Talk:Number sign

Alternate name

I recall some data processing people whom I knew in the 1970s pronounced it "box" when reading it aloud in a string of characters. It was one of just two (along with "@") non-letter non-numerals that was permitted in computer program and file names in the IBM-360 era. WHPratt (talk) 06:53, 31 January 2020 (UTC)

Do you have a source UnsungHistory (Questions?) (Did I mess up?) 18:54, 21 November 2024 (UTC)

It is not productive to reply to years-old comments simply to ask if they have a source. Remsense ‥ 论 01:24, 22 November 2024 (UTC)

Do I have a source? Only my recollection that it was used in vocal confirmation of program names at our site (representing the Technical Colleges throughout South Caroline). The article lists as alternate names "crosshatch, crunch, fence, flash, garden fence, garden gate, gate, grid, hak, mesh, oof, pig-pen, punch mark, rake, score, scratch, scratch mark, tic-tac-toe, and unequal," but somehow "box" needs expert witness to qualify? WHPratt (talk) 14:41, 18 April 2025 (UTC)

Fair comment, so I have removed the list of uncited random makey-uppy nicknames as not enclopedic. Even a dictionary would want to see evidence of actual use. --𝕁𝕄𝔽 (talk) 16:02, 18 April 2025 (UTC)

Do we want to have the random makey-uppy names? https://web.archive.org/web/20100721065416/http://www.nytimes.com/1991/03/24/magazine/on-language-hit-the-pound-sign.html, which I found through https://languagelog.ldc.upenn.edu/nll/?p=2461 (although for something else it's already cited on our page), is an "expert witness" (I suppose) which claims these alternate names:

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." Willard Espy, the Valentino of word lovers, has built on this thought in Harvard Magazine to suggest proditio , Latin for "betrayal," or double-cross. Nice try; won't fly.

And also https://web.archive.org/web/20100723025356/http://www.nytimes.com/1991/03/24/magazine/on-language-hit-the-pound-sign.html?pagewanted=2

Privacy Journal surveyed its readers in 1986 for ideas and discovered the symbol is called bradgard , or "lumberyard" in Swedish, because it looks like the mark for a lumberyard on maps for hiking and other outdoor activities. Score was also recommended, perhaps by a card player. The journal concluded that thingamajig was too general, octothorpe was most authentic, and gridlet was the most intriguing.

There is also http://catb.org/jargon/html/A/ASCII.html with

Common: number sign; pound; pound sign; hash; sharp; crunch  ; hex; [mesh]. Rare: grid; cross�hatch; oc�to�thorpe; flash; <square>, pig-pen; tic�tac�toe; scratchmark; thud; thump; splat .

Dingolover6969 (talk) 14:28, 14 January 2026 (UTC)

double-struck N

User:Moyogo, thanks for your recent edits. The page now describes that german number sign as involving a "double-struck N". Is this supposed to be the same sense of "double-struck" that is Blackboard bold? Or rather just an N with two lines struck across it? If the former, I would like to wikilink that term, and if the latter I would like to call it something else. However, I am not familiar with German cursive so I'm not sure which it is. Dingolover6969 (talk) 18:26, 11 January 2026 (UTC)

The letter is not "double-struck" like blackboard bold, so maybe that’s not the most adequate description and "with two bars" may be clearer. At the same time Gothic script capitals often have extra strokes and the capital N is one of them, so the capital letter with two bars is not semantically distinct, the bars are only stylistic. But overall I’m wondering if the illustration is not misleading: unless there’s a reference pointing to a simplification from that Nº to #º, maybe the illustration is leading to believe such simplification took place (which may be the case but it’s not clear). We can remove the illustration for now unless there’s a reference making it more relevant, if that make sense. --Moyogo/ (talk) 19:13, 11 January 2026 (UTC)

Thanks! That all makes sense to me. Dingolover6969 (talk) 19:55, 11 January 2026 (UTC)

This from pound sign may explain the crossing lines:

According to the Royal Mint Museum:

It is not known for certain when the horizontal line or lines, which indicate an abbreviation, first came to be drawn through the L. However, there is in the Bank of England Museum a cheque dated 7 January 1661 with a clearly discernible £ sign. By the time the Bank was founded in 1694 the £ sign was in common use.

so it is an abbreviation of nummer? 𝕁𝕄𝔽 (talk) 21:05, 11 January 2026 (UTC)

Interesting theory! Dingolover6969 (talk) 12:55, 12 January 2026 (UTC)

References

Paucity of historical information about the #

Today I was reading https://www.britannica.com/topic/number-sign and https://languagelog.ldc.upenn.edu/nll/?p=2461 to see if they knew anything we didn't, and it struck me that this page is currently lacking in real historical information about the origins of #.

