Woke up this morning feeling fine / The Hermann Grid was on my mind / Look where the lines cross and you’ll see a grey blob right here / But look at it closely and you’ll see it disappear
What causes the Hermann Grid illusion? Although it is almost 150 years old, we are still not quite sure. What is almost certain is that the explanation given in most Psychology textbooks is complete bunkum. The traditional story goes like this:
The retina – the back of the eye, onto which an image of the world is projected – is made up of special light-sensitive cells called photoreceptors. Obviously, these cells send signals to the brain, to tell it what we’re looking at. But these cells aren’t connected to the brain directly. Rather, cells in the eye called retinal ganglion cells pool the signals from lots of photoreceptors, and send the result on to the brain. These ganglion cells have a centre and a surround. The centre is well behaved: when it sees white, it shouts “white”. The surround is a contrary so-and-so: When it sees white it shouts “black”, and vice versa.
So – the story goes – when faced with a line (left), the centre shouts “white”. As for the surround, roughly half of it – the bit that sees black – wants to shout “white”, whilst the other half – the bit that sees white – wants to shout “black”. As a compromise, the surround says nothing at all. So, because the centre shouts “white”, the overall message from the cell to the brain is a resounding “white!”. This is why the lines (where there is no intersection) look white.
When faced with a cross (right), the centre again shouts “white”. This time, though, most of the surround sees white and – being a contrary so-and-so – shouts “black”. So – claims the standard theory – the centre’s shout of “white” and the surround’s shout of “(mostly) black” combine, meaning that the overall message from the cell to the brain is “grey”. This – supposedly – is why we see grey blobs.
But it isn’t. Here are two reasons why this theory is bunkum (there are quite a few more).
First, the illusion still works if the lines are much thinner. The theory predicts that, because most of the surround is now seeing black – and therefore shouting “white” – the grey blobs should disappear. Second – and conversely – the illusion disappears almost completely if you make the lines serrated or wavy, which has virtually no effect on the overall amount of black or white falling on either the centre or surround.
So what is really going on? What the standard textbook account fails to mention is the existence of brain cells that are on the lookout solely for horizontal edges (there are also others that look out for vertical edges). So when faced with a horizontal line (left) the horizontal-edge detector cell confidently announces “yes, black edge here!”. But when faced with a cross (right), this cell is thrown into a tizzy. For most of the area that it is monitoring, there is no horizontal edge, as the one that would have been present is wiped out by the vertical line. On the other hand, there is a bit of a horizontal edge on the left and right hand sides. So the cell hedges its bets, and goes all Hugh Grant “Er…crikey…no, not a black edge really, but um, kind of, you know a bit of one… gosh….there’s not no edge…but, on the other hand, there’s not an edge…”.
You get the picture. Or, rather, you don’t. As a result of all this fence sitting, you can’t quite decide whether you can see a black edge or a white space, and end up seeing something in between: a grey smudge.
This theory easily handles all the facts that the old one can’t. Illusion still works with very thin lines? No problem, the edge detectors can still see the edges. Illusion disappears with serrated or wavey lines? Of course! Since there are no straight horizontal or vertical edges, the edge detector cells keep quiet.
This talk of lines, edges and smudges is all very well, but is there a deeper lesson? In my view, there is, and one that applies not just to vision research, but to psychological and scientific research more broadly. The moral is this: If we try to explain a single phenomenon in isolation (here, the original Hermann grid), we risk coming up with a theory that is barking up the wrong tree entirely. It is only by trying to explain the conditions under which a phenomenon does and does not occur (for example, the Hermann grid with thin, serrated and wavey lines) that we can hope to uncover its true underlying causes.
Source: Schiller, P.H. & Carvey, C.E. (2005). The Hermann grid illusion revisited. Perception, 34 (11): 1375–97
This is a sample chapter in the style of Psy-Q by Ben Ambridge, forthcoming from Profile Books (UK) and Penguin (US). For details click here