Picturing Vision

Joel Snyder, ‘Picturing Vision’, pp.157-171, in:
Yates, Steve. ed. Poetics of Space: A Critical Photographic Anthology. University of New Mexico Press, Albequerque.


Our willingness to accept photographs as natural and mechanical records of what we see underscores the power of our belief that certain kinds of pictures achieve significance because they are “natural” – meaning that such pictures are related to what they depict in exactly (or roughly) the same way vision is related to what we see.

But the history of the camera and its ancient forerunner, the camera obscura, demonstrates beyond any question that camera imagery is entirely conventional, demonstrates that cameras have been designed to achieve specific kinds of pictorial results.

The question that is hidden beneath the modern (and futile) discourse about the ontological status of photographs is far more interesting than any that has yet been posed by theorists of photographic representation: How is it that we ever came to think of photographs as natural phenomena at all?

Our passive and unquestioning acceptance of the commensurability of certain kinds of pictures with what we see is the source of our unshakeable belief in the congruence of certain pictures and the world we view.

The history of Western painting from the early-fifteenth to the late-nineteenth century was marked by the attempt to secure a scientific basis (both mechanical and physiological) for picture construction that serves, in turn, to warrant the viewer’s belief in the fidelity of the picture to what it represents.

The primary condition for this kind of picture-making is the belief that vision is amenable to depiction because it is itself pictorial. New theories of vision lead to new “facts” concerning what we “really see”.


The artist can depict what we see because what we see is pictorial. And yet, in paintings, the artist can achieve fidelity to his own vision based upon his knowledge of vision and depiction, and we will accept the picture as credible and warranted even though we might insist at the same time that we never quite saw things that way before.

[…] the joining of artistic practice to scientific theory in the early Renaissance gave a new rationale and impetus to artists who wished to depict what they saw. And it provided rhetorical assurance to the audience that what they saw in paintings was related by the sure methods of science to what they saw when looking at the world.

The first text on linear perspective, De Pictura, written in Florence in 1435, by Leon Battisa Alberti, also marks the first effort by a painter to establish the certainty of his method of picture construction by deriving it from a scientific account of vision.

The system continues to be used today in its purest form, in many kinds of handmade illustrations, and, of course, in nearly all applications of photography, including motion pictures and television.

De Pictura lays out procedures that permit an artist to paint what is seen by means of rules derived from a mechanical and psychological account of how one sees.


Linear perspective, by definition, requires the painter to “fix” the eye in a determined and unvarying relation to the picture surface in order to recreate within the picture the rational structure of perceptual judgements.


[further discussion of Alberti’s rationale]

The scientific account of vision adopted by him provides a basis for explaining how we are able to make “certified” judgements about the sensible things of the world. It is not an account of momentary glances or “impressions,” nor is it, strictly speaking, an account of “appearances.” A completed perceptual judgement, that is, a unified one in which we correctly identify objects, their attributes and their interrelations, can be made only under specified observation conditions through time, by means of discrimination, comparison, and integration.

What is fragmentary or unsure in perception cannot be certified, unified or identified. Such fragments have no place in depiction because they are irrational and incomplete; they fail to achieve the purpose of vision.

We still resort to this mode of depiction, or modes very closely related to it, when we wish to make “literal” pictures. But when we do so, we adopt a thoroughly medieval notion of vision and early-Renaissance conception of depiction.