The Scent of Tears
Why do people shed tears when they cry?
Scientists believe that crying evolved as a means of signaling vulnerability or distress, but what's the purpose of the tears?
In other words, when a person cries, we respond to the way they look and sound. There's a lot to see and hear. What else is conveyed by the image of water trickling down their cheek?
Well, perhaps it's not the image we respond to. A new study, published in a leading journal, seems to show that the mere scent of a woman's tears reduces male aggression.
When I first heard about this study, I was skeptical. I mean, do tears even have a scent?
As I read the study, I was impressed by its thoroughness and rigor. Statistics played a critical role in demonstrating a behavioral effect (reduced aggression) as well as its neurological substrates.
So, can we trust the findings? If so, what exactly do they tell us?
Lab studies are sometimes too contrived to be of much use. This is a common challenge for educational research, for instance, where a carefully designed intervention may work well in the lab but can't be implemented in the more or less chaotic conditions of an actual classroom.
This study actually benefits from its contrivances. What may seem like contrived and weird experimental procedures actually make the data more believable.
However, as you'll see, those contrivances do have some unexpected effects on what we can learn from the study.
The new study
This study, published just over three weeks ago in PLOS Biology, was conducted by a team led by Drs. Shani Agron (Weizmann Institute of Science, Israel) and Claire de March (Duke University).
The researchers explored two questions: Does the the scent of a woman's tears reduce male aggression? And, if so, what's the neurological mechanism?
Why even suspect that the scent of tears would have such effects?
Like other animals, humans learn about each other from olfactory cues, even when we're not aware of what's happening. For instance, studies show that people can detect fear in body odor. When exposed to the sweat of a person who'd been watching a horror movie, our brains and bodies show subtle changes associated with fear. We don't respond this way to the sweat of someone who's calm.
These findings suggest that we evolved to sense fear in others. If someone else is afraid, they probably have reason to be, and you're more likely to survive if you become a bit fearful yourself (and therefore more vigiliant about your surroundings).
As for tears, they contain hormones and other chemicals that our brains may detect, and there's evidence that tears reduce aggression among rodents.
Of course, we're not rodents. Hence the scientific need for this study.
As for the practical need, that's painfully obvious. Male aggression against women continues to be as inevitable as death and taxes. The statistics on this are abundant, nuanced, and relentlessly grim. Here's just one example: Roughly 1 in 3 American women will experience sexual violence, physical violence, and/or stalking by an intimate male partner at least once during her lifetime. Any study that explores how to reduce male-on-female violence is almost automatically worthwhile.
(Of course, same-sex aggression occurs too, and gender isn't a binary characteristic. I use binary language in this newsletter because the researchers only looked at the effects of women's tears on men's aggression.)
Methods
Agron and colleagues identified six women who could "cry at ease" and asked them to watch sad film clips. Small vials were then used to capture any tears that trickled down.
Saline was also collected after being applied to the women's cheeks.
The researchers now had two salty fluids, one consisting of tears, the other saline, both identical except for their exact chemical composition.
In the first experiment, each man sniffed one of the fluids 10 times. The content of the fluid was not described. A pad saturated with the fluid was then pasted under each man's nostrils, and the man then played a computer game. (By the end of the experiment, each man had played the game with a tear-saturated pad once and a saline-saturated pad once.)
I did mention that the experimental procedures seem contrived, right? Here are some additional details illustrating why these contrivances make for a strong study:
1. Why sniff the fluid 10 times?
After each sniff, the men quickly rated the intensity, pleasantness, and familiarity of the fluid. No perceptual differences between the tears and the saline were found.
This is important, because it means that any impact of tears would result from the hormones and other chemicals they contain. The brain would be registering the presence of these chemicals without men's awareness, analogous to the way we respond unconsciously to the smell of anxiety or fear.
2. Why paste a pad under each man's nostrils?
By saturating each pad with exactly 100 milliliters of fluid and pasting it close, the researchers guaranteed that men were exposed to tears or saline throughout the experiment, and that the amount of exposure was the same each time. Standardization in this respect increases the credibility of the findings (but also limits their generalizability to real-world settings, as I'll explain later).
3. Why measure aggression by means of a computer game?
This game, the Point Subtraction Aggression Paradigm (PSAP), is commonly used in studies of aggressive behavior. If you're interested, I explain how the PSAP works in the Appendix.
Here, you just need to know that in Agron and colleagues' study, the PSAP is played on a computer against an unseen opponent.
During the game, the men are provoked by having money stolen from them. They can choose to ignore the provocation or take revenge. Aggression was defined as the ratio of revenge responses to provocations.
The lowest possible score would be 0, which would mean that the man never deducted money from his opponent. If the man experienced 9 provocations (i.e., thefts) and he deducted money from his opponent each time, then his aggression score would be 1.0 (9 divided by 9).
Scores could also be greater than 1.0. Suppose the man experienced 9 provocations and got so aggravated he spent lots time deducting money from his opponent instead of earning it. If he deducted money 18 times, for instance, his aggression score would be 2.0 (18 divided by 9).
