The "pound sign" mystery

« previous post | next post »

Yesterday, in discussing Kevin Fowler's song Pound Sign, there was some debate about the origin of the term "pound sign" for the symbol #.  I suggested that it all started with the substitution of # for £ on American typewriter keyboards, but others argued that # was a standard symbol for pound(s) avoirdupois. I've heard this theory before, but I expressed skepticism about it because I've never actually seen the symbol used that way.

Today, after some further research, I'm still not completely sure. But I've found a new theory, which I think has a better chance to be correct: it's all Emile Baudot's fault.

William Safire asked about this usage in his On Language column for 1/20/1991, "MAIL CALL (ALL CAPS)":

… you and I rarely use suite or even apartment number , abbreviated to Apt. No. ; most often, we use the crosshatch symbol for "number," which has come to be known as the pound sign (the origin and etymology of which I seek Lex Irreg help in finding).

Others writing about the same time indicate that this usage was unfamiliar to them as well. Thus Ron Alexander, "Metropolitan Diary", NYT 6/20/1990:

The responses are delivered in a reassuring, synthesized male voice that never hesitates or sneezes or coughs. It says to "push the pound sign after entering the deposit amount." The pound sign, the voice explains, "is to the right of the zero; the asterisk is to the left of the zero."

I believe that AT&T used "pound sign" for this symbol from the first introduction of the touch-tone keypad in 1963, but it's clear that in 1990-1991 the term was still new to many Americans. I left AT&T in 1990, and at that point, "touch-tone penetration" (as we used to call it) was still low enough to be an argument for outfitting voice response systems with speech recognition technology, even though that technology was then very expensive and not very robust.

But touch-tone penetration (the proportion of households with touch-tone as opposed to rotary phones) started to pick up after 1984, when customers could buy their own phones rather than leasing them from AT&T as part of their service package. And by 1990 or so, voice response systems were starting to take off, and people throughout the country were forced to learn what a "pound sign" was.

Safire reported the results of his query a couple of months laters,  in "Hit the Pound Sign", 3/24/1991. He starts with a fictionalized voice-mail interaction:

"Please enter your password," the synthetic-syrupy recorded voice directs, "followed by the pound sign."

"When was I born?" is my first internal question. The second, which I asked only once, was: "What and where is the pound sign ?"

I always thought the pound sign was the symbol for the British pound sterling, a script capital L with a hyphen through it, a stylized representation of the Latin libra . (The Roman pound was several ounces lighter than the modern pound; some things stick in my mind, but numbers fade fast.)

No; the sign that the recorded Miss Syrup refers to is what some of us remember as the tick-tack-toe sign, or the crosshatch, or the sign of the double-cross, similar to the symbol Charlie Chaplin wore as a parody of the swastika in his 1940 movie, "The Great Dictator."

After some more stuff making fun of voice menus, Safire gives some of the things that are not answers:

When #, as we shall call it here, is placed before a number, it is called a number sign ; #1 pencils have a softer lead than #2 pencils, and nothing fits into a #10 envelope anymore. When a musician sees it, the meaning is "sharp"; doctors often use it as a symbol for "fracture."

When a proofreader uses the #, the meaning is "insert space here": two words incorrectly runtogether are happily separated by a #.

And finally he gives the answer that he's settled on:

The origin [of the touch-tone keypad term] may be from the use of # to mean "pound," as in "a 5# bag of sugar," written by someone unhappy with the abbreviation lb. to stand for "pound." (We know that pound is from the Latin libra pondo , "a pound by weight," which accounts for the lb. ; not everybody knows that) […]

A more remote possibility is that pound evokes a mashing of the desired button, as one pounds on a door, but I go for the 5#-bag theory.

The trouble with this theory, as I observed in my earlier post, is that the evidence for # meaning "pound avoirdupois" is so thin. It's clear that Safire himself had never encountered it, or he wouldn't have needed to ask for help from his readers. I myself have never seen it in a grocery store or a recipe or a newspaper ad — and I've been shopping and cooking for myself since 1963.

Lists of abbreviations from the late 19th and early 20th century don't use it:

Thus N.A. Calkins, Manual of object-teaching, New York, Harper & Brothers, 1882:

Or Bernard Fantus, A text book on prescription-writing and pharmacy, 1905 ("Second Edition, Thoroughly Revised and Adapted to the Eighth (1905) Edition of the the United States Pharmacopoeia"):

That last reference is especially relevant because it uses non-standard typography for several other measures, but gives "pound" as simple "lb."

The earliest example that I've been able to find of the symbol # apparently used as shorthand for "pound (as a measure of weight)" is the OED's citation:

1923 W. E. HARNED Typewriting Stud. II. 29/1 Special Signs and Characters..#..Number or pound sign; # 10 (No. 10); 10# (ten pounds).

A similar reference can be found in Nancy Lawrence et al., Correlated studies in stenography:
the correlation of business correspondence, English, office practice, and shorthand
,
1932:

But these might literally refer to some rather specialized kind of business shorthand, rather than a general usage.  There's a more helpful example in Milton Kaufman, Radio operator's license Q & A Manual, 1955:

Similarly, the number or pound sign (£) cannot be reproduced in transmission and in lieu thereof the sender should use the words NUMBER (or NO), POUND (or LB) or POUNDS (or LBS), as the case may be.

