We hear different music depending on how we dance to it

Listen to this short recording:

Blogging on Peer-Reviewed ResearchIt's a sequence that repeats every sixth beat. But when we're listening to music, we usually prefer to divide rhythm into two- or three-beat patterns (duple or triple rhythm). In this case, the sequence doesn't make it obvious which pattern is correct. A traditional duple rhythm, like a march, would accent every other beat -- the musicians play every other note a little louder. Similarly, a traditional triple rhythm, like a waltz, accents every third beat.

When we're dancing (or marching), we move in time to these accented beats. Indeed, it's often difficult to resist moving when we hear music. The Western tradition of sitting still to listen to "classical" music, which only originated in the past few hundred years, may be the only setting in which movement is not considered an integral part of experiencing music.

It may not be too much of a stretch to suggest that movement itself may also affect the sounds we perceive. Jessica Phillips-Sliver and Laurel Trainor devised a set of experiments to test that notion. While volunteers listened to the recording above, the experimenter showed them how to "bounce" along with the rhythm, either every other beat, or every third beat. After bouncing along for two minutes, the listeners heard two additional recordings, which added a minor modification to first recording.

In one recording, every other beat was accented, creating a duple rhythm:

In the other recording, every third beat was accented: a triple rhythm:

Then they were asked which recording matched the recording they had bounced along with. Here are the results:


Nearly everyone who bounced every second beat thought they actually heard a duple rhythm, while those who bounced every third beat thought they had heard a triple rhythm -- even though the identical recording had been played for every listener during the bouncing phase.

In two subsequent experiments, Phillips-Silver and Trainor eliminated two potential explanations for the results. Wearing headphones had no effect on the results (thus showing that the motion of the listener relative to the sound source didn't cause the effect). Neither did wearing a blindfold (eliminating the possibility that a visual, not a movement cue, was responsible for the effect).

In their final experiment, listeners sat still while watching the experimenter bounce at either a duple or a triple rhythm. Here are those results:


Now there was no significant difference between the different bouncing conditions. Phillips-Silver and Trainor argue that it must be the listener's motion that causes different perceptions of the sound.

Phillips-Silver, J., Trainor, L. (2007). Hearing what the body feels: Auditory encoding of rhythmic movementâËâ . Cognition, 105(3), 533-546. DOI: 10.1016/j.cognition.2006.11.006


More like this

When Greta earned her Ph.D. 13 years ago, Jim was two and a half years old, and Nora was just 10 months old. Jim knew a few words, and Nora couldn't talk at all. You might think a baby as young as Nora wouldn't have an appreciation for music or dance. If you can't walk, what good is dancing? But…
Two weeks ago, we challenged our readers to see if they could discern the difference between MP3 recordings at different sampling data rates. Nearly 700 completed our study. So does a very high data rate result in a noticeable difference? Here our are basic results: Respondents rated two…
Snowball, the sulphur-crested cockatoo, is an internet superstar. He's known for his penchant for grooving to music, notably Everybody by the Backstreet Boys. As the music plays, Snowball bobs his head and taps his feet in perfect time with it. If it speeds up or slows down, his rhythm does too. He…
There was some doubt as to whether the "tone-deafness" test I linked to Monday really tests for amusia. The defining trait of amusia is the inability to discern the difference between different musical pitches. So here's a test that might generate a more clear-cut result. The following track plays…

I didn't think either of the modified ones sounded right. I bounced on my own in duple, but the original almost has an odd-meter feel to it.

By marciepooh (not verified) on 11 Dec 2007 #permalink

"I didn't think either of the modified ones sounded right"
Its just a matter of accenting, you can hear exactly the same rhythm being played in all the examples but as it is in 6/4 you can accent different beats to change the feel.

Maybe the full description of the procedure would include this, but I would think that the participants would need to be able to select "neither track sounds correct". Without that choice, they could be selecting the track that matches their movement regardless of what they hear?
Or allowed to select a track from all three: six-beat, two-beat or three beat? Interesting experiment!

I had the same thought as Marcy did. Why weren't they given the choice of the original recording?


This is very similar to Hemiola - a common compositional technique which is used to induce uncertainty in a piece of music, a good method for enhancing expectation before a resolution - although there are other uses. It's caused by a change in emphasis - which can be initiated by the listener (a change in focus), or by the music itself (either melodically or through volume).

American minimalist music uses these and other techniques to invoke a feeling of illusionary stasis - nothing changes yet there is the uncanny feeling that all is shifting. In the literature the technique is related to/described as 'phase shifting'. Steve Reich, Philip Glass, John Adams etc.

The ambiguity caused by the shifting of emphasis is similar to the visual phenomena of figure ground problems..like the vase/profile.

I've heard a lot about how perception shapes reality, but this is the first case in which I've heard of our movement shaping our reality as well. I like it!

I'm always amazed at the versatility and adaptability of the brain. I wonder why this phenomenon occurs. I've also read that when you think you hear music or you play a song in your head, that your brain is doing the exact same thing as when you actually hear it.

Could there be an actual cognitive stage involved in the bouncing, where people actually start counting the beats? In that case, it wouldn't necessarily be the movement per se that affected the interpretation, but rather the additional thought that the bouncing induced?

I wonder if having formal musical training skewed the results (I could only read the abstract)? I just question the proposition in the abstract that the piece is "an ambiguous rhythm with no accented beats". I guess you could argue that there no accented beats because the volume is normalized and rhythm is completely quantized.

However, the reference clip just sounds to me like 3/4 time, but it uses a (very "square") triplet between the 2nd and 3rd beat... which just sounds very odd. The first comparison clip just sounds like straight 2/4 to me and the last clip sounds like 3/4 again with an accent on the down beat as well as the "and" of 2nd beat... again rather "odd" sounding.

So I guess -- to me -- the second comparison clip sounds like the best match simply due to the matching time signatures. I'm just not sure the reference clip would sound "ambiguous" to anyone who plays an instrument.