Berlin, Akademie der Künste, 12. 10. 2012
Whenever I think about the relation between film and theoretical physics I feel quite alone. I have yet to meet someone in the film scene who understands what I'm aiming at and why it could be of any relevance. The only exception was the wonderful Hollis Frampton1, but too soon he passed away, and that was that. On the other hand, when I talk to physicists they have serious problems understanding the complexity of the film medium. Let’s see if we can change that today. So the theme of my lecture is film and theoretical physics. Both are highly complex phenomena. And to give a short overview of what "physics" is, is meanwhile beyond anybody's capabilities. But "film" is a strange invention, too.
The first thing I want to point out is that it has in the meantime completely changed its material base. It started with something we call celluloid, a mechanical thing, mechanical stuff. And now it's an electronic phenomenon. The material basis has been completely transformed. Despite that, our idea of film has stayed more or less the same. That's strange. Generally when technologies become obsolete the content the technology used to transport undergoes significant changes. So that's interesting.
The second point is: When film was invented everybody was impressed by its ability to present motion. That was obviously a big plus in relation to photography. But it in fact just added the attribute of motion. The more important invention was installed three or four years later when different film strips were glued together. That happened by pure chance. If another material carrier had been the basis it wouldn't have been possible. But as film had sprocket holes and could be cut with scissors and glued together it became possible to generate "film programs". And when those programs, in which different space-time sections had been put together following some "idea", were presented to an audience, the audience had to cope with it. That was the origin of imaginative editing. Today we can clearly see that editing was a much more far reaching invention than the reproduction of motion was. Because by editing and discovering the potential of what I call "linear cuts", time constructions became possible that were very hard to achieve in other art forms. Literature has a bit of that potential, too, but that's all. What is a linear cut?
Consider a person sitting in one space and you film it. Then you place him in a different location and film that, too. Splicing those shots together, one generates a linear cut. Seeing it, you think the person has moved from space one to space two. Because we inject the laws of physical causality into film for some reason, we assume that the person has moved and that he needed some time for that, a time not directly presented in the film. At the cut point there is a time jump. And that's something really interesting, because it made larger amounts of time accessible for the imagination. One could say that by film's potential of being edited it became suddenly possible to depict all the time in the world. And because it furthermore introduced the concept that one can depict large scale temporal developments it paved the way for the usual interpretation of Schwarzschild's symmetrical solution of Einstein's field equation, which is known to us as the big bang theory. Before that, everybody was thinking in terms of a static universe. So it might not have been pure coincidence that cinema came before the theory of general relativity.
Something similar happened to Husserl's phenomenology. When he first described how we see objects - that was in 1911 - meaning the way we imagine for example the invisible backside of a cup while actually seeing only the front, his formulations were very insecure, almost obscure. The second formulation2 of 1926 was much clearer, and I think not only because he was a better writer then, but also because of the movies. For in most movies one has the experience that the camera looks behind an already familiar object in a later shot. And we compare that with what we have imagined about the object's backside before. So the phenomenological thought that we see more than what we actually see was meanwhile a widely shared commonplace experience, and formulating it in the framework of phenomenology had become much easier.
That as an introduction. In both cases the movies made the access to new modes of thinking possibly a bit easier. But from now on we want to argue the other way around, and walk into areas where theoretical physics can offer some help, if we want to understand certain complex film phenomena. Before I really start I have to point out, however, that the physics I'll be discussing will not really "explain" those phenomena, at least not in a mathematical way. If you expect that, you will be disappointed. I will just present certain concepts, and then give some idea of how those concepts might be applied to certain confusing aspects of film, hoping to arrive at different, maybe fresher, ways of looking at them. So keep in mind that the transfer of ideas into the film medium will mostly be understood metaphorically.
--- looped projection of Lumière's "Arrival of a train"
So the new semester has just started, a reason to show Lumiere's "Arrival of a train" again. I don't know how often I have seen it, fifty times maybe, not as often as I've seen my own films, of course, but pretty often. The last time was probably 10, maybe 20 years ago, though. Showing it again last week I was amazed to what extent it was different from what I remembered of it. It's not that I suddenly saw things that I had never seen in it before, for in the course of the film I slowly remembered that I might have seen the stuff in previous viewings – but in the very moment I encountered the images again, I had completely forgotten most of the details. And that is very interesting because it leads us to the question: What do we remember of a film while we see it? Or what do we see at all when we see a film?