This one pretty clearly comes from the equivalence established in (various versions of) the Baudot Code. As the Wikipedia article explains,

Baudot invented his original code during 1870 and patented it during 1874. It was a 5-bit code, with equal on and off intervals, which allowed telegraph transmission of the Roman alphabet and punctuation and control signals. It was based on an earlier code developed by Carl Friedrich Gauss and Wilhelm Weber in 1834.

The code was entered on a keyboard which had just five piano type keys, operated with two fingers of the left hand and three fingers of the right hand. Once the keys had been pressed they were locked down until mechanical contacts in a distributor unit passed over the sector connected to that particular keyboard, when the keyboard was unlocked ready for the next character to be entered, with an audible click (known as the "cadence signal") to warn the operator. Operators had to maintain a steady rhythm, and the usual speed of operation was 30 words per minute. […]

Since it was a five-bit code, 32 symbols could be represented. In order to extend this set, there were two mode symbols, allowing the system to shift into "letters mode" or "figures mode".

(The following details are taken from this page, perhaps with some scribal or interpretive errors on my part.)

In the version of the code adopted as ITA 1, several of the "figures mode" symbols were left "at the disposal of each administration for its internal service". One of these was 11011, which in the version of Baudot used by the British Post Office was N in letters mode, and £ in figures mode. I believe that in U.S. versions of ITA 1, this was # (though I haven't been able to find a reference to nail this down).

In the version created by Donald Murray in 1901, the five-bit pattern 10111 was the letter X in "letters" mode, but was the pound sign £ in figures mode (that is, after a figures shift sign 11011 had been sent).  In America, the Teletype Corporation adopted Murray's system with some changes, one of which was to substitute # for £ as the meaning of 10111 in figures mode.

In the ITA 2 five-bit code (in use since 1930 or so),  the details are again different, but the relevant correspondences are the same. In letters mode, 10100 is the letter H, but after a figures shift shift 11011 has been sent, it's £ in international versions of the code, and # in American versions — called "pound sign" in both cases.

So I'm quite sure that this is why the engineers at Bell Labs called # "pound sign" — it corresponded to a Baudot code-point that had been used for £ in the UK and # in the U.S., probably since the late 19th century and certainly since the early 20th century. Those guys would have memorized the ITA 2 tables, without any question.

This is probably also why typewriter keyboards shared this meaning for shift-3.

A few questions remain.

1. When was the Baudot figures-mode correspondence between £ and # first established?

2. When the correspondence was first established, was it essentially random? That is, did some Americans say to themselves, "We don't need that £ symbol, so what should we use that code-point for? How about §? Naah. Well then, what about #? Sure, why not." Or instead, was there already an established usage of # for pound(s) avoirdupois, so that it was a natural pun?

My money is on the "random code-point re-use" theory. And on that view, any sporadic use of # to mean "pound(s) avoirdupois" would be a secondary and derivative effect, not a cause.

[By the way, I once spent a week in Denver recording the young woman that Bill Safire called "Miss Syrup". But that's a story for another time.]



77 Comments

  1. Mel said,

    July 18, 2010 @ 1:09 pm

    I have definitely seen # used after a number to mean pounds, although not frequently and I'm having trouble remembering specific contexts. Archery is one, I think (to refer to draw-weight, e.g. a 50# bow).

    Doesn't shed any light on the history, though.

  2. onymous said,

    July 18, 2010 @ 1:18 pm

    I didn't realize that the name "octothorpe" for this symbol also had its origins at Bell Labs, but this seems to be the consensus of the internet, despite some disagreement over whether the original was "octatherp".

  3. ~flow said,

    July 18, 2010 @ 1:30 pm

    one possible explanation that "#" has been made to stand in for "lb.", "lbs.", "℔" and so on and be called pound sign might be that in olden days signs of measurement were often written in italics or had an italics-like appearance. "£" and "§" have a distinct hand-written appearance; traditionally, "1ℓ" was often preferred over "1L" or the unreadable "1l", and "1,—ℳ" was once used in germany for one mark (of currency) in preference over the straight "M".

    now there is a cursive variant of "℔" (nicely illustrated at http://upload.wikimedia.org/wikipedia/commons/7/7d/Gewichtmaße1.jpg and http://de.wikipedia.org/wiki/Pfund) where the stroke that forms the belly of the "b" is elongated and drawn to the left so that it crosses the preceding "l", then takes an upward turn to the right and forms a flourish that hovers above both "l" and "b". i guess variants where that final flourish crosses the stems of the letters might have been commonplace, too.

    now, the resulting figure is remarkably similar to "#": connect both the endings on the lower right as well as those on the left and the tictactoe morphs into a libra pondo sign. "#" is also somewhat easier to write than "", so it could be an abbreviation of a shorthand. which, to me, offers the hypothesis that "#" really *originated* from libra pondo (instead of being repurposed). in which case we have to wonder: how did it come to be understood as the number sign?

  4. Mark said,

    July 18, 2010 @ 1:51 pm

    Nice piece; thanks!

  5. Bruce Rusk said,

    July 18, 2010 @ 2:20 pm

    Here's a 1930 example:It's from a shorthand book, but the instruction is for a simplified form of #, suggesting that # was already a familiar sign (at least for readers of the books).

  6. Eric Fischer said,

    July 18, 2010 @ 2:21 pm

    I don't have a reference either for N being # in US Baudot, but it seems believable because the original French Baudot paired N with "No" for number.

