Just What Is It I’m Doing?

Every scientist should be able to explain their research to someone who’s not in the field. Here’s my go.


Header image: Spectrogram of a rhesus macaque scream

Every scientist should be able to explain their research to someone who’s not in the field. I owe that much to my friends and family who have expressed interest in what I do, and the public whose tax dollars have gone to fund my work. With that in mind, this post is devoted to describing my dissertation, which is titled “Vocal Emotion Expression Across Contexts, Vocalization Types, and Species: Implications for General Processes of Vocal Evolution.”

I wanted to include all the background someone would need to understand why my dissertation research is important and what it all means, so this post is fairly lengthy; if you absolutely need the short version, skip ahead to the “what I’m actually doing” section. But I know that because you love me, you will read the whole thing…

Finally, as you might know, I am also working on a line of research investigating human screams—that research involves some of the themes I’ll cover here, but it is different enough from my dissertation that I’ve decided to cover it in a separate, future post. Stay tuned for that.

Animal Emotions, Feelings, and Evolution

I find emotions really fascinating; I want to understand what they are, what they do, and why we have them. And by “we,” I mean not just humans, but also other species. Charles Darwin believed that animals’ expressions betrayed an emotional life similar to that of humans. This idea caught on, leading some scientists to anthropomorphize animals in ways bordering on absurd (including examples I’ve talked about in a previous blog post). This had a chilling effect: for the better part of the 20th century, animal researchers were highly skeptical of the idea of animal emotions. While their skepticism was justified, they went too far in the opposite direction. The topic of animal emotion became taboo, and some scientists are still very hesitant to even talk about it. But the evidence is clear: advances in biology and neuroscience have shown that many nonhuman species (particularly mammals) do in fact have emotions, and these are very similar, biologically, to human emotions.

Illustrations from Darwin’s book, The Expression of the Emotions in Man and Animals

A tangential, but important point: by “emotions,” I and other scientists mostly mean neural and physiological activity: measurable things going on in your body. For example, heart rate increases, firing of neurons in your amygdala, and release of cortisol into the blood stream are all aspects of the emotion known as fear. Emotions do not equal our subjective experience of them—what we feel—which we call “feelings” (yes, that is the actual scientific term). The fact that nonhuman species have human-like emotions doesn’t mean they have human-like feelings. Scientific consensus is lacking on this. However, my professional opinion is that, because we share evolutionarily old emotions with other mammals, we should assume that feelings are also evolutionarily old, and therefore that other species have them. You might be thinking, “Of course! You just have to look into a dog’s eyes to know that.” I would respond by cautioning against relying so heavily on your intuition. We evolved to anthropomorphize others, that is, to look around and see other beings like ourselves, with minds and experiences like our own; just because we apply that evolved tendency to other species (and even inanimate objects!) doesn’t mean we are right to. That anthropomorphic bias is just that—a bias—and late-19th-century scientists going too far with that very bias is what caused later scientists to be overly skeptical of animal emotions in the first place. I believe that other mammals have feelings because the available evidence and reason suggest it is likely, irrespective of my own anthropomorphic biases.

I may be interested in emotion, but my roots are firmly in evolutionary biology, specifically ethology: the study of the evolution of animal behavior. So within the study of animal emotion, I am primarily interested in how animals’ emotions can and have influenced the course of behavior evolution. Broadly speaking, ethologists have acknowledged the emerging evidence of animal emotions with a collective shrug: “The question of emotions is interesting in its own right, but ultimately it concerns how an animal’s behavior happens. My research is focused on why, evolutionarily, the behavior happens, and so emotions are not really relevant to me.” I disagree. In fact, I believe that it’s impossible to fully understand how a behavior evolved, or is evolving, or will evolve, without understanding how the behavior occurs—including the role of emotions. I will come back to this.

The Sound and the Fury: Emotion and Vocalization

In addition to emotions (and some other things that won’t make it into this post), I’m also interested in sounds; that we perceive such complex sounds from mere vibrations in the air is one of the most mind-blowing things in the universe to me. So it comes as little surprise that my main area of research is how emotion influences the sounds that we (humans as well as other species) make.

