What makes a picture similar to a mirror image? A pragmatic assump-tion whereby a dark room should be as truthful as a mirror and, at any rate, testify to the presence of an imprinter (present in the case of mir-rors, past in that of photography). The difference lies in the fact that the exposed plate is indeed an imprint or a trace,
A trace differs under certain respects from a mirror image, even disre-garding image reversal on the plate, its further reversal on the printed picture, and the recovery of its inverse symmetry, that is, the actual inversion of the congruence characterizing mirror images.
The main point is that imprints are motivated but heteromaterial (Eco 1976, 3.6): the plate turns light rays into different matter. We no longer perceive light rays, but pure intensity relations as well as pigmentation relations. Thus there has been a projection from matter to matter. The channel tends to lose its importance, the picture can be transferred on different materials, while relations remain unchanged. The image is not independent of its channel as the Morse code is of the material em-ployed for its standard signals. However, some kind of liberation is foreshadowed.
Probably because of the above phenomenon, the ‘photograph stage’ comes much later than the ‘mirror stage’ in the subject’s ontogenesis. A baby has no problems in recognizing his image reflected in a mirror, whereas a child up to five years of age finds it very difficult (and requires some sort of training) to identify photographed objects. Indeed, he will perceive images as expressions referring to a generic content and, only through this connection with the universal, will he refer to the improper subject. He sees the picture of a woman X, considers it the picture of a type-woman, applies it to a token-woman Y and finally states that it is a picture of his mother. He actually fails to refer that proper-improper name, that slack designator represented by a photographic image. We are , witnessing a semiosic phenomenon.
Our pragmatics of photography reflects the effects of those early mis-takes. While testifying that the plate has been exposed to something (and, in view of that, photographic images can be used as evidence), it nevertheless arouses the suspicion that something has not been there at all. We know that, through staging, optical tricks, emulsion, solarization, and the like, someone could have produced the image of something that did not exist, had not existed, and will never exist. A photograph can lie. We realize that, even when we assume, naively, if not under the influ-ence of a fideistic attitude, that it does not.
The objective referent is conjectured and yet, at every moment, it risks dissolving in pure con-tent. Is a photograph the photograph of a man or the photograph of that man? It depends on how we use it (see Goodman’s remarks, 1968). Oc-casionally, on the basis of a surreptitious reference to general (universal) content, we take the photograph of X for that of Y. It is not just an error of perception, that is, as if we saw in a mirror the image of X coming in and thought it was Y. In fact, there is more to it: in any imprint, how-ever well defined, as that of an exposed plate, generic characters ultimately prevail over specific ones.
Except for catoptric theaters, the choice of the shot in the mirror is left to me, even when I am spying on someone: I need only move. Incidentally, if I see a half-length image of myself in a mirror, I need only get closer and look inside, downward, to see, within limits, that portion of my body the previous image did not show. The object is there, to produce the image, even where I did not see it at first. In contrast, with photographs, the shot is strictly set. I will never get the chance of seeing those legs if they are not in the image from the start: I just have to assume their existence (and still it could be the photograph of a cul-de-jatte). Again, the legs I presuppose are not one’s legs, but just one’s two-footedness. The impression of actual reference immediately faces into clusters of content. A photograph is already a semiosic phe-nomenon.
Second magic experiment: the frozen image moves. Motion pictures, obviously, to which all the remarks made on photography apply, plus the actual grammar of editing, with all the deceptive and generalizing effects it involves. Imprints, but moving ones.
Third experiment: the imprint has a very low definition, the mirror looks like an image freezer, and, on top of that, there is no longer a guarantee of the existence of a mirror and of a referent for the image. What I see is not only staging, shot, a selected visual angle, but also the result of the work done on the surface so that the latter seems to reflect the rays coming from an object: it is, in fact, a painting. In this case, all the requirements of semiosic phenomena are met; the physics of pro-duction combines with the pragmatics of interpretation in an utterly dif-ferent way from that of the mirror image.
Our three imaginary experiments have led us to imagine phenomena no longer related to mirrors. Despite that, when dealing with such phenomena, we can never totally abandon the memory of the mirror images of which they are monkeys (the same as art is always simia naturae).
However, it is worth reconsidering for a moment our experiments in-volving a sequence of mirrors placed at regular intervals along a row of hills. Let us assume we replace the sequence of mirrors with other de-vices turning the light rays coming from the initial object into electric signals which are then transformed into optical signals by a final device. The resulting image would have the same characteristics as imprints such as photographs and motion pictures; in other words, they would enjoy a lower definition than the mirror image (anyway, we decided to consider such an inconvenience as only temporary), they would be heteromaterial and retranslated (reinverted). And yet, like the chain of mirrors, such a system would seem to involve a rigid designation: the image would be determined by the present referent which causes it, and the relation would be from occurrence to occurrence.
Obviously, such a system, where a schematic model of TV transmis-sion can be detected, would only have this characteristic in case of live emission. As to the pragmatic attitude, a recorded TV broadcast does not differ from a film show, except for image definition and type of sensory stimulus. Only live TV broadcasts would share with mirrors their abso-lute relation with the referent.
The point is (and this may also apply to the set of mirrors reflecting a distant image) that just the space interval between referent and image arouses, more or less consciously, a suspicion of potential absence. The object should be there, but it may even not be. In addition, a further basic element should be taken into account: recorder emissions give rise to distrust in the audience as to the truthfulness of live emissions. From a pragmatic point of view, TV images share the advantages of mirror images as well as the disadvantages of the other photographic and motion-picture imprints. It is occurrence, acting as a parasite to the re-ferent, but not necessarily. Who can be sure? And how many and what manipulationns may have taken place along the channel? And what is the role played, not only by the shot, but also by the editing, which influences live broadcasts, too, and through which the camera decides which aspects of the real referent to explore and the mixing may produce Kuleshov effects at any moment?
However, such comparisons between photosensitive imprints and mir-ror images tell us at least something which is highly important for the semiotics of photographic, motion-picture, and television images. The latter lie within the boundaries of semiotics, but certainly not within those of linguistics. Each imprint is a projection working as a toposensi-tive whole, not as a sequence of discrete elements replicable by ratio facilis (Eco 1976, 3.4.9). The way all imprints (which are actual signs) can be interpreted is similar to the way we interpret a distorted or low-defined mirror image (which, on the contrary, is not a sign). The process develops by projective relations, a given dimension must correspond to the same dimension in the image, if not in the object-occurrence (re-ferent), at least in the object-type (content) the image ‘tells’ me about.
The actual grammatical categories come into play only in connection with shooting and editing. Imprints are not mirror images, but we read them almost as if they were. At a certain level of analysis —when, for instance, one is concerned only with iconographic conventions — one is entitled to look at photographic imprints as if they were real mirror images, that is, the immediate result of a reflection tout court, and their semiotic strategies will be investigated only at the highest manipulatory levels (staging, framing, and so on). In other cases, it would be, on the contrary, indispensable to cast in doubt their presumed ‘innocence’, to discuss their cultural origin, the non-naturality of their supposed causal relation with the referent.
7.16. The experimentum crucis
However strong illusions, ambiguities, confusions On the threshold’, and the temptation to rank