    I'm not sure where that claim that Teletype's modified Murray code paired # with X comes from though. The Morkrum code of 1912 (Teletype was Morkrum before it was Teletype) made these pairings:

    – @ : $ 3 % & £ 8 ' ( ) ? c 9 0 1 4 # 5 7 , 2 / 6 "
    A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

    and included both £ and #, paired with H and S, respectively, and this remained the case through at least 1916.

    In 1918 Western Union was using this same code but got rid of # for a control code, "THRU", and 1919 Western Electric (by then the parent company of Teletype) was doing the same, using the former # for "BELL" but preserving £ paired with H. This was still being used by Western Union in 1926 according to the documents submitted to the CCIT convention.

    The use of X for £ seems to have started in the UK. That is the practice cited in their submittals to the 1929 CCIT convention. The ITA2 emerging from the convention actualy standardized "/" to be paired with X, and at least some US users did follow this — for example the promotional material for the Teletype Model 28 shows this pairing, as does an article on Western Union's code standardization in 1956 (in both before and after versions).

    If there is anything that ever mixed up £ and # on the same code, I think it may be that 1956 Western Union reorganization that did it. Prior to 1956 they had been pairing £ with H as they had been doing for decades. Beginning in 1956 they paired # with H instead and abandoned £. So there was probably a transitional period where operators were sending # and having it received as £ and vice versa.

    But the "Pound" name for "#" is older than that. Google Book Search turns up the a reference in "Correlated studies in stenography" from 1932 and in "Boyd's syllabic shorthand text book" from 1903. I don't think telegraphic codes can be blamed or credited for the calling # "Pound".

  7. Bruce Rusk said,

    July 18, 2010 @ 2:21 pm

    The previous attempt at embedding obviously didn't work; here's a link:
    http://books.google.com/books?id=T5JAAAAAYAAJ&pg=PA58&img=1&zoom=3&hl=en&sig=ACfU3U0KAmpCFyczpF7SP0qC3cD2qeR8YA&ci=77%2C1039%2C847%2C314&edge=0

  8. Noumenon said,

    July 18, 2010 @ 2:22 pm

    I work in a plastics factory and sometimes abbreviate a tare weight as 65#, but it might be a usage I got from growing up with touch tones!

  9. Lucy Kemnitzer said,

    July 18, 2010 @ 2:30 pm

    This all makes sense. It would have the hotshot entrepreneurs of groceries and dry goods adopting the # sign in the 30s, and it would explain how the usage became ordinary enough to seem obvious to some of us while still remaining unusual enough to be mysterious to others.

    This could be tested by laboriously looking over old photographs of places that sell things by weight, and seeing when the earliest ones occured.

  10. Stephen Jones said,

    July 18, 2010 @ 2:59 pm

    [By the way, I once spent a week in Denver recording the young woman that Bill Safire called "Miss Syrup". But that's a story for another time.]

    Like next week?

    In English voicephone messages in Saudi the symbol is always referred to as hash. I think this is the same in Sri Lanka.

    The COCA doesn't seem to have any examples of hash followed by a cardinal number, The BNC has at least five examples.

  11. sollersuk said,

    July 18, 2010 @ 3:13 pm

    At least the introduction of the euro for most of Europe has simplified one thing.

    Previously, strictly speaking, script capital L with two lines across = pounds sterling. Script capital L with one line across = Italian lira. Both derived from the Roman "libra" but a colossal difference in monetary value.

  12. Nicholas Waller said,

    July 18, 2010 @ 3:26 pm

    @ Stephen Jone – In the UK it's also usually, as far as I am aware, known as the hash key, as in instructions here for changing the voicemail PIN on an Orange mobile phone:

    "3. Enter a PIN between 4 and 10 digits followed by the hash (#) key"

  13. Martinus Scriblerus said,

    July 18, 2010 @ 3:37 pm

    When touch tone phones were first introduced in the UK it was known as the square key. I remember this from working through PORs (Post Office Requirements) from the days before BT was separated from the GPO – about 1980. The term hash key came later.

  14. mollymooly said,

    July 18, 2010 @ 4:04 pm

    Is

    the number or pound sign (£) cannot be reproduced in transmission

    really what Kaufman's 1955 book says, or is that a typo for

    the number or pound sign (#) cannot be reproduced in transmission

    ?

  15. Ran Ari-Gur said,

    July 18, 2010 @ 4:27 pm

    @~flow: Indeed, that is the English Wikipedia's explanation.

    Perhaps the use of "#" as a number sign results from a similar process applied to the other number sign, (abbreviated from "numero").

  16. Kutsuwamushi said,

    July 18, 2010 @ 4:43 pm

    When I did some bulk ordering for a coffeeshop I worked at, I came across "#" used to mean pounds, as in weight, for the first time. I'd never come across it outside of that context, but when I asked my manager about it, he didn't think it was at all unusual. It wasn't only one of our suppliers that used it either.

    You wouldn't find these uses by Googling, because they were only on the wholesale order forms, which you often can only get if you have a business license. I got the impression that it's a symbol that's often used on the backend of the business, but not presented to customers because they might find it confusing.

  17. Lugubert said,

    July 18, 2010 @ 4:46 pm

    Wiki gives close to 900 hits for "British dollar sign", meaning £.

