Aug 18 2009

Recognition of Facial Expressions Not Universal

A recent study suggests that the recognition of certain facial expressions of emotions may not be universal, as was previously thought. This finding, if confirmed, raises some interesting questions.

Sociological studies have shown that in humans facial expressions and their recognition are universal – an American can easily interpret the expressions of an Australian aborigine. But more careful recent study suggests that there are differences.

Specifically, Rachael Jack studied 13 caucasian and 13 east Asian men and women for their ability to identify fear, surprise, anger, and disgust in a standardized set of facial expressions. They found that the caucasians had no difficulty, while the east Asians confused surprise with fear and anger with disgust.

The reason for the disparity is also interesting – Jack found that the Asian subjects focused their gaze on the eyes, while the caucasians shifted their gaze between the eyes, the mouth, and other parts of the face. Therefore, caucasians seem to be taking in more emotional information from the whole face, while the asian subjects were relying mostly on the eyes alone.

To follow up on this idea computer modeling was used to see what part of facial expression are most reliable, and they found that the mouth is often better than the eyes for distinguishing surprise from fear and anger from disgust.

So the evidence all seems to fit – Asians rely on the eyes to read emotions and the eyes alone can be misleading.

That Asians rely more on the eyes for reading emotions, it has been noted, is also reflected in their emoticons – those cute little facial expressions made out of keyboard characters that are so useful for online communication.

Here are the western emoticons for happy, sad, and surprise respectively:

: )     : (      : o

these are meant to be mentally rotated 90 degrees clockwise.

Here are the Asian emoticons for happy, sad, and surprise:

(^_^)    (T_T)   (*_*)

The obvious difference is that the western emoticons vary the mouth while the Asian emoticons vary the eyes.

Now the big question is – what are the roots of these difference? Are they genetic or cultural. Jack seems pretty certain that they are cultural, but plans to test this hypothesis by studying Asians raised in the West.

I am uncertain if this difference is cultural or genetic. Asians and Europeans have been genetically separated (although not completely isolated) for thousands of years, which seems long enough for this kind of difference to evolve. So a genetic difference is plausible, but so is a cultural difference. We can only await the evidence to see which hypothesis is more likely to be true.

The related question is why would this particular difference arise. It could be random genetic or cultural drift, and not specifically adaptive. But the recognition of human emotions seems to be a highly adaptive trait and therefore it is reasonable to think about an adaptive cause.

Jack has a hypothesis:

East Asian cultures tend to frown on the display of negative emotions in public, Jack says. It’s possible that east Asians have learned that by paying close attention to another person’s eyes, they can spot facial giveaways of muzzled feelings of disgust or fear, she says.

That’s interesting. Would, then, westerners have more difficulty detecting hidden emotions in and Asian culture? Is it more difficult to hide these emotions from an Asian subject? These sound like testable questions.

I look forward to seeing how this fascinating research progresses.

22 responses so far

22 Responses to “Recognition of Facial Expressions Not Universal”

  1. Diogoon 18 Aug 2009 at 8:57 am

    It seems to me that the more straighforward hypothesis to raise is that the difference is genetic. If it was cultural you’d expect a big variation among the Asian subjects, since their “environments” are almost certainly very different. Anyway, a study with 26 people isn’t sample enough to suggest much. I’m surprised about all the media coverage of this study, more so that the issue of the small sample is buried deep in all the stories I’ve seen, if at all mentioned.

  2. neokortexon 18 Aug 2009 at 9:29 am

    My problem with the study as it is described is how the terms “Asian” and “Western” or “Caucasians” are defined, especially when used as an adjective for “culture.” I know Westerners like to lump all “Asians” together but do they have enough cultural artifacts in common to say the observed differences are socially constructed? If the emotional cues really do differ among (narrowing this down a bit) east Asians and American/Europeans as a group, would that not tend to indicate a genetic differences? Even in Western cultures there are cultural variations. The British are often considered reserved and stoic in public, while Americans are often considered obnoxious by just about everyone else East and West, so we obviously don’t have problems expression any emotions in public. What I’m saying here is that the tendency to stereotype should be considered a variable in the experiments. It would be interesting to have the studies duplicated by various East Asian scientist as well. And what about the differences between public and private displays of emotions? Shouldn’t the cues work either way? Other questions to consider: If the differences are genetic, then it follows they would tend to become reflected in cultural make-up. Conversely, if a purely cultural trait stays around long enough, would selection eventually tend to favor individuals better suited to conform, leading to a genetic variation?

