A post from Shady Characters

The Octothorpe, part 2 of 2

This is the last of two posts in a series on The Octothorpe (#). Start at PART 1 or view ALL POSTS in the series.


Bell Telephone Laboratories, one-time research subsidiary of American telecoms giant AT&T, has produced some of the 20th century’s most influential developments in the worlds of science and technology. It boasts seven Nobel Prizes in Physics awarded for, among other things, a demonstration of the wave nature of matter, the invention of the transistor and the discovery of background cosmic radiation. Other notable products of this storied research centre include the laser, radio astronomy, the first communications satellite and the UNIX operating system,* which forms a key component of the internet and of modern computing in general.2

Most relevant in terms of punctuation, though, is a small linguistic innovation which emerged in the wake of a much larger technological one. Engaged in the 1960s in reinventing the world’s ageing telephone dialling system, one of the many Bell Labs engineers working on this mammoth task was almost certainly responsible for coining the term ‘octothorpe’. And though the octothorpe’s birthplace is well recorded, the question of exactly who created it is somewhat more contentious.


Since the arrival in the late 19th century of the earliest automated exchanges, telephone calls had been routed from caller to callee by a method known as ‘pulse dialling’. Each digit of the recipient’s number was transmitted to the exchange by interrupting the line a number of times equal to the desired number, producing a series of characteristic “ticks” as the telephone’s rotary dial is released (dialling the number zero produces ten pulses, thus avoiding the otherwise inevitable logical paradox). At roughly ten pulses per second, with pauses between digits of a significant fraction of a second,3 the time taken to dial a number tied up expensive call routing equipment at central exchanges for too long;4 as telephones became more and more widespread, the problem was compounded. Pulse dialling was overdue for replacement.

Devised in the late 1940s and refined a decade later with the advent of affordable transistors, Bell Labs’ new system consisted of a grid of buttons where each row and column was associated with a different frequency.5 When a button was pressed, it sent an audible tone comprising the frequencies of its row and column, and when received by the telephone exchange, the two frequencies were decoded to determine the position of the button on the grid.6

Early Touch-Tone or DTMF button layout with the assigned frequency of each row and column.7
Hertz 1209 1336 1477
697 1 2 3
770 4 5 6
852 7 8 9
941   0  

Named the ‘dual-tone multi-frequency’ system, or the friendlier ‘Touch-Tone’ for public consumption,8 this apparently simple design belied a high degree of sophistication. The combination of two frequencies in each tone reduced ‘talk-off’, where a human voice is accidentally interpreted as a button press, while the frequencies themselves were chosen at intervals which minimized the chance of confusing them.9 By transmitting tones in the normal range of human hearing, the new system could be used over existing networks’ copper wiring without the need for costly upgrades. And lastly, unlike pulse dialling, where the tick-tick-tick of each digit travelled no further than the local exchange, the audible notes of a Touch-Tone handset reached all the way to a call’s recipient, allowing callers to remotely control systems such as voicemail or telephone banking.8 Whether this latter feature constitutes an advancement or step back in the world of telephony would no doubt be contested by anyone who has suffered through an interactive voice menu.

Though the underlying system supported a 4 x 4 grid of sixteen buttons, the first consumer handsets had only ten: three rows carried the digits 123, 456 and 789, with 0 orphaned at the centre of a fourth row.10 The missing fourth column, labelled ABCD in the original design,11 was reserved for special applications; the US military’s Automatic Voice Network, for example, repurposed these extra buttons with priority codes so that a caller could override a lower priority call if there were no free lines.7

Military Touch-Tone telephone keypad
A Touch-Tone or DTMF keypad as shown on a military telephone. The fourth column of keys were normally absent from consumer handsets and were used chiefly in special-purpose handsets such as those procured by the US military. Labelled with priority codes, the caller could assign a precedence to their call and cause the system to terminate a lower priority call if there were no free lines. One of the buttons, labelled ‘FO’ for ‘Flash Override’, was reserved for use by the President only. (Photography by nesnet on Flickr.)

This layout, with ‘0’ isolated at the bottom of the keypad, was controversial. Accountants complained that their calculators already had an established keypad where the numbers ran in a strict descending order, and they felt that the proposed placement of the ‘0’ key was an affront to mathematical consistency. The (winning) counter-argument was that on rotary dial phones, each number doubled as a set of letters for mnemonic purposes — 2 = ABC, 3 = DEF and so forth — and that changing their positions would ruin the corresponding alphabetical order.5

The next problem faced was linguistic rather than mathematical, when, in 1968, the two unused buttons either side of the ‘0’ were finally made available for public use in controlling menus and other special services provided by the exchange.5 There were two questions to be answered: what symbols should these buttons carry, and what should they be called? Here the plot thickens.