Roaring behavior by black and gold howler monkeys, a species closely related to the ones I studied for my Masters. Photo by Steve

Our emotions influence not only what we say, but also how we say it. When you are especially angry, especially happy, especially afraid, etc., certain changes happen to your voice: the pitch goes up, you talk faster, the tone changes, your voice gets louder, harsher, etc. These changes occur involuntarily, and often without conscious awareness. In the last couple of decades, scientists have begun to study whether other mammals’ emotions influence their voices in similar ways. Turns out, for the most part, they do. This suggests that these emotional changes in the voice are probably inherited, in humans and other mammals alike, from a common ancestor, likely one living over 100 million years ago. That said, this view is still tentative, and more research is needed.

What about emotions cause these changes in the voice? Our scientific understanding of this is murky, but we do know some things. Before getting into the specifics, it’s important to understand the effects emotions have on the body more generally.

Emotions involve changes not only in your psychological state, but also in your body. The best-understood example of this is the phenomenon of “arousal.” Contrary to colloquial usage, arousal doesn’t mean the experience of being turned on sexually, although that is a particular kind of arousal (called, as you might guess, “sexual arousal”). Arousal simply refers to a state of excitation, wakefulness, and alertness. Many emotions involve increased arousal, including joy, fear, and anger, while other emotions, such as sadness, are typically low-arousal. In fact, a prominent scientific view of the different emotions is that they are not actually distinct states, but just different points along the spectrum of arousal (along with 1-2 other spectra—no need to get into the details here).

Among other things, increased arousal causes changes to respiration, heart rate, and muscle tension. These changes can be helpful. Imagine an early mammal in the presence of a predator: sensing the predator nearby causes it to enter a fearful, and hence aroused, state. Increased heart rate, respiration, and muscle tension all prepare the mammal’s body to escape or fight for its life—the so-called “fight-or-flight” response. These bodily changes have, throughout time, helped our evolutionary ancestors and cousins survive and reproduce, and have therefore been favored by natural selection.

Importantly, the effects of arousal are not limited to the muscles involved in fight and flight, but are instead quite global—including, you guessed it, the muscles of the voice. Check out this illustration of the “vocal apparatus” (all the parts of your throat, nose, and mouth that contribute to the sound of your voice).

Image result for vocal tract]
The vocal apparatus, composed of dozens of muscles

Increased arousal involves tension in many muscles in and around the vocal apparatus, and hence, changes to sound of the voice. Tension in the vocal fold muscles (formerly known as the vocal cords) can increase the pitch of the voice, while tension in the diaphragm can increase its volume, and changes in the pharynx can alter the tone of your voice, etc. There are dozens of acoustic variables like these that have been shown to change in association with emotional states, and many are associated with arousal.

We know how arousal and other aspects of emotions influence the voice in many species, but by no means all. And while there seem to be some similarities across species in terms of vocal emotion expression, we can’t assume that they’re all the same. Expanding our view of the relationship between emotion and vocal communication across species is important for a few reasons. First, there is the basic intellectual merit intrinsic in learning more facts about the natural world. Second, decoding the emotional significance of vocalizations in a species can allow human caretakers to better know what emotions an animal is experiencing, and therefore, help them provide better care. Finally, by studying more and more species, we can get a better idea of which aspects of the relationship between emotions and vocalizations are universal and which are unique. This influences our view of the evolutionary history of vocal communication, in ways that I’ll get into below. But first, it’s probably time I stop burying the lead and get into what I’m actually doing.

What I’m Actually Doing

File:Rhesus macaque (Macaca mulatta mulatta) female.jpg
A young female rhesus macaque. Photo by Charles Sharp

For my dissertation, I’m exploring the link between arousal and vocalizations in rhesus macaques, an Old World monkey species native to South, Central, and Southeast Asia. These monkeys live in large, matrilineal social groups, meaning groups are structured around female families (matrilines)—sisters, daughters, aunts, and female cousins stick up for one another. Research institutions house more rhesus macaques than any other primate species, making it especially important to understand their emotions and behavior, so that scientists can care for them properly.

