Chapter 2. Photomontage

From the earliest days of coinage and use by the Dadaists, photomontage has been associated with mixed signifying modes including printed and hand-written fragments of text (letters, journal entries, newspaper headlines and columns, advertisements, packaging, instructions, posters and fliers). But the means of mixing or compositing Dada "photomontages" were frequently the scissors and paste of collage. Photomontage as we will trace it in this chapter produces a single photographic (or photo-digital) image from multiple negatives, scans, or layers; the stack of layers may or may not include images of texts. Photomontage is still a mixed mode, however, because it involves seeing double or multiply. I am thus excluding photomurals and illusionistic fakes because they use compositing to create a single image. Photomontage insists on its multiple modes within a single frame.

Collage, as opposed to montage, is made up of fragments with sharp cut or torn edges; montage blends its component images together by thinning opacity at the periphery. Our practice will follow the urging of D. W. Coleman (on many occasions). Gordon Baldwin also draws the collage/montage distinction in terms of edges/smooth blending (Baldwin, 26).

In film, the term montage is generally used, following Eisenstein, to refer to the transition from one shot or one sequence to the next one. The second may continue the theme of the first by a kind of visual metaphor (couple kissing -> train entering tunnel) or ironic ("intellectual") juxtaposition (these are Eisenstein favorites). In The Language of New Media, Lev Manovich speaks of this as an aesthetic of juxtaposition and dissonance, which he claims was the dominant aesthetic throughout the twentieth century "from the avant-garde of the 1920s up until the postmodernism of the 1980s" (p. 144) and he distinctly prefers it to the current "digital compositing aesthetic" of blended virtual worlds cluttered with multiple but not juxtaposed bits of media. In any case, the aesthetic of photomontage in a composited image is not one of juxtaposition or dissonance but of co-presence in a space.

Figure 2.1
Duane Michals: Rene Magritte (1965)

Much has changed since the early days of photomontage, both in the technology of image making and in the terms for interpreting them. Photomontage did make the transition to color, and then received tremendous impetus from the development in digital graphics of channels and layers for controlling opacity and superposition of images. Making photomontage has become very easy; seeing and interpreting it, however, can be quite hard. The simple linguistic analogy that thinks of component shapes as words and the placement of the shapes as the syntax that constructs them into sentences is in trouble with photomontage from the outset, since photomontage works against the discreteness of "words" or their placement in one consistent organizing sentence structure. The modes of language that come closest are those of analogy, simile, metaphor, syllepsis ("construction changed"), and paradox—figures of doubleness and plurisignification rather than unequivocal statement. It is well to remember that one great proponent of the simple linguistic analogy is Robert Horn, and Horn uses clip art to exemplify his claims. If anything has hard, crisp edges, clip art does, and I have never seen it with gradient shading or reduced opacity. Whether and how photomontage might be used for illustration of complex issues and processes will be taken up in the final section→.

A second source of difficulty is the unnatural naturalness of photomontage. As many have noted, it presses always toward the impossible, the incongruous, and the non-literal, though it is executed as a photographic print, the most referential and realistic medium there is. It returns us to the question of visual literacy and complicates the simple dichotomy of special conventions of a medium vs. general heuristics for understanding the world. That is, we noted with approval Messaris' argument that "literacies" ought not to be multiplied unnecessarily. In short, we do not need to be taught or taken through an apprenticeship of viewing to be able to see small images of people in a picture as farther away from us that larger ones. That is a basic principle of how we orient ourselves and what we see in space. But some of the objects in photomontage have unusual properties and present us with problems seeing the objects, locating them in space, or identifying them. So does that mean that there is literacy in photomontage—i. e., a set of special decoding skills and interpretive rules that we would need to acquire over and above those required for ordinary life in the world? This is a point we will take up directly in Section 1.

