On 23 August 2007, a young web developer named Chris Messina typed a short question into a new service that most people had never heard of.
“how do you feel about using # (pound) for groups. As in #barcamp [msg]?”
It was a modest proposal. Twitter was barely a year old, and it had no way of grouping messages by subject. If you wanted to follow a conversation about a conference, a news event or a hobby, you had to hope the right people happened to be in your feed. Messina suggested a workaround: put a pound sign in front of a word, and anyone could search for it.
According to Messina’s own later account, Twitter wasn’t impressed. In a 2013 interview with The Wall Street Journal, he recalled being told that these things were “for nerds” and would never catch on.
Within a few years, the symbol was printed on billboards, shouted in late-night comedy sketches, added to the Oxford English Dictionary and given an official French translation. It named political movements, and turned TV shows, protests and product launches into searchable public conversations.
Then, almost as quietly as it arrived, the hashtag began to retreat.
This is the story of a symbol older than the telephone, adopted by a company that didn’t want it, and now being politely shown the door by the platforms it helped build.
Eight loose ends
Long before it organised conversations, the # had a much more physical job: it measured weight.
The most widely accepted explanation traces the symbol to the Latin libra pondo, “a pound by weight”. Scribes abbreviated it as “lb” and drew a horizontal stroke across the letters to mark the abbreviation. Over centuries of hurried handwriting, that crossed “lb” collapsed into something simpler: two horizontal lines over two slanted ones.
That is why, in the United States, many people still call it the “pound sign”. In the UK and Australia it’s usually a “hash”, a name probably related to cross-hatching. Officially, it has a much duller name: the Unicode Consortium lists the character as NUMBER SIGN.
The symbol’s real break came in 1968, when Bell Telephone Laboratories put it in the bottom-right corner of the touch-tone telephone keypad. Engineers had added two extra keys, the asterisk and the #, and needed names for them. One of those names became a small legend of its own: the octothorpe.
Nobody agrees where it came from. “Octo” clearly refers to the eight points where the four strokes end. The “thorpe” is the mystery. One account credits a Bell engineer named Don Macpherson, who supposedly chose it to honour the Olympian Jim Thorpe. Another names two different engineers who coined it as a joke. Others link it to the founder of the colony of Georgia, or to an Old Norse word for village. Lexicographers have found none of these stories backed by contemporary evidence, and the earliest known printed use of the word dates only to 1974. The most honest answer is that “octothorpe” was probably a piece of engineering humour that escaped the lab, and that every origin story is, at best, plausible.
For years the key itself saw little use. It only became truly familiar in the early 1980s, when large voicemail systems began asking callers to “press pound”.
Borrowed from the chat rooms
By the time Messina sent his tweet, the # already had a life online.
Around 1988, Internet Relay Chat, one of the earliest real-time chat systems, used the symbol to name channels that spanned an entire network. A room called #linux was a place where people talked about Linux. For a generation of programmers and early internet users, a word starting with # already meant a shared space organised around a topic.
Messina’s idea was to carry that habit into Twitter, which had no rooms at all. Twitter was a stream: everything everyone you followed said, in reverse chronological order. A tag wouldn’t create a separate room. It would simply let people find related messages in the stream, without Twitter having to build anything.
Two days after his tweet, Messina explained on his blog that he wanted “a better eavesdropping experience” on Twitter. On 26 August 2007, the writer Stowe Boyd published a post titled “Hash Tags = Twitter Groupings”. According to the lexicographer Ben Zimmer, it is the first published use of the term that would eventually become a single word.
So within three days, the symbol had a proposal, a purpose and a name. What it didn’t have was a reason for ordinary people to use it.
The fire that made it stick
The reason arrived in October 2007, in the form of wildfires.
As fires spread across San Diego County, people on Twitter began sharing updates about evacuations, road closures and the movement of the flames. Messina suggested that they tag their updates #sandiegofire. Many did. For the first time, a hashtag wasn’t a geeky convention among conference-goers. It was a practical tool for finding urgent information from strangers during an emergency.
This is the pattern that would repeat for the rest of the hashtag’s life. It spread fastest when events were happening right now, when information was scattered across thousands of people who didn’t follow each other, and when a shared tag was the quickest way to gather it in one place.
It also showed something about Twitter’s early culture. Many of the platform’s most recognisable habits were invented by users before the company built them in. Users addressed each other with an @ sign before Twitter treated it as a feature. They copied each other’s posts with “RT”, for retweet, before Twitter added an official retweet button in late 2009. The hashtag belonged to this tradition of improvised conventions, and the company eventually accepted all of them.
Twitter gives in
On 2 July 2009, almost two years after Messina’s proposal, Twitter began turning hashtags into links. Clicking one now led straight to search results for the tag. The convention had become infrastructure.