    Contrary to what was stated for "European keyboards" in the previous set of comments, my Swedish kbd has £ at AltGr 3, $ at AltGr 4 and # at Shift-3.

    My cellphone, made by a partially Swedish company, has a hash sign that's neither horizontally nor vertically slanted. It is normally referred to as "fyrkant", lit. four-edge, 'square'. I prefer to think of it as "brädgård", 'lumberyard', from our standard map symbol.

  18. Kim W. said,

    July 18, 2010 @ 5:07 pm

    This is a pretty low-tech (and probably folk etymology) explanation, but I always thought the pound sign was named as such because it resembles jail bars. That, and it's easier to remember than 'octothorpe'.

  19. Jim said,

    July 18, 2010 @ 5:39 pm

    And here appears to be usage in a nautical-industrial context? (beam at top of frame: "Gross Wt. 10.000#"):

    http://spectrum.ieee.org/energy/environment/can-kevin-costners-machines-really-help-the-gulf-cleanup

    (specific image):
    http://spectrum.ieee.org/image/1641822

  20. Sili said,

    July 18, 2010 @ 6:18 pm

    I have before me the 8th edition of the Merck Index from 1968. I can't access it in Google Books, but on p1264 there's a list of prescription notation like the one quoted. But it adds exactly the stylised figure mentioned ~flow: a handwritten lb with the loop of the b elongated back across the l. With an additional flourish at the onset of the l it does look a bit like #. It's also pretty much the sign used by nan in the occasional recipe.

    The Merck Index goes back to 1889, so it should be easy to check how long that particular table has been included.

    In case noöne else can access Google Books either, I can scan the relevant page.

  21. army1987 said,

    July 18, 2010 @ 6:33 pm

    I think that is the most plausible origin for #.

  22. army1987 said,

    July 18, 2010 @ 6:35 pm

    (The explanation in Wikipedia http://en.wikipedia.org/wiki/Number_sign seems to make perfect sense to me.)

    [(myl) It makes sense, but it's not clear that it's true. No citation is given for the claimed development from lb with a horizontal stroke to #. (And note that the 1905 citation shows ℔ as the symbol for the obsolete Troy pound, with lb as the symbol for the Avoirdupois pound.) Nor is there any clear history given for the "number sign" usage, which seems equally murky to me at present — I can't find any clear examples earlier than the 1923 OED citation.]

  23. B K said,

    July 18, 2010 @ 8:18 pm

    I'm a little suspicious that referring to the double cross as 'the pound sign' was "still new to many Americans" in 1991. My first syntax class (with Jim McCawley) would have been round about 1989, and I remember being surprised he called the pound sign 'the double-cross.' My memory is that everyone I knew called it the pound sign and my father and uncles certainly used it as an abbreviation for pounds when they wrote down large quantities.

    Never a good idea to trust one's memory, but it does seem like google (or any other digital document collection) will not be a great source of information about how common this usage was.

    [(myl) I don't have any statistics (though the archives of AT&T marketing probably do). But it's clear that in 1990-91 some writers at the NYT (including William Safire) were among those innocent of this knowledge. And at that point, AT&T and other makers of voice-response systems felt that users needed to be told that the "pound sign" was that thing to the right of the zero. They certainly didn't feel the need to tell people what and where the zero was — that's something that everyone really did know.]

  24. Julie said,

    July 18, 2010 @ 8:20 pm

    Touch-tone phones may be spreading the pound sign now, but my 80-year-old mother did not learn it from them. I just asked her, and she has no memory of learning the symbol…meaning she's probably been using it since the 1930's or 40's, long before she had a phone. (And nobody in our little town had touchtone till the 80's.)

    I'm pretty sure she took typing and shorthand in high school, though. Wonder if it was taught in "business" classes? I wouldn't expect to see a migration from business classes to recipes and shopping lists.

  25. Nijma said,

    July 18, 2010 @ 8:48 pm

    Here is a pound sign in our family Thanksgiving recipe for sweet potatoes written out for me by a home ec major around 1980 or so. I remember it from recipes as a kid, also in supermarket signs. Also, per wikipedia:

    At first "lb." was used; however, printers later designed a font containing a special symbol of an "lb" with a line through the verticals so that the lowercase letter "l" would not be mistaken for the numeral/digit "1". Unicode character U+2114 (℔) is called the "L B Bar Symbol", and it is a cursive development of this symbol. Ultimately, the symbol was reduced for clarity as an overlay of two horizontal strokes "=" across two forward-slash-like strokes "//".

    [(myl) Again, the Wikipedia entry cites no source for the development of ℔ (which at least at one time was the symbol for Troy pound as opposed to Avoirdupois pound) into #. So maybe it's true, and maybe it's a semiotic/typographical folk etymology.]

  26. Sili said,

    July 18, 2010 @ 9:08 pm

    I see Lars Mathiesen has dug up the Danish pound sign that looks like a capital Π.

    The prescription one I found in Merck, has the overbar displaced down.

  27. Chris said,

    July 18, 2010 @ 9:42 pm

    I'm Australian, and this post was the first I've heard of that symbol being called a pound sign. My reaction to the last few posts on bleeping in comics was "Pound sign? Wouldn't a 'hash' be a more common bleep-character than a '£'?"