Early in the development of the new keypad, a number of handsets had been produced where for testing purposes the two new keys had been labelled with a non-standard five-pointed star and a diamond symbol. The fact that these characters did not appear on standard typewriter keyboards proved problematic and Doug Kerr, a Bell engineer working on the new system, was tasked with selecting more suitable characters to replace them.4 Kerr was Bell Labs’ representative on the committee responsible for creating the American Standard Code for Information Interchange, or ASCII,12 the verbose title given to a standard character set to be adopted by computer manufacturers. Despite its name, ASCII was intended to be an international standard, and with this in mind Kerr rejected the existing symbols before choosing instead the asterisk and the hash symbol, satisfied that they both existed within the ASCII character set and appeared on the ubiquitous typewriter keyboard.

This done, the new characters needed names. ‘Asterisk’ was considered too difficult to pronounce and spell; ‘#’, as already discussed, had no commonly agreed name. To avoid rocking the boat any further, Kerr suggested that the names of the earlier test figures could be carried over: ‘star’ was an obvious choice, being easier to pronounce than ‘asterisk’ and accurately describing the character’s visual appearance, while adopting ‘diamond’ — a tenuous reference to the centre of the ‘#’ symbol — would avoid the confusion between ‘pound’ or ‘number’ symbol.

Two of Kerr’s peers, however, were not to be swayed. Howard Eby and Lauren Asplund had been involved in testing the original star/diamond keypad and were mildly piqued by the rejection of what they felt had been perfectly reasonable symbols. They sent Kerr a memo explaining how they had decided to help ‘solve’ this problem:

They told me that they had read with interest the part of my report in which I regretted the absence of a unique typographical name for the character “#”, and said they had solved my problem by coining one, “octatherp”. They said that it had no etymological basis, but they had been guided by one principle. They said they were irritated that I had rejected some candidate characters they thought were good on the basis of lack of compatibility with emerging international standards (with which the Bell System had a tradition at the time of little interest). Thus, they said, as a way of getting even, they had included in the name the diphthong “th”, which of course does not appear in German and several other languages and thus might be difficult for users of those languages to pronounce, which would serve them right.4

In personal correspondence with Kerr, Lauren Asplund elaborated on the etymology of their invented word:

Howard and I thought we needed a better name […] so one day early in 1964 we did some brainstorming at lunch. We finally decided in a jocular way to call the pound sign an “Octotherp”. That was because it had eight points and “therp” sounded Greek and also seemed to go well with the “octo” portion of the word.13

According to Kerr and Asplund, then, ‘octatherp’ rather than ‘octothorpe’ was the original form of the hash mark’s new name. Kerr decided to join Asplund and Eby’s light-hearted crusade and incorporated the word in his own documents, adding footnotes to the effect that ‘#’ was “sometimes called octatherp.”13 ‘Octatherp’ and its relatives started appearing throughout Bell literature and eventually made the leap into the wider world; as AT&T‘s new telephones became commonplace in the 1970s, newspapers stories about the new handsets often included asides about the ‘#’ symbol’s new name.14

The story is not quite as straightforward as it first seems, however, as the Oxford English Dictionary’s etymologies for the word ‘octothorpe’ (mentioned previously on Shady Characters) reveal. A decade earlier than the Kerr’s 2006 account, another Bell Labs engineer had come forward with his own recollection. Writing in the journal of the Telephone Heritage Group,15 Ralph Carlsen averred that ‘*’ and ‘#’ had been chosen because of their presence on the typewriter keyboard and that again the ‘#’ symbol’s multitude of names had proved problematic. Carlsen recounted the actions of one Don MacPherson, a Bell Labs supervisor sent to train customers in the use of the new system, who decided to create an unambiguous name for the symbol and inject some levity into his presentations at the same time.

