Alfred Yarbus' Eye Movements and Vision

The Mark Wexler Free Library presents the seminal work on eye movements, “Eye Movements and Vision” by Alfred Yarbus,1967.


Al’fred Luk’yanovich Yarbus was born in Moscow in 1914. He was graduated from the Faculty of Physics of Moscow University in 1941 and was a scientific assistant at the Institute of Biophysics of the Academy of Sciences of the USSR until 1963. He was a senior scientist at the Institute for Problems of Information Transmission of the Academy. In 1964 he received the degree of Doctor of Biological Sciences for his work on “The Role of Eye Movements in Vision.”


Having recently read this book I wanted to share a few thoughts on it.

First of all, this book was very difficult to read. I had expected it would be a sociological treatise, but in reality it was heavily scientific, and not in a good way. What I mean by that is that it was very technical, written in the sort of precise, legalistic language that you find in academia, all so that truth is conveyed and an expert wouldn’t be able to pick holes in it; but sadly with the consequence that a layman is left infuriated.

I would not recommend anyone read this book. Instead I would suggest simply digesting the information provided in Anthony’s article, given just below, as I found there was very little more to learn than what I had already gleaned from said article.

If you must however, read chapter 7, pages 171-196. After that Chapters 3 and 4 are quite relevant.

What follows are a few nuggets of information I thought interesting, besides what I already knew from the article above. I take no responsibility for the fact I might have misunderstood certain bits. This is very complicated I’m afraid!

Chapter 1

1) To understand vision it is not enough to understand the eye. One must understand eye movements as well.
To give you a metaphor, suppose a TV channel is covering a soccer match. They take their cameras to the game and set them up. However, it’s important that throughout the game those cameras move according to where the action is.

This may seem a simple point, but there is an important flaw in this metaphor. Whilst we can appreciate that great technology is needed to produce a TV camera, the act of moving it about seems easy.

With the human eye this is not so. Like the camera, the eye is very complex. But the movement of it, unlike with the movement of the camera, is also highly complex and not easily understood. Like the eye itself, eye-movements are also a great feat of nature.

How the eye moves is largely an unconscious activity. These movements however are an integral part of the processing of visual information. Subconscious eye movements are just as ingenious as the eye.

2) A stationary retinal image disappears after 3 seconds.
This sounds ridiculous but is testimony to the fact that our eyes move around intensely. If you can genuinely get someone’s eyes to stay still, they will fail to see anything after 3 seconds (provided also that the scene is motionless). It’s a bit like when you’re computer automatically switches off because you haven’t pressed a button in a while.

If you put coloured paper on your nose, you can put your retinal image out of focus. This makes it easier, apparently, for the vanishing act to occur.

3) If the angular dimensions of an object are smaller than the eye’s involuntary movements, the observer has difficulty seeing it.
This is an incredibly complex issue, but one worth consideration.

Let’s ask the question why we can’t see an electron. Suppose we have a single electron in space and ask ourselves what are the impediments to seeing it?

First of all it probably doesn’t produce enough light to register with a receptive cell in the eye. But let’s ignore that and suppose it is an electron giving off huge quantities of light.

The next issue concerns the size of a receptor cell in the eye. Let’s think of a receptor cell as being a darts board. Now whilst the electron would have enough energy to hit a bull’s eye onto the receptor cell, the receptor cell would not be able to read this as a bullseye. Instead it would just know that the electron hit the dart board as a whole.

The upshot of this is that the eye would potentially see the electron, but it would see it as much bigger than it actually was.

But that’s not the whole story. The receptor cell is really composed of two dart boards. Or at least an inner dart board and an outer dart board. Now whilst the receptor cell can’t tell if a bullseye has been scored, it can tell whether the electron hit the inner board or the outer board. So in this example the receptor cell would know that the electron hit the inner board.

What the eye wants to do next is make an involuntary small movement, to see if it can get the electron to throw out another beam of light that this time will hit the outer board of the same receptor cell. The eye now switches off and jumps to another location and switches back on again.

However in this new location, the electron no longer hits the board. Not the outer board or the inner board. Nothing. Because the electron is now ‘miles away’.

This mechanism of the eye, of switching off and jumping locations, is a mechanism that allows it to ignore being triggered by small objects. All objects which are worthy of consideration to the eye, must be big enough that they can survive this jump in perspective, and register on the dart board on both goes.

