on

    Very interesting theme and good for me to catch some english expressions as you have written in a way not too hard to understand. Thank you!
    During my four years of apprenticeship as a typesetter in the early sixties, I learned about the Aline-Sign which looked much the same, but like a reversed D, not as reversed P. For the same purpose I still should have a single type with a face something like a big (centered) cross which also has a square body (or shank) and fills this completely. About both I was told, that they were mainly used in gothic times to show where a new alinea starts without interrupting the edge of the column. No-one asked at this time where it comes from, and I was really surprised when, changing from typesetting manually to the then rather new Mac, I found the sign again. At least on typewriters with the Swiss-German layout I never saw it. This may be due to the French accents we often need and therefore have no spare types to fill with special signs. Now you gave me the insight, that what I greeted as dating from some hundred years ago really goes back some thousand years. I think, I am growing older than I thought. And particularly nowadays it is worthwile to learn where we come from and that we have ancestors that knew how to think better than most of us today. Thank you for this part of a foundation to stand on.

    1. Comment posted by Keith Houston on

      Hi Christian — I’m glad you enjoyed the article. The pilcrow never really made it onto the typewriter keyboard — it skipped a generation of typesetting devices, so to speak. We’re fortunate that it survived in hot and cold metal type to take root again in digital typesetting.

      Many thanks for the comment!

Leave a comment

Required fields are marked *. Your email address will not be published. If you prefer to contact me privately, please see the Contact page.

Leave a blank line for a new paragraph. You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>. Learn how your com­ment data is pro­cessed.