Surprising findings
The average aggression score while exposed to saline was 1.67. The average score during exposure to tears was 0.94.
In short, tears reduced aggressive responses by 43.7%, a statistically significant difference.
To put it concretely, in the saline condition, the average response to being stolen from 10 times was 16.7 acts of revenge. When exposed to tears, the average response was 9.4 acts of revenge, or slightly less than the total number of thefts.
This is not to say that tears made aggression stop. They simply reduced the extent of it. (Or, you might say they kept aggression proportional to the provocation.)
However you interpret the results, they're surprising. Male aggression is reduced by the scent of women's tears, even though the scent isn't consciously perceptible.
More surprising findings
Just as our taste buds respond selectively to different flavors, so we have more than 350 olfactory receptor cells in our noses (and further back in our airways) that respond to specific odors.
In their second experiment, Agron and colleagues took olfactory receptors in vitro (think of cells in a petri dish) and exposed them to the fluids used in their study. Four of these receptors responded to tears but not saline.
In other words, our nervous system is wired to respond selectively to tears. And there's more:
Agron and colleagues repeated their first experiment with a new group of men who were hooked up to a functional MRI device. (The fMRI detects changes in blood flow that indicate brain activity.) This time, being provoked increased activity in several areas associated with aggressive behavior. Tears, but not saline, reduced activity in these areas.
Not all studies are this thorough. Agron and colleagues demonstrated a behavioral effect (a reduction in aggression) as well as the neurological mechanism underlying this effect (olfactory receptors that respond selectively to chemicals in tears and message the brain to reduce aggression). Statistical procedures were essential to both the behavioral and the neurological findings.
The fundamental question
I talked to a number of people about this study. All of them, with or without prompting, had opinions on the same fundamental question: Does this actually happen in real life?
In other words, outside of the lab, does the scent of a woman's tears actually reduce male aggression?
Agron and colleagues speculated that this is a "viable possibility", but they didn't say much more than that in their article. (Senior author Noam Sobel was a little bolder in an interview with The Guardian, arguing there that chemosensory information in tears "actually lowers aggression").
Two of the experts I talked with were receptive to the new findings, but they also touched on some of the reasons why the scent of tears may have little or no impact in everyday settings.
(Please note: Some comments in the following section may be disturbing.)
Expert perspectives
Earlier this week I reached out to Liz Carrasco, clinical director at a rape crisis center in Nevada, and formerly the head of an interpersonal violence program at a university counseling center. I asked, in effect, about her gut reaction to the study.
Ms. Carrasco acknowledged some tension between what data tells us and what she has observed as an experienced clinician and trauma specialist:
[T]he scholar in me says that I need to keep an open mind toward whatever evidence is uncovered by research. However, my gut tells me that this does not translate from the laboratory setting...into human interactions.
I have known and treated plenty of clients that cried during their rape or beatings and their tears did not influence their perpetrators' aggression enough to stop what they were doing. So I do not know if they were not close enough to detect the scent of tears or if their beatings or attacks were less severe with the tears. Hard to tell...
I find it admirable that Ms. Carrasco remains open-minded about the findings, though her gut says they don't capture what happens outside the lab. She also touches on two crucial issues:
1. If the scent of tears could reduce male aggression, the man would have to be close enough to smell them. The exact distance would depend on many factors, including the quantity of tears, environmental conditions such as air flow, and whether or not there are competing scents (e.g., the soap that a woman uses). It's unclear from Agron and colleagues' study, and from the literature, what that distance might be, even under ideal circumstances.
Here's one thing that is clear: Agron and colleagues affixed pads saturated with 100 milliliters of tears under men's noses for the duration of the experiment. This seems hopelessly unrealistic. Each woman produced an average of 1.6 milliliters of tears each time they cried. The researchers then arranged for more than 60 times that amount to remain about an inch away from men's nostrils. Although this allowed for standardization, it would rarely if ever reflect real-world conditions.
2. Ms. Carrasco notes that it's uncertain whether the attacks experienced by her clients were less severe when they cried. Logically speaking, this is uncertain because we can't know the counterfactual. At the moment a specific man commits a specific act of violence against a crying woman, we can't know whether the extent of his violence would've been greater if she hadn't crying.
Ms. Carrasco added that aggression might not be exactly the right focus when we inquire about the effects of tears:
"[S]ince the continuum of intimate partner and sexual violence has more to do with power and control than aggression per se, aggression is just the tactic the perpetrators have learned has been effective in making the other person do what they want to maintain the perpetrators' sense of power, entitlement, and control over their subject. It is a tactic as opposed to the end goal."
In other words, even if the scent of tears reduced aggression, it might not reduce the overall extent of violence and domination that some men perpetrate against women. They would just shift tactics.
A slightly different perspective was provided by Carly Wielstein, a licensed marriage and family therapist who specializes in trauma. Ms. Wielstein found the results at least somewhat plausible, "especially under the theory that emotions are to a degree “contagious”."
