Cartoon Faces

In Jonathan Franzen's latest book, The Discomfort Zone [1] (highly recommended), I found a nice couple of paragraphs dwelling on the psychology of cartoon faces. Franzen is reminiscing on his fascination for Charles M. Schulz's Peanuts. Here's what he writes:

Our visual cortexes are wired to quickly recognize faces and then quickly substract massive amounts of detail from them, zeroing in on their essential message: Is this person happy? Angry? Fearful? Individual faces may vary greatly, but a smirk on one is like a smirk on another. Smirks are conceptual, not pictorial. Our brains are like cartoonists - and cartoonists are like our brain, simplifying and exaggerating, subordinating facial detail to abstract comic concepts.

Scott McCloud, in his cartoon treatise Understanding Comics, argues that the image you have of yourself when you're conversing is very different from your image of the person you're conversing with. Your interlocutor may produce universal smiles and universal frowns, and they may help you to identify with him emotionally, but he also has a particular nose and particular skin and particular hair that continually remind you that he's an Other. The image you have of your own face, by contrast, is highly cartoonish. When you feel yourself smile, you imagine a cartoon of smiling, not the complete skin-and-and-hair package. It's precisely the simplicity and universality of cartoon faces, the absence of Otherly particulars, that invite us to love them as we love ourselves. The most widely loved (and profitable) faces in the modern world tend to be exceptionally basic and abstract cartoons: Mickey Mouse, the Simpsons, Tintin, and - simplest of all, barely more than a circle, two dots, and a horizontal line - Charlie Brown.

Pop-psychology already has a theory of what makes cartoon faces attractive, but it is completely different: it was spread by Stephen Jay Gould's 1980 paper on Mickey Mouse [2]. Gould, who never refrained from telling a just-so-story when he had a good one in store, proposes that the reason why we find Mickey Mouse attractive is due to our innate attraction for all things baby-like. Gould dwells on Konrad Lorenz's theory of neoteny. Neoteny is the set of facial characteristics peculiar to babies. The theory states that, in the course of evolution, our species (and many others besides) have evolved neotenic features for our youngsters, in order to tap this innate mechanism that attaches us to round faces, big eyes and soft features - what so many languages have a word for: the cute, the kawaii, the mignon, the moutik. Gould spectacularly illustrated his point with the ontogeny of Mickey Mouse, who evolved, in a somewhat spooky trajectory that reminds me of Mickael Jackson, from a real character with features and peculiarities to a big simplified balloon of niceness.

The evolution of Mickey Mouse, reproduced from Gould's paper.

To give these observations the cachet of quantitative science, I applied my best pair of dial calipers to three stages of the official phylogeny--the thin-nosed, ears forward figure of the early 1930s (stage 1), the latter-day jack of Mickey and the Beanstalk (1947, stage 2), and the modern mouse (stage 3). I measured three signs of Mickey's creeping juvenility: increasing eye size maximum height) as a percentage of head length (base of the nose to the top of rear ear); increasing head length as a percentage of body length; and increasing cranial vault size measured by rearward displacement of the front ear (base of the nose to top of front ear as a percentage of base of the nose to top of rear ear).

All three percentages increased steadily--eye size from 27 to 42 percent of head length; head length from 42.7 to 48. 1 percent of body length; and nose to front ear from 71.7 to a whopping 95.6 percent of nose to rear ear. For comparison, I measured Mickey's young "nephew" Morty Mouse. In each case, Mickey has clearly been evolving toward youthful stages of his stock, although he still has a way to go for head length.

The theory of Franzen and McCloud seems to me much more interesting and testable: popular cartoon faces seem popular to us not because they are particularly neotenic - many of them are not - but because they are transparent, which makes it easier to project our mental states on them. This in turn may render them sympathetic, but not because of their natural cuteness. I have aligned below some very popular comics characters - some of them arguably kawaii, like Astroboy and Dora, some others not so much - it's difficult to be less neotenic than Fido Dido, that adolescent growth spurt made into an advertisement icon.

