This Typefoundry post on the history and current state of the number that appears on the door of 10 Downing Street is pretty far afield from the usual concerns of Languagehat, but I happen to know that I have among my commentariat architects and other persons with an interest in this sort of thing, and besides, it’s a fascinating (and infuriating) piece, so here it is. There’s even a picture of “a small boy on a visit to London in 1924 who was allowed to pose for a photograph in the doorway. He was Harold Wilson, who became Prime Minister.”


  1. What a great site too. I’ll enjoy this piece tomorrow. Thanks!

  2. marie-lucie says:

    It’s amazing how much there is to say about a zero.

  3. Architect Erith: Dear Mr Mosley…It is nice of you to be so interested, but all I want to do is forget about it.
    Mr Mosley: notwithstanding the courtesy of his first letter to me, I have the sense that he did not really want the involvement of an outsider

  4. Thanks for posting this. I have a feeling that this will become one of my regular blogs to read.
    The passage quoted from Mr Erith’s letter (Dear Mr Mosley… It is nice of you to be so interested, but all I want to do is forget about it) seems a little misleading, in that it gives the impression of a brush-off, whereas the entire letter is about the most forthright I’ve ever read from someone in such a position, making it clear that he was truly furious (even long afterwards) with the incompetents he had to work with. I know the feeling.

  5. I didn’t think it was a brush-off, all he wanted was to forget about it. As an architect myself, I know that feeling.

  6. LH: It’s very nitpicky, perhaps a bit obsessive to someone not so deeply interested in typefaces, but why is it infuriating ?

  7. It is infuriating that such a simple thing as getting a number on a door right turned out to be impossible thanks to the usual nightmare caused by the intersection of politics, bureaucracy, and money. The people who know and care about such things are the least important people involved, and their input is the first to be jettisoned.

  8. araucaria says:

    The post title maybe works in US English, but in Britain it has to be “10 Downing Street” – you can’t omit the “Street”. I think that’s because it’s relatively common in Britain to find a whole series of streets, avenues, terraces, squares and soon with the same name in the same locality, which maybe doesn’t happen in the USA?

  9. Quite right, araucaria.
    P.S. When John Lewis opened a new store in Cambridge a couple of years ago, they waggishly made some play of their address: “10 Downing Street”.

  10. Araucaria, having streets, places, roads, and other things with the same name is fairly common in the US, but that doesn’t prevent people from omitting that part of the name when there is no ambiguity. In Washington, DC, it would not be unusual in informal contexts for someone to refer to 3701 Connecticut Avenue as simply 3701 Connecticut, or to 3030 K Street as 3030 K.

  11. But he’s not writing in British English.

  12. And I wouldn’t dare try.

  13. Actually, my first thought was to call the post simply “10,” but I was afraid that might bring up distracting associations.

  14. The article had led me to read about the adjoining 9 Downing Street* where the Chief Whip currently lives. 9 Downing Street would make a good title for a book — many different books, actually.
    *(as well as the better-known number 11. And 12)

  15. Another transatlantic difference, AJP? In the US, 11 Downing Street would normally be across the street from 10, and the adjoining addresses would be 8 and 12.

  16. Usually the same in the UK, KC, but not in some pre-19th C streets.

  17. The US also tends not to use street-numbers smaller than 100.
    I know of quite a few localities where all street-numbers obligatorily have five digits.
    I’m curious, does this reflect the origins of street-numbering? That is, the UKs’ go back so far they were the first house on the high street, or the tenth house on the lane with the maple, etc., until it became conventionalized. The US meanwhile tended to plan (somewhat) these things out, such that in many cities you can tell from the street-number what the nearest cross-street is.

  18. Trond Engen says:

    “No. 10″ alone would be unambiguous, wouldn’t it?

  19. Ahem, that should’ve been UK’s. I guess I was a little eager to show the plural explicitly.

  20. I grew up at 18 Pembridge Place, right next door to number 17 Pembridge Place, in London. Both were from the 1850s. Later I moved to 169 Sullivan Street in Greenwich Village, a building from about 1900 that’s next door to number 171.
    One reason the US sometimes has such huge numbers on its streets is because they (the USians) sometimes increase the numbers by up to one hundred in consecutive blocks, so they have the 100-block, the 200-block, and so on. That doesn’t happen in Britain, or in Europe at all as far as I know.
    One of the things about Norway that would drive me nuts if I let it is that I live at Deiligveien 162, rather than 162 Deiligveien. I think addresses ought to be written from the smallest to the largest (…18 Pembridge Place, London W.2., England, Europe, The World, Solar System, as we used to write as children).