In the case of "Arrival of a train" I definitely remembered the basic geometrical structure. I knew the train was going from right to left and to the front, I remembered people climbing out, I even remembered reflections on the train and I think a few trees in the background. Not much more. Last week I was above all overwhelmed by the multitude of people moving about the station, and the interesting ways people were dressed. I had thought there were only ten or a maximum of twenty. I had actually been at the station, where the train had been filmed in I guess 1896, in La Ciotat. While shooting my film "The Open Universe" in Marseille3 I undertook the effort to go there. For filmmakers it’s a magic place. It's hard to tell why, because the film was not part of the legendary first film presentation on the Boulevard des Capucines4, but that doesn't seem to matter. For most filmmakers it serves as a metaphor for the beginning of cinema nevertheless. It was a boring station, though, not very interesting, eventually a train came in but it was electrically driven, and just one person got out. But even after the visit the film kept its magic. So I was thinking of that while I was watching it last week.
I also realized another effect: While looking at specific details which I momentarily thought I had never seen before, I could not see any other details, I mean: at the same time. That is also very interesting. We can only look at one thing at a time. And that’s a problem, because when we see a film, we only see its images for a certain period. With paintings it's different: At the paintings in a museum we can look as long as we want. In narrative films the usual shot length is maybe three seconds. The Lumière film, well, despite the fact that I've seen it at least fifty times, I really don't know how long it is, well, maybe it's forty, fifty seconds long. That’s quite long for one shot, but we can look at paintings much longer. So the limitation of time forces us to be satisfied with an imperfect perception of the whole thing; it's never complete.
One other thing is obvious: when something is moving in a film image, our eye almost by reflex gets attached to this moving thing. On the other hand we know that space is a very important quality of images. So we also try to at least get an idea of where the space might be, in which the picture had been recorded. But unfortunately we can only see one thing at a time when we watch a film image. So when we concentrate on the movement, we have only vague ideas about the space in which it takes place. And to concentrate on the space is only possible when we forget about the movement. That’s reminiscent of something one encounters in physics, something all of you have heard of. It's Heisenberg's uncertainty principle: If you want to locate a particle you can't at the same time determine its, to put it perhaps over-simply, velocity. Because if you want to locate the particle perfectly, you practically have to stop it. So in this case you have an uncertainty in the velocity. On the other hand, if you want to measure its velocity, it is not in one spot anymore. So there is a certain spatial uncertainty, and the product of the two, which was Heisenberg's discovery, is a constant. That's quite remarkable. For some reason something similar seems to happen in cinema. If your attention concentrates on a certain thing you have to neglect other things. If you concentrate on the movement you forget about the space. And the other way around: You can only concentrate on space when there is little movement in an image, let's say if you have a wide angle shot of a forest. Only when there is no spectacular motion-aspect is one capable of concentrating on the spatial quality. So I think there is an uncertainty principle at work when we watch images appear in a film.
-- end of looped projection of "Arrival of a train"
One could of course think, that the uncertainty will vanish if the shot is long enough. Now it’s a pleasure to have among us the magnificent movie master Michael Snow. Some of his films have very long lasting shots indeed.5 But for some reason one can't exhaust them, they become more and more complex the longer you watch. There are many miniscule things happening, and they all seem to superimpose. It's very hard to describe what's taking place inside you while you watch his films.
In the avant-garde there was (and still is) a direction that concentrated on films with long lasting shots and quite slow developments within the image, especially in the ‘60s and early ‘70s. I guess it started with Warhol's work, but significant contributions were made by – to drop a few more names in addition to Michael Snow – Ernie Gehr, Barry Gerson and Larry Gottheim.6 And it was astonishing to realize how interesting they were, how inexhaustible. Another, almost opposite, direction used very fast editing, and in these films – I’d like to mention the work of Kurt Kren7 – it is amazing how much information we can collect when we see something that's edited very fast.
To give you an example of that, I’ll show you a piece I shot in 1973, when I was interested in narrative editing. I wanted to explore how narrative editing worked. At that time there were no video recorders yet, so whenever you started to theorize about editing, you had to work with the memory of films you had seen, trying to recall how their editing operated. That was very unsatisfactory. So to overcome this handicap I sat in front of a TV with a Super-8 camera on a tripod and watched films on the regular programs. And whenever there was a cut in a chosen film I pushed the trigger and stopped the camera immediately after. So each time, at each cut of a movie, the camera registered three or four frames of the appearing shot. Each shot was thus kind of recorded, and at the end you can see the complete montage system of a chosen feature film within a few minutes.
