"A chameleon only has the color of a chameleon when put on top of another chameleon"
- F. Cavanna
Any tool - a hammer, a dagger, a stone axe, a brush, a pencil, a bookshelf, a plastic bag - is supposed to have one or several functions, that are the socially agreed upon purposes it has been assigned.
For instance, you can - but you are not supposed to - kill your neighbour with a hammer, a pencil, a brush, a bookshelf or a plastic bag. However, a dagger or a stone axe are considered more adequate for such a function, whereas a car - although cars kill much more people today than daggers - is not agreed upon to be built for killing people. Every tool and machine has a set of functionalities and this set of functionalities is a convention.
Because each tool - each machine - is socially built for serving its socially agreed upon purposes, it socially appears as determined and limited by these functions or purposes. Actually, this does not apply only to what we usually call "tools" or "machines". It applies just as well to a number of concrete and abstract objects that are usually not considered as tools or machines although they are not different in their essence, such as words, languages, magical recipes, theories, techniques, game rules* and more generally any sort of social rules* ...
In other terms, anything that has ever been created, built, modified or even destroyed to serve a determined purpose, socially appears as determined and limited by these functions or purposes.
Hence it is critical to establish and keep in mind a clear distinction between
1 - the functionality of a machine, which is plainly a set of conventions about its usage,
2 - the machine itself which, although built by human beings, is an integrated part of the Real and as such unknown to a very large extent,
3 - the effects of a machine which of course include the intended effects that belong to its functionality, but also include a wide range of unknown and non intended effects, some of them being only discovered after centuries or millenaries of use* .
In other terms, a machine cannot be reduced to its socially agreed upon aspects, it cannot be reduced to its functionality. A machine is part of the Real and as such, the consequences of introducing it in the Universal History are largely unpredictable.
Realism is quite precisely this belief that the World is what it is agreed upon to be and Surrealism is the experimental process making obvious that the World is always much more than that* .
Michael Richardson is a realist because he believes that a machine is limited to and by its functionality. When he writes that surrealists in the past have been "undermining the machines very functionality" he is right of course. Unfortunately, what he actually means is that the surrealists purpose was undermining the machines themselves.
This is yet highly erroneous since Duchamp, who is by far the best and deepest surrealist theoricist as regards machines, is also known for having built several devices, made several inventions and even to have participated in at least one inventors show in Paris (Le concours Lépine).
As a matter of fact, Surrealism has been using the same approach as regards machines and technique, as it did regarding perception and forms. This approach, as is indicated by the "sur" in the word "surrealism", has always been a movement of dialectical negation and hence of extension.
Indeed, while the shortest and most efficient way to undermine the functionality of a machine is quite simply to break it, the surrealist approach, following the most dialectical paths, has always been to extend the functionalities of machines and tools up to the point where these functionalities appeared for what they were : an arbitrary and usually rather comical set of conventions and social decisions.
The major effect of this dialectical approach was to make the sovereignty of human freedom obvious also in the technical area. As Duchamp himself explained, the choice that lies at the core of the ready-made represents a key point of the artist's freedom.
I hope I succeeded in making clear that :
(a) - the functionality of a machine is nothing else than a set of conventions, decisions and points of views regarding this machine.
(b) - As such the functionality of a machine, tool or object is nothing else than Realism applied to the technical area, nothing else than a point of view according to which this object, tool or machine is "just that". In a surrealist perspective no machine, tool or object is ever "just that". Quite on the opposite, it is always "much more than that". Thousands of points of view may be cast on any object, tool or machine. Hundreds of surrealist usages may be found for each object and man is granted 24 hours of freedom per day as Scutenaire used to say.
(c) - No machine can ever be reduced to its functionality. If this was the case then the ecological critique of this world would have no object. The least that may be said is that there are huge evidences that this is not the case... And more yet to come.
This is the situation on basis of which Michael Richardson feels entitled to conclude that "The computer is in its very essence anti-surrealist" for being "the machine of functionality par excellence".
According to the above developed argumentation, Michael Richardson's point of view is essentially a realist non-sense because from a surrealist point of view, nothing can be said to be or to have ever been "of functionality par excellence".