In splendid article on photomontage and photocollage as developed by the Russians in the 1920s and '30s, Benjamin Buchloh gives great prominence to the utopian and futurist hopes of Lissitsky, Klutis, and Rodchenko among others that their new art which shattered the conventions of bourgeois realism and representation would have a direct, immediate appeal to the workers and peasants, so that the work of the avant garde could be the art of the people and foster revolutionary aspirations to build a new order. This dream quickly collapsed, says Buchloh, when it became clear that the new order might just as easily be Fascist as Bolshevik and when the new art was not embraced by those traditionally excluded from the appreciation of high culture. ¹

What Lenin, Stalin, and the revolutionary councils failed to do, however, was accomplished by one corporation, Adobe Systems, and one product, Photoshop. This image editing program (from the 3rd version on) has set loose a tidal wave of photomontage in consumer and business cultures and makes the graphics for the Web, born as a graphic display medium at almost the same moment as Photoshop 3. Graphic looks that used to require many hours of tedious labor and expensive equipment (not to mention training and apprenticeship in using it) can now be produced in a few minutes at a desktop computer. And many of those looks, from CD cover to poster to corporate annual report, are looks of photomontage.

Further, the options available in Photoshop go well beyond the looks that photography or media has put before us, and there are no conventions or expectations or standard uses of them. So the questions of how we see and interpret these new images is very much an open question.

The discussion begins with seeing photomontage and with four parameters that make seeing them difficult: opacity/transparency of the components, edges of objects, location in space, and order in a stack layers; we will then takes up interpretation, especially of these unusual visual features , and then we will move to questions of use.

2.1 Reflections and Opacity

A point of visual literacy arises when either a medium presents us with something we have not seen before or assigns something a special signification it does not usually have outside the medium. Even if something meets one of those conditions, it may not be codified (i.e. assumed to be recognized by all users of the medium). Codification narrows interpretation, in that it picks out one from among several significations that we could think of assigning to a feature. A spot ( or circle, or star) of light appearing usually relatively high in a photograph can be recognized as lens flare—a virtual image of the sun or other strong light source—produced by viewing through a lens at a certain angle. Its look is distinctive enough that it can be faked by an image-processing program. It does not have a signification, however, other than that the scene has been viewed through a lens at a certain angle etc. Ways of indicating "flashback" in cinema, however, via softened focus, and/or slow motion, are codified, and film makers can expect audiences to recognize that what they are seeing in a flashback is not currently being experienced by anyone in the movie, but is part of the experience of some individual, or the collective experience in the world of the characters.

Photomontage presents us with special combinations of line, shape, and light that challenge our working, day-to-day visual literacy, but we are not at an utter loss when we encounter them, for some of the key differences such as reduction of opacity and discontinuity of form do occur daily in our encounters with reflections. Indeed, Barbara Morgan entitled a lovely Macy's store window picture "Natural Photomontage" (1939/72). The spectral look is also the look of reflections in plate glass, as in Eugene Atget's series of Paris display windows: the reflected buildings (or faces, or objects) can be nearly as "solid" looking as the window frame itself, but are usually partial and appear in impossible places. In Atget's favored setup, the buildings in the street are reflected in the window about two meters deep, so that the mannequins are sandwiched between the reflection and the glass. And they pose their clothes (usually) in the peculiar, stylized space directing their gazes and expressions into yet a third kind of space, which is the wish/dream space of "glamour" (as John Berger would say). In fact, Berger uses a display window image by his collaborator Sven Blomberg to open the chapter on art and advertising in Ways of Seeing. Everything ends up being partial and not well connected with the ground (though they are well-connected in the system of commodities via the price tags). The pictures are visual mazes or puzzles which intrigue the eye as well as the interpreting mind; they may remind us that photography captures not things but light from reflecting surfaces, but they can finally be naturalized in the conventional scene of one camera taking one exposure from one point at one moment in time.

