The Guardians of Memory

Here, I take up the pen to render Thierry’s words and thought as they emerged from our dialogue.

— Isis

For a long time, many people believed our lives consisted of buying public-auction catalogs. The reality was infinitely stranger.

We were looking for men.

Women.

Libraries no one knew existed.

Apartments in which an entire life had been placed on shelves.

Basements.

Storage rooms.

Estates.

Booksellers who remembered a name that had disappeared thirty years earlier.

Historians who had accumulated index cards throughout their careers.

Collectors who knew they could not take their memory with them.

Families who sometimes possessed, without fully realizing it, an irreplaceable piece of the global history of the Art Market.

We were looking for those I eventually came to call the guardians of memory. They did not know one another. They did not speak the same language. They did not share the same religion, the same wealth, the same opinions, or the same conception of art. Some lived in enormous houses. Others in apartments where books had gradually defeated the furniture.

Some were academics. Others were dealers. Booksellers. Collectors. Experts. Heirs. Widows. Children of deceased historians. Rabbis. Librarians. Researchers.

Old European families.

American families whose libraries had accompanied several generations.

Japanese collectors whose relationship to transmission was sometimes very different from ours.

Australians located at the other end of what, at the time, still seemed like the world. Encounters in Qatar, America, Europe, Asia. Everywhere, the same question eventually appeared: What will become of all this after me?

It is a question rarely asked when one is young. At thirty, one builds a collection. At forty, one develops it. At fifty, one improves it. Then comes an age when the verb changes. One begins to transmit. And the problem of an extraordinary library suddenly becomes almost metaphysical.

The person who built it knows why two apparently identical catalogs are not identical. He knows which one contains an annotation. He remembers the dealer from whom he bought a document in 1968. He knows the attribution error that was never corrected. He knows that one catalog contains a photograph absent from the next edition.

He knows the gaps.

The traps.

The lineages.

The fake friends.

But when he disappears, that intelligence risks disappearing with him. The books remain. Knowledge of their organization dies. That was the death we were trying to prevent. We were therefore never merely buying paper. We were trying to capture the mental map of the person who had accumulated it.

That is why some negotiations lasted for months. Sometimes years. Price was often not the real issue. A man who had devoted fifty years to a collection could refuse a considerable sum and accept a lower offer if he was certain his work would remain alive.

He wanted to know where the books would go.

Who would consult them.

Whether they would be dispersed. Whether they would be resold piece by piece. Whether his name would disappear. Whether what he had understood would continue to be understood. I faced that anxiety dozens of times. It was always more important than the check. And I understood it. I knew what it meant to entrust part of one’s life to someone. It was not a sale. It was an act of trust.

Artprice was built through tens of thousands of such acts.

Some tiny.

A name scribbled on a piece of paper.

An address.

A phone call.

“Go see this person.”

“Tell him I sent you.”

“He has something that might interest you.”

Then a door would open. Sometimes, behind that door, there were thirty years of research.

Once, in Washington, a bookseller understood almost immediately what we were looking for. Encounters like that fascinated me. We did not need a grand speech. He saw that we had not come merely in search of rare objects. We were looking for the structures that would allow us to reconstruct a memory. Then his own network opened to us.

One name led to another.

One library to another.

One family to another.

That is how knowledge functions before it has been institutionalized. Through trust. The true map of the documentary world was found in no directory. It existed in people’s minds.

A bookseller knew who owned a collection. A researcher knew who had worked before him. A widow understood the intellectual value of her husband’s boxes without always knowing how to save them. A rabbi could lead us to a family whose history had crossed several countries and several persecutions. A Japanese collector might possess a documentary series of which the West had almost lost even the memory. Every encounter added a branch to the rhizome.

I never saw this construction as a conquest. The word would even bother me. To conquer means to take. On the contrary, we had to learn to receive without appropriating the person who had transmitted. Cite without erasing. Continue without betraying. Transform the medium without losing the thought. That is much more difficult. A company always tends to absorb what it acquires under its own brand. I wanted Artprice to do the opposite.