I suspected of course, that it might be impossible to see anything if you have just 3 or 4 frames of each shot. But then I was surprised that one could understand the whole film even in this concentrated form lasting just 2 or 3 minutes instead of the 90 minutes of the original. Let's look at one of the short versions now. It's part of a 100-minute work called "Elementary Film History", which I finished in 1974.8
----- projection of the compact version of "Kiss of Death" (2 minutes)
The original film of this section is Henry Hathaway's "The Kiss of Death", made in 1947, with Victor Mature. What do we see in it? We see how our attention is attracted by movement, of course. Like a reflex: whenever something moves – 4 frames seem to be enough – our eyes want to follow. But we see many other things. Mainly the montage figure you are watching now, which is called "shot-reverse-shot". It occurs very often, almost in an inflationary mode, I would say. And we see also that when images reappear, we go back to an almost identical framing. I call this the "dominance of the return-cut" in the narrative film formula. If you look at it in this fast form, there seems to be a formula behind it indeed. And the return cut is by far the most frequent cut. Now the physicist comes into play; as a physicist you want to find out what the formula might be. It's not that difficult to get a grasp of it. The regularity of it, that you cut back to the same framing, that the same patterns appear in different guises, all that makes it possible to develop an understanding of what narrative film editing is about.
-- the film is still running
Well, of course first of all it's about faces. Ah, look, Richard Widmark is in it, too. The face, and that you recognize faces, is the most dominant structure in narrative films. That has little to do with physics of course. But then there are always residues behind the faces, residues of something the physicist would call "space". A film is something that modulates "space" in the course of time. Oh, now we have reached the end: very important this title "The End", because it makes it clear that the film is finished. But let's get into loop mode and look at the film again.
-- compact version of "Kiss of Death" is running from the start in loop mode now
When mathematicians or physicists think about space, they usually think about "coordinates". Up to Einstein it was not complicated to attach coordinate systems to each space you saw. In wide angle shots, let's say here, in this room, it's easy to establish a coordinate system. And here you see that these total shots get decomposed into close-ups. And that there are two different kinds of cuts: most of them are "return-cuts", which go back to the same face again and again and again. And sometimes there is a type of cut in which the space is switched. These are the "linear cuts" we have already talked about. They are linear in the person we follow when he or she changes his or her location. See, here we are in a prison, and now, one of the persons moves to another space in a car. That was a linear cut, connecting 2 spaces. As a physicist you would immediately say that these cuts represent coordinate-transformations. Coordinate transformations are very simple mathematic structures. They have the wonderful quality that they are additive. So when you make a coordinate-transformation you can add another one. And you have a result. If you have a series of let’s say 5 coordinate-transformations, you can replace it by a single cut. You can calculate with it, so to speak. And you can make films much shorter. So it's not important how the coordinate-transformation actually happens. If you have a camera in position one, you can put in a box, take a taxi, make a big detour, go to location two, record location two. When you cut the two together, the viewer directly connects the two and it does not matter which path the camera took in between. Coordinate-transformations split up into simple translations and rotations. And another one is called "zoom", with which you magnify things. To a certain extent you can describe all the cuts you are just seeing on screen as space translations, space rotations, and zooms.
There wouldn't be much of a mystery about it, if there were not the factor of time. Time makes the whole thing very complicated, because there is a time segment attached to each of these images. Sometimes you have time jumps, and sometimes you have the feeling that time passes smoothly, that it's continuous. In narrative films you usually have the impression, that there are hardly any time jumps. But these you register carefully. Most of the time, however, the action appears to be continuous. For example in shot-reverse-shot systems you seem to have zero time loss at each cut. Whereas in linear cuts, when you change locations, you often encounter time jumps. Because Richard Widmark's body needs some time to go from space A to space B. And because he can't be in two spaces at the same time. Quantum physics taught us that there are some things that have different properties. But narrative films usually describe classical bodies which behave properly.