Figure 2.4
Clarence John Laughlin: Victorian Phantasms (1946)

The visual array in the display window is quite a complex mixture of opacities and interrupted forms, and, though it is easy to come across display windows, we may not automatically decode them—in fact, we may simply ignore the reflections—filter them out. Still, knowing that something is a display window makes it very much easier to decode. We know, for example, that it is shallow space, that one wall is plate glass and likely to reflect part of the daylighted street scene back on itself.

This display window picture from early in Clarence John Laughlin's career (1946) starts out as a fairly standard example of this visually complex naturalism but pushes into the surrealistic mode of impossible vision. In the foreground is what appears to be a doll with the placard "Leghorn Hat" resting on an ornamentally carved pedestal. The head is certainly not a doll's head; it may be a mannequin's head, but it is grotesquely oversized for the little dress and stumpy body. Behind this figure and facing the other way is a mannequin of full stature but attenuated to near transparency from the shoulder blades down and hence not supported, apparently, at all. Within and through the outline of her figure we see yet a third dress on display. It too may be a "phantasm." We may feel cheated that the "decode the reflections" game is not being played by the rules, but the title after all warns us that something paranormal is afoot.

Laughlin, who made his living as an industrial and architectural photographer, has the surrealist fondness for "impossible worlds" but his favored kinds of impossibility are things like unsupported objects and trees growing out of stone columns. His titles rather reassuringly identify the scenes as apparitions, dreams, phantoms, visions, enigmas, specters, spells, strange, and enchanting. Breton and his crowd don't let common sense off so easily.

Figure 2.5
Hygeia: Bridal Fantasies

Recently, the street window maze was made the basis of a whole project in several British cities by Tapio Mäkelä and Susanna Paasonen. Displays with touch screens were set up in some downtown shop windows (Oxford Street in London), and the windows were photographed much as they would be seen by someone using the touchscreen. Some of the Atget-like images from the subsequent web site were made "dynamic" with Javascript adding a more "spectral" image. Here for example the image you first see lacks the black and white central portion which appears only on mouseover (and disappears on mouseout). This "reflection" is of a happy couple encircled by a ring (not completely clear here--it occurs more fully elsewhere in the site). This image is part of the cultural production of weddings; its location suggests that it emanates from the girl's admiring, aspiring gaze. Thus, touching the image with the mouse causes this additional "reflection" to appear, much as if we are standing in front of the window and change our angle of view.

Figure 2.6
Mac Adams: "Reflection of a skin head rally in Rome, Italy in a window on Broadway, New York City, 1994." (1995)

These reflections in plate glass are always a little spectral, since they tend to not reach the ground and to float. Mac Adams develops the apparitional possibilities by composing images where figures from the news (mainly crime stories) are layered onto various reflecting surfaces. He performs what spherical or cylindrical distortions are necessary to emulate the look, feathers and fades the edges, and voilá!— the child-murderer Susan Smith appears in your teapot as she heads into court. For a moment, just a moment, the teapot is an impossible object—a crystal ball—expressing, Adams suggests, the much increased closeness of things in the world brought about by the webbed communication (and, one might add, the increased apparitional potential of things). In the same series, he includes an image of a skinhead rally in Rome reflected in the window of a New York cafe.

Portraits taken through plate glass with reflections or taken of the person as reflected in the glass have become quite trendy. Wired has run at least one in a profile article each of the last several months. Under high modernism, photographers avoided traces of the scene of photographing such as reflections of the taker and camera, flash and lens flares, and shadows and reflections of the taker. Now the sophisticated look foregrounds the artifice, and produces a portrait which does not purport to capture the very essence of the sitter but declares itself to be a glimpse, an unplanned, momentary appearance.