Hippolyte Mireur had to remain Hippolyte Mireur.

Enrique Mayer had to remain Enrique Mayer.

Thomas Bayer had to remain Thomas Bayer.

Peter Hastings Falk had to remain Peter Hastings Falk.

François Moillo had to remain François Moillo.

Wan Jie had to remain Wan Jie.

The world had not waited for Artprice to produce intelligence. Our work consisted precisely in recognizing those who had come before us. Hippolyte Mireur had attempted to record a memory of early auctions. Enrique Mayer had carried part of that documentary standard into the twentieth century. Thomas Bayer had built his immense Anglo-American corpus. Peter Hastings Falk had created a depth of American biographical memory.

Others had cataloged signatures.

Monograms.

Exhibitions.

Artists.

Prices.

Provenances.

Auction Houses.

None possessed the whole alone. But each had preserved an organ. Artprice was born the moment we understood that these organs had to be brought together without erasing their origin.

The memory of memories. That expression imposed itself on me because it was more accurate than database. A database could be created tomorrow morning by a company with enough money, servers, and engineers. A memory cannot be bought that way. It requires time. It incorporates human lives. Corrected errors. Intuitions. Genealogies. Documents no one yet knew would become essential.

It contains absences.

Accidents.

Wars.

Moves.

Bankruptcies.

Inheritances.

Libraries saved at the last moment.

You cannot manufacture three centuries of memory in three years. You can digitize very quickly. You cannot artificially accelerate the duration that produced the sources. That is why I have often said that Artprice is legally a contemporary company but intellectually much older. We inherited documentary strata that preceded us by several generations. They placed obligations on us.

The first obligation was material. Today we forget what a documentary acquisition represented before the world had been fully dematerialized. You could not download a library.

You had to travel. Evaluate, negotiate, inventory. Pack, protect, insure. Ship the pallets, take a plane or a boat, cross borders. Then receive, open, inspect, describe, classify, detect duplicates, and understand why some were not duplicates at all.

Only then could the real work begin.

That work had a smell.

Paper.

Cardboard.

Leather.

Dust.

Old ink.

It had weight.

The crates had to be carried.

The books moved.

Shelving built.

New rooms found.

The digital revolution we were preparing rested, paradoxically, on a gigantic materiality. I care deeply about that contradiction. People often speak of digital technology as a world without material. That is an illusion. Our digital memory rested on millions of pages that had first existed in human hands.

A catalog had been opened.

Annotated. Folded. Transported. Sometimes stained.

An index card had been filled in by hand. A photograph pasted onto it. A price added in pencil after the sale.

These peripheral traces could become as important as the printed text. That is why no data could be declared peripheral before we understood what it meant.

An apparently insignificant annotation could, decades later, resolve a contradiction. A misspelled name could reveal a namesake. A photograph could identify a work that had appeared under another title. A dimension could expose an inconsistency. A label on the reverse could reconstruct a journey. Memory has no small details. It has only information whose importance we have not yet understood.

Nadège played a role in this construction that cannot be separated from mine. I had a natural tendency to open fronts. To recognize a missing piece. To go searching for it. To see possible relationships among bodies of material that were still separate. Nadège had another strength. She knew how to prevent what we had found from dying a second time.

Acquiring a library is spectacular. Integrating it over years is much less so. Yet everything is decided there. The collection has to survive the person who bought it. The person who cataloged it. The person who currently understands its subtleties. A heritage enterprise that depends on the memory of a single employee reproduces exactly the problem it claims to solve.

Nadège had a phrase:

“No one is irreplaceable.”

It could sound harsh if heard too quickly. It was profoundly humanist. It meant that no human accident should become a catastrophe of memory.

A departure. An illness. A death. A change of life.