  21. Stephen Dedalus
    Class of Elements
    Clongowes Wood College
    County Kildare
    The World
    The Universe

  22. We never got anything from further away than Australia.

  23. komfo,amonan says:

    The US also tends not to use street-numbers smaller than 100.
    Umm. I think this varies widely, probably depending on how old the town in question is. Much of NYC, for example, uses the European system wherein the first lot on the street is numbered 1, & the subsequent lots go up by 2 regardless of the intersecting streets. (Then again, much of NYC goes by the n-hundred block system like much of the USA.)
    An odd instance: in Brooklyn, Fifth Ave starts out using the former system, until it crosses 39th St, when it switches to the latter system.

  24. Yes, lots of the US uses the n-hundred block system, but often the first block is known as “the unit block” and has numbers starting at 1.

  25. It’s a good blog, I wish they’d posted more often. Some time ago I read their Roman Tragedy on how serifs developed. It also has some examples of painted lettering in Pompeii.

  26. Thanks for linking to that post; I recommend it to everyone interested in lettering (though it too is infuriating in a couple of places—why can’t we take better care of precious remnants of the past?).

  27. They should really do something about the 1 O in 10 Downing Street before it becomes an election issue. I would even volunteer to do it myself if I were in the country more often. I don’t buy Mr Mosley’s argument that the wide spacing is done to avoid hitting a molding, it would be easy to move the 1 a bit to the right.

  28. it too is infuriating
    I was in love with Optima-Palatino combination when I first read it. It almost makes you cry: putting a replica in a museum room and leaving the real thing erode outside. And the Pompeii writing washed away after surviving nearly two thousand years under ground, unbelievable.
    By the way, I am not entirely sure that it only was paintbrushes they used. Similar effect can be achieved with chalk and carbon.

  29. I’m curious, does this reflect the origins of street-numbering? That is, the UKs’ go back so far they were the first house on the high street
    There are some streets in Hong Kong – IIRC Repulse Bay Road, for example – where the houses are numbered in the order they were built. So 1 is in the middle, 2 is down one end, 3 is close to 1 but on the other side, 4 is off the other direction and so on. Hong Kong postmen are men of iron.

  30. Men?

  31. 1015. In a churchyard not far from where I live there is a piece of neglected stonework with this number, a broken tombstone or architrave. I’ve seen it several times before realising something isn’t right: it’s in Arabic numerals, but they didn’t spread in Europe until 1200s. The church sits on the famous medieval pilgrimage route to Santiago-de-Compostela and was built around 1015. Does anyone know somebody who is really passionate about Arabic numerals and their spread in Europe? If the stone is period it could be the first instance/evidence of their use in non-Moorish Europe.

  32. I would assume the carving doesn’t actually date from that year.

  33. that’s the crux of the matter: dating the carving (and the stone). The place was a Plantagenet stronghold, so even if the carving was done in Angevin times it is still a very early example of the use of Arabic numerals. And a valuable relic of the Norman-Arab-Byzantine civilization – the linking point between the Dark Ages and pre-Renaissance. I’ve been trying to find someone who could, at least roughly, date the carving.

  34. marie-lucie says:

    What does the carving look like? what type of stone is it on? with what tools does it seem to have been carved? is there a debate about the date among specialists in medieval architecture or sculpture?

  35. It’s just the number with a small floral cross above under a pointed gothic arch. The numerals are rounded, no serifs, the ones don’t have strokes to the left and the 5 looks a bit like S. It looks like a small architrave or the top of a tombstone. The stone looks like grey granite and it seems to have been chiseled. No, there is no debate, nobody seems to know about it or take any notice. The story of the numerals itself is fascinating, from the sorcerer Pope Sylvester II to Fibonacci. I want to find someone of authority who could tell me if the find is indeed interesting.

  36. marie-lucie says:

    I am not sure what a “floral cross” is, but a “pointed gothic arch” suggests a date definitely later than 1015.
    A “5″ looking like an “S” is not uncommon in medieval inscriptions, but since 1015 looks very much like IOIS (with “1″ of “I” drawn as a single line), could the characters be letters (initials) rather than numbers?
    Mr Crown, our resident architect, are you able to help?

