This text brings together personal reflections arising from systems built since 2009. It is neither a universal doctrine nor an additional regulation of the Bureau des Protocoles. It proposes an inquiry and, at the end, an experiment that others may attempt.
I. Before the system, an observation
I build poetic systems and put them into operation. Yet I never begin by setting out to make poetry. I begin with an observation: a practice that takes time, an expression people repeat without performing it, a rule thought self-evident, sometimes a scientific reading that displaces an ordinary question.
I learn that movement affects the measurement of time under the conditions described by relativity. I turn this into an instruction: go to a post office and walk until closing time. Walking does not slow time for everyone. The physical effect is minuscule and depends on the comparison being made. But something I read has become a precisely situated action, actually performed.
I notice that people already pray on behalf of others. Religious sisters carry out this task. If prayer may be delegated to a person, why not to a machine? I know the French expression “jeter un pavé dans la mare” — to throw a cobblestone into a pond, meaning to cause a stir. If someone has no time to get angry, why not have another person perform the gesture?
These questions do not begin as jokes. They begin with a possibility taken seriously. The next task is to build the means of carrying it out.
From observation to operation
This route describes possible experiments, not a law that all works must follow.
II. Transferring an operation
I built a machine for answering prayers. Requests are encoded in binary and converted into flashes of light sent towards the sky. Users choose a religious operator: Catholic, Islamic or Jewish. The machine has physically separate cells for these operators.
Users are then asked whether their prayer was answered. Their responses provide a rate of self-reported answers for each operator. These replies do not prove divine intervention; they nevertheless form part of the operation of the service I constructed.
One day, a cell failed. While it was being repaired and its equipment replaced, Catholic prayers were transferred to the Islamic cell. A religious practice had encountered an ordinary problem of maintenance and service continuity. None of that is necessary for praying. All of it becomes necessary when one decides to operate a prayer machine.
The original function was not invented: prayer predated the machine. What I invented was another way to organize its execution. Transferring the practice creates constraints, incidents and possibilities specific to the new system.
III. Producing without a product
I also built a system for producing Nothing. On a website, a button reads “Produce Nothing”. When clicked, nothing visible happens. The system automatically generates a manufacturing number corresponding to the Unix time of the operation. A unit of Nothing is then produced and identified according to the rules of the device.
The system does not prove that a material called Nothing has appeared. It executes a production process that delivers no physical object. Its distinctive feature is that the absence of a product may be the expected result of a real operation.
If the program failed and generated no number, the absence of a product would look much the same. Yet production would not have taken place. We must distinguish producing nothing in accordance with a rule from failing to execute the rule.
Units were subsequently sold through CHEZBelette, and the tax authorities claimed VAT on them. These events did not create the production process. They show that an operation instituted by an artist can encounter the ordinary workings of commerce and administration. The sale is an additional fact. The invention, here, was production.
IV. Can absence perform a function?
I made L'Internet, a box plugged into the electrical supply. A diode flashes: the device is powered. It provides no access to the Internet. No page can be consulted, no video watched; no digital activity is offered. In this sense, I say that it connects its user to reality.
It does not create a new physical connection to the world or guarantee the user's attention. It makes available a device whose stated function is precisely to provide no digital stream.
I also built an inverse box, whose effect is to trip an electrical protection device. In the first case, the lack of service is expected; in the second, interruption is the event. These devices do not prove that every inaction is an action. They investigate the conditions under which absence or interruption may be intended, organized and recognized as a result.
Operation does not necessarily mean a useful service. And what appears to be a fault in one context may be the intended operation in another.
V. Giving zero a form
Paslaz began with a different question: could I hold in my hand something that physically exists but, in another sense, does not exist? On the same post are the entry and exit signs of a town. The post and signs can be touched; the distance between entry and exit is zero.
The image I had in mind was subtraction: 1 − 1 = 0. Zero can be written, but not grasped as one grasps a sign.
From that operation, Paslaz acquired other attributes: signs installed by the public works department, and mail deliveries. This does not make Paslaz a legally recognized municipality. It makes certain attributes of a town operate without assigning it an area. Here, third parties are more than witnesses: their actions are part of the system's operation.
VI. Taking laughter seriously
My constructions often make people laugh. I do not, however, ask them to be jokes. They interest me when it becomes possible to say, without qualification about their execution: “this was built”, “this operates”, “this act took place”.