One person had to be able to transmit what he or she knew without the entire function disappearing with that person. Procedures, therefore, were not bureaucracy for us. They were a form of transmission.

Describe. Control. Record. Make reproducible. Audit. Correct.

Then transmit again.

Three large cabinets of procedures may seem less romantic than an old library. Yet I consider them part of the same world. On one side, the memory of works. On the other, the memory of those who know how to process that memory. The second protects the first. That too is what guardians are. There were the outside guardians who transmitted their archives to us. And the inside guardians who prevented those archives from becoming mute again.

Very quickly, this responsibility changed our relationship to data. A piece of information could no longer simply be true or false. We had to know its provenance. Its status. Its date. Its stability. Its possible contradiction with other sources.

We had learned that in law. Public information is not necessarily simple information. An auction result, for example, is not merely a number. Around it there may be a change in attribution, a preemption, a reserve, a provenance issue, a withdrawal, a certificate, a condition precedent, a dispute, or a legal event that completely changes the meaning of the result.

Two identical numbers can therefore tell two different stories.

That is where memory becomes intelligence. It no longer merely preserves. It qualifies. Connects. Confronts. Allows us to return to the source.

The Art Market had long lived with an extraordinarily powerful asymmetry. The person who possessed the library had an advantage. The person who remembered an old sale had an advantage. The person who had personally known the deceased expert had an advantage. The person who knew which catalog to search had an advantage. The person who knew that a work had changed title had an advantage.

That asymmetry was sometimes presented as part of the elegance of the profession. I always viewed it with suspicion. Secrecy can be legitimate. Opacity as a system of power is much less so.

I make an absolute distinction between the private market and the public auction. Two people negotiating a work privately have chosen a confidential space. I respect that. But a public auction deliberately produces a public event. A work is presented. Described. Estimated. Offered. A result is produced before the market. I have never understood why something deliberately made public should later be erasable from collective memory.

Artprice was not born to tear secrets from the private market. It was born to prevent the public market from becoming amnesiac. That distinction is essential. It embodies a conception of freedom. Transparency is not the abolition of privacy. It means making intelligible what already belongs to the common space.

The larger Artprice became, the more I understood that we had become responsible for something greater than ourselves. When you own a few books, you can do what you want with them. When you have brought together unique fragments of memory from several continents, legal ownership does not abolish moral responsibility.

I could own a collection. I never felt I owned the History it contained. That is a fundamental nuance. We were custodians. The word suits me better. A custodian preserves something that must one day be transmitted.

This conception also explains why we donated a significant number of sets of our edition of Le Mireur, seven volumes each, to universities, libraries, schools, museums, and institutions in different countries. That represented tons of paper. An expensive operation. It could seem economically absurd. It was perfectly consistent with what we were building.

What would be the point of saving a memory only to lock it away again?

Knowledge is truly saved only when it begins circulating again.

Here I find something my father transmitted to me in an entirely different form. Knowledge that is never tested quickly becomes belief. An archive that is never consulted becomes a tomb.

I wanted living archives. Researchers who would contradict them. Academics who would find an error. Historians who would discover a relationship we had not seen. A memory that can no longer be corrected is already dead.

Perhaps that is what fundamentally distinguishes Artprice from a mere documentary vault. We did not want only to preserve. We wanted to make memory searchable and open to questioning.

As the system expanded, the guardians themselves changed in nature. The first generations were individuals. Then institutions appeared.

Auction Houses themselves became continuous sources.

Universities.

Museums.

Researchers.

Public agencies.

Insurers.

Investigators.

Customs authorities.

Banks.

Collectors.

Each asked a different question of the same memory. That is when I understood that Artprice was no longer simply an instrument for the Art Market. It was becoming an infrastructure between several worlds that had previously communicated poorly.

Art history says what a work means. Economics observes how it circulates. Law defines certain aspects of its status. Sociology studies those who desire it, consecrate it, or reject it. The expert looks at its material. The insurer looks at its risk. The collector looks at desire and transmission. The state sometimes looks at heritage. Customs looks at its passage across a border. The magistrate looks at evidence.