  37. marie-lucie says:

    I mean “l” or “I”.

  38. Sashura, maybe a photo would help?
    I’ve always worried about the paint on Viking rune stones. It seems this is being done by museums (and I think it’s ugly), but even if it’s an accurate restoration, is it reversible and does it damage the stone?

  39. ok, I’ve put the photo here, click on the image to enlarge it. I would be grateful for any comments. No mercy, please. And Marie-Lucie, you are right as always: I published the photo two years ago in my magazine and one English reader suggested it was Lois rather than 1015.

  40. Could it be a headstone? Or part of a wall? I’ve seen carved stones in England on the inside of cathedral used for burial–the person was buried inside under the floor and the stone with name and sometimes their likeness in medieval battle gear etc. carved on it was part of the floor.
    If you click to enbiggen, it also seems to have something like serifs, but not identical ones.

  41. marie-lucie says:

    Thank you, Sashura.
    The stone does look old, but the “pointed arch” does not seem very pointed, so Gothic it is not (but that does not mean much, since the carver had to fit the arch within the dimensions of the stone).
    To me the characters look more like letters than numbers: IOIS (that is i, o, i, s), which could be the initials of a common Latin inscription (but I have no idea what that could be). The last letter does not look very much like an S, but even less like a 5.
    Even if the carved characters are indeed 1,0,1,5, I am skeptical that that those characters would have been used on such a stone, at such an early date. First of all, grave markers or other religious objects usually follow a traditional pattern, even if there is innovation in other aspects of the culture, so in the early medieval period that would mean that dates were recorded in Roman numerals. Second, even though the “Arabic” numbers were introduced in Europe fairly early, knowledge of the use of those numbers was not at all widespread, but restricted to accounting professionals. Even the positional notation took centuries to be fully adopted, and that made keeping accounts very difficult. So it is unlikely that many people in 1015 would have known how to write or read that sequence of numbers. Third, if there is only one inscription on a tombstone, it is likely to be the name of the deceased, and/or a traditional formula (the tomb of a member of a religious order might have just a formula, like “RIP”, without a name). Writing just a date would be strange: for instance, the pictures I have seen of Gaulish cemeteries of the Roman period (which often bear Latin-Gaulish bilingual inscriptions), do not seem to mention dates, even though they are quite precise about the deceased person.
    On the picture you can see another stone with a design, below and slightly to the right the one with the inscription, and another, smaller, less remarkable stone. It is not clear if those various stones were all part of the same structure, but perhaps there is a further clue in those stones.
    Is there a local society devoted to the history of the region, where someone might have a clue?

  42. oh thanks for the comments and the idea, Marie-Lucie, I’ll see if there is a local historical association, maybe they know more. Eleanore of Aquitaine stopped at the place several times, so it may be that a learned Moor accompanied her, or someone before her, hence, methinks, 1015 is not entirely out of the question.
    Nijma, I didn’t notice the serifs, but it does look like the symbols are slightly flared at the ends.

  43. Could the first I stand for Iesus? I’m then tempted towards Salvator for the S, but that’s ignorance speaking.

  44. marie-lucie says:

    Picky, those guesses sound good. Then O is probably Omnis (or a form of it).

  45. Or Orbis?

  46. marie-lucie says:

    Orbis, yes, possibly.

  47. and what is the second I for?

  48. Well, inglese, obviously

  49. sorry, I don’t understand.

  50. Sorry. Bad joke.

  51. aqilluqqaaq says:

    Not Jois, as a family name, then?

  52. It looks like these Arabic numerals did arrive in Spain and France at an early date, here are some examples, but they didn’t look like the numbers we use now. As far as Christian abbreviations, so far I’ve found ICHTHYS (Iēsous Christos, Theou Huios, Sōtēr-Jesus anointed god’s son savior), INRI (Iesus Nazarenus, Rex Iudaeorum-Jesus of Nazareth king of the Judeans), and IHS, a monogram of the name of Jesus (used from the 3rd c.), derived from the first three letters of the Greek version of his name: Iota, Eta and Sigma. Assuming the language is Latin.