Rhesus macaques have a roughly average-sized vocal repertoire made up of around 10-20 distinct calls (depending on how you define a distinct call—that’s a whole can of worms). My research is focused on two of these calls, “coos” and “screams.” Coos sound like “hoo,” and range from jovial to forlorn in melody (these are very unscientific terms but they do the job here). Coos are an example of what’s called a contact call, meaning they are used in a wide variety of contexts, but mostly they function to convey that “I’m over here.” As you might guess, screams—which sound to me more like a screech than a human-like scream—are limited to intense situations, but unlike in humans, rhesus screams don’t usually have to do with extreme danger. Instead, individuals of this species scream when they are being threatened or attacked by a member of their social group. The function of screams is thought to be to let female kin know that “I’m being bullied, come get my back!”


Spectrograms of a rhesus macaque scream…
and coo





My central question research question is, how does arousal relate to vocalizations in this species, and how consistent is this relationship across different vocalization types and different contexts? The first half of that is something that many researchers have looked at in a variety of species (although not rhesus macaques), but the second half is more unusual.

I’m addressing this question using a collection of coos and screams from juvenile rhesus macaques living at the Yerkes National Primate Research Center Field Station, located in Lawrenceville, GA. All my monkeys live in a social group of around 60 adults and their immature offspring, in a large (~180×180 ft) outdoor enclosure. I collected my body of screams by hanging out outside the enclosure and watching the monkeys through the fence with a pair of binoculars and a digital recorder and mic—every time I picked up a scream, I would note which monkey it came from and what the context was. The types of bullying that elicit screams range in intensity from a mere lunge forward to a full-on attack, and while all of these interactions probably evoke increased arousal in the victim, I’m comfortable assuming that an attack victim would experience a greater arousal increase than the victim of a mere threat. So, what I have amounts to a bunch of scream recordings representing a range of different arousal levels. This is going to allow me to examine the correlation between arousal and sound qualities of screams. To do this, I’ll be measuring around 20 sound-related parameters of all these screams, relating to pitch, tone, length, rate, etc.

What it’s like inside the enclosures at the Yerkes Field Station

My collection of coo vocalizations comes from a collaboration with the Stress, Obesity, and Diabetes project at Yerkes. One of the things these researchers are studying is the relationship between stress and behavior in rhesus macaques; to study this, they administered behavioral tests to monkeys, videotaped their behaviors, and measured their levels of cortisol, a hormone associated with arousal and stress. One of the behaviors you see a lot during these tests is cooing; as a result, I have a body of coos all associated with biological measurements of arousal. Like with the screams, I can examine the correlation between arousal and sound qualities of coos. And, as I mentioned above, I can compare the results with coos to the results with the screams, to see how consistent arousal’s effects on the voice really are in this species.

Finally, I’m also working on another project investigating how humans perceive arousal from rhesus macaque vocalizations—the same vocalizations I described above. As I live out of state, my fabulous undergraduate assistant Emma is currently running participants through this experiment. The question here is, do people know, just from listening, which of two coos or two screams represents greater arousal? Even if they’ve never seen or heard a rhesus macaque before? If so, that would suggest one of two possibilities: either we have some innate understanding of the emotional significance of the vocalizations of other species, or can use our learned understanding of emotion in the vocalizations of the species we are familiar with (humans, pets, etc.) and apply it to the vocalizations of an unfamiliar species. In either case, the takeaway would be that there is some cross-species evolutionary continuity in how arousal is communicated through vocalizations.

This idea makes sense if you think about it: the basic anatomy of the vocal apparatus is not that different across primates and other mammals, and neither is the way arousal works on muscles throughout the body, so it shouldn’t be shocking that we should see evolutionary similarities. But that remains a hypothesis that needs to be tested. By doing so in my dissertation, I’m extending the central research question—how consistent is the relationship between arousal and vocalizations across different vocalization types and different contexts?—to include different species as well.

File:2008-06-28 Ruby begging.jpg
Humans seem to be sensitive to the arousal level in dog vocalizations, but it’s hard to know how much of this is because dogs are domesticated. Asking similar questions with undomesticated species like macaques can help address this. Photo by Ildar Sagdejev

Vocal emotion expression and evolution: Bringing it all together

It is tempting to ask, what are the evolutionary benefits of a particular sound in a particular situation? Perhaps a high-pitched call during an attack can startle a predator long enough to allow for escape? Maybe a harsher tone of voice in aggressive situations could intimidate potential rivals? Ethologists interested in vocal communication have made plenty of conclusions like these. My view is that these explanations, although perfectly plausible, are missing something—something that my research will help address.

