Photography, aesthetics of

2. Photography as mechanism

Claims that photography is a mechanical medium are common; clear expositions of the thesis are rare. Perhaps the clearest has been offered by Kendall Walton (1984).Compare a painting and a photograph of a rose. In both cases there is a dependence between the appearance of the rose and the appearance of the picture; if the rose had been different in various ways, the picture would have been different, and so, correspondingly, would the viewer’s experience of looking at the picture. But there is a difference between the painting and the photograph: the patterns of dependence that hold in the case of the painting hold in virtue of facts about the painter’s mental state. If the rose had had more petals, the painting would have shown more – but only because the painter would have mentally registered the greater number and incorporated it into the painting. In testing for dependence in the case of painting, if we vary the appearance of the rose, but hold fixed the painter’s beliefs about its appearance, the appearance of the painting does not vary. Things are different with the photograph; holding fixed the photographer’s mental state leaves the dependence between scene and photograph intact. Had the rose had more petals, the photograph would have shown more, regardless of whether the photographer noticed their number. A hallucinating painter will paint what they think is there; a photograph will show what is there. Photography is mechanical in that the relevant counterfactuals are independent, in the sense just described, of mental states. Here we shall concentrate on this version of the claim that photography is mechanical.

Having explicated the mechanicity thesis, let us now make the reality thesis comparably plausible and precise. Some versions of the reality thesis seem to assert obvious falsehoods. André Bazin’s claim (1967) that ‘the photographic image is the object itself’ is hard to reconcile with the fact that the Washington Monument and a photograph of it are utterly distinct and possibly very distant from each other in space and time. A better explication is provided once again by Kendall Walton, according to whom photographs are ‘transparent’: seeing a photograph of X is a case of really seeing X, just as seeing X through a window or through a telescope and seeing X’s reflection in a mirror are such cases. On the other hand, seeing a painting or drawing of X is not a case of seeing X. Photographs are ‘aids to vision’, whereas ‘handmade’ images are ways of representing. Roger Scruton (1981) has likened photography to the artful arrangement of a frame around a street scene; what we see inside the frame may be pleasing, but it is not a representation. It is the same, he says, with the photograph. Scruton is best understood as asserting, not that the photograph of the street is, or contains, the street, but that seeing the photograph is a way of seeing the street. This way of understanding the reality thesis does not require the photograph to be identified with or even overlap with its subject; we may see someone in a mirror without having to suppose that the mirror is part of the person. Nor does it require spatiotemporal proximity between photograph and subject; we now see stars that ceased to exist millions of years ago. It is this version of the reality thesis which I shall consider. I shall call it the transparency thesis.

Let us see how, on Walton’s construal, the mechanicity thesis supports the transparency thesis. Recall that the patterns of dependence holding between the rose and the photograph of the rose hold also between the rose and my experience of seeing the photograph of the rose, and that these dependencies hold independently of the beliefs of the photographer. Let us say that in that case my experience of the photograph is ‘mechanically dependent’ on the state of the rose. By contrast, my experience of looking at the painting of the rose is ‘intentionally dependent’ on the state of the rose; if we imagine the appearance of the rose to vary but hold fixed the painter’s mental states, my visual experience of seeing the painting does not vary. Now consider the case where I directly see a rose in front of my eyes, through a window, through a lens or in a mirror. All these are cases of mechanical dependence; they are also cases of genuinely seeing the rose. Seeing the photograph of the rose shares this mechanical dependence, but seeing the painting of one does not. So seeing the photograph of the rose is to count as genuinely seeing the rose, but seeing the painting is not.

This nexus of views about the mechanicity and transparency of photography has the advantage of not denying certain facts about photography which are sometimes cited in favour of the view that there is not, after all, any principled difference between photography and other representational media like painting: for example, that people typically choose the subject, location and lighting conditions of the photographs they take, and that these choices can all be made with more or less skill, aesthetic sensibility and expressive effect. While true, these considerations are largely irrelevant to the thesis that photographs are mechanical and transparent. When we look at something with the naked eye, how it looks to us depends on lighting conditions and on spatial relations between the viewer and the object viewed, and we can choose to look at this rather than at that. None of this impugns the claim that seeing things with the naked eye is in a certain sense a mechanical process: that when we look in a certain direction it is not up to us or to anyone else to decide what we see; what we see is determined ‘mechanically’ by what is in front of our eyes and not by decisions or preferences. Nor is it an objection to the claims of mechanicity and transparency that photographs do not always look very much like – and never look exactly like – the things they are photographs of. A photograph may grossly distort the appearance of a man by being created with the help of a distorting lens. But you can look at a man through a distorting lens, thereby seeing him in a grossly distorted way: you are still seeing the man. Nor need the advocate of mechanicity and transparency deny that skills are exercised in the production of photographs, and that photographs so produced may exhibit these skills.

Does the mechanicity of photography provide a convincing argument for its transparency? Imagine two indiscernible clocks, A and B, where mechanical connection ensures that the time told by the hands of A causally determines the time told by those of B. If I am looking at clock B, then it is true of my current visual state that, if the hands of A told a different time, my current visual awareness of B would have been correspondingly different. And these counterfactuals, we may assume, are independent of the intentions of agents in the sense described above. But my seeing B is not also a case of my seeing A. So mechanicity is not sufficient for transparency.

Walton grants the insufficiency of mechanicity for transparency; there might, he concedes, be a machine which examines objects and churns out linguistic descriptions of them in such a way that there is mechanical dependency between the object and the descriptions produced; but reading the descriptions would not be a case of seeing, or of perceiving in any way, the thing described. What stops reading the description from being a case of seeing the object described is, Walton claims, the lack of relevant similarities between the experience of reading and the experience of seeing the object itself. Here we are close to the assertion that mechanicity and likeness (in some relevant respect) jointly constitute a sufficient condition for transparency. But that is not so: the experience of seeing clock B from our earlier example is, in every relevant respect, very much like the experience of seeing clock A, but seeing clock B is not a case of seeing clock A. This is not to say that the transparency thesis is false, but merely that it is not supported by the mechanicity thesis.