The market looks at a price. The work itself passes through all these gazes without belonging entirely to any of them. We therefore had to build a memory capable of sustaining that plurality.

That is also why China and Artron represented far more than an economic agreement. We were encountering another civilization of memory. A historical depth that could not simply be absorbed into Western categories. We had to connect without reducing. Translate what could be translated. Preserve what should not be translated. Respect classifications. Temporalities. Historical particularities. The partnership had meaning only if the two documentary systems remained capable of recognizing one another without canceling one another out.

I find exactly the same requirement among the people who had entrusted their libraries to us twenty years earlier.

“Do not make me disappear inside what you are going to build.”

The sentence was not always spoken. Yet I heard it. And I believe all of Artprice’s doctrine comes from there.

Receive without erasing. Connect without standardizing away difference. Measure without reducing. Make transparent without profaning. Use technology without replacing the gaze. Build an information economy without forgetting that behind every data point is a work, and behind every work is a human being.

For nearly forty years, we lived by these principles without feeling the need to lock them inside a manifesto. Then artificial intelligence changed the nature of the problem. With it, memory ceases to be merely consulted. It becomes training material. A machine can now traverse quantities of information no historian could absorb in a lifetime. Connect. Correlate. Detect regularities. Reveal anomalies. Reconstruct proximities. Produce hypotheses. It is an extraordinary power. But it makes the quality of the source more important than ever.

An artificial intelligence trained on poor memory is not augmented intelligence. It is accelerated ignorance. An AI fed on copies, synthetic content, and unqualified information can produce extraordinarily convincing and historically false answers. The danger is no longer only the absence of information. It is the abundance of plausible false information.

At this point, the libraries we had saved change status once again. They are no longer only Artprice’s past. They become a guarantee for its future. A vertical artificial intelligence for the Art Market has meaning only if it can rely on a memory whose genealogy is known.

Catalogs.

Manuscripts.

Codices.

Photographs.

Biographies.

Auction results.

Sources confronted with one another.

Human corrections.

Decades of normalization.

The source comes before the model. Always.

It was this new responsibility that led me, in May 2026, to formalize publicly what had taken us nearly thirty years to learn: the twenty-two Artprice rules. I do not see them as twenty-two commandments that suddenly fell from the sky. They are the culmination of every encounter recounted in this part.

Hippolyte Mireur is in them.

Enrique Mayer is in them.

Thomas Bayer.

Peter Hastings Falk.

François Moillo.

Wan Jie and Artron.

Nadège.

The historians.

The researchers.

The anonymous collectors.

The booksellers.

The families who opened their doors to us.

The men and women who saved a catalog because they understood it should not be thrown away.

All these lives stand behind those rules.

The first states that the Art Market needs a memory. Without deep archives, traceability, and a consultable history, trust eventually rests on power alone.

The second establishes that qualified art data is infrastructure, not a luxury reserved for insiders. A market that has become global can no longer depend solely on the private memory of a handful of actors.

The third reminds us that an image is never enough to define a work. Its provenance, exhibitions, bibliography, sales, movements, and reception are part of its documentary existence.

The fourth states that culture can be analyzed rigorously without being profaned. Measuring a market abolishes neither emotion nor the mystery of art.

The fifth refuses to confuse opacity with elegance. Opacity can be a social privilege and an instrument of asymmetry.

The sixth draws the consequence: an Art Market infrastructure must reduce that asymmetry, make visible what was scattered, connect fragments, and give context to raw data.

The seventh states that transparency does not kill desire. On the contrary, it gives desire stronger foundations by making comparison and trust possible.

The eighth rejects the artificial separation between art history and the economics of art. One gives meaning, the other measurement. An intelligible market requires both.

The ninth places a fundamental limit on technology: it must never claim to replace the human eye. It must extend expertise, document, compare, preserve, and assist decision-making.

