Friday, May 6, 2011

Gaze Direction

Which way are these eyes looking? Do they seem to be looking right at you, or off to the side?

Our perception of gaze direction is influenced by the position of the iris inside the visible surface of the sclera (the white of the eye), but that’s only part of the story.

In 1824 William Henry Wollaston made exact duplicates of an engraved plate of eyes and eyebrows. He then placed them in two different facial contexts. One face is turned one way, and the other points in the opposite direction.

Surprisingly the exact same set of eyes appears to be looking in different directions solely because of the surrounding facial cues.

Even if you take a matched pair of eyes and eyebrows and just shift the nose beneath them from one side to the other, you can shift the apparent direction of gaze.

Why this happens is still not completely understood. Ophthalmologists Michael F. Marmor and James G. Ravine, authors of the new book “The Artist’s Eyes: Vision and the History of Art,” suggest a psychological cause: “Our judgment of the direction of someone’s eyes is linked, in part, to the direction we believe that person to be looking.”

This might explain why cartoon eyes drawn "crosseyed" still seem to be looking at us.
The Artist’s Eyes: Vision and the History of Art by Marmor and Ravine
William Hyde Wollaston, Apparent Direction of Eyes in a Portrait
Donald Duck is copyright Disney.
Don’t forget to vote in the video contest (scroll down to view the videos).


LandPainter said...

Wow! Very neat observation.

ivanka k said...

In other words nothing what we see is really real. It's scary how much things become subjective, eyes are very complex mechanism indeed yet they still have a nonchalant way of cheating us.

kev ferrara said...
This comment has been removed by the author.
jake gumbleton said...

that's amazing, where do you find this stuff!?

James Gurney said...

Jake, I was just as surprised by this stuff. I learned about it from reading the book called "The Artist's Eyes," linked at the end of the post. I couldn't find much about it online.

Ivanka and Kev, yes, apparently there's a lot more subjectivity involved here that meets the eye.

Hippo said...

I'm delurking after a couple months of reading your blog (and quite a few years of generally being a fan) just to say a big ol' thank you!

Today a patron came into the library where I work and asked for art books. She was going to teach herself to paint and wanted tips and inspiration - so after helping her choose a stack of books I showed her this blog, and she was thrilled!
There's so much fascinating material here, and so much art stuff I probably wouldn't have had found otherwise.


irinapictures said...

James, I am completely agree with Flodo. Your blog keeps the highest level, informs, teachs, surprises, presents ideas. Amazing. Thank you so much!
Eyes! How could guess??

Sarah Stevenson said...

Fascinating! Before I read your post, I was speculating that maybe each eye was pointing in a slightly different direction and that was why the overall direction felt ambiguous (like when you're talking to someone with a glass eye or who's blind in one eye).

Adam Nowak said...

This is an interesting entry. It's been really buggin me why quite a few drawings and paintings I see seem to have the eyes ever so slightly cross eyed but where if you do a double take they look fine. I've seen it in drawings and oil paintings. I thought that they may have just been mistakes made by the artist, but I see it every so often to think it really might have been done on purpose. One example is Vanderpoel's drawings. I couldn't find the actual drawing online so I scanned it in myself from my copy.

Another one is Sargent:

Bruce Jensen said...

utterly fascinating, what a great blog! thanks!

Anonymous said...

Fascinating! Wow...

James Gurney said...

Thanks for the kind words everyone, and welcome. No such thing as a lurker here--It's totally fine to read without leaving comments.

Here's another weird thing about that second illustration. If you look at the two "faces" with the eyes appearing to look in different directions, and then cover up the bottom part of the face, the two sets of eyes keep on looking in different directions, every with those accessory clues removed. Very strange!

Carolyn A. Pappas said...

I found this interesting because my own eyes often cross (without my knowledge) and people sometimes ask me what I'm looking at.

jonathanpaulmayer said...

I think it has more to do with geometry than psychology. The surface of the eyeball is curved, not flat. The iris is also curved—but inward—which compounds the subtleties. A slight change in the angle between the viewer and the subject's eye will cause the relationship between pupil, iris, and sclera to change. Thus, we should be able to tell which way the eyes are pointed even without context.

For me, when I physically cover the features in the second illustration, I can only see the face as pointing left. I think this may be because of the shading beneath the eyebrows. But the context of the nose pointing the opposite direction must override the subtleties of the eyeball and brows.

Erik Bongers said...

This has an important implication for photoshop painting and, to some extend classic painting.
In photoshop you can easily zoom in on the eyes to work on the details. However, for the direction of the eyes, you really need to zoom out again.
As I said, this is probably also true for fairly large size portrait paintings. Often step back when working on the eyes!

Erik Bongers said...

Aha! There's another reason!
For the first example, I noticed that there's something confusing with these eyes anyway. Regardless the face. Couldn't quite nail it down.

But I just played around a little in Photoshop with it and I discovered that the left eye looks to the right, while the right eye looks straight at you.

My guess is that the eye that is closest to you is the one that you use to determine direction of gaze.

So, in the left example, the eye closest to you is the left eye - which is looking to the right.
In the right example the right eye is closest, and this one is looking straight at me.

Problem solved? Or not quite?

Erik Bongers said...

Just checked: the same goes for the 2nd example, although not quite as strong.
If you isolate the eyes, make sure to mask or erase the shadow of the nose as well, as it's already enough a hint to be able to guess the direction of the face.

Unknown said...

I was just thinking about this recently about the cross-eyed thing in cartoons and why it works so well. I guess it's like the character is focusing on us as if we're close to them. Our eyes get crossed when we focus on things up close.

James Gurney said...

Raphael Kretz doesn't have a Google account, and wanted me to share his thoughts:

"The first thing that sprang to my mind at the copied eyes in different heads is how things relate to their place in the bigger scheme:

"Eyes focussing on something near have that slightly convergent orientation (for exactly parallel would be a gaze towards the horizon, "through a person", or staring like a hypnotized bunny), just as the donald duck picture you posted. so, in most cases, eyes will have their own hint of directionality. wollaston's set of eyes obviously veers towards the right of the picture, having more white to the left of the iris in both eyes.

"Now if you place this slight to-the-rightness onto a face that is already looking to the right, its pronouncedly looking to the right - like at someone whos standing to our right, ignoring us. if you place those same eyes onto a head that indicates looking to the left, the head is indicating looking over our left shoulder, but he is looking out through the sides of his eyes towards (our) right - i.e. straight at us.
right + right = very much to the right, whereas left + right = neutral, so to say.

"I took your quote from marmor/ravine just that way: what our guess is what direction someones eyes are directed happens in relation to the direction their face is pointed. (that the eyes are depicted from an angle where the head orientation could be both ways, seems to strengthen the effect. eyes seen from almost-side-on seem less ambiguous)"

--Raphael Kretz via JG
Thanks, Raphael and to all of you for such interesting insights.