The Bureau des Protocoles arose from that demand for seriousness. It receives requests, carries out acts, maintains a register and seals executions in 1.3 kg lead cubes. The protocol Colère (“Anger”) may involve throwing a cobblestone into a pond. The Bureau certifies the execution of the gesture; it does not certify anyone's emotional state.
The Bureau has its own doctrine of ballast, set out in seven articles. It imposes specific conditions on its own acts. It is not a general rule to which all my earlier inventions must conform.
Seriousness does not correct laughter. It removes the refuge of “it was only an idea”. The operations have been carried out; their meanings remain open to discussion.
VII. What is said, what is done, what can be established
An assigned function, an actual operation and a proof are not the same thing. They are tempting to conflate because they often coincide in ordinary machinery. A lock closes; the door resists; the result can be directly perceived.
The systems described here sometimes separate these dimensions. The Nothing program creates a number in accordance with its rules. The prayer machine transmits signals and collects responses without proving that prayers have been answered. L'Internet powers a diode without measuring a connection to reality. The Bureau carries out acts and retains traces of them without supplying psychological proof.
Three levels to distinguish
An operation may be real without proving every effect associated with the function's name.
This distinction does not take a function away from its system. It clarifies what its operation establishes. It also requires a definition of failure: when an intended operation does not occur, its absence cannot be confused with an absence deliberately produced.
VIII. Why make things work?
I do not necessarily seek to improve existing systems or prove mine superior. Sometimes I take an existing function and change its operator; elsewhere I move a rule, take an expression literally or treat an absence as a result.
What interests me is that a proposition which could have remained an idea becomes available as an operation. It receives conditions of execution. It can be requested, ignored, found to have failed, or unexpectedly come into contact with institutions.
I call this proposing a possible alternative. Not a solution everyone should adopt, but a way of operating that was not needed and whose existence can now be tested.
This practice does not provide a general answer to “what is a function?” It raises another, more shareable question: what would need to be built for a proposition, however superfluous, to be genuinely put into operation?
IX. An open experimental protocol
This document can do more than describe my own systems. I therefore suggest trying the method. The following steps are neither an artistic doctrine nor a recipe that guarantees poetry. They are a working procedure, open to criticism and modification.
- Observe. Start from a fact, practice, rule, expression or piece of knowledge. State exactly what caught your attention, without immediately turning it into a metaphor.
- Displace. Examine what happens when an operator, material, scale, context or condition is changed. Sometimes simply taking words literally is enough.
- Name the function. Write a sentence that can be evaluated at the level of the protocol: “this system is intended to…”. Distinguish that function from effects one might hope for but be unable to verify.
- Build. Specify actual means of execution: device, program, organization, people, duration, access, instructions. A description alone does not constitute putting a system into operation.
- Define failure. State what counts as a complete execution and what would count as a fault. If the outcome is an absence, decide how to distinguish it from a failure of the device.
- Put into operation. Check the device, then make it usable under real and safe conditions. The system need not be useful; it must be able to carry out its rules.
- Observe third parties. Distinguish their roles: user, operator, recipient, institution, witness. Their participation may be part of the operation or add an unforeseen event.
- Record and revise. Keep traces of what was done, failures, discrepancies and unanswered questions. Amend the protocol when trials reveal its limits.
Poetry does not authorize deception, endangering others or appropriating their work. Safety, consent and legal requirements continue to apply to any experiment.
X. A worksheet to reuse
Anyone wishing to build a system of their own can begin by filling in this worksheet. Answers need not be spectacular. They should be precise.
| Question | To be completed |
|---|---|
| Initial observation | What practice, rule or fact started the inquiry? |
| Displacement | What changes: operator, place, material, aim or condition? |
| Proposed function | What is the system intended to accomplish? |
| Device | What material, human or software resources are needed? |
| Success condition | What operations must have been executed? |
| Failure condition | What fault or missing execution could be observed? |
| Third-party roles | Who requests, operates, observes or receives? |
| Trace | What should be kept to describe what took place, without claiming to prove more? |
XI. An inquiry to continue
This method grows out of work I began in 2009, but it is not its final truth. It does not entirely explain why one observation attracts my attention rather than another, nor what gives an experiment its poetic force. It makes it possible to examine the passage from an observation to a device and its operation.
If someone builds a system using this protocol, they can show the results as well as the dead ends. A method that never faced an objection would be less interesting than one others could correct.
I build poetic systems and put them into operation. This text proposes trying, elsewhere and differently, what becomes possible when an idea is taken seriously enough to be made to work.