Roar
July 24, 2018 8:50 AM   Subscribe

Jordan Raine had a hunch that humans might use roaring to communicate information about size and strength, so he scienced it. He found that we're pretty good at determining whether someone is stronger than us based on their roar, with one exception: In cases where males were weaker than female listeners, they were only correctly identified as weaker 25% of the time.
posted by clawsoon (37 comments total) 9 users marked this as a favorite
 
Interesting choice of measurement of strength (from the methods PDF):
Flexed bicep circumference and handgrip strength were also measured, and these measurements were aggregated to produce a single, equally weighted, z-scored strength value for each subject (following Sell et al. 2009; Puts et al. 2011, and others). These measures explain approximately 55% and 24% of the variance in strength as measured by weight-lifting machines in male college students, 60 respectively (Sell et al., 2009).
I can't access that Sell et al. paper to see more about this, but I'm not sure that I would use bicep circumference as a measurement of strength when comparing females and males.
posted by exogenous at 9:02 AM on July 24, 2018 [4 favorites]


Here it is: Sell et al., 2009
posted by chrchr at 9:21 AM on July 24, 2018 [1 favorite]


Just downloaded the sample vocalization recordings. Super worth it to get an idea of the range.
posted by es_de_bah at 9:27 AM on July 24, 2018


I heard about it on CBC Radio's Quirks and Quarks, which has an audio interview with the author here.
posted by clawsoon at 9:37 AM on July 24, 2018 [1 favorite]


I mean, I guess these are not professional actors, but it's not easy to make "that's enough, I'm coming for you!" sound convincing. If you really want to scare someone, try yelling "That's my purse! I don't know you!"
posted by sfenders at 9:38 AM on July 24, 2018 [7 favorites]


Dangit, Bobby!
posted by darkstar at 9:42 AM on July 24, 2018 [2 favorites]


Hmmm. It's an interesting study, but I have reservations.

The biggest, to my mind, is the question of mechanism. In the original paper, the authors propose no mechanistic account whatsoever of how listeners might be able to judge a person's strength and height from their roar. In the popular press article, Raine does mention the relationship between vocal tract length and voice pitch, but this doesn't really address the question of strength. Without a plausible mechanism, this is merely an interesting correlation in a small sample size, which could quite possibly just be statistical noise.

This is a particularly strange oversight considering that the perception of social dominance, something which in nonhuman animals is closely linked to size and strength, from vocal cues has in fact been studied in a mechanistic fashion in humans quite extensively. In the companion nontechnical article, the author writes:
Nor had anyone studied whether we can judge from a voice alone whether someone is more powerful than us.
I don't think this is true, though I don't have a specific citation for physical size or strength. However, it is known that vocal pitch is correlated with androgen levels in men, and androgens promote muscle development in men. So for men, at least, there is prior evidence to support the idea of a relationship between vocal features and physical strength, albeit at a couple of correlations removed [1].

Assuming the result is real, my guess is the mechanism is something like the following. Androgen levels are correlated with physical strength and size, as well as many other things. When listeners are asked to estimate strength and size on the basis of a roar, they mentally lean on their learned associations between vocal features and physical characteristics, the most salient of which are governed by androgen levels. Thus the task effectively becomes "estimate relative androgen levels," which is also tightly coupled with numerous other factors, including perceived gender. The decreased ability to judge relative strength and size from a roar across gender lines is probably partly explained by this.

In other words, the relationship between physical strength and vocal features is probably mediated by hormones. The relationship between vocal features and listeners' perceptions is probably more complex and governed by learned associations and cultural norms about gender and strength.

[1] Note that if A is correlated with B and B is correlated with C, A is not necessarily correlated with C. This is a common statistical fallacy. However, finding A<>B and B<>C gives at least a motivation to look for A<>C, so it's weird not to discuss it.
posted by biogeo at 9:47 AM on July 24, 2018 [6 favorites]


Jordan Raine had a hunch that humans might use roaring to communicate information about size and strength

Unfortunately he had been banned from the University library and was therefore restricted to online sources.
posted by Segundus at 9:54 AM on July 24, 2018 [7 favorites]


The study had 27 cases where a woman listened to a "weaker" (see above and below) man. That's already a somewhat low N, but it's also not independent observations; we're talking about pairs including something like a half dozen each of the weakest men and strongest women among the study subjects.