    And as far as the emoticons you described, these obviously apply to the Japanese at least, as anyone who has played a Japanese RPG video game can recognize (^_^)

  3. JPentonon 18 Aug 2009 at 9:44 am

    It seems cultural to me.

    I thought about Japanese animation and Manga when I read this. The characters all have very large (some might say western) eyes that are extremely expressive.

    I always wondered why they wouldn’t draw characters more in their own image and I think this study makes that more clear.

    The “eyes” have it.

  4. Justin L.on 18 Aug 2009 at 10:18 am

    This study doesn’t necessarily suggest that facial expressions are not universal. If anything, it seems to suggest that some cultures merely place greater emphasis than others on facial expressions that are nevertheless universal. For example, I have no problem interpreting the meaning of the Asian emoticons even though I have never seen them before, and I suspect this is also true of Asians with respect to western emoticons. Still, I find the study interesting if not surprising.

    Bye the way, I only discovered this blog last month, and it is already the one I check first each day.


  5. mindmeon 18 Aug 2009 at 10:25 am

    I noticed in the west we talk about a “winning smile”. Good teeth are important in the west. During my time in Korea, Koreans were all about eyes. Making their eyes bigger was a popular form of cosmetic surgery. I noticed my Korean GF’s nephew would always try to make his eyes as big as possible when he looked at me and wanted me to carry him on my shoulders.

  6. durnetton 18 Aug 2009 at 10:41 am

    I guess this means that the TV show Lie To Me is not going to have a big audience in Hong Kong.

  7. Steven Novellaon 18 Aug 2009 at 11:32 am

    Justin – I agree, which is why I was careful to write that the “recognition of” facial expressions may not be universal.

    We can speculate endlessly about the culture vs genetic debate, but on this one I think I’ll just wait for the evidence.

    I also agree that we are dealing with nestled categories (Asian vs Chinese, for example) and so there are subquestions about how much this applies to subgroups, and whether those subgroups vary more by culture or genes.

  8. Justin L.on 18 Aug 2009 at 12:14 pm

    Point taken. I obviously wasn’t reading very carefully.

  9. Michael Hutzleron 18 Aug 2009 at 1:21 pm

    It’s very interesting and some cultural variations show: “East Asian cultures tend to frown on the display of negative emotions in public…” Apparently they don’t (at least in public). There might be some linguistic analysis that could support these concepts. Do the frequencies of facial features in idioms (e.g. “frown upon” vs. “look down upon”) differ?* There are measurable, testable questions.

    Additionally, the epicanthic folds found in the East Asian population offer a potential anatomic explanation. Any studies will have a challenge sorting cultural from developmental aspects. Do the faces you scan while growing up impact the features you key on. Again, this is a testable question.

    *I’m aware my second example does not mention the feature the eyes directly, but a thesaurus search turned up no matches to “frown [up]on” which did. There are English expressions, like “pull the wool over their eyes”, which do.

  10. artfulDon 18 Aug 2009 at 1:24 pm

    In every culture we learn to mask our “natural” expressions so as not to offend our various masters. In what were formerly known as Oriental cultures, these masks were historically more extreme than those normally required in the more Western areas.

    There was also the tendency in some cultures or their social strata to conceal true feelings through exaggerated expressions.

    And note that here in the US there was that same masking effect observed in Southern slave cultures, which has still carried forward in parts of that area today.
    Plus in each culture there are ways of masking emotions that are mutually acceptable and thus interpreted accordingly, where the same maskings in a different culture will be in themselves offensive.
    So yes, the differences observed are overwhelmingly cultural.
    (Not that some microevolutionary changes here and there would have been impossible.)

  11. mindmeon 18 Aug 2009 at 2:04 pm

    ||It’s very interesting and some cultural variations show: “East Asian cultures tend to frown on the display of negative emotions in public…” Apparently they don’t (at least in public).||

    Which reminds me, in Korea (and I think Japan), smiling was usually a response to shame or embarrassment.