This is all anecdotal evidence, of course, but it would be quite easy to test - does the sympathy a character elicits vary in function of its neotenical features or in function of its simplicity? The experimental testing seems straightforward. I also bet we could find a similar dynamics in other artistic fields - I remember my professor of Commedia Dell'Arte explaining why the good guys have the simpler masks (or no mask at all), using arguments similar to Franzen's.

One might also argue that these simple faces owe their success to their simplicity itself, which makes them easier to process and to memorise, a crucial feature for a character who struggles for life in a market full of other interesting comics characters. That would be another story, but not, I suspect, an entirely different one.


[1] Franzen, J. (2010). The discomfort zone: a personal history. Farrar, Straus and Giroux

[2] Gould, S. J. (1979). Mickey Mouse meets Konrad Lorenz. Natural History88(5), 30-36.

4 Comments

  • Lucien Dontask 8 December 2007 (00:00)

    Dear Olivier, This hypothesis looks indeed more promising than the neoteny hypothesis by Gould. What seems to me incongruous is the following: if we find simple expressive faces with facial expressions that are easy to read so attractive, why then are such representations quite rare in early art or in children’s art? The average 4 to 5 year old draws humans with static, almost hieratic features. Similarly, if we think of Palaeolithic human imagery, facial features are even almost entirely absent. And in later art (Egyptian, Greek, Mesopotamia, Islamic miniatures), faces are usually emotionless. Indeed, expressing emotion in faces only appears in very specific periods of art history (e.g., late Hellenic sculpture). So, my question is then, why are cartoon faces so time and culture bound? (After all, the manga did not originate in Japan, but was inspired by American comics).

  • Lucien Dontask 10 December 2007 (00:00)

    Evening Helen, it’s nice to chat with you! Why are emotional faces so rare in general art history, and in children’s art? Well, I’d say, quite simply, because they are so difficult to draw. The fact that some cultural invention is attractive and easy to like does not make it easy to produce. In order to draw an emotion you have to render the coordinated activity of so many muscles... The first emotion psychologists, like Darwin or Duchenne de Boulogne, often criticized the art of their time for being unprecise. Now, try to make your drawing not only accurate, but fast, easy, standardized, easy to recognize and reproduce... That’s next to impossible when you have no model. That’s why Charles M. Schulz is a genius - and, come to think of it, smileys might be one of the most precious conquests of civilization! ;o)

  • Olivier Morin 28 December 2008 (16:16)

    Why are emotional faces so rare in general art history, and in children's art? Well, I'd say, quite simply, because they are so difficult to draw. The fact that some cultural invention is attractive and easy to like does not make it easy to produce. In order to draw an emotion you have to render the coordinated activity of so many muscles... The first emotion psychologists, like Darwin or Duchenne de Boulogne, often criticized the art of their time for being unprecise. Now, try to make your drawing not only accurate, but fast, easy, standardized, easy to recognize and reproduce... That's next to impossible when you have no model. That's why Charles M. Schulz is a genius - and, come to think of it, smileys might be one of the most precious conquests of civilization! ;o)

  • Olivier Morin 2 February 2009 (11:04)

    Thank you Roberto for your interest in this post. Actually I suspect that your rejection of the simplicity hypothesis in favour of the neoteny hypothesis springs from a misunderstanding. You argue that cartoon faces are not simplified representations of real faces, but extremely stimulating visual objects. I fully agree with you that cartoon drawers are not simplifying their perception of real faces when drawing: they are not simplifying anything. But cartoon faces may be more or less simple without that simplicity having anything to do with a simplification of real features. The Nike logo is simple, yet it is not a simplified version of a natural object (or if it does, it doesn\'t owe its success to this). Simple, but not simplified, cartoon faces would have more success than less simple faces; in my view it\'s because they are easy to process and consequently likeable. In Franzen and McCloud\'s view it\'s because they are easy to identify with. Both these theories may be wrong, but they are real theories. On the other hand, the weakness of the neoteny hypothesis is, in my view, blatant when you consider that 1) many popular cartoon faces are the opposite of neotenic (consider Fido Dido) and 2) many cartoon characters that are not supposed to attract sympathy and never become popular typically have all the features of a cartoon face (consider the extremely neotenic Rastapopoulos).