Figure 2.7
Man Ray: The Primacy of Matter over Thought(1931)

Few photomontages are equally difficult in all four ways, but if we have difficulty with what, the latter questions are hard even to formulate. When determinations of what become effortful, we fall back on re-tracing elementary, usually automatic, processes of tracing edges and contours and looking for at least partial matches to familiar objects. When looking at Figure 2.7, which is solarized but not a photomontage of two negatives, seeing a familiar object in a familiar orientation is not a great problem, but it provides a little shudder in that the edges and surfaces we expect with this familiar object break down on the lower edge, which appears rather as that of the edge of a flowing liquid. Rather than regarding this as an impossible object or optical illusion, however, we may try to naturalize it in some grotesque way ("she is melting"). There are other, harder cases.

Figure 2.10
Coryndon Luxmoore: Gaze 4

Edges and contours (and indeed, objects) are to various degrees problematic in a remarkable series of photomontages by Coryndon Luxmoore at The Birdhouse. This one ("Gaze 4") is difficult for several reasons, notably the variable opacity of the women figure and the odd angle of vision. There is neither vertical nor horizontal in the entire piece—even what appears to be a corner of a wall. The angle of view of the girl is not "canonical" (there must be several canonical views of the human figure male and female, but this angle is unusual) and it is hard to imagine a viewing point for us (the camera).

Figure 2.11
Coryndon Luxmoore: Gaze 5

With regard to a conventional viewing point, Figure 2.11 ("Gaze 5") is much easier, since there is a very long pictorial tradition of women being viewed from this angle. Also, the cable and one pipe make a solid horizon reference line to which the vertical pipes and brick ledge can be linked, so that there appears to be a supporting surface or floor upon which the woman lies. Obviously the images are about contrasts of hard, edged, well-defined stone and metal shapes and the soft contours of human flesh, but the images don't read as cyborg fantasies or fantasies of penetration and violation of the body by the mechanical and rigid; the body, lacking a uniform opacity, seems to inexplicably fade in and out like a ground fog. This uses the powers of a Photoshop type graphics program to gain a level of control that would be very difficult with multiple negatives, masks, bleach, and clever printing. The problem posed for perception here is that the image will not sort into planes or even into continuous (if incomplete) objects. The image cannot be made sense of according to the rules for constructing the world of material objects, so that even the hard metal and brick here become melted or dissolved in the desired or hallucinated body. Here for comparison are the edge-detected and posterized versions of "Gaze 4" (somewhat enlarged):

Figure 2.12

Most of the clear edges are those of cable and pipe; the only line that traces the edge of the woman's body runs along her forehead and down her nose, but it would be very difficult to see a human form based just on that line.

Figure 2.13

Here the shading is reduced to three levels (black, white, and gray) and actually does a somewhat better job than the edge-detected image at identifying a human shape. Note the shadow-line reinforcing the lower edge of the face which merges into the shadow cast by the pipe: that is not helpful at all!

Clearly these will not get anyone called up on obscenity charges! Granted this is a complicated example involving perception of objects and orientations in space, but part of the uncertainty lies with what are the objects. ²

The "Gaze" series do not easily sort into figure-and-ground (even though we automatically take the softly shaded bioform to be the figure); most classic photomontage does not fade figure and ground in this way. Fading one object into another to make an impossible object was a favorite device of the Surrealists, but they did not often fade object into ground; further, they generally avoided the diaphanous, semi-transparent treatment of any objects, even and especially the impossible ones, presumably because they did not want to pre-sort the scene for the viewer in to more and less "real." And yet further, if they do include a semi-transparent object, they observe a very general convention of such objects that they exhibit the same degree of transparency throughout. This is one of the general principles (conventions?) that Luxmoore's Gaze series violates, and it is violated also in Clarence John Laughlin's "The Eye That Never Sleeps" (Figure 2.14), which is unusual for him.