  12. Willon 18 Aug 2009 at 7:54 pm

    Steve. Your articles are top notch science reporting. You should be writing for newspapers and science mags. Why hasn’t this happened? 😛

  13. ausduckon 19 Aug 2009 at 2:53 am

    Interesting and thought provoking. Easy to see the cultural/genetic causes. But I think there may also be some anatomical and anthropological causes as well – would be great to see a bigger study that looks at all of these aspects.

  14. SteveAon 19 Aug 2009 at 7:17 am

    mindme “Which reminds me, in Korea (and I think Japan), smiling was usually a response to shame or embarrassment.”

    I remember reading (years ago) about a study that measured the reaction of western and eastern subjects to sexually explicit and/or violent images. Essentially the subjects were put in a room and given some very strong hard-core porn to watch. Some people were told they were being observed while they watched, others that they were unobserved. In both groups the westerners tended to react with the same facial expressions (ie they behaved the same if they thought they were being observed or not) – mostly betraying distaste. In contrast, the easterners (Japanese, I think) exhibited the same facial expressions of distaste as the westerners if they thought they were unobserved, but if they were in an ‘observed’ group most of them put on tight grins and did a lot of vigorous head nodding. To western eyes the ‘observed easterners’ looked like they were really enjoying their porn-fest, wheras the opposite was true.

  15. SteveAon 19 Aug 2009 at 7:19 am

    Sorry, re. my previous post, I forgot to mention that the subjects were alone while watching their movies.

  16. HHCon 19 Aug 2009 at 1:35 pm

    Based on the basic studies in personality, it has always been difficult to identify cross-cultural emotions from a print picture. Human Genetics is the wrong direction for studying identification differences in emotions. Such researchers have their own preferred genetic racial bias when developing thoughts for hypotheses. Cultural diversity is interesting, as shown by cultural patterns of recognition in faces. Medical Biological bases of pattern recognition are the same in the human family.

  17. daedalus2uon 19 Aug 2009 at 4:09 pm

    Visual recognition of an emotional state is communication. All communication is the communication of mental states, including the communication of emotional states. All communication requires the translation of the mental state into a communication medium, transfer of the data stream of that communication medium to another individual, and then conversion of that data stream back into a mental representation of the initial mental state.

    For a communication mechanism to be “innate”, the internal mental representation of the mental state being communicated must be “innate” also. If the internal representation of the mental state is learned, then mechanisms to communicate that mental state must be learned also because it must have a one-to-one correspondence with the internal mental state and can only develop after the neuroanatomy that supports the internal mental state develops.

  18. HHCon 19 Aug 2009 at 11:43 pm

    Emotions expressed in print pictures range along a continuum. How expert is the photographer in catching an emotional moment in time with an actor? “Mistaking” fear for only surprise and disgust for supposed real anger implies a lack of intensity in the actor’s part or the photograher. If we wish to comment on cultural difference among groups of 13 volunteers each, why not assume that the Causacian viewers lack focus to comment on intensities in emotional spectrum, while the Asian viewers expect more intense facial expression to communicate true fear and real anger. How honest are the actors in the print pictures in portraying intense human emotion along a continuum?

  19. HHCon 22 Aug 2009 at 1:39 am

    “Asians and Europeans have been genetically separated (although not completely isolated for thousands of years, which seems long enough for this kind of difference to evolve”. If you substitute the words “racially separated” for “genetically separated”, I believe you will get a clearer image of Novella’s hypothesis. Now, if you substitute the words “culturally separated” for “genetically separated” you will have a clearer understanding of the research.

  20. daedalus2uon 25 Aug 2009 at 9:01 pm

    I came across this paper

    Which shows that there is a window of time, that if infants are exposed to monkey faces, that their ability to discriminate between the faces of different individual monkeys is enhanced. This pretty strongly indicates that there is plasticity in face recognition (and presumably communication mediated by facial expressions).

  21. artfulDon 25 Aug 2009 at 9:36 pm

    The paper is about differences in facial construction, differences in ethnicity, ability to recognize one’s mother, etc., etc. It’s not about the commonality of expressions among all ethnic groups that cultures may alter on the surface but not on the instinctive reaction level.

  22. juliakindleon 13 Jun 2011 at 7:30 am

    I agree facial expressions are different in people of different origin. There might be some people who could identify it correctly but most don’t. It gets even more difficult with people who are tough and can control their expressions. I have found that plastic surgery to some extent can also change these expressions is it so?

Trackback URI | Comments RSS

Leave a Reply

You must be logged in to post a comment.