Why music helps you remember vocabulary
A song holds the same word in several places at once, then replays it without being asked. That is the short answer, and the research behind it is more specific than most people expect.
Music helps you remember vocabulary because a sung phrase carries a melody, a rhythm, a stress pattern and a fixed line length along with the word itself. Every one of those is an extra retrieval cue, so when the meaning alone will not surface, the tune can still pull the word back. Three findings support this. Text is recalled better when it is sung than when it is spoken, but only when the melody repeats (Wallace, 1994). Adults who learned Hungarian phrases by singing them scored higher than adults who learned the same phrases by speaking (Ludke, Ferreira and Overy, 2014). And songs lodge themselves in the head unprompted, producing rehearsal nobody scheduled (Jakubowski and colleagues, 2017). Music does not replace review. It makes each word denser in memory, and it smuggles in repetitions you never planned.
Four things are going on: a second, non-verbal code attached to the word; a melodic scaffold that chunks the line; the act of producing the word by singing it; and involuntary replay afterwards. The first three make the memory stronger. The fourth makes it more frequent.
A song encodes the word twice
Allan Paivio's dual coding theory describes two cooperating memory systems, one verbal and one non-verbal, and argues that material represented in both is retrieved more reliably than material represented in only one, because either code can trigger recall (Paivio, 1991).
A flashcard gives you one route to the word: the word itself. The same word inside a sung line gives you the word plus a melodic contour, a rhythmic slot and a position in a phrase you have heard before. Paivio's non-verbal code was studied mostly through visual imagery rather than melody, so this is a reasonable extension rather than a direct result. The useful part is that it makes a testable prediction, and someone tested it.
The melody has to repeat
Wanda Wallace ran the experiment that matters most here, and the result has a catch in it that a lot of popular writing skips (Wallace, 1994).
When people heard three verses of a text sung to the same melody, they recalled the text better than people who heard it spoken. But when they heard only a single sung verse, or when each verse came with a different melody, the spoken version won. A tune that keeps changing is not a memory aid. It is extra material to learn.
So melody works as a scaffold, and a scaffold only helps once it is standing. The music chunks words into phrases, fixes line lengths, marks stress and tells you roughly how many syllables belong in a slot. All of that constrains what you could possibly be trying to remember, which is exactly why the chorus is the part of a song you can still sing years later, and exactly where your hardest vocabulary should live.
Singing beat speaking, under controlled conditions
The clearest experimental evidence for using songs in language learning comes from Karen Ludke, Fernanda Ferreira and Katie Overy (Ludke, Ferreira and Overy, 2014). Sixty adults were randomly assigned to one of three listen-and-repeat conditions for learning phrases in Hungarian, a language none of them knew: speaking, rhythmic speaking, or singing.
After a fifteen minute learning period, the singing group performed better overall, and the difference was statistically significant on the two tests that required participants to recall and produce spoken Hungarian phrases. The effect was not explained by age, gender, mood, phonological working memory ability, or musical ability and training.
Two details make this stronger than it first looks. Rhythmic speaking was its own condition, so the advantage cannot be written off as rhythm alone. And the tests that separated the groups were production tests, not recognition tests, which is the harder direction and the one most learners are weakest at.
The honest limits: sixty people, one language, one session, short-term paired-associate phrase learning. This is a real result, not a law of nature.
An earworm is review you did not schedule
Everything above is about making one exposure count for more. The other half of the story is that songs come back on their own.
Kelly Jakubowski and colleagues studied tunes named as involuntary musical imagery by 3,000 survey participants, then matched 100 frequently named earworm tunes against 100 tunes nobody named, controlling for popularity and style, and compared them across 83 melodic features (Jakubowski, Finkel, Stewart and Müllensiefen, 2017). Earworm tunes had more common global melodic contours, less common average gradients between melodic turning points, and faster average tempi. In other words, stickiness is partly a property of the melody itself, not only of how often you have heard it.
A later review of 47 studies across 33 peer reviewed articles concluded that involuntary musical imagery is a universal psychological phenomenon rather than an oddity (Liikkanen and Jakubowski, 2020).
