Where did the @ symbol come from?

Computer programmer Ray Tomlinson died this week at the age of 74.  He definitely left his mark.

In 1971, Tomlinson invented email. As if that isn’t enough, he also first used @ – the at sign – to separate the username from the domain in the first electronic mail, now the standard symbol around the world.

The at sign, also known as the commercial at, has many other colorful names across the globe, including words for monkey, snail, and puppy in a number of languages.

Thanks to Tomlinson, @ now prevails in digital correspondence, but the symbol previously served commerce: accountants once used @ to abbreviate “at a rate of.” For instance, 10 pencils @ $1: 10 pencils at a rate of $1, or $10.

Tomlinson invented email as a way to send messages from one computer to another, a problem the US government recruited him to solve. To do so, the user and the host names needed some sort of punctuation mark to separate them. Tomlinson has since explained he chose @ because not only was it available, as it was not widely used, but also because it handily communicated a sense of location.

Hicks at Utah, or HICKS@UTAH, is an early example of @’s usage, as the Oxford English Dictionary (OED) records. The OED first dates the email @ to 1972, while the commercial @ is cited yet earlier in 1969.

But as a symbol, @ is much older.

The OED notes the “earliest evidence so far found for the symbol is in 16th-cent. European mercantile records.” The dictionary points to its usage as an Italian unit of measurement, called the anfora, as well as a Spanish and Portuguese one, the arroba.

Anfora, or amphora in English, measured about 9 gallons for the Greeks, apparently, and 6 gallons, 7 pints for the Romans, the OED explains. The two-handled vessel, the amphora, is the inspiration for the name and the unit of measurement. As used in this sense, amphora is attested in English by 1607. The Spanish arroba actually derives from Arabic – al-rub, a weight one “quarter” of a Spanish quintal – and thus typically measured about 25 Spanish pounds. Arroba, the OED tells us, is recored by 1598.

Like so many unread emails, these amphorae piled up in Pula, Croatia. Image from freeimages.com/Michalina Piotrowksi

Several theories attempt to explain @’s distinctive spiral. Most converge on medieval manuscript shorthand for high-frequency words, like at, a significant efficiency when we consider the labor and expense required to produce and copy manuscripts. Some say @ wraps the e in each around the a in at; others, the curve abbreviates the t in at’s a. Yet others look to Latin’s ad (“to” or “at), with the symbol’s loop preserving an earlier way of writing lowercase d. In French, this at would be à, so @ saves writers from lifting their pens due to the accent mark, as it apparently can today.

If @ is first evidenced in medieval manuscripts, I, for one, would look to the Romance languages for the origin of this symbol and abbreviation.

Wherever @ comes from, one thing’s for sure: thanks to his technological (and typographical) genius, Tomlinson has made sure this once obscure and obsolescent symbol won’t be going anywhere anytime soon.

m ∫ r ∫


So often, we don’t hear about protocols until they’ve been breached, as in recent efforts to control Ebola in the United States. It turns out, though, that protocols really may be the “glue” that holds it all together.

"Protocol." Doodle by @andrescalo.
“Protocol.” Doodle by @andrescalo.


If we follow protocol back to its roots, we get a Greek expression for “first glued on”: protokollon, or πρωτόκολλον, a manuscript technology that was all about authorship and authenticity.

With MS. meaning “manuscript” and MSS., its plural, Walter Skeat explains:

[Protocol] meant, in Byzantine authors, orig. the first leaf glued on to MSS., in order to register under whose administration, and by whom the MS. was written; it was afterwards particularly applied to documents drawn up by notaries, because, by a decree of Justinian, such documents were always to be accompanied by such a first leaf or fly-leaf. It meant ‘first glued-on,’ i.e., glued on at the beginning.

The Oxford English Dictionary (OED) is helpful for its clarity here: The “first leaf of a papyrus-roll, bearing the official authentication and date of manufacture of the papyrus.” If you’ve ever flipped through the front matter of a book to assemble your “Works Cited,” you’ll see that the tradition continues.

The word was rolled out into Latin, then French, where it took on senses of a “draft” of a document, a proto-table of contents, or, like docket, the “minutes” of a meeting. Specialized meanings–written formulas used in drawing up church documents, special provisions and etiquette in diplomatic proceedings, records of medical and scientific procedures–eventually propelled it as the “accepted or established code of behaviour” (OED) by 1915.

If we unglue protocol, we find two elements:

  1. The Greek protos (combining form, proto-, or πρωτο-), meaning “first” or “foremost.” We’ve seen the word before in protein. At root is the prodigiously prolific Proto-Indo-European *per-, with a base meaning of “forward” or “through.”
  2. The Greek kolla (κόλλα)“glue,” as you may suspect. The word lives on in science as colloid and collagen–and in art with collage, French for “pasting” or “gluing.”

Speaking of which, though your teacher will tell you not to eat your paste or glue, you kind of already are. Glue is from the Latin gluten, in wheat, rye, and barley products, though we might all now know it best by its absence. And paste originally referred to kneaded, moistened flour (Oxford Dictionary of English Etymology)–hence its close relationship to words like paste and pastry.

As for the deeper origins of kolla, I have found nothing certain. I wonder if it may be a regional term for tree tar or plant gum, which were used in some of the earliest glue technologies. But whatever the origin of kolla, “glue” turns out to be instructive: Some meanings of words stick while others fall off, making the history of a word quite the collage.

m ∫ r ∫