If you watch this 2-minute film a few times you get a clearer and clearer understanding of what a movie is. Actually after you have seen it a few thousand times I bet, ten percent of you could formulate a complete editing theory of the narrative cinema. Something you can discover in almost every film that can be seen on the screen, even today, something you can also discover in TV series. Of course most of the narrative stuff in TV is not as ambitious as an art film, but nevertheless you will observe most of the structures there, too. You could call it a universal pattern, one might even dare say (as many do) that there is a "language" hidden behind it. A universal language. I once met a person however – Grahame Weinbren9 – with whom I talked about "film language", and that it might have a grammar, and yes, that I possibly had discovered a grammar there, in this short film, which you are just seeing. But he said: "No, Klaus, that's impossible. In a grammar you have the word "no". It requires the word "no" and there is no word for "no" in cinema." And I think Grahame was right. A film image says "yes" all the time. No film image says "no". Film is affirmative. You trust that it's real. At least it's very difficult to construct the word "no" in cinema. Much more difficult than in a language, where saying "No" is the easiest thing. Maybe all languages started with the word "No", the rest of it is embroidery, are attempts to rationalize excuses. So the narrative structure I had discovered is just a structure, nothing more. Just as music is a structure, a certain way of structuring sounds in very specific ways, and not a language. There is no language of music and there is no language of film. So much for that.
--- end of the looped projection of the compact version of "Kiss of Death"
So this was an example of how theoretical physics can come into play when you analyze certain phenomena of cinema.
Now we come to a field which is more complex. Going back in my memories, I remember several conversations with my friend Paul Sharits10, most of them conducted casually while drinking in bars and so forth. In such talks filmmakers usually don't talk much about film, and if they do, it's generally just in the form of short sentences, very short remarks. But for some reason one keeps some of those remarks as precious jewels. And one of Paul Sharits's casual remarks, I think from the early ‘70s, was: "I hate cameras, I love projectors." That astonished me, because I thought the camera was at the heart of film production. But he said: "No, the camera bores me. Projectors are the really interesting thing." Well, at that time I of course thought he was just being provocative, but his remark had such a paradoxical ring that I kept thinking about it. And slowly I came to think Paul was right. Projectors are really the heart of cinema.
So I came to think that a film is something in which a projector shoots particles at the viewer. The terminology is strikingly fitting: a projector is a "thrower", and films are ordered ensembles of so called "shots" whose impact on the audience is carefully calculated by filmmakers. All the time something is hitting the viewer from the screen. This "something" is modulated light that emerges from the projector, is reflected on the screen and reaches your eye from there. Exactly that was operating in Tony Conrad's and Paul Sharits's first films. In some of them there were even warnings that they might induce epileptic seizures. I refer of course to "The Flicker" and "Ray Gun Virus".11 And just as Paul's title suggests: in those films the projector was perceived as a "gun", shooting colored light rays (or rather light particles) at the viewer in order to see what would happen. Of course, one can't be satisfied with just, let's say in the extreme case, destroying the viewer’s brain by high energy flicker.
As a trained physicist this reminded me of something. As a student I was working at "DESY" for a while, a so called electron synchrotron,12 an early form of the hadron collider that recently made the news in connection with the Higgs particle. In DESY we were accelerating electrons to high speeds and shot them on selected specimens of matter. Just to see what happens. So there was some parallel, for in Tony's and Paul's films we also have particles shot at a target, the particles being light particles, of course. But what was the target? First of all it is the human eye, naturally. But it's more than just the eye, it is the human brain, human consciousness, human perception.
So I thought: "Hey, what if Paul's ‘ray gun’ does more than just throw colored light at the viewer, infecting his brain with a virus? What if the projector ejects particles which are images?" And what does a viewer make of it when those "image particles" contain objects he wants to perceive in the very moment they hit his brain? Could there be a connection between this type of setup and the scattering experiments I helped set up as a physicist? So in order to find out a few things about it I had to produce a film that could be used in such an "image-particle" experiment. At that time I also became interested in Beethoven, so I suddenly – it was 1977 – decided to make a film out of his last piano sonata13, taking the music score Beethoven wrote and transferring its rhythmical structures into a visual form. Meaning I generated a shooting score of it, following Beethoven's tempo values plus his accellerandos, meno allegros and ritardandos. So the unmodulated allegro con brio with metronome 120 would mean a quarter note is 12 frames long, an eighth 6 and a sixteenth note 3 frames, etc. Using the single frame features of my Super-8 camera, I then made a film of it, right in the camera, so that for each note of the music a picture of the same length would pop up.
