Choosing images for a walk

People sometimes assume the photographs in THE WALK were chosen because they are good photographs. Some are. That was not the criterion. A picture earns its place in a walk by what it does to the person standing in front of it with a question beneath it, and that is a different test from the one a gallery applies. Here is the checklist we used for the twenty, and now use when hosts ask whether a local image could join an edition.

Does it withhold? The picture must leave something unresolved that the viewer has to supply. A figure too distant to read. A door open onto light with nothing visible through it. Hands holding a cup, and no face. If the picture answers its own question, there is nothing for the viewer to do but agree or disagree with it. We rejected many beautiful images at this stage because they were complete.

Is there a person, and are they far enough away? Almost every image in the walk has a human figure. A landscape with no one in it invites admiration; a landscape with one small figure in it invites identification. But the figure must be small or turned away or half-lit. A face looking at the camera makes the picture about that person. A back, or a silhouette, lets it be about the viewer.

Is it quiet? No text in the frame, no logos, no recognisable places, no strong colour. The twenty are all black and white, which was a practical decision as much as an aesthetic one. Colour carries mood. A red room tells you how to feel before you have looked. Grey does not.

Can the question sit under it without fighting it? We tried every finalist with several questions and kept the pairings where the picture changed the question rather than illustrated it. The corridor could have carried "Where are you going?" That is an illustration. It carries "Where in your day do you move without noticing?" because the corridor makes the viewer think about habit, and habit is where the question lands.

Does it survive being seen for ten minutes? Some images are strong for five seconds and empty after thirty. We sat with each finalist for ten minutes, as participants would, and kept the ones that gave something new in the last two minutes. The still-water picture passed this test better than any other, which is why it sits at the hinge between the second and third movements.

Does it do its work in the sequence? A walk is not a selection; it is an order. The first movement settles the viewer with familiar things. The second unsettles gently. The third goes to what matters. The fourth brings the viewer out towards a step. A picture can be right in itself and wrong in position. Our fourth image was moved three times before it stopped fighting its neighbours.

Is it safe? Nothing that depicts harm, nothing that could be read as an image of a specific traumatic event, nothing with a child in it. We are not a therapy room and we do not know who will walk in. The picture that unsettles should do so the way a question does, not the way a shock does.

One last thing we check that has nothing to do with the image. Could the host explain it? If the answer is yes, we look at it harder, because an image the host can explain is an image the host will be tempted to explain, and the walk depends on nobody doing that.