The first time I heard a marmot pup scream I almost dropped it. The sound was like I had just stepped on a cat, or a baby! It was loud, whiney, and noisy. The frequency went up and down rapidly. Physically, I heard it in my ears but I felt it in my gut -- I had an uncontrolled emotional response. The baby rodent was smaller than an adult squirrel, but its teeth were just erupting and it was defenseless. I was holding it in gently in one hand and petting its back, and with its furry tail and big eyes, it could have passed for a small puppy. How could such a cute little fuzzy rodent elicit such a response?
Let me explain. I'm a marmoteer -- a field biologist who's spent much of my career observing and listening to these large, cat-sized ground squirrels. I sometimes have to catch them in order to mark them so that we can observe their behavior from a distance, and to collect samples (hair for DNA, feces for hormonal and parasite studies, and a small blood sample for parasite and hormonal studies). The vast majority of animals are quite docile when we trap them, and while I've captured marmots thousands of times, I'd never heard an adult scream.
Marmots, I learned, scream mostly when they're really young. And they're not the only animals that shriek like this. I once heard a young deer scream as it was being killed by a coyote, and there are descriptions and recordings of the screams from many species of birds and mammals -- all of these screams are emitted under duress. And, of course, humans scream.
What's fascinating about this is that a human listener would realize that these are screams -- they sound very similar even though they are produced by a variety of different species. And, like people, animals scream for a very good reason: to get help. Indeed, while Darwin didn't know about marmots (nowhere in his writings are marmots mentioned), he did know about screams. Darwin said that screams were emitted by young animals to recruit help from their parents.
Screams are characterized by rapidly changing frequencies and noise. Noise is chaotic energy that is found at a variety of frequencies and sounds raspy or distorted to us. Noise and rapidly changing frequencies are two types of things that physicists and bioacousticians refer to as non-linear phenomena.
Nonlinearities are properties of systems, like the sound production system in your stereo, or that found in a wind instrument, mammalian larynx, or an avian syrinx. All work similarly in the sense that as you increase the amplitude on your stereo, at some point the sound coming out doesn't just get louder, it gets distorted. The same thing happens when you overblow a trumpet, a larynx or a syrinx -- blow a little, it gets louder but at some point the sounds get noisy and distorted.
While much work has focused on precisely describing the structure of these sounds, there are only a few hypotheses that have focused on their function. One compelling hypothesis is that because nonlinear sounds are somewhat unpredictable, they capture our attention. I have found that noise added to normal marmot alarm calls makes them more evocative, but the idea should be generalizable to many species.
There aren't many marmots in LA, but there are certainly a lot of talented musicians. While giving a public talk at UCLA I mentioned that I bet that musicians and film score composers and audio engineers capitalize on the especially evocative nature of nonlinear sounds when they compose music and create compelling film scores. After my talk, Peter Kaye, a musician and film score composer who was studying the biological basis of emotion in music, came up to me and said that he thought I was right. We immediately began collaborating.
Our first finding (published in the scientific journal Biology Letters in 2010) was that film soundtracks either enhance or suppress -- depending on the film's genre -- what we refer to as "simulated nonlinearities." I say "simulated" because these are not naturally produced sounds, but rather emerge from the mix of voice, music, Foley (sound effects), and background noise. For instance, noisy (female) screams are more common than would be expected in horror films (think Psycho), while noise is much less common than would be expected in keystone scenes in sad dramatic films (think The Green Mile).
These results were correlative in the sense that we looked at associations between films and certain types of sounds; we didn't specifically compose music and see how people felt. Experiments are always better than correlative results, so back at UCLA I chatted with my colleague Greg Bryant, a musician and professor in the Department of Communication Studies and an expert on how people respond to emotional sounds. Greg, Peter and I set out to conduct an experiment to formally test our hypothesis, which involved seeing if we could scare humans with biologically-inspired music!
Greg and Peter used a computer synthesizer and composed 10-second clips of rather bland music (think of Musak or the sounds you'd hear when you're placed on hold at your dentist's office). Then they modified them after five seconds to include musical noise (the sort of distortion that Jimi Hendrix made famous) or very rapid frequency changes.
Greg then clapped headphones on 42 UCLA undergraduates and played the pieces to them. We asked the undergraduates to score the sounds on two dimensions -- arousal (how emotionally stimulating or active a sound was) and valance (how positive or negative -- happy or sad -- a sound was).
We found, as we've just published in the journal Biology Letters, that compared to the Musaky music, music with noise and rapid frequency shifts up, were more arousing, and music with noise and upward frequency shifts were perceived as sadder or more frightening. Essentially, these young twenty-somethings were responding to nonlinear music the same way that marmots respond to screams.
This finding is the first experimental finding (that we are aware of) to show that the addition of nonlinearities to music modifies emotional judgments by humans.
This is an important finding because it's essentially reverse engineering how emotions are conveyed in music. There is a very large literature on music that focuses on the chords used (minor chords evoke sad feelings), the tempo (rapid tempos are arousing), and the amplitude (sudden increases in volume are arousing), but none that really start from the perspective of how emotions are communicated by non-humans, and I think that there's a lot to learn by studies of non-humans.
This has broader implications because it illustrates a biologically-tested way that musicians can manipulate our emotions. Indeed, I think that there is ample opportunity for more studies that start from the perspective of capitalizing on 3.5 billion years of life to identify nature's rules. At this point, we seem to have identified one--that noise and other nonlinearities are particularly evocative and are used to convey emotions.
Successful musicians, like Jimi Hendrix and film score composers, may not have known this explicitly, but they certainly tapped into our inner animal when they created emotionally evocative sounds. We should all listen to our inner marmot more for inspiration, and our on-going work explores these and other hypotheses about how and why music, as Greg says, brings out the animal in us.