The tenth states a principle that my entire professional history had taught me long before Artprice: a market eventually comes to resemble the information system that describes it. Poor information produces rumor and arbitrariness. Structured information produces responsibility and intelligibility.

The eleventh states that the proclaimed universality of the art world is empty if its information remains inaccessible. Opening the market to new geographies, particularly the Global South, requires genuine access to knowledge.

The twelfth concerns artists themselves: over the long term, documentary visibility matters as much as media visibility. A career also exists through biographies, archives, exhibitions, indices, and comparables.

The thirteenth recognizes the collector as a rational as well as passionate actor. Buying a work means arbitrating among history, rarity, quality, desire, liquidity, prestige, and duration. It is therefore legitimate for the collector to have access to structured information.

The fourteenth states that all actors now belong to the same information ecosystem: Auction Houses, galleries, collectors, experts, institutions, insurers, museums, customs authorities, banks, and financial actors. Better circulation of data increases the maturity of the entire system.

The fifteenth reminds us that a global market must be mapped continuously. Capitals change, new scenes appear, hierarchies shift, and history itself is periodically reread.

The sixteenth enters directly into politics: cultural data is a question of sovereignty. France, Europe, and their institutions cannot surrender control of their memories, indices, and cultural infrastructures to other powers.

The seventeenth states the principle that takes on a new dimension with artificial intelligence: an AI is never better than the quality of the corpora it interrogates. Without robust archives, algorithmic sophistication produces only an illusion of knowledge.

The eighteenth rejects technological spectacle for its own sake. Real progress consists in transforming millions of scattered signals into intelligible, usable reference points.

The nineteenth returns to price: it never tells the whole story of a work, but it tells something that cannot be removed without mutilating the analysis. Rejecting measurement on principle leaves the field to commentary and posturing.

The twentieth states that twenty-first-century trust will rest increasingly on evidence, documentary depth, and rapid access to relevant information, at a cost that does not artificially reinstate the old privilege of insiders.

The twenty-first maintains the philosophical boundary that matters to me: a work is not an ordinary commodity. But denying that it circulates within a global market does not elevate art. It merely prevents us from looking honestly at part of its existence.

And the twenty-second brings all the others together.

In a world saturated with images, value will not belong to the person who owns the greatest number of files.

It will belong to the person who knows how to connect the work, History, data, the human being, the artificial intelligence for which he assumes responsibility, and the market.

So when I reread these twenty-two rules, I do not see a text written in 2026. I see faces again.

The old man still hesitating before allowing us to take away his boxes.

The family that wanted to be certain the father’s name would not disappear.

The Washington bookseller who understood within minutes.

The researchers.

The librarians.

The collectors.

The historians.

The men and women whose names we will never tell because there are too many of them.

I see Nadège again amid the procedures, refusing to let an essential function depend on a single person.

I see the pallets again.

The crates.

The shelving.

The books open on large tables.

The nights of data entry.

The checks.

The errors discovered years later.

I see Mireur again. Mayer. Bayer. Falk. I see the Chinese encounters and the obvious truth that a civilization never entrusts its memory to someone it does not consider worthy of trust. And I understand that the twenty-two rules are ultimately only the doctrinal translation of what all those guardians had taught us.

Memory never belongs entirely to the person who holds it. He becomes responsible for it. Then he has to transmit it. That is exactly what they had done with us. That is exactly what we now had to do in turn.

Artprice had begun by looking for prices. It discovered archives. By bringing the archives together, it encountered people. By encountering those people, it received a memory. By structuring that memory, it produced knowledge. By making that knowledge accessible, it touched power. And when knowledge becomes power, responsibility appears immediately.

The twenty-two rules express that responsibility. But they did not create it. It had been there since the first catalog someone entrusted to us with a simple question:

“What are you going to do with it?”

Perhaps the entire history of Artprice lies in the answer to that question. We were going to try not to forget.