Shady Characters

I’m almost ashamed to say it, but I never really gave the backslash a second thought.

The backslash’s forward-leaning counterpart is everywhere, especially in computing. It lives in network and web addresses such as https://shadycharacters.co.uk; in file paths, such as /home/keith; and it introduces human-readable “comments” in any number of programming languages, often like /* this */ or // this.

But the backslash? It’s a rarer species. Windows users might recognise it from the command prompt, where it occurs in file paths such as C:\Users\Keith (although Windows does also recognise Unix-style forward slashes). The backslash is also sometimes used to “escape”, or neutralise, characters that would otherwise have some special meaning in a given context. For example, if an ampersand has some special meaning in a particular type of file, prefixing it with a backslash (\&) will often cause it to be treated as a normal ampersand instead. (Ironically, in writing that last sentence I had to escape the backslash character itself, \\, since it has a special meaning in the software that runs shadycharacters.co.uk.)

It was with interest, then, that I read a post on Mastodon* from a user named Modulux on the subject of the backslash:

Friend of mine was commenting that the origins of the backslash character are unclear. It was included in the teletype character set and IBM put it on ASCII for that reason but it is not known what it was for. Searching around I found some unreliable info that it was used in typography from the 16th century in order to represent line or paragraph breaks. Can anyone confirm or disconfirm, preferably with source?

To set one thing straight, the backslash wasn’t historically used for line or paragraph breaks. Instead, that was one of the functions of the forward slash, which, once upon a time, had an important role in punctuating texts: in the medieval and early modern periods, a forward slash could be used to indicate a pause of a lesser or greater length, anywhere between a comma and a paragraph depending on its user. It was often called a virgula or virgule, and it gave rise to the modern comma. It is covered along with many other such marks in the Shady Characters book.

Historical wobbles aside, on the question of where the backslash, or ‘\’, had come from, a lively discussion ensued. Stewart Russell responded:

According to Mackenzie (“Coded Character Sets, History and Development”) it seems to have appeared in the IBM Stretch design proposal, sometime in the mid 1950s (even though the first Stretch machine was delivered to Los Alamos in 1961) – see the chapter “Early Codes”, around page 80. The first name given to it was “Reverse Divide”.

It was also used in the text representation of ALGOL, with /\ for “∧” (AND) and \/ for “∨” (OR).

I had picked up a copy of Charles Mackenzie’s Coded-character Sets: History and Development during the writing of Shady Characters (never say writing isn’t a glamorous occupation), so I looked up Stewart’s reference. And indeed, the modern backslash does seem to have appeared first in IBM’s so-called Stretch character set,1 as used in an eponymous computer delivered to Los Alamos National Laboratory† in 1961.2

I talk more about character sets in my upcoming book, Face with Tears of Joy, but the long and short of it is that character sets are lexicons of individual characters known to one computer or another. To communicate successfully, any two computers must agree on which character set to use. Nowadays, that is a very simple negotiation, since effectively all computers use the all-encompassing Unicode character set. Back in the ’50s, however, things were different — electronic computers were still new, and many of them boasted their own specialised character sets. IBM’s Stretch computer was a case in point.

Why, though, had IBM chosen to incorporate a backslash in this new character set? Even before the Stretch computer had been delivered to Los Alamos, an influential IBMer named Bob Bemer3 had written to the Association of Computing Machinery, the industry body for America’s computer manufacturers and computer scientists, to make the case for developing a standard character set.4 His proposal kickstarted a drive towards just such a standard, and eventually resulted in the creation of ASCII — the American Standard Code for Information Interchange, a forerunner of Unicode, and whose name you may well have heard before.

