Value of the original

Why does a perfect copy of a painting feel worthless when no one can tell it from the original?

Suppose a copy so exact that no eye, and no instrument, can separate it from the painting it copies. Every visible property is identical. If what we valued were the picture — the arrangement of colour that produces the experience — the two should be worth the same. They are not: one is priceless and the other is a curiosity. Either we are being irrational, or the thing we were valuing was never entirely visible. Which is it?

Let us test the first option. Is it snobbery? Then discovering a beloved painting is a forgery should change nothing, since the experience it gives is unchanged. But it changes everything, including for viewers with no financial stake — and the case is not hypothetical. Han van Meegeren painted Vermeers that experts accepted as Vermeers in the 1930s and 1940s; when he confessed in 1945, the pictures did not alter by a single brushstroke, yet they moved from masterpieces to museum oddities. Whatever collapsed was not visual.

So ask what a painting is besides an image. It is also an event: a particular object that a particular person made, at a particular moment, by hand, and which has travelled through time since. Call the record of that travel its provenance — the documented chain of ownership and location running back to the artist's studio. The original has a causal link to the artist's hand. The copy has a causal link to the copyist's.

Does that link do any work, or is it sentiment? Consider what we actually want from looking. Part of it is the image, but part is contact: this is the surface the artist worked, these are the corrections they made, this is what survived. Walter Benjamin called the original's presence in time and space its aura, and argued that mechanical reproduction is precisely what strips it away. You can accept his description of the phenomenon without accepting his whole verdict about art's future — and it does describe what a viewer notices in front of a real object.

Now the economic half, which is more concrete. Appearance is not scarce; a good copy can be produced, and a photograph of it distributed endlessly at no cost. Authenticity is scarce, and irreducibly so: there is exactly one object with that history, and no process can manufacture a second. So a market pricing paintings will price what is scarce, not what is abundant. That is why prices track attribution — who made it — rather than beauty, and why a reattribution can multiply or erase a value overnight without anyone repainting anything.

Notice the two accounts are not rivals. One says the original carries a history the copy lacks; the other says histories, unlike appearances, cannot be duplicated. The scarcity claim explains the price; the aura claim explains why anyone cares in the first place.

Now the honest complication. This whole picture assumes a work with one authentic object — and many works have no such thing. A photograph is printed from a negative, many times over, so what would "the original" mean? For prints, etchings, and cast sculpture, the authorised edition is the work, and the copy/original line lands somewhere different: on whether the maker or estate authorised this instance. Digital works have bit-identical copies by construction, and what is sold is a record of entitlement, not a distinguishable object. And in conceptual art, some works consist of instructions plus a certificate — a Sol LeWitt wall drawing can be executed by others on a new wall, and the certificate, not any painted surface, is what is owned. There, provenance has not been added to the object; it has replaced it.

Three refinements connect the steps.

First, "the copy is worthless" is too strong. A copy has the value of a copy: it can teach, decorate, and reproduce the image faithfully. What it cannot do is be the thing that was there. Losing a category of value is not the same as having none.

Second, the forgery case is contaminated by deceit, which is a separate wrong, so it is worth isolating. Take an openly made, unsigned replica: no fraud, nobody misled — and it still is not the painting. That isolates the point cleanly, because what remains missing is only the history.

Third, a caveat about where this reasoning applies. Everything above concerns works whose identity is tied to a made object. For music, literature, and choreography the question barely arises: a novel's identity is its text, and a well-printed copy is not a lesser novel. Whether visual art's attachment to the object is a deep fact about art or a contingent fact about how visual art has been made and sold is a live argument, not a settled one — and the certificate-based works are the clearest evidence that it can be arranged otherwise.

• How attribution is actually decided, and what happens to a picture when the experts change their minds.
• Why reproductions of famous paintings increased, rather than reduced, the crowds at the originals.

Every term the collection defines is gathered in the glossary.