This next part is going to be a little lofty and theoretical, but bear with me. Remember earlier when I said that you can’t understand how a behavior evolved without understanding how it happens? That holds true for vocal communication: I believe you can’t understand the evolution of the sound quality of a vocalization in a specific situation, without understanding how that vocalization is produced, including both the anatomy of the vocal apparatus and how it is acted upon by the animal’s emotions. To illustrate why, here is a hypothetical example: Imagine a population of early mammals. Some have this thing called arousal—when they are in intense situations, their whole body tenses up—and others have nothing of the sort. The individuals with this arousal thing are going to be better at escaping predators etc., and so natural selection will favor the tendency to become aroused in intense situations, right? Now, arousal may result in some bodily changes that have nothing to do with survival or reproduction—including, perhaps, changes to the voice (increased pitch, volume, etc.). These vocal changes don’t need to be beneficial, they could just happen to be consequences of the way arousal works. Nonetheless, as the link between intense situations and arousal is strengthened by natural selection, so too will be the link between those situations and these vocal changes. In other words, it could well be that natural selection has favored arousal in certain intense situations due to a few specific beneficial effects (e.g., muscle tension in the limbs), and then the vocal changes associated with arousal are simply coming along for the ride.

File:FMIB 47726 Features in the Life History of the Ribbed Clam (Paphia staminea).jpeg
Gould and Lewontin (1979) argued vehemently that scientists should pay more attention to the idea of nonadaptive byproducts, using the ridges on bivalve shells as an example.

This process is referred to in evolutionary science as “correlated response to selection”: natural selection acts on one trait, and then as that trait evolves, other traits evolve as well even though they may not be beneficial in any way. We call the latter trait a “non-adaptive byproduct.” Correlated response to selection can only happen when the two traits are biologically linked in some way. In this case, emotion—specifically, arousal—is the critical link between the naturally selected trait (increased limb muscle tension in an intense situation) and the byproduct (changes to the voice in similar situations).

Taking the above idea further, another possibility is that any given vocal change in an intense situation might be a nonadaptive byproduct of natural selection on some other vocal change. As a hypothetical example, perhaps higher pitch is useful for escaping predators, so natural selection favors increased arousal in an intense situation, due to the effects of arousal on pitch. Because arousal also makes for a noisier tone, you might see that vocalizations during predator attacks are noisy. But it could be that the noisiness isn’t itself beneficial, but rather is a nonadaptive byproduct of natural selection favoring increased pitch (and hence increased arousal). As you can probably imagine, there are all kinds of possibilities like this; I’m tempted to keep giving examples, but I’ll leave it there as I think I’ve gotten the general idea across.

Geneticists and paleontologists have been talking about correlated response to selection for a long time—as you might guess, I think it’s one of the coolest ideas in evolution—but the concept has been much less common in ethology, and totally absent from the scientific literature around animal vocal communication. Why is that? Well, I believe that it comes back to ethologists’ tendency to dismiss the idea of emotion as irrelevant to our work. Remember, the critical link in this correlated response to selection idea is arousal, which is an aspect of emotions. If you think emotions are irrelevant, then it’s little surprise you wouldn’t consider the potential evolutionary significance of arousal.

So now I return to the central research question of my dissertation: how does arousal relate to vocalizations in rhesus macaques, and how consistent is this relationship across different vocalization types, different contexts, and different species? Another way of putting this is, how strong is the link between the different effects of arousal? Remember, this link is the critical piece in the hypotheses about correlated response to selection that I have outlined above. So among other impacts, my research will give us a much better sense of how likely this correlated-response-to-selection idea is for vocal changes, relative to the more common “oh, it must be beneficial in some way” types of conclusions that currently dominate the scientific literature in this area.

And there you have it! If you’ve stayed with me through this whole post, I commend your tenacity and thank you for reading. I hope that this has given you a better sense of what it is I’m doing, and I hope I’ve gotten you just the littlest bit excited about my research. Of course, please feel free to reach out to me with any lingering questions or comments you might have. Otherwise, see you next time.

Author: Jay W. Schwartz

I'm a Ph.D. candidate in Psychology at Emory University. My research focuses on emotion, vocal communication, and evolution in nonhuman primates and humans. Visit my blog to learn about exciting animal research and issues: https://jaywschwartz.wordpress.com/blog/

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s