If the eye did move continuously, its scanning systems would continuously be triggered by a litany of small objects. By scanning an object at a discrete bunch of locations, separated from each other by a minimum distance, small irrelevant objects are excluded from view.

NB, this interpretation might be wrong, but the bottom line is that this is a very complex process to try and understand.

EDIT: According to Wikipedia there are such things as smooth eye movements used when the eye tracks a moving object. I am therefore not sure if this contradicts what I have said.

4) You can see the shadows of your own eye’s blood vessels.

Chapters 3 & 4

1) Eye movements consists of points of fixation and involuntary jumps known as saccades.

2) We see nothing during one of these involuntary jumps. They are very fast.

3) The purpose of an involuntary jump is to put upon the fovea (the centre of the eye) a novel and particular element of the object being looked at.
This is important. It means that these involuntary jumps are not random. Rather they are done so as to put the eye into a specific spatial relationship with a specific part of the object. But to repeat, these involuntary jumps are subconscious, again speaking to the cleverness of the eye-movement system.

4) Saccades are equal for both eyes.
Thus these involuntary jumps seem to happen to both eyes equally and simultaneously. There is another involuntary movement of the eye, known as drift, that affects each eye differently.

5) Left to their own devices, eyes do not follow lines.
This was spelt out in Anthony’s article but is worth repeating. It is a huge misnomer to imagine that when viewing a painting, an observer’s eyes follow the lines of the artwork.

Chapter 6

1) Without a moving object, the systems of the eye used for tracking moving objects cannot be consciously put into use.
This is very important for anyone hoping to portray motion in an artwork. Unfortunately evolution is against you. Because unless something actually moves, the human system of following an object will simply not kick in. This will be a huge clue to the viewer that what they are seeing is not really moving.

Chapter 7

1) When viewers are shown a painting, and then the same painting in black and white, how they examine the painting is the same.

2) Viewers are drawn to people, and particularly to the eyes and mouth.
The mouth tells of mood and gives cues about health. It also helps decipher language. The eyes are the mirrors of the soul and tell you everything about a person’s mood and intentions.

3) Individuals differ at how they look at a painting.
There is not a one size fits all unfortunately. There are however common trends, like the eyes and mouth business mentioned above.

4) Should we give viewers a primer, a task, before looking at a painting?
As with many things in art, perhaps we need to change the conventional set up, if we really want to recreate reality when a viewer interacts with an artwork. By giving a viewer a mission, before letting them see a piece, a lot of their evolutionary systems would then kick in, which would perhaps heighten the realism of the painting.

I suppose in some ways titles can have this effect, but no one is obliged to read them before looking at an artwork.

5) Many elements of a painting are not given foveal (central) vision but only peripheral vision. An artist could take advantage of this. For example, a man hiding in a dark corner of a painting might not be seen until the tenth viewing of the artwork. Or even just when pointed out to the viewer. This could give them a fright.

6) Bright objects appear bigger than dark objects of the same size.

1 Like

Thank you so much for this thoughtful and insightful glimpse into your experience with this book Thomas. It’s no joke that it is a pretty dry and technical read—but like you point out—there are some truly valuable insights there for the taking.

One idea in regards to " 2) We see nothing during one of these involuntary jumps. They are very fast." that you may find interesting is often referred to as “Chronostasis.”

Almost everyone has had the slightly odd experience of looking at their watch, and believing that the watch had stopped. Then, after a perceptibly longer time than a second , the seconds hand of the watch (or its digital display) starts moving again and all subsequent measured seconds last for, well, a second.

This well known effect happens because the brain fills backwards in time the period of time when it was blind with more of the same, to wit, with the image it saw first when the cascade ceased. Some argue against this explanation for chronostasis by pointing out such experiences occur across different modalities (like the “dead phone illusion”)—but I think it’s still a pretty cool idea to kick around that ties in which saccadic blindness.

1 Like

Thanks Anthony. Wow that is an interesting hypothesis. It’s actually quite scary when you start to think about it! I guess in some ways that hypothesis would suggest there is a delay on what we see and that we are not quite ‘in real time’. Although that has some holes in it as well, namely that if the brain can ‘fill backwards’, why does it not just operate on real time in the first place.

EDIT: Although actually because of processing, it’s clear the brain can’t run on real time.

I don’t think we’re on real anything. LOL! But it’s super fun to ponder these things.

1 Like