This is an interesting study, and I'm not trying to criticize someone for not having an infinite army of test subjects to play with, just pointing out that the error bars on that "one exception" are large enough that I wouldn't be completely confident in swearing it's an exception.

Taken at face value it does seem to fit a general pattern that also exists in the male-listener and larger-N data columns, though: people seem to be more likely to overestimate rather than underestimate the strength of a male vocalizer. Cynically I'm reminded of the old story about excessive pattern-matching: people who looked at empty grass and thought they saw a tiger hiding there still got to be among our ancestors, but people who looked at a hidden tiger and mistook it for empty grass did not.
I'm not sure that I would use bicep circumference as a measurement of strength when comparing females and males.
Even grip strength will add a bit of bias for the sake of ease of measurement. Adult women on average have something like half the upper-body and two-thirds the lower-body strength of adult men, and grip strength is in the former category. It's a matter of definitions to say what "really" counts as strength, but I'd put the weights at closer to 50-50 than 100-0, in which case the ratio of female to male strength in this study is getting underestimated by about a sixth.
posted by roystgnr at 10:02 AM on July 24, 2018 [2 favorites]


biogeo: In other words, the relationship between physical strength and vocal features is probably mediated by hormones.

It could also be a direct result of physical strength. The timbre and loudness of the voice change when you are squeezing your core muscles with all your strength, and it might be the case that both timbre and loudness change more when you have more strength. The correlation chain would then be grip strength<>core strength<>vocal changes.
posted by clawsoon at 10:03 AM on July 24, 2018


*mournfully walks over to the "days since having a Katy Perry song stuck in head" sign*

*erases previous value, writes in 0*
posted by Reclusive Novelist Thomas Pynchon at 10:04 AM on July 24, 2018 [11 favorites]


My boss sent this to us when it came out a month ago, and I spent an enjoyable few hours playing the recordings in the supplementals and giggling to myself. Astounding.

Personally, as biogeo mentions this has indeed been exhaustively studied in nonhumans--red deer roaring, for example. The thing is, in those species, usually we are studying highly ritualized and socially structured vocalizations that appear primarily in specific forms of display.

When was the last time you heard someone stand there and roar at you? In any context? How does the perceived social unpredictability of someone who's mad enough to make that kind of vocalization weigh into your perception of threat--and how does sex weigh into that, given relative likelihoods of attack by men versus women?

It just struck me as totally bizarre to see those considerations omitted from the article.
posted by sciatrix at 10:07 AM on July 24, 2018 [7 favorites]


It's also weird to focus on relative threat levels by measuring bicep size and grip strength, given human attack styles. Like, squeezing is not the way that most of the people I know who have tried to damage another human without tools have gone about it. Slapping, sure; punching, maybe; kicking, biting, scratching, body-slamming, generally knocking someone down and kicking is pretty nasty. Grabbing limbs and twisting them or yanking them, that's another common maneuver. And of course there are weapons. Women are especially likely to use tactics like this, in my experience--as witness that woman who got groped and who twisted around and body-slammed a dude in the FPP a few links down.

Grip strength is totally unrelated to all of these kinds of attacks, with the possible exception of grabbing a limb and twisting it. Pity, because it's like the one form of strength I'm actually any good at.
posted by sciatrix at 10:12 AM on July 24, 2018 [1 favorite]


sciatrix: The thing is, in those species, usually we are studying highly ritualized and socially structured vocalizations that appear primarily in specific forms of display. When was the last time you heard someone stand there and roar at you? In any context?

Semi-ritualized roaring as a prelude to battle is/was a thing, wasn't it? That seems like it would fit your description.

In high school football, we'd always start the game in a circle as a team doing an aggressive chant a couple of times and then a collective roar/cheer.

Do professional wrestlers ever roar before matches?
posted by clawsoon at 10:22 AM on July 24, 2018 [1 favorite]


Here and here are random football teams to give you a sense of the chant+roar if you've never been in a pre-game huddle. There's often a call-and-response statement of purpose (typically "band of brothers" and "never give up" stuff), followed by a more straightforward roar, "ooh-ooh-ooh", "charge!", etc.