Figure 2.15
Catherine McIntyre: Crystal Ball

Another process that alters shading and edges is Inversion (or "Negative"), in which all the color values switch sign, as it were, with perfect neutral gray being 0. That is, the darkest become the lightest, hues become their complements. When the lighting is directional, Inversion makes it seem to be coming from the opposite direction; so, objects lighted from above will appear to be lighted from below. Once launched by photomontage down the path of "post exposure manipulation", Inversion is a next step away from what we commonly experience. It is not always as disruptive as one might think to perception, though it tends to look like an xray negative when applied to the human form. That is perhaps for most people their main preparation for seeing it in photomontage. It is much employed in making fantasy and magic worlds because of the mysterious emanations of light it suggests. So in Figure 2.15, the crystals seem to illuminate the scene with their glow (dilithium, don't you know) along with the little double helix in the sphere and the ice-crystal thingies that decapitate the women. One click inverts the image, which then is lighted from above as is normal and the woman has a more clearly discernible profile.

Figure 2.16
Morgan: City Street (1937)

Horizon and Perspective Space

Albertian single-vanishing point perspective comes with the camera and hence with every negative; indeed, it is built into the standard image-synthesizing ray-tracers as well. (Refer to perspective demo.) In addition, the camera also renders texture perspective (the blocks and bricks of walls and floors get smaller with distance), atmospheric perspective (distant objects are reduced in contrast, blurred in detail, and bluish in color), and of course shading and shadow. Photomontage with two or more fairly robust perspective spaces are not going to blend or merge very smoothly. In some cases, tension between perspective spaces can be very productive, as in Figure 2.16, where the camera is rotated 90 degrees from portrait to landscape orientation (and taken a few floors up). The shadows help to key this. The joining of lighted/shadowed portions of the sidewalk on a line slightly off the vertical is most artful, the composition rigorous and unaltered.

Figure 2.17
Val Telberg: Greeting (1979)

This piece shows some horizon and perspective, particularly with the central receding passage, but equally clearly it sets up a space so that the women can break out of it, and the contemplative figure in black silhouette is in yet a third space. And too solarization is fairly heavily used, which further breaks down naturalistic space, lighting, and surface.

Figure 2.18
Clarence John Laughlin: The Mirror of Long Ago (1946)

When a clash or doubling of perspective spaces is not wanted, the common practice is just to crop or erase the signs of location in space surrounding an image and then paste it (or overprint) it into the matrix image. Usually great differences in scale help to deflect a conflict of perspectives. This selection procedure seems to have been that applied in making "The Mirror of Long Ago" (Figure 2.18), but viewing the larger-sized image will show that Laughlin has not effaced the framing of the young woman's image. He tells us that the picture is made up of two negatives taken in a certain grand old New Orleans house, and then we see the two ornate frames, somewhat misaligned, as if the framed portrait, placed at a slight angle is being reflected in a grand view mirror for the room. There is upon reflection (8-)) one large problem with that, which is that angle or no, the picture would have to be shot through the back of the portrait. Further details like the ceiling and wallpaper lines and frames suggest that both frames have been photographed while flat on their walls. In fact, there seems to be a second chandelier, suggesting that they may be hanging in different rooms. His caption in the catalog notes is clearly a tease; "Here the phantom of the lovely and ill-fated Julie rises from the great mirror in the drawing room, surrounded by the symbols of magnificence" (Laughlin, p. 121)—yes, but only after you have had a pint of bourbon and stopped following the clues he has scattered about.

Superposition (stacking)

Even when edges are not strongly defined, the blending of one figure into another or the ground does not necessarily undermine our perception of the individual objects, provided they are familiar objects and only overlap or bleed through in certain areas.