We also have no explanation for why this mark represents both numbers and weight.

We assert that it probably comes from ℔ based on some speculation that ultimately all comes from Keith Houston (we cite his book Shady Characters and also a New Yorker article he wrote on the same subjects) — which is fine speculation, but I don't think we have any real evidence for it. I don't find the Newton example convincing. (cf: https://web.archive.org/web/20100909043853/http://www.chemheritage.org/community/periodic-tabloid/2010-08-31-pounds-ampersands-and-skulls.aspx) We should have better attestation of this given how much written material there is from the time. In fact, the fact that there were nicely-printed ℔s in books represent the thing Newton wrote kind of makes me doubt that this mutated into #. Why would that happen? Well, the New Yorker article does call the form a "missing link", so I take it the author is aware of this problem.

Furthermore, Shady Characters is maybe relying for this claim entirely on Keith Gordon Irwin's book, the romance of writing, which we already cite, but which is just the same speculation. I have procured a copy of this book (1956 edition, it seems, in case that matters) and there seems to be no evidence in it, just speculation. Again, a fine speculation, but the only diagram of the character changing over time seems to be illustrative of the idea, not actual evidence. (maybe he'd seen a lot of evidence in person somehow, or is citing books that have) He also does not explain the number usage. Note that this book does not have the Newton example, so that's definitely a connection

We know that there's a real # by the time of the Remington Standard typewriter (c. 1886), but that's about as far back as we can go on our current evidence, unless the ducat mark (mentioned in the previous talk page section) isn't a coincidence.

Britannica and Language Log mention a telegraph-based theory (is Britannica just going off Language Log? well, they don't cite any sources), but it doesn't seem very well-supported to me and it wouldn't explain where # comes from, only why it's called "pound" sometimes maybe.

I find it difficult to find good sources on this, but my source-finding abilities have always been weak. Dingolover6969 (talk) 12:54, 12 January 2026 (UTC)

Irwin 1967’s statement is odd to me as it seems to say ℔ was an Italian abbreviation, while it would be more accurate to describe it as a Latin abbreviation widely used in various European languages. But the "hurried way of writing the abbreviation appear[ing] to have been responsible for the # sign" part checks out.

Italian arithmetic manual manuscript from 1428 (here ℔ is used as a currency in the libra solidus denarius system.).

Maybe the Newton’s sample ℔ isn’t that clear as it uses looped l. A few other variations of the sign do seem closer to #, sometimes in print:

1825 Karl Tauchnitz type specimen with various shapes of ℔ sign (on the right).

--Moyogo/ (talk) 13:17, 14 January 2026 (UTC)

Unfortunately, this is taking us unambiguously into WP:OR, we really need an RS. It does seem astonishing that a symbol so heavily used in the US does not seem to have any literature about it. (btw, I could be convinced that ℔ became # as in "Beans 2¢︍︌︋︊︉/#" but how did it become a number sign? To postulate convergent evolution from a crossed N (N) as well would really need a very solid source.)

A pity as this article is almost GA but this speculation is a show stopper. 𝕁𝕄𝔽 (talk) 13:48, 14 January 2026 (UTC)

Ooo, very cool!

Although, it's not really clear to me that the #-like symbol here in the 1825 Karl Tauchnitz type specimen is supposed to represent an lb. It's right next to an ℔, for sure, but it's also next to other currency signs... Dingolover6969 (talk) 13:53, 14 January 2026 (UTC)

Reading the section again, I think that, although we may be teetering on the edge, I don't believe that we have actually breached WP:SYNTH or WP:OR. We note the existence of examples of marks that are increasingly close to # from either end but we don't make any claim that this was how the mark evolved. I wonder, though, whether we should include an explicit statement up front to say that no reliable source has made a definitive declaration of its origin? 𝕁𝕄𝔽 (talk) 18:34, 17 January 2026 (UTC)

I think it would be fine to do something like that. One problem is that various RSs do provide declarations that seem definitive — they just disagree and have little support. For instance the New Yorker source states the lb theory but if you chase down his citation in the full form of his argument it's just the Romance Of Writing making a speculative call. There is also the source saying it's from a symbol for "village" that our other sources pooh-pooh. But we could still do something along these lines. Dingolover6969 (talk) 21:34, 18 January 2026 (UTC)

Surely that just means that the sources are not reliable. If so, can we cite them anywhere? We cite the New Yorker rather heavily... 𝕁𝕄𝔽 (talk) 11:30, 21 January 2026 (UTC)

Hmm, I dunno! I think it makes them reliable, but only on the level of "scholars speculate", whether or not they themselves use hedge words.