Yet Bemer also had his day job at IBM, and one of his tasks there was to figure out which characters should be available on the as-yet unfinished Stretch computer.5 Enter the backslash — or rather, enter the “reverse divide” character, which is what Bemer had called it in his 1959 paper for the ACM.4

Before his death in 2004, Bemer wrote an short article entitled “How ASCII Got its Backslash” on, well, how ASCII got its backslash, and it does not have a lot to do with division, reverse or otherwise. Bemer recalls that he wanted to be able to write certain logical expressions in a programming language called Algol.6 At the time, Algol was still very new, though it had already been anointed by the ACM as its standard language for computer algorithms. Among other things, Algol allowed programmers to use binary logic, in which values could be combined and manipulated using operations called AND, OR and NOT. The first of those two operations were typically written using A- and V-shaped characters (‘∧’ and ‘∨’ respectively), and Algol adopted that same convention.7

The problem, as Bemer saw it, was that extant character sets had no way to render those crucial Boolean operations. Being all too aware of the limited memory available to early computers, Bemer realised that with the addition of a single new character — ‘\’ — he could write Algol’s AND and OR operators by combining it with the ‘/’ already present in many other character sets: /\, \/.6 Thus the backslash appeared, first for use by programs running on the Stretch computer and later in ASCII itself.

As to the name of the ‘\’; well, it’s all a little unsatisfying. In his earliest published writing on the subject, that 1959 article for the ACM, Bemer calls the backslash the “reverse divide”.4 His reasoning is obvious enough — ‘\’ is the mirror image of ‘/’, which is commonly used to indicate mathematical division — although he never spelled it out as such. “Reverse slant” is another early favourite, again for fairly obvious reasons, cropping up in a contemporary account of the IBM Stretch project8 and going on to eclipse “reverse divide” in popularity. The existence of two different names would suggest that ‘\’ had no commonly-held name at the time, and that a new one had had to be invented.

Yet neither name would last very long. Soon after Bemer minted the ‘\’, some anonymous programmer decided that “reverse slant” and “reverse divide” were too wordy, or too obscure, and came up with a pithier name for it. Born in the ’50s as the “reverse divide”, in the ’60s the ‘\’ was reinvented as the “backslash”‡ (who wasn’t reinventing themselves in that decade?), and the name has stuck ever since.

Except, except! That is not the whole story. The backslash, it turn out, is a few decades older that it seems — or rather, it has a doppelganger from an earlier technological era. There’s an intriguing story behind that one too, but it will have to wait for another time!

1.

MacKenzie, Charles E. “Early Codes : The Stretch Code”. In Coded Character Sets: History and Development, 67-75. Addison-Wesley Pub (Sd), 1980.

 ↢

2.

CHM Revolution. “IBM Stretches Its Capabilities”. Accessed March 13, 2025.

 ↢

3.

Hardison, Erica. “Robert W. Bemer”. IEEE Computer Society.

 ↢

4.

Bemer, R. W. “A Proposal for a Generalized Card Code for 256 Characters”. Commun. ACM 2, no. 9 (September 1, 1959): 19–23. https://doi.org/10.1145/368424.368435.

 ↢

5.

Bemer, Bob. A Story of ASCII.

 ↢

6.

Bemer, Bob. “How ASCII Got Its Backslash”.

 ↢

7.

Perlis, A. J., and K. Samelson. “Report on the Algorithmic Language ALGOL the ACM Committee on Programming Languages and the GAMM Committee on Programming”. Numerische Mathematik 1, no. 1 (December 1959): 41-60. https://doi.org/10.1007/BF01386372.

 ↢

8.

Ballance, Richard S., and Werner Buchholz. “Chapter 6: Character Set”. In Planning a computer system; Project Stretch. New York: McGraw-Hill, 1962.

 ↢

*

This is a convenient simplification: Modulux posted on a personal server compatible with Mastodon but not actually running Mastodon itself — so although I first read Modulux’s post on Mastodon, it did not originate there. This is the strength of the so-called fediverse, where anyone can run a social network and connect it, or “federate”, to other networks. It’s a bit like email — you can use one of the big providers such as Gmail or Apple iCloud, or you can run your own personal mail server. It is a brave new world, and a welcome one, too. Shady Characters also lives there, at mastodon.social/@shadychars. ↢

†

Nuclear energy once again! This is turning into a theme. ↢

‡

Google Ngrams shows 1967 as the inflection point. ↢