That bloody dress again – vision and choice in perception and values

I haven’t updated this blog in a while, mainly because of being very busy. I am still very busy, but since I have been gripped by the same thing that everyone is getting by turns obsessed and infuriated about over the past day or so, it seemed worth getting it off my chest.

For some reason, everyone went totally mad over a fairly ordinary quality photograph of a not very nice dress. The reason for this insanity was that people disagreed dramatically over its colour. Some people saw it as blue and black, while others saw it as white and gold. I was closest to the blue and black team, although it was more like a vibrant blue and a dark warm brown colour.

Very quickly, people started offering scientific explanations of the phenomenon. Most explanations put it in terms of a basic contrast effect. When displayed in a particular visual context, different colours appear in different ways. This XKCD cartoon demonstrates the effect.

Now this is part of the explanation, but it doesn’t give us everything that we need to know. Take the famous checkerboard illusion. In this illusion, two squares that are coloured in the image with the same pigment appear as different colours because of the visual context. But with this illusion, pretty much everyone looking at it will perceive it the same way. With the dress, people were pretty close to coming to blows over being contradicted on what just seemed so obviously the colour. Even people looking at the picture in the same conditions disagreed about the colour, so we can’t just put it down to surrounding lighting conditions, the colours of other things on the screen or in the surrounding environment and so on.

We then got other explanations of the difference that referred to stuff going on in people’s visual systems that led them to differ in their colour perception. Different arrangements of rod and cone cells, or different neuronal activity might privilege a certain way of seeing the dress.

It seems that the best explanation is likely to be a mixture of the two, combined with other factors. Certain particularities about the specific configuration of our eyes, our nervous systems and the surrounding environment will all have a role to play. Other things might come into it too: I wouldn’t dismiss the possibility that certain moods might incline us to imagine the context as a lighter one or a darker one for example, or that memories of clothes that we have seen or owned might come into it. There are also features about other perceived qualities of the dress. I was eventually able to make the switch to seeing it as a white/gold dress, but only once I started to think of it as much more shiny and satiny in texture than I had previously done.

This is the sort of thing that the phenomenologist Maurice Merleau-Ponty talks about. In his view, the perception of a quality is not something that happens in isolation. I don’t just see a red, for example, but rather a warm woolly red. Colours are not just colours in the abstract, but they come with textures, smells, occasions, memories and meanings. Perception is a feature of the way that we operate in the world, and the meanings that things have for us. I saw the dress as blue/black when I was imagining it as a matte textured garment that I might squeeze myself into in the morning, with sunlight streaming through the curtains. When I managed (after a bit of work) to see it as white/gold, I was seeing it as a glitzy evening affair that I would put on in preparation for a night out, while my hair was in curlers and I was listening to ABBA. This is a bit of a dramatization because I probably wouldn’t ever wear that dress, but you get the point. Things appear in the way that they do because they have certain kinds of practical significance.

But actually it wasn’t so much the peculiar visual differences around the dress that interested me so much. What intrigued me was quite how mental everyone went about it. Not only was there radical disagreement between people’s visual perceptions, but this was extremely vehement, and people were quite bothered by the fact that other people saw things so differently. People seemed anxious that this was leading them to question objective reality, or the reliability of their senses. Other people flooded to online tests for colour blindness with the conviction that if other people saw things so differently, perhaps there was something wrong with their visual system. The response to the effect seemed very often either to be that either you or I must have some pathology, or the basis of the world that we inhabit is somehow threatened.

This is interesting because people disagree about things on the internet all the time. This is often vehement and entrenched, but it doesn’t usually lead to the same kind of anxiety or questioning. Why not? Probably because most of these disagreements are about politics, aesthetics or ethics. If I say that you need to get your ears tested if you claim not to have enjoyed the latest Björk album (I haven’t listened to it yet actually, but you know) you would not actually think that it was something that you needed to rush to the GP about. If I said that you are wrong in the head for voting UKIP, you probably wouldn’t take that remark seriously (actually, you should). These are value judgments of various kinds, and most people today tend to think about value as subjective: even if our moral, aesthetic or political reaction seems to us to be immediate, obvious or visceral, we won’t usually be surprised that other people don’t see the world in the same way.

Iris Murdoch has an interesting line on facts and values. For her, an ethical outlook is a way of seeing things. People or objects appear to us in certain ways because they matter to us in relevant respects. This is close, in some ways, to Merleau-Ponty’s ideas about visual perception: a face, for example, might appear to us completely differently according to the attitude that we have towards the person concerned. There hasn’t been much work, so far as I can tell, on the connections between Merleau-Ponty and Murdoch, but I think there should be. Perhaps that will be my next project.

Another thing that Murdoch talks about is the way in which the classic distinction between facts and values is not as clear-cut as it might first appear. It is certainly useful to make fact/value distinctions in some contexts: for example, it is helpful to challenge people who think that something is automatically good because it is ‘natural’, and bad because it is not. But this approach has its limits. The stark distinction, Murdoch thinks, is itself based on a certain moral take on the world: one that is rooted (she thinks) in Protestantism and Liberalism. Why would we say that? I find the Protestantism aspect a bit harder to grasp (although, informed by my recent viewing of Wolf Hall, I am starting to get some inkling of it) but the liberal bit has two components:

  1. Meaningful discourse must be possible, so we must basically agree on the nature of a public external world that we all inhabit, which furnishes us with ‘objective facts’ that are established through our phenomenal experience.
  2. Disagreement must not be irrational, so ways that things matter or have meaning for us (‘values’) must be private, subjective, and available as choices.

Now, if we live by this kind of framework, of course disagreements about something like the dress will disturb us in ways that disagreements about ethics, politics or aesthetics will not. The ‘objective world’ which we thought that we shared with other people has come under threat.

The dress is an interesting case because (a) it is not set up as an illusion – it isn’t something that has been deliberately set up to trick us, but rather takes the form of the photos that we usually regard as reliable, which makes it more threatening; and (b) many people did not feel that they had any choice in the matter of how they saw it – disagreement seems to make it fit in the ‘subjective’ box, but the absence of choice over how to see it (in contrast with, say, the Necker Cube) seems to make it belong in the ‘objective’ box. The idea that we see things differently, and this is not something that we have any obvious power to change, is understandably threatening to many people.

Interestingly, there seemed in this case to be much more agreement about an aesthetic issue. Most people seemed to be clearly of the opinion that it was a horrible dress. In this case it was the ‘subjective’ issue where there appeared to be the most agreement.

Of course, despite a failure to see the colours differently, we can still come to scientific and philosophical understandings of how the differences come about. Even if we can’t see it how others do, understanding why their perceptual experience might differ from ours might make us refrain from shouting at them that they must be mad, blind or deceitful. And eventually, by trying out different conditions, squinting a bit, and returning to the image, not to mention a bit of imagination, we might actually start to see it differently, even if just for a moment. Here I think the issue of value is no different.