You might find it interesting from the highly-ritualized-and-socially-structured point of view. :-)
posted by clawsoon at 10:37 AM on July 24, 2018


This entire line of inquiry is strange. The paper I posted earlier, Sell et al., is about how well a person can estimate another person's fighting ability by looking at a photo of their face. Why is this a research topic?

I guess it would be "unethical" or whatever to have the test subjects engage in hand-to-hand combat, so researchers are using strength as a proxy. Ok, science. Sure.
posted by chrchr at 10:39 AM on July 24, 2018


Why is this a research topic?

Signalling theory, I suspect.
posted by clawsoon at 10:43 AM on July 24, 2018 [1 favorite]


All of those things are communal shouts, though--how are you going to size up whether any given individual is a potential threat vis-a-vis you? It's really hard to pick voices out from a crowd by design during those kinds of vocalizations, since they're generally the sort of thing that people engage in to drum up team unity as much as to intimidate other teams. In that case, I suspect the intimidation factor is probably based on numbers and amplitude more than individual pitches. (IIRC, they probably were able to control for absolute amplitude in this study. I'm jealous--my animals never sing directly into the microphone!)

I think professional wrestlers are probably not very representative of any situation an individual human might find themselves in. The kinds of ritualized behaviors I'm discussing are probably best compared to replacing one's Tindr profile by screaming directly into the void, and then having everyone sort through the shouts to find people who are attractive.

....perhaps we aren't so different after all!

No, seriously, in many of these more competition oriented species where ritualized vocalizations are used in a sexual selection context, female choice might or might not exist in a clearly visible way and males are directly, one-on-one competing for territories or matings with one another rather than appealing to females. It's a less common system than folks used to think, but competition-oriented systems do exist and this kind of assessment is fairly common in them. Red deer are probably the best studied ones, as I mentioned. You'll also see these kinds of evaluative threat vocalizations at the boundaries of territorial disputes or sometimes over dominance (in the sense of access to resources) disputes.

We aren't actually one of those systems, but there you go. Human territories and strength to me seem more likely to be defended with relational strength and relative social power, given the intensely social context of humans across all known cultures. I'm really dubious about the concept of single-individual ritualized vocalizations intended to demonstrate relative physical strength, particularly because social power is so incredibly important to humans.
posted by sciatrix at 10:44 AM on July 24, 2018 [5 favorites]


Thanks for that paper, chrchr. I only gave it a quick read but I don't think it supports bicep circumference as a general measure of strength to compare females and males as it seems to be used in the FPP paper. As I told a woman friend of mine who was lamenting the appearance of her upper arms, "I wish I had arms like that!"
posted by exogenous at 10:48 AM on July 24, 2018 [1 favorite]


I guess it would be "unethical" or whatever to have the test subjects engage in hand-to-hand combat, so researchers are using strength as a proxy. Ok, science. Sure.

It's less that it's unethical, I suspect, and more that it's difficult to recruit people to sign up to be thrown into a cage match. Bear in mind, people for studies in psych departments have to come from somewhere.

That being said, I also think the field is a bit silly. This isn't the first time I've come across this sort of study, and by and large I tend to feel like the human work tends to be bad at evaluating human aggression in a species-appropriate context--probably because there's this strange desire to focus on humans in an oddly "animalistic" context without reference to aggression in other socially similar systems or what we know about human aggression in contextually realistic situations. I don't just mean contextually realistic situations for WEIRD humans, either; I would be happy to see more of this research conducted on humans whose cultures don't fit that category. But it always seems to be these kinds of very specifically envisioned contexts that I don't think have a ton of relevance outside the experimental room.
posted by sciatrix at 10:48 AM on July 24, 2018 [3 favorites]


When was the last time you heard someone stand there and roar at you? In any context?

I don't have the ability to question the worth or accuracy of this study, but I agree the setup does seem a bit odd to start with. Literally the only times in my life I've ever had someone roar at me, a movie immediately followed. I imagine that if someone roared at me in real life, I'd just get the hell away from the crazy person before I even bothered to consider their perceived strength.
posted by Greg_Ace at 10:50 AM on July 24, 2018 [1 favorite]


I don't know about the rest of you, perhaps modern society has lulled you into a Disney-like sense of constant security, but I've had people roar fiercely at me and then expect a fight more often than I can remember. Mostly in the judo class I took in middle school.
posted by sfenders at 11:16 AM on July 24, 2018 [1 favorite]


Personally I wonder how strong people think this guy is.
posted by biogeo at 11:19 AM on July 24, 2018


sciatrix: In that case, I suspect the intimidation factor is probably based on numbers and amplitude more than individual pitches.