But even from within Michael Richardson's realist point of view, we shall see further on that such statements are rather quick.
Computers are sometimes called "universal calculators". This expression refers to one of the first computers ever created : Turing's Machine.
Turing's Machine is an imaginary sort of machine, only supposed to be able to read instructions from a paper tape* and to go forward and backwards along another paper tape in order to read or write data. As a matter of fact, Turing's Machine is nothing else than a given set of rules and the only thing that ever operated such a machine I fear, is the human mind.
Now, why did Turing invent such an extremely simplistic and apparently limited sort of mechanism? In other terms, what was the intended function of his machine?
Well, what Turing had in mind was just a simple model that would allow him to study what can be calculated and what cannot be calculated. Something that would help him to draw a map that would show the frontiers of machines and determine what could be enclosed within these frontiers and also what could not be enclosed within them.
Alan Turing's project was after all not so different from the project that Michael Richardson suggests with the title of his article: "Far from the frontiers of machines". There is nevertheless a huge difference between the two approaches and I would tend to call it work!
While Alan Turing sets up a model, plays with it, patiently studies its limits, reaches strange conclusions and results of a yet unknown nature and finally points out where some of the possible limits of machines may actually be located, Michael Richardson proudly stands still on a position strongly built from ready made answers.
Demonstrating an amazing ignorance or a pretentious contempt of more than half a century of quickly developing cognitive sciences, Michael Richardson does not produce one single new argument, and to be honest does nothing more than collect worn out statements according to which the human mind would essentially* be "far from the frontiers of machines".
Michael Richardson apparently does not even notice that he did not make any effort to identify such frontiers, a prerequisite to which on the opposite Alan Turing contributed* in the brightest way. Actually Michael Richardson's article rather suggests that he never had the slightest intention to seriously explore the domain of machines.
The first thing to take into account as regards understanding the nature of machines and even of simple tools, is the critical importance of the view point. For instance, a hammer has a function in a carpenter's vision, but none from a bee's point of view.
Similarly, a computer can only be said to calculate from a human being's mind perspective, a perspective in which calculations and their results have a meaning within a network of mental metaphors linking a model to what it is supposed to represent in the Reality.
Calculations actually only operate at the level of models and if the model is abstract enough, such as the set of axioms defining integer numbers for instance, then the result as such essentially has no concrete meaning (e.g. 10 + 10 = 20 for change).
It is only based on a network of metaphors linking the model to its counterpart in Reality that the result may be given some semantics (e.g. 10 kg of extra weight + 10 kg of extra weight = 20 kg of extra weight which is beyond what I am allowed to have on this cheap flight).
From the the starts and the Moon's perspective a computer does nothing more than a leaf falling from a tree and when considered from such a non human point of view, one should say that the computer "flows", like water does in a river, rather than it "calculates".
So, although when considering a tool we mainly happen to see its function, we should keep in mind that in a certain way, it is some sort of hallucination. A lion or a bear would probably not see the function.
Also, Turing's machine was indeed quite similar to a game. If we consider it from this point of view, telling in which way such a game could essentially be different from a surrealist game as regards its primary orientation - exploration of the human mind - may prove to be difficult. But of course, in their shrinking understanding of Surrealism, some people who claim to valiantly work for the liberation of the human mind provided it is amputated of its mathematical, predictive and calculation abilities will stick to the opinion that Turing and Breton can by no means deal with different aspects of the same fundamental quest.
Yet, beyond the fear and blind hatred of anything like human scientific or technical imagination as it is promoted by people of whom one would normally expect to support and encourage - as a bare minimum - unleashed freedom of thinking, it should be clear that what Turing expected from his model machine cannot properly be called a "function", or else surrealism could said to "have a function" as well, which is exactly not what I think of surrealism..
Now of course, behind Alan Turing, stood people who were actually not so deeply interested in the possible limits of calculation as they were of the results of calculations. And quite obviously among them were the Intelligence Service departments where Alan Turing worked during World War II. These people were utilitarian* and of course, for them* , a computer had a function : providing results. Answers!
A computer is essentially an interpret of any possible function (and of a lot of other things indeed, that have nothing to do with functions), just like a theater actor for instance is potentially an interpret for any possible character in theater plays, or just like a shaman is potentially an interpret for any mental trip, or just like a chameleon is an interpret for any specific color you put it on.