... Reading through WP:RS it seems like reliability applies to the source as a whole, not just one dodgy article/claim they might make. So, since the New Yorker probably has "a reputation for fact-checking and accuracy", I guess it's still a reliable source and citable. There may be other dimensions to this that I'm not considering. Dingolover6969 (talk) 00:28, 22 January 2026 (UTC)

"No worries, mate" . We can just conveniently overlook the single one that we have disproved. 𝕁𝕄𝔽 (talk) 20:45, 22 January 2026 (UTC)

@Dingolover6969: FWIW the arithmetic booklet manuscript LBC Cas A 1086#1 has several examples of ℔ looking like an ℔ sign with two strokes and often an open b. So it’s seems the 3 signs (2 previous to last one first line of symbols, 1 previous to last on third lines of symbols) in the 1825 Karl Tauchnitz type specimen are variants of the ℔ symbol. File:Andrew Jackson papers, 1775-1874, Jean Baptiste Plauché & Co. to Andrew Jackson, February 19, 1843, price of 2¢ per ℔ of cotton.jpg is another example of ℔ with two strokes. --Moyogo/ (talk) 08:15, 21 January 2026 (UTC)

Good one! I had to look at that image three times to stop seeing a # and start seeing an ℔ sign with two strokes. Unfortunately we can't make the call. 𝕁𝕄𝔽 (talk) 11:27, 21 January 2026 (UTC)

@JMF: The author of LBC Cas A 1086#1 writes out Pfund (pound) in the sentences and ℔ with an extra stroke in the equations. I’m not sure what you mean by "make the call". --Moyogo/ (talk) 19:43, 22 January 2026 (UTC)

Yes that is true but, as obvious as it may be, we are not permitted to declare that it evolved into a # mark (used to mean "pounds"). We can certainly show a thread of logic that would lead the reader to draw that conclusion but we can't be the ones to declare it to be true. We need an RS to have made that call. 𝕁𝕄𝔽 (talk) 20:21, 22 January 2026 (UTC)

First appearance on telephone keypad

Unfortunately Google no longer reveals the page from Houston. Can anyone confirm that it says 1968? because the first reference in a patent is 1975.

• US patent 3920926, George Victor Lenaerts & Eric Egils Auzins, "TELEPHONE DATA SET INCLUDING VISUAL DISPLAY MEANS", issued November 18, 1975, assigned to Northern Electric Company Ltd, Montreal, Canada  "The pad 1 provides keys for numerals 0 to 9, while the sextile or asterisk ( * ) key is decoded to provide a decimal point and the octothorp ( # ) key generates a command [etc]"

. 𝕁𝕄𝔽 (talk) 17:11, 15 January 2026 (UTC)

For that info ("in 1968, two unused buttons either side of the zero were finally made available for public use in controlling menus and other special services provided by the exchange"), Houston cites Sheldon Hochheiser, “Pressing Matters: Touch-Tone Phones Spark Debate”, Encore, first and second quarters, 1999, pp. 13–14 which is copied online: https://www.svpal.org/~dickel/octothorp/Encore_magazine.pdf --Moyogo/ (talk) 19:55, 15 January 2026 (UTC)

Thank you, best to have an answer for the begged question. (I liked the Neither number nor octothorp won the heated battle. Today most people call it the pound key. That was then, this is now and most people call it the hash key .) 𝕁𝕄𝔽 (talk) 23:55, 15 January 2026 (UTC)

More interesting old catalogue examples

Searching for something else, I came across this discussion which has some interesting examples of # in metal type catalogues from the 1920s (all "number (sign)"). Not clear if they have any encyclopedic value, but I leave the link here in case: https://typedrawers.com/discussion/3900/the-number-sign-now-and-then-across-languages/p2 Dingolover6969 (talk) 07:13, 30 January 2026 (UTC)