When the film was finished (and screened of course at 24 frames per second) it turned out to be quite spectacular. For a while I was showing a silent version14 of it, because it was not so easy to play the Beethoven Sonata in a way that it would fit the tempo the changing film-images adopted. It actually took me almost thirty years of practice till I finally managed to perform something that could be approximately synced with my visual interpretation of Beethoven's score. But I finally managed, and if you look it at now, you will experience that the images really hit your brain like particles. So let's look at it, it’s the beginning of a 80-minute film called "Hommage to Ludwig van Beethoven", which I finished in 2006.15
--- projection of "op. 111-1" (8 min.), the first part of "Hommage to Ludwig van Beethoven"
Well, thank you. So now you were part of a scattering experiment. Picture-particles got shot into your brains and generated all kinds of brain-particles. But what do you make of it? That's really the question. Even I have no answer, although I've made the film. And even though I've seen it maybe a hundred times. Also in this special case we encounter the question that arose in Lumière’s "Arrival": What does one keep in memory of a "film"? In this film the question becomes even more dramatic: So what have you kept in your memory? I think it's impossible to answer. Every one of you probably has a different memory of it. The only reliable answer that all of us can admit to is: "I remember that I've seen this thing." But you can't really say what you saw. One can of course remember a few structures, and denote some of them by certain words, but one knows immediately that these words describe just a minuscule amount of what was offered. In a certain way the idea of memory of a film becomes completely ridiculous in this case. What you remember most is maybe the surprising experience you had, and how it affected your body.
In this film the unit is the shot: a more or less representational shot in which you can recognize a few things. Then the next shot appears, sometimes four frames long, sometimes three, two, sometimes twenty frames or even more. Of course one recognizes hardly anything in the very short ones unless they have some drastic graphic quality, only then you register recognizable stuff. In the shots lasting longer the images can be subtler. And having observed this we are able to proceed to a new chain of arguments, turning the screw a bit tighter: Let's try to find out what all films have in common. This film definitely had very fast editing. In narrative films the editing is, as we all know, considerably slower, you have cuts at an average of let's say three seconds. In some films the average shot length might be longer, of course, in action sequences the editing is often faster.
Trying to find out what all films have in common I am again grateful for Paul Sharits's remark "I hate cameras, I love projectors." Nowadays projectors are almost museum pieces. The knowledge of how they work is not common knowledge anymore. On projectors you generally see 2 reels. On one, called the feed reel, is the part of the film which has not yet been projected. The other one is called take-up reel, on it you have the part of the film that was already shown. From the feed reel the film is transported by sprocket wheels that give stability to the transport, into something called "the gate". In the gate the light emitted by the light source hits and penetrates the film and from there the image is thrown at the screen, from where it hits your brain. In a certain sense the image in the gate is the present tense of a film. So you have the future on the feed reel, the projected past on the take-up reel, and the present tense in the gate. Now, what exactly is the present tense? It's something philosophers have argued about for a few thousand years. Even physicists are surprisingly insecure about it, as we have seen when we mentioned Heisenberg's uncertainty principle. In film it seems to be comparatively simple. The present tense is what is in the gate, it is presented to us by the picture you see at this very moment.
Discussing Lumière's film, we have already found out, however, that we only get hold of sections of the projected image, our attention does not go beyond certain parts. That’s true also for films that have hundreds of shots, of course. But while we look at the parts that catch our attention in the present shot, we also have a memory of what is on the take-up reel, meaning of what has already been shown. In our brains we have definitely stored residues of that. The strongest memory is of course the memory of the preceding shot. To get an idea of what is taking place when we watch a film, we can use a simple model. Just assume that the new "picture particle" which is in the "gate of presence" is aimed at a "pool of impressions" that your brain has distilled from all the shots that have appeared before. And on the surface of this "pool of collected impressions" floats, as if on a raft, the last image you have seen, the preceding shot, the preceding event. The impressions of the other ones, of the shots having appeared before (meaning the ones that are now on the take-up reel), are much less distinct, they seem to have a somehow fluid aspect, that’s why we choose the metaphor of a "pool". As the memory trace of a film is not concentrated in an area with clear boundaries, the term "pool" of course lacks precision. So "sea of impressions" (which alludes to the well know "electron sea model" that proved useful in understanding electric conductivity) might be in some respect more appropriate, but then the model would lose some handiness. So let's stick for now with the "pool".