  53. whatsthisthen says:

    It looks like these Arabic numerals did arrive in Spain and France at an early date, here are some examples, but they didn’t look like the numbers we use now. As far as Christian abbreviations, so far I’ve found ICHTHYS (Iēsous Christos, Theou Huios, Sōtēr-Jesus anointed god’s son savior), INRI (Iesus Nazarenus, Rex Iudaeorum-Jesus of Nazareth king of the Judeans), and IHS, a monogram of the name of Jesus (used from the 3rd c.), derived from the first three letters of the Greek version of his name: Iota, Eta and Sigma. Assuming the language is Latin.

  54. Trond Engen says:

    Never worked so well, apparently.

  55. Is the blog broken, or is it me?

  56. Oh, crap. Sorry for the mess, Hat. That was totally invisible on my other laptop, along with Trond’s comment, even though I was able to post and view a comment on my own blog with no problem. Cache problem or public WIFI connection problem? Kept returning “zero message error” from the WIZ
    (City of Chicago hotspot server). I suspect some days are just bad karma for blogging. It’s a sign.

  57. marie-lucie says:

    Nijma, thank you for digging up those number characters. Judging from them, there is no way that an inscription looking like modern “1015″ would be from that early date. Also, those characters all come from manuscripts, and except for “1″ they look too complex overall to be intended for carving in stone: Roman characters are mostly composed of straight lines, much more suitable for this purpose. So if the inscription does represent the date 1015, it was carved much later. But why should a date alone be written prominently on a single stone? (unless it was done in a pseudo-medieval style, centuries later).
    aqilluqqaaq, your suggestion of a family name Ioisis a good one, but 1015 (when the church was built) is before the Norman conquest, and Iois could be “Joyce”, definitely a name from after the conquest (“I” and “J” were not differentiated in spelling for many centuries after the Romans).

  58. if the inscription does represent the date 1015, it was carved much later.
    exactly, but how much later? That’s why I am so curious as to whether it is possible to date the stone. Also, we tend to think land-wise about the spread of knowledge. But, remember, it was the height of Norman-Viking maritime power, the strongest link connecting all of Europe, from Normandy to Sicily to Constantinople to Kievan Rus. There were learned Arabs at the Norman court in Sicily.
    I found a historical association there, going to contact them.
    why should a date alone be written prominently on a single stone?
    It looks like it was part of something bigger – the lower part is jagged.

  59. You know, to my thoroughly untutored eye the problem with the lettering is that it’s difficult to place it by period – in fact the only period that rings true to me is … C19th – C20th

  60. yeah, Beardsley-Waterhouse kind of style. But, then, they modeled their style on something medieval: Beardsley illustrated Malory’s Arthur.

  61. Yep, well, with a sort of conception of something medieval, anyway. But can you see a medieval mason doing those stylised characters? And (OK I can only see a photo) once you discount the stylising of the characters, and the matching stylisation of the cross, isn’t the work on the arch a bit crisp for something many centuries old?

  62. @Nijma: if these are early numerals, it may be the presence of the zero which is most surprising.

  63. exactly: early sets of new numerals in Europe didn’t have one, Sylvester’s abacus didn’t have one either.

  64. Picky, I’m more surprised by the absence of zero. The Arabs had one. The WP article is frustratingly slim on detail, and a quick glance at Amazon and Google Books shows more attention to the concept of zero than to the development and spread of the symbol itself. This article credits Leonardo Fibonacci’s book Liber Abaci (1202) with the introduction of zero to Europe,

    It was used in European mathematics from the 12th century, and entered common use from the 15th century. Robert Chester translated the Latin into English.[citation needed]
    The familiar shape of the Western Arabic glyphs as now used with the Latin alphabet, (0, 1, 2, 3, 4, 5, 6, 7, 8, 9) are the product of the late 15th to early 16th century, when they enter early typesetting.

    Fibonacci (“Leonardo of Pisa”) studied with Arabs while living in North Africa with his father, Guglielmo Bonaccio, who wished for him to become a merchant.
    As for the crispness of the arches, it seems to disappear with a closer view of the stone. It’s not new, the lichens have been there a long time and the stone has probably been in that position a long time too, judging by the extra moss growth where the shadows fall. The surface is pitted enough to look like concrete, but I have never seen moss grow on concrete. You can see some sort of serif or tail or perhaps merely a design element on the lower left side of the first upright character.