  28. Aaron Toivo said,

    July 18, 2010 @ 11:23 pm

    Before just a few years ago, it was not yet clear to me there was anyone who would call £ a pound sign – singular – and this still seems bizarre and alien to me. It is the pounds sign, plural, indicating units of British currency, which are countable entities, while "pound" is a non-count name for a telephony function. The pound sign # and the pounds sign £ are different symbols with different names – as far as I knew for the first 30-ish years of my life.

  29. Peter said,

    July 19, 2010 @ 12:47 am

    I'm also in Australia. The only name for the symbol other than "hash" that I can recall hearing is from mainframe programmers, who call it "crunch".

  30. Mike said,

    July 19, 2010 @ 1:22 am

    I remember my father, who owned a print shop, using the hash mark to denote paper weight. For example he would write 20# for twenty pound paper. This was in Florida in the 1950's. I'm pretty sure that he never studied shorthand.

  31. C Thornett said,

    July 19, 2010 @ 1:23 am

    I also remember # being used to indicate pounds in weight when I was growing up on the West Coast, California and Alaska, in the 50s and 60s. The specific context that comes to mind is the fruit and vegetable section of supermarkets. The feminist in me suspects that perhaps Safire wasn't sufficiently familiar with grocery shopping, but is willing to concede the possibility of regional variation.

    It was also commonly used to indicate 'number' at that period.

    My memory is that it was a 'hatch' mark, as in cross-hatching, rather than a 'hash' mark, although it was much less likely to be refered to by its name as a symbol; in other words it would usually have been called a pound or number sign.

  32. Bread & roses said,

    July 19, 2010 @ 1:49 am

    pound
    hash
    number
    crunch
    double-cross
    crosshatch
    sharp
    octothorpe
    octatherp

    Somehow it seems like this list is oddly violent.

  33. Yuval said,

    July 19, 2010 @ 1:58 am

    Hebrew also has excessive naming for this symbol:
    Sharp, Diaz (both of musical origin, used for the slightly-different ♯), Hash, Pound (all direct loanwords so far), Rishtit (from reshet, which means "net", but rishtit is also used for "retina"), Sarig (from the fabric knit), and the most acceptable – Sulamit (from sulam, "ladder").
    7 understandable forms – not too shabby!
    (I posted about this, in Hebrew, at http://tinyurl.com/26zjdv9)

  34. Joyce Melton said,

    July 19, 2010 @ 3:14 am

    Just to add another datapoint, I remember the hash symbol being used as an abbreviation for pounds in groceries, carpentry and printing during the sixties and perhaps earlier. Bell calling it a pound sign did surprise me a bit, though, since I thought of the thing's name as being "number sign".

  35. Chris Hunt said,

    July 19, 2010 @ 5:29 am

    Thinking back to my mis-spent youth, I remember seeing # used to denote pounds weight in my Dungeons and Dragons rulebooks of the early 80's. Perhaps it's specific to certain regions of the US? Or maybe Gary Gygax was son of a grocer?

  36. Max said,

    July 19, 2010 @ 8:33 am

    Aaron Toivo, do you call "$" the dollars sign? For me this is definitely called a "dollar sign". And people have "dollar signs in their eyes".

    Note that my home country uses neither pounds (of either kind) nor dollars.

  37. Luke said,

    July 19, 2010 @ 8:55 am

    Just a note – # is not the symbol for "sharp" in music. ♯ is a sharp. The first has slanted verticals, the second slanted horizontals. A right-angle variety might be possible for the pound sign, but I've never seen a sharp without the appropriate angles.

  38. Arnold Zwicky said,

    July 19, 2010 @ 10:29 am

    To Aaron Toivo and Max: In most N+N compounds in English, the first N is not actually in SG form, but is an unmarked N stem (which of course usually is phonologically identical to a SG form). This is true whether the first N is mass or count; a count N in first position is usually interpreted semantically as plural (as in job market).

    PL-in-form first Ns do occur, and there's a considerable literature (some of it referred to on LLog) on the choice between the two variants for count Ns (Raider rooter vs. Raiders rooter). But in any case pounds sign isn't forced by generalizations about English morphology.

  39. Arnold Zwicky said,

    July 19, 2010 @ 10:35 am

    Luke: "I've never seen a sharp without the appropriate angles."

    Just look at ascii-ized text. Pound-sign for sharp-sign and lower-case-b for flat-sign (F# minor, Bb major) are all over the net (see the way these key names are displayed here), though some sites (like iTunes) go to the trouble to use sharp and flat rather than get the symbols "wrong": F-sharp minor, B-flat major.

  40. Andrew Foland said,

    July 19, 2010 @ 11:40 am

    I can attest to usage of # for actual pound weight by restaurant prep cooks since the 1980's.

  41. Skullturf Q. Beavispants said,

    July 19, 2010 @ 11:42 am

    I share Chris Hunt's memory. I can't verify this now because I no longer have the books, but my memory is that # was used for pounds in the 1978 edition of the Player's Handbook and/or the 1979 edition of the Dungeon Master's Guide. If anybody has a paper or electronic copy, this would be easy enough to check.

  42. Richard Pfeiffer said,

    July 19, 2010 @ 12:36 pm

    One more (U.S.) Westerner who remembers the use of the "number sign" as a stand-in for "lb." in the produce section of grocery stores in the 1960s and thereafter.