So whenever a new shot gets into the gate, an "image particle" is ejected from there. Reaching the brain it hits the pool of impressions with a big splash. Doing this it hits the raft floating on top first. The raft (presenting our memory of the preceding shot) gets destroyed or it at least loses its distinctness, so that most of its structures disappear within a fraction of a second, while some remnants start sinking down. Meanwhile the present particle already works havoc in the memory-liquid, where it modifies and destroys a considerable amount of the impressions deposited there. As if a meteorite hits the sea a lot of the memory-stuff splashes out and gets lost. Having finished its destructive job – all this happens within half a second (and it keeps on going) – the picture particle drifts up to the pool's surface, forming a new raft there, which now floats on a "sea of changed impressions", getting more and more structure within its remaining projection time – till the next picture particle will be in the gate, by which the present raft will also be destroyed and the pool modified anew. So that works as model, an approximate model of what happens, when we see a film. I think it's true for all kinds of films, also the Beethoven film with the extremely fast editing. But it's also usable for narrative motion pictures. In each case you get impressions from images and they get modified or destroyed by the next shot. Somehow a pool of those impressions vaguely remains and when the film is over, the remaining pool plus the impressions of the last shot (preferably a happy end, because it feels good when it's superimposed on a film) is what you think you have seen when you leave the cinema.
And this again is very interesting if you are a trained physicist. Because this model also has some resemblance with how scattering experiments in physics are set up. In them you also have particles being shot on something which often has a complex structure, and this structure gets destroyed or at least modified by the process. We have already observed that even watching a seemingly simple film like Lumière's "Arrival" is a complicated affair. You can't see the whole thing completely, because there are several uncertainty principles at work. In particle physics similar effects occur. In many cases the uncertainty principle also restricts the precision there. The processes taking place when high energy particles hit a target often become so complicated that it's practically impossible to describe them by a compact mathematical formula. That's why one restricts the physical analysis to an analysis of the incoming particles and the outgoing ones. The function that compares incoming and outgoing particles is called "scattering matrix". And in many areas of contemporary physics the scattering matrix represents the ultimate knowledge you can get out of an experiment. Because the uncertainty principle is in the heart of matter and it's impossible to really describe what is taking place in all its minute details. At the core of all things is an unstable, unimaginable boiling process, which by principle is not describable.
Generally one can say that whenever you realize that something like an uncertainty principle is at work, then the only way to get to a reliable understanding of what's taking place is by examining the scattering matrix. Now, what is the scattering matrix in the case of film? The simplest model we find in early film history again, in the period where you had only total shots of the locations in which the action takes place and no close-ups. In those early films you usually see the present location, let's say a room, then a few people would enter, they interact, and then they leave. Obviously that is a scattering process, the incoming particles being the persons entering the room, the outgoing particles the people leaving it. The scattering matrix would describe how they have been changed by their interactions. These changes can be obvious, for example if a person gets terribly wounded, or so subtle that one hardly notices a change. But even then there must have been some change, otherwise the shot was useless. One can say, the reason for a scene appearing in a film is always that it induces subtle changes in the acting persons. Otherwise the scattering matrix of this particular shot would be trivial and the shot superfluous.
In the case of the Beethoven film the impression pool is quite chaotic. The new shots almost destroy the past impressions, only very little remains of them. Also every viewer has his own pool, so to speak, everybody sucks different impressions out of the film, dependent on his personal capabilities and his momentary psychological disposition. In narrative features it's different, the impression pools of narrative films have a certain stability, and that’s why every viewer stores approximately the same impressions of certain aspects appearing on screen. Narrative stability is exhibited in 2 layers. In the first, all the locations the viewer has seen in the preceding shots form a compositum, usually a connected fabric of spaces, almost in the form of a geographical map. And in the second layer he registers the characters that have entered those entangled premises and how their dispositions get changed by their interactions. When a new interaction takes place for the same person, the changes add up.
So what is the scattering matrix of a whole film? Very simple. In our deliberations of what happens in a single shot, we have to replace the location, in which the interaction takes place with the entangled fabric of spaces generated in the course of the film. And then you have to consider the changes that have affected the characters acting in it. Again you have their state when they first entered the movie, and the state in which they left it. So each movie is a scattering experiment with actors, and the changes in their dispositions are described by a scattering matrix. If you like, it presents the "content" of a film.