That's a good point. I suspect that an interesting discussion could be had about within-group vs. between-group selection in humans if you wanted to maintain the hypothesis that roaring is a useful signal for humans. You mention edge-of-territory roaring in (some) other species, and that immediately makes me think of intergroup human conflicts.

Or maybe I'm just thinking of Braveheart...

I'm really dubious about the concept of single-individual ritualized vocalizations intended to demonstrate relative physical strength, particularly because social power is so incredibly important to humans.

There are cases in which people can raise their social standing by demonstrating individual strength and/or skill (see: every sport ever), but, yeah, that's only one thread in the fabric of social power.
posted by clawsoon at 11:21 AM on July 24, 2018


Or this guy.
posted by biogeo at 11:26 AM on July 24, 2018


interesting discussion could be had about within-group vs. between-group selection in humans

*visible wince* I'm not sure I can think of a sentence more finely calculated to induce headaches and cause fights among behavioral ecologists, to be honest!

(Within-group vs. between-group selection is, of course, a deeply contentious topic in its own right.)
posted by sciatrix at 11:46 AM on July 24, 2018 [2 favorites]


More generally, and building off what sciatrix said, there's a common, and fallacious, tendency towards hyperadaptationism when interpreting this type of research. So if we do take the results at face value and accept that this is evidence that humans are capable of inferring strength and size from the sound of another human's roar, how do we interpret this? Raine's nontechnical article suggests that this is why humans roar: to produce an honest signal of strength. But that simply doesn't follow. A claim of evolutionary function is actually a really strong claim which should at least attempt to meet certain standards.

If humans roar to honestly communicate strength, then the following things should be true:
  • The relationship between vocal features in roars and strength should form an honest signal. There should be mechanistic constraints on the ability to produce strong-sounding roars without actually being strong. This is plausible, but not really addressed in the paper.
  • The communication of that signal should have a positive fitness consequence, and the failure to roar should have a negative fitness consequence. This is extremely hard to test directly, so the best way to address it is by analogy to other animals for which we can observe or estimate the effects on fitness of communicative traits more directly. As sciatrix points out, though, it's not clear that the animals we're looking to in these cases are really good analogies for human social structures.
  • On the receiver side, the ability to accurately estimate strength on the basis of vocal features in a roar should also have an effect on fitness.
  • The trait of roaring should have a clear evolutionary history. It's reasonable to postulate that that human roaring may be homologous to certain vocal displays in nonhuman primates, particularly in other great apes. If this is true, then roaring-as-strength-signal may be true in them as well. We should be able to interrogate this using field experiments, using audio playback and observing behavioral responses, similar to Cheney & Seyfarth's experiments in vervet monkeys. If the roaring-as-strength-signal trait is an evolutionary innovation in humans not found in other primates, there should be an evolutionary mechanism to account for that.
  • Manipulating the relationship between vocalization and strength should have a significant effect on the behavior and physiology of the listener. E.g., changes in the decision to engage, differences in galvanic skin response due to different levels of parasympathetic arousal, etc. These changes should show the same pattern as when the same information is communicated using another modality (e.g., vision).
  • The roaring-as-strength-signal trait should not simply be a phylogenetic constraint, that is, something that evolved earlier in the primate lineage which we retain but which serves no particular function in modern humans.
posted by biogeo at 11:59 AM on July 24, 2018 [4 favorites]


biogeo: More generally, and building off what sciatrix said, there's a common, and fallacious, tendency towards hyperadaptationism when interpreting this type of research.

There'd also be the challenge of explaining why it would be advantageous for women to overestimate the strength of men.

I think that it's a potentially interesting result even if you don't care about the (evolutionary or not) cause. Is this result part of a larger pattern of women overestimating the physical strength of men? Is the author correct to suggest that it has implications for self-defense recommendations for women?
posted by clawsoon at 12:29 PM on July 24, 2018 [2 favorites]


sciatrix: *visible wince* I'm not sure I can think of a sentence more finely calculated to induce headaches and cause fights among behavioral ecologists, to be honest!