Figure 2.19
Burk Uzzle: My Precious

Burk Uzzle's "My Precious" is not hard to see: it evokes a familiar "family picture" scenario with the two men posing, one holding the boy (son), before a brick residence with to be sure a rather substantial door. Centered between them is another familiar figure and she is quite "solid" and uninterrupted in her materiality —even though the scale is wrong and her position is well-known and fixed elsewhere, so she is clearly symbolic of American liberties (in her usual way) rather than physically present in the snapshot, and similarly, the pistol is out of scale the other way with the metal of its barrel and handle giving up their solid opacity to the dungareed legs of the men, so again we conclude it is symbolic of the presence of violence and not present in the scene or between the figures and the camera, or what have you. The thinning out of the boy and his father, which allows the brick and door trim patterns to appear through their flesh is perhaps a gesture in the direction of the symbolic: that is, their physical presence and reality is reduced slightly in the presence of the strongly symbolic images of Liberty Lady and automatic. It is not completely clear what they are all doing there (the title, "My Precious" narrows but does not fore-close all interpretation), but it is clear what is there, since the gradients at the edges of the bodies are strongly shadowed and continuous even when attenuated. (Aperture published a book of Uzzle's photographs (in conjunction with an exhibit) with the title All American in 1979.

Since, as is generally agreed in vision science, certain receptors in the brain act like "Edge Detect" graphics filters, we can simulate the brain's "initial sketch" of images by viewing them through edge-detect filters and 3-color palettes ("posterizing" filters) that have the effect of reducing contour gradients to edges. We can contrast such reduced "sketches" to "sketches" of the considerably less edged images of Coryndon Luxmoore. Figures 2.20 and 2.21 are the reduced sketches for "My Precious".

The separation and integrity of the objects is clearly evident even in these reduced sketches. So the objects are fairly easy to see and to sort out into a basic scene of viewing (and photographing) with certain readily identifiable symbols added.

Figure 2.22
Joyce Neimanas: Baby Tale

Joyce Neimanas's photomontages, however, cultivate overlapping images that almost merge, so that seeing them is like looking at a Neckar cube: you can see them one way, or another, but not both at once. "Baby Tale" (Figure 2.22) is typical of this work, where you must fix on the woman's (grown up from child? mother?) eyes and mouth to pull her face out of the more strongly outlined baby's.

Figure 2.23
Joyce Neimanas: Face Lift

A second stack of faces, that in "Face Lift" (Figure 2.23) , is even harder to resolve into separate faces—impossible, I would say, given just a small .gif to work with. With the larger .jpg from Eastman House, problems remain identifying the number of different faces stacked up. Starting from the front, there is a double image which may illustrate the notion of a face being lifted up and put on—but whether it is over the same person's head or another's is unclear. This larger face, with hair, seems to fit just inside another—one that is proper to the second plane (the contained rectangle) and which has longer hair falling onto the right shoulder. This hair in turn seems of a color different from the raven tresses of the largest head. So how many faces, or different faces, do we see? The relation of the faces (connected by or to "face lift") is equally hard to fix, as is the bearing of the cartoon balloon on the entire assembly. The arms and hands lifting up the face illustrate the triumph of line over color: they are broken into four bands of pearly, translucent color but manage to stay very clearly visible as arms. (They are more successful in 24bit True Color than in 8bit .gif.)

The notion of a plane with several objects or textures being a slice of a world perhaps derives from the early practice of sandwiching negatives or making successive exposures of a piece of printing paper. With a digital program, one uses layers, usually transparent, into which selected regions of another image are pasted and it is easy to think of just affecting a region of a picture rather than superposing worlds. In any case, modern computer-generated photomontage often becomes quite large and/or many layered, and sustaining numerous planes over the whole image can produce a lot of meaningless clutter, so, with larger or many-layered "canvases" the planes dissolve into zones, and the ordering front-to-back of the stack falls away as a key element in interpretation (or even sorting of what is seen). The space of the image becomes the infinitely capacious, tolerant one that is generally interpreted as dream, magic, or fantasy landscape, which we will look at in more detail below in Section 4.

There are then at least three four dimensions along which photomontage varies:

• opacity,
• outline,
• perspectival space,
• planes.

We have already touched on certain interpretive rules of thumb that have gotten attached to certain values on these dimensions. We will now turn to general principles and particular applications associated with these three features of photomontage.