So, just like the chameleon is an animal and not a color, just like an actor is a human being and not a theater play, just like a shaman is a human being and not the mental trip he does, a computer can simulate or interpret any functionality, but it is not functionality as such. The functionality is in the program executed by the computer, in other terms, a functionality is made of language* , which seems consistent with the nature of functionalities as conventions that I attempted to describe previously.
Roughly said, since a computer can interpret or simulate a lot of things and situations and it is more or less adequate to answer any generic question starting with "What if...". And yes, this is a limit because, while answering questions is something computers can do, computers are currently not so clever at inventing or discovering new interesting questions. This situation has been summarized in a both strange and funny way by the adepts of the Artificial Intelligence language called Prolog : "if Prolog is the answer, then what is the question ?"
Are we then facing here an essential limit of computers or just a temporary stage that would be due to some weaknesses the current state of the art as regards the understanding of the human ability to invent new sorts of questions ? Worse... Would not such a situation be the bare result of utilitarianism itself ?
Anyway, it appears that in a pleasant and humorous manner, the so said "machine of functionality par excellence" turns back towards the utilitarian and his stupid ideas of mechanical slavery and ironically whispers into his ear : "Yes Master, I shall do as your ordered and provide you with the suitably intelligent answers, provided you feed me with suitably intelligent questions...". And then of course starts the utilitarian sort of quest that J.L. Borgès described so well in "The Library of Babel", a quest to find the question, the answer of which will solve all problems* .
Anyone who believes a tool or machine to be a slave should seriously consider wiping the last shadow of the spirit of slavery out of his own mind. The original Sanskrit root "uti" for "tool" only meant "a help".
Systems evolving beyond prediction
Also the concept of function, while reasonably useful in engineering areas, is a bit too easy going in most of the others - biology included. Being essentially a question of (human) point of view, it is not very surprising that it may be suitable in some situations and inadequate in lots of others.
Let's assume that - like is now quite ordinary - a computer has been programmed for simulating the evolution of a population of living beings...Let's go a step further... A computer can also simulate systems that are not computers but are of a different nature such as neural networks or cellular automata.
These systems are made of numerous interlinked nodes (or cells), each node reacting, accordingly to a given set of (simple) rules, both to its own previous status and to the current status of its neighbouring nodes. Conway's game also called the Game of Life is a classical example of such systems.
It has been mathematically proven, that in certain configurations and under certain circumstances, neural networks - although totally deterministic in their behavior - can evolve in such a way that the succession of their states can absolutely not be predicted - except of course by re-executing the entire history of the network step by step.
What happens is that the shortest way to tell the story of their evolution is to follow this evolution again step by step. Such evolution paths are actually irreducible to anything else than themselves
But then, what sort of functionality could humanly be assigned to such a system, since in a veryy absolute manner, no one can tell where it goes ?
And it is that bad indeed, that engineers easily come to consider that neural networks cannot be of any interest because, since their behavior cannot be predicted, they cannot be used for any purpose.
Then, what could possibly be the function of something that cannot be used ?
Worse... (and that will establish that the nature of computers is not so easy to determine) Not only a computer can simulate a neural network or cellular automata, but the reverse is true.
It has been proven, that in other configurations and under other circumstances the evolution of even the simple Game of Life can build the components of a computer.
In other terms, it has been proven that the Game of Life itself is equivalent to a universal calculator. And this, although the program fed into the computer was never been for such a purpose, but just to simulate any network of cellular automata.
I shall not go further about neural networks fascinating abilities - but I should, as they are capable of discovering/creating forms, of learning, of creating analogies etc... - and I shall not go further either about the extremely interesting conclusions that can be drawn from some of their evolution paths regarding the nature of freedom.
For the time being, I shall simply conclude that the only way to pretend to be "far beyond the frontiers of machines" is to experiment with machines, study them and play with them like Kolmogoroff, Alan Turing, Gregory Chaitin and biologists like Francisco Varela, Gerald Edelman, Henri Atlan and a lot of others in the various branches of modern sciences did, and certainly not to lazily run away from machines pretending to be "far beyond" something you did not have the guts - or more plainly the occasion - to barely identify.