  65. Oops, link to history of Hindu–Arabic numeral system article.

  66. marie-lucie says:

    The stone looks old, but the carving is not necessarily as old as the squaring of the stone. The larger image makes the final character clearer: it definitely looks like an S, not a 5. This S is somewhat pitted, like the right bottom edge of the stone, which suggests that end of the stone might have spent time in the soil at one point (therefore that the stone stood on its small end, used for a purpose different from the one it was meant for, after the destruction of the structure it was originally part of).
    The cross inside a kind of niche with a rounded (not pointed) arch seems to have been a popular motif at a certain time, although the cross pushed upwards in order to make room for an inscription seems unusual. If indeed the stone was part of a larger structure, it would have been more traditional to have the cross in its niche on the top stone, and the inscription on the stone below.

  67. Here is a closer view of the whole end of the stone. I don’t know that much about dressing stone–I imagine it’s done with chisels–but in carving something like that from wood, it would be much easier to carve the arch. The chisel could be used at an angle and with a jig to guide it, shearing the unwanted material away from the block in a series of smooth motions. The area between the characters would have to be done with more of an up and down chipping motion and would not look as smooth. The area between the letters does look more weathered but not the rest. I wonder if this could be explained by the shade as well, the part in shadow not being dried by the sun as much. Also, the photograph doesn’t show the angle with respect to the ground, whether there might have been some pooling of water on some areas of the stone that would have been a factor in weathering the stone while it was in the position it is in now.

  68. marie-lucie, can you pinpoint the ‘certain period’?
    But can you see a medieval mason doing those stylised characters?
    oh, yes, here in France, every other crossroads has a stone cross or a chapel, and in my parts, the granite country, many are medieval or even older with with beautiful curving ecriture. Granite is a strong stone and withstands well the harsh weather.
    I’ve dug out ‘Mathematics’ by Jan Gullberg (1997).
    With the zero, it’s not entirely clear what happened. Both West Arabic (Maghreb, with the circle for zero)numerals, now extinct, and East Arabic numerals (Egypt, with the dot for zero, like in modern Arabic numbers), had zero. And number 5 looks remarkably like S in West Arabic set. That’s assuming numerals came via Maghreb-Spain.
    On the other hand Robert of Chester is credited with the first translation into Latin, 1120, of the 825 Persian work with a fully developed numerical system by al-Khowarizmi. My book says it was Chester who wrongly assumed them to be Arabic, and the term stuck.

  69. marie-lucie says:

    I wrote “a certain period” because I could not determine what it was (such is the ambiguity of the word “certain”). In looking for a picture of something similar to your stone I saw a few tallish monuments with a similar design at the top (but with just the cross), but I did not save the pages and have been unable to go back to them. They looked old, but not 1000 years old.
    I think that the definitive reference on numbers and their origin is by Georges Ifrah: Histoire universelle des chiffres, translated into English as A Universal History of Numbers, a massive work which covers a vast number of cultures.

  70. The 5 is all wrong too, at least compared to the manuscript versions.
    The “passing Moor” theory–digressing perhaps, here are the Arabic numbers. Zero is a dot, five looks like “o”. The WP article is incorrect about Arabic numbers being reversed like the letters. The place values, at least as they are written now in Arabic, are the same as ours. So the year 2001 stamped in a passport looks like ٢٠٠١.
    It seems as if there is about 500 missing years of writing numerals from the time they appear in Europe to when they spring into print in a completely different form. Was there no intervening tradition of carving on stone? Maybe a complete tangent, but there was this Scandinavian pentimal number system for carving runes on stone, with a character for zero, possibly a secret cipher used by the tailors guild, oldest authenticated use in 1885. Sacrobosco’s 1230 Algorismus Vulgaris was the first text to introduce the “Indian” zero into the European university curriculum; in Scandinavia it was Erlendsson’s c. 1320 Hauksbók.

  71. Holy cow, zero was a big no-no as late as 1594.

  72. When writing modern Standard Arabic, it’s normal to inscribe the digits “1234″ in the order 4, 3, 2, 1, right to left. This corresponds to the canonical MSA spoken form, literally “four and thirty and two hundred and one thousand”. However, the most significant (thousands) digit stands leftmost, just as in all modern cultures using a place-value decimal notation system.
    This is not true in writing Persian, where the hand skips leftward leaving enough room to write the digits, inscribes them left to right, and then skips leftward again to continue writing, just as when Roman-script text is embedded. That is why Unicode has entirely different representations for the Arabic and “Arabic-Indic” digit sets, even though only a few of them actually look different; they have quite different bidirectionality properties.

Speak Your Mind