Having selected the self-evident prefix octo-, MacPherson was in need of a second syllable to form a more convincing complete word, and his selection of ‘thorpe’ was almost as leftfield as Asplund and Eby’s conjuration of ‘therp’ out of thin air: the Native American athlete Jim Thorpe, who had died in 1953, was considered by some to be one of the best American athletes of all time. Having won both the decathlon and pentathlon at the 1912 Olympic Games at Stockholm, Thorpe’s gold medals were subsequently stripped from him when it was discovered that he had contravened the Games’ strict rules by having played baseball professionally during 1909 and 1910.16 MacPherson, an avid supporter of the campaign to have Thorpe’s medals posthumously restored, chose Thorpe’s name to form the second part of his new word, and began to include ‘octothorpe’ in his customer presentations and memoranda. Just as in Kerr’s account, MacPherson’s ‘octothorpe’ made its way first into AT&T literature and then more broadly into the outside world.15

Despite being given a lustrous sheen of credibility through its inclusion in the Oxford English Dictionary,17 Carlsen’s account alone was not deemed sufficiently authoritative, and the OED goes on to to muddy the waters yet further by citing a third possible origin at Bell Labs. Referring this time to a 1996 issue of New Scientist,18 it echoes Robert Bringhurst’s statement that ‘octothorpe’ comes from octo- for ‘eight’ and ‘thorpe’ for ‘village’. That this third Bell Labs source is not identified is emblematic of the trouble with the word ‘octothorpe’: created to address the ambiguities of naming the ‘#’ mark, it has instead raised a new host of related questions.


The octothorpe finds itself as relevant today as it was when Doug Kerr, Don MacPherson et al first got to grips with it on the keypad of their telephone handsets. Having been adopted by Twitter to identify ‘hashtags’, terms which allow tweets to be grouped together according to common themes, the character is enjoying a second day in the sun.19,20 Even if it does not quite match the scope of their achievements in physics and technology, Bell Labs’ creation is still with us; sadly for the octothorpe, though, no Nobel Prize in Literature has yet been forthcoming.

Nimble Books uses the recursive hashtag '#octothorpe'
Nimble Books creates the recursive hashtag ‘#octothorpe’. (Image courtesy of Nimble Books on Twitter.)
1.
E. Siever, S. Spainhour, and N. Patwardhan, “Perl in A Nutshell: A Desktop Quick Reference,” , 2 ed., O’Reilly Media, 2002. ↩︎
2.
Bell Laboratories (American Company),” Encyclopaedia Britannica, 2011. ↩︎
3.
T. Viswanathan, “Telecommunication Switching Systems and Networks.” Prentice-Hall of India Pvt.Ltd, 2004. ↩︎
4.
D. A. Kerr, “The ASCII Character "Octatherp",” , 2006. ↩︎
5.
S. Hochheiser, “Pressing Matters: Touch-tone phones spark debate,” Encore, vol. First and, pp. 13-14, 1999. ↩︎
6.
T. Viswanathan, “Crossbar Switching.” Prentice-Hall of India Pvt.Ltd, 2004, pp. 67-68. ↩︎
7.
M. Weik, “Computer Science and Communications Dictionary,” , vol. 1, 2000. ↩︎
8.
R. J. Bates and D. W. Gregory, “Voice & Data Communications Handbook (Standards & Protocols),” , 4 ed., McGraw-Hill Osborne Media, 2001. ↩︎
9.
L. Schenker, “Pushbutton Calling with a Two-Group Voice Frequency Code,” The Bell System Technical Journal, pp. 235-255, 1960. ↩︎
10.
Milestones in AT&T History – 1963,” AT&T, 2011. ↩︎
11.
R. A. Williams, “Communication Systems Analysis and Design. A Systems Approach.” Prentice Hall, 1987. ↩︎
12.
{. S. A. S. C. on Computers, X3. Information Processing, and B. E. M. A. (U.S.)}, American standard code for information interchange: Sponsor: Business Equipment Manufacturers Association. Approved June 17, 1963, , 1963. ↩︎
13.
D. A. Kerr, “\Personal correspondence\.” Keith Houston, 2011. ↩︎
14.
“You Asked Us: About The * And \# On The New Phones,” The Calgary Herald, 1972. ↩︎
15.
R. Carlsen, “What the \#\#\#\#?,” Telecoms Heritage Journal, iss. 28, pp. 52-53, 1996. ↩︎
16.
Jim Thorpe (American athlete),” Encyclopaedia Britannica, 2011. ↩︎
17.
J. A. Simpson, E. S. C. Weiner, and Oxford University Press, The Oxford English dictionary., Clarendon Press; Oxford University Press, 1989. ↩︎
18.
K. Dekker, “Letters: Internet Hash,” New Scientist, vol. 54, iss. 2023, 1996. ↩︎
19.
T. Dowling, “How the \# became the sign of our times,” The Guardian, 2010. ↩︎
20.
R. Fulford, “How Twitter saved the octothorpe,” National Post, 2010. ↩︎
*
In certain UNIX programs, the hash symbol is forced into a shotgun marriage with the exclamation mark — the ‘bang’ in printers’ parlance — to yield ‘#!’, the ‘hash-bang’ or ‘shebang’.1 ↩︎
Doug Kerr’s 2006 essay names two different engineers — John C. Schaak and Herbert T. Uthlaut — as the originators of the word ‘octatherp’. However, in personal correspondence with the author, Kerr quotes a message from Lauren Asplund which gives credit instead to Asplund and Eby.13 ↩︎