For a language learner this is the quietly valuable part. When a line loops in your head on the bus, the line contains the word. That is rehearsal you did not plan, did not schedule and did not have to feel motivated for. No flashcard app can offer that.
What music does not do
Music strengthens a memory and it repeats a memory. It does not schedule one, and scheduling turns out to be where long-term retention is actually decided.
The largest synthesis on this covers 839 assessments of distributed practice across 317 experiments in 184 articles (Cepeda, Pashler, Vul, Wixted and Rohrer, 2006). Its central finding is that the gap between study sessions and the length of time you need to remember something work together: the interval that produced maximum retention grew as the retention interval grew. If you want a word for a year, reviewing it three times tonight is close to worthless.
A song replays at random. A review schedule replays on purpose. They solve different problems, and the sensible move is to run both.
One more limit worth naming. Learning a phrase inside a song teaches you that phrase in that grammatical shape. Recognising a word when the chorus hands it to you is not the same skill as producing it in a conversation, which is why self-testing still belongs in the loop even when the tune is stuck in your head.
How to use this on your own words
- Put the hard words in the chorus. Wallace's result was driven by the same melody returning, so the repeated section is where difficult vocabulary earns the most.
- Sing along instead of only listening. The Ludke group produced the phrases out loud, and production is where their advantage showed up.
- Keep one tune per set of words. A new melody for every verse cost people recall. Familiarity is the mechanism, not novelty.
- Use the earworm, then break it. When a line starts looping, try producing the word without the music behind it. If it only comes with the tune, it is not learned yet.
- Review on a schedule anyway. Spread the repetitions out over days and weeks rather than stacking them into one evening.
If you want a concrete place to start in one language, we have longer guides for learning German with music, learning Spanish with songs, learning French with music, learning Turkish with music and learning Italian with music.
How Lingotify is built around this
Lingotify was built from these findings, not loosely inspired by them, so it is worth being specific about which part does what.
- You choose the words. The song is generated from your own vocabulary list, so the lines that repeat are made of the items you actually need, not whatever a songwriter happened to pick.
- The Hook preset repeats your weakest words. It takes the one or two words your review history says you are struggling with most and builds a chorus phrase that returns them again and again, which is the condition Wallace found was necessary.
- Karaoke runs word by word. Each word lights up as it is sung, so the sound and the written form arrive together instead of one after the other.
- Quizzes appear inside the song. A vocabulary word is blanked out of the line you just heard and you fill it in, which turns a listening pass into a production test.
- Every answer feeds an SM-2 spaced repetition schedule. That is the part music cannot do for you: deciding when the word should come back.
You can hear the chorus loop running on a real generated song, and try the tap-a-word and in-song quiz behaviour yourself, in the demo on the homepage. It works for German, Spanish, French, Italian, Turkish and English, and the first song is free.
Sources
- Paivio, A. (1991). Dual coding theory: Retrospect and current status. Canadian Journal of Psychology, 45(3), 255 to 287. doi.org/10.1037/h0084295
- Wallace, W. T. (1994). Memory for music: Effect of melody on recall of text. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20(6), 1471 to 1485. doi.org/10.1037/0278-7393.20.6.1471
- Ludke, K. M., Ferreira, F., and Overy, K. (2014). Singing can facilitate foreign language learning. Memory & Cognition, 42(1), 41 to 52. doi.org/10.3758/s13421-013-0342-5
- Jakubowski, K., Finkel, S., Stewart, L., and Müllensiefen, D. (2017). Dissecting an earworm: Melodic features and song popularity predict involuntary musical imagery. Psychology of Aesthetics, Creativity, and the Arts, 11(2), 122 to 135. doi.org/10.1037/aca0000090
- Liikkanen, L. A., and Jakubowski, K. (2020). Involuntary musical imagery as a component of ordinary music cognition: A review of empirical evidence. Psychonomic Bulletin & Review, 27(6), 1195 to 1217. pubmed.ncbi.nlm.nih.gov/32583211
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., and Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354 to 380. doi.org/10.1037/0033-2909.132.3.354