    I also recall being were taught about the number sign in elementary school, back in the late '50s. When I was introduced to the use of the term "pound sign" in the late '70s or early '80s, in the context of computer science, I remember being kind of weirded out, since the other term was so deeply engrained.

  43. Richard Pfeiffer said,

    July 19, 2010 @ 12:37 pm

    That is, "I also recall being taught about the number sign in elementary school, back in the late '50s."

  44. mollymooly said,

    July 19, 2010 @ 1:29 pm

    Some GoogleBooks-ing finds a 1903 antedating in "Boyd's syllabic shorthand text book" p.58:

    Make the (#) pound or number sign thus: 37# = 37 Ibs.  #37 = No. 37.

    The 2nd and 3rd # glyphs in the original are Boyd's own shorthand symbols; the 1st is a genuine # but looks hand-made, so I presume it wasn't yet part of a typesetter's common stock.

    BTW this is also the earliest cite for "number sign" in this sense I found. Dunno what OED offers there.

  45. Will said,

    July 19, 2010 @ 1:57 pm

    For £, "pound sign" and "pounds sign" both sound fine to me.
    Similarly, for $ both "dollar sign" and "dollars sign" sound fine.

    In either case I'd probably opt to write/say it in the plural myself, but wouldn't even notice if it was singular or plural when reading it.

    But I agree that calling # a "pounds sign" seems definitely wrong.

    Then again it might feel right to people who use # as shorthand for lb. (regardless of whether this usage is regional, jargon, derivative, or rare, it seems clear that there it is plenty attested).

  46. Morten Juhl-Johansen Zölde-Fejér said,

    July 19, 2010 @ 2:09 pm

    It is interesting to see it used in a wholesale and maritime context – maybe it would be relevant to go back to old ships' cargo manifests?

  47. Stephen Jones said,

    July 19, 2010 @ 2:59 pm

    "pound sign" and "pounds sign" both sound fine to me.
    Similarly, for $ both "dollar sign" and "dollars sign" sound fine.

    So do we talk about 'bookscase' and 'biscuitstin'?

  48. Rubrick said,

    July 19, 2010 @ 3:12 pm

    I suddenly want to open a head shop so I can advertise it as the area's "#1#| dealer".

  49. Dennis Paul Himes said,

    July 19, 2010 @ 3:23 pm

    When I first heard someone refer to the uncertain origin of the term "pound sign" for # I was surprised, because it seemed obvious to me that it was from its use as an indicator of weight.

  50. Lamar said,

    July 19, 2010 @ 3:27 pm

    When I was starting out in the printing business about 30 years ago, paper was commonly marked as "20#" for 20-pound bond, or "60#" or whatever the weight was, so this practice goes back at least that far.

    L.

  51. Mr Punch said,

    July 19, 2010 @ 3:45 pm

    I recall the "pound sign" being used in the late '50s/early '60s to mark containers of, e,g., 10-pound nails — that is, indicating a measure based on weight rather than weight itself. It seems to me that the usages noted above with regard to paper, archery, and perhaps in some applications fruits and vegetables (indicating size?) are similar.

  52. J. W. Brewer said,

    July 19, 2010 @ 6:33 pm

    I'm a bit puzzled by the sequencing implied in "it all started with the substitution of # for £ on American typewriter keyboards." A bit of googling suggests that it is common ground that the QWERTY typewriter keyboard originated in the U.S. But alas the Ur-layout apparently lacked a shift key. Without having quickly found online confirmation, I would tentatively assume that the U.S. convention of # resulting from hitting 3 with the shift key down was likely developed quite early on, rather than waiting for the Brits to stick the £ somewhere and then figure out what else to do with the equivalent location for the U.S. market. So I would think the historical question is more likely to be what led the Brits to stick the £ where we have the # and not, for example, where we have the $. And indeed if for whatever reason the development sequence ran the other way you'd think it would have been more obvious for the U.S. designers to substitute $ for £, since # would be more likely to have *some* use in both markets.

  53. J. W. Brewer said,

    July 19, 2010 @ 6:43 pm

    Should have added: Prof. Liberman's more refined thesis is that the £ to # correspondence in Baudot (whatever its own origin) then leaked over to typewriter keyboard layout on the respective sides of the Atlantic. Is that direction of influence plausible in terms of either timing or which industry/technology would be likely to influence which? I note that the old teletype keyboard shown at the wikipedia article for Baudot looks very untypewriter-like because it doesn't do lowercase (and I am just old enough myself to have done radio newscasts in the mid-1980's with an old-fashioned teletype printing out wire service copy in ALL CAPS on rolls of very low-grade paper, so that feature of that technology remained to the bitter end). 3 seems to appear on that keyboard as shift-E, meaning there's no shift-3.

  54. lynneguist said,

    July 19, 2010 @ 7:44 pm

    This post has inspired me to do the post I've been meaning to do about £, #, and their names. In case it's of interest: http://bit.ly/a0xLKB

  55. Darla-Jean Weatherford said,

    July 19, 2010 @ 9:22 pm

    I'm sure I remember # as both "number" (#2 pencils) and "weight" (10# sack of flour) from my childhood in Texas from the 1950s; I'm more surprised that others are surprised by the use for weight than that we southerners appear to use it that way.

    On the other hand, I also have a vague sense that I stopped seeing it at the grocery store by sometime in the 1970s, but I believe Istill saw it at least on reams of paper (20#, 60#) for some time after that.

