The word arrives late in the book. Richard Dawkins, the esteemed evolutionary biologist, spends most of The Selfish Gene, published by Oxford University Press in 1976, making his case: that the useful unit for thinking about evolution is not the organism or the species but the gene, the thing that gets copied. Only in the final chapter, almost as a coda, does he ask whether the same logic might apply to something other than DNA. He decides that it might, and he needs a name for it.
That chapter is titled “Memes: the new replicators.” Dawkins takes the Greek root mimeme, meaning roughly that which is imitated, and files it down to a single syllable that rhymes with gene. He notes in passing that it can also call to mind “memory,” or the French même. His examples are ordinary: tunes, ideas, catch-phrases, clothes fashions, ways of making pots or of building arches. A meme, in his account, is a unit of cultural transmission, something that passes from one mind to another by imitation.
His framing is deliberately large for an idea introduced on the way out the door. A new kind of replicator, he writes, has recently appeared on this planet, still clumsy and drifting about in the primeval soup of human culture. The claim is structural rather than mystical. Darwinian selection does not care what its unit is made of. Give it variation, copying, and differential survival, and it will run on genes, or on ideas, or on anything else that reproduces imperfectly.
He even reached, in that same chapter, for the metaphor that would later look prophetic: ideas spreading the way a virus spreads.
What the word actually claimed
Precision matters here, because most later use of the word throws the distinction away. Dawkins was not describing pictures of cats. He was proposing that culture evolves through a process analogous to biological evolution, with ideas competing for the limited resource of human attention and memory. The ones that get copied accurately and often persist. The ones that do not, fade. Longevity, fecundity, and copying-fidelity were his three properties of a successful replicator, borrowed straight from the gene chapter and pointed at the contents of people’s heads.
This was a metaphor doing analytical work, and Dawkins was careful at first about how far to push it. In later years he grew more comfortable with the stronger version, endorsing the psychologist Susan Blackmore’s attempt to build a full theory out of it in The Meme Machine (1999), for which he wrote the foreword. Others were less convinced. The field that grew up around the idea, memetics, struggled to produce measurable, testable results of its own, and its main outlet, the Journal of Memetics, stopped publishing in 2005. One of its own contributors, Bruce Edmonds, marked the exit with a paper titled “The revealed poverty of the gene-meme analogy”, arguing the project had failed to deliver substantive results. Whether a meme is a real unit or merely a convenient way of talking has never been fully settled.
What the internet did to it
For its first two decades the word kept fairly specialist company. The bridge to its current life was built earlier than most people assume. In a 1994 essay for Wired, “Meme, Counter-meme,” the lawyer Mike Godwin, better known for Godwin’s law, used the term for the way ideas spread and mutate across Usenet, one of the first applications of “meme” to online circulation. What most people mean by the word now arrived later. Across the mid-to-late 2000s the format-driven version took over: an image, a phrase, or a template that spreads online, gets copied, gets altered, and spreads again. LOLcats. Advice animals. Reaction images with the caption swapped out. By the time the word was common currency, this was what it meant, and the 1976 book had become a footnote to the thing it named.
On the surface, this looks like vindication. Dawkins had described ideas replicating and mutating as they move from brain to brain, and here was a technology built for exactly that, running the process faster than print ever could. He noticed the fit himself. Speaking to Wired in 2013, as documented in Science Friday’s account of the word’s origin, he said the internet meaning was not far from his own, that a meme is more or less anything that goes viral, and that online usage had taken a subset of what he meant and treated it as the whole.
But he also drew a line.
Where the analogy breaks
In a 2013 talk staged for advertising executives at the Cannes Lions festival, a piece of theatre later released as a short film called “Just for Hits”, Dawkins made the distinction plainly. The internet meme, he argued, is a hijacking of the original idea. His memes mutated by accident, through the small errors that creep in whenever something is copied imperfectly, and then survived or died by a blind selection process. Internet memes work differently. They are altered on purpose. Someone opens an image editor and changes the caption, fully aware of what they are doing. The mutations are designed, not random.
That difference is not cosmetic. Dawkins’s engine runs on undirected variation, the same feature that lets biological evolution work without a designer. Strip that out and replace it with deliberate human choice, and you have something that resembles his process from a distance while operating on a different principle underneath. The internet did make the metaphor literal, in the sense that ideas now visibly reproduce, compete, and spread. It did so by removing the exact mechanism that made the metaphor Darwinian in the first place.
The concept that got away
A word coined to describe how ideas escape their origins and take on lives of their own went and did precisely that to its author. That symmetry is tidy enough to be a little suspicious. Dawkins can stand on a stage and call the internet version a hijacking, but coining a word carries no right to govern it afterwards. Meaning is settled by use, not by the person who struck the coin. The meme concept became, by its own definition, a meme: copied, mutated, and now understood in a way its originator did not pick.
Scholars did not wait for permission. The media researcher Limor Shifman, in her 2013 paper “Memes in a Digital World” and the 2014 book Memes in Digital Culture, treated the online sense as a serious object of study rather than a corruption, and drew a distinction Dawkins had not. A viral hit and a meme are not the same thing. A viral video spreads as a single unit, copied and passed along intact. A meme gets remade, with people rebuilding the template as they go, which is closer to what the word was already doing on Usenet in 1994 than to a clip that merely travels.
This is the part most retellings skip. The popular version holds that Dawkins predicted internet culture, which is not quite right. He described a general mechanism for cultural transmission, and a technology arrived that fit the description while quietly changing how the mechanism ran. The word survived the transition better than the theory did. Most people who use “meme” fluently every day have never read the chapter, could not say what a replicator is, and would be mildly surprised to learn the term came out of a book about genes.
Dawkins appears to find the whole thing both funny and a little irritating, which is a fair way to feel about it. The word he made has travelled further than almost anything else he ever wrote. The argument it was built to carry mostly stayed behind on the page.