That's an important observation, given the substantial evidence for emotional contagion – including the studies I mentioned earlier showing that subtle fear responses result from smelling the sweat of a frightened person. Chemosignals in tears (and sweat) might have a deterrent effect if they cause an aggressor to become fearful of getting caught, or causing excessive harm, etc.
At the same time, Ms. Wielstein aired an important qualification about the psychology of aggression:
"My sense is that the sight of women crying would create a diverse array of psychosocial reactions from men, especially in times of high conflict, based on any number of factors that would not be biological; i.e, perception of emotional expression, level of emotional dysregulation, power dynamics/imbalances, perception of women crying..."
In other words, men will be affected on many levels by the sight (and sound) of a woman crying. Whether their behavior changes will depend on many variables other than biological ones. Tears might cause aggression to diminish, stay the same, or grow worse, depending on the man. Meanwhile, it would be difficult or impossible to disentangle the influence of the scent of tears from the influence of the sights and sounds of the woman crying.
What does the study show?
I haven't expressed any skepticism about the data per se, because I don't have any. The methodology is strong and I trust the findings. I am persuaded that the scent of a woman's tears can reduce male aggression. In the lab. Under conditions that may never be attained the real world.
My main concern is that pads containing 100 milliliters of tears were pasted under the mens' noses for the duration of the experiment. That's about 3.4 ounces, but the women, each time they cried, produced an average of 1.6 milliliters, or about 1/20th of an ounce.
Without further data, I'm would conclude that the scent of tears reduces aggression, but only if an impossibly large quantity of them remain impossibly close to a man's nostrils.
In the real world, there are fewer tears, more distance, and a host of psychosocial complexities such as those that Ms. Wielstein mentioned.
Senior author Noam Sobel, in his interview with The Guardian, concedes that real-life effects might not be observed, but he adds that tears may play a role in protecting infants from physical abuse.
I find this deeply implausible too.
(a) Infants may cry a lot, but they're not going to produce anything close to 100 milliliters of tears at any one time. So, again, any protective function of crying probably wouldn't arise from olfactory cues.
(b) Infant cries are a known trigger of abusive behavior from male parents and other adult males. In other words, if a man is already predisposed to be abusive, crying may not protect the infant but rather increase the risk of abuse.
Bottom line
Darwin referred to human tears as "purposeless". Agron and colleagues argue that the scent of tears reduces aggression. Here's an alternative view: The perceptual characteristics of crying evolved as a means of signaling vulnerability and distress. These characteristics include the sights, sounds, and smells of people crying. They may even include the image of tears. But in ordinary circumstances, the olfactory information doesn't register.
In other words, only in a lab study like this one, where the exposure to tears is excessive, and there's literally no visual or auditory information about the crying person, can the effects of scent be detected and possibly affect behavior.
Meanwhile, the literature on intimate partner violence, for instance, doesn't show that crying is automatically helpful, as it may trigger reactive abuse or other undesirable consequences. Certainly I wouldn't infer from the new study that it's wise to move closer to an aggressor so that he can smell your tears.
Instead, if you or someone you know has concerns about domestic or intimate partner violence, you might consider calling the National Domestic Violence hotline (800-799-7233 – that's 800-799-SAFE) or visiting websites such as this or this.
The scent of crying… Its affects are another remnant of our evolutionary development, like the tails that embryos have until about eight weeks, which then become the coccyx, or tailbone. Your tailbone is always there. We all have one. We just don't notice.
Thanks for reading!
Appendix: The Point Subtraction Aggression Paradigm (PSAP).
In Agron and colleagues' version of the PSAP, there are two players.
Player A is able to steal money from player B.
Player B can deduct money from player A but can't keep the money.
All of the men in this study were assigned the role of player B. Why? Because player A doesn't exist. Player A is a computer program that's set up to periodically steal money from player B (i.e., the male participant).
The actual game was extremely simple. Briefly, the men pressed a button repeatedly to earn small amounts of money. The amount they earned was displayed on the screen. Periodically, they would notice that player A had stolen money from them. The men could choose to continue earning money, or pause and press another button to deduct money from player A.
Each theft of money is called a provocation. The program was set up so that there are many provocations. (Evil computer!)
Each deduction of money by player B is called a revenge response. This is assumed to reflect aggression, because player B doesn't get to keep the money deducted.
As I mentioned in the text, aggression was defined as the ratio of revenge responses to provocations. The lowest possible aggression score would be 0 (no deduction of money). If the participant deducted money each time he was provoked, but at no other time, his aggression score would be 1.0. Higher scores are possible too.
Methodologically, the PSAP has two advantages:
(a) No need to deal with the ethical challenges of measuring aggression against an actual person. Although human-on-human aggression is still studied in lab settings, fallout from the Stanford Prison Experiment and others like it impose many limits on what's permissible nowadays.
(b) It's hard to measure aggression. The PSAP yields a simple numerical score.
(Of course, the PSAP is far from ideal, since it relies on simply counting aggressive acts. This can be misleading. For instance, it's aggressive to poke someone in the chest with a finger in order to emphasize a point, but I wouldn't say that doing this two or three times is more aggressive than punching them, just once, in the face. As always, standardization is not very compatible with nuance.)