Heh.

Do we know of any species other than humans where group roaring is used in ritualized-territory-dispute (or similar) situations?
posted by clawsoon at 12:36 PM on July 24, 2018


(I'm not thinking of "we are here" group roaring, but more specifically of "we are about to have a fight with you, let's roar first and see if you still want to fight" group roaring.)
posted by clawsoon at 12:39 PM on July 24, 2018


When was the last time you heard someone stand there and roar at you? In any context?

oh god, they don't stand still to do it and it's been years since I lived in a place this was popular, but more groups of repulsive young men have made long wordless guttural bellows at and around me than I care to count. it's one of the reasons that using "catcalling" as an all-purpose term for auditory harassment annoys me so much, because it makes it sound so comparatively genteel. the roaring thing isn't compliment or insult, just raw aggression in its most unpleasant possible form.

but it does always happen in groups. no solitary male yawps or barks have I heard. used to come across to me as some kind of chest-bumping equivalent, a bonding exercise of some kind. I would gladly see every roarer dead but they certainly exist. and in spite of the patronizing tone of the article, I don't believe women overestimate male roarer-strength because of some internalized inferiority complex or stereotyped assumption or whatever, but rather because in real life men only scream wordlessly at you when they want to startle or terrify you. and you don't have to have big muscles to do violence or be a threat.
posted by queenofbithynia at 12:52 PM on July 24, 2018 [8 favorites]


Do we know of any species other than humans where group roaring is used in ritualized-territory-dispute (or similar) situations?

Loads, depending on how exactly you define "roar". Howler monkeys come to mind as a particularly clear example. Perhaps arguably chimpanzees in certain inter-group conflicts.
posted by biogeo at 3:42 PM on July 24, 2018 [1 favorite]


When was the last time you heard someone stand there and roar at you? In any context?

I've observed a correlation between roaring  and having taken a lot of the wrong kinds of drugs.
posted by justsomebodythatyouusedtoknow at 3:54 PM on July 24, 2018 [1 favorite]


Gunnery Sergeant Hartman: Private Joker, why did you join my beloved Corps?
Private Joker: Sir, to kill, sir!
Gunnery Sergeant Hartman: So you're a killer!
Private Joker: Sir, yes sir!
Gunnery Sergeant Hartman: Let me see your war face!
Private Joker: Sir?
Gunnery Sergeant Hartman: You got a war face? RAAAAAAAAAGH! That's a war face. Now let me see your war face!
Private Joker: Aaaaaaaagh!
Gunnery Sergeant Hartman: Bullshit! You didn't convince me! Let me see your real war face!
Private Joker: Aaaaaaaaaaaaaaaaagh!
Gunnery Sergeant Hartman: You don't scare me! Work on it!
posted by Mr.Encyclopedia at 5:38 PM on July 24, 2018


> I've observed a correlation between roaring  and having taken a lot of the wrong kinds of drugs.

This is a good a place as any for a link to a David Lee Roth soundboard.
posted by klarck at 5:43 PM on July 24, 2018


Who of us has not bellowed a hearty “Woo hoo” in appreciation of a particularly choice guitar solo?

I can think of lots of instances of individuals spontaneously yelling, but it’s in celebration rather than for the purposes of intimidation.
posted by chrchr at 6:02 PM on July 24, 2018 [1 favorite]


Interestingly, in our study women tended to overestimate men’s strength. That is, female listeners tended to rate male vocalisers of similar strength to themselves as stronger than them. Even in cases where males were weaker than female listeners, they were only correctly identified as weaker 25% of the time.

Gee, I wonder what possible reason might women have for presuming a generally higher level of potential violence and aggression from male voices.

SO INTERESTING, THIS FINDING.

Huh.

Maybe this could be useful in something having to do with teaching women to fight off rapists?

Like, women don't even know their own strength.

That probably explains it.
posted by desuetude at 8:18 AM on July 25, 2018 [2 favorites]


« Older "The stage is a living-room set: couch, TV, coffee...   |   How I've Navigated Sex And Consent As A... Newer »


This thread has been archived and is closed to new comments