That’s of course just the final situation, when the film is finished, meaning when it's completely on the take-up reel. When it's still entirely on the feed reel, one doesn't even know the actors and of course all the spaces that will appear are also still unknown. You might have a speculative idea of it though, because you might have heard about the film, in an advertisement or from friends, or because you have an idea of what kind of movie a specific actor prefers to perform in. But there is no actual scattering matrix yet. Just as the impression pool is a function of the projection time so is the scattering matrix. If you stop the film in the middle, you have only memories of the actors that have appeared up to this point. And how they have changed until "now". All this is expressed by the "momentary scattering matrix". As soon as the projector starts again, this momentary matrix gets modified by what happens in the next shots, more or less according to our "pool model". Interesting is how the composite space-fabric gets constructed in the course of the film, thus giving us a "realistic" playground for what is affecting the actors. This space construction can of course be described with the help of space-time coordinates. Which brings us back to "The Kiss of Death" and the editing theory we talked about. The developing space-time fabric is constructed mainly by linear cuts. Return cuts use this fabric and give a ground for what is described in the scattering matrix. Well, I hope I haven't overstrained your attention. If you became interested in it, you can read more in my editing theory book. The last 100 pages of it deal with the aspects we have just talked about, relating them to Feynman Graphs and a mathematical field called topology16, but that’s beyond the scope of this talk.
One might of course ask if that's not a bit too complicated. After all, film is just an industrial commodity, put out by thousands of small companies generally not run by geniuses. So there shouldn't be much of a mystery behind it. And the rules used to fabricate those films can't be very complex. Well, I think that’s true with respect to "making" films, at least to a wide extent. Most films are made following very simple rules. But don't forget that we have described what a "viewer" makes of it. And that is something that takes place in our brains. So what we described is our perception of a projection of what was "made" following simple rules. There the whole territory of our brain comes into play. And one of the great mysteries of cinema is that each viewer has to perform the act of seeing without the help of others. Brakhage has expressed that very precisely by calling one of his films "The act of seeing with one's own eyes".17 Well, at present nobody knows how a film is stored in our brain and how the brain manages to make something out of it. But we have at least developed a plausible model. I doubt that neuroscientists are capable of presenting a simpler one. And more than that: I'm afraid a simpler one will never be available.
But again: why bother about something that will become obsolete soon? Meanwhile it's obvious that the mechanically based cinema with its celluloid, its projectors, showrooms and reels will have no future. And here I come back to the opening statement of my talk, where I mentioned that "film" has managed to change into an electronic phenomenon without changing its character. I'm afraid the narrative film-form that was developed in the last century, with all its editing strategies and cuts and scattering matrices and so forth, will remain with us, will remain with humanity for at least a few thousand years, experiencing (and surviving) several more transitions of its material base. To a great extent its future is still hidden from us on the feed reel, so to speak, in the future of our civilization. But in the gate of presence we continuously get glimpses of it – and we leave them behind us on the take-up reel of our past achievements. Thank you.
OK, so I think we agree that it's quite complicated. The film was shot in 1977/88 with a single frame Super-8 camera in locations where I grew up as a boy. That made shooting easy because I felt at home. It looks like a lot of work but it was actually not that much. The work energy put in was nothing compared to what an experimental physicist has to invest if he wants to prove a hypothesis. It took me hardly two weeks. Each day I went to the locations I had chosen at ten o'clock in the morning. With me I had my script, based on Beethoven's score, and then I was just, you know, a little bureaucrat, clicking on my camera, all the time counting single frames, hoping not to make mistakes while putting filters in front of the lens for certain shots, making fades, all that kind of stuff. The next day I went there again and so forth, and within three weeks it was finished. After this I had enough confidence to get busy with the second movement of "opus 111", which of course is also part of the "Homage to Ludwig van Beethoven". For that I filmed "three voices", three layers of rhythmically edited images, whose rhythms I derived from Beethoven's score. After the shooting was finished I superimposed them with an optical printer. The result turned out to be even more spectacular than the film you just saw, of course. But that is not the point, at the moment. Let’s rather come back to our question "What do we see in a film and what do we remember?"