    Maybe it was used regionally, faded out, and drifted back in—with a vengeance when we had to start "pounding" cell phone keys.

  56. Kaleberg said,

    July 19, 2010 @ 11:47 pm

    You see the pound sign used a lot in 19th century ledgers and accounting books. I have no idea as to its origin.

  57. Ron Stack said,

    July 20, 2010 @ 12:45 am

    @Nicholas Waller @Stephen Jones: That would explain why in Twitter a group, topic or search term with the pound sign prepended (e.g., #linguistics) is called a "hashtag".

  58. Keith M Ellis said,

    July 20, 2010 @ 7:39 am

    "I didn't realize that the name 'octothorpe' for this symbol also had its origins at Bell Labs, but this seems to be the consensus of the internet, despite some disagreement over whether the original was 'octatherp'."

    That's a bit of 'net lore—but much less widely known is that "octothorpe" was a spur-of-the-moment joke that, IIRC, someone at Bell Labs took more seriously than they ought.

  59. Jill Lundquist said,

    July 20, 2010 @ 8:37 am

    I took a photo of a sign using # to mean per-pound and sent it to an English friend who had asked me why on earth Americans called that thing a pound sign. The sign was posted in a shop window, either in the grocery store or next door, and I took the picture in the summer of 2006.

    In case my embedded link doesn't work:
    http://ofb.net/~jill/US_Pound_Sign.jpg

    Since that picture will eventually vanish, I'll record what the sign says:
    ——————-
    Fresh Lake
    Superior Whitefish
    Available Now!

    $4.79 #

  60. Stephanie said,

    July 20, 2010 @ 12:13 pm

    This thread is extremely interesting, but then, I am a (Canadian) postgraduate student, & my research focuses on the history of punctuation & annotation.
    For what it's worth, '#' seems to be a very new symbol — which may explain why it apparently lacks a fixed name or usage :
    1) it is (obviously) not a medieval scribal siglum, notae Tironianae, or alchemical symbol, nor is it a descendent of such a symbol (no real surprise there);
    2) it does not make an appearance in any of the Renaissance or early modern printers' manuals or type specimens, nor in any tachygraphy or cryptography manuals from those periods (my search spanned 1000-1750).
    3) It is not mentioned in any of the modern authorities on the history of punctuation, typography, scripts/scribal practices, etc — not even in Parkes' "Pause & Effect" or Updike's 2 vol. history of "Printing Types".

    Bringhurst asserts that it is a traditional cartographic symbol properly called an 'octothorpe', & provides a probably spurious etymology ('8 fields'; "Elements of Typographic Style", 2005, p.314). Although this has been disputed, sec. 4.43 of "Scientific Style & Format : the CBE Manual for Authors, Editors, & Publishers" makes the same claim : "The octothorpe ('8 fields') has been used in cartography as a symbol for 'village' & in the avoirdupois system of weights as a symbol for 'pound'. These uses are archaic for scientific publishing, but the sign is still used by proofreaders & copyeditors [. . .] & by printers [. . .]" (1994, p.67).
    It is also often used as a logic symbol.

    You can take the '8 fields' etymology or leave it. I can say with some confidence that '#' isn't a traditional or (for lack of a better term) 'proper' typographic symbol, but more of a very modern, somewhat ad hoc technical symbol. Contemporary type & graphic design manuals rarely include it; on the odd occasions when they do, they are usually silent as to the name/origin. There is no formal consensus on either point, & even the accepted use of the symbol is fairly 'loose'. Bringhurst includes it among the miscellaneous modern legal and commercial logograms (p.76), & I'm inclined to agree with his classification (if not his etymology). So, I suppose we can say that '#' is properly called whatever you want to call it.

    I don't know if any of that is useful, but it might help your next pub trivia game.

    [CBE Manual link : http://books.google.com/books?id=PoFJ-OhE63UC&printsec=frontcover&dq=scientific+style+and+format:+the+CBE+manual+for+authors,+editors,+and+publishers&hl=en&ei=kLtFTJSOI8H68AaiyZXwBA&sa=X&oi=book_result&ct=result&resnum=1&ved=0CCwQ6AEwAA#v=onepage&q=cartography&f=false ].

  61. Stephanie said,

    July 20, 2010 @ 12:43 pm

    I should add, its inclusion in Boyd's shorthand was unexpected. My mother (a now-retired food broker) went to secretarial school in Halifax, Nova Scotia, in the early 1960s. They learned Pitman shorthand — it is one of the older systems — and the # sign is *not* in Pitman. She remembers the teachers calling it the "number sign".
    After graduation, my mom went to work for an off-shore shipping company. Her boss told her to *never* use # : in all the records, ships' manifestoes, etc. the preferred form was "No.".

  62. SeekTruthFromFacts said,

    July 20, 2010 @ 10:01 pm

    China Mobile have recently changed their recorded announcements (in English) from "and press the pound sign" to "and press the number key". Presumably they've discovered that the "pound sign" usage is peculiar to North Americans.

  63. Terminologia etc. » » # nomi inglesi del cancelletto # said,

    July 23, 2010 @ 3:03 am

    […] comuni associati al simbolo #) e scoprire, in particolare dai commenti a questo e a un altro post, The "pound sign" mystery, che il cancelletto viene chiamato in molti altri modi in inglese, spesso con differenze d’uso […]

  64. Frans said,

    July 24, 2010 @ 7:25 am

    In Dutch we call it hekje (little fence), or at least that's what the carriers call it. Before I that I was primarily familiar with the kruis (cross) from music (sharp in English).