Twitter later added Trending Topics to its front page, a list of the subjects and tags surging in use. That created a new incentive: getting a hashtag to trend became a goal in itself, for fans, activists, television networks and advertisers.
The effect on how events played out was striking. Television shows began displaying hashtags on screen during broadcasts, inviting viewers to talk about the episode in real time. Conferences printed their tags on badges. Brands launched campaigns around them, sometimes with spectacular backfires when the public used the tag to mock the brand instead. Messina’s simple search trick had turned into a way to measure attention as it happened.
A word of the year
The language caught up quickly.
In January 2013, the American Dialect Society named “hashtag” its word of the year for 2012. That same month, France’s official terminology body published a French replacement, mot-dièse, using the musical sharp sign as its image, even though the sharp sign (♯) and the number sign (#) are, strictly speaking, different symbols. In June 2014, the Oxford English Dictionary added “hashtag”, defining it as a word or phrase preceded by the # symbol and used on social media to identify messages on a specific topic.
Pop culture made fun of it too. In September 2013, Jimmy Fallon and Justin Timberlake performed a sketch on Late Night in which they talked to each other entirely in spoken hashtags, saying “hashtag” out loud before each phrase and making finger-gestures in the shape of the symbol. The joke only worked because everyone in the audience already understood the convention.
By then the symbol had spread far beyond Twitter. Instagram users tagged photos with dozens of hashtags to help strangers discover them. Facebook added support in 2013. The # that once told a phone exchange “I’ve finished dialling” now told every search box “this is what I’m about”.
From tag to movement
The hashtag’s most consequential use came when people stopped treating it as a search tool and started treating it as a banner.
In 2013, after the acquittal of George Zimmerman in the killing of Trayvon Martin, three organisers, Alicia Garza, Patrisse Cullors and Opal Tometi, began using the phrase Black Lives Matter. As a hashtag, it grew into one of the largest social movements of the decade. In October 2017, the actor Alyssa Milano invited people who had experienced sexual harassment or assault to reply “me too”. The phrase had been used for years by the activist Tarana Burke, and as #MeToo it spread across platforms and countries within days.
In these cases, the hashtag did something no earlier technology had done so cheaply. It let millions of strangers attach their individual posts to a shared name, visible to anyone, without any organisation approving them. It created a public archive in real time. And because anyone could use it, it could also be used against itself: critics, opponents and spammers could flood the same tag with their own messages.
That double edge became part of the hashtag’s identity. A tag could gather a movement, but it couldn’t control it.
The slow retreat
The hashtag was built for a particular kind of internet: chronological feeds, keyword search and users who had to find content themselves. That internet has largely been replaced by recommendation systems, which decide what to show each user based on behaviour, interests and the content itself, whether or not it carries a tag.
The platforms have noticed, and each is retreating in its own way.
When Meta’s Threads introduced its version of hashtags in December 2023, it called them topic tags and changed almost every rule. Users don’t type the # symbol. Tags can contain several words and spaces. And each post can carry only one tag. Adam Mosseri, the head of Instagram, presented them as a way to connect with people interested in the same subjects. In practice, the design ruled out the long strings of hashtags that had cluttered so many posts.
On 17 December 2024, Elon Musk, X’s owner, went further. “Please stop using hashtags,” he posted. “The system doesn’t need them anymore and they look ugly.” The platform where the hashtag had first become a convention was now telling users that its recommendation systems could work out what a post was about on their own.
A year later, in December 2025, Instagram announced that it would gradually limit captions on posts and reels to five hashtags. It said a few targeted hashtags would serve content better than a long list of generic ones, and that the cap would also reduce misuse by spammers and scammers. On the platform where people had once stuffed captions with dozens of tags in the hope of being discovered, the hashtag was being turned into a short label.
None of this means the symbol has disappeared. People still use hashtags to name events, campaigns, communities and jokes, and many platforms still let users search and follow them. What has changed is its role. It began as the main way to find conversations on a platform that couldn’t do it for you. It is becoming an optional signal in systems that decide for you.
Press pound to continue
There is something fitting about where the symbol has ended up.
On a telephone keypad, the # never carried a message. It was a signal to the machine: I’ve finished entering my number, now do something. On Twitter in 2007, it became a signal to other people: this is what I’m talking about, come and find it. For a decade, that human signal shaped how news broke, how fans watched television and how movements were named.
Now the machines have learned to read our posts directly, and they don’t need us to point at the topic. The # is drifting back towards something like its old role, a small instruction in a system most people never see.
But every time someone types that little grid of four lines before a word, they are repeating a 2007 suggestion from a developer who was told it was “for nerds”. The nerds, it turns out, were right for about fifteen years. That is a remarkably long time on the internet.
