The state nobody describes

The acquisition paperwork for this impression contains a phrase that looks like a fact: state III of IV, without signature. Six words. Two of them turn out to be checkable, and neither checks out cleanly.

“Of IV”

Start with the easier half. Does this plate have four states?

In one catalogue, yes. The New Hollstein volumes divide it into four, and two major print rooms confirm it independently: the British Museum cites impressions as 170.II and 170.IV, and the Rijksmuseum encodes its three sheets as 170-1(4), 170-2(4) and 170-4(4). That is real corroboration from institutions that had the objects in front of them.

But Bartsch, writing in 1797, recognised one state. Hollstein's original series: one. White & Boon in 1969: one — and the National Gallery of Art's record puts it in words, “White/Boon 1969, no. 20, State only.” The same National Gallery of Art then cites the New Hollstein itself as “no. 170, state ii/iii”, a three-state reading of the four-state catalogue.

So “of IV” is not invented. It is also not a fact about the plate. It is a fact about which book you are holding.

“State III”

Here it gets harder, and more interesting.

We went looking for a description of state III. Not an impression catalogued as state III — a description: what changes, what to look for, what distinguishes it from the state on either side.

State I is documented. State II is documented, and vividly: the British Museum's cataloguer writes “second state with jaw delineated” — the chin below the left ear becomes clearly defined. State IV is documented, and its description is the most consequential sentence in this whole enquiry: “fourth state with signature redrawn; from an edition of the Recueil published by Henri Louis Basan around 1810.”

State III: nothing. No collection record we could reach describes it. No impression catalogued as 170.III surfaced in any of the collections searched. The best that can be said is bounded by its neighbours — state II is defined as being before the cross-hatching in the beret above the plume, so state III is presumably where that cross-hatching appears. That is an inference from an adjacent definition, and we have published it as an inference.

What follows from state IV

Read that fourth-state description again. The signature was redrawn. Somebody re-cut Rembrandt's signature into the plate, and they did it for a commercial edition published in Paris around 1810 — a hundred and forty years after his death.

A signature gets redrawn when it has worn away. So the wear happened before the rework, which places it in the working life of the third state. An impression showing a faint or effectively absent signature would be entirely consistent with a late third state.

Which is to say: the two halves of that six-word description may be consistent with each other after all, and what they are jointly consistent with is a sheet printed late in the plate's life, well after Rembrandt.

We want to be precise about the status of that paragraph. It is a hypothesis. It is not on the object record, and it will not go there. It is here because the reasoning should be visible and available to be knocked down.

The competing explanations, which have to be excluded first

A signature can be missing for reasons that have nothing to do with state. The sheet is described as trimmed — trimmed sheets are common for this print, and one institutional impression is recorded as cut within the platemark. The British Museum holds three sheets it catalogues as copies of this composition, not impressions of it. And nineteenth-century photomechanical reproductions of Rembrandt etchings exist and are convincing at a glance.

None of these can be ruled out from paperwork. All of them can be ruled out from a photograph.

Where this leaves the record

The object page carries the state claim at possible, attributed to the source that made it, with a note saying it has not been checked. It carries the signature question as an open contradiction rather than resolving it in the object's favour. And the printing date is recorded as unknown, which is what it is.

One colour-managed photograph of the recto resolves the signature question in an afternoon. That photograph waits on a conservator, and the conservator is the right thing to wait on.