  65. Sili said,

    July 25, 2010 @ 7:19 pm

    Finally got around to firing up the scanner for that reference I gave upthread.

    Pound sign from the Merck Index.

  66. Craig said,

    August 28, 2010 @ 10:39 pm

    It seems that here at LL, when a topic is dead, it's dead. But this prompted me to respond.

    I'm 49 and grew up in central New York State ("Upstate" to you New Yorkers,) and my entire life I have known the symbol '#' as representing pounds, as in weight, e.g. "5# sugar" on a grocery list. This predates not just computer keyboards, but also push-button telephones (thanks for making me cop to that!) so my memory isn't influenced by those, unless it's completely addled – which is looking less and less implausible these days, but that's another story.

    I believe I also know enough about my own mental process that I would not have this strong association simply from someone (read: my parents) telling me this. I must have seen it in print enough times to confirm it, such that when I read here that the association was in doubt, my immediate response was, "Huh? Mystery? What's all the fuss about? Of course it stands for pounds." Not that I can conjure any mental images from decades past of actually seeing this in print.

    Also, it makes sense that designers, or more likely marketers, of touch-tone phones and computer keyboards would name the symbols on the keys according to designations that would be familiar to people at the time, and not just make up random or obscure names. They would want to gain acceptance for the new technology in any way they could. Computer terminology is full of such examples – sometimes whimsical (like "mouse"), but usually utterly pragmatic.

  67. Julie said,

    September 2, 2010 @ 5:53 pm

    Craig…I accidentally stumbled on your post…it seemed very weird to me that so many people had never heard of it. Like you, I grew up with it and assumed that everyone understood it as I did. There are two threads, and I found myself defending it against people who thought it must be brand-new.

    Turns out, there are a lot of us, but huge swaths of America were never exposed to the #.

  68. John said,

    October 31, 2010 @ 3:29 pm

    I certainly grew up seeing it in grocery stores and printers in the 1950s, in both Detroit and western MA. I recall, too, a class, probably third grade, in which the point was explicitly made that # meant either 'pound' or 'number', depending on whether it came after or before a number.

    Today, in SW Florida, I continue to see the symbol used both ways.

  69. Alan Shaw said,

    September 30, 2011 @ 9:29 pm

    Simple logic indicates that Safire's not having heard of something is no disproof of its existence. But beyond that, he had a notorious(ly) tin ear for spoken usage. Worst. Language. Columnist. Ever.

  70. The many lives of # | Word Geeks said,

    December 4, 2012 @ 7:44 am

    […] only three months until #grammarday?) But how did it get the name "pound" sign? This Language Log article explores that question, and concludes that "pound" for # may have its likely origin in […]

  71. The Hashtag: A History Deeper than Twitter | CopyPressed said,

    July 1, 2013 @ 9:30 am

    […] http://languagelog.ldc.upenn.edu/nll/?p=2461 […]

  72. Anonymouse said,

    May 25, 2014 @ 6:44 am

    In Australia, # is almost always called hash, The only place I have heard it called 'pound sign' is in some textbooks and foreign publications.
    As an example, in the late 1990's a telephone service was introduced to call back the last unanswered number. The number for the service was *10# and advertised (spoken) as 'star ten hash'.

  73. Steven Harper said,

    June 20, 2014 @ 8:58 am

    I have a recipe written by my grandmother (who was born in 1900 and died in 1986) in which she used # to denote pound (i.e., "1 # sugar"). She transcribed this recipe to give to my mother some years before her death. It could have been written as early as 1970, when I first remember asking my grandmother for this particular recipe. I seem to recall that my mother had to explain the meaning of the symbol to me at that time. Since other people have suggested the abbreviation was used in the wholesale trades, it may be relevant that she grew up in a family that operated grocery stores northeast Ohio in the early 1900s. On the other hand, mother was apparently familiar with this abbreviation–had she learned it from her mother-in-law, or did she learn it independently?

  74. Trent said,

    June 21, 2014 @ 1:03 pm

    If the British money symbol was substituted in a typewriter, then why was the $ symbol not replaced?

    Hmm…teletype, maybe

  75. James Wimberley said,

    June 25, 2014 @ 6:11 am

    A prescription. The "pound" usage for # is ambiguous, and in the days of automated voice instructions to operate number pads it's confusing to non-Americans and could be dangerous in emergencies. (Confession: I'm a 68-year-old Oxford graduate and I wasn't familiar with it before today). Twitter has made "hashtag" universal, so let's please standardize on "hash".

  76. David Crosbie said,

    July 1, 2014 @ 9:25 am

    I have read in various places that # is derived from lb with a cross-stroke and refers to pounds weight. Of course, that don't mean that it's true, but I'm very much encouraged to believe it on learning of bona fide character ℔ — Unicode 2114.

  77. David Crosbie said,

    July 2, 2014 @ 5:09 am

    Oops!

    Just seen Nijma's posting of July 18, 2010.

    Well, it could be a made-up etymology for the term pound sign, but rather too sophisticated for a 'folk etymology'. It's would be more like the invention of a well informed individual than the collective decision of speakers who were sensible but uninformed.

RSS feed for comments on this post