Take a listen to this brief audio clip of “Unforgettable.”
Aside from the fact that it’s a computer-generated MIDI performance, do you hear anything unusual?
If you’re a non-musician like me, you might not have noticed anything. It sounds basically like the familiar song, even though the synthesized sax isn’t nearly as pleasing as the familiar Nat King Cole version of the song. But most trained musicians can’t listen to a song like this without cringing. Why? Because the music has been made “bitonal” by moving the accompanying piano part up two semitones (a semitone is the difference between a “natural” note and a sharp or flat). Here’s the original, unaltered piece:
Can you tell the difference? A 2000 study led by R.S. Wolpert found that non-musicians couldn’t distinguish between monotonal and bitonal music played side-by-side. Meanwhile musicians found artificially-created bitonal music to be almost unlistenable. For most non-musicians, if they heard anything wrong with the clips, they typically said they were being played too fast, or mentioned some other unrelated concept.
But Mayumi Hamamoto, Mauro Bothelo, and Margaret Munger (AKA Greta) wondered if years of musical training were really necessary for non-musicians to hear bitonal music. Bitonality is actually a bit controversial in the world of music, and it can be a little hard to define. In principle, there’s a difference between bitonality and just playing or singing off-key, but in practice, the difference may not even exist. Advocates of bitonality like to point to the works of composers like Milhaud, Bartók, Prokofiev, and Strauss. These composers deliberately wrote in two different musical keys. But how is that different from occasionally or regularly writing dissonant chords? After all, all the same notes can be written using any musical key. To be truly bitonal, advocates say the two separate parts must unfold independently in different keys. This results in a distinctive “crunch” when the music is played. The separate question is, is this noticeable? Wolpert’s work shows that it is, at least for trained musicians.
Hamamoto’s team replicated Wolpert’s study by playing altered and original clips of familiar songs like the above example to three groups of undergraduates: “Musicians” with more than 5 years of training, “Amateur Musicians” with 1 to 5 years of training, and “Non-Musicians” with less than a year of training. There were 14 students in each group. Musicians were significantly better at noticing that the modified clips were bitonal or “out of tune.”
Next, everyone was given brief training session, where instead of modifying monotonal music to be bitonal, some of Milhaud’s music originally intended to be bitonal was modified to be monotonal. Here’s an example bitonal piece (Milhaud’s “Botafogo”):
After hearing the clip and seeing it identified as bitonal, the students were told
Notice sometimes there is a “crunch” in the sound. This should sound somewhat unpleasant and feel like it shouldn’t be that way.
Then they listened to a manipulated version of the same clip:
Again, they were told this clip was monotonal and directed to notice how the sound seems smoother and more pleasant (to my mind, it’s not nearly as interesting as the original — but that wasn’t part of the study). Next they were trained with feedback, listening and identifying clips until they could accurately label four in a row. This took just a few minutes.
Finally, the respondents were tested on four new clips, all songs by Milhaud. This graph shows the results:
As you can see, for all the songs except “Ipanema,” the students were quite accurate at identifying both bitonal and monotonal songs (error bars are 95 percent confidence intervals). More important, however, was that there was no significant difference in the results for Musicians, Amateur Musicians, and Non-Musicians. All three groups fared equally well.
The authors conclude the identifying bitonal music isn’t a matter of years of musical instruction; it can be achieved with just a brief training session. In fact, the Non-Musicians took no longer than Musicians to complete the training session, so years of experience don’t even help with learning about bitonality.
It also may suggest that the controversy about whether bitonality actually exists may not be warranted. If nearly everyone can hear the difference, then it’s probably a genuine musical phenomenon.
Note: More of the clips from the study can be found on Greta’s website here.
Hamamoto, M., Bothelho, M., & Munger, M.P. (In press). Non-musicians’ and musicians’ perception of bitonality. Psychology of Music. DOI: 10.1177/0305735609351917