You have sung the same chorus for ten years. In the car. In the shower. At a concert, with the confidence of someone who could replace the lead singer at short notice. Then a friend asks one devastating question: “What did you just say?”
You check the lyrics. Your favourite line does not exist. The strange little story you thought the song was telling has apparently been your own contribution all along.
Why do we mishear song lyrics? Listening involves interpreting sounds using what we already know about language. When a vocal phrase is ambiguous, more than one interpretation may seem to fit. Expectations can help one of those versions win. Crucially, the alternative still needs to sound sufficiently close to the actual words.
There is a word for this: mondegreen. It describes a phrase created by mishearing something spoken or sung. The name comes from writer Sylvia Wright, who remembered misunderstanding a Scottish ballad. She heard “laid him on the green” as “Lady Mondegreen”, accidentally adding a character to the story. The term dates to 1954, long before lyric videos and comment sections gave these mistakes a worldwide audience.
A famous Jimi Hendrix example shows how little needs to change. “Kiss the sky” becomes “kiss this guy”. A few similar sounds produce a very different scene. The joke works because both versions can feel recognisable in the same stretch of singing.
Consider the familiar wordplay “ice cream” and “I scream”. On a page, the spaces tell you how to divide the words. In a flowing vocal line, deciding where one word ends can be trickier. As linguist Andrew Nevins explains in his discussion of misheard speech, expectations influence these mistakes within the limits set by the sound itself. That helps explain why an alternative lyric can be both ridiculous in meaning and surprisingly convincing to the ear.
Singing adds another complication: the words have musical work to do. A vowel may last several beats. A syllable may land somewhere unexpected in the phrase. The performance changes the timing and shape of language, which can make words harder to identify.
In a 2008 study comparing spoken and sung phrases, researchers Lauren Collister and David Huron tested 20 listeners using recordings by three trained vocalists. Listeners misheard roughly a quarter of the target words in the unaccompanied singing, making more errors than with the spoken versions. That figure belongs to this particular experiment; it does not mean we misunderstand a quarter of every song. It does show that words can become harder to recognise through singing even without a loud band competing for attention.
This makes the familiar accusation that a singer is “just mumbling” feel a little too easy. Listen to a passage you find unclear and ask what you do hear: a long note, a rhythmic pattern, a particular tone of voice. Those may be the very things you love about it. Perfectly clear diction is one possible musical priority among several.
The recording can stay exactly the same while the words you hear seem to change.
A 2014 study published in PLOS ONE explored that effect with 23 German-speaking participants. Researchers played song excerpts and showed listeners written alternatives to the lyrics. In a later listening round without those written cues, participants reported hearing some of the suggested versions. The effect varied considerably between people and excerpts.
Knowing the song did not reliably protect listeners from the effect in that experiment. Previous knowledge of the alternative lyric, however, was associated with more frequent misperceptions. This offers a possible explanation for the frustration of knowing the correct words while still hearing the joke version. Learning what the singer actually said and changing how an ambiguous passage sounds to you can be different experiences.
The study does not establish that a wrong lyric becomes permanent. It does show why a friend's suggestion can do more than make you laugh: it can influence what you subsequently report hearing.
The story gets even stranger when two languages are involved. A line sung in one language can resemble words in another, producing the kind of cross-language mishearing often called soramimi. The result can sound uncannily meaningful to a listener who recognises the suggested phrase.
A 2022 study of English–Japanese misheard lyrics analysed 60 examples selected from the Japanese television programme Soramimi Hour. Researchers found recurring patterns in how English sounds were reinterpreted through Japanese, including changes in where listeners placed word boundaries. These were curated entertainment examples, so they cannot tell us how often the general population mishears lyrics. They do illustrate that the apparent nonsense can follow linguistic patterns. The languages we know help shape the alternatives available to us.
If you want to settle an argument about a lyric, start with the artist's published lyrics, an album booklet or an official lyric video. Check that you are comparing the same version of the song. A live performance and a studio recording need not contain identical words.
For a more entertaining experiment, play the disputed line to a friend before revealing either interpretation. Ask them to write down what they hear, then compare answers. This is an informal listening game, of course. The interesting part is seeing whether your apparently obvious sentence was obvious to anybody else.
There is also something affectionate about keeping a private, incorrect version of a favourite song. You might choose to sing the right words at the next gig and still remember the old version every time. It belongs to the years when you knew the melody by heart and had never thought to check the rest.
Which lyric did you sing with complete confidence before discovering you had invented it?