*game rules : such as the rules of surrealist games, "which no matter how strange they may be are still rules, and even especially rigid ones" as Michael Richardson would say.
*social rules : such as the written or unwritten rules based on which a surrealist group exists, "which no matter how strange they may be are still rules, and even especially rigid ones" as Richardson would say too.
*the realist painter : F. Nietzsche
*centuries or millenaries of use : to the extent that a living beaing - bacteria, plant or animal - is considered from an utilitarian point of view, it has the same function as a tool or a machine. As such, the neolithic is a technological revolution that established a technological system or actually several of them (Ivan Illitch used the megamachine terminology for such systems).
We now know that most of the diseases we get are connected to the technological systems we are using, which - among other effects they have - create new ecological conditions for the germs, hereby sometimes selecting the worst bacterial or virus lines.
Plague, for instance is propagated by rats. But rats do not tend to live together with human groups as long as these human groups have to move permanently because their way of living is based on hunting-collecting technologies and hence on natural resources that would quickly disappear if the human groups were to stay too long in the same places (Cf. Marshall Sahlins "Stone Age Economics").
Things of course change drastically when human groups sedentarize and survive based on a agricultural practices. Then rats and human beings live together and plague may be transmitted from rats to men.
The same is true regarding influenza which remains hidden within pigs or chicken populations before it "skips" to human beings. When human populations do not live in close contact with pigs and chickens, there are little chances that influenza viruses get adapted to the conditions they meet in human bodies. This is why influenza still kills entire indian tribes in Amazonia.
And I would like to add that if the cow disease called "vaccine" could provide a protection against the human smallpox, it's probably not by chance.
Similarly, if horse blood extracts (serum) were of some efficiency against tetanus, it's certainly not by chance either.
I guess that I do not need to further insist on the role of technology as regards tuberculosis, cancers, etc...
Indeed, ecological disasters are by no ways new. They begun with the introduction of the neolithic production system which itself was the result of the over exploitation of wild game resources by wide scale hunting in some areas of the world.
Going back to the hunting and collecting production system would certainly be a good solution if it did not unfortunately imply a reduction of probably more than 99% of the current world population.
*paper tape : containing the "program" and "data" of Turing's machine
*The World is always much more than that : because the Real is always much more than the Reality, Realism is a permanent disappointment while Surrealism on the opposite expects and often discovers and creates marvels.
*essentially : since Richardson provides no details to establish or just explain such an obvious "surrealist" truth , we are left wondering whether the human mind is supposed to be "far from the frontiers of machines" due to its platonician essence or just by God's will ?
While a strong debate has been going on on this point for decades in the field of Artificial Intelligence among people who know what they are speaking of, Richardson pops up and pretends to clarify it all in just one sentence ! I feel happy that he did not choose to prove in the same "surrealist" fashion that virgin Mary was the mother of God or of a cow, a hen, or a pig.
*contributed : for instance, there are interesting connections (equivalences) between some aspects of Turing's work and Gödel's theorem.
*utilitarian : based on the fact that utilitarianism is one of the most official ideologies of the current capitalist system, I always found odd that revolutionaries persist in accepting to consider tools and machines under the sole aspect of utilities.
Based on ecological considerations proving that any of our so said "utilities" can turn into plain nuisances or even frank enemies of our species, I personally long have come to the point of considering the old animist approach as highly superior and far preferable.
As long as you consider your tools with enough care and respect to understand and keep in mind that they are "not only that" - I mean, that they are not only the function they have socially been assigned - you unconsciously stay on your guards about possibly unexpected aspects that they may take at any moment.
As I attempted to make clear, you may somewhat contemptuously consider your refrigerator "just as" a utility for cooling things, but if you care a bit and respect this machine, you may also learn what sort of fluid it uses and understand that your innocent machine can also be a fierce enemy of the ozone layer and hence of yourself.
*language: yes, in language and not Matter. Which tells a lot about the actual power of poetry in this (physical) world
*problems : the utilitarian essentially does not understand that questions are not meant to be solved but on the opposite to open new problems. As Duchamp clearly stated, there is no solution, because there is no problem.