Tuesday, October 25, 2005

Let me tell you how sad I am, slowly

Dave Munger has a good post describing a recent study by Isabelle Peretz and Lise Gagnon on emotional effects of music. It turns out that tempo is more important than mode (major vs. minor keys) in creating the sensation of happiness or sadness. However, I find the tempo choices to be quite bizarre. The slowest tempo was 110 beats per minute, well above the accepted normal range. Dave doesn't say how the primary beat was determined, a very complex matter as Peter Martens could tell you. But typical music definitions of slow tempi are in the range of 40-60 beats per minute, half of the study's speed.


Michael J. West said...

I'm listening to a Captain Beefheart track right now that I'd estimate is in the 110 bpm range. Being Beefheart, it doesn't easily lend itself to description, but I'd wager it ain't "sad."

Maybe you disagree with me, Scott, having much more experience with this, but I find that the 1 to 10 scale can be misleading. I've taken part in similar experiments before and find that when I'm forced to quantify that rigidly, I get confused and to a sloppy job with the classification.

Does that make sense?

Scott Spiegelberg said...


I don't disagree, though for different reasons. Psychologists have found that scales with too many steps can be confusing because there are too many choices. Especially when the steps are categories rather than quantities, having fewer choices are more reliable.

Dave Munger said...


The point isn't so much the absolute tempo but the relative tempo -- the same sequence of notes played at different tempi. Also, since they were using synthesized music, with no opportunity for expression, I'd submit that going much slower than 110 b.p.m. would put participants to sleep.

Scott Spiegelberg said...


That is a valid point, yet I still think it is an odd decision. 110 is barely less than a march tempo of 120, and marches are not associated with slowness!

Peter (the other) said...

I agree with you, Scott, that 110 is barely slow.

I think what I might consider from this study, is that a slow (truly slow) tempo will seriously inhibit any other musical, emotional cues from carrying "happiness", and the other way around.

My own work involves the emotional affect of the smallest "museme" of harmonic morphology. That is, that the basic emotion is created in the change from one chord to another, not as much, in the chords themselves. Tempi is a modifier of this, for example, sad/fast might be "angry".

Another problem I have with the Peretz (as reported on the blog), is the lack of semiotic cleanliness. What do they mean, happy/sad? This is why I think that Phil Tagg's semiotic analysis has to be run on all material before testing.