12 comments on “The Octothorpe, part 2 of 2

  1. Comment posted by Petter on

    At least in Sweden, dialing ‘0’ is one pulse, ‘1’ is two pulses and so on.

    1. Comment posted by Keith Houston on

      Hi Petter,

      Yup, quite right — I think Sweden, Oslo and New Zealand were the only places to opt out of the ‘ten pulses for zero’ system.

      Thanks for the comment! I hope you enjoyed the rest of the piece.

  2. Comment posted by Leonardo Boiko on

    This is from Kerr’s article and not your text, and I hate to be that guy, but someone has to point that “th” is not a diphthong. A diphthong is a kind of sound; specifically, two vowels gliding one into other in a single syllable (as in the start of the English word “ice”). “Th” is a pair of letters, not sounds (and it’s not even a vowel).

    I suppose what Eby and Asplund were getting at is that they wanted the character’s name to use one of the sounds that English represents as “th” (presumably the one in “thorn”). This is a simple, noncomposite sound, despite being written as two letters; pairs of letters used in this way are called “digraphs”, and Kerr or his colleagues might have confused the word with “diphthong”. But even then, digraphs are a feature of written representations, not of languages.

    1. Comment posted by Keith Houston on

      Hi Leonardo,

      Thanks for that! I think you’re right, and that Doug Kerr meant ‘digraph’ — still, I’m sure you’d agree that his recollections make for a great addition to the story of the octothorpe.

      Apparently the sound produced by the digraph ‘th’ in this context is called the ‘voiceless dental fricative’, rendered /θ/.

  3. Comment posted by Jason Black on

    Another great article! I have to say, though, the whole “octatherp” side-plot puts me strongly in mind of the children’s book “Frindle,” a charming little story about the intentional, arbitrary coinage of a synonym for “pen.” Arbitrary coinage for the win!

    1. Comment posted by Keith Houston on

      Hi Jason — glad you’re still enjoying the blog. I do enjoy the idea of some Bell Labs engineers conspiring to invent a new word, essentially for the hell of it. And to do it before the advent of the internet is quite an achievement!

  4. Comment posted by starry diadem on

    I love this column. You have just brightened up my Monday by several kilowatts.

  5. Comment posted by annie morgan on

    To say nothing about the fact that the unhuman voices tell you to enter your (payment or whatever) “followed by the number sign”.

  6. Comment posted by Bill Stewart on

    A somewhat related discussion that went on for many years at Bell Labs was “Where should we put the Q and the Z on the TouchTone keypad?” There were enough applications developed that involved “press a key one, two, or three times to pick one of the letters on it” that would break if you put the Q in the middle of the PRS on the 7 key or fail to notice the Z if you added it on the 9 key, and some people made keyboards with Q and Z on the 1 key, which had other problems, and it was the argument that Just Wouldn’t Die. Eventually everybody settled on the 7 and 9, but it took a long time.

    1. Comment posted by Keith Houston on

      Hi Bill — intriguing stuff! This might be something to look into for a future revision. Thanks for the comment!

  7. Comment posted by Brian on

    Man, this is good stuff. If you told me I’d be interested in the history of punctuation, I’d give you the most sideways of glances. I never would have guessed that the octothorpe (now my word of the day) had such a tortuous history.

    Do you have plans to collect these writings in a book some day?

    1. Comment posted by Keith Houston on

      Hi Brian — I’m glad you’re enjoying the site. When I started looking into all this I didn’t really know what would come out of it, but I’ve been amazed at the stories behind some of these marks.

      I’d love to do a book at some point. We’ll see what happens!

Leave a comment

Required fields are marked *. Your email address will not be published.

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>

